5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Modify behavior of ALPN to invoke callback after SNI/servername
8 callback, such that updates to the SSL_CTX affect ALPN.
11 *) Add SSL_CIPHER queries for authentication and key-exchange.
13 *) Modify behavior of ALPN to invoke callback after SNI/servername
14 callback, such that updates to the SSL_CTX affect ALPN.
17 *) Changes to the DEFAULT cipherlist:
18 - Prefer (EC)DHE handshakes over plain RSA.
19 - Prefer AEAD ciphers over legacy ciphers.
20 - Prefer ECDSA over RSA when both certificates are available.
21 - Prefer TLSv1.2 ciphers/PRF.
22 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
26 *) Change the ECC default curve list to be this, in order: x25519,
27 secp256r1, secp521r1, secp384r1.
30 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
31 disabled by default. They can be re-enabled using the
32 enable-weak-ssl-ciphers option to Configure.
35 *) If the server has ALPN configured, but supports no protocols that the
36 client advertises, send a fatal "no_application_protocol" alert.
37 This behaviour is SHALL in RFC 7301, though it isn't universally
38 implemented by other servers.
41 *) Add X25519 support.
42 Integrate support for X25519 into EC library. This includes support
43 for public and private key encoding using the format documented in
44 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
45 OID from that draft, encodes public keys using little endian
46 format in the ECPoint structure and private keys using
47 little endian form in the privateKey field of the ECPrivateKey
48 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
51 Note: the current version supports key generation, public and
52 private key encoding and ECDH key agreement using the EC API.
53 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
57 *) Deprecate SRP_VBASE_get_by_user.
58 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
59 In order to fix an unavoidable memory leak (CVE-2016-0798),
60 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
61 seed, even if the seed is configured.
63 Users should use SRP_VBASE_get1_by_user instead. Note that in
64 SRP_VBASE_get1_by_user, caller must free the returned value. Note
65 also that even though configuring the SRP seed attempts to hide
66 invalid usernames by continuing the handshake with fake
67 credentials, this behaviour is not constant time and no strong
68 guarantees are made that the handshake is indistinguishable from
72 *) Configuration change; it's now possible to build dynamic engines
73 without having to build shared libraries and vice versa. This
74 only applies to the engines in engines/, those in crypto/engine/
75 will always be built into libcrypto (i.e. "static").
77 Building dynamic engines is enabled by default; to disable, use
78 the configuration option "disable-dynamic-engine".
80 The only requirements for building dynamic engines are the
81 presence of the DSO module and building with position independent
82 code, so they will also automatically be disabled if configuring
83 with "disable-dso" or "disable-pic".
85 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
86 are also taken away from openssl/opensslconf.h, as they are
90 *) Configuration change; if there is a known flag to compile
91 position independent code, it will always be applied on the
92 libcrypto and libssl object files, and never on the application
93 object files. This means other libraries that use routines from
94 libcrypto / libssl can be made into shared libraries regardless
95 of how OpenSSL was configured.
97 If this isn't desirable, the configuration options "disable-pic"
98 or "no-pic" can be used to disable the use of PIC. This will
99 also disable building shared libraries and dynamic engines.
102 *) Removed JPAKE code. It was experimental and has no wide use.
105 *) The INSTALL_PREFIX Makefile variable has been renamed to
106 DESTDIR. That makes for less confusion on what this variable
107 is for. Also, the configuration option --install_prefix is
111 *) Heartbeat for TLS has been removed and is disabled by default
112 for DTLS; configure with enable-heartbeats. Code that uses the
113 old #define's might need to be updated.
114 [Emilia Käsper, Rich Salz]
116 *) Rename REF_CHECK to REF_DEBUG.
119 *) New "unified" build system
121 The "unified" build system is aimed to be a common system for all
122 platforms we support. With it comes new support for VMS.
124 This system builds supports building in a different directory tree
125 than the source tree. It produces one Makefile (for unix family
126 or lookalikes), or one descrip.mms (for VMS).
128 The source of information to make the Makefile / descrip.mms is
129 small files called 'build.info', holding the necessary
130 information for each directory with source to compile, and a
131 template in Configurations, like unix-Makefile.tmpl or
134 We rely heavily on the perl module Text::Template.
137 *) Added support for auto-initialisation and de-initialisation of the library.
138 OpenSSL no longer requires explicit init or deinit routines to be called,
139 except in certain circumstances. See the OPENSSL_init_crypto() and
140 OPENSSL_init_ssl() man pages for further information.
143 *) The arguments to the DTLSv1_listen function have changed. Specifically the
144 "peer" argument is now expected to be a BIO_ADDR object.
146 *) Rewrite of BIO networking library. The BIO library lacked consistent
147 support of IPv6, and adding it required some more extensive
148 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
149 which hold all types of addresses and chains of address information.
150 It also introduces a new API, with functions like BIO_socket,
151 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
152 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
153 have been adapted accordingly.
156 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
160 *) CRIME protection: disable compression by default, even if OpenSSL is
161 compiled with zlib enabled. Applications can still enable compression
162 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
163 using the SSL_CONF library to configure compression.
166 *) The signature of the session callback configured with
167 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
168 was explicitly marked as 'const unsigned char*' instead of
172 *) Always DPURIFY. Remove the use of uninitialized memory in the
173 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
176 *) Removed many obsolete configuration items, including
177 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
178 MD2_CHAR, MD2_INT, MD2_LONG
180 IDEA_SHORT, IDEA_LONG
181 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
182 [Rich Salz, with advice from Andy Polyakov]
184 *) Many BN internals have been moved to an internal header file.
185 [Rich Salz with help from Andy Polyakov]
187 *) Configuration and writing out the results from it has changed.
188 Files such as Makefile include/openssl/opensslconf.h and are now
189 produced through general templates, such as Makefile.in and
190 crypto/opensslconf.h.in and some help from the perl module
193 Also, the center of configuration information is no longer
194 Makefile. Instead, Configure produces a perl module in
195 configdata.pm which holds most of the config data (in the hash
196 table %config), the target data that comes from the target
197 configuration in one of the Configurations/*.conf files (in
201 *) To clarify their intended purposes, the Configure options
202 --prefix and --openssldir change their semantics, and become more
203 straightforward and less interdependent.
205 --prefix shall be used exclusively to give the location INSTALLTOP
206 where programs, scripts, libraries, include files and manuals are
207 going to be installed. The default is now /usr/local.
209 --openssldir shall be used exclusively to give the default
210 location OPENSSLDIR where certificates, private keys, CRLs are
211 managed. This is also where the default openssl.cnf gets
213 If the directory given with this option is a relative path, the
214 values of both the --prefix value and the --openssldir value will
215 be combined to become OPENSSLDIR.
216 The default for --openssldir is INSTALLTOP/ssl.
218 Anyone who uses --openssldir to specify where OpenSSL is to be
219 installed MUST change to use --prefix instead.
222 *) The GOST engine was out of date and therefore it has been removed. An up
223 to date GOST engine is now being maintained in an external repository.
224 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
225 support for GOST ciphersuites (these are only activated if a GOST engine
229 *) EGD is no longer supported by default; use enable-egd when
231 [Ben Kaduk and Rich Salz]
233 *) The distribution now has Makefile.in files, which are used to
234 create Makefile's when Configure is run. *Configure must be run
235 before trying to build now.*
238 *) The return value for SSL_CIPHER_description() for error conditions
242 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
244 Obtaining and performing DNSSEC validation of TLSA records is
245 the application's responsibility. The application provides
246 the TLSA records of its choice to OpenSSL, and these are then
247 used to authenticate the peer.
249 The TLSA records need not even come from DNS. They can, for
250 example, be used to implement local end-entity certificate or
251 trust-anchor "pinning", where the "pin" data takes the form
252 of TLSA records, which can augment or replace verification
253 based on the usual WebPKI public certification authorities.
256 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
257 continues to support deprecated interfaces in default builds.
258 However, applications are strongly advised to compile their
259 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
260 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
261 or the 1.1.0 releases.
263 In environments in which all applications have been ported to
264 not use any deprecated interfaces OpenSSL's Configure script
265 should be used with the --api=1.1.0 option to entirely remove
266 support for the deprecated features from the library and
267 unconditionally disable them in the installed headers.
268 Essentially the same effect can be achieved with the "no-deprecated"
269 argument to Configure, except that this will always restrict
270 the build to just the latest API, rather than a fixed API
273 As applications are ported to future revisions of the API,
274 they should update their compile-time OPENSSL_API_COMPAT define
275 accordingly, but in most cases should be able to continue to
276 compile with later releases.
278 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
279 0x10000000L and 0x00908000L, respectively. However those
280 versions did not support the OPENSSL_API_COMPAT feature, and
281 so applications are not typically tested for explicit support
282 of just the undeprecated features of either release.
285 *) Add support for setting the minimum and maximum supported protocol.
286 It can bet set via the SSL_set_min_proto_version() and
287 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
288 MaxProtcol. It's recommended to use the new APIs to disable
289 protocols instead of disabling individual protocols using
290 SSL_set_options() or SSL_CONF's Protocol. This change also
291 removes support for disabling TLS 1.2 in the OpenSSL TLS
292 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
295 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
298 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
299 and integrates ECDSA and ECDH functionality into EC. Implementations can
300 now redirect key generation and no longer need to convert to or from
303 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
304 include the ec.h header file instead.
307 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
308 ciphers who are no longer supported and drops support the ephemeral RSA key
309 exchange. The LOW ciphers currently doesn't have any ciphers in it.
312 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
313 opaque. For HMAC_CTX, the following constructors and destructors
316 HMAC_CTX *HMAC_CTX_new(void);
317 void HMAC_CTX_free(HMAC_CTX *ctx);
319 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
320 destroy such methods has been added. See EVP_MD_meth_new(3) and
321 EVP_CIPHER_meth_new(3) for documentation.
324 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
325 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
326 EVP_MD_CTX_reset() should be called instead to reinitialise
327 an already created structure.
328 2) For consistency with the majority of our object creators and
329 destructors, EVP_MD_CTX_(create|destroy) were renamed to
330 EVP_MD_CTX_(new|free). The old names are retained as macros
331 for deprecated builds.
334 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
335 cryptographic operations to be performed asynchronously as long as an
336 asynchronous capable engine is used. See the ASYNC_start_job() man page for
337 further details. Libssl has also had this capability integrated with the
338 introduction of the new mode SSL_MODE_ASYNC and associated error
339 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
340 pages. This work was developed in partnership with Intel Corp.
343 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
344 always enabled now. If you want to disable the support you should
345 exclude it using the list of supported ciphers. This also means that the
346 "-no_ecdhe" option has been removed from s_server.
349 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
350 SSL_{CTX_}set1_curves() which can set a list.
353 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
354 curve you want to support using SSL_{CTX_}set1_curves().
357 *) State machine rewrite. The state machine code has been significantly
358 refactored in order to remove much duplication of code and solve issues
359 with the old code (see ssl/statem/README for further details). This change
360 does have some associated API changes. Notably the SSL_state() function
361 has been removed and replaced by SSL_get_state which now returns an
362 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
363 altogether. The previous handshake states defined in ssl.h and ssl3.h have
367 *) All instances of the string "ssleay" in the public API were replaced
368 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
369 Some error codes related to internal RSA_eay API's were renamed.
372 *) The demo files in crypto/threads were moved to demo/threads.
375 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
377 [Matt Caswell, Rich Salz]
379 *) New ASN.1 embed macro.
381 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
382 structure is not allocated: it is part of the parent. That is instead of
390 This reduces memory fragmentation and make it impossible to accidentally
391 set a mandatory field to NULL.
393 This currently only works for some fields specifically a SEQUENCE, CHOICE,
394 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
395 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
399 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
402 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
403 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
404 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
405 DES and RC4 ciphersuites.
408 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
409 This changes the decoding behaviour for some invalid messages,
410 though the change is mostly in the more lenient direction, and
411 legacy behaviour is preserved as much as possible.
414 *) Fix no-stdio build.
415 [ David Woodhouse <David.Woodhouse@intel.com> and also
416 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
418 *) New testing framework
419 The testing framework has been largely rewritten and is now using
420 perl and the perl modules Test::Harness and an extended variant of
421 Test::More called OpenSSL::Test to do its work. All test scripts in
422 test/ have been rewritten into test recipes, and all direct calls to
423 executables in test/Makefile have become individual recipes using the
424 simplified testing OpenSSL::Test::Simple.
426 For documentation on our testing modules, do:
428 perldoc test/testlib/OpenSSL/Test/Simple.pm
429 perldoc test/testlib/OpenSSL/Test.pm
433 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
434 are used; the latter aborts on memory leaks (usually checked on exit).
435 Some undocumented "set malloc, etc., hooks" functions were removed
436 and others were changed. All are now documented.
439 *) In DSA_generate_parameters_ex, if the provided seed is too short,
441 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
443 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
444 from RFC4279, RFC4785, RFC5487, RFC5489.
446 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
447 original RSA_PSK patch.
450 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
451 era flag was never set throughout the codebase (only read). Also removed
452 SSL3_FLAGS_POP_BUFFER which was only used if
453 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
456 *) Changed the default name options in the "ca", "crl", "req" and "x509"
457 to be "oneline" instead of "compat".
460 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
461 not aware of clients that still exhibit this bug, and the workaround
462 hasn't been working properly for a while.
465 *) The return type of BIO_number_read() and BIO_number_written() as well as
466 the corresponding num_read and num_write members in the BIO structure has
467 changed from unsigned long to uint64_t. On platforms where an unsigned
468 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
472 *) Given the pervasive nature of TLS extensions it is inadvisable to run
473 OpenSSL without support for them. It also means that maintaining
474 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
475 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
478 *) Removed support for the two export grade static DH ciphersuites
479 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
480 were newly added (along with a number of other static DH ciphersuites) to
481 1.0.2. However the two export ones have *never* worked since they were
482 introduced. It seems strange in any case to be adding new export
483 ciphersuites, and given "logjam" it also does not seem correct to fix them.
486 *) Version negotiation has been rewritten. In particular SSLv23_method(),
487 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
488 and turned into macros which simply call the new preferred function names
489 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
490 should use the new names instead. Also as part of this change the ssl23.h
491 header file has been removed.
494 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
495 code and the associated standard is no longer considered fit-for-purpose.
498 *) RT2547 was closed. When generating a private key, try to make the
499 output file readable only by the owner. This behavior change might
500 be noticeable when interacting with other software.
502 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
506 *) Added HTTP GET support to the ocsp command.
509 *) Changed default digest for the dgst and enc commands from MD5 to
513 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
516 *) Added support for TLS extended master secret from
517 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
518 initial patch which was a great help during development.
521 *) All libssl internal structures have been removed from the public header
522 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
523 now redundant). Users should not attempt to access internal structures
524 directly. Instead they should use the provided API functions.
527 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
528 Access to deprecated functions can be re-enabled by running config with
529 "enable-deprecated". In addition applications wishing to use deprecated
530 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
531 will, by default, disable some transitive includes that previously existed
532 in the header files (e.g. ec.h will no longer, by default, include bn.h)
535 *) Added support for OCB mode. OpenSSL has been granted a patent license
536 compatible with the OpenSSL license for use of OCB. Details are available
537 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
538 for OCB can be removed by calling config with no-ocb.
541 *) SSLv2 support has been removed. It still supports receiving a SSLv2
542 compatible client hello.
545 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
546 done while fixing the error code for the key-too-small case.
547 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
549 *) CA.sh has been removmed; use CA.pl instead.
552 *) Removed old DES API.
555 *) Remove various unsupported platforms:
561 Sinix/ReliantUNIX RM400
566 16-bit platforms such as WIN16
569 *) Clean up OPENSSL_NO_xxx #define's
570 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
571 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
572 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
573 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
574 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
575 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
576 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
577 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
578 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
579 Remove MS_STATIC; it's a relic from platforms <32 bits.
582 *) Cleaned up dead code
583 Remove all but one '#ifdef undef' which is to be looked at.
586 *) Clean up calling of xxx_free routines.
587 Just like free(), fix most of the xxx_free routines to accept
588 NULL. Remove the non-null checks from callers. Save much code.
591 *) Add secure heap for storage of private keys (when possible).
592 Add BIO_s_secmem(), CBIGNUM, etc.
593 Contributed by Akamai Technologies under our Corporate CLA.
596 *) Experimental support for a new, fast, unbiased prime candidate generator,
597 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
598 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
600 *) New output format NSS in the sess_id command line tool. This allows
601 exporting the session id and the master key in NSS keylog format.
602 [Martin Kaiser <martin@kaiser.cx>]
604 *) Harmonize version and its documentation. -f flag is used to display
606 [mancha <mancha1@zoho.com>]
608 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
609 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
610 [mancha <mancha1@zoho.com>]
612 *) Fix some double frees. These are not thought to be exploitable.
613 [mancha <mancha1@zoho.com>]
615 *) A missing bounds check in the handling of the TLS heartbeat extension
616 can be used to reveal up to 64k of memory to a connected client or
619 Thanks for Neel Mehta of Google Security for discovering this bug and to
620 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
621 preparing the fix (CVE-2014-0160)
622 [Adam Langley, Bodo Moeller]
624 *) Fix for the attack described in the paper "Recovering OpenSSL
625 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
626 by Yuval Yarom and Naomi Benger. Details can be obtained from:
627 http://eprint.iacr.org/2014/140
629 Thanks to Yuval Yarom and Naomi Benger for discovering this
630 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
631 [Yuval Yarom and Naomi Benger]
633 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
634 this fixes a limitation in previous versions of OpenSSL.
637 *) Experimental encrypt-then-mac support.
639 Experimental support for encrypt then mac from
640 draft-gutmann-tls-encrypt-then-mac-02.txt
642 To enable it set the appropriate extension number (0x42 for the test
643 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
645 For non-compliant peers (i.e. just about everything) this should have no
648 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
652 *) Add EVP support for key wrapping algorithms, to avoid problems with
653 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
654 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
655 algorithms and include tests cases.
658 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
662 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
663 MGF1 digest and OAEP label.
666 *) Make openssl verify return errors.
667 [Chris Palmer <palmer@google.com> and Ben Laurie]
669 *) New function ASN1_TIME_diff to calculate the difference between two
670 ASN1_TIME structures or one structure and the current time.
673 *) Update fips_test_suite to support multiple command line options. New
674 test to induce all self test errors in sequence and check expected
678 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
679 sign or verify all in one operation.
682 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
683 test programs and fips_test_suite. Includes functionality to parse
684 the minimal script output of fipsalgest.pl directly.
687 *) Add authorisation parameter to FIPS_module_mode_set().
690 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
693 *) Use separate DRBG fields for internal and external flags. New function
694 FIPS_drbg_health_check() to perform on demand health checking. Add
695 generation tests to fips_test_suite with reduced health check interval to
696 demonstrate periodic health checking. Add "nodh" option to
697 fips_test_suite to skip very slow DH test.
700 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
704 *) More extensive health check for DRBG checking many more failure modes.
705 New function FIPS_selftest_drbg_all() to handle every possible DRBG
706 combination: call this in fips_test_suite.
709 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
710 and POST to handle Dual EC cases.
713 *) Add support for canonical generation of DSA parameter 'g'. See
716 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
717 POST to handle HMAC cases.
720 *) Add functions FIPS_module_version() and FIPS_module_version_text()
721 to return numerical and string versions of the FIPS module number.
724 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
725 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
726 outside the validated module in the FIPS capable OpenSSL.
729 *) Minor change to DRBG entropy callback semantics. In some cases
730 there is no multiple of the block length between min_len and
731 max_len. Allow the callback to return more than max_len bytes
732 of entropy but discard any extra: it is the callback's responsibility
733 to ensure that the extra data discarded does not impact the
734 requested amount of entropy.
737 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
738 information in FIPS186-3, SP800-57 and SP800-131A.
741 *) CCM support via EVP. Interface is very similar to GCM case except we
742 must supply all data in one chunk (i.e. no update, final) and the
743 message length must be supplied if AAD is used. Add algorithm test
747 *) Initial version of POST overhaul. Add POST callback to allow the status
748 of POST to be monitored and/or failures induced. Modify fips_test_suite
749 to use callback. Always run all selftests even if one fails.
752 *) XTS support including algorithm test driver in the fips_gcmtest program.
753 Note: this does increase the maximum key length from 32 to 64 bytes but
754 there should be no binary compatibility issues as existing applications
755 will never use XTS mode.
758 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
759 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
760 performs algorithm blocking for unapproved PRNG types. Also do not
761 set PRNG type in FIPS_mode_set(): leave this to the application.
762 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
763 the standard OpenSSL PRNG: set additional data to a date time vector.
766 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
767 This shouldn't present any incompatibility problems because applications
768 shouldn't be using these directly and any that are will need to rethink
769 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
772 *) Extensive self tests and health checking required by SP800-90 DRBG.
773 Remove strength parameter from FIPS_drbg_instantiate and always
774 instantiate at maximum supported strength.
777 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
780 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
783 *) New function DH_compute_key_padded() to compute a DH key and pad with
784 leading zeroes if needed: this complies with SP800-56A et al.
787 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
788 anything, incomplete, subject to change and largely untested at present.
791 *) Modify fipscanisteronly build option to only build the necessary object
792 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
795 *) Add experimental option FIPSSYMS to give all symbols in
796 fipscanister.o and FIPS or fips prefix. This will avoid
797 conflicts with future versions of OpenSSL. Add perl script
798 util/fipsas.pl to preprocess assembly language source files
799 and rename any affected symbols.
802 *) Add selftest checks and algorithm block of non-fips algorithms in
803 FIPS mode. Remove DES2 from selftests.
806 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
807 return internal method without any ENGINE dependencies. Add new
808 tiny fips sign and verify functions.
811 *) New build option no-ec2m to disable characteristic 2 code.
814 *) New build option "fipscanisteronly". This only builds fipscanister.o
815 and (currently) associated fips utilities. Uses the file Makefile.fips
816 instead of Makefile.org as the prototype.
819 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
820 Update fips_gcmtest to use IV generator.
823 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
824 setting output buffer to NULL. The *Final function must be
825 called although it will not retrieve any additional data. The tag
826 can be set or retrieved with a ctrl. The IV length is by default 12
827 bytes (96 bits) but can be set to an alternative value. If the IV
828 length exceeds the maximum IV length (currently 16 bytes) it cannot be
832 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
833 underlying do_cipher function handles all cipher semantics itself
834 including padding and finalisation. This is useful if (for example)
835 an ENGINE cipher handles block padding itself. The behaviour of
836 do_cipher is subtly changed if this flag is set: the return value
837 is the number of characters written to the output buffer (zero is
838 no longer an error code) or a negative error code. Also if the
839 input buffer is NULL and length 0 finalisation should be performed.
842 *) If a candidate issuer certificate is already part of the constructed
843 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
846 *) Improve forward-security support: add functions
848 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
849 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
851 for use by SSL/TLS servers; the callback function will be called whenever a
852 new session is created, and gets to decide whether the session may be
853 cached to make it resumable (return 0) or not (return 1). (As by the
854 SSL/TLS protocol specifications, the session_id sent by the server will be
855 empty to indicate that the session is not resumable; also, the server will
856 not generate RFC 4507 (RFC 5077) session tickets.)
858 A simple reasonable callback implementation is to return is_forward_secure.
859 This parameter will be set to 1 or 0 depending on the ciphersuite selected
860 by the SSL/TLS server library, indicating whether it can provide forward
862 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
864 *) New -verify_name option in command line utilities to set verification
868 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
869 Add CMAC pkey methods.
872 *) Experimental renegotiation in s_server -www mode. If the client
873 browses /reneg connection is renegotiated. If /renegcert it is
874 renegotiated requesting a certificate.
877 *) Add an "external" session cache for debugging purposes to s_server. This
878 should help trace issues which normally are only apparent in deployed
879 multi-process servers.
882 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
883 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
884 BIO_set_cipher() and some obscure PEM functions were changed so they
885 can now return an error. The RAND changes required a change to the
886 RAND_METHOD structure.
889 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
890 a gcc attribute to warn if the result of a function is ignored. This
891 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
892 whose return value is often ignored.
895 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
896 These allow SCTs (signed certificate timestamps) to be requested and
897 validated when establishing a connection.
898 [Rob Percival <robpercival@google.com>]
900 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
902 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
903 Builds that are not configured with "enable-weak-ssl-ciphers" will not
904 provide any "EXPORT" or "LOW" strength ciphers.
907 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
908 is by default disabled at build-time. Builds that are not configured with
909 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
910 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
911 will need to explicitly call either of:
913 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
915 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
917 as appropriate. Even if either of those is used, or the application
918 explicitly uses the version-specific SSLv2_method() or its client and
919 server variants, SSLv2 ciphers vulnerable to exhaustive search key
920 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
921 ciphers, and SSLv2 56-bit DES are no longer available.
925 *) Fix a double-free in DSA code
927 A double free bug was discovered when OpenSSL parses malformed DSA private
928 keys and could lead to a DoS attack or memory corruption for applications
929 that receive DSA private keys from untrusted sources. This scenario is
932 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
937 *) Disable SRP fake user seed to address a server memory leak.
939 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
941 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
942 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
943 was changed to ignore the "fake user" SRP seed, even if the seed
946 Users should use SRP_VBASE_get1_by_user instead. Note that in
947 SRP_VBASE_get1_by_user, caller must free the returned value. Note
948 also that even though configuring the SRP seed attempts to hide
949 invalid usernames by continuing the handshake with fake
950 credentials, this behaviour is not constant time and no strong
951 guarantees are made that the handshake is indistinguishable from
952 that of a valid user.
956 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
958 In the BN_hex2bn function the number of hex digits is calculated using an
959 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
960 large values of |i| this can result in |bn_expand| not allocating any
961 memory because |i * 4| is negative. This can leave the internal BIGNUM data
962 field as NULL leading to a subsequent NULL ptr deref. For very large values
963 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
964 In this case memory is allocated to the internal BIGNUM data field, but it
965 is insufficiently sized leading to heap corruption. A similar issue exists
966 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
967 is ever called by user applications with very large untrusted hex/dec data.
968 This is anticipated to be a rare occurrence.
970 All OpenSSL internal usage of these functions use data that is not expected
971 to be untrusted, e.g. config file data or application command line
972 arguments. If user developed applications generate config file data based
973 on untrusted data then it is possible that this could also lead to security
974 consequences. This is also anticipated to be rare.
976 This issue was reported to OpenSSL by Guido Vranken.
980 *) Fix memory issues in BIO_*printf functions
982 The internal |fmtstr| function used in processing a "%s" format string in
983 the BIO_*printf functions could overflow while calculating the length of a
984 string and cause an OOB read when printing very long strings.
986 Additionally the internal |doapr_outch| function can attempt to write to an
987 OOB memory location (at an offset from the NULL pointer) in the event of a
988 memory allocation failure. In 1.0.2 and below this could be caused where
989 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
990 could be in processing a very long "%s" format string. Memory leaks can
993 The first issue may mask the second issue dependent on compiler behaviour.
994 These problems could enable attacks where large amounts of untrusted data
995 is passed to the BIO_*printf functions. If applications use these functions
996 in this way then they could be vulnerable. OpenSSL itself uses these
997 functions when printing out human-readable dumps of ASN.1 data. Therefore
998 applications that print this data could be vulnerable if the data is from
999 untrusted sources. OpenSSL command line applications could also be
1000 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1001 as command line arguments.
1003 Libssl is not considered directly vulnerable. Additionally certificates etc
1004 received via remote connections via libssl are also unlikely to be able to
1005 trigger these issues because of message size limits enforced within libssl.
1007 This issue was reported to OpenSSL Guido Vranken.
1011 *) Side channel attack on modular exponentiation
1013 A side-channel attack was found which makes use of cache-bank conflicts on
1014 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1015 of RSA keys. The ability to exploit this issue is limited as it relies on
1016 an attacker who has control of code in a thread running on the same
1017 hyper-threaded core as the victim thread which is performing decryptions.
1019 This issue was reported to OpenSSL by Yuval Yarom, The University of
1020 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1021 Nadia Heninger, University of Pennsylvania with more information at
1022 http://cachebleed.info.
1026 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1027 if no keysize is specified with default_bits. This fixes an
1028 omission in an earlier change that changed all RSA/DSA key generation
1029 apps to use 2048 bits by default.
1032 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1033 *) DH small subgroups
1035 Historically OpenSSL only ever generated DH parameters based on "safe"
1036 primes. More recently (in version 1.0.2) support was provided for
1037 generating X9.42 style parameter files such as those required for RFC 5114
1038 support. The primes used in such files may not be "safe". Where an
1039 application is using DH configured with parameters based on primes that are
1040 not "safe" then an attacker could use this fact to find a peer's private
1041 DH exponent. This attack requires that the attacker complete multiple
1042 handshakes in which the peer uses the same private DH exponent. For example
1043 this could be used to discover a TLS server's private DH exponent if it's
1044 reusing the private DH exponent or it's using a static DH ciphersuite.
1046 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1047 TLS. It is not on by default. If the option is not set then the server
1048 reuses the same private DH exponent for the life of the server process and
1049 would be vulnerable to this attack. It is believed that many popular
1050 applications do set this option and would therefore not be at risk.
1052 The fix for this issue adds an additional check where a "q" parameter is
1053 available (as is the case in X9.42 based parameters). This detects the
1054 only known attack, and is the only possible defense for static DH
1055 ciphersuites. This could have some performance impact.
1057 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1058 default and cannot be disabled. This could have some performance impact.
1060 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1064 *) SSLv2 doesn't block disabled ciphers
1066 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1067 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1068 been disabled, provided that the SSLv2 protocol was not also disabled via
1071 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1072 and Sebastian Schinzel.
1076 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1078 *) BN_mod_exp may produce incorrect results on x86_64
1080 There is a carry propagating bug in the x86_64 Montgomery squaring
1081 procedure. No EC algorithms are affected. Analysis suggests that attacks
1082 against RSA and DSA as a result of this defect would be very difficult to
1083 perform and are not believed likely. Attacks against DH are considered just
1084 feasible (although very difficult) because most of the work necessary to
1085 deduce information about a private key may be performed offline. The amount
1086 of resources required for such an attack would be very significant and
1087 likely only accessible to a limited number of attackers. An attacker would
1088 additionally need online access to an unpatched system using the target
1089 private key in a scenario with persistent DH parameters and a private
1090 key that is shared between multiple clients. For example this can occur by
1091 default in OpenSSL DHE based SSL/TLS ciphersuites.
1093 This issue was reported to OpenSSL by Hanno Böck.
1097 *) Certificate verify crash with missing PSS parameter
1099 The signature verification routines will crash with a NULL pointer
1100 dereference if presented with an ASN.1 signature using the RSA PSS
1101 algorithm and absent mask generation function parameter. Since these
1102 routines are used to verify certificate signature algorithms this can be
1103 used to crash any certificate verification operation and exploited in a
1104 DoS attack. Any application which performs certificate verification is
1105 vulnerable including OpenSSL clients and servers which enable client
1108 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1112 *) X509_ATTRIBUTE memory leak
1114 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1115 memory. This structure is used by the PKCS#7 and CMS routines so any
1116 application which reads PKCS#7 or CMS data from untrusted sources is
1117 affected. SSL/TLS is not affected.
1119 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1124 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1125 This changes the decoding behaviour for some invalid messages,
1126 though the change is mostly in the more lenient direction, and
1127 legacy behaviour is preserved as much as possible.
1130 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1132 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1134 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1136 *) Alternate chains certificate forgery
1138 During certificate verfification, OpenSSL will attempt to find an
1139 alternative certificate chain if the first attempt to build such a chain
1140 fails. An error in the implementation of this logic can mean that an
1141 attacker could cause certain checks on untrusted certificates to be
1142 bypassed, such as the CA flag, enabling them to use a valid leaf
1143 certificate to act as a CA and "issue" an invalid certificate.
1145 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1149 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1151 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1152 incompatibility in the handling of HMAC. The previous ABI has now been
1156 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1158 *) Malformed ECParameters causes infinite loop
1160 When processing an ECParameters structure OpenSSL enters an infinite loop
1161 if the curve specified is over a specially malformed binary polynomial
1164 This can be used to perform denial of service against any
1165 system which processes public keys, certificate requests or
1166 certificates. This includes TLS clients and TLS servers with
1167 client authentication enabled.
1169 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1173 *) Exploitable out-of-bounds read in X509_cmp_time
1175 X509_cmp_time does not properly check the length of the ASN1_TIME
1176 string and can read a few bytes out of bounds. In addition,
1177 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1180 An attacker can use this to craft malformed certificates and CRLs of
1181 various sizes and potentially cause a segmentation fault, resulting in
1182 a DoS on applications that verify certificates or CRLs. TLS clients
1183 that verify CRLs are affected. TLS clients and servers with client
1184 authentication enabled may be affected if they use custom verification
1187 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1188 independently by Hanno Böck.
1192 *) PKCS7 crash with missing EnvelopedContent
1194 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1195 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1196 with missing content and trigger a NULL pointer dereference on parsing.
1198 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1199 structures from untrusted sources are affected. OpenSSL clients and
1200 servers are not affected.
1202 This issue was reported to OpenSSL by Michal Zalewski (Google).
1206 *) CMS verify infinite loop with unknown hash function
1208 When verifying a signedData message the CMS code can enter an infinite loop
1209 if presented with an unknown hash function OID. This can be used to perform
1210 denial of service against any system which verifies signedData messages using
1212 This issue was reported to OpenSSL by Johannes Bauer.
1216 *) Race condition handling NewSessionTicket
1218 If a NewSessionTicket is received by a multi-threaded client when attempting to
1219 reuse a previous ticket then a race condition can occur potentially leading to
1220 a double free of the ticket data.
1224 *) Only support 256-bit or stronger elliptic curves with the
1225 'ecdh_auto' setting (server) or by default (client). Of supported
1226 curves, prefer P-256 (both).
1229 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1231 *) ClientHello sigalgs DoS fix
1233 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1234 invalid signature algorithms extension a NULL pointer dereference will
1235 occur. This can be exploited in a DoS attack against the server.
1237 This issue was was reported to OpenSSL by David Ramos of Stanford
1240 [Stephen Henson and Matt Caswell]
1242 *) Multiblock corrupted pointer fix
1244 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1245 feature only applies on 64 bit x86 architecture platforms that support AES
1246 NI instructions. A defect in the implementation of "multiblock" can cause
1247 OpenSSL's internal write buffer to become incorrectly set to NULL when
1248 using non-blocking IO. Typically, when the user application is using a
1249 socket BIO for writing, this will only result in a failed connection.
1250 However if some other BIO is used then it is likely that a segmentation
1251 fault will be triggered, thus enabling a potential DoS attack.
1253 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1257 *) Segmentation fault in DTLSv1_listen fix
1259 The DTLSv1_listen function is intended to be stateless and processes the
1260 initial ClientHello from many peers. It is common for user code to loop
1261 over the call to DTLSv1_listen until a valid ClientHello is received with
1262 an associated cookie. A defect in the implementation of DTLSv1_listen means
1263 that state is preserved in the SSL object from one invocation to the next
1264 that can lead to a segmentation fault. Errors processing the initial
1265 ClientHello can trigger this scenario. An example of such an error could be
1266 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1269 This issue was reported to OpenSSL by Per Allansson.
1273 *) Segmentation fault in ASN1_TYPE_cmp fix
1275 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1276 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1277 certificate signature algorithm consistency this can be used to crash any
1278 certificate verification operation and exploited in a DoS attack. Any
1279 application which performs certificate verification is vulnerable including
1280 OpenSSL clients and servers which enable client authentication.
1284 *) Segmentation fault for invalid PSS parameters fix
1286 The signature verification routines will crash with a NULL pointer
1287 dereference if presented with an ASN.1 signature using the RSA PSS
1288 algorithm and invalid parameters. Since these routines are used to verify
1289 certificate signature algorithms this can be used to crash any
1290 certificate verification operation and exploited in a DoS attack. Any
1291 application which performs certificate verification is vulnerable including
1292 OpenSSL clients and servers which enable client authentication.
1294 This issue was was reported to OpenSSL by Brian Carpenter.
1298 *) ASN.1 structure reuse memory corruption fix
1300 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1301 memory corruption via an invalid write. Such reuse is and has been
1302 strongly discouraged and is believed to be rare.
1304 Applications that parse structures containing CHOICE or ANY DEFINED BY
1305 components may be affected. Certificate parsing (d2i_X509 and related
1306 functions) are however not affected. OpenSSL clients and servers are
1311 *) PKCS7 NULL pointer dereferences fix
1313 The PKCS#7 parsing code does not handle missing outer ContentInfo
1314 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1315 missing content and trigger a NULL pointer dereference on parsing.
1317 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1318 otherwise parse PKCS#7 structures from untrusted sources are
1319 affected. OpenSSL clients and servers are not affected.
1321 This issue was reported to OpenSSL by Michal Zalewski (Google).
1325 *) DoS via reachable assert in SSLv2 servers fix
1327 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1328 servers that both support SSLv2 and enable export cipher suites by sending
1329 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1331 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1332 (OpenSSL development team).
1336 *) Empty CKE with client auth and DHE fix
1338 If client auth is used then a server can seg fault in the event of a DHE
1339 ciphersuite being selected and a zero length ClientKeyExchange message
1340 being sent by the client. This could be exploited in a DoS attack.
1344 *) Handshake with unseeded PRNG fix
1346 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1347 with an unseeded PRNG. The conditions are:
1348 - The client is on a platform where the PRNG has not been seeded
1349 automatically, and the user has not seeded manually
1350 - A protocol specific client method version has been used (i.e. not
1351 SSL_client_methodv23)
1352 - A ciphersuite is used that does not require additional random data from
1353 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1355 If the handshake succeeds then the client random that has been used will
1356 have been generated from a PRNG with insufficient entropy and therefore the
1357 output may be predictable.
1359 For example using the following command with an unseeded openssl will
1360 succeed on an unpatched platform:
1362 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1366 *) Use After Free following d2i_ECPrivatekey error fix
1368 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1369 could cause a use after free condition. This, in turn, could cause a double
1370 free in several private key parsing functions (such as d2i_PrivateKey
1371 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1372 for applications that receive EC private keys from untrusted
1373 sources. This scenario is considered rare.
1375 This issue was discovered by the BoringSSL project and fixed in their
1380 *) X509_to_X509_REQ NULL pointer deref fix
1382 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1383 the certificate key is invalid. This function is rarely used in practice.
1385 This issue was discovered by Brian Carpenter.
1389 *) Removed the export ciphers from the DEFAULT ciphers
1392 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1394 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1395 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1396 So far those who have to target multiple plaforms would compromise
1397 and argue that binary targeting say ARMv5 would still execute on
1398 ARMv8. "Universal" build resolves this compromise by providing
1399 near-optimal performance even on newer platforms.
1402 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1403 (other platforms pending).
1404 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1406 *) Add support for the SignedCertificateTimestampList certificate and
1407 OCSP response extensions from RFC6962.
1410 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1411 for corner cases. (Certain input points at infinity could lead to
1412 bogus results, with non-infinity inputs mapped to infinity too.)
1415 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1416 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1417 common cases are optimized and there still is room for further
1418 improvements. Vector Permutation AES for Altivec is also added.
1421 *) Add support for little-endian ppc64 Linux target.
1422 [Marcelo Cerri (IBM)]
1424 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1425 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1426 are optimized and there still is room for further improvements.
1427 Both 32- and 64-bit modes are supported.
1428 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1430 *) Improved ARMv7 NEON support.
1433 *) Support for SPARC Architecture 2011 crypto extensions, first
1434 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1435 SHA256/512, MD5, GHASH and modular exponentiation.
1436 [Andy Polyakov, David Miller]
1438 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1440 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1442 *) Support for new and upcoming Intel processors, including AVX2,
1443 BMI and SHA ISA extensions. This includes additional "stitched"
1444 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1447 This work was sponsored by Intel Corp.
1450 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1451 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1452 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1455 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1456 this fixes a limiation in previous versions of OpenSSL.
1459 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1460 MGF1 digest and OAEP label.
1463 *) Add EVP support for key wrapping algorithms, to avoid problems with
1464 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1465 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1466 algorithms and include tests cases.
1469 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1471 [Douglas E. Engert, Steve Henson]
1473 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1474 difference in days and seconds between two tm or ASN1_TIME structures.
1477 *) Add -rev test option to s_server to just reverse order of characters
1478 received by client and send back to server. Also prints an abbreviated
1479 summary of the connection parameters.
1482 *) New option -brief for s_client and s_server to print out a brief summary
1483 of connection parameters.
1486 *) Add callbacks for arbitrary TLS extensions.
1487 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1489 *) New option -crl_download in several openssl utilities to download CRLs
1490 from CRLDP extension in certificates.
1493 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1496 *) New function X509_CRL_diff to generate a delta CRL from the difference
1497 of two full CRLs. Add support to "crl" utility.
1500 *) New functions to set lookup_crls function and to retrieve
1501 X509_STORE from X509_STORE_CTX.
1504 *) Print out deprecated issuer and subject unique ID fields in
1508 *) Extend OCSP I/O functions so they can be used for simple general purpose
1509 HTTP as well as OCSP. New wrapper function which can be used to download
1510 CRLs using the OCSP API.
1513 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1516 *) SSL_CONF* functions. These provide a common framework for application
1517 configuration using configuration files or command lines.
1520 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1521 message callback and prints the results. Needs compile time option
1522 "enable-ssl-trace". New options to s_client and s_server to enable
1526 *) New ctrl and macro to retrieve supported points extensions.
1527 Print out extension in s_server and s_client.
1530 *) New functions to retrieve certificate signature and signature
1534 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1538 *) New Suite B modes for TLS code. These use and enforce the requirements
1539 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1540 only use Suite B curves. The Suite B modes can be set by using the
1541 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1544 *) New chain verification flags for Suite B levels of security. Check
1545 algorithms are acceptable when flags are set in X509_verify_cert.
1548 *) Make tls1_check_chain return a set of flags indicating checks passed
1549 by a certificate chain. Add additional tests to handle client
1550 certificates: checks for matching certificate type and issuer name
1554 *) If an attempt is made to use a signature algorithm not in the peer
1555 preference list abort the handshake. If client has no suitable
1556 signature algorithms in response to a certificate request do not
1557 use the certificate.
1560 *) If server EC tmp key is not in client preference list abort handshake.
1563 *) Add support for certificate stores in CERT structure. This makes it
1564 possible to have different stores per SSL structure or one store in
1565 the parent SSL_CTX. Include distint stores for certificate chain
1566 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1567 to build and store a certificate chain in CERT structure: returing
1568 an error if the chain cannot be built: this will allow applications
1569 to test if a chain is correctly configured.
1571 Note: if the CERT based stores are not set then the parent SSL_CTX
1572 store is used to retain compatibility with existing behaviour.
1576 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1577 mask based on the current session, check mask when sending client
1578 hello and checking the requested ciphersuite.
1581 *) New ctrls to retrieve and set certificate types in a certificate
1582 request message. Print out received values in s_client. If certificate
1583 types is not set with custom values set sensible values based on
1584 supported signature algorithms.
1587 *) Support for distinct client and server supported signature algorithms.
1590 *) Add certificate callback. If set this is called whenever a certificate
1591 is required by client or server. An application can decide which
1592 certificate chain to present based on arbitrary criteria: for example
1593 supported signature algorithms. Add very simple example to s_server.
1594 This fixes many of the problems and restrictions of the existing client
1595 certificate callback: for example you can now clear an existing
1596 certificate and specify the whole chain.
1599 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1600 the certificate can be used for (if anything). Set valid_flags field
1601 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1602 to have similar checks in it.
1604 Add new "cert_flags" field to CERT structure and include a "strict mode".
1605 This enforces some TLS certificate requirements (such as only permitting
1606 certificate signature algorithms contained in the supported algorithms
1607 extension) which some implementations ignore: this option should be used
1608 with caution as it could cause interoperability issues.
1611 *) Update and tidy signature algorithm extension processing. Work out
1612 shared signature algorithms based on preferences and peer algorithms
1613 and print them out in s_client and s_server. Abort handshake if no
1614 shared signature algorithms.
1617 *) Add new functions to allow customised supported signature algorithms
1618 for SSL and SSL_CTX structures. Add options to s_client and s_server
1622 *) New function SSL_certs_clear() to delete all references to certificates
1623 from an SSL structure. Before this once a certificate had been added
1624 it couldn't be removed.
1627 *) Integrate hostname, email address and IP address checking with certificate
1628 verification. New verify options supporting checking in opensl utility.
1631 *) Fixes and wildcard matching support to hostname and email checking
1632 functions. Add manual page.
1633 [Florian Weimer (Red Hat Product Security Team)]
1635 *) New functions to check a hostname email or IP address against a
1636 certificate. Add options x509 utility to print results of checks against
1640 *) Fix OCSP checking.
1641 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1643 *) Initial experimental support for explicitly trusted non-root CAs.
1644 OpenSSL still tries to build a complete chain to a root but if an
1645 intermediate CA has a trust setting included that is used. The first
1646 setting is used: whether to trust (e.g., -addtrust option to the x509
1650 *) Add -trusted_first option which attempts to find certificates in the
1651 trusted store even if an untrusted chain is also supplied.
1654 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1655 platform support for Linux and Android.
1658 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1661 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1662 When in FIPS mode the approved implementations are used as normal,
1663 when not in FIPS mode the internal unapproved versions are used instead.
1664 This means that the FIPS capable OpenSSL isn't forced to use the
1665 (often lower performance) FIPS implementations outside FIPS mode.
1668 *) Transparently support X9.42 DH parameters when calling
1669 PEM_read_bio_DHparameters. This means existing applications can handle
1670 the new parameter format automatically.
1673 *) Initial experimental support for X9.42 DH parameter format: mainly
1674 to support use of 'q' parameter for RFC5114 parameters.
1677 *) Add DH parameters from RFC5114 including test data to dhtest.
1680 *) Support for automatic EC temporary key parameter selection. If enabled
1681 the most preferred EC parameters are automatically used instead of
1682 hardcoded fixed parameters. Now a server just has to call:
1683 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1684 support ECDH and use the most appropriate parameters.
1687 *) Enhance and tidy EC curve and point format TLS extension code. Use
1688 static structures instead of allocation if default values are used.
1689 New ctrls to set curves we wish to support and to retrieve shared curves.
1690 Print out shared curves in s_server. New options to s_server and s_client
1691 to set list of supported curves.
1694 *) New ctrls to retrieve supported signature algorithms and
1695 supported curve values as an array of NIDs. Extend openssl utility
1696 to print out received values.
1699 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1700 between NIDs and the more common NIST names such as "P-256". Enhance
1701 ecparam utility and ECC method to recognise the NIST names for curves.
1704 *) Enhance SSL/TLS certificate chain handling to support different
1705 chains for each certificate instead of one chain in the parent SSL_CTX.
1708 *) Support for fixed DH ciphersuite client authentication: where both
1709 server and client use DH certificates with common parameters.
1712 *) Support for fixed DH ciphersuites: those requiring DH server
1716 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1718 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1719 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1720 X509_CINF_get_signature were reverted post internal team review.
1722 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1724 *) Build fixes for the Windows and OpenVMS platforms
1725 [Matt Caswell and Richard Levitte]
1727 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1729 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1730 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1731 dereference. This could lead to a Denial Of Service attack. Thanks to
1732 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1736 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1737 dtls1_buffer_record function under certain conditions. In particular this
1738 could occur if an attacker sent repeated DTLS records with the same
1739 sequence number but for the next epoch. The memory leak could be exploited
1740 by an attacker in a Denial of Service attack through memory exhaustion.
1741 Thanks to Chris Mueller for reporting this issue.
1745 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1746 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1747 method would be set to NULL which could later result in a NULL pointer
1748 dereference. Thanks to Frank Schmirler for reporting this issue.
1752 *) Abort handshake if server key exchange message is omitted for ephemeral
1755 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1756 reporting this issue.
1760 *) Remove non-export ephemeral RSA code on client and server. This code
1761 violated the TLS standard by allowing the use of temporary RSA keys in
1762 non-export ciphersuites and could be used by a server to effectively
1763 downgrade the RSA key length used to a value smaller than the server
1764 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1765 INRIA or reporting this issue.
1769 *) Fixed issue where DH client certificates are accepted without verification.
1770 An OpenSSL server will accept a DH certificate for client authentication
1771 without the certificate verify message. This effectively allows a client to
1772 authenticate without the use of a private key. This only affects servers
1773 which trust a client certificate authority which issues certificates
1774 containing DH keys: these are extremely rare and hardly ever encountered.
1775 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1780 *) Ensure that the session ID context of an SSL is updated when its
1781 SSL_CTX is updated via SSL_set_SSL_CTX.
1783 The session ID context is typically set from the parent SSL_CTX,
1784 and can vary with the CTX.
1787 *) Fix various certificate fingerprint issues.
1789 By using non-DER or invalid encodings outside the signed portion of a
1790 certificate the fingerprint can be changed without breaking the signature.
1791 Although no details of the signed portion of the certificate can be changed
1792 this can cause problems with some applications: e.g. those using the
1793 certificate fingerprint for blacklists.
1795 1. Reject signatures with non zero unused bits.
1797 If the BIT STRING containing the signature has non zero unused bits reject
1798 the signature. All current signature algorithms require zero unused bits.
1800 2. Check certificate algorithm consistency.
1802 Check the AlgorithmIdentifier inside TBS matches the one in the
1803 certificate signature. NB: this will result in signature failure
1804 errors for some broken certificates.
1806 Thanks to Konrad Kraszewski from Google for reporting this issue.
1808 3. Check DSA/ECDSA signatures use DER.
1810 Reencode DSA/ECDSA signatures and compare with the original received
1811 signature. Return an error if there is a mismatch.
1813 This will reject various cases including garbage after signature
1814 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1815 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1816 (negative or with leading zeroes).
1818 Further analysis was conducted and fixes were developed by Stephen Henson
1819 of the OpenSSL core team.
1824 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1825 results on some platforms, including x86_64. This bug occurs at random
1826 with a very low probability, and is not known to be exploitable in any
1827 way, though its exact impact is difficult to determine. Thanks to Pieter
1828 Wuille (Blockstream) who reported this issue and also suggested an initial
1829 fix. Further analysis was conducted by the OpenSSL development team and
1830 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1831 the OpenSSL core team.
1835 *) Do not resume sessions on the server if the negotiated protocol
1836 version does not match the session's version. Resuming with a different
1837 version, while not strictly forbidden by the RFC, is of questionable
1838 sanity and breaks all known clients.
1839 [David Benjamin, Emilia Käsper]
1841 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1842 early CCS messages during renegotiation. (Note that because
1843 renegotiation is encrypted, this early CCS was not exploitable.)
1846 *) Tighten client-side session ticket handling during renegotiation:
1847 ensure that the client only accepts a session ticket if the server sends
1848 the extension anew in the ServerHello. Previously, a TLS client would
1849 reuse the old extension state and thus accept a session ticket if one was
1850 announced in the initial ServerHello.
1852 Similarly, ensure that the client requires a session ticket if one
1853 was advertised in the ServerHello. Previously, a TLS client would
1854 ignore a missing NewSessionTicket message.
1857 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1859 *) SRTP Memory Leak.
1861 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1862 sends a carefully crafted handshake message, to cause OpenSSL to fail
1863 to free up to 64k of memory causing a memory leak. This could be
1864 exploited in a Denial Of Service attack. This issue affects OpenSSL
1865 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1866 whether SRTP is used or configured. Implementations of OpenSSL that
1867 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1869 The fix was developed by the OpenSSL team.
1873 *) Session Ticket Memory Leak.
1875 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1876 integrity of that ticket is first verified. In the event of a session
1877 ticket integrity check failing, OpenSSL will fail to free memory
1878 causing a memory leak. By sending a large number of invalid session
1879 tickets an attacker could exploit this issue in a Denial Of Service
1884 *) Build option no-ssl3 is incomplete.
1886 When OpenSSL is configured with "no-ssl3" as a build option, servers
1887 could accept and complete a SSL 3.0 handshake, and clients could be
1888 configured to send them.
1890 [Akamai and the OpenSSL team]
1892 *) Add support for TLS_FALLBACK_SCSV.
1893 Client applications doing fallback retries should call
1894 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1896 [Adam Langley, Bodo Moeller]
1898 *) Add additional DigestInfo checks.
1900 Reencode DigestInto in DER and check against the original when
1901 verifying RSA signature: this will reject any improperly encoded
1902 DigestInfo structures.
1904 Note: this is a precautionary measure and no attacks are currently known.
1908 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1910 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1911 SRP code can be overrun an internal buffer. Add sanity check that
1912 g, A, B < N to SRP code.
1914 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1915 Group for discovering this issue.
1919 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1920 TLS 1.0 instead of higher protocol versions when the ClientHello message
1921 is badly fragmented. This allows a man-in-the-middle attacker to force a
1922 downgrade to TLS 1.0 even if both the server and the client support a
1923 higher protocol version, by modifying the client's TLS records.
1925 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1926 researching this issue.
1930 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1931 to a denial of service attack. A malicious server can crash the client
1932 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1933 ciphersuite and sending carefully crafted handshake messages.
1935 Thanks to Felix Gröbert (Google) for discovering and researching this
1940 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1941 to leak memory. This can be exploited through a Denial of Service attack.
1942 Thanks to Adam Langley for discovering and researching this issue.
1946 *) An attacker can force openssl to consume large amounts of memory whilst
1947 processing DTLS handshake messages. This can be exploited through a
1948 Denial of Service attack.
1949 Thanks to Adam Langley for discovering and researching this issue.
1953 *) An attacker can force an error condition which causes openssl to crash
1954 whilst processing DTLS packets due to memory being freed twice. This
1955 can be exploited through a Denial of Service attack.
1956 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1961 *) If a multithreaded client connects to a malicious server using a resumed
1962 session and the server sends an ec point format extension it could write
1963 up to 255 bytes to freed memory.
1965 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1970 *) A malicious server can crash an OpenSSL client with a null pointer
1971 dereference (read) by specifying an SRP ciphersuite even though it was not
1972 properly negotiated with the client. This can be exploited through a
1973 Denial of Service attack.
1975 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1976 discovering and researching this issue.
1980 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1981 X509_name_oneline, X509_name_print_ex et al. to leak some information
1982 from the stack. Applications may be affected if they echo pretty printing
1983 output to the attacker.
1985 Thanks to Ivan Fratric (Google) for discovering this issue.
1987 [Emilia Käsper, and Steve Henson]
1989 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1990 for corner cases. (Certain input points at infinity could lead to
1991 bogus results, with non-infinity inputs mapped to infinity too.)
1994 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1996 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1997 handshake can force the use of weak keying material in OpenSSL
1998 SSL/TLS clients and servers.
2000 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2001 researching this issue. (CVE-2014-0224)
2002 [KIKUCHI Masashi, Steve Henson]
2004 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2005 OpenSSL DTLS client the code can be made to recurse eventually crashing
2008 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2010 [Imre Rad, Steve Henson]
2012 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2013 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2014 client or server. This is potentially exploitable to run arbitrary
2015 code on a vulnerable client or server.
2017 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2018 [Jüri Aedla, Steve Henson]
2020 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2021 are subject to a denial of service attack.
2023 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2024 this issue. (CVE-2014-3470)
2025 [Felix Gröbert, Ivan Fratric, Steve Henson]
2027 *) Harmonize version and its documentation. -f flag is used to display
2029 [mancha <mancha1@zoho.com>]
2031 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2032 in i2d_ECPrivateKey.
2033 [mancha <mancha1@zoho.com>]
2035 *) Fix some double frees. These are not thought to be exploitable.
2036 [mancha <mancha1@zoho.com>]
2038 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2040 *) A missing bounds check in the handling of the TLS heartbeat extension
2041 can be used to reveal up to 64k of memory to a connected client or
2044 Thanks for Neel Mehta of Google Security for discovering this bug and to
2045 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2046 preparing the fix (CVE-2014-0160)
2047 [Adam Langley, Bodo Moeller]
2049 *) Fix for the attack described in the paper "Recovering OpenSSL
2050 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2051 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2052 http://eprint.iacr.org/2014/140
2054 Thanks to Yuval Yarom and Naomi Benger for discovering this
2055 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2056 [Yuval Yarom and Naomi Benger]
2058 *) TLS pad extension: draft-agl-tls-padding-03
2060 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2061 TLS client Hello record length value would otherwise be > 255 and
2062 less that 512 pad with a dummy extension containing zeroes so it
2063 is at least 512 bytes long.
2065 [Adam Langley, Steve Henson]
2067 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2069 *) Fix for TLS record tampering bug. A carefully crafted invalid
2070 handshake could crash OpenSSL with a NULL pointer exception.
2071 Thanks to Anton Johansson for reporting this issues.
2074 *) Keep original DTLS digest and encryption contexts in retransmission
2075 structures so we can use the previous session parameters if they need
2076 to be resent. (CVE-2013-6450)
2079 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2080 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2081 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2082 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2083 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2084 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2085 [Rob Stradling, Adam Langley]
2087 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2089 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2090 supporting platforms or when small records were transferred.
2091 [Andy Polyakov, Steve Henson]
2093 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2095 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2097 This addresses the flaw in CBC record processing discovered by
2098 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2099 at: http://www.isg.rhul.ac.uk/tls/
2101 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2102 Security Group at Royal Holloway, University of London
2103 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2104 Emilia Käsper for the initial patch.
2106 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2108 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2109 ciphersuites which can be exploited in a denial of service attack.
2110 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2111 and detecting this bug and to Wolfgang Ettlinger
2112 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2116 *) Return an error when checking OCSP signatures when key is NULL.
2117 This fixes a DoS attack. (CVE-2013-0166)
2120 *) Make openssl verify return errors.
2121 [Chris Palmer <palmer@google.com> and Ben Laurie]
2123 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2124 the right response is stapled. Also change SSL_get_certificate()
2125 so it returns the certificate actually sent.
2126 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2127 [Rob Stradling <rob.stradling@comodo.com>]
2129 *) Fix possible deadlock when decoding public keys.
2132 *) Don't use TLS 1.0 record version number in initial client hello
2136 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2138 *) Sanity check record length before skipping explicit IV in TLS
2139 1.2, 1.1 and DTLS to fix DoS attack.
2141 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2142 fuzzing as a service testing platform.
2146 *) Initialise tkeylen properly when encrypting CMS messages.
2147 Thanks to Solar Designer of Openwall for reporting this issue.
2150 *) In FIPS mode don't try to use composite ciphers as they are not
2154 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2156 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2157 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2158 mean any application compiled against OpenSSL 1.0.0 headers setting
2159 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2160 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2161 0x10000000L Any application which was previously compiled against
2162 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2163 will need to be recompiled as a result. Letting be results in
2164 inability to disable specifically TLS 1.1 and in client context,
2165 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2168 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2169 disable just protocol X, but all protocols above X *if* there are
2170 protocols *below* X still enabled. In more practical terms it means
2171 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2172 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2173 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2177 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2179 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2180 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2181 in CRYPTO_realloc_clean.
2183 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2184 issue and to Adam Langley <agl@chromium.org> for fixing it.
2186 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2188 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2191 *) Workarounds for some broken servers that "hang" if a client hello
2192 record length exceeds 255 bytes.
2194 1. Do not use record version number > TLS 1.0 in initial client
2195 hello: some (but not all) hanging servers will now work.
2196 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2197 the number of ciphers sent in the client hello. This should be
2198 set to an even number, such as 50, for example by passing:
2199 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2200 Most broken servers should now work.
2201 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2202 TLS 1.2 client support entirely.
2205 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2208 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2210 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2211 STRING form instead of a DigestInfo.
2214 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2215 and the RSA_sign/RSA_verify functions. This was made more apparent when
2216 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2217 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2218 the correct format in RSA_verify so both forms transparently work.
2221 *) Some servers which support TLS 1.0 can choke if we initially indicate
2222 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2223 encrypted premaster secret. As a workaround use the maximum permitted
2224 client version in client hello, this should keep such servers happy
2225 and still work with previous versions of OpenSSL.
2228 *) Add support for TLS/DTLS heartbeats.
2229 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2231 *) Add support for SCTP.
2232 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2234 *) Improved PRNG seeding for VOS.
2235 [Paul Green <Paul.Green@stratus.com>]
2237 *) Extensive assembler packs updates, most notably:
2239 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2240 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2241 - x86_64: bit-sliced AES implementation;
2242 - ARM: NEON support, contemporary platforms optimizations;
2243 - s390x: z196 support;
2244 - *: GHASH and GF(2^m) multiplication implementations;
2248 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2249 (removal of unnecessary code)
2250 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2252 *) Add TLS key material exporter from RFC 5705.
2255 *) Add DTLS-SRTP negotiation from RFC 5764.
2258 *) Add Next Protocol Negotiation,
2259 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2260 disabled with a no-npn flag to config or Configure. Code donated
2262 [Adam Langley <agl@google.com> and Ben Laurie]
2264 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2265 NIST-P256, NIST-P521, with constant-time single point multiplication on
2266 typical inputs. Compiler support for the nonstandard type __uint128_t is
2267 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2268 Code made available under Apache License version 2.0.
2270 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2271 line to include this in your build of OpenSSL, and run "make depend" (or
2272 "make update"). This enables the following EC_METHODs:
2274 EC_GFp_nistp224_method()
2275 EC_GFp_nistp256_method()
2276 EC_GFp_nistp521_method()
2278 EC_GROUP_new_by_curve_name() will automatically use these (while
2279 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2281 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2283 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2284 all platforms. Move ssize_t definition from e_os.h to the public
2285 header file e_os2.h as it now appears in public header file cms.h
2288 *) New -sigopt option to the ca, req and x509 utilities. Additional
2289 signature parameters can be passed using this option and in
2293 *) Add RSA PSS signing function. This will generate and set the
2294 appropriate AlgorithmIdentifiers for PSS based on those in the
2295 corresponding EVP_MD_CTX structure. No application support yet.
2298 *) Support for companion algorithm specific ASN1 signing routines.
2299 New function ASN1_item_sign_ctx() signs a pre-initialised
2300 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2301 the appropriate parameters.
2304 *) Add new algorithm specific ASN1 verification initialisation function
2305 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2306 handling will be the same no matter what EVP_PKEY_METHOD is used.
2307 Add a PSS handler to support verification of PSS signatures: checked
2308 against a number of sample certificates.
2311 *) Add signature printing for PSS. Add PSS OIDs.
2312 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2314 *) Add algorithm specific signature printing. An individual ASN1 method
2315 can now print out signatures instead of the standard hex dump.
2317 More complex signatures (e.g. PSS) can print out more meaningful
2318 information. Include DSA version that prints out the signature
2322 *) Password based recipient info support for CMS library: implementing
2326 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2327 neatly separates the code into cipher and PBE sections and is required
2328 for some algorithms that split PBES2 into separate pieces (such as
2329 password based CMS).
2332 *) Session-handling fixes:
2333 - Fix handling of connections that are resuming with a session ID,
2334 but also support Session Tickets.
2335 - Fix a bug that suppressed issuing of a new ticket if the client
2336 presented a ticket with an expired session.
2337 - Try to set the ticket lifetime hint to something reasonable.
2338 - Make tickets shorter by excluding irrelevant information.
2339 - On the client side, don't ignore renewed tickets.
2340 [Adam Langley, Bodo Moeller (Google)]
2342 *) Fix PSK session representation.
2345 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2347 This work was sponsored by Intel.
2350 *) Add GCM support to TLS library. Some custom code is needed to split
2351 the IV between the fixed (from PRF) and explicit (from TLS record)
2352 portions. This adds all GCM ciphersuites supported by RFC5288 and
2353 RFC5289. Generalise some AES* cipherstrings to include GCM and
2354 add a special AESGCM string for GCM only.
2357 *) Expand range of ctrls for AES GCM. Permit setting invocation
2358 field on decrypt and retrieval of invocation field only on encrypt.
2361 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2362 As required by RFC5289 these ciphersuites cannot be used if for
2363 versions of TLS earlier than 1.2.
2366 *) For FIPS capable OpenSSL interpret a NULL default public key method
2367 as unset and return the appropriate default but do *not* set the default.
2368 This means we can return the appropriate method in applications that
2369 switch between FIPS and non-FIPS modes.
2372 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2373 ENGINE is used then we cannot handle that in the FIPS module so we
2374 keep original code iff non-FIPS operations are allowed.
2377 *) Add -attime option to openssl utilities.
2378 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2380 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2383 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2384 FIPS EC methods unconditionally for now.
2387 *) New build option no-ec2m to disable characteristic 2 code.
2390 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2391 all cases can be covered as some introduce binary incompatibilities.
2394 *) Redirect RSA operations to FIPS module including keygen,
2395 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2398 *) Add similar low level API blocking to ciphers.
2401 *) Low level digest APIs are not approved in FIPS mode: any attempt
2402 to use these will cause a fatal error. Applications that *really* want
2403 to use them can use the private_* version instead.
2406 *) Redirect cipher operations to FIPS module for FIPS builds.
2409 *) Redirect digest operations to FIPS module for FIPS builds.
2412 *) Update build system to add "fips" flag which will link in fipscanister.o
2413 for static and shared library builds embedding a signature if needed.
2416 *) Output TLS supported curves in preference order instead of numerical
2417 order. This is currently hardcoded for the highest order curves first.
2418 This should be configurable so applications can judge speed vs strength.
2421 *) Add TLS v1.2 server support for client authentication.
2424 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2428 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2429 FIPS modules versions.
2432 *) Add TLS v1.2 client side support for client authentication. Keep cache
2433 of handshake records longer as we don't know the hash algorithm to use
2434 until after the certificate request message is received.
2437 *) Initial TLS v1.2 client support. Add a default signature algorithms
2438 extension including all the algorithms we support. Parse new signature
2439 format in client key exchange. Relax some ECC signing restrictions for
2440 TLS v1.2 as indicated in RFC5246.
2443 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2444 to new signature format when needed using client digest preference.
2445 All server ciphersuites should now work correctly in TLS v1.2. No client
2446 support yet and no support for client certificates.
2449 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2450 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2451 ciphersuites. At present only RSA key exchange ciphersuites work with
2452 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2453 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2454 and version checking.
2457 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2458 with this defined it will not be affected by any changes to ssl internal
2459 structures. Add several utility functions to allow openssl application
2460 to work with OPENSSL_NO_SSL_INTERN defined.
2464 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2466 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2469 *) Permit abbreviated handshakes when renegotiating using the function
2470 SSL_renegotiate_abbreviated().
2471 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2473 *) Add call to ENGINE_register_all_complete() to
2474 ENGINE_load_builtin_engines(), so some implementations get used
2475 automatically instead of needing explicit application support.
2478 *) Add support for TLS key exporter as described in RFC5705.
2479 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2481 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2482 a few changes are required:
2484 Add SSL_OP_NO_TLSv1_1 flag.
2485 Add TLSv1_1 methods.
2486 Update version checking logic to handle version 1.1.
2487 Add explicit IV handling (ported from DTLS code).
2488 Add command line options to s_client/s_server.
2491 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2493 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2494 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2495 content decryption and always return the same error. Note: this attack
2496 needs on average 2^20 messages so it only affects automated senders. The
2497 old behaviour can be reenabled in the CMS code by setting the
2498 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2499 an MMA defence is not necessary.
2500 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2501 this issue. (CVE-2012-0884)
2504 *) Fix CVE-2011-4619: make sure we really are receiving a
2505 client hello before rejecting multiple SGC restarts. Thanks to
2506 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2509 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2511 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2512 Thanks to Antonio Martin, Enterprise Secure Access Research and
2513 Development, Cisco Systems, Inc. for discovering this bug and
2514 preparing a fix. (CVE-2012-0050)
2517 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2519 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2520 of the Vaudenay padding oracle attack on CBC mode encryption
2521 which enables an efficient plaintext recovery attack against
2522 the OpenSSL implementation of DTLS. Their attack exploits timing
2523 differences arising during decryption processing. A research
2524 paper describing this attack can be found at:
2525 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2526 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2527 Security Group at Royal Holloway, University of London
2528 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2529 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2530 for preparing the fix. (CVE-2011-4108)
2531 [Robin Seggelmann, Michael Tuexen]
2533 *) Clear bytes used for block padding of SSL 3.0 records.
2535 [Adam Langley (Google)]
2537 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2538 Kadianakis <desnacked@gmail.com> for discovering this issue and
2539 Adam Langley for preparing the fix. (CVE-2011-4619)
2540 [Adam Langley (Google)]
2542 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2543 [Andrey Kulikov <amdeich@gmail.com>]
2545 *) Prevent malformed RFC3779 data triggering an assertion failure.
2546 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2547 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2548 [Rob Austein <sra@hactrn.net>]
2550 *) Improved PRNG seeding for VOS.
2551 [Paul Green <Paul.Green@stratus.com>]
2553 *) Fix ssl_ciph.c set-up race.
2554 [Adam Langley (Google)]
2556 *) Fix spurious failures in ecdsatest.c.
2557 [Emilia Käsper (Google)]
2559 *) Fix the BIO_f_buffer() implementation (which was mixing different
2560 interpretations of the '..._len' fields).
2561 [Adam Langley (Google)]
2563 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2564 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2565 threads won't reuse the same blinding coefficients.
2567 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2568 lock to call BN_BLINDING_invert_ex, and avoids one use of
2569 BN_BLINDING_update for each BN_BLINDING structure (previously,
2570 the last update always remained unused).
2571 [Emilia Käsper (Google)]
2573 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2574 [Bob Buckholz (Google)]
2576 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2578 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2579 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2580 [Kaspar Brand <ossl@velox.ch>]
2582 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2583 for multi-threaded use of ECDH. (CVE-2011-3210)
2584 [Adam Langley (Google)]
2586 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2589 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2590 signature public key algorithm by using OID xref utilities instead.
2591 Before this you could only use some ECC ciphersuites with SHA1 only.
2594 *) Add protection against ECDSA timing attacks as mentioned in the paper
2595 by Billy Bob Brumley and Nicola Tuveri, see:
2597 http://eprint.iacr.org/2011/232.pdf
2599 [Billy Bob Brumley and Nicola Tuveri]
2601 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2603 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2604 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2606 *) Fix bug in string printing code: if *any* escaping is enabled we must
2607 escape the escape character (backslash) or the resulting string is
2611 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2613 *) Disable code workaround for ancient and obsolete Netscape browsers
2614 and servers: an attacker can use it in a ciphersuite downgrade attack.
2615 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2618 *) Fixed J-PAKE implementation error, originally discovered by
2619 Sebastien Martini, further info and confirmation from Stefan
2620 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2623 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2625 *) Fix extension code to avoid race conditions which can result in a buffer
2626 overrun vulnerability: resumed sessions must not be modified as they can
2627 be shared by multiple threads. CVE-2010-3864
2630 *) Fix WIN32 build system to correctly link an ENGINE directory into
2634 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2636 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2638 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2640 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2642 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2643 context. The operation can be customised via the ctrl mechanism in
2644 case ENGINEs want to include additional functionality.
2647 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2650 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2651 output hashes compatible with older versions of OpenSSL.
2652 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2654 *) Fix compression algorithm handling: if resuming a session use the
2655 compression algorithm of the resumed session instead of determining
2656 it from client hello again. Don't allow server to change algorithm.
2659 *) Add load_crls() function to apps tidying load_certs() too. Add option
2660 to verify utility to allow additional CRLs to be included.
2663 *) Update OCSP request code to permit adding custom headers to the request:
2664 some responders need this.
2667 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2669 [Julia Lawall <julia@diku.dk>]
2671 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2672 needlessly dereferenced structures, used obsolete functions and
2673 didn't handle all updated verify codes correctly.
2676 *) Disable MD2 in the default configuration.
2679 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2680 indicate the initial BIO being pushed or popped. This makes it possible
2681 to determine whether the BIO is the one explicitly called or as a result
2682 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2683 it handles reference counts correctly and doesn't zero out the I/O bio
2684 when it is not being explicitly popped. WARNING: applications which
2685 included workarounds for the old buggy behaviour will need to be modified
2686 or they could free up already freed BIOs.
2689 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2690 renaming to all platforms (within the 0.9.8 branch, this was
2691 done conditionally on Netware platforms to avoid a name clash).
2692 [Guenter <lists@gknw.net>]
2694 *) Add ECDHE and PSK support to DTLS.
2695 [Michael Tuexen <tuexen@fh-muenster.de>]
2697 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2701 *) Add "missing" function EVP_MD_flags() (without this the only way to
2702 retrieve a digest flags is by accessing the structure directly. Update
2703 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2704 or cipher is registered as in the "from" argument. Print out all
2705 registered digests in the dgst usage message instead of manually
2706 attempting to work them out.
2709 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2710 this allows the use of compression and extensions. Change default cipher
2711 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2712 by default unless an application cipher string requests it.
2715 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2716 key ids to find matching certificates and keys but some PKCS#12 files
2717 don't follow the (somewhat unwritten) rules and this strategy fails.
2718 Now just gather all certificates together and the first private key
2719 then look for the first certificate that matches the key.
2722 *) Support use of registered digest and cipher names for dgst and cipher
2723 commands instead of having to add each one as a special case. So now
2730 openssl dgst -sha256 foo
2732 and this works for ENGINE based algorithms too.
2736 *) Update Gost ENGINE to support parameter files.
2737 [Victor B. Wagner <vitus@cryptocom.ru>]
2739 *) Support GeneralizedTime in ca utility.
2740 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2742 *) Enhance the hash format used for certificate directory links. The new
2743 form uses the canonical encoding (meaning equivalent names will work
2744 even if they aren't identical) and uses SHA1 instead of MD5. This form
2745 is incompatible with the older format and as a result c_rehash should
2746 be used to rebuild symbolic links.
2749 *) Make PKCS#8 the default write format for private keys, replacing the
2750 traditional format. This form is standardised, more secure and doesn't
2751 include an implicit MD5 dependency.
2754 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2755 committed to OpenSSL should pass this lot as a minimum.
2758 *) Add session ticket override functionality for use by EAP-FAST.
2759 [Jouni Malinen <j@w1.fi>]
2761 *) Modify HMAC functions to return a value. Since these can be implemented
2762 in an ENGINE errors can occur.
2765 *) Type-checked OBJ_bsearch_ex.
2768 *) Type-checked OBJ_bsearch. Also some constification necessitated
2769 by type-checking. Still to come: TXT_DB, bsearch(?),
2770 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2774 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2775 seconds to a tm structure directly, instead of going through OS
2776 specific date routines. This avoids any issues with OS routines such
2777 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2778 and X509_time_adj_ex() to cover the extended range. The existing
2779 X509_time_adj() is still usable and will no longer have any date issues.
2782 *) Delta CRL support. New use deltas option which will attempt to locate
2783 and search any appropriate delta CRLs available.
2785 This work was sponsored by Google.
2788 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2789 code and add additional score elements. Validate alternate CRL paths
2790 as part of the CRL checking and indicate a new error "CRL path validation
2791 error" in this case. Applications wanting additional details can use
2792 the verify callback and check the new "parent" field. If this is not
2793 NULL CRL path validation is taking place. Existing applications wont
2794 see this because it requires extended CRL support which is off by
2797 This work was sponsored by Google.
2800 *) Support for freshest CRL extension.
2802 This work was sponsored by Google.
2805 *) Initial indirect CRL support. Currently only supported in the CRLs
2806 passed directly and not via lookup. Process certificate issuer
2807 CRL entry extension and lookup CRL entries by bother issuer name
2808 and serial number. Check and process CRL issuer entry in IDP extension.
2810 This work was sponsored by Google.
2813 *) Add support for distinct certificate and CRL paths. The CRL issuer
2814 certificate is validated separately in this case. Only enabled if
2815 an extended CRL support flag is set: this flag will enable additional
2816 CRL functionality in future.
2818 This work was sponsored by Google.
2821 *) Add support for policy mappings extension.
2823 This work was sponsored by Google.
2826 *) Fixes to pathlength constraint, self issued certificate handling,
2827 policy processing to align with RFC3280 and PKITS tests.
2829 This work was sponsored by Google.
2832 *) Support for name constraints certificate extension. DN, email, DNS
2833 and URI types are currently supported.
2835 This work was sponsored by Google.
2838 *) To cater for systems that provide a pointer-based thread ID rather
2839 than numeric, deprecate the current numeric thread ID mechanism and
2840 replace it with a structure and associated callback type. This
2841 mechanism allows a numeric "hash" to be extracted from a thread ID in
2842 either case, and on platforms where pointers are larger than 'long',
2843 mixing is done to help ensure the numeric 'hash' is usable even if it
2844 can't be guaranteed unique. The default mechanism is to use "&errno"
2845 as a pointer-based thread ID to distinguish between threads.
2847 Applications that want to provide their own thread IDs should now use
2848 CRYPTO_THREADID_set_callback() to register a callback that will call
2849 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2851 Note that ERR_remove_state() is now deprecated, because it is tied
2852 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2853 to free the current thread's error state should be replaced by
2854 ERR_remove_thread_state(NULL).
2856 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2857 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2858 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2859 application was previously providing a numeric thread callback that
2860 was inappropriate for distinguishing threads, then uniqueness might
2861 have been obtained with &errno that happened immediately in the
2862 intermediate development versions of OpenSSL; this is no longer the
2863 case, the numeric thread callback will now override the automatic use
2865 [Geoff Thorpe, with help from Bodo Moeller]
2867 *) Initial support for different CRL issuing certificates. This covers a
2868 simple case where the self issued certificates in the chain exist and
2869 the real CRL issuer is higher in the existing chain.
2871 This work was sponsored by Google.
2874 *) Removed effectively defunct crypto/store from the build.
2877 *) Revamp of STACK to provide stronger type-checking. Still to come:
2878 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2879 ASN1_STRING, CONF_VALUE.
2882 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2883 RAM on SSL connections. This option can save about 34k per idle SSL.
2886 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2887 STACK, TXT_DB, bsearch, qsort.
2890 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2891 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2892 support for data, signedData, compressedData, digestedData and
2893 encryptedData, envelopedData types included. Scripts to check against
2894 RFC4134 examples draft and interop and consistency checks of many
2895 content types and variants.
2898 *) Add options to enc utility to support use of zlib compression BIO.
2901 *) Extend mk1mf to support importing of options and assembly language
2902 files from Configure script, currently only included in VC-WIN32.
2903 The assembly language rules can now optionally generate the source
2904 files from the associated perl scripts.
2907 *) Implement remaining functionality needed to support GOST ciphersuites.
2908 Interop testing has been performed using CryptoPro implementations.
2909 [Victor B. Wagner <vitus@cryptocom.ru>]
2911 *) s390x assembler pack.
2914 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2918 *) Implement Opaque PRF Input TLS extension as specified in
2919 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2920 official specification yet and no extension type assignment by
2921 IANA exists, this extension (for now) will have to be explicitly
2922 enabled when building OpenSSL by providing the extension number
2923 to use. For example, specify an option
2925 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2927 to the "config" or "Configure" script to enable the extension,
2928 assuming extension number 0x9527 (which is a completely arbitrary
2929 and unofficial assignment based on the MD5 hash of the Internet
2930 Draft). Note that by doing so, you potentially lose
2931 interoperability with other TLS implementations since these might
2932 be using the same extension number for other purposes.
2934 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2935 opaque PRF input value to use in the handshake. This will create
2936 an interal copy of the length-'len' string at 'src', and will
2937 return non-zero for success.
2939 To get more control and flexibility, provide a callback function
2942 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2943 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2947 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2950 Callback function 'cb' will be called in handshakes, and is
2951 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2952 Argument 'arg' is for application purposes (the value as given to
2953 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2954 be provided to the callback function). The callback function
2955 has to return non-zero to report success: usually 1 to use opaque
2956 PRF input just if possible, or 2 to enforce use of the opaque PRF
2957 input. In the latter case, the library will abort the handshake
2958 if opaque PRF input is not successfully negotiated.
2960 Arguments 'peerinput' and 'len' given to the callback function
2961 will always be NULL and 0 in the case of a client. A server will
2962 see the client's opaque PRF input through these variables if
2963 available (NULL and 0 otherwise). Note that if the server
2964 provides an opaque PRF input, the length must be the same as the
2965 length of the client's opaque PRF input.
2967 Note that the callback function will only be called when creating
2968 a new session (session resumption can resume whatever was
2969 previously negotiated), and will not be called in SSL 2.0
2970 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2971 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2972 for applications that need to enforce opaque PRF input.
2976 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2979 [Victor B. Wagner <vitus@cryptocom.ru>]
2981 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2982 RFC4507bis. The encrypted ticket format is an encrypted encoded
2983 SSL_SESSION structure, that way new session features are automatically
2986 If a client application caches session in an SSL_SESSION structure
2987 support is transparent because tickets are now stored in the encoded
2990 The SSL_CTX structure automatically generates keys for ticket
2991 protection in servers so again support should be possible
2992 with no application modification.
2994 If a client or server wishes to disable RFC4507 support then the option
2995 SSL_OP_NO_TICKET can be set.
2997 Add a TLS extension debugging callback to allow the contents of any client
2998 or server extensions to be examined.
3000 This work was sponsored by Google.
3003 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3004 OpenSSL should now compile cleanly on gcc 4.2
3005 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3007 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3008 support including streaming MAC support: this is required for GOST
3009 ciphersuite support.
3010 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3012 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3013 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3014 to output in BER and PEM format.
3017 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3018 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3019 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3020 ENGINE support for HMAC keys which are unextractable. New -mac and
3021 -macopt options to dgst utility.
3024 *) New option -sigopt to dgst utility. Update dgst to use
3025 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3026 alternative signing parameters such as X9.31 or PSS in the dgst
3030 *) Change ssl_cipher_apply_rule(), the internal function that does
3031 the work each time a ciphersuite string requests enabling
3032 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3033 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3034 the order of disabled ciphersuites such that those ciphersuites
3035 that most recently went from enabled to disabled not only stay
3036 in order with respect to each other, but also have higher priority
3037 than other disabled ciphersuites the next time ciphersuites are
3040 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3041 the same ciphersuites as with "HIGH" alone, but in a specific
3042 order where the PSK ciphersuites come first (since they are the
3043 most recently disabled ciphersuites when "HIGH" is parsed).
3045 Also, change ssl_create_cipher_list() (using this new
3046 funcionality) such that between otherwise identical
3047 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3051 *) Change ssl_create_cipher_list() so that it automatically
3052 arranges the ciphersuites in reasonable order before starting
3053 to process the rule string. Thus, the definition for "DEFAULT"
3054 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3055 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3056 This makes it much easier to arrive at a reasonable default order
3057 in applications for which anonymous ciphers are OK (meaning
3058 that you can't actually use DEFAULT).
3059 [Bodo Moeller; suggested by Victor Duchovni]
3061 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3062 processing) into multiple integers instead of setting
3063 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3064 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3065 (These masks as well as the individual bit definitions are hidden
3066 away into the non-exported interface ssl/ssl_locl.h, so this
3067 change to the definition of the SSL_CIPHER structure shouldn't
3068 affect applications.) This give us more bits for each of these
3069 categories, so there is no longer a need to coagulate AES128 and
3070 AES256 into a single algorithm bit, and to coagulate Camellia128
3071 and Camellia256 into a single algorithm bit, which has led to all
3074 Thus, among other things, the kludge introduced in 0.9.7m and
3075 0.9.8e for masking out AES256 independently of AES128 or masking
3076 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3078 With the change, we also introduce new ciphersuite aliases that
3079 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3083 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3084 Use the leftmost N bytes of the signature input if the input is
3085 larger than the prime q (with N being the size in bytes of q).
3088 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3089 it yet and it is largely untested.
3092 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3095 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3096 some compilers (gcc 4.2 and later) reject their use. Safestack is
3097 reimplemented. Update ASN1 to avoid use of legacy functions.
3100 *) Win32/64 targets are linked with Winsock2.
3103 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3104 to external functions. This can be used to increase CRL handling
3105 efficiency especially when CRLs are very large by (for example) storing
3106 the CRL revoked certificates in a database.
3109 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3110 new CRLs added to a directory can be used. New command line option
3111 -verify_return_error to s_client and s_server. This causes real errors
3112 to be returned by the verify callback instead of carrying on no matter
3113 what. This reflects the way a "real world" verify callback would behave.
3116 *) GOST engine, supporting several GOST algorithms and public key formats.
3117 Kindly donated by Cryptocom.
3120 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3121 partitioned by DP are handled but no indirect CRL or reason partitioning
3122 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3123 selected via a scoring technique which handles IDP and AKID in CRLs.
3126 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3127 will ultimately be used for all verify operations: this will remove the
3128 X509_STORE dependency on certificate verification and allow alternative
3129 lookup methods. X509_STORE based implementations of these two callbacks.
3132 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3133 Modify get_crl() to find a valid (unexpired) CRL if possible.
3136 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3137 this would be called X509_CRL_cmp() but that name is already used by
3138 a function that just compares CRL issuer names. Cache several CRL
3139 extensions in X509_CRL structure and cache CRLDP in X509.
3142 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3143 this maps equivalent X509_NAME structures into a consistent structure.
3144 Name comparison can then be performed rapidly using memcmp().
3147 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3151 *) Allow digests to supply their own micalg string for S/MIME type using
3152 the ctrl EVP_MD_CTRL_MICALG.
3155 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3156 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3157 ctrl. It can then customise the structure before and/or after signing
3161 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3162 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3163 to free up any added signature OIDs.
3166 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3167 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3168 digest and cipher tables. New options added to openssl utility:
3169 list-message-digest-algorithms and list-cipher-algorithms.
3172 *) Change the array representation of binary polynomials: the list
3173 of degrees of non-zero coefficients is now terminated with -1.
3174 Previously it was terminated with 0, which was also part of the
3175 value; thus, the array representation was not applicable to
3176 polynomials where t^0 has coefficient zero. This change makes
3177 the array representation useful in a more general context.
3180 *) Various modifications and fixes to SSL/TLS cipher string
3181 handling. For ECC, the code now distinguishes between fixed ECDH
3182 with RSA certificates on the one hand and with ECDSA certificates
3183 on the other hand, since these are separate ciphersuites. The
3184 unused code for Fortezza ciphersuites has been removed.
3186 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3187 (not "ECDHE"). For consistency with the code for DH
3188 certificates, use of ECDH certificates is now considered ECDH
3189 authentication, not RSA or ECDSA authentication (the latter is
3190 merely the CA's signing algorithm and not actively used in the
3193 The temporary ciphersuite alias "ECCdraft" is no longer
3194 available, and ECC ciphersuites are no longer excluded from "ALL"
3195 and "DEFAULT". The following aliases now exist for RFC 4492
3196 ciphersuites, most of these by analogy with the DH case:
3198 kECDHr - ECDH cert, signed with RSA
3199 kECDHe - ECDH cert, signed with ECDSA
3200 kECDH - ECDH cert (signed with either RSA or ECDSA)
3201 kEECDH - ephemeral ECDH
3202 ECDH - ECDH cert or ephemeral ECDH
3208 AECDH - anonymous ECDH
3209 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3213 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3214 Use correct micalg parameters depending on digest(s) in signed message.
3217 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3218 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3221 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3222 an engine to register a method. Add ENGINE lookups for methods and
3223 functional reference processing.
3226 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3227 EVP_{Sign,Verify}* which allow an application to customise the signature
3231 *) New -resign option to smime utility. This adds one or more signers
3232 to an existing PKCS#7 signedData structure. Also -md option to use an
3233 alternative message digest algorithm for signing.
3236 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3237 create PKCS7 structures containing multiple signers. Update smime
3238 application to support multiple signers.
3241 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3245 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3246 Reorganize PBE internals to lookup from a static table using NIDs,
3247 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3248 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3249 PRF which will be automatically used with PBES2.
3252 *) Replace the algorithm specific calls to generate keys in "req" with the
3256 *) Update PKCS#7 enveloped data routines to use new API. This is now
3257 supported by any public key method supporting the encrypt operation. A
3258 ctrl is added to allow the public key algorithm to examine or modify
3259 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3263 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3264 a default digest type to use. In most cases this will be SHA1 but some
3265 algorithms (such as GOST) need to specify an alternative digest. The
3266 return value indicates how strong the preference is 1 means optional and
3267 2 is mandatory (that is it is the only supported type). Modify
3268 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3269 use the default md. Update openssl utilities to use the default digest
3270 type for signing if it is not explicitly indicated.
3273 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3274 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3275 signing method from the key type. This effectively removes the link
3276 between digests and public key types.
3279 *) Add an OID cross reference table and utility functions. Its purpose is to
3280 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3281 rsaEncryption. This will allow some of the algorithm specific hackery
3282 needed to use the correct OID to be removed.
3285 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3286 structures for PKCS7_sign(). They are now set up by the relevant public
3290 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3293 *) Add support for key derivation (agreement) in the API, DH method and
3297 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3298 public and private key formats. As a side effect these add additional
3299 command line functionality not previously available: DSA signatures can be
3300 generated and verified using pkeyutl and DH key support and generation in
3305 [Oliver Tappe <zooey@hirschkaefer.de>]
3307 *) New make target "install_html_docs" installs HTML renditions of the
3309 [Oliver Tappe <zooey@hirschkaefer.de>]
3311 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3312 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3313 support key and parameter generation and add initial key generation
3314 functionality for RSA.
3317 *) Add functions for main EVP_PKEY_method operations. The undocumented
3318 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3319 EVP_PKEY_{encrypt,decrypt}_old.
3322 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3323 key API, doesn't do much yet.
3326 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3327 public key algorithms. New option to openssl utility:
3328 "list-public-key-algorithms" to print out info.
3331 *) Implement the Supported Elliptic Curves Extension for
3332 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3335 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3336 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3339 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3340 utilities such as rsa, dsa, dsaparam etc except they process any key
3344 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3345 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3346 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3350 *) Initial support for pluggable public key ASN1.
3351 De-spaghettify the public key ASN1 handling. Move public and private
3352 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3353 algorithm specific handling to a single module within the relevant
3354 algorithm directory. Add functions to allow (near) opaque processing
3355 of public and private key structures.
3358 *) Implement the Supported Point Formats Extension for
3359 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3362 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3363 for the psk identity [hint] and the psk callback functions to the
3364 SSL_SESSION, SSL and SSL_CTX structure.
3367 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3371 SSL_CTX_use_psk_identity_hint
3372 SSL_get_psk_identity_hint
3373 SSL_get_psk_identity
3374 SSL_use_psk_identity_hint
3376 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3378 *) Add RFC 3161 compliant time stamp request creation, response generation
3379 and response verification functionality.
3380 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3382 *) Add initial support for TLS extensions, specifically for the server_name
3383 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3384 have new members for a host name. The SSL data structure has an
3385 additional member SSL_CTX *initial_ctx so that new sessions can be
3386 stored in that context to allow for session resumption, even after the
3387 SSL has been switched to a new SSL_CTX in reaction to a client's
3388 server_name extension.
3390 New functions (subject to change):
3392 SSL_get_servername()
3393 SSL_get_servername_type()
3396 New CTRL codes and macros (subject to change):
3398 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3399 - SSL_CTX_set_tlsext_servername_callback()
3400 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3401 - SSL_CTX_set_tlsext_servername_arg()
3402 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3404 openssl s_client has a new '-servername ...' option.
3406 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3407 '-key2 ...', '-servername_fatal' (subject to change). This allows
3408 testing the HostName extension for a specific single host name ('-cert'
3409 and '-key' remain fallbacks for handshakes without HostName
3410 negotiation). If the unrecognized_name alert has to be sent, this by
3411 default is a warning; it becomes fatal with the '-servername_fatal'
3414 [Peter Sylvester, Remy Allais, Christophe Renou]
3416 *) Whirlpool hash implementation is added.
3419 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3420 bn(64,32). Because of instruction set limitations it doesn't have
3421 any negative impact on performance. This was done mostly in order
3422 to make it possible to share assembler modules, such as bn_mul_mont
3423 implementations, between 32- and 64-bit builds without hassle.
3426 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3427 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3431 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3432 dedicated Montgomery multiplication procedure, is introduced.
3433 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3434 "64-bit" performance on certain 32-bit targets.
3437 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3438 in SSL structures. New SSL ctrl to set maximum send fragment size.
3439 Save memory by seeting the I/O buffer sizes dynamically instead of
3440 using the maximum available value.
3443 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3444 in addition to the text details.
3447 *) Very, very preliminary EXPERIMENTAL support for printing of general
3448 ASN1 structures. This currently produces rather ugly output and doesn't
3449 handle several customised structures at all.
3452 *) Integrated support for PVK file format and some related formats such
3453 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3454 these in the 'rsa' and 'dsa' utilities.
3457 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3460 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3461 place for the (very old) "NETSCAPE" format certificates which are now
3462 handled using new ASN1 code equivalents.
3465 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3466 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3467 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3470 *) Modify CRL distribution points extension code to print out previously
3471 unsupported fields. Enhance extension setting code to allow setting of
3475 *) Add print and set support for Issuing Distribution Point CRL extension.
3478 *) Change 'Configure' script to enable Camellia by default.
3481 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3483 *) When rejecting SSL/TLS records due to an incorrect version number, never
3484 update s->server with a new major version number. As of
3485 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3486 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3487 the previous behavior could result in a read attempt at NULL when
3488 receiving specific incorrect SSL/TLS records once record payload
3489 protection is active. (CVE-2010-0740)
3490 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3492 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3493 could be crashed if the relevant tables were not present (e.g. chrooted).
3494 [Tomas Hoger <thoger@redhat.com>]
3496 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3498 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3499 [Martin Olsson, Neel Mehta]
3501 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3502 accommodate for stack sorting, always a write lock!).
3505 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3506 excessive delays in the RAND_poll(): over a minute. As a workaround
3507 include a time check in the inner Heap32Next loop too.
3510 *) The code that handled flushing of data in SSL/TLS originally used the
3511 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3512 the problem outlined in PR#1949. The fix suggested there however can
3513 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3514 of Apache). So instead simplify the code to flush unconditionally.
3515 This should be fine since flushing with no data to flush is a no op.
3518 *) Handle TLS versions 2.0 and later properly and correctly use the
3519 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3520 off ancient servers have a habit of sticking around for a while...
3523 *) Modify compression code so it frees up structures without using the
3524 ex_data callbacks. This works around a problem where some applications
3525 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3526 restarting) then use compression (e.g. SSL with compression) later.
3527 This results in significant per-connection memory leaks and
3528 has caused some security issues including CVE-2008-1678 and
3532 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3533 change when encrypting or decrypting.
3536 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3537 connect and renegotiate with servers which do not support RI.
3538 Until RI is more widely deployed this option is enabled by default.
3541 *) Add "missing" ssl ctrls to clear options and mode.
3544 *) If client attempts to renegotiate and doesn't support RI respond with
3545 a no_renegotiation alert as required by RFC5746. Some renegotiating
3546 TLS clients will continue a connection gracefully when they receive
3547 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3548 waiting for a server hello which it will never receive. Now we treat a
3549 received no_renegotiation alert as a fatal error. This is because
3550 applications requesting a renegotiation might well expect it to succeed
3551 and would have no code in place to handle the server denying it so the
3552 only safe thing to do is to terminate the connection.
3555 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3556 peer supports secure renegotiation and 0 otherwise. Print out peer
3557 renegotiation support in s_client/s_server.
3560 *) Replace the highly broken and deprecated SPKAC certification method with
3561 the updated NID creation version. This should correctly handle UTF8.
3564 *) Implement RFC5746. Re-enable renegotiation but require the extension
3565 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3566 turns out to be a bad idea. It has been replaced by
3567 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3568 SSL_CTX_set_options(). This is really not recommended unless you
3569 know what you are doing.
3570 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3572 *) Fixes to stateless session resumption handling. Use initial_ctx when
3573 issuing and attempting to decrypt tickets in case it has changed during
3574 servername handling. Use a non-zero length session ID when attempting
3575 stateless session resumption: this makes it possible to determine if
3576 a resumption has occurred immediately after receiving server hello
3577 (several places in OpenSSL subtly assume this) instead of later in
3581 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3582 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3583 fixes for a few places where the return code is not checked
3585 [Julia Lawall <julia@diku.dk>]
3587 *) Add --strict-warnings option to Configure script to include devteam
3588 warnings in other configurations.
3591 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3592 makes it possible to install openssl libraries in locations which
3593 have names other than "lib", for example "/usr/lib64" which some
3595 [Steve Henson, based on patch from Jeremy Utley]
3597 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3598 X690 8.9.12 and can produce some misleading textual output of OIDs.
3599 [Steve Henson, reported by Dan Kaminsky]
3601 *) Delete MD2 from algorithm tables. This follows the recommendation in
3602 several standards that it is not used in new applications due to
3603 several cryptographic weaknesses. For binary compatibility reasons
3604 the MD2 API is still compiled in by default.
3607 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3611 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3612 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3614 [Guenter <lists@gknw.net>]
3616 *) Fix the server certificate chain building code to use X509_verify_cert(),
3617 it used to have an ad-hoc builder which was unable to cope with anything
3618 other than a simple chain.
3619 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3621 *) Don't check self signed certificate signatures in X509_verify_cert()
3622 by default (a flag can override this): it just wastes time without
3623 adding any security. As a useful side effect self signed root CAs
3624 with non-FIPS digests are now usable in FIPS mode.
3627 *) In dtls1_process_out_of_seq_message() the check if the current message
3628 is already buffered was missing. For every new message was memory
3629 allocated, allowing an attacker to perform an denial of service attack
3630 with sending out of seq handshake messages until there is no memory
3631 left. Additionally every future messege was buffered, even if the
3632 sequence number made no sense and would be part of another handshake.
3633 So only messages with sequence numbers less than 10 in advance will be
3634 buffered. (CVE-2009-1378)
3635 [Robin Seggelmann, discovered by Daniel Mentz]
3637 *) Records are buffered if they arrive with a future epoch to be
3638 processed after finishing the corresponding handshake. There is
3639 currently no limitation to this buffer allowing an attacker to perform
3640 a DOS attack with sending records with future epochs until there is no
3641 memory left. This patch adds the pqueue_size() function to determine
3642 the size of a buffer and limits the record buffer to 100 entries.
3644 [Robin Seggelmann, discovered by Daniel Mentz]
3646 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3647 parent structure is freed. (CVE-2009-1379)
3650 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3651 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3654 [Ilya O. <vrghost@gmail.com>]
3656 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3658 *) Disable renegotiation completely - this fixes a severe security
3659 problem (CVE-2009-3555) at the cost of breaking all
3660 renegotiation. Renegotiation can be re-enabled by setting
3661 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3662 run-time. This is really not recommended unless you know what
3666 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3668 *) Don't set val to NULL when freeing up structures, it is freed up by
3669 underlying code. If sizeof(void *) > sizeof(long) this can result in
3670 zeroing past the valid field. (CVE-2009-0789)
3671 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3673 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3674 checked correctly. This would allow some invalid signed attributes to
3675 appear to verify correctly. (CVE-2009-0591)
3676 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3678 *) Reject UniversalString and BMPString types with invalid lengths. This
3679 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3680 a legal length. (CVE-2009-0590)
3683 *) Set S/MIME signing as the default purpose rather than setting it
3684 unconditionally. This allows applications to override it at the store
3688 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3689 to handle some structures.
3692 *) Improve efficiency of mem_gets: don't search whole buffer each time
3694 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3696 *) New -hex option for openssl rand.
3699 *) Print out UTF8String and NumericString when parsing ASN1.
3702 *) Support NumericString type for name components.
3705 *) Allow CC in the environment to override the automatically chosen
3706 compiler. Note that nothing is done to ensure flags work with the
3710 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3712 *) Properly check EVP_VerifyFinal() and similar return values
3714 [Ben Laurie, Bodo Moeller, Google Security Team]
3716 *) Enable TLS extensions by default.
3719 *) Allow the CHIL engine to be loaded, whether the application is
3720 multithreaded or not. (This does not release the developer from the
3721 obligation to set up the dynamic locking callbacks.)
3722 [Sander Temme <sander@temme.net>]
3724 *) Use correct exit code if there is an error in dgst command.
3725 [Steve Henson; problem pointed out by Roland Dirlewanger]
3727 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3728 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3731 *) Add experimental JPAKE support, including demo authentication in
3732 s_client and s_server.
3735 *) Set the comparison function in v3_addr_canonize().
3736 [Rob Austein <sra@hactrn.net>]
3738 *) Add support for XMPP STARTTLS in s_client.
3739 [Philip Paeps <philip@freebsd.org>]
3741 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3742 to ensure that even with this option, only ciphersuites in the
3743 server's preference list will be accepted. (Note that the option
3744 applies only when resuming a session, so the earlier behavior was
3745 just about the algorithm choice for symmetric cryptography.)
3748 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3750 *) Fix NULL pointer dereference if a DTLS server received
3751 ChangeCipherSpec as first record (CVE-2009-1386).
3754 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3755 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3758 *) The fix in 0.9.8c that supposedly got rid of unsafe
3759 double-checked locking was incomplete for RSA blinding,
3760 addressing just one layer of what turns out to have been
3761 doubly unsafe triple-checked locking.
3763 So now fix this for real by retiring the MONT_HELPER macro
3764 in crypto/rsa/rsa_eay.c.
3766 [Bodo Moeller; problem pointed out by Marius Schilder]
3768 *) Various precautionary measures:
3770 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3772 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3773 (NB: This would require knowledge of the secret session ticket key
3774 to exploit, in which case you'd be SOL either way.)
3776 - Change bn_nist.c so that it will properly handle input BIGNUMs
3777 outside the expected range.
3779 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3782 [Neel Mehta, Bodo Moeller]
3784 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3785 the load fails. Useful for distros.
3786 [Ben Laurie and the FreeBSD team]
3788 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3791 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3794 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3796 This work was sponsored by Logica.
3799 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3800 keystores. Support for SSL/TLS client authentication too.
3801 Not compiled unless enable-capieng specified to Configure.
3803 This work was sponsored by Logica.
3806 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3807 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3808 attribute creation routines such as certifcate requests and PKCS#12
3812 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3814 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3815 handshake which could lead to a cilent crash as found using the
3816 Codenomicon TLS test suite (CVE-2008-1672)
3817 [Steve Henson, Mark Cox]
3819 *) Fix double free in TLS server name extensions which could lead to
3820 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3823 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3825 Clear the error queue to ensure that error entries left from
3826 older function calls do not interfere with the correct operation.
3827 [Lutz Jaenicke, Erik de Castro Lopo]
3829 *) Remove root CA certificates of commercial CAs:
3831 The OpenSSL project does not recommend any specific CA and does not
3832 have any policy with respect to including or excluding any CA.
3833 Therefore it does not make any sense to ship an arbitrary selection
3834 of root CA certificates with the OpenSSL software.
3837 *) RSA OAEP patches to fix two separate invalid memory reads.
3838 The first one involves inputs when 'lzero' is greater than
3839 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3840 before the beginning of from). The second one involves inputs where
3841 the 'db' section contains nothing but zeroes (there is a one-byte
3842 invalid read after the end of 'db').
3843 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3845 *) Partial backport from 0.9.9-dev:
3847 Introduce bn_mul_mont (dedicated Montgomery multiplication
3848 procedure) as a candidate for BIGNUM assembler implementation.
3849 While 0.9.9-dev uses assembler for various architectures, only
3850 x86_64 is available by default here in the 0.9.8 branch, and
3851 32-bit x86 is available through a compile-time setting.
3853 To try the 32-bit x86 assembler implementation, use Configure
3854 option "enable-montasm" (which exists only for this backport).
3856 As "enable-montasm" for 32-bit x86 disclaims code stability
3857 anyway, in this constellation we activate additional code
3858 backported from 0.9.9-dev for further performance improvements,
3859 namely BN_from_montgomery_word. (To enable this otherwise,
3860 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3862 [Andy Polyakov (backport partially by Bodo Moeller)]
3864 *) Add TLS session ticket callback. This allows an application to set
3865 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3866 values. This is useful for key rollover for example where several key
3867 sets may exist with different names.
3870 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3871 This was broken until now in 0.9.8 releases, such that the only way
3872 a registered ENGINE could be used (assuming it initialises
3873 successfully on the host) was to explicitly set it as the default
3874 for the relevant algorithms. This is in contradiction with 0.9.7
3875 behaviour and the documentation. With this fix, when an ENGINE is
3876 registered into a given algorithm's table of implementations, the
3877 'uptodate' flag is reset so that auto-discovery will be used next
3878 time a new context for that algorithm attempts to select an
3880 [Ian Lister (tweaked by Geoff Thorpe)]
3882 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3883 implemention in the following ways:
3885 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3888 Lack of BER streaming support means one pass streaming processing is
3889 only supported if data is detached: setting the streaming flag is
3890 ignored for embedded content.
3892 CMS support is disabled by default and must be explicitly enabled
3893 with the enable-cms configuration option.
3896 *) Update the GMP engine glue to do direct copies between BIGNUM and
3897 mpz_t when openssl and GMP use the same limb size. Otherwise the
3898 existing "conversion via a text string export" trick is still used.
3899 [Paul Sheer <paulsheer@gmail.com>]
3901 *) Zlib compression BIO. This is a filter BIO which compressed and
3902 uncompresses any data passed through it.
3905 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3906 RFC3394 compatible AES key wrapping.
3909 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3910 sets string data without copying. X509_ALGOR_set0() and
3911 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3912 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3913 from an X509_ATTRIBUTE structure optionally checking it occurs only
3914 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3918 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3919 to get the expected BN_FLG_CONSTTIME behavior.
3920 [Bodo Moeller (Google)]
3924 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3925 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3926 - added some more tests to do_tests.pl
3927 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3928 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3929 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3930 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3931 - various changes to netware.pl to enable gcc-cross builds on Win32
3933 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3934 - various changes to fix missing prototype warnings
3935 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3936 - added AES, WHIRLPOOL and CPUID assembler code to build files
3937 - added missing AES assembler make rules to mk1mf.pl
3938 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3939 [Guenter Knauf <eflash@gmx.net>]
3941 *) Implement certificate status request TLS extension defined in RFC3546.
3942 A client can set the appropriate parameters and receive the encoded
3943 OCSP response via a callback. A server can query the supplied parameters
3944 and set the encoded OCSP response in the callback. Add simplified examples
3945 to s_client and s_server.
3948 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3950 *) Fix various bugs:
3951 + Binary incompatibility of ssl_ctx_st structure
3952 + DTLS interoperation with non-compliant servers
3953 + Don't call get_session_cb() without proposed session
3954 + Fix ia64 assembler code
3955 [Andy Polyakov, Steve Henson]
3957 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3959 *) DTLS Handshake overhaul. There were longstanding issues with
3960 OpenSSL DTLS implementation, which were making it impossible for
3961 RFC 4347 compliant client to communicate with OpenSSL server.
3962 Unfortunately just fixing these incompatibilities would "cut off"
3963 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3964 server keeps tolerating non RFC compliant syntax. The opposite is
3965 not true, 0.9.8f client can not communicate with earlier server.
3966 This update even addresses CVE-2007-4995.
3969 *) Changes to avoid need for function casts in OpenSSL: some compilers
3970 (gcc 4.2 and later) reject their use.
3971 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3974 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3975 RFC4507bis. The encrypted ticket format is an encrypted encoded
3976 SSL_SESSION structure, that way new session features are automatically
3979 If a client application caches session in an SSL_SESSION structure
3980 support is transparent because tickets are now stored in the encoded
3983 The SSL_CTX structure automatically generates keys for ticket
3984 protection in servers so again support should be possible
3985 with no application modification.
3987 If a client or server wishes to disable RFC4507 support then the option
3988 SSL_OP_NO_TICKET can be set.
3990 Add a TLS extension debugging callback to allow the contents of any client
3991 or server extensions to be examined.
3993 This work was sponsored by Google.
3996 *) Add initial support for TLS extensions, specifically for the server_name
3997 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3998 have new members for a host name. The SSL data structure has an
3999 additional member SSL_CTX *initial_ctx so that new sessions can be
4000 stored in that context to allow for session resumption, even after the
4001 SSL has been switched to a new SSL_CTX in reaction to a client's
4002 server_name extension.
4004 New functions (subject to change):
4006 SSL_get_servername()
4007 SSL_get_servername_type()
4010 New CTRL codes and macros (subject to change):
4012 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4013 - SSL_CTX_set_tlsext_servername_callback()
4014 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4015 - SSL_CTX_set_tlsext_servername_arg()
4016 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4018 openssl s_client has a new '-servername ...' option.
4020 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4021 '-key2 ...', '-servername_fatal' (subject to change). This allows
4022 testing the HostName extension for a specific single host name ('-cert'
4023 and '-key' remain fallbacks for handshakes without HostName
4024 negotiation). If the unrecognized_name alert has to be sent, this by
4025 default is a warning; it becomes fatal with the '-servername_fatal'
4028 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4030 *) Add AES and SSE2 assembly language support to VC++ build.
4033 *) Mitigate attack on final subtraction in Montgomery reduction.
4036 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4037 (which previously caused an internal error).
4040 *) Squeeze another 10% out of IGE mode when in != out.
4043 *) AES IGE mode speedup.
4044 [Dean Gaudet (Google)]
4046 *) Add the Korean symmetric 128-bit cipher SEED (see
4047 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4048 add SEED ciphersuites from RFC 4162:
4050 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4051 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4052 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4053 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4055 To minimize changes between patchlevels in the OpenSSL 0.9.8
4056 series, SEED remains excluded from compilation unless OpenSSL
4057 is configured with 'enable-seed'.
4058 [KISA, Bodo Moeller]
4060 *) Mitigate branch prediction attacks, which can be practical if a
4061 single processor is shared, allowing a spy process to extract
4062 information. For detailed background information, see
4063 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4064 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4065 and Necessary Software Countermeasures"). The core of the change
4066 are new versions BN_div_no_branch() and
4067 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4068 respectively, which are slower, but avoid the security-relevant
4069 conditional branches. These are automatically called by BN_div()
4070 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4071 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4072 remove a conditional branch.
4074 BN_FLG_CONSTTIME is the new name for the previous
4075 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4076 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4077 in the exponent causes BN_mod_exp_mont() to use the alternative
4078 implementation in BN_mod_exp_mont_consttime().) The old name
4079 remains as a deprecated alias.
4081 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4082 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4083 constant-time implementations for more than just exponentiation.
4084 Here too the old name is kept as a deprecated alias.
4086 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4087 the BN_BLINDING structure gets an independent copy of the
4088 modulus. This means that the previous "BIGNUM *m" argument to
4089 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4090 essentially becomes "const BIGNUM *m", although we can't actually
4091 change this in the header file before 0.9.9. It allows
4092 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4093 enable BN_FLG_CONSTTIME.
4095 [Matthew D Wood (Intel Corp)]
4097 *) In the SSL/TLS server implementation, be strict about session ID
4098 context matching (which matters if an application uses a single
4099 external cache for different purposes). Previously,
4100 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4101 set. This did ensure strict client verification, but meant that,
4102 with applications using a single external cache for quite
4103 different requirements, clients could circumvent ciphersuite
4104 restrictions for a given session ID context by starting a session
4105 in a different context.
4108 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4109 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4110 authentication-only ciphersuites.
4113 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4114 not complete and could lead to a possible single byte overflow
4115 (CVE-2007-5135) [Ben Laurie]
4117 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4119 *) Since AES128 and AES256 (and similarly Camellia128 and
4120 Camellia256) share a single mask bit in the logic of
4121 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4122 kludge to work properly if AES128 is available and AES256 isn't
4123 (or if Camellia128 is available and Camellia256 isn't).
4126 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4127 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4128 When a point or a seed is encoded in a BIT STRING, we need to
4129 prevent the removal of trailing zero bits to get the proper DER
4130 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4131 of a NamedBitList, for which trailing 0 bits need to be removed.)
4134 *) Have SSL/TLS server implementation tolerate "mismatched" record
4135 protocol version while receiving ClientHello even if the
4136 ClientHello is fragmented. (The server can't insist on the
4137 particular protocol version it has chosen before the ServerHello
4138 message has informed the client about his choice.)
4141 *) Add RFC 3779 support.
4142 [Rob Austein for ARIN, Ben Laurie]
4144 *) Load error codes if they are not already present instead of using a
4145 static variable. This allows them to be cleanly unloaded and reloaded.
4146 Improve header file function name parsing.
4149 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4150 or CAPABILITY handshake as required by RFCs.
4153 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4155 *) Introduce limits to prevent malicious keys being able to
4156 cause a denial of service. (CVE-2006-2940)
4157 [Steve Henson, Bodo Moeller]
4159 *) Fix ASN.1 parsing of certain invalid structures that can result
4160 in a denial of service. (CVE-2006-2937) [Steve Henson]
4162 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4163 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4165 *) Fix SSL client code which could crash if connecting to a
4166 malicious SSLv2 server. (CVE-2006-4343)
4167 [Tavis Ormandy and Will Drewry, Google Security Team]
4169 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4170 match only those. Before that, "AES256-SHA" would be interpreted
4171 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4172 the same strength classification in 0.9.7h) as we currently only
4173 have a single AES bit in the ciphersuite description bitmap.
4174 That change, however, also applied to ciphersuite strings such as
4175 "RC4-MD5" that intentionally matched multiple ciphersuites --
4176 namely, SSL 2.0 ciphersuites in addition to the more common ones
4177 from SSL 3.0/TLS 1.0.
4179 So we change the selection algorithm again: Naming an explicit
4180 ciphersuite selects this one ciphersuite, and any other similar
4181 ciphersuite (same bitmap) from *other* protocol versions.
4182 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4183 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4185 Since SSL 2.0 does not have any ciphersuites for which the
4186 128/256 bit distinction would be relevant, this works for now.
4187 The proper fix will be to use different bits for AES128 and
4188 AES256, which would have avoided the problems from the beginning;
4189 however, bits are scarce, so we can only do this in a new release
4190 (not just a patchlevel) when we can change the SSL_CIPHER
4191 definition to split the single 'unsigned long mask' bitmap into
4192 multiple values to extend the available space.
4196 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4198 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4199 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4201 *) Add AES IGE and biIGE modes.
4204 *) Change the Unix randomness entropy gathering to use poll() when
4205 possible instead of select(), since the latter has some
4206 undesirable limitations.
4207 [Darryl Miles via Richard Levitte and Bodo Moeller]
4209 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4210 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4211 cannot be implicitly activated as part of, e.g., the "AES" alias.
4212 However, please upgrade to OpenSSL 0.9.9[-dev] for
4213 non-experimental use of the ECC ciphersuites to get TLS extension
4214 support, which is required for curve and point format negotiation
4215 to avoid potential handshake problems.
4218 *) Disable rogue ciphersuites:
4220 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4221 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4222 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4224 The latter two were purportedly from
4225 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4228 Also deactivate the remaining ciphersuites from
4229 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4230 unofficial, and the ID has long expired.
4233 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4234 dual-core machines) and other potential thread-safety issues.
4237 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4238 versions), which is now available for royalty-free use
4239 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4240 Also, add Camellia TLS ciphersuites from RFC 4132.
4242 To minimize changes between patchlevels in the OpenSSL 0.9.8
4243 series, Camellia remains excluded from compilation unless OpenSSL
4244 is configured with 'enable-camellia'.
4247 *) Disable the padding bug check when compression is in use. The padding
4248 bug check assumes the first packet is of even length, this is not
4249 necessarily true if compresssion is enabled and can result in false
4250 positives causing handshake failure. The actual bug test is ancient
4251 code so it is hoped that implementations will either have fixed it by
4252 now or any which still have the bug do not support compression.
4255 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4257 *) When applying a cipher rule check to see if string match is an explicit
4258 cipher suite and only match that one cipher suite if it is.
4261 *) Link in manifests for VC++ if needed.
4262 [Austin Ziegler <halostatue@gmail.com>]
4264 *) Update support for ECC-based TLS ciphersuites according to
4265 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4266 TLS extensions, which are supported starting with the 0.9.9
4267 branch, not in the OpenSSL 0.9.8 branch).
4270 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4271 opaque EVP_CIPHER_CTX handling.
4274 *) Fixes and enhancements to zlib compression code. We now only use
4275 "zlib1.dll" and use the default __cdecl calling convention on Win32
4276 to conform with the standards mentioned here:
4277 http://www.zlib.net/DLL_FAQ.txt
4278 Static zlib linking now works on Windows and the new --with-zlib-include
4279 --with-zlib-lib options to Configure can be used to supply the location
4280 of the headers and library. Gracefully handle case where zlib library
4284 *) Several fixes and enhancements to the OID generation code. The old code
4285 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4286 handle numbers larger than ULONG_MAX, truncated printing and had a
4287 non standard OBJ_obj2txt() behaviour.
4290 *) Add support for building of engines under engine/ as shared libraries
4291 under VC++ build system.
4294 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4295 Hopefully, we will not see any false combination of paths any more.
4298 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4300 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4301 (part of SSL_OP_ALL). This option used to disable the
4302 countermeasure against man-in-the-middle protocol-version
4303 rollback in the SSL 2.0 server implementation, which is a bad
4304 idea. (CVE-2005-2969)
4306 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4307 for Information Security, National Institute of Advanced Industrial
4308 Science and Technology [AIST], Japan)]
4310 *) Add two function to clear and return the verify parameter flags.
4313 *) Keep cipherlists sorted in the source instead of sorting them at
4314 runtime, thus removing the need for a lock.
4317 *) Avoid some small subgroup attacks in Diffie-Hellman.
4318 [Nick Mathewson and Ben Laurie]
4320 *) Add functions for well-known primes.
4323 *) Extended Windows CE support.
4324 [Satoshi Nakamura and Andy Polyakov]
4326 *) Initialize SSL_METHOD structures at compile time instead of during
4327 runtime, thus removing the need for a lock.
4330 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4331 attempting to decrypt each encrypted key in turn. Add support to
4335 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4337 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4340 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4343 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4344 key into the same file any more.
4347 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4350 *) Add -utf8 command line and config file option to 'ca'.
4351 [Stefan <stf@udoma.org]
4353 *) Removed the macro des_crypt(), as it seems to conflict with some
4354 libraries. Use DES_crypt().
4357 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4358 involves renaming the source and generated shared-libs for
4359 both. The engines will accept the corrected or legacy ids
4360 ('ncipher' and '4758_cca' respectively) when binding. NB,
4361 this only applies when building 'shared'.
4362 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4364 *) Add attribute functions to EVP_PKEY structure. Modify
4365 PKCS12_create() to recognize a CSP name attribute and
4366 use it. Make -CSP option work again in pkcs12 utility.
4369 *) Add new functionality to the bn blinding code:
4370 - automatic re-creation of the BN_BLINDING parameters after
4371 a fixed number of uses (currently 32)
4372 - add new function for parameter creation
4373 - introduce flags to control the update behaviour of the
4374 BN_BLINDING parameters
4375 - hide BN_BLINDING structure
4376 Add a second BN_BLINDING slot to the RSA structure to improve
4377 performance when a single RSA object is shared among several
4381 *) Add support for DTLS.
4382 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4384 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4385 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4388 *) Remove buggy and incomplete DH cert support from
4389 ssl/ssl_rsa.c and ssl/s3_both.c
4392 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4393 the apps/openssl applications.
4396 *) Compile clean with "-Wall -Wmissing-prototypes
4397 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4398 DEBUG_SAFESTACK must also be set.
4401 *) Change ./Configure so that certain algorithms can be disabled by default.
4402 The new counterpiece to "no-xxx" is "enable-xxx".
4404 The patented RC5 and MDC2 algorithms will now be disabled unless
4405 "enable-rc5" and "enable-mdc2", respectively, are specified.
4407 (IDEA remains enabled despite being patented. This is because IDEA
4408 is frequently required for interoperability, and there is no license
4409 fee for non-commercial use. As before, "no-idea" can be used to
4410 avoid this algorithm.)
4414 *) Add processing of proxy certificates (see RFC 3820). This work was
4415 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4416 EGEE (Enabling Grids for E-science in Europe).
4419 *) RC4 performance overhaul on modern architectures/implementations, such
4420 as Intel P4, IA-64 and AMD64.
4423 *) New utility extract-section.pl. This can be used specify an alternative
4424 section number in a pod file instead of having to treat each file as
4425 a separate case in Makefile. This can be done by adding two lines to the
4428 =for comment openssl_section:XXX
4430 The blank line is mandatory.
4434 *) New arguments -certform, -keyform and -pass for s_client and s_server
4435 to allow alternative format key and certificate files and passphrase
4439 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4440 update associated structures and add various utility functions.
4442 Add new policy related verify parameters, include policy checking in
4443 standard verify code. Enhance 'smime' application with extra parameters
4444 to support policy checking and print out.
4447 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4448 Nehemiah processors. These extensions support AES encryption in hardware
4449 as well as RNG (though RNG support is currently disabled).
4450 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4452 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4455 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4456 [Andy Polyakov and a number of other people]
4458 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4459 implementation contributed by IBM.
4460 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4462 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4463 exponent rather than 'unsigned long'. There is a corresponding change to
4464 the new 'rsa_keygen' element of the RSA_METHOD structure.
4465 [Jelte Jansen, Geoff Thorpe]
4467 *) Functionality for creating the initial serial number file is now
4468 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4470 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4471 number file to 1, which is bound to cause problems. To avoid
4472 the problems while respecting compatibility between different 0.9.7
4473 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4474 CA.pl for serial number initialization. With the new release 0.9.8,
4475 we can fix the problem directly in the 'ca' utility.)
4478 *) Reduced header interdepencies by declaring more opaque objects in
4479 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4480 give fewer recursive includes, which could break lazy source code - so
4481 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4482 developers should define this symbol when building and using openssl to
4483 ensure they track the recommended behaviour, interfaces, [etc], but
4484 backwards-compatible behaviour prevails when this isn't defined.
4487 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4490 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4491 This will generate a random key of the appropriate length based on the
4492 cipher context. The EVP_CIPHER can provide its own random key generation
4493 routine to support keys of a specific form. This is used in the des and
4494 3des routines to generate a key of the correct parity. Update S/MIME
4495 code to use new functions and hence generate correct parity DES keys.
4496 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4497 valid (weak or incorrect parity).
4500 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4501 as looking them up. This is useful when the verified structure may contain
4502 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4503 present unless the new PKCS7_NO_CRL flag is asserted.
4506 *) Extend ASN1 oid configuration module. It now additionally accepts the
4509 shortName = some long name, 1.2.3.4
4512 *) Reimplemented the BN_CTX implementation. There is now no more static
4513 limitation on the number of variables it can handle nor the depth of the
4514 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4515 information can now expand as required, and rather than having a single
4516 static array of bignums, BN_CTX now uses a linked-list of such arrays
4517 allowing it to expand on demand whilst maintaining the usefulness of
4518 BN_CTX's "bundling".
4521 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4522 to allow all RSA operations to function using a single BN_CTX.
4525 *) Preliminary support for certificate policy evaluation and checking. This
4526 is initially intended to pass the tests outlined in "Conformance Testing
4527 of Relying Party Client Certificate Path Processing Logic" v1.07.
4530 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4531 remained unused and not that useful. A variety of other little bignum
4532 tweaks and fixes have also been made continuing on from the audit (see
4536 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4537 associated ASN1, EVP and SSL functions and old ASN1 macros.
4540 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4541 and this should never fail. So the return value from the use of
4542 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4543 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4546 *) BN_CTX_get() should return zero-valued bignums, providing the same
4547 initialised value as BN_new().
4548 [Geoff Thorpe, suggested by Ulf Möller]
4550 *) Support for inhibitAnyPolicy certificate extension.
4553 *) An audit of the BIGNUM code is underway, for which debugging code is
4554 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4555 is considered valid when processing BIGNUMs, and causes execution to
4556 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4557 further steps are taken to deliberately pollute unused data in BIGNUM
4558 structures to try and expose faulty code further on. For now, openssl will
4559 (in its default mode of operation) continue to tolerate the inconsistent
4560 forms that it has tolerated in the past, but authors and packagers should
4561 consider trying openssl and their own applications when compiled with
4562 these debugging symbols defined. It will help highlight potential bugs in
4563 their own code, and will improve the test coverage for OpenSSL itself. At
4564 some point, these tighter rules will become openssl's default to improve
4565 maintainability, though the assert()s and other overheads will remain only
4566 in debugging configurations. See bn.h for more details.
4567 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4569 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4570 that can only be obtained through BN_CTX_new() (which implicitly
4571 initialises it). The presence of this function only made it possible
4572 to overwrite an existing structure (and cause memory leaks).
4575 *) Because of the callback-based approach for implementing LHASH as a
4576 template type, lh_insert() adds opaque objects to hash-tables and
4577 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4578 to clean up those corresponding objects before destroying the hash table
4579 (and losing the object pointers). So some over-zealous constifications in
4580 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4581 objects as "const" and the lh_doall[_arg] callback wrappers are not
4582 prototyped to have "const" restrictions on the object pointers they are
4583 given (and so aren't required to cast them away any more).
4586 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4587 (speed) prefers to use its own implementation. The two implementations
4588 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4589 its object type properly exposed (MS_TM) instead of casting to/from "char
4590 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4591 aren't necessarily the greatest nomenclatures - but this is what was used
4592 internally to the implementation so I've used that for now.
4595 *) Ensure that deprecated functions do not get compiled when
4596 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4597 the self-tests were still using deprecated key-generation functions so
4598 these have been updated also.
4601 *) Reorganise PKCS#7 code to separate the digest location functionality
4602 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4603 New function PKCS7_set_digest() to set the digest type for PKCS#7
4604 digestedData type. Add additional code to correctly generate the
4605 digestedData type and add support for this type in PKCS7 initialization
4609 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4610 structure of type "other".
4613 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4614 sure the loop does correctly stop and breaking ("division by zero")
4615 modulus operations are not performed. The (pre-generated) prime
4616 table crypto/bn/bn_prime.h was already correct, but it could not be
4617 re-generated on some platforms because of the "division by zero"
4618 situation in the script.
4619 [Ralf S. Engelschall]
4621 *) Update support for ECC-based TLS ciphersuites according to
4622 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4623 SHA-1 now is only used for "small" curves (where the
4624 representation of a field element takes up to 24 bytes); for
4625 larger curves, the field element resulting from ECDH is directly
4626 used as premaster secret.
4627 [Douglas Stebila (Sun Microsystems Laboratories)]
4629 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4630 curve secp160r1 to the tests.
4631 [Douglas Stebila (Sun Microsystems Laboratories)]
4633 *) Add the possibility to load symbols globally with DSO.
4634 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4636 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4637 control of the error stack.
4640 *) Add support for STORE in ENGINE.
4643 *) Add the STORE type. The intention is to provide a common interface
4644 to certificate and key stores, be they simple file-based stores, or
4645 HSM-type store, or LDAP stores, or...
4646 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4649 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4650 pass a list of arguments to any function as well as provide a way
4651 for a function to pass data back to the caller.
4654 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4655 works like BUF_strdup() but can be used to duplicate a portion of
4656 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4660 *) Add the function sk_find_ex() which works like sk_find(), but will
4661 return an index to an element even if an exact match couldn't be
4662 found. The index is guaranteed to point at the element where the
4663 searched-for key would be inserted to preserve sorting order.
4666 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4667 takes an extra flags argument for optional functionality. Currently,
4668 the following flags are defined:
4670 OBJ_BSEARCH_VALUE_ON_NOMATCH
4671 This one gets OBJ_bsearch_ex() to return a pointer to the first
4672 element where the comparing function returns a negative or zero
4675 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4676 This one gets OBJ_bsearch_ex() to return a pointer to the first
4677 element where the comparing function returns zero. This is useful
4678 if there are more than one element where the comparing function
4682 *) Make it possible to create self-signed certificates with 'openssl ca'
4683 in such a way that the self-signed certificate becomes part of the
4684 CA database and uses the same mechanisms for serial number generation
4685 as all other certificate signing. The new flag '-selfsign' enables
4686 this functionality. Adapt CA.sh and CA.pl.in.
4689 *) Add functionality to check the public key of a certificate request
4690 against a given private. This is useful to check that a certificate
4691 request can be signed by that key (self-signing).
4694 *) Make it possible to have multiple active certificates with the same
4695 subject in the CA index file. This is done only if the keyword
4696 'unique_subject' is set to 'no' in the main CA section (default
4697 if 'CA_default') of the configuration file. The value is saved
4698 with the database itself in a separate index attribute file,
4699 named like the index file with '.attr' appended to the name.
4702 *) Generate muti valued AVAs using '+' notation in config files for
4706 *) Support for nameConstraints certificate extension.
4709 *) Support for policyConstraints certificate extension.
4712 *) Support for policyMappings certificate extension.
4715 *) Make sure the default DSA_METHOD implementation only uses its
4716 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4717 and change its own handlers to be NULL so as to remove unnecessary
4718 indirection. This lets alternative implementations fallback to the
4719 default implementation more easily.
4722 *) Support for directoryName in GeneralName related extensions
4726 *) Make it possible to link applications using Makefile.shared.
4727 Make that possible even when linking against static libraries!
4730 *) Support for single pass processing for S/MIME signing. This now
4731 means that S/MIME signing can be done from a pipe, in addition
4732 cleartext signing (multipart/signed type) is effectively streaming
4733 and the signed data does not need to be all held in memory.
4735 This is done with a new flag PKCS7_STREAM. When this flag is set
4736 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4737 is done after the data is output (and digests calculated) in
4738 SMIME_write_PKCS7().
4741 *) Add full support for -rpath/-R, both in shared libraries and
4742 applications, at least on the platforms where it's known how
4746 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4747 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4748 will now compute a table of multiples of the generator that
4749 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4750 faster (notably in the case of a single point multiplication,
4751 scalar * generator).
4752 [Nils Larsch, Bodo Moeller]
4754 *) IPv6 support for certificate extensions. The various extensions
4755 which use the IP:a.b.c.d can now take IPv6 addresses using the
4756 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4760 *) Added an ENGINE that implements RSA by performing private key
4761 exponentiations with the GMP library. The conversions to and from
4762 GMP's mpz_t format aren't optimised nor are any montgomery forms
4763 cached, and on x86 it appears OpenSSL's own performance has caught up.
4764 However there are likely to be other architectures where GMP could
4765 provide a boost. This ENGINE is not built in by default, but it can be
4766 specified at Configure time and should be accompanied by the necessary
4767 linker additions, eg;
4768 ./config -DOPENSSL_USE_GMP -lgmp
4771 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4772 testing availability of engines with "-t" - the old behaviour is
4773 produced by increasing the feature's verbosity with "-tt".
4776 *) ECDSA routines: under certain error conditions uninitialized BN objects
4777 could be freed. Solution: make sure initialization is performed early
4778 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4782 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4783 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4784 software implementations. For DSA and DH, parameter generation can
4785 also be overriden by providing the appropriate method callbacks.
4788 *) Change the "progress" mechanism used in key-generation and
4789 primality testing to functions that take a new BN_GENCB pointer in
4790 place of callback/argument pairs. The new API functions have "_ex"
4791 postfixes and the older functions are reimplemented as wrappers for
4792 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4793 declarations of the old functions to help (graceful) attempts to
4794 migrate to the new functions. Also, the new key-generation API
4795 functions operate on a caller-supplied key-structure and return
4796 success/failure rather than returning a key or NULL - this is to
4797 help make "keygen" another member function of RSA_METHOD etc.
4799 Example for using the new callback interface:
4801 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4805 BN_GENCB_set(&my_cb, my_callback, my_arg);
4807 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4808 /* For the meaning of a, b in calls to my_callback(), see the
4809 * documentation of the function that calls the callback.
4810 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4811 * my_callback should return 1 if it wants BN_is_prime_ex()
4812 * to continue, or 0 to stop.
4817 *) Change the ZLIB compression method to be stateful, and make it
4818 available to TLS with the number defined in
4819 draft-ietf-tls-compression-04.txt.
4822 *) Add the ASN.1 structures and functions for CertificatePair, which
4823 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4825 CertificatePair ::= SEQUENCE {
4826 forward [0] Certificate OPTIONAL,
4827 reverse [1] Certificate OPTIONAL,
4828 -- at least one of the pair shall be present -- }
4830 Also implement the PEM functions to read and write certificate
4831 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4833 This needed to be defined, mostly for the sake of the LDAP
4834 attribute crossCertificatePair, but may prove useful elsewhere as
4838 *) Make it possible to inhibit symlinking of shared libraries in
4839 Makefile.shared, for Cygwin's sake.
4842 *) Extend the BIGNUM API by creating a function
4843 void BN_set_negative(BIGNUM *a, int neg);
4844 and a macro that behave like
4845 int BN_is_negative(const BIGNUM *a);
4847 to avoid the need to access 'a->neg' directly in applications.
4850 *) Implement fast modular reduction for pseudo-Mersenne primes
4851 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4852 EC_GROUP_new_curve_GFp() will now automatically use this
4854 [Nils Larsch <nla@trustcenter.de>]
4856 *) Add new lock type (CRYPTO_LOCK_BN).
4859 *) Change the ENGINE framework to automatically load engines
4860 dynamically from specific directories unless they could be
4861 found to already be built in or loaded. Move all the
4862 current engines except for the cryptodev one to a new
4864 The engines in engines/ are built as shared libraries if
4865 the "shared" options was given to ./Configure or ./config.
4866 Otherwise, they are inserted in libcrypto.a.
4867 /usr/local/ssl/engines is the default directory for dynamic
4868 engines, but that can be overriden at configure time through
4869 the usual use of --prefix and/or --openssldir, and at run
4870 time with the environment variable OPENSSL_ENGINES.
4871 [Geoff Thorpe and Richard Levitte]
4873 *) Add Makefile.shared, a helper makefile to build shared
4874 libraries. Adapt Makefile.org.
4877 *) Add version info to Win32 DLLs.
4878 [Peter 'Luna' Runestig" <peter@runestig.com>]
4880 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4881 can be added using this API to created arbitrary PKCS#12
4882 files while avoiding the low level API.
4884 New options to PKCS12_create(), key or cert can be NULL and
4885 will then be omitted from the output file. The encryption
4886 algorithm NIDs can be set to -1 for no encryption, the mac
4887 iteration count can be set to 0 to omit the mac.
4889 Enhance pkcs12 utility by making the -nokeys and -nocerts
4890 options work when creating a PKCS#12 file. New option -nomac
4891 to omit the mac, NONE can be set for an encryption algorithm.
4892 New code is modified to use the enhanced PKCS12_create()
4893 instead of the low level API.
4896 *) Extend ASN1 encoder to support indefinite length constructed
4897 encoding. This can output sequences tags and octet strings in
4898 this form. Modify pk7_asn1.c to support indefinite length
4899 encoding. This is experimental and needs additional code to
4900 be useful, such as an ASN1 bio and some enhanced streaming
4903 Extend template encode functionality so that tagging is passed
4904 down to the template encoder.
4907 *) Let 'openssl req' fail if an argument to '-newkey' is not
4908 recognized instead of using RSA as a default.
4911 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4912 As these are not official, they are not included in "ALL";
4913 the "ECCdraft" ciphersuite group alias can be used to select them.
4914 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4916 *) Add ECDH engine support.
4917 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4919 *) Add ECDH in new directory crypto/ecdh/.
4920 [Douglas Stebila (Sun Microsystems Laboratories)]
4922 *) Let BN_rand_range() abort with an error after 100 iterations
4923 without success (which indicates a broken PRNG).
4926 *) Change BN_mod_sqrt() so that it verifies that the input value
4927 is really the square of the return value. (Previously,
4928 BN_mod_sqrt would show GIGO behaviour.)
4931 *) Add named elliptic curves over binary fields from X9.62, SECG,
4932 and WAP/WTLS; add OIDs that were still missing.
4934 [Sheueling Chang Shantz and Douglas Stebila
4935 (Sun Microsystems Laboratories)]
4937 *) Extend the EC library for elliptic curves over binary fields
4938 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4941 EC_GF2m_simple_method
4945 EC_GROUP_new_curve_GF2m
4946 EC_GROUP_set_curve_GF2m
4947 EC_GROUP_get_curve_GF2m
4948 EC_POINT_set_affine_coordinates_GF2m
4949 EC_POINT_get_affine_coordinates_GF2m
4950 EC_POINT_set_compressed_coordinates_GF2m
4952 Point compression for binary fields is disabled by default for
4953 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4956 As binary polynomials are represented as BIGNUMs, various members
4957 of the EC_GROUP and EC_POINT data structures can be shared
4958 between the implementations for prime fields and binary fields;
4959 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4960 are essentially identical to their ..._GFp counterparts.
4961 (For simplicity, the '..._GFp' prefix has been dropped from
4962 various internal method names.)
4964 An internal 'field_div' method (similar to 'field_mul' and
4965 'field_sqr') has been added; this is used only for binary fields.
4967 [Sheueling Chang Shantz and Douglas Stebila
4968 (Sun Microsystems Laboratories)]
4970 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4971 through methods ('mul', 'precompute_mult').
4973 The generic implementations (now internally called 'ec_wNAF_mul'
4974 and 'ec_wNAF_precomputed_mult') remain the default if these
4975 methods are undefined.
4977 [Sheueling Chang Shantz and Douglas Stebila
4978 (Sun Microsystems Laboratories)]
4980 *) New function EC_GROUP_get_degree, which is defined through
4981 EC_METHOD. For curves over prime fields, this returns the bit
4982 length of the modulus.
4984 [Sheueling Chang Shantz and Douglas Stebila
4985 (Sun Microsystems Laboratories)]
4987 *) New functions EC_GROUP_dup, EC_POINT_dup.
4988 (These simply call ..._new and ..._copy).
4990 [Sheueling Chang Shantz and Douglas Stebila
4991 (Sun Microsystems Laboratories)]
4993 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4994 Polynomials are represented as BIGNUMs (where the sign bit is not
4995 used) in the following functions [macros]:
4998 BN_GF2m_sub [= BN_GF2m_add]
4999 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5000 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5001 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5003 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5004 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5005 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5006 BN_GF2m_cmp [= BN_ucmp]
5008 (Note that only the 'mod' functions are actually for fields GF(2^m).
5009 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5011 For some functions, an the irreducible polynomial defining a
5012 field can be given as an 'unsigned int[]' with strictly
5013 decreasing elements giving the indices of those bits that are set;
5014 i.e., p[] represents the polynomial
5015 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5017 p[0] > p[1] > ... > p[k] = 0.
5018 This applies to the following functions:
5023 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5024 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5026 BN_GF2m_mod_sqrt_arr
5027 BN_GF2m_mod_solve_quad_arr
5031 Conversion can be performed by the following functions:
5036 bntest.c has additional tests for binary polynomial arithmetic.
5038 Two implementations for BN_GF2m_mod_div() are available.
5039 The default algorithm simply uses BN_GF2m_mod_inv() and
5040 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5041 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5042 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5044 [Sheueling Chang Shantz and Douglas Stebila
5045 (Sun Microsystems Laboratories)]
5047 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5048 functionality is disabled at compile-time.
5049 [Douglas Stebila <douglas.stebila@sun.com>]
5051 *) Change default behaviour of 'openssl asn1parse' so that more
5052 information is visible when viewing, e.g., a certificate:
5054 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5055 mode the content of non-printable OCTET STRINGs is output in a
5056 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5057 avoid the appearance of a printable string.
5058 [Nils Larsch <nla@trustcenter.de>]
5060 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5062 EC_GROUP_set_asn1_flag()
5063 EC_GROUP_get_asn1_flag()
5064 EC_GROUP_set_point_conversion_form()
5065 EC_GROUP_get_point_conversion_form()
5066 These control ASN1 encoding details:
5067 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5068 has been set to OPENSSL_EC_NAMED_CURVE.
5069 - Points are encoded in uncompressed form by default; options for
5070 asn1_for are as for point2oct, namely
5071 POINT_CONVERSION_COMPRESSED
5072 POINT_CONVERSION_UNCOMPRESSED
5073 POINT_CONVERSION_HYBRID
5075 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5078 EC_GROUP_get0_seed()
5079 EC_GROUP_get_seed_len()
5080 This is used only for ASN1 purposes (so far).
5081 [Nils Larsch <nla@trustcenter.de>]
5083 *) Add 'field_type' member to EC_METHOD, which holds the NID
5084 of the appropriate field type OID. The new function
5085 EC_METHOD_get_field_type() returns this value.
5086 [Nils Larsch <nla@trustcenter.de>]
5091 EC_POINT_point2hex()
5092 EC_POINT_hex2point()
5093 providing useful interfaces to EC_POINT_point2oct() and
5094 EC_POINT_oct2point().
5095 [Nils Larsch <nla@trustcenter.de>]
5097 *) Change internals of the EC library so that the functions
5098 EC_GROUP_set_generator()
5099 EC_GROUP_get_generator()
5100 EC_GROUP_get_order()
5101 EC_GROUP_get_cofactor()
5102 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5103 to methods, which would lead to unnecessary code duplication when
5104 adding different types of curves.
5105 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5107 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5108 arithmetic, and such that modified wNAFs are generated
5109 (which avoid length expansion in many cases).
5112 *) Add a function EC_GROUP_check_discriminant() (defined via
5113 EC_METHOD) that verifies that the curve discriminant is non-zero.
5115 Add a function EC_GROUP_check() that makes some sanity tests
5116 on a EC_GROUP, its generator and order. This includes
5117 EC_GROUP_check_discriminant().
5118 [Nils Larsch <nla@trustcenter.de>]
5120 *) Add ECDSA in new directory crypto/ecdsa/.
5122 Add applications 'openssl ecparam' and 'openssl ecdsa'
5123 (these are based on 'openssl dsaparam' and 'openssl dsa').
5125 ECDSA support is also included in various other files across the
5126 library. Most notably,
5127 - 'openssl req' now has a '-newkey ecdsa:file' option;
5128 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5129 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5130 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5131 them suitable for ECDSA where domain parameters must be
5132 extracted before the specific public key;
5133 - ECDSA engine support has been added.
5134 [Nils Larsch <nla@trustcenter.de>]
5136 *) Include some named elliptic curves, and add OIDs from X9.62,
5137 SECG, and WAP/WTLS. Each curve can be obtained from the new
5139 EC_GROUP_new_by_curve_name(),
5140 and the list of available named curves can be obtained with
5141 EC_get_builtin_curves().
5142 Also add a 'curve_name' member to EC_GROUP objects, which can be
5144 EC_GROUP_set_curve_name()
5145 EC_GROUP_get_curve_name()
5146 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5148 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5149 was actually never needed) and in BN_mul(). The removal in BN_mul()
5150 required a small change in bn_mul_part_recursive() and the addition
5151 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5152 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5153 bn_sub_words() and bn_add_words() except they take arrays with
5157 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5159 *) Cleanse PEM buffers before freeing them since they may contain
5161 [Benjamin Bennett <ben@psc.edu>]
5163 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5164 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5165 authentication-only ciphersuites.
5168 *) Since AES128 and AES256 share a single mask bit in the logic of
5169 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5170 kludge to work properly if AES128 is available and AES256 isn't.
5173 *) Expand security boundary to match 1.1.1 module.
5176 *) Remove redundant features: hash file source, editing of test vectors
5177 modify fipsld to use external fips_premain.c signature.
5180 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5181 run algorithm test programs.
5184 *) Make algorithm test programs more tolerant of whitespace.
5187 *) Have SSL/TLS server implementation tolerate "mismatched" record
5188 protocol version while receiving ClientHello even if the
5189 ClientHello is fragmented. (The server can't insist on the
5190 particular protocol version it has chosen before the ServerHello
5191 message has informed the client about his choice.)
5194 *) Load error codes if they are not already present instead of using a
5195 static variable. This allows them to be cleanly unloaded and reloaded.
5198 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5200 *) Introduce limits to prevent malicious keys being able to
5201 cause a denial of service. (CVE-2006-2940)
5202 [Steve Henson, Bodo Moeller]
5204 *) Fix ASN.1 parsing of certain invalid structures that can result
5205 in a denial of service. (CVE-2006-2937) [Steve Henson]
5207 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5208 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5210 *) Fix SSL client code which could crash if connecting to a
5211 malicious SSLv2 server. (CVE-2006-4343)
5212 [Tavis Ormandy and Will Drewry, Google Security Team]
5214 *) Change ciphersuite string processing so that an explicit
5215 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5216 will no longer include "AES128-SHA"), and any other similar
5217 ciphersuite (same bitmap) from *other* protocol versions (so that
5218 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5219 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5220 changes from 0.9.8b and 0.9.8d.
5223 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5225 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5226 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5228 *) Change the Unix randomness entropy gathering to use poll() when
5229 possible instead of select(), since the latter has some
5230 undesirable limitations.
5231 [Darryl Miles via Richard Levitte and Bodo Moeller]
5233 *) Disable rogue ciphersuites:
5235 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5236 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5237 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5239 The latter two were purportedly from
5240 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5243 Also deactive the remaining ciphersuites from
5244 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5245 unofficial, and the ID has long expired.
5248 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5249 dual-core machines) and other potential thread-safety issues.
5252 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5254 *) Adapt fipsld and the build system to link against the validated FIPS
5255 module in FIPS mode.
5258 *) Fixes for VC++ 2005 build under Windows.
5261 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5262 from a Windows bash shell such as MSYS. It is autodetected from the
5263 "config" script when run from a VC++ environment. Modify standard VC++
5264 build to use fipscanister.o from the GNU make build.
5267 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5269 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5270 The value now differs depending on if you build for FIPS or not.
5271 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5272 safely run with a non-FIPSed libcrypto, as it may crash because of
5273 the difference induced by this change.
5276 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5278 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5279 (part of SSL_OP_ALL). This option used to disable the
5280 countermeasure against man-in-the-middle protocol-version
5281 rollback in the SSL 2.0 server implementation, which is a bad
5282 idea. (CVE-2005-2969)
5284 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5285 for Information Security, National Institute of Advanced Industrial
5286 Science and Technology [AIST], Japan)]
5288 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5289 mainly for FIPS compliance and not fully integrated at this stage.
5292 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5293 the exponentiation using a fixed-length exponent. (Otherwise,
5294 the information leaked through timing could expose the secret key
5295 after many signatures; cf. Bleichenbacher's attack on DSA with
5299 *) Make a new fixed-window mod_exp implementation the default for
5300 RSA, DSA, and DH private-key operations so that the sequence of
5301 squares and multiplies and the memory access pattern are
5302 independent of the particular secret key. This will mitigate
5303 cache-timing and potential related attacks.
5305 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5306 and this is automatically used by BN_mod_exp_mont() if the new flag
5307 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5308 will use this BN flag for private exponents unless the flag
5309 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5310 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5312 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5314 *) Change the client implementation for SSLv23_method() and
5315 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5316 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5317 (Previously, the SSL 2.0 backwards compatible Client Hello
5318 message format would be used even with SSL_OP_NO_SSLv2.)
5321 *) Add support for smime-type MIME parameter in S/MIME messages which some
5325 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5326 a threadsafe manner. Modify rsa code to use new function and add calls
5327 to dsa and dh code (which had race conditions before).
5330 *) Include the fixed error library code in the C error file definitions
5331 instead of fixing them up at runtime. This keeps the error code
5332 structures constant.
5335 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5337 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5340 *) Fixes for newer kerberos headers. NB: the casts are needed because
5341 the 'length' field is signed on one version and unsigned on another
5342 with no (?) obvious way to tell the difference, without these VC++
5343 complains. Also the "definition" of FAR (blank) is no longer included
5344 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5345 some needed definitions.
5348 *) Undo Cygwin change.
5351 *) Added support for proxy certificates according to RFC 3820.
5352 Because they may be a security thread to unaware applications,
5353 they must be explicitly allowed in run-time. See
5354 docs/HOWTO/proxy_certificates.txt for further information.
5357 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5359 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5360 server and client random values. Previously
5361 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5362 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5364 This change has negligible security impact because:
5366 1. Server and client random values still have 24 bytes of pseudo random
5369 2. Server and client random values are sent in the clear in the initial
5372 3. The master secret is derived using the premaster secret (48 bytes in
5373 size for static RSA ciphersuites) as well as client server and random
5376 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5379 [Stephen Henson, reported by UK NISCC]
5381 *) Use Windows randomness collection on Cygwin.
5384 *) Fix hang in EGD/PRNGD query when communication socket is closed
5385 prematurely by EGD/PRNGD.
5386 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5388 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5391 *) Back-port of selected performance improvements from development
5392 branch, as well as improved support for PowerPC platforms.
5395 *) Add lots of checks for memory allocation failure, error codes to indicate
5396 failure and freeing up memory if a failure occurs.
5397 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5399 *) Add new -passin argument to dgst.
5402 *) Perform some character comparisons of different types in X509_NAME_cmp:
5403 this is needed for some certificates that reencode DNs into UTF8Strings
5404 (in violation of RFC3280) and can't or wont issue name rollover
5408 *) Make an explicit check during certificate validation to see that
5409 the CA setting in each certificate on the chain is correct. As a
5410 side effect always do the following basic checks on extensions,
5411 not just when there's an associated purpose to the check:
5413 - if there is an unhandled critical extension (unless the user
5414 has chosen to ignore this fault)
5415 - if the path length has been exceeded (if one is set at all)
5416 - that certain extensions fit the associated purpose (if one has
5420 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5422 *) Avoid a race condition when CRLs are checked in a multi threaded
5423 environment. This would happen due to the reordering of the revoked
5424 entries during signature checking and serial number lookup. Now the
5425 encoding is cached and the serial number sort performed under a lock.
5426 Add new STACK function sk_is_sorted().
5429 *) Add Delta CRL to the extension code.
5432 *) Various fixes to s3_pkt.c so alerts are sent properly.
5433 [David Holmes <d.holmes@f5.com>]
5435 *) Reduce the chances of duplicate issuer name and serial numbers (in
5436 violation of RFC3280) using the OpenSSL certificate creation utilities.
5437 This is done by creating a random 64 bit value for the initial serial
5438 number when a serial number file is created or when a self signed
5439 certificate is created using 'openssl req -x509'. The initial serial
5440 number file is created using 'openssl x509 -next_serial' in CA.pl
5441 rather than being initialized to 1.
5444 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5446 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5447 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5448 [Joe Orton, Steve Henson]
5450 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5452 [Joe Orton, Steve Henson]
5454 *) Make it possible to have multiple active certificates with the same
5455 subject in the CA index file. This is done only if the keyword
5456 'unique_subject' is set to 'no' in the main CA section (default
5457 if 'CA_default') of the configuration file. The value is saved
5458 with the database itself in a separate index attribute file,
5459 named like the index file with '.attr' appended to the name.
5462 *) X509 verify fixes. Disable broken certificate workarounds when
5463 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5464 keyUsage extension present. Don't accept CRLs with unhandled critical
5465 extensions: since verify currently doesn't process CRL extensions this
5466 rejects a CRL with *any* critical extensions. Add new verify error codes
5470 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5471 A clarification of RFC2560 will require the use of OCTET STRINGs and
5472 some implementations cannot handle the current raw format. Since OpenSSL
5473 copies and compares OCSP nonces as opaque blobs without any attempt at
5474 parsing them this should not create any compatibility issues.
5477 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5478 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5479 this HMAC (and other) operations are several times slower than OpenSSL
5483 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5484 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5486 *) Use the correct content when signing type "other".
5489 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5491 *) Fix various bugs revealed by running the NISCC test suite:
5493 Stop out of bounds reads in the ASN1 code when presented with
5494 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5496 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5498 If verify callback ignores invalid public key errors don't try to check
5499 certificate signature with the NULL public key.
5503 *) New -ignore_err option in ocsp application to stop the server
5504 exiting on the first error in a request.
5507 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5508 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5512 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5513 extra data after the compression methods not only for TLS 1.0
5514 but also for SSL 3.0 (as required by the specification).
5515 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5517 *) Change X509_certificate_type() to mark the key as exported/exportable
5518 when it's 512 *bits* long, not 512 bytes.
5521 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5522 blocks during encryption.
5525 *) Various fixes to base64 BIO and non blocking I/O. On write
5526 flushes were not handled properly if the BIO retried. On read
5527 data was not being buffered properly and had various logic bugs.
5528 This also affects blocking I/O when the data being decoded is a
5532 *) Various S/MIME bugfixes and compatibility changes:
5533 output correct application/pkcs7 MIME type if
5534 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5535 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5536 of files as .eml work). Correctly handle very long lines in MIME
5540 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5542 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5543 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5544 a protocol version number mismatch like a decryption error
5545 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5548 *) Turn on RSA blinding by default in the default implementation
5549 to avoid a timing attack. Applications that don't want it can call
5550 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5551 They would be ill-advised to do so in most cases.
5552 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5554 *) Change RSA blinding code so that it works when the PRNG is not
5555 seeded (in this case, the secret RSA exponent is abused as
5556 an unpredictable seed -- if it is not unpredictable, there
5557 is no point in blinding anyway). Make RSA blinding thread-safe
5558 by remembering the creator's thread ID in rsa->blinding and
5559 having all other threads use local one-time blinding factors
5560 (this requires more computation than sharing rsa->blinding, but
5561 avoids excessive locking; and if an RSA object is not shared
5562 between threads, blinding will still be very fast).
5565 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5566 ENGINE as defaults for all supported algorithms irrespective of
5567 the 'flags' parameter. 'flags' is now honoured, so applications
5568 should make sure they are passing it correctly.
5571 *) Target "mingw" now allows native Windows code to be generated in
5572 the Cygwin environment as well as with the MinGW compiler.
5575 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5577 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5578 via timing by performing a MAC computation even if incorrect
5579 block cipher padding has been found. This is a countermeasure
5580 against active attacks where the attacker has to distinguish
5581 between bad padding and a MAC verification error. (CVE-2003-0078)
5583 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5584 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5585 Martin Vuagnoux (EPFL, Ilion)]
5587 *) Make the no-err option work as intended. The intention with no-err
5588 is not to have the whole error stack handling routines removed from
5589 libcrypto, it's only intended to remove all the function name and
5590 reason texts, thereby removing some of the footprint that may not
5591 be interesting if those errors aren't displayed anyway.
5593 NOTE: it's still possible for any application or module to have it's
5594 own set of error texts inserted. The routines are there, just not
5595 used by default when no-err is given.
5598 *) Add support for FreeBSD on IA64.
5599 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5601 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5602 Kerberos function mit_des_cbc_cksum(). Before this change,
5603 the value returned by DES_cbc_cksum() was like the one from
5604 mit_des_cbc_cksum(), except the bytes were swapped.
5605 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5607 *) Allow an application to disable the automatic SSL chain building.
5608 Before this a rather primitive chain build was always performed in
5609 ssl3_output_cert_chain(): an application had no way to send the
5610 correct chain if the automatic operation produced an incorrect result.
5612 Now the chain builder is disabled if either:
5614 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5616 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5618 The reasoning behind this is that an application would not want the
5619 auto chain building to take place if extra chain certificates are
5620 present and it might also want a means of sending no additional
5621 certificates (for example the chain has two certificates and the
5625 *) Add the possibility to build without the ENGINE framework.
5626 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5628 *) Under Win32 gmtime() can return NULL: check return value in
5629 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5632 *) DSA routines: under certain error conditions uninitialized BN objects
5633 could be freed. Solution: make sure initialization is performed early
5634 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5635 Nils Larsch <nla@trustcenter.de> via PR#459)
5638 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5639 checked on reconnect on the client side, therefore session resumption
5640 could still fail with a "ssl session id is different" error. This
5641 behaviour is masked when SSL_OP_ALL is used due to
5642 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5643 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5644 followup to PR #377.
5647 *) IA-32 assembler support enhancements: unified ELF targets, support
5648 for SCO/Caldera platforms, fix for Cygwin shared build.
5651 *) Add support for FreeBSD on sparc64. As a consequence, support for
5652 FreeBSD on non-x86 processors is separate from x86 processors on
5653 the config script, much like the NetBSD support.
5654 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5656 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5658 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5661 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5662 code (06) was taken as the first octet of the session ID and the last
5663 octet was ignored consequently. As a result SSLv2 client side session
5664 caching could not have worked due to the session ID mismatch between
5666 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5670 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5671 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5675 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5676 seems that in spite of existing for more than a year, many application
5677 author have done nothing to provide the necessary callbacks, which
5678 means that this particular engine will not work properly anywhere.
5679 This is a very unfortunate situation which forces us, in the name
5680 of usability, to give the hw_ncipher.c a static lock, which is part
5682 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5683 appear in 0.9.8 or later. We EXPECT application authors to have
5684 dealt properly with this when 0.9.8 is released (unless we actually
5685 make such changes in the libcrypto locking code that changes will
5686 have to be made anyway).
5689 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5690 octets have been read, EOF or an error occurs. Without this change
5691 some truncated ASN1 structures will not produce an error.
5694 *) Disable Heimdal support, since it hasn't been fully implemented.
5695 Still give the possibility to force the use of Heimdal, but with
5696 warnings and a request that patches get sent to openssl-dev.
5699 *) Add the VC-CE target, introduce the WINCE sysname, and add
5700 INSTALL.WCE and appropriate conditionals to make it build.
5701 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5703 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5704 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5705 edit numbers of the version.
5706 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5708 *) Introduce safe string copy and catenation functions
5709 (BUF_strlcpy() and BUF_strlcat()).
5710 [Ben Laurie (CHATS) and Richard Levitte]
5712 *) Avoid using fixed-size buffers for one-line DNs.
5713 [Ben Laurie (CHATS)]
5715 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5716 resizing buffers containing secrets, and use where appropriate.
5717 [Ben Laurie (CHATS)]
5719 *) Avoid using fixed size buffers for configuration file location.
5720 [Ben Laurie (CHATS)]
5722 *) Avoid filename truncation for various CA files.
5723 [Ben Laurie (CHATS)]
5725 *) Use sizeof in preference to magic numbers.
5726 [Ben Laurie (CHATS)]
5728 *) Avoid filename truncation in cert requests.
5729 [Ben Laurie (CHATS)]
5731 *) Add assertions to check for (supposedly impossible) buffer
5733 [Ben Laurie (CHATS)]
5735 *) Don't cache truncated DNS entries in the local cache (this could
5736 potentially lead to a spoofing attack).
5737 [Ben Laurie (CHATS)]
5739 *) Fix various buffers to be large enough for hex/decimal
5740 representations in a platform independent manner.
5741 [Ben Laurie (CHATS)]
5743 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5744 resizing buffers containing secrets, and use where appropriate.
5745 [Ben Laurie (CHATS)]
5747 *) Add BIO_indent() to avoid much slightly worrying code to do
5749 [Ben Laurie (CHATS)]
5751 *) Convert sprintf()/BIO_puts() to BIO_printf().
5752 [Ben Laurie (CHATS)]
5754 *) buffer_gets() could terminate with the buffer only half
5756 [Ben Laurie (CHATS)]
5758 *) Add assertions to prevent user-supplied crypto functions from
5759 overflowing internal buffers by having large block sizes, etc.
5760 [Ben Laurie (CHATS)]
5762 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5764 [Ben Laurie (CHATS)]
5766 *) Eliminate unused copy of key in RC4.
5767 [Ben Laurie (CHATS)]
5769 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5770 [Ben Laurie (CHATS)]
5772 *) Fix off-by-one error in EGD path.
5773 [Ben Laurie (CHATS)]
5775 *) If RANDFILE path is too long, ignore instead of truncating.
5776 [Ben Laurie (CHATS)]
5778 *) Eliminate unused and incorrectly sized X.509 structure
5780 [Ben Laurie (CHATS)]
5782 *) Eliminate unused and dangerous function knumber().
5783 [Ben Laurie (CHATS)]
5785 *) Eliminate unused and dangerous structure, KSSL_ERR.
5786 [Ben Laurie (CHATS)]
5788 *) Protect against overlong session ID context length in an encoded
5789 session object. Since these are local, this does not appear to be
5791 [Ben Laurie (CHATS)]
5793 *) Change from security patch (see 0.9.6e below) that did not affect
5794 the 0.9.6 release series:
5796 Remote buffer overflow in SSL3 protocol - an attacker could
5797 supply an oversized master key in Kerberos-enabled versions.
5799 [Ben Laurie (CHATS)]
5801 *) Change the SSL kerb5 codes to match RFC 2712.
5804 *) Make -nameopt work fully for req and add -reqopt switch.
5805 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5807 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5808 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5810 *) Make sure tests can be performed even if the corresponding algorithms
5811 have been removed entirely. This was also the last step to make
5812 OpenSSL compilable with DJGPP under all reasonable conditions.
5813 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5815 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5816 to allow version independent disabling of normally unselected ciphers,
5817 which may be activated as a side-effect of selecting a single cipher.
5819 (E.g., cipher list string "RSA" enables ciphersuites that are left
5820 out of "ALL" because they do not provide symmetric encryption.
5821 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5822 [Lutz Jaenicke, Bodo Moeller]
5824 *) Add appropriate support for separate platform-dependent build
5825 directories. The recommended way to make a platform-dependent
5826 build directory is the following (tested on Linux), maybe with
5829 # Place yourself outside of the OpenSSL source tree. In
5830 # this example, the environment variable OPENSSL_SOURCE
5831 # is assumed to contain the absolute OpenSSL source directory.
5832 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5833 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5834 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5835 mkdir -p `dirname $F`
5836 ln -s $OPENSSL_SOURCE/$F $F
5839 To be absolutely sure not to disturb the source tree, a "make clean"
5840 is a good thing. If it isn't successful, don't worry about it,
5841 it probably means the source directory is very clean.
5844 *) Make sure any ENGINE control commands make local copies of string
5845 pointers passed to them whenever necessary. Otherwise it is possible
5846 the caller may have overwritten (or deallocated) the original string
5847 data when a later ENGINE operation tries to use the stored values.
5848 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5850 *) Improve diagnostics in file reading and command-line digests.
5851 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5853 *) Add AES modes CFB and OFB to the object database. Correct an
5854 error in AES-CFB decryption.
5857 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5858 allows existing EVP_CIPHER_CTX structures to be reused after
5859 calling EVP_*Final(). This behaviour is used by encryption
5860 BIOs and some applications. This has the side effect that
5861 applications must explicitly clean up cipher contexts with
5862 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5865 *) Check the values of dna and dnb in bn_mul_recursive before calling
5866 bn_mul_comba (a non zero value means the a or b arrays do not contain
5867 n2 elements) and fallback to bn_mul_normal if either is not zero.
5870 *) Fix escaping of non-ASCII characters when using the -subj option
5871 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5874 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5875 form for "surname", serialNumber has no short form.
5876 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5877 therefore remove "mail" short name for "internet 7".
5878 The OID for unique identifiers in X509 certificates is
5879 x500UniqueIdentifier, not uniqueIdentifier.
5880 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5883 *) Add an "init" command to the ENGINE config module and auto initialize
5884 ENGINEs. Without any "init" command the ENGINE will be initialized
5885 after all ctrl commands have been executed on it. If init=1 the
5886 ENGINE is initailized at that point (ctrls before that point are run
5887 on the uninitialized ENGINE and after on the initialized one). If
5888 init=0 then the ENGINE will not be iniatialized at all.
5891 *) Fix the 'app_verify_callback' interface so that the user-defined
5892 argument is actually passed to the callback: In the
5893 SSL_CTX_set_cert_verify_callback() prototype, the callback
5894 declaration has been changed from
5897 int (*cb)(X509_STORE_CTX *,void *);
5898 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5899 i=s->ctx->app_verify_callback(&ctx)
5900 has been changed into
5901 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5903 To update applications using SSL_CTX_set_cert_verify_callback(),
5904 a dummy argument can be added to their callback functions.
5905 [D. K. Smetters <smetters@parc.xerox.com>]
5907 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5908 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5910 *) Add and OPENSSL_LOAD_CONF define which will cause
5911 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5912 This allows older applications to transparently support certain
5913 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5914 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5915 load the config file and OPENSSL_add_all_algorithms_conf() which will
5916 always load it have also been added.
5919 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5920 Adjust NIDs and EVP layer.
5921 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5923 *) Config modules support in openssl utility.
5925 Most commands now load modules from the config file,
5926 though in a few (such as version) this isn't done
5927 because it couldn't be used for anything.
5929 In the case of ca and req the config file used is
5930 the same as the utility itself: that is the -config
5931 command line option can be used to specify an
5935 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5936 use "openssl_conf" if filename is NULL use default openssl config file.
5939 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5940 config section name. Add a new flag to tolerate a missing config file
5941 and move code to CONF_modules_load_file().
5944 *) Support for crypto accelerator cards from Accelerated Encryption
5945 Processing, www.aep.ie. (Use engine 'aep')
5946 The support was copied from 0.9.6c [engine] and adapted/corrected
5947 to work with the new engine framework.
5948 [AEP Inc. and Richard Levitte]
5950 *) Support for SureWare crypto accelerator cards from Baltimore
5951 Technologies. (Use engine 'sureware')
5952 The support was copied from 0.9.6c [engine] and adapted
5953 to work with the new engine framework.
5956 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5957 make the newer ENGINE framework commands for the CHIL engine work.
5958 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5960 *) Make it possible to produce shared libraries on ReliantUNIX.
5961 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5963 *) Add the configuration target debug-linux-ppro.
5964 Make 'openssl rsa' use the general key loading routines
5965 implemented in apps.c, and make those routines able to
5966 handle the key format FORMAT_NETSCAPE and the variant
5968 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5970 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5971 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5973 *) Add -keyform to rsautl, and document -engine.
5974 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5976 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5977 BIO_R_NO_SUCH_FILE error code rather than the generic
5978 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5981 *) Add new functions
5983 ERR_peek_last_error_line
5984 ERR_peek_last_error_line_data.
5985 These are similar to
5988 ERR_peek_error_line_data,
5989 but report on the latest error recorded rather than the first one
5990 still in the error queue.
5991 [Ben Laurie, Bodo Moeller]
5993 *) default_algorithms option in ENGINE config module. This allows things
5995 default_algorithms = ALL
5996 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5999 *) Preliminary ENGINE config module.
6002 *) New experimental application configuration code.
6005 *) Change the AES code to follow the same name structure as all other
6006 symmetric ciphers, and behave the same way. Move everything to
6007 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6008 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6010 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6011 [Ben Laurie and Theo de Raadt]
6013 *) Add option to output public keys in req command.
6014 [Massimiliano Pala madwolf@openca.org]
6016 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6017 (up to about 10% better than before for P-192 and P-224).
6020 *) New functions/macros
6022 SSL_CTX_set_msg_callback(ctx, cb)
6023 SSL_CTX_set_msg_callback_arg(ctx, arg)
6024 SSL_set_msg_callback(ssl, cb)
6025 SSL_set_msg_callback_arg(ssl, arg)
6027 to request calling a callback function
6029 void cb(int write_p, int version, int content_type,
6030 const void *buf, size_t len, SSL *ssl, void *arg)
6032 whenever a protocol message has been completely received
6033 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6034 protocol version according to which the SSL library interprets
6035 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6036 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6037 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6038 specification (change_cipher_spec(20), alert(21), handshake(22)).
6039 'buf' and 'len' point to the actual message, 'ssl' to the
6040 SSL object, and 'arg' is the application-defined value set by
6041 SSL[_CTX]_set_msg_callback_arg().
6043 'openssl s_client' and 'openssl s_server' have new '-msg' options
6044 to enable a callback that displays all protocol messages.
6047 *) Change the shared library support so shared libraries are built as
6048 soon as the corresponding static library is finished, and thereby get
6049 openssl and the test programs linked against the shared library.
6050 This still only happens when the keyword "shard" has been given to
6051 the configuration scripts.
6053 NOTE: shared library support is still an experimental thing, and
6054 backward binary compatibility is still not guaranteed.
6055 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6057 *) Add support for Subject Information Access extension.
6058 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6060 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6061 additional bytes when new memory had to be allocated, not just
6062 when reusing an existing buffer.
6065 *) New command line and configuration option 'utf8' for the req command.
6066 This allows field values to be specified as UTF8 strings.
6069 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6070 runs for the former and machine-readable output for the latter.
6073 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6074 of the e-mail address in the DN (i.e., it will go into a certificate
6075 extension only). The new configuration file option 'email_in_dn = no'
6076 has the same effect.
6077 [Massimiliano Pala madwolf@openca.org]
6079 *) Change all functions with names starting with des_ to be starting
6080 with DES_ instead. Add wrappers that are compatible with libdes,
6081 but are named _ossl_old_des_*. Finally, add macros that map the
6082 des_* symbols to the corresponding _ossl_old_des_* if libdes
6083 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6084 desired, the des_* symbols will be mapped to DES_*, with one
6087 Since we provide two compatibility mappings, the user needs to
6088 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6089 compatibility is desired. The default (i.e., when that macro
6090 isn't defined) is OpenSSL 0.9.6c compatibility.
6092 There are also macros that enable and disable the support of old
6093 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6094 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6095 are defined, the default will apply: to support the old des routines.
6097 In either case, one must include openssl/des.h to get the correct
6098 definitions. Do not try to just include openssl/des_old.h, that
6101 NOTE: This is a major break of an old API into a new one. Software
6102 authors are encouraged to switch to the DES_ style functions. Some
6103 time in the future, des_old.h and the libdes compatibility functions
6104 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6105 default), and then completely removed.
6108 *) Test for certificates which contain unsupported critical extensions.
6109 If such a certificate is found during a verify operation it is
6110 rejected by default: this behaviour can be overridden by either
6111 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6112 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6113 X509_supported_extension() has also been added which returns 1 if a
6114 particular extension is supported.
6117 *) Modify the behaviour of EVP cipher functions in similar way to digests
6118 to retain compatibility with existing code.
6121 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6122 compatibility with existing code. In particular the 'ctx' parameter does
6123 not have to be to be initialized before the call to EVP_DigestInit() and
6124 it is tidied up after a call to EVP_DigestFinal(). New function
6125 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6126 EVP_MD_CTX_copy() changed to not require the destination to be
6127 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6128 requires the destination to be valid.
6130 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6131 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6134 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6135 so that complete 'Handshake' protocol structures are kept in memory
6136 instead of overwriting 'msg_type' and 'length' with 'body' data.
6139 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6140 [Massimo Santin via Richard Levitte]
6142 *) Major restructuring to the underlying ENGINE code. This includes
6143 reduction of linker bloat, separation of pure "ENGINE" manipulation
6144 (initialisation, etc) from functionality dealing with implementations
6145 of specific crypto iterfaces. This change also introduces integrated
6146 support for symmetric ciphers and digest implementations - so ENGINEs
6147 can now accelerate these by providing EVP_CIPHER and EVP_MD
6148 implementations of their own. This is detailed in crypto/engine/README
6149 as it couldn't be adequately described here. However, there are a few
6150 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6151 were changed in the original introduction of ENGINE code have now
6152 reverted back - the hooking from this code to ENGINE is now a good
6153 deal more passive and at run-time, operations deal directly with
6154 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6155 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6156 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6157 they were not being used by the framework as there is no concept of a
6158 BIGNUM_METHOD and they could not be generalised to the new
6159 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6160 ENGINE_cpy() has been removed as it cannot be consistently defined in
6164 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6167 *) Change mkdef.pl to sort symbols that get the same entry number,
6168 and make sure the automatically generated functions ERR_load_*
6169 become part of libeay.num as well.
6172 *) New function SSL_renegotiate_pending(). This returns true once
6173 renegotiation has been requested (either SSL_renegotiate() call
6174 or HelloRequest/ClientHello received from the peer) and becomes
6175 false once a handshake has been completed.
6176 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6177 sends a HelloRequest, but does not ensure that a handshake takes
6178 place. SSL_renegotiate_pending() is useful for checking if the
6179 client has followed the request.)
6182 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6183 By default, clients may request session resumption even during
6184 renegotiation (if session ID contexts permit); with this option,
6185 session resumption is possible only in the first handshake.
6187 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6188 more bits available for options that should not be part of
6189 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6192 *) Add some demos for certificate and certificate request creation.
6195 *) Make maximum certificate chain size accepted from the peer application
6196 settable (SSL*_get/set_max_cert_list()), as proposed by
6197 "Douglas E. Engert" <deengert@anl.gov>.
6200 *) Add support for shared libraries for Unixware-7
6201 (Boyd Lynn Gerber <gerberb@zenez.com>).
6204 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6205 be done prior to destruction. Use this to unload error strings from
6206 ENGINEs that load their own error strings. NB: This adds two new API
6207 functions to "get" and "set" this destroy handler in an ENGINE.
6210 *) Alter all existing ENGINE implementations (except "openssl" and
6211 "openbsd") to dynamically instantiate their own error strings. This
6212 makes them more flexible to be built both as statically-linked ENGINEs
6213 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6214 Also, add stub code to each that makes building them as self-contained
6215 shared-libraries easier (see README.ENGINE).
6218 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6219 implementations into applications that are completely implemented in
6220 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6221 commands that can be used to configure what shared-library to load and
6222 to control aspects of the way it is handled. Also, made an update to
6223 the README.ENGINE file that brings its information up-to-date and
6224 provides some information and instructions on the "dynamic" ENGINE
6225 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6228 *) Make it possible to unload ranges of ERR strings with a new
6229 "ERR_unload_strings" function.
6232 *) Add a copy() function to EVP_MD.
6235 *) Make EVP_MD routines take a context pointer instead of just the
6236 md_data void pointer.
6239 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6240 that the digest can only process a single chunk of data
6241 (typically because it is provided by a piece of
6242 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6243 is only going to provide a single chunk of data, and hence the
6244 framework needn't accumulate the data for oneshot drivers.
6247 *) As with "ERR", make it possible to replace the underlying "ex_data"
6248 functions. This change also alters the storage and management of global
6249 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6250 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6251 index counters. The API functions that use this state have been changed
6252 to take a "class_index" rather than pointers to the class's local STACK
6253 and counter, and there is now an API function to dynamically create new
6254 classes. This centralisation allows us to (a) plug a lot of the
6255 thread-safety problems that existed, and (b) makes it possible to clean
6256 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6257 such data would previously have always leaked in application code and
6258 workarounds were in place to make the memory debugging turn a blind eye
6259 to it. Application code that doesn't use this new function will still
6260 leak as before, but their memory debugging output will announce it now
6261 rather than letting it slide.
6263 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6264 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6265 has a return value to indicate success or failure.
6268 *) Make it possible to replace the underlying "ERR" functions such that the
6269 global state (2 LHASH tables and 2 locks) is only used by the "default"
6270 implementation. This change also adds two functions to "get" and "set"
6271 the implementation prior to it being automatically set the first time
6272 any other ERR function takes place. Ie. an application can call "get",
6273 pass the return value to a module it has just loaded, and that module
6274 can call its own "set" function using that value. This means the
6275 module's "ERR" operations will use (and modify) the error state in the
6276 application and not in its own statically linked copy of OpenSSL code.
6279 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6280 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6281 the operation, and provides a more encapsulated way for external code
6282 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6283 to use these functions rather than manually incrementing the counts.
6285 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6288 *) Add EVP test program.
6291 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6294 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6295 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6296 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6297 These allow a CRL to be built without having to access X509_CRL fields
6298 directly. Modify 'ca' application to use new functions.
6301 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6302 bug workarounds. Rollback attack detection is a security feature.
6303 The problem will only arise on OpenSSL servers when TLSv1 is not
6304 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6305 Software authors not wanting to support TLSv1 will have special reasons
6306 for their choice and can explicitly enable this option.
6307 [Bodo Moeller, Lutz Jaenicke]
6309 *) Rationalise EVP so it can be extended: don't include a union of
6310 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6311 (similar to those existing for EVP_CIPHER_CTX).
6316 EVP_MD_CTX_init(&md); /* new function call */
6317 EVP_DigestInit(&md, EVP_sha1());
6318 EVP_DigestUpdate(&md, in, len);
6319 EVP_DigestFinal(&md, out, NULL);
6320 EVP_MD_CTX_cleanup(&md); /* new function call */
6324 *) Make DES key schedule conform to the usual scheme, as well as
6325 correcting its structure. This means that calls to DES functions
6326 now have to pass a pointer to a des_key_schedule instead of a
6327 plain des_key_schedule (which was actually always a pointer
6330 des_key_schedule ks;
6332 des_set_key_checked(..., &ks);
6333 des_ncbc_encrypt(..., &ks, ...);
6335 (Note that a later change renames 'des_...' into 'DES_...'.)
6338 *) Initial reduction of linker bloat: the use of some functions, such as
6339 PEM causes large amounts of unused functions to be linked in due to
6340 poor organisation. For example pem_all.c contains every PEM function
6341 which has a knock on effect of linking in large amounts of (unused)
6342 ASN1 code. Grouping together similar functions and splitting unrelated
6343 functions prevents this.
6346 *) Cleanup of EVP macros.
6349 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6350 correct _ecb suffix.
6353 *) Add initial OCSP responder support to ocsp application. The
6354 revocation information is handled using the text based index
6355 use by the ca application. The responder can either handle
6356 requests generated internally, supplied in files (for example
6357 via a CGI script) or using an internal minimal server.
6360 *) Add configuration choices to get zlib compression for TLS.
6363 *) Changes to Kerberos SSL for RFC 2712 compliance:
6364 1. Implemented real KerberosWrapper, instead of just using
6365 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6366 2. Implemented optional authenticator field of KerberosWrapper.
6368 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6369 and authenticator structs; see crypto/krb5/.
6371 Generalized Kerberos calls to support multiple Kerberos libraries.
6372 [Vern Staats <staatsvr@asc.hpc.mil>,
6373 Jeffrey Altman <jaltman@columbia.edu>
6374 via Richard Levitte]
6376 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6377 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6378 values for each of the key sizes rather than having just
6379 parameters (and 'speed' generating keys each time).
6382 *) Speed up EVP routines.
6385 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6386 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6387 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6388 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6390 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6391 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6392 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6395 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6397 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6400 *) Added the OS2-EMX target.
6401 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6403 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6404 to support NCONF routines in extension code. New function CONF_set_nconf()
6405 to allow functions which take an NCONF to also handle the old LHASH
6406 structure: this means that the old CONF compatible routines can be
6407 retained (in particular wrt extensions) without having to duplicate the
6408 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6411 *) Enhance the general user interface with mechanisms for inner control
6412 and with possibilities to have yes/no kind of prompts.
6415 *) Change all calls to low level digest routines in the library and
6416 applications to use EVP. Add missing calls to HMAC_cleanup() and
6417 don't assume HMAC_CTX can be copied using memcpy().
6418 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6420 *) Add the possibility to control engines through control names but with
6421 arbitrary arguments instead of just a string.
6422 Change the key loaders to take a UI_METHOD instead of a callback
6423 function pointer. NOTE: this breaks binary compatibility with earlier
6424 versions of OpenSSL [engine].
6425 Adapt the nCipher code for these new conditions and add a card insertion
6429 *) Enhance the general user interface with mechanisms to better support
6430 dialog box interfaces, application-defined prompts, the possibility
6431 to use defaults (for example default passwords from somewhere else)
6432 and interrupts/cancellations.
6435 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6436 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6439 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6440 tidy up some unnecessarily weird code in 'sk_new()').
6441 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6443 *) Change the key loading routines for ENGINEs to use the same kind
6444 callback (pem_password_cb) as all other routines that need this
6448 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6449 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6450 than this minimum value is recommended.
6453 *) New random seeder for OpenVMS, using the system process statistics
6454 that are easily reachable.
6457 *) Windows apparently can't transparently handle global
6458 variables defined in DLLs. Initialisations such as:
6460 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6462 wont compile. This is used by the any applications that need to
6463 declare their own ASN1 modules. This was fixed by adding the option
6464 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6465 needed for static libraries under Win32.
6468 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6469 setting of purpose and trust fields. New X509_STORE trust and
6470 purpose functions and tidy up setting in other SSL functions.
6473 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6474 structure. These are inherited by X509_STORE_CTX when it is
6475 initialised. This allows various defaults to be set in the
6476 X509_STORE structure (such as flags for CRL checking and custom
6477 purpose or trust settings) for functions which only use X509_STORE_CTX
6478 internally such as S/MIME.
6480 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6481 trust settings if they are not set in X509_STORE. This allows X509_STORE
6482 purposes and trust (in S/MIME for example) to override any set by default.
6484 Add command line options for CRL checking to smime, s_client and s_server
6488 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6489 are set then the CRL is looked up in the X509_STORE structure and
6490 its validity and signature checked, then if the certificate is found
6491 in the CRL the verify fails with a revoked error.
6493 Various new CRL related callbacks added to X509_STORE_CTX structure.
6495 Command line options added to 'verify' application to support this.
6497 This needs some additional work, such as being able to handle multiple
6498 CRLs with different times, extension based lookup (rather than just
6499 by subject name) and ultimately more complete V2 CRL extension
6503 *) Add a general user interface API (crypto/ui/). This is designed
6504 to replace things like des_read_password and friends (backward
6505 compatibility functions using this new API are provided).
6506 The purpose is to remove prompting functions from the DES code
6507 section as well as provide for prompting through dialog boxes in
6508 a window system and the like.
6511 *) Add "ex_data" support to ENGINE so implementations can add state at a
6512 per-structure level rather than having to store it globally.
6515 *) Make it possible for ENGINE structures to be copied when retrieved by
6516 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6517 This causes the "original" ENGINE structure to act like a template,
6518 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6519 operational state can be localised to each ENGINE structure, despite the
6520 fact they all share the same "methods". New ENGINE structures returned in
6521 this case have no functional references and the return value is the single
6522 structural reference. This matches the single structural reference returned
6523 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6527 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6528 needs to match any other type at all we need to manually clear the
6532 *) Changes to the "openssl engine" utility to include;
6533 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6534 about an ENGINE's available control commands.
6535 - executing control commands from command line arguments using the
6536 '-pre' and '-post' switches. '-post' is only used if '-t' is
6537 specified and the ENGINE is successfully initialised. The syntax for
6538 the individual commands are colon-separated, for example;
6539 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6542 *) New dynamic control command support for ENGINEs. ENGINEs can now
6543 declare their own commands (numbers), names (strings), descriptions,
6544 and input types for run-time discovery by calling applications. A
6545 subset of these commands are implicitly classed as "executable"
6546 depending on their input type, and only these can be invoked through
6547 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6548 can be based on user input, config files, etc). The distinction is
6549 that "executable" commands cannot return anything other than a boolean
6550 result and can only support numeric or string input, whereas some
6551 discoverable commands may only be for direct use through
6552 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6553 pointers, or other custom uses. The "executable" commands are to
6554 support parameterisations of ENGINE behaviour that can be
6555 unambiguously defined by ENGINEs and used consistently across any
6556 OpenSSL-based application. Commands have been added to all the
6557 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6558 control over shared-library paths without source code alterations.
6561 *) Changed all ENGINE implementations to dynamically allocate their
6562 ENGINEs rather than declaring them statically. Apart from this being
6563 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6564 this also allows the implementations to compile without using the
6565 internal engine_int.h header.
6568 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6569 'const' value. Any code that should be able to modify a RAND_METHOD
6570 should already have non-const pointers to it (ie. they should only
6571 modify their own ones).
6574 *) Made a variety of little tweaks to the ENGINE code.
6575 - "atalla" and "ubsec" string definitions were moved from header files
6576 to C code. "nuron" string definitions were placed in variables
6577 rather than hard-coded - allowing parameterisation of these values
6578 later on via ctrl() commands.
6579 - Removed unused "#if 0"'d code.
6580 - Fixed engine list iteration code so it uses ENGINE_free() to release
6581 structural references.
6582 - Constified the RAND_METHOD element of ENGINE structures.
6583 - Constified various get/set functions as appropriate and added
6584 missing functions (including a catch-all ENGINE_cpy that duplicates
6585 all ENGINE values onto a new ENGINE except reference counts/state).
6586 - Removed NULL parameter checks in get/set functions. Setting a method
6587 or function to NULL is a way of cancelling out a previously set
6588 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6589 and doesn't justify the extra error symbols and code.
6590 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6591 flags from engine_int.h to engine.h.
6592 - Changed prototypes for ENGINE handler functions (init(), finish(),
6593 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6596 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6597 to the algorithm using long division. The binary algorithm can be
6598 used only if the modulus is odd. On 32-bit systems, it is faster
6599 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6600 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6601 up to 450 bits. In 64-bit environments, the binary algorithm
6602 appears to be advantageous for much longer moduli; here we use it
6603 for moduli up to 2048 bits.
6606 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6607 could not support the combine flag in choice fields.
6610 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6611 extensions from a certificate request to the certificate.
6614 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6615 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6616 file: this allows the display of the certificate about to be
6617 signed to be customised, to allow certain fields to be included
6618 or excluded and extension details. The old system didn't display
6619 multicharacter strings properly, omitted fields not in the policy
6620 and couldn't display additional details such as extensions.
6623 *) Function EC_POINTs_mul for multiple scalar multiplication
6624 of an arbitrary number of elliptic curve points
6625 \sum scalars[i]*points[i],
6626 optionally including the generator defined for the EC_GROUP:
6627 scalar*generator + \sum scalars[i]*points[i].
6629 EC_POINT_mul is a simple wrapper function for the typical case
6630 that the point list has just one item (besides the optional
6634 *) First EC_METHODs for curves over GF(p):
6636 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6637 operations and provides various method functions that can also
6638 operate with faster implementations of modular arithmetic.
6640 EC_GFp_mont_method() reuses most functions that are part of
6641 EC_GFp_simple_method, but uses Montgomery arithmetic.
6643 [Bodo Moeller; point addition and point doubling
6644 implementation directly derived from source code provided by
6645 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6647 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6648 crypto/ec/ec_lib.c):
6650 Curves are EC_GROUP objects (with an optional group generator)
6651 based on EC_METHODs that are built into the library.
6653 Points are EC_POINT objects based on EC_GROUP objects.
6655 Most of the framework would be able to handle curves over arbitrary
6656 finite fields, but as there are no obvious types for fields other
6657 than GF(p), some functions are limited to that for now.
6660 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6661 that the file contains a complete HTTP response.
6664 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6665 change the def and num file printf format specifier from "%-40sXXX"
6666 to "%-39s XXX". The latter will always guarantee a space after the
6667 field while the former will cause them to run together if the field
6668 is 40 of more characters long.
6671 *) Constify the cipher and digest 'method' functions and structures
6672 and modify related functions to take constant EVP_MD and EVP_CIPHER
6676 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6677 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6680 *) Modify EVP_Digest*() routines so they now return values. Although the
6681 internal software routines can never fail additional hardware versions
6685 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6687 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6688 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6691 ERR_R_NESTED_ASN1_ERROR
6693 ERR_R_MISSING_ASN1_EOS
6694 were 4 .. 9, conflicting with
6695 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6697 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6698 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6700 Add new error code 'ERR_R_INTERNAL_ERROR'.
6703 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6707 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6708 sets the subject name for a new request or supersedes the
6709 subject name in a given request. Formats that can be parsed are
6710 'CN=Some Name, OU=myOU, C=IT'
6712 'CN=Some Name/OU=myOU/C=IT'.
6714 Add options '-batch' and '-verbose' to 'openssl req'.
6715 [Massimiliano Pala <madwolf@hackmasters.net>]
6717 *) Introduce the possibility to access global variables through
6718 functions on platform were that's the best way to handle exporting
6719 global variables in shared libraries. To enable this functionality,
6720 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6721 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6722 is normally done by Configure or something similar).
6724 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6725 in the source file (foo.c) like this:
6727 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6728 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6730 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6731 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6733 OPENSSL_DECLARE_GLOBAL(int,foo);
6734 #define foo OPENSSL_GLOBAL_REF(foo)
6735 OPENSSL_DECLARE_GLOBAL(double,bar);
6736 #define bar OPENSSL_GLOBAL_REF(bar)
6738 The #defines are very important, and therefore so is including the
6739 header file everywhere where the defined globals are used.
6741 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6742 of ASN.1 items, but that structure is a bit different.
6744 The largest change is in util/mkdef.pl which has been enhanced with
6745 better and easier to understand logic to choose which symbols should
6746 go into the Windows .def files as well as a number of fixes and code
6747 cleanup (among others, algorithm keywords are now sorted
6748 lexicographically to avoid constant rewrites).
6751 *) In BN_div() keep a copy of the sign of 'num' before writing the
6752 result to 'rm' because if rm==num the value will be overwritten
6753 and produce the wrong result if 'num' is negative: this caused
6754 problems with BN_mod() and BN_nnmod().
6757 *) Function OCSP_request_verify(). This checks the signature on an
6758 OCSP request and verifies the signer certificate. The signer
6759 certificate is just checked for a generic purpose and OCSP request
6763 *) Add OCSP_check_validity() function to check the validity of OCSP
6764 responses. OCSP responses are prepared in real time and may only
6765 be a few seconds old. Simply checking that the current time lies
6766 between thisUpdate and nextUpdate max reject otherwise valid responses
6767 caused by either OCSP responder or client clock inaccuracy. Instead
6768 we allow thisUpdate and nextUpdate to fall within a certain period of
6769 the current time. The age of the response can also optionally be
6770 checked. Two new options -validity_period and -status_age added to
6774 *) If signature or public key algorithm is unrecognized print out its
6775 OID rather that just UNKNOWN.
6778 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6779 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6780 ID to be generated from the issuer certificate alone which can then be
6781 passed to OCSP_id_issuer_cmp().
6784 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6785 ASN1 modules to export functions returning ASN1_ITEM pointers
6786 instead of the ASN1_ITEM structures themselves. This adds several
6787 new macros which allow the underlying ASN1 function/structure to
6788 be accessed transparently. As a result code should not use ASN1_ITEM
6789 references directly (such as &X509_it) but instead use the relevant
6790 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6791 use of the new ASN1 code on platforms where exporting structures
6792 is problematical (for example in shared libraries) but exporting
6793 functions returning pointers to structures is not.
6796 *) Add support for overriding the generation of SSL/TLS session IDs.
6797 These callbacks can be registered either in an SSL_CTX or per SSL.
6798 The purpose of this is to allow applications to control, if they wish,
6799 the arbitrary values chosen for use as session IDs, particularly as it
6800 can be useful for session caching in multiple-server environments. A
6801 command-line switch for testing this (and any client code that wishes
6802 to use such a feature) has been added to "s_server".
6803 [Geoff Thorpe, Lutz Jaenicke]
6805 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6806 of the form '#if defined(...) || defined(...) || ...' and
6807 '#if !defined(...) && !defined(...) && ...'. This also avoids
6808 the growing number of special cases it was previously handling.
6811 *) Make all configuration macros available for application by making
6812 sure they are available in opensslconf.h, by giving them names starting
6813 with "OPENSSL_" to avoid conflicts with other packages and by making
6814 sure e_os2.h will cover all platform-specific cases together with
6816 Additionally, it is now possible to define configuration/platform-
6817 specific names (called "system identities"). In the C code, these
6818 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6819 macro with the name beginning with "OPENSSL_SYS_", which is determined
6820 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6824 *) New option -set_serial to 'req' and 'x509' this allows the serial
6825 number to use to be specified on the command line. Previously self
6826 signed certificates were hard coded with serial number 0 and the
6827 CA options of 'x509' had to use a serial number in a file which was
6831 *) New options to 'ca' utility to support V2 CRL entry extensions.
6832 Currently CRL reason, invalidity date and hold instruction are
6833 supported. Add new CRL extensions to V3 code and some new objects.
6836 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6837 disable standard block padding (aka PKCS#5 padding) in the EVP
6838 API, which was previously mandatory. This means that the data is
6839 not padded in any way and so the total length much be a multiple
6840 of the block size, otherwise an error occurs.
6843 *) Initial (incomplete) OCSP SSL support.
6846 *) New function OCSP_parse_url(). This splits up a URL into its host,
6847 port and path components: primarily to parse OCSP URLs. New -url
6848 option to ocsp utility.
6851 *) New nonce behavior. The return value of OCSP_check_nonce() now
6852 reflects the various checks performed. Applications can decide
6853 whether to tolerate certain situations such as an absent nonce
6854 in a response when one was present in a request: the ocsp application
6855 just prints out a warning. New function OCSP_add1_basic_nonce()
6856 this is to allow responders to include a nonce in a response even if
6857 the request is nonce-less.
6860 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6861 skipped when using openssl x509 multiple times on a single input file,
6862 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6865 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6866 set string type: to handle setting ASN1_TIME structures. Fix ca
6867 utility to correctly initialize revocation date of CRLs.
6870 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6871 the clients preferred ciphersuites and rather use its own preferences.
6872 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6873 Internet Explorer by ensuring unchanged hash method during stepup.
6874 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6877 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6878 to aes and add a new 'exist' option to print out symbols that don't
6882 *) Additional options to ocsp utility to allow flags to be set and
6883 additional certificates supplied.
6886 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6887 OCSP client a number of certificate to only verify the response
6891 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6892 handle the new API. Currently only ECB, CBC modes supported. Add new
6895 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6896 Encryption Standard (AES) Ciphersuites for Transport Layer
6897 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6898 not enabled by default and were not part of the "ALL" ciphersuite
6899 alias because they were not yet official; they could be
6900 explicitly requested by specifying the "AESdraft" ciphersuite
6901 group alias. In the final release of OpenSSL 0.9.7, the group
6902 alias is called "AES" and is part of "ALL".)
6903 [Ben Laurie, Steve Henson, Bodo Moeller]
6905 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6906 request to response.
6909 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6910 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6911 extract information from a certificate request. OCSP_response_create()
6912 creates a response and optionally adds a basic response structure.
6913 OCSP_basic_add1_status() adds a complete single response to a basic
6914 response and returns the OCSP_SINGLERESP structure just added (to allow
6915 extensions to be included for example). OCSP_basic_add1_cert() adds a
6916 certificate to a basic response and OCSP_basic_sign() signs a basic
6917 response with various flags. New helper functions ASN1_TIME_check()
6918 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6919 (converts ASN1_TIME to GeneralizedTime).
6922 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6923 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6924 structure from a certificate. X509_pubkey_digest() digests the public_key
6925 contents: this is used in various key identifiers.
6928 *) Make sk_sort() tolerate a NULL argument.
6929 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6931 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6932 passed by the function are trusted implicitly. If any of them signed the
6933 response then it is assumed to be valid and is not verified.
6936 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6937 to data. This was previously part of the PKCS7 ASN1 code. This
6938 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6939 [Steve Henson, reported by Kenneth R. Robinette
6940 <support@securenetterm.com>]
6942 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6943 routines: without these tracing memory leaks is very painful.
6944 Fix leaks in PKCS12 and PKCS7 routines.
6947 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6948 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6949 effectively meant GeneralizedTime would never be used. Now it
6950 is initialised to -1 but X509_time_adj() now has to check the value
6951 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6952 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6953 [Steve Henson, reported by Kenneth R. Robinette
6954 <support@securenetterm.com>]
6956 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6957 result in a zero length in the ASN1_INTEGER structure which was
6958 not consistent with the structure when d2i_ASN1_INTEGER() was used
6959 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6960 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6961 where it did not print out a minus for negative ASN1_INTEGER.
6964 *) Add summary printout to ocsp utility. The various functions which
6965 convert status values to strings have been renamed to:
6966 OCSP_response_status_str(), OCSP_cert_status_str() and
6967 OCSP_crl_reason_str() and are no longer static. New options
6968 to verify nonce values and to disable verification. OCSP response
6969 printout format cleaned up.
6972 *) Add additional OCSP certificate checks. These are those specified
6973 in RFC2560. This consists of two separate checks: the CA of the
6974 certificate being checked must either be the OCSP signer certificate
6975 or the issuer of the OCSP signer certificate. In the latter case the
6976 OCSP signer certificate must contain the OCSP signing extended key
6977 usage. This check is performed by attempting to match the OCSP
6978 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6979 in the OCSP_CERTID structures of the response.
6982 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6983 and related routines. This uses the standard OpenSSL certificate
6984 verify routines to perform initial checks (just CA validity) and
6985 to obtain the certificate chain. Then additional checks will be
6986 performed on the chain. Currently the root CA is checked to see
6987 if it is explicitly trusted for OCSP signing. This is used to set
6988 a root CA as a global signing root: that is any certificate that
6989 chains to that CA is an acceptable OCSP signing certificate.
6992 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6993 extensions from a separate configuration file.
6994 As when reading extensions from the main configuration file,
6995 the '-extensions ...' option may be used for specifying the
6997 [Massimiliano Pala <madwolf@comune.modena.it>]
6999 *) New OCSP utility. Allows OCSP requests to be generated or
7000 read. The request can be sent to a responder and the output
7001 parsed, outputed or printed in text form. Not complete yet:
7002 still needs to check the OCSP response validity.
7005 *) New subcommands for 'openssl ca':
7006 'openssl ca -status <serial>' prints the status of the cert with
7007 the given serial number (according to the index file).
7008 'openssl ca -updatedb' updates the expiry status of certificates
7010 [Massimiliano Pala <madwolf@comune.modena.it>]
7012 *) New '-newreq-nodes' command option to CA.pl. This is like
7013 '-newreq', but calls 'openssl req' with the '-nodes' option
7014 so that the resulting key is not encrypted.
7015 [Damien Miller <djm@mindrot.org>]
7017 *) New configuration for the GNU Hurd.
7018 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7020 *) Initial code to implement OCSP basic response verify. This
7021 is currently incomplete. Currently just finds the signer's
7022 certificate and verifies the signature on the response.
7025 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7026 value of OPENSSLDIR. This is available via the new '-d' option
7027 to 'openssl version', and is also included in 'openssl version -a'.
7030 *) Allowing defining memory allocation callbacks that will be given
7031 file name and line number information in additional arguments
7032 (a const char* and an int). The basic functionality remains, as
7033 well as the original possibility to just replace malloc(),
7034 realloc() and free() by functions that do not know about these
7035 additional arguments. To register and find out the current
7036 settings for extended allocation functions, the following
7037 functions are provided:
7039 CRYPTO_set_mem_ex_functions
7040 CRYPTO_set_locked_mem_ex_functions
7041 CRYPTO_get_mem_ex_functions
7042 CRYPTO_get_locked_mem_ex_functions
7044 These work the same way as CRYPTO_set_mem_functions and friends.
7045 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7046 extended allocation function is enabled.
7047 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7048 a conventional allocation function is enabled.
7049 [Richard Levitte, Bodo Moeller]
7051 *) Finish off removing the remaining LHASH function pointer casts.
7052 There should no longer be any prototype-casting required when using
7053 the LHASH abstraction, and any casts that remain are "bugs". See
7054 the callback types and macros at the head of lhash.h for details
7055 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7058 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7059 If /dev/[u]random devices are not available or do not return enough
7060 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7062 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7063 /etc/entropy will be queried once each in this sequence, quering stops
7064 when enough entropy was collected without querying more sockets.
7067 *) Change the Unix RAND_poll() variant to be able to poll several
7068 random devices, as specified by DEVRANDOM, until a sufficient amount
7069 of data has been collected. We spend at most 10 ms on each file
7070 (select timeout) and read in non-blocking mode. DEVRANDOM now
7071 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7072 (previously it was just the string "/dev/urandom"), so on typical
7073 platforms the 10 ms delay will never occur.
7074 Also separate out the Unix variant to its own file, rand_unix.c.
7075 For VMS, there's a currently-empty rand_vms.c.
7078 *) Move OCSP client related routines to ocsp_cl.c. These
7079 provide utility functions which an application needing
7080 to issue a request to an OCSP responder and analyse the
7081 response will typically need: as opposed to those which an
7082 OCSP responder itself would need which will be added later.
7084 OCSP_request_sign() signs an OCSP request with an API similar
7085 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7086 response. OCSP_response_get1_basic() extracts basic response
7087 from response. OCSP_resp_find_status(): finds and extracts status
7088 information from an OCSP_CERTID structure (which will be created
7089 when the request structure is built). These are built from lower
7090 level functions which work on OCSP_SINGLERESP structures but
7091 wont normally be used unless the application wishes to examine
7092 extensions in the OCSP response for example.
7094 Replace nonce routines with a pair of functions.
7095 OCSP_request_add1_nonce() adds a nonce value and optionally
7096 generates a random value. OCSP_check_nonce() checks the
7097 validity of the nonce in an OCSP response.
7100 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7101 This doesn't copy the supplied OCSP_CERTID and avoids the
7102 need to free up the newly created id. Change return type
7103 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7104 This can then be used to add extensions to the request.
7105 Deleted OCSP_request_new(), since most of its functionality
7106 is now in OCSP_REQUEST_new() (and the case insensitive name
7107 clash) apart from the ability to set the request name which
7108 will be added elsewhere.
7111 *) Update OCSP API. Remove obsolete extensions argument from
7112 various functions. Extensions are now handled using the new
7113 OCSP extension code. New simple OCSP HTTP function which
7114 can be used to send requests and parse the response.
7117 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7118 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7119 uses the special reorder version of SET OF to sort the attributes
7120 and reorder them to match the encoded order. This resolves a long
7121 standing problem: a verify on a PKCS7 structure just after signing
7122 it used to fail because the attribute order did not match the
7123 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7124 it uses the received order. This is necessary to tolerate some broken
7125 software that does not order SET OF. This is handled by encoding
7126 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7127 to produce the required SET OF.
7130 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7131 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7132 files to get correct declarations of the ASN.1 item variables.
7135 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7136 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7137 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7138 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7139 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7140 ASN1_ITEM and no wrapper functions.
7143 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7144 replace the old function pointer based I/O routines. Change most of
7145 the *_d2i_bio() and *_d2i_fp() functions to use these.
7148 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7149 lines, recognice more "algorithms" that can be deselected, and make
7150 it complain about algorithm deselection that isn't recognised.
7153 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7154 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7155 to use new functions. Add NO_ASN1_OLD which can be set to remove
7156 some old style ASN1 functions: this can be used to determine if old
7157 code will still work when these eventually go away.
7160 *) New extension functions for OCSP structures, these follow the
7161 same conventions as certificates and CRLs.
7164 *) New function X509V3_add1_i2d(). This automatically encodes and
7165 adds an extension. Its behaviour can be customised with various
7166 flags to append, replace or delete. Various wrappers added for
7167 certifcates and CRLs.
7170 *) Fix to avoid calling the underlying ASN1 print routine when
7171 an extension cannot be parsed. Correct a typo in the
7172 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7175 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7176 entries for variables.
7179 *) Add functionality to apps/openssl.c for detecting locking
7180 problems: As the program is single-threaded, all we have
7181 to do is register a locking callback using an array for
7182 storing which locks are currently held by the program.
7185 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7186 SSL_get_ex_data_X509_STORE_idx(), which is used in
7187 ssl_verify_cert_chain() and thus can be called at any time
7188 during TLS/SSL handshakes so that thread-safety is essential.
7189 Unfortunately, the ex_data design is not at all suited
7190 for multi-threaded use, so it probably should be abolished.
7193 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7194 [Broadcom, tweaked and integrated by Geoff Thorpe]
7196 *) Move common extension printing code to new function
7197 X509V3_print_extensions(). Reorganise OCSP print routines and
7198 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7201 *) New function X509_signature_print() to remove duplication in some
7205 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7206 set (this was treated exactly the same as SET OF previously). This
7207 is used to reorder the STACK representing the structure to match the
7208 encoding. This will be used to get round a problem where a PKCS7
7209 structure which was signed could not be verified because the STACK
7210 order did not reflect the encoded order.
7213 *) Reimplement the OCSP ASN1 module using the new code.
7216 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7217 for its ASN1 operations. The old style function pointers still exist
7218 for now but they will eventually go away.
7221 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7222 completely replaces the old ASN1 functionality with a table driven
7223 encoder and decoder which interprets an ASN1_ITEM structure describing
7224 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7225 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7226 has also been converted to the new form.
7229 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7230 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7231 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7232 for negative moduli.
7235 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7236 of not touching the result's sign bit.
7239 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7243 *) Changed the LHASH code to use prototypes for callbacks, and created
7244 macros to declare and implement thin (optionally static) functions
7245 that provide type-safety and avoid function pointer casting for the
7246 type-specific callbacks.
7249 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7251 [Veers Staats <staatsvr@asc.hpc.mil>,
7252 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7254 *) Reformat the FAQ so the different questions and answers can be divided
7255 in sections depending on the subject.
7258 *) Have the zlib compression code load ZLIB.DLL dynamically under
7262 *) New function BN_mod_sqrt for computing square roots modulo a prime
7263 (using the probabilistic Tonelli-Shanks algorithm unless
7264 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7265 be handled deterministically).
7266 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7268 *) Make BN_mod_inverse faster by explicitly handling small quotients
7269 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7270 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7273 *) New function BN_kronecker.
7276 *) Fix BN_gcd so that it works on negative inputs; the result is
7277 positive unless both parameters are zero.
7278 Previously something reasonably close to an infinite loop was
7279 possible because numbers could be growing instead of shrinking
7280 in the implementation of Euclid's algorithm.
7283 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7284 sign of the number in question.
7286 Fix BN_is_word(a,w) to work correctly for w == 0.
7288 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7289 because its test if the absolute value of 'a' equals 'w'.
7290 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7291 it exists mostly for use in the implementations of BN_is_zero(),
7292 BN_is_one(), and BN_is_word().
7295 *) New function BN_swap.
7298 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7299 the exponentiation functions are more likely to produce reasonable
7300 results on negative inputs.
7303 *) Change BN_mod_mul so that the result is always non-negative.
7304 Previously, it could be negative if one of the factors was negative;
7305 I don't think anyone really wanted that behaviour.
7308 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7309 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7310 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7311 and add new functions:
7320 BN_mod_lshift1_quick
7324 These functions always generate non-negative results.
7326 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7327 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7329 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7330 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7331 be reduced modulo m.
7332 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7335 The following entry accidentally appeared in the CHANGES file
7336 distributed with OpenSSL 0.9.7. The modifications described in
7337 it do *not* apply to OpenSSL 0.9.7.
7339 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7340 was actually never needed) and in BN_mul(). The removal in BN_mul()
7341 required a small change in bn_mul_part_recursive() and the addition
7342 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7343 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7344 bn_sub_words() and bn_add_words() except they take arrays with
7349 *) In 'openssl passwd', verify passwords read from the terminal
7350 unless the '-salt' option is used (which usually means that
7351 verification would just waste user's time since the resulting
7352 hash is going to be compared with some given password hash)
7353 or the new '-noverify' option is used.
7355 This is an incompatible change, but it does not affect
7356 non-interactive use of 'openssl passwd' (passwords on the command
7357 line, '-stdin' option, '-in ...' option) and thus should not
7361 *) Remove all references to RSAref, since there's no more need for it.
7364 *) Make DSO load along a path given through an environment variable
7365 (SHLIB_PATH) with shl_load().
7368 *) Constify the ENGINE code as a result of BIGNUM constification.
7369 Also constify the RSA code and most things related to it. In a
7370 few places, most notable in the depth of the ASN.1 code, ugly
7371 casts back to non-const were required (to be solved at a later
7375 *) Make it so the openssl application has all engines loaded by default.
7378 *) Constify the BIGNUM routines a little more.
7381 *) Add the following functions:
7383 ENGINE_load_cswift()
7385 ENGINE_load_atalla()
7387 ENGINE_load_builtin_engines()
7389 That way, an application can itself choose if external engines that
7390 are built-in in OpenSSL shall ever be used or not. The benefit is
7391 that applications won't have to be linked with libdl or other dso
7392 libraries unless it's really needed.
7394 Changed 'openssl engine' to load all engines on demand.
7395 Changed the engine header files to avoid the duplication of some
7396 declarations (they differed!).
7399 *) 'openssl engine' can now list capabilities.
7402 *) Better error reporting in 'openssl engine'.
7405 *) Never call load_dh_param(NULL) in s_server.
7408 *) Add engine application. It can currently list engines by name and
7409 identity, and test if they are actually available.
7412 *) Improve RPM specification file by forcing symbolic linking and making
7413 sure the installed documentation is also owned by root.root.
7414 [Damien Miller <djm@mindrot.org>]
7416 *) Give the OpenSSL applications more possibilities to make use of
7417 keys (public as well as private) handled by engines.
7420 *) Add OCSP code that comes from CertCo.
7423 *) Add VMS support for the Rijndael code.
7426 *) Added untested support for Nuron crypto accelerator.
7429 *) Add support for external cryptographic devices. This code was
7430 previously distributed separately as the "engine" branch.
7431 [Geoff Thorpe, Richard Levitte]
7433 *) Rework the filename-translation in the DSO code. It is now possible to
7434 have far greater control over how a "name" is turned into a filename
7435 depending on the operating environment and any oddities about the
7436 different shared library filenames on each system.
7439 *) Support threads on FreeBSD-elf in Configure.
7442 *) Fix for SHA1 assembly problem with MASM: it produces
7443 warnings about corrupt line number information when assembling
7444 with debugging information. This is caused by the overlapping
7446 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7449 NCONF_get_number() has no error checking at all. As a replacement,
7450 NCONF_get_number_e() is defined (_e for "error checking") and is
7451 promoted strongly. The old NCONF_get_number is kept around for
7452 binary backward compatibility.
7453 Make it possible for methods to load from something other than a BIO,
7454 by providing a function pointer that is given a name instead of a BIO.
7455 For example, this could be used to load configuration data from an
7459 *) Fix for non blocking accept BIOs. Added new I/O special reason
7460 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7461 with non blocking I/O was not possible because no retry code was
7462 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7466 *) Added the beginnings of Rijndael support.
7469 *) Fix for bug in DirectoryString mask setting. Add support for
7470 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7471 to allow certificate printing to more controllable, additional
7472 'certopt' option to 'x509' to allow new printing options to be
7476 *) Clean old EAY MD5 hack from e_os.h.
7479 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7481 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7482 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7483 [Joe Orton, Steve Henson]
7485 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7487 *) Fix additional bug revealed by the NISCC test suite:
7489 Stop bug triggering large recursion when presented with
7490 certain ASN.1 tags (CVE-2003-0851)
7493 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7495 *) Fix various bugs revealed by running the NISCC test suite:
7497 Stop out of bounds reads in the ASN1 code when presented with
7498 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7500 If verify callback ignores invalid public key errors don't try to check
7501 certificate signature with the NULL public key.
7505 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7506 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7510 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7511 extra data after the compression methods not only for TLS 1.0
7512 but also for SSL 3.0 (as required by the specification).
7513 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7515 *) Change X509_certificate_type() to mark the key as exported/exportable
7516 when it's 512 *bits* long, not 512 bytes.
7519 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7521 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7522 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7523 a protocol version number mismatch like a decryption error
7524 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7527 *) Turn on RSA blinding by default in the default implementation
7528 to avoid a timing attack. Applications that don't want it can call
7529 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7530 They would be ill-advised to do so in most cases.
7531 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7533 *) Change RSA blinding code so that it works when the PRNG is not
7534 seeded (in this case, the secret RSA exponent is abused as
7535 an unpredictable seed -- if it is not unpredictable, there
7536 is no point in blinding anyway). Make RSA blinding thread-safe
7537 by remembering the creator's thread ID in rsa->blinding and
7538 having all other threads use local one-time blinding factors
7539 (this requires more computation than sharing rsa->blinding, but
7540 avoids excessive locking; and if an RSA object is not shared
7541 between threads, blinding will still be very fast).
7544 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7546 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7547 via timing by performing a MAC computation even if incorrrect
7548 block cipher padding has been found. This is a countermeasure
7549 against active attacks where the attacker has to distinguish
7550 between bad padding and a MAC verification error. (CVE-2003-0078)
7552 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7553 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7554 Martin Vuagnoux (EPFL, Ilion)]
7556 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7558 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7559 memory from it's contents. This is done with a counter that will
7560 place alternating values in each byte. This can be used to solve
7561 two issues: 1) the removal of calls to memset() by highly optimizing
7562 compilers, and 2) cleansing with other values than 0, since those can
7563 be read through on certain media, for example a swap space on disk.
7566 *) Bugfix: client side session caching did not work with external caching,
7567 because the session->cipher setting was not restored when reloading
7568 from the external cache. This problem was masked, when
7569 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7570 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7573 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7574 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7575 [Zeev Lieber <zeev-l@yahoo.com>]
7577 *) Undo an undocumented change introduced in 0.9.6e which caused
7578 repeated calls to OpenSSL_add_all_ciphers() and
7579 OpenSSL_add_all_digests() to be ignored, even after calling
7583 *) Change the default configuration reader to deal with last line not
7584 being properly terminated.
7587 *) Change X509_NAME_cmp() so it applies the special rules on handling
7588 DN values that are of type PrintableString, as well as RDNs of type
7589 emailAddress where the value has the type ia5String.
7590 [stefank@valicert.com via Richard Levitte]
7592 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7593 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7594 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7595 the bitwise-OR of the two for use by the majority of applications
7596 wanting this behaviour, and update the docs. The documented
7597 behaviour and actual behaviour were inconsistent and had been
7598 changing anyway, so this is more a bug-fix than a behavioural
7600 [Geoff Thorpe, diagnosed by Nadav Har'El]
7602 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7603 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7606 *) Fix initialization code race conditions in
7607 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7608 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7609 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7610 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7611 ssl2_get_cipher_by_char(),
7612 ssl3_get_cipher_by_char().
7613 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7615 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7616 the cached sessions are flushed, as the remove_cb() might use ex_data
7617 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7618 (see [openssl.org #212]).
7619 [Geoff Thorpe, Lutz Jaenicke]
7621 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7622 length, instead of the encoding length to d2i_ASN1_OBJECT.
7625 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7627 *) [In 0.9.6g-engine release:]
7628 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7629 [Lynn Gazis <lgazis@rainbow.com>]
7631 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7633 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7634 and get fix the header length calculation.
7635 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7636 Alon Kantor <alonk@checkpoint.com> (and others),
7639 *) Use proper error handling instead of 'assertions' in buffer
7640 overflow checks added in 0.9.6e. This prevents DoS (the
7641 assertions could call abort()).
7642 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7644 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7646 *) Add various sanity checks to asn1_get_length() to reject
7647 the ASN1 length bytes if they exceed sizeof(long), will appear
7648 negative or the content length exceeds the length of the
7650 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7652 *) Fix cipher selection routines: ciphers without encryption had no flags
7653 for the cipher strength set and where therefore not handled correctly
7654 by the selection routines (PR #130).
7657 *) Fix EVP_dsa_sha macro.
7661 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7662 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7663 that was added in OpenSSL 0.9.6d.
7665 As the countermeasure turned out to be incompatible with some
7666 broken SSL implementations, the new option is part of SSL_OP_ALL.
7667 SSL_OP_ALL is usually employed when compatibility with weird SSL
7668 implementations is desired (e.g. '-bugs' option to 's_client' and
7669 's_server'), so the new option is automatically set in many
7673 *) Changes in security patch:
7675 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7676 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7677 Air Force Materiel Command, USAF, under agreement number
7680 *) Add various sanity checks to asn1_get_length() to reject
7681 the ASN1 length bytes if they exceed sizeof(long), will appear
7682 negative or the content length exceeds the length of the
7683 supplied buffer. (CVE-2002-0659)
7684 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7686 *) Assertions for various potential buffer overflows, not known to
7688 [Ben Laurie (CHATS)]
7690 *) Various temporary buffers to hold ASCII versions of integers were
7691 too small for 64 bit platforms. (CVE-2002-0655)
7692 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7694 *) Remote buffer overflow in SSL3 protocol - an attacker could
7695 supply an oversized session ID to a client. (CVE-2002-0656)
7696 [Ben Laurie (CHATS)]
7698 *) Remote buffer overflow in SSL2 protocol - an attacker could
7699 supply an oversized client master key. (CVE-2002-0656)
7700 [Ben Laurie (CHATS)]
7702 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7704 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7705 encoded as NULL) with id-dsa-with-sha1.
7706 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7708 *) Check various X509_...() return values in apps/req.c.
7709 [Nils Larsch <nla@trustcenter.de>]
7711 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7712 an end-of-file condition would erronously be flagged, when the CRLF
7713 was just at the end of a processed block. The bug was discovered when
7714 processing data through a buffering memory BIO handing the data to a
7715 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7716 <ptsekov@syntrex.com> and Nedelcho Stanev.
7719 *) Implement a countermeasure against a vulnerability recently found
7720 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7721 before application data chunks to avoid the use of known IVs
7722 with data potentially chosen by the attacker.
7725 *) Fix length checks in ssl3_get_client_hello().
7728 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7729 to prevent ssl3_read_internal() from incorrectly assuming that
7730 ssl3_read_bytes() found application data while handshake
7731 processing was enabled when in fact s->s3->in_read_app_data was
7732 merely automatically cleared during the initial handshake.
7733 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7735 *) Fix object definitions for Private and Enterprise: they were not
7736 recognized in their shortname (=lowercase) representation. Extend
7737 obj_dat.pl to issue an error when using undefined keywords instead
7738 of silently ignoring the problem (Svenning Sorensen
7739 <sss@sss.dnsalias.net>).
7742 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7743 generators, i.e. generators other than 2 and 5. (Previously, the
7744 code did not properly initialise the 'add' and 'rem' values to
7745 BN_generate_prime().)
7747 In the new general case, we do not insist that 'generator' is
7748 actually a primitive root: This requirement is rather pointless;
7749 a generator of the order-q subgroup is just as good, if not
7753 *) Map new X509 verification errors to alerts. Discovered and submitted by
7754 Tom Wu <tom@arcot.com>.
7757 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7758 returning non-zero before the data has been completely received
7759 when using non-blocking I/O.
7760 [Bodo Moeller; problem pointed out by John Hughes]
7762 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7763 [Ben Laurie, Lutz Jaenicke]
7765 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7766 Yoram Zahavi <YoramZ@gilian.com>).
7769 *) Add information about CygWin 1.3 and on, and preserve proper
7770 configuration for the versions before that.
7771 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7773 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7774 check whether we deal with a copy of a session and do not delete from
7775 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7776 <izhar@checkpoint.com>.
7779 *) Do not store session data into the internal session cache, if it
7780 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7781 flag is set). Proposed by Aslam <aslam@funk.com>.
7784 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7788 *) [In 0.9.6d-engine release:]
7789 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7790 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7792 *) Add the configuration target linux-s390x.
7793 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7795 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7796 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7797 variable as an indication that a ClientHello message has been
7798 received. As the flag value will be lost between multiple
7799 invocations of ssl3_accept when using non-blocking I/O, the
7800 function may not be aware that a handshake has actually taken
7801 place, thus preventing a new session from being added to the
7804 To avoid this problem, we now set s->new_session to 2 instead of
7805 using a local variable.
7806 [Lutz Jaenicke, Bodo Moeller]
7808 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7809 if the SSL_R_LENGTH_MISMATCH error is detected.
7810 [Geoff Thorpe, Bodo Moeller]
7812 *) New 'shared_ldflag' column in Configure platform table.
7815 *) Fix EVP_CIPHER_mode macro.
7816 ["Dan S. Camper" <dan@bti.net>]
7818 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7819 type, we must throw them away by setting rr->length to 0.
7820 [D P Chang <dpc@qualys.com>]
7822 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7824 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7825 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7826 worked incorrectly for those cases where range = 10..._2 and
7827 3*range is two bits longer than range.)
7830 *) Only add signing time to PKCS7 structures if it is not already
7834 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7835 OBJ_ld_ce should be OBJ_id_ce.
7836 Also some ip-pda OIDs in crypto/objects/objects.txt were
7837 incorrect (cf. RFC 3039).
7838 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7840 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7841 returns early because it has nothing to do.
7842 [Andy Schneider <andy.schneider@bjss.co.uk>]
7844 *) [In 0.9.6c-engine release:]
7845 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7846 [Andy Schneider <andy.schneider@bjss.co.uk>]
7848 *) [In 0.9.6c-engine release:]
7849 Add support for Cryptographic Appliance's keyserver technology.
7850 (Use engine 'keyclient')
7851 [Cryptographic Appliances and Geoff Thorpe]
7853 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7854 is called via tools/c89.sh because arguments have to be
7855 rearranged (all '-L' options must appear before the first object
7857 [Richard Shapiro <rshapiro@abinitio.com>]
7859 *) [In 0.9.6c-engine release:]
7860 Add support for Broadcom crypto accelerator cards, backported
7862 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7864 *) [In 0.9.6c-engine release:]
7865 Add support for SureWare crypto accelerator cards from
7866 Baltimore Technologies. (Use engine 'sureware')
7867 [Baltimore Technologies and Mark Cox]
7869 *) [In 0.9.6c-engine release:]
7870 Add support for crypto accelerator cards from Accelerated
7871 Encryption Processing, www.aep.ie. (Use engine 'aep')
7872 [AEP Inc. and Mark Cox]
7874 *) Add a configuration entry for gcc on UnixWare.
7875 [Gary Benson <gbenson@redhat.com>]
7877 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7878 messages are stored in a single piece (fixed-length part and
7879 variable-length part combined) and fix various bugs found on the way.
7882 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7883 instead. BIO_gethostbyname() does not know what timeouts are
7884 appropriate, so entries would stay in cache even when they have
7886 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7888 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7889 faced with a pathologically small ClientHello fragment that does
7890 not contain client_version: Instead of aborting with an error,
7891 simply choose the highest available protocol version (i.e.,
7892 TLS 1.0 unless it is disabled). In practice, ClientHello
7893 messages are never sent like this, but this change gives us
7894 strictly correct behaviour at least for TLS.
7897 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7898 never resets s->method to s->ctx->method when called from within
7899 one of the SSL handshake functions.
7900 [Bodo Moeller; problem pointed out by Niko Baric]
7902 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7903 (sent using the client's version number) if client_version is
7904 smaller than the protocol version in use. Also change
7905 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7906 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7907 the client will at least see that alert.
7910 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7914 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7915 client receives HelloRequest while in a handshake.
7916 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7918 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7919 should end in 'break', not 'goto end' which circumvents various
7920 cleanups done in state SSL_ST_OK. But session related stuff
7921 must be disabled for SSL_ST_OK in the case that we just sent a
7924 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7925 before just sending a HelloRequest.
7926 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7928 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7929 reveal whether illegal block cipher padding was found or a MAC
7930 verification error occurred. (Neither SSLerr() codes nor alerts
7931 are directly visible to potential attackers, but the information
7932 may leak via logfiles.)
7934 Similar changes are not required for the SSL 2.0 implementation
7935 because the number of padding bytes is sent in clear for SSL 2.0,
7936 and the extra bytes are just ignored. However ssl/s2_pkt.c
7937 failed to verify that the purported number of padding bytes is in
7941 *) Add OpenUNIX-8 support including shared libraries
7942 (Boyd Lynn Gerber <gerberb@zenez.com>).
7945 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7946 'wristwatch attack' using huge encoding parameters (cf.
7947 James H. Manger's CRYPTO 2001 paper). Note that the
7948 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7949 encoding parameters and hence was not vulnerable.
7952 *) BN_sqr() bug fix.
7953 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7955 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7956 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7957 followed by modular reduction.
7958 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7960 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7961 equivalent based on BN_pseudo_rand() instead of BN_rand().
7964 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7965 This function was broken, as the check for a new client hello message
7966 to handle SGC did not allow these large messages.
7967 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7970 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7973 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7974 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7977 *) Rework the configuration and shared library support for Tru64 Unix.
7978 The configuration part makes use of modern compiler features and
7979 still retains old compiler behavior for those that run older versions
7980 of the OS. The shared library support part includes a variant that
7981 uses the RPATH feature, and is available through the special
7982 configuration target "alpha-cc-rpath", which will never be selected
7984 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7986 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7987 with the same message size as in ssl3_get_certificate_request().
7988 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7989 messages might inadvertently be reject as too long.
7990 [Petr Lampa <lampa@fee.vutbr.cz>]
7992 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7995 *) Modified SSL library such that the verify_callback that has been set
7996 specificly for an SSL object with SSL_set_verify() is actually being
7997 used. Before the change, a verify_callback set with this function was
7998 ignored and the verify_callback() set in the SSL_CTX at the time of
7999 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8000 to allow the necessary settings.
8003 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8004 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8005 done automatically (in contradiction to the requirements of the C
8006 standard). This made problems when used from OpenSSH.
8009 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8010 dh->length and always used
8012 BN_rand_range(priv_key, dh->p).
8014 BN_rand_range() is not necessary for Diffie-Hellman, and this
8015 specific range makes Diffie-Hellman unnecessarily inefficient if
8016 dh->length (recommended exponent length) is much smaller than the
8017 length of dh->p. We could use BN_rand_range() if the order of
8018 the subgroup was stored in the DH structure, but we only have
8023 BN_rand(priv_key, l, ...)
8025 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8031 RSA_eay_public_encrypt
8032 RSA_eay_private_decrypt
8033 RSA_eay_private_encrypt (signing)
8034 RSA_eay_public_decrypt (signature verification)
8036 (default implementations for RSA_public_encrypt,
8037 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8038 always reject numbers >= n.
8041 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8042 to synchronize access to 'locking_thread'. This is necessary on
8043 systems where access to 'locking_thread' (an 'unsigned long'
8044 variable) is not atomic.
8047 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8048 *before* setting the 'crypto_lock_rand' flag. The previous code had
8049 a race condition if 0 is a valid thread ID.
8050 [Travis Vitek <vitek@roguewave.com>]
8052 *) Add support for shared libraries under Irix.
8053 [Albert Chin-A-Young <china@thewrittenword.com>]
8055 *) Add configuration option to build on Linux on both big-endian and
8057 [Ralf Baechle <ralf@uni-koblenz.de>]
8059 *) Add the possibility to create shared libraries on HP-UX.
8062 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8064 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8065 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8066 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8067 PRNG state recovery was possible based on the output of
8068 one PRNG request appropriately sized to gain knowledge on
8069 'md' followed by enough consecutive 1-byte PRNG requests
8070 to traverse all of 'state'.
8072 1. When updating 'md_local' (the current thread's copy of 'md')
8073 during PRNG output generation, hash all of the previous
8074 'md_local' value, not just the half used for PRNG output.
8076 2. Make the number of bytes from 'state' included into the hash
8077 independent from the number of PRNG bytes requested.
8079 The first measure alone would be sufficient to avoid
8080 Markku-Juhani's attack. (Actually it had never occurred
8081 to me that the half of 'md_local' used for chaining was the
8082 half from which PRNG output bytes were taken -- I had always
8083 assumed that the secret half would be used.) The second
8084 measure makes sure that additional data from 'state' is never
8085 mixed into 'md_local' in small portions; this heuristically
8086 further strengthens the PRNG.
8089 *) Fix crypto/bn/asm/mips3.s.
8092 *) When only the key is given to "enc", the IV is undefined. Print out
8093 an error message in this case.
8096 *) Handle special case when X509_NAME is empty in X509 printing routines.
8099 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8100 positive and less than q.
8103 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8104 used: it isn't thread safe and the add_lock_callback should handle
8106 [Paul Rose <Paul.Rose@bridge.com>]
8108 *) Verify that incoming data obeys the block size in
8109 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8113 [Ulf Möller, Bodo Möller]
8115 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8116 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8117 when fixing the server behaviour for backwards-compatible 'client
8118 hello' messages. (Note that the attack is impractical against
8119 SSL 3.0 and TLS 1.0 anyway because length and version checking
8120 means that the probability of guessing a valid ciphertext is
8121 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8124 Before 0.9.5, the countermeasure (hide the error by generating a
8125 random 'decryption result') did not work properly because
8126 ERR_clear_error() was missing, meaning that SSL_get_error() would
8127 detect the supposedly ignored error.
8129 Both problems are now fixed.
8132 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8133 (previously it was 1024).
8136 *) Fix for compatibility mode trust settings: ignore trust settings
8137 unless some valid trust or reject settings are present.
8140 *) Fix for blowfish EVP: its a variable length cipher.
8143 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8144 parameters in DSA public key structures and return an error in the
8145 DSA routines if parameters are absent.
8148 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8149 in the current directory if neither $RANDFILE nor $HOME was set.
8150 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8151 caused some confusion to Windows users who haven't defined $HOME.
8152 Thus RAND_file_name() is changed again: e_os.h can define a
8153 DEFAULT_HOME, which will be used if $HOME is not set.
8154 For Windows, we use "C:"; on other platforms, we still require
8155 environment variables.
8157 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8158 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8159 having multiple threads call RAND_poll() concurrently.
8162 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8163 combination of a flag and a thread ID variable.
8164 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8165 flag), *other* threads can enter ssleay_add_bytes without obeying
8166 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8167 that they do not hold after the first thread unsets add_do_not_lock).
8170 *) Change bctest again: '-x' expressions are not available in all
8174 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8176 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8177 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8179 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8180 the default extension for executables, if any. Also, make the perl
8181 scripts that use symlink() to test if it really exists and use "cp"
8182 if it doesn't. All this made OpenSSL compilable and installable in
8186 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8187 If SEQUENCE is length is indefinite just set c->slen to the total
8188 amount of data available.
8189 [Steve Henson, reported by shige@FreeBSD.org]
8190 [This change does not apply to 0.9.7.]
8192 *) Change bctest to avoid here-documents inside command substitution
8193 (workaround for FreeBSD /bin/sh bug).
8194 For compatibility with Ultrix, avoid shell functions (introduced
8195 in the bctest version that searches along $PATH).
8198 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8199 with des_encrypt() defined on some operating systems, like Solaris
8203 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8204 On the Importance of Eliminating Errors in Cryptographic
8205 Computations, J. Cryptology 14 (2001) 2, 101-119,
8206 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8209 *) MIPS assembler BIGNUM division bug fix.
8212 *) Disabled incorrect Alpha assembler code.
8215 *) Fix PKCS#7 decode routines so they correctly update the length
8216 after reading an EOC for the EXPLICIT tag.
8218 [This change does not apply to 0.9.7.]
8220 *) Fix bug in PKCS#12 key generation routines. This was triggered
8221 if a 3DES key was generated with a 0 initial byte. Include
8222 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8223 (but broken) behaviour.
8226 *) Enhance bctest to search for a working bc along $PATH and print
8228 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8230 *) Fix memory leaks in err.c: free err_data string if necessary;
8231 don't write to the wrong index in ERR_set_error_data.
8234 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8238 *) Replace rdtsc with _emit statements for VC++ version 5.
8239 [Jeremy Cooper <jeremy@baymoo.org>]
8241 *) Make it possible to reuse SSLv2 sessions.
8244 *) In copy_email() check for >= 0 as a return value for
8245 X509_NAME_get_index_by_NID() since 0 is a valid index.
8246 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8248 *) Avoid coredump with unsupported or invalid public keys by checking if
8249 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8250 PKCS7_verify() fails with non detached data.
8253 *) Don't use getenv in library functions when run as setuid/setgid.
8254 New function OPENSSL_issetugid().
8257 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8258 due to incorrect handling of multi-threading:
8260 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8262 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8264 3. Count how many times MemCheck_off() has been called so that
8265 nested use can be treated correctly. This also avoids
8266 inband-signalling in the previous code (which relied on the
8267 assumption that thread ID 0 is impossible).
8270 *) Add "-rand" option also to s_client and s_server.
8273 *) Fix CPU detection on Irix 6.x.
8274 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8275 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8277 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8280 [This change does not apply to 0.9.7.]
8282 *) Use the cached encoding of an X509_NAME structure rather than
8283 copying it. This is apparently the reason for the libsafe "errors"
8284 but the code is actually correct.
8287 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8288 Bleichenbacher's DSA attack.
8289 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8290 to be set and top=0 forces the highest bit to be set; top=-1 is new
8291 and leaves the highest bit random.
8292 [Ulf Moeller, Bodo Moeller]
8294 *) In the NCONF_...-based implementations for CONF_... queries
8295 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8296 a temporary CONF structure with the data component set to NULL
8297 (which gives segmentation faults in lh_retrieve).
8298 Instead, use NULL for the CONF pointer in CONF_get_string and
8299 CONF_get_number (which may use environment variables) and directly
8300 return NULL from CONF_get_section.
8303 *) Fix potential buffer overrun for EBCDIC.
8306 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8307 keyUsage if basicConstraints absent for a CA.
8310 *) Make SMIME_write_PKCS7() write mail header values with a format that
8311 is more generally accepted (no spaces before the semicolon), since
8312 some programs can't parse those values properly otherwise. Also make
8313 sure BIO's that break lines after each write do not create invalid
8317 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8318 macros previously used would not encode an empty SEQUENCE OF
8319 and break the signature.
8321 [This change does not apply to 0.9.7.]
8323 *) Zero the premaster secret after deriving the master secret in
8327 *) Add some EVP_add_digest_alias registrations (as found in
8328 OpenSSL_add_all_digests()) to SSL_library_init()
8329 aka OpenSSL_add_ssl_algorithms(). This provides improved
8330 compatibility with peers using X.509 certificates
8331 with unconventional AlgorithmIdentifier OIDs.
8334 *) Fix for Irix with NO_ASM.
8335 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8337 *) ./config script fixes.
8338 [Ulf Moeller, Richard Levitte]
8340 *) Fix 'openssl passwd -1'.
8343 *) Change PKCS12_key_gen_asc() so it can cope with non null
8344 terminated strings whose length is passed in the passlen
8345 parameter, for example from PEM callbacks. This was done
8346 by adding an extra length parameter to asc2uni().
8347 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8349 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8350 call failed, free the DSA structure.
8353 *) Fix to uni2asc() to cope with zero length Unicode strings.
8354 These are present in some PKCS#12 files.
8357 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8358 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8359 when writing a 32767 byte record.
8360 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8362 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8363 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8365 (RSA objects have a reference count access to which is protected
8366 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8367 so they are meant to be shared between threads.)
8368 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8369 "Reddie, Steven" <Steven.Reddie@ca.com>]
8371 *) Fix a deadlock in CRYPTO_mem_leaks().
8374 *) Use better test patterns in bntest.
8377 *) rand_win.c fix for Borland C.
8380 *) BN_rshift bugfix for n == 0.
8383 *) Add a 'bctest' script that checks for some known 'bc' bugs
8384 so that 'make test' does not abort just because 'bc' is broken.
8387 *) Store verify_result within SSL_SESSION also for client side to
8388 avoid potential security hole. (Re-used sessions on the client side
8389 always resulted in verify_result==X509_V_OK, not using the original
8390 result of the server certificate verification.)
8393 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8394 SSL3_RT_APPLICATION_DATA, return 0.
8395 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8399 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8400 releases, have been re-implemented by renaming the previous
8401 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8402 and ssl3_read_internal, respectively, and adding 'peek' parameters
8403 to them. The new ssl[23]_{read,peek} functions are calls to
8404 ssl[23]_read_internal with the 'peek' flag set appropriately.
8405 A 'peek' parameter has also been added to ssl3_read_bytes, which
8406 does the actual work for ssl3_read_internal.
8409 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8410 the method-specific "init()" handler. Also clean up ex_data after
8411 calling the method-specific "finish()" handler. Previously, this was
8412 happening the other way round.
8415 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8416 The previous value, 12, was not always sufficient for BN_mod_exp().
8419 *) Make sure that shared libraries get the internal name engine with
8420 the full version number and not just 0. This should mark the
8421 shared libraries as not backward compatible. Of course, this should
8422 be changed again when we can guarantee backward binary compatibility.
8425 *) Fix typo in get_cert_by_subject() in by_dir.c
8426 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8428 *) Rework the system to generate shared libraries:
8430 - Make note of the expected extension for the shared libraries and
8431 if there is a need for symbolic links from for example libcrypto.so.0
8432 to libcrypto.so.0.9.7. There is extended info in Configure for
8435 - Make as few rebuilds of the shared libraries as possible.
8437 - Still avoid linking the OpenSSL programs with the shared libraries.
8439 - When installing, install the shared libraries separately from the
8443 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8445 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8446 and not in SSL_clear because the latter is also used by the
8447 accept/connect functions; previously, the settings made by
8448 SSL_set_read_ahead would be lost during the handshake.
8449 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8451 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8452 Previously, it would create entries for disabled algorithms no
8456 *) Added several new manual pages for SSL_* function.
8459 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8461 *) In ssl23_get_client_hello, generate an error message when faced
8462 with an initial SSL 3.0/TLS record that is too small to contain the
8463 first two bytes of the ClientHello message, i.e. client_version.
8464 (Note that this is a pathologic case that probably has never happened
8465 in real life.) The previous approach was to use the version number
8466 from the record header as a substitute; but our protocol choice
8467 should not depend on that one because it is not authenticated
8468 by the Finished messages.
8471 *) More robust randomness gathering functions for Windows.
8472 [Jeffrey Altman <jaltman@columbia.edu>]
8474 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8475 not set then we don't setup the error code for issuer check errors
8476 to avoid possibly overwriting other errors which the callback does
8477 handle. If an application does set the flag then we assume it knows
8478 what it is doing and can handle the new informational codes
8482 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8483 a general "ANY" type, as such it should be able to decode anything
8484 including tagged types. However it didn't check the class so it would
8485 wrongly interpret tagged types in the same way as their universal
8486 counterpart and unknown types were just rejected. Changed so that the
8487 tagged and unknown types are handled in the same way as a SEQUENCE:
8488 that is the encoding is stored intact. There is also a new type
8489 "V_ASN1_OTHER" which is used when the class is not universal, in this
8490 case we have no idea what the actual type is so we just lump them all
8494 *) On VMS, stdout may very well lead to a file that is written to
8495 in a record-oriented fashion. That means that every write() will
8496 write a separate record, which will be read separately by the
8497 programs trying to read from it. This can be very confusing.
8499 The solution is to put a BIO filter in the way that will buffer
8500 text until a linefeed is reached, and then write everything a
8501 line at a time, so every record written will be an actual line,
8502 not chunks of lines and not (usually doesn't happen, but I've
8503 seen it once) several lines in one record. BIO_f_linebuffer() is
8506 Currently, it's a VMS-only method, because that's where it has
8507 been tested well enough.
8510 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8511 it can return incorrect results.
8512 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8513 but it was in 0.9.6-beta[12].)
8516 *) Disable the check for content being present when verifying detached
8517 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8518 include zero length content when signing messages.
8521 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8522 BIO_ctrl (for BIO pairs).
8525 *) Add DSO method for VMS.
8528 *) Bug fix: Montgomery multiplication could produce results with the
8532 *) Add RPM specification openssl.spec and modify it to build three
8533 packages. The default package contains applications, application
8534 documentation and run-time libraries. The devel package contains
8535 include files, static libraries and function documentation. The
8536 doc package contains the contents of the doc directory. The original
8537 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8540 *) Add a large number of documentation files for many SSL routines.
8541 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8543 *) Add a configuration entry for Sony News 4.
8544 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8546 *) Don't set the two most significant bits to one when generating a
8547 random number < q in the DSA library.
8550 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8551 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8552 the underlying transport is blocking) if a handshake took place.
8553 (The default behaviour is needed by applications such as s_client
8554 and s_server that use select() to determine when to use SSL_read;
8555 but for applications that know in advance when to expect data, it
8556 just makes things more complicated.)
8559 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8563 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8564 work better on such systems.
8565 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8567 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8568 Update PKCS12_parse() so it copies the friendlyName and the
8569 keyid to the certificates aux info.
8572 *) Fix bug in PKCS7_verify() which caused an infinite loop
8573 if there was more than one signature.
8574 [Sven Uszpelkat <su@celocom.de>]
8576 *) Major change in util/mkdef.pl to include extra information
8577 about each symbol, as well as presenting variables as well
8578 as functions. This change means that there's n more need
8579 to rebuild the .num files when some algorithms are excluded.
8582 *) Allow the verify time to be set by an application,
8583 rather than always using the current time.
8586 *) Phase 2 verify code reorganisation. The certificate
8587 verify code now looks up an issuer certificate by a
8588 number of criteria: subject name, authority key id
8589 and key usage. It also verifies self signed certificates
8590 by the same criteria. The main comparison function is
8591 X509_check_issued() which performs these checks.
8593 Lot of changes were necessary in order to support this
8594 without completely rewriting the lookup code.
8596 Authority and subject key identifier are now cached.
8598 The LHASH 'certs' is X509_STORE has now been replaced
8599 by a STACK_OF(X509_OBJECT). This is mainly because an
8600 LHASH can't store or retrieve multiple objects with
8601 the same hash value.
8603 As a result various functions (which were all internal
8604 use only) have changed to handle the new X509_STORE
8605 structure. This will break anything that messed round
8606 with X509_STORE internally.
8608 The functions X509_STORE_add_cert() now checks for an
8609 exact match, rather than just subject name.
8611 The X509_STORE API doesn't directly support the retrieval
8612 of multiple certificates matching a given criteria, however
8613 this can be worked round by performing a lookup first
8614 (which will fill the cache with candidate certificates)
8615 and then examining the cache for matches. This is probably
8616 the best we can do without throwing out X509_LOOKUP
8617 entirely (maybe later...).
8619 The X509_VERIFY_CTX structure has been enhanced considerably.
8621 All certificate lookup operations now go via a get_issuer()
8622 callback. Although this currently uses an X509_STORE it
8623 can be replaced by custom lookups. This is a simple way
8624 to bypass the X509_STORE hackery necessary to make this
8625 work and makes it possible to use more efficient techniques
8626 in future. A very simple version which uses a simple
8627 STACK for its trusted certificate store is also provided
8628 using X509_STORE_CTX_trusted_stack().
8630 The verify_cb() and verify() callbacks now have equivalents
8631 in the X509_STORE_CTX structure.
8633 X509_STORE_CTX also has a 'flags' field which can be used
8634 to customise the verify behaviour.
8637 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8638 excludes S/MIME capabilities.
8641 *) When a certificate request is read in keep a copy of the
8642 original encoding of the signed data and use it when outputing
8643 again. Signatures then use the original encoding rather than
8644 a decoded, encoded version which may cause problems if the
8645 request is improperly encoded.
8648 *) For consistency with other BIO_puts implementations, call
8649 buffer_write(b, ...) directly in buffer_puts instead of calling
8652 In BIO_puts, increment b->num_write as in BIO_write.
8653 [Peter.Sylvester@EdelWeb.fr]
8655 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8656 BN_zero, we may not return a BIGNUM with an array consisting of
8660 *) Avoid calling abort() from within the library when problems are
8661 detected, except if preprocessor symbols have been defined
8662 (such as REF_CHECK, BN_DEBUG etc.).
8665 *) New openssl application 'rsautl'. This utility can be
8666 used for low level RSA operations. DER public key
8667 BIO/fp routines also added.
8670 *) New Configure entry and patches for compiling on QNX 4.
8671 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8673 *) A demo state-machine implementation was sponsored by
8674 Nuron (http://www.nuron.com/) and is now available in
8675 demos/state_machine.
8678 *) New options added to the 'dgst' utility for signature
8679 generation and verification.
8682 *) Unrecognized PKCS#7 content types are now handled via a
8683 catch all ASN1_TYPE structure. This allows unsupported
8684 types to be stored as a "blob" and an application can
8685 encode and decode it manually.
8688 *) Fix various signed/unsigned issues to make a_strex.c
8690 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8692 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8693 length if passed a buffer. ASN1_INTEGER_to_BN failed
8694 if passed a NULL BN and its argument was negative.
8695 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8697 *) Modification to PKCS#7 encoding routines to output definite
8698 length encoding. Since currently the whole structures are in
8699 memory there's not real point in using indefinite length
8700 constructed encoding. However if OpenSSL is compiled with
8701 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8704 *) Added BIO_vprintf() and BIO_vsnprintf().
8707 *) Added more prefixes to parse for in the the strings written
8708 through a logging bio, to cover all the levels that are available
8709 through syslog. The prefixes are now:
8711 PANIC, EMERG, EMR => LOG_EMERG
8712 ALERT, ALR => LOG_ALERT
8713 CRIT, CRI => LOG_CRIT
8714 ERROR, ERR => LOG_ERR
8715 WARNING, WARN, WAR => LOG_WARNING
8716 NOTICE, NOTE, NOT => LOG_NOTICE
8717 INFO, INF => LOG_INFO
8718 DEBUG, DBG => LOG_DEBUG
8720 and as before, if none of those prefixes are present at the
8721 beginning of the string, LOG_ERR is chosen.
8723 On Win32, the LOG_* levels are mapped according to this:
8725 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8726 LOG_WARNING => EVENTLOG_WARNING_TYPE
8727 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8731 *) Made it possible to reconfigure with just the configuration
8732 argument "reconf" or "reconfigure". The command line arguments
8733 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8734 and are retrieved from there when reconfiguring.
8738 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8740 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8743 *) The obj_dat.pl script was messing up the sorting of object
8744 names. The reason was that it compared the quoted version
8745 of strings as a result "OCSP" > "OCSP Signing" because
8746 " > SPACE. Changed script to store unquoted versions of
8747 names and add quotes on output. It was also omitting some
8748 names from the lookup table if they were given a default
8749 value (that is if SN is missing it is given the same
8750 value as LN and vice versa), these are now added on the
8751 grounds that if an object has a name we should be able to
8752 look it up. Finally added warning output when duplicate
8753 short or long names are found.
8756 *) Changes needed for Tandem NSK.
8757 [Scott Uroff <scott@xypro.com>]
8759 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8760 RSA_padding_check_SSLv23(), special padding was never detected
8761 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8762 version rollback attacks was not effective.
8764 In s23_clnt.c, don't use special rollback-attack detection padding
8765 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8766 client; similarly, in s23_srvr.c, don't do the rollback check if
8767 SSL 2.0 is the only protocol enabled in the server.
8770 *) Make it possible to get hexdumps of unprintable data with 'openssl
8771 asn1parse'. By implication, the functions ASN1_parse_dump() and
8772 BIO_dump_indent() are added.
8775 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8776 these print out strings and name structures based on various
8777 flags including RFC2253 support and proper handling of
8778 multibyte characters. Added options to the 'x509' utility
8779 to allow the various flags to be set.
8782 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8783 Also change the functions X509_cmp_current_time() and
8784 X509_gmtime_adj() work with an ASN1_TIME structure,
8785 this will enable certificates using GeneralizedTime in validity
8786 dates to be checked.
8789 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8790 negative public key encodings) on by default,
8791 NO_NEG_PUBKEY_BUG can be set to disable it.
8794 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8795 content octets. An i2c_ASN1_OBJECT is unnecessary because
8796 the encoding can be trivially obtained from the structure.
8799 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8800 not read locks (CRYPTO_r_[un]lock).
8803 *) A first attempt at creating official support for shared
8804 libraries through configuration. I've kept it so the
8805 default is static libraries only, and the OpenSSL programs
8806 are always statically linked for now, but there are
8807 preparations for dynamic linking in place.
8808 This has been tested on Linux and Tru64.
8811 *) Randomness polling function for Win9x, as described in:
8812 Peter Gutmann, Software Generation of Practically Strong
8816 *) Fix so PRNG is seeded in req if using an already existing
8820 *) New options to smime application. -inform and -outform
8821 allow alternative formats for the S/MIME message including
8822 PEM and DER. The -content option allows the content to be
8823 specified separately. This should allow things like Netscape
8824 form signing output easier to verify.
8827 *) Fix the ASN1 encoding of tags using the 'long form'.
8830 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8831 STRING types. These convert content octets to and from the
8832 underlying type. The actual tag and length octets are
8833 already assumed to have been read in and checked. These
8834 are needed because all other string types have virtually
8835 identical handling apart from the tag. By having versions
8836 of the ASN1 functions that just operate on content octets
8837 IMPLICIT tagging can be handled properly. It also allows
8838 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8839 and ASN1_INTEGER are identical apart from the tag.
8842 *) Change the handling of OID objects as follows:
8844 - New object identifiers are inserted in objects.txt, following
8845 the syntax given in objects.README.
8846 - objects.pl is used to process obj_mac.num and create a new
8848 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8851 This is currently kind of a hack, and the perl code in objects.pl
8852 isn't very elegant, but it works as I intended. The simplest way
8853 to check that it worked correctly is to look in obj_dat.h and
8854 check the array nid_objs and make sure the objects haven't moved
8855 around (this is important!). Additions are OK, as well as
8856 consistent name changes.
8859 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8862 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8863 The given file adds to whatever has already been seeded into the
8864 random pool through the RANDFILE configuration file option or
8865 environment variable, or the default random state file.
8868 *) mkstack.pl now sorts each macro group into lexical order.
8869 Previously the output order depended on the order the files
8870 appeared in the directory, resulting in needless rewriting
8874 *) Patches to make OpenSSL compile under Win32 again. Mostly
8875 work arounds for the VC++ problem that it treats func() as
8876 func(void). Also stripped out the parts of mkdef.pl that
8877 added extra typesafe functions: these no longer exist.
8880 *) Reorganisation of the stack code. The macros are now all
8881 collected in safestack.h . Each macro is defined in terms of
8882 a "stack macro" of the form SKM_<name>(type, a, b). The
8883 DEBUG_SAFESTACK is now handled in terms of function casts,
8884 this has the advantage of retaining type safety without the
8885 use of additional functions. If DEBUG_SAFESTACK is not defined
8886 then the non typesafe macros are used instead. Also modified the
8887 mkstack.pl script to handle the new form. Needs testing to see
8888 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8889 the default if no major problems. Similar behaviour for ASN1_SET_OF
8890 and PKCS12_STACK_OF.
8893 *) When some versions of IIS use the 'NET' form of private key the
8894 key derivation algorithm is different. Normally MD5(password) is
8895 used as a 128 bit RC4 key. In the modified case
8896 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8897 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8898 as the old Netscape_RSA functions except they have an additional
8899 'sgckey' parameter which uses the modified algorithm. Also added
8900 an -sgckey command line option to the rsa utility. Thanks to
8901 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8902 algorithm to openssl-dev.
8905 *) The evp_local.h macros were using 'c.##kname' which resulted in
8906 invalid expansion on some systems (SCO 5.0.5 for example).
8907 Corrected to 'c.kname'.
8908 [Phillip Porch <root@theporch.com>]
8910 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8911 a STACK of email addresses from a certificate or request, these look
8912 in the subject name and the subject alternative name extensions and
8913 omit any duplicate addresses.
8916 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8917 This makes DSA verification about 2 % faster.
8920 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8921 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8922 plus overhead for 1024 bit moduli).
8923 This makes exponentiations about 0.5 % faster for 1024 bit
8924 exponents (as measured by "openssl speed rsa2048").
8927 *) Rename memory handling macros to avoid conflicts with other
8929 Malloc => OPENSSL_malloc
8930 Malloc_locked => OPENSSL_malloc_locked
8931 Realloc => OPENSSL_realloc
8932 Free => OPENSSL_free
8935 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8936 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8939 *) CygWin32 support.
8940 [John Jarvie <jjarvie@newsguy.com>]
8942 *) The type-safe stack code has been rejigged. It is now only compiled
8943 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8944 by default all type-specific stack functions are "#define"d back to
8945 standard stack functions. This results in more streamlined output
8946 but retains the type-safety checking possibilities of the original
8950 *) The STACK code has been cleaned up, and certain type declarations
8951 that didn't make a lot of sense have been brought in line. This has
8952 also involved a cleanup of sorts in safestack.h to more correctly
8953 map type-safe stack functions onto their plain stack counterparts.
8954 This work has also resulted in a variety of "const"ifications of
8955 lots of the code, especially "_cmp" operations which should normally
8956 be prototyped with "const" parameters anyway.
8959 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8960 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8961 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8962 where all of 'md' is used each time the PRNG is used, but 'state'
8963 is used only indexed by a cyclic counter. As entropy may not be
8964 well distributed from the beginning, 'md' is important as a
8965 chaining variable. However, the output function chains only half
8966 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8967 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8968 in all of 'state' being rewritten, with the new values depending
8969 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8972 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8973 the handshake is continued after ssl_verify_cert_chain();
8974 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8975 can lead to 'unexplainable' connection aborts later.
8976 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8978 *) Major EVP API cipher revision.
8979 Add hooks for extra EVP features. This allows various cipher
8980 parameters to be set in the EVP interface. Support added for variable
8981 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8982 setting of RC2 and RC5 parameters.
8984 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8987 Remove lots of duplicated code from the EVP library. For example *every*
8988 cipher init() function handles the 'iv' in the same way according to the
8989 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8990 for CFB and OFB modes they zero ctx->num.
8992 New functionality allows removal of S/MIME code RC2 hack.
8994 Most of the routines have the same form and so can be declared in terms
8997 By shifting this to the top level EVP_CipherInit() it can be removed from
8998 all individual ciphers. If the cipher wants to handle IVs or keys
8999 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9002 Change lots of functions like EVP_EncryptUpdate() to now return a
9003 value: although software versions of the algorithms cannot fail
9004 any installed hardware versions can.
9007 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9008 this option is set, tolerate broken clients that send the negotiated
9009 protocol version number instead of the requested protocol version
9013 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9014 i.e. non-zero for export ciphersuites, zero otherwise.
9015 Previous versions had this flag inverted, inconsistent with
9016 rsa_tmp_cb (..._TMP_RSA_CB).
9017 [Bodo Moeller; problem reported by Amit Chopra]
9019 *) Add missing DSA library text string. Work around for some IIS
9020 key files with invalid SEQUENCE encoding.
9023 *) Add a document (doc/standards.txt) that list all kinds of standards
9024 and so on that are implemented in OpenSSL.
9027 *) Enhance c_rehash script. Old version would mishandle certificates
9028 with the same subject name hash and wouldn't handle CRLs at all.
9029 Added -fingerprint option to crl utility, to support new c_rehash
9033 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9036 *) Fix for SSL server purpose checking. Server checking was
9037 rejecting certificates which had extended key usage present
9038 but no ssl client purpose.
9039 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9041 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9042 is a little unclear about how a blank password is handled.
9043 Since the password in encoded as a BMPString with terminating
9044 double NULL a zero length password would end up as just the
9045 double NULL. However no password at all is different and is
9046 handled differently in the PKCS#12 key generation code. NS
9047 treats a blank password as zero length. MSIE treats it as no
9048 password on export: but it will try both on import. We now do
9049 the same: PKCS12_parse() tries zero length and no password if
9050 the password is set to "" or NULL (NULL is now a valid password:
9051 it wasn't before) as does the pkcs12 application.
9054 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9055 perror when PEM_read_bio_X509_REQ fails, the error message must
9056 be obtained from the error queue.
9059 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9060 it in ERR_remove_state if appropriate, and change ERR_get_state
9061 accordingly to avoid race conditions (this is necessary because
9062 thread_hash is no longer constant once set).
9065 *) Bugfix for linux-elf makefile.one.
9068 *) RSA_get_default_method() will now cause a default
9069 RSA_METHOD to be chosen if one doesn't exist already.
9070 Previously this was only set during a call to RSA_new()
9071 or RSA_new_method(NULL) meaning it was possible for
9072 RSA_get_default_method() to return NULL.
9075 *) Added native name translation to the existing DSO code
9076 that will convert (if the flag to do so is set) filenames
9077 that are sufficiently small and have no path information
9078 into a canonical native form. Eg. "blah" converted to
9079 "libblah.so" or "blah.dll" etc.
9082 *) New function ERR_error_string_n(e, buf, len) which is like
9083 ERR_error_string(e, buf), but writes at most 'len' bytes
9084 including the 0 terminator. For ERR_error_string_n, 'buf'
9086 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9088 *) CONF library reworked to become more general. A new CONF
9089 configuration file reader "class" is implemented as well as a
9090 new functions (NCONF_*, for "New CONF") to handle it. The now
9091 old CONF_* functions are still there, but are reimplemented to
9092 work in terms of the new functions. Also, a set of functions
9093 to handle the internal storage of the configuration data is
9094 provided to make it easier to write new configuration file
9095 reader "classes" (I can definitely see something reading a
9096 configuration file in XML format, for example), called _CONF_*,
9097 or "the configuration storage API"...
9099 The new configuration file reading functions are:
9101 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9102 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9104 NCONF_default, NCONF_WIN32
9106 NCONF_dump_fp, NCONF_dump_bio
9108 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9109 NCONF_new creates a new CONF object. This works in the same way
9110 as other interfaces in OpenSSL, like the BIO interface.
9111 NCONF_dump_* dump the internal storage of the configuration file,
9112 which is useful for debugging. All other functions take the same
9113 arguments as the old CONF_* functions wth the exception of the
9114 first that must be a `CONF *' instead of a `LHASH *'.
9116 To make it easer to use the new classes with the old CONF_* functions,
9117 the function CONF_set_default_method is provided.
9120 *) Add '-tls1' option to 'openssl ciphers', which was already
9121 mentioned in the documentation but had not been implemented.
9122 (This option is not yet really useful because even the additional
9123 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9126 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9127 OpenSSL-based applications) load shared libraries and bind to
9128 them in a portable way.
9129 [Geoff Thorpe, with contributions from Richard Levitte]
9131 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9133 *) Make sure _lrotl and _lrotr are only used with MSVC.
9135 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9136 (the default implementation of RAND_status).
9138 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9139 to '-clrext' (= clear extensions), as intended and documented.
9140 [Bodo Moeller; inconsistency pointed out by Michael Attili
9143 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9144 was larger than the MD block size.
9145 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9147 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9148 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9149 using the passed key: if the passed key was a private key the result
9150 of X509_print(), for example, would be to print out all the private key
9154 *) des_quad_cksum() byte order bug fix.
9155 [Ulf Möller, using the problem description in krb4-0.9.7, where
9156 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9158 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9160 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9162 *) For easily testing in shell scripts whether some command
9163 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9164 returns with exit code 0 iff no command of the given name is available.
9165 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9166 the output goes to stdout and nothing is printed to stderr.
9167 Additional arguments are always ignored.
9169 Since for each cipher there is a command of the same name,
9170 the 'no-cipher' compilation switches can be tested this way.
9172 ('openssl no-XXX' is not able to detect pseudo-commands such
9173 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9176 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9179 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9180 is set; it will be thrown away anyway because each handshake creates
9182 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9183 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9184 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9185 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9188 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9189 'Q' and 'R' lose their special meanings (quit/renegotiate).
9190 This is part of what -quiet does; unlike -quiet, -ign_eof
9191 does not suppress any output.
9194 *) Add compatibility options to the purpose and trust code. The
9195 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9196 accepts a certificate or CA, this was the previous behaviour,
9197 with all the associated security issues.
9199 X509_TRUST_COMPAT is the old trust behaviour: only and
9200 automatically trust self signed roots in certificate store. A
9201 new trust setting X509_TRUST_DEFAULT is used to specify that
9202 a purpose has no associated trust setting and it should instead
9203 use the value in the default purpose.
9206 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9207 and fix a memory leak.
9210 *) In util/mkerr.pl (which implements 'make errors'), preserve
9211 reason strings from the previous version of the .c file, as
9212 the default to have only downcase letters (and digits) in
9213 automatically generated reasons codes is not always appropriate.
9216 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9217 using strerror. Previously, ERR_reason_error_string() returned
9218 library names as reason strings for SYSerr; but SYSerr is a special
9219 case where small numbers are errno values, not library numbers.
9222 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9223 converts DSA parameters into DH parameters. (When creating parameters,
9224 DSA_generate_parameters is used.)
9227 *) Include 'length' (recommended exponent length) in C code generated
9228 by 'openssl dhparam -C'.
9231 *) The second argument to set_label in perlasm was already being used
9232 so couldn't be used as a "file scope" flag. Moved to third argument
9236 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9237 instead of RAND_bytes for encryption IVs and salts.
9240 *) Include RAND_status() into RAND_METHOD instead of implementing
9241 it only for md_rand.c Otherwise replacing the PRNG by calling
9242 RAND_set_rand_method would be impossible.
9245 *) Don't let DSA_generate_key() enter an infinite loop if the random
9246 number generation fails.
9249 *) New 'rand' application for creating pseudo-random output.
9252 *) Added configuration support for Linux/IA64
9253 [Rolf Haberrecker <rolf@suse.de>]
9255 *) Assembler module support for Mingw32.
9258 *) Shared library support for HPUX (in shlib/).
9259 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9261 *) Shared library support for Solaris gcc.
9262 [Lutz Behnke <behnke@trustcenter.de>]
9264 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9266 *) PKCS7_encrypt() was adding text MIME headers twice because they
9267 were added manually and by SMIME_crlf_copy().
9270 *) In bntest.c don't call BN_rand with zero bits argument.
9271 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9273 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9274 case was implemented. This caused BN_div_recp() to fail occasionally.
9277 *) Add an optional second argument to the set_label() in the perl
9278 assembly language builder. If this argument exists and is set
9279 to 1 it signals that the assembler should use a symbol whose
9280 scope is the entire file, not just the current function. This
9281 is needed with MASM which uses the format label:: for this scope.
9282 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9284 *) Change the ASN1 types so they are typedefs by default. Before
9285 almost all types were #define'd to ASN1_STRING which was causing
9286 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9290 *) Change names of new functions to the new get1/get0 naming
9291 convention: After 'get1', the caller owns a reference count
9292 and has to call ..._free; 'get0' returns a pointer to some
9293 data structure without incrementing reference counters.
9294 (Some of the existing 'get' functions increment a reference
9295 counter, some don't.)
9296 Similarly, 'set1' and 'add1' functions increase reference
9297 counters or duplicate objects.
9300 *) Allow for the possibility of temp RSA key generation failure:
9301 the code used to assume it always worked and crashed on failure.
9304 *) Fix potential buffer overrun problem in BIO_printf().
9305 [Ulf Möller, using public domain code by Patrick Powell; problem
9306 pointed out by David Sacerdote <das33@cornell.edu>]
9308 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9309 RAND_egd() and RAND_status(). In the command line application,
9310 the EGD socket can be specified like a seed file using RANDFILE
9314 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9315 Some CAs (e.g. Verisign) distribute certificates in this form.
9318 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9319 list to exclude them. This means that no special compilation option
9320 is needed to use anonymous DH: it just needs to be included in the
9324 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9325 EVP_MD_type. The old functionality is available in a new macro called
9326 EVP_MD_md(). Change code that uses it and update docs.
9329 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9330 where the 'void *' argument is replaced by a function pointer argument.
9331 Previously 'void *' was abused to point to functions, which works on
9332 many platforms, but is not correct. As these functions are usually
9333 called by macros defined in OpenSSL header files, most source code
9334 should work without changes.
9337 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9338 sections with information on -D... compiler switches used for
9339 compiling the library so that applications can see them. To enable
9340 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9341 must be defined. E.g.,
9342 #define OPENSSL_ALGORITHM_DEFINES
9343 #include <openssl/opensslconf.h>
9344 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9345 [Richard Levitte, Ulf and Bodo Möller]
9347 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9351 *) Change the 'other' type in certificate aux info to a STACK_OF
9352 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9353 the required ASN1 format: arbitrary types determined by an OID.
9356 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9357 argument to 'req'. This is not because the function is newer or
9358 better than others it just uses the work 'NEW' in the certificate
9359 request header lines. Some software needs this.
9362 *) Reorganise password command line arguments: now passwords can be
9363 obtained from various sources. Delete the PEM_cb function and make
9364 it the default behaviour: i.e. if the callback is NULL and the
9365 usrdata argument is not NULL interpret it as a null terminated pass
9366 phrase. If usrdata and the callback are NULL then the pass phrase
9367 is prompted for as usual.
9370 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9371 the support is automatically enabled. The resulting binaries will
9372 autodetect the card and use it if present.
9373 [Ben Laurie and Compaq Inc.]
9375 *) Work around for Netscape hang bug. This sends certificate request
9376 and server done in one record. Since this is perfectly legal in the
9377 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9378 the bugs/SSLv3 entry for more info.
9381 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9384 *) Add -rand argument to smime and pkcs12 applications and read/write
9388 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9391 *) Add command line password options to the remaining applications.
9394 *) Bug fix for BN_div_recp() for numerators with an even number of
9398 *) More tests in bntest.c, and changed test_bn output.
9401 *) ./config recognizes MacOS X now.
9404 *) Bug fix for BN_div() when the first words of num and divsor are
9405 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9408 *) Add support for various broken PKCS#8 formats, and command line
9409 options to produce them.
9412 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9413 get temporary BIGNUMs from a BN_CTX.
9416 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9420 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9421 include a #define from the old name to the new. The original intent
9422 was that statically linked binaries could for example just call
9423 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9424 link with digests. This never worked because SSLeay_add_all_digests()
9425 and SSLeay_add_all_ciphers() were in the same source file so calling
9426 one would link with the other. They are now in separate source files.
9429 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9432 *) Use a less unusual form of the Miller-Rabin primality test (it used
9433 a binary algorithm for exponentiation integrated into the Miller-Rabin
9434 loop, our standard modexp algorithms are faster).
9437 *) Support for the EBCDIC character set completed.
9438 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9440 *) Source code cleanups: use const where appropriate, eliminate casts,
9441 use void * instead of char * in lhash.
9444 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9445 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9446 this the server could overwrite ephemeral keys that the client
9450 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9451 using 50 iterations of the Rabin-Miller test.
9453 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9454 iterations of the Rabin-Miller test as required by the appendix
9455 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9456 As BN_is_prime_fasttest includes trial division, DSA parameter
9457 generation becomes much faster.
9459 This implies a change for the callback functions in DSA_is_prime
9460 and DSA_generate_parameters: The callback function is called once
9461 for each positive witness in the Rabin-Miller test, not just
9462 occasionally in the inner loop; and the parameters to the
9463 callback function now provide an iteration count for the outer
9464 loop rather than for the current invocation of the inner loop.
9465 DSA_generate_parameters additionally can call the callback
9466 function with an 'iteration count' of -1, meaning that a
9467 candidate has passed the trial division test (when q is generated
9468 from an application-provided seed, trial division is skipped).
9471 *) New function BN_is_prime_fasttest that optionally does trial
9472 division before starting the Rabin-Miller test and has
9473 an additional BN_CTX * argument (whereas BN_is_prime always
9474 has to allocate at least one BN_CTX).
9475 'callback(1, -1, cb_arg)' is called when a number has passed the
9476 trial division stage.
9479 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9483 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9486 *) New function BN_pseudo_rand().
9489 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9490 bignum version of BN_from_montgomery() with the working code from
9491 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9495 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9496 made it impossible to use the same SSL_SESSION data structure in
9497 SSL2 clients in multiple threads.
9500 *) The return value of RAND_load_file() no longer counts bytes obtained
9501 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9502 to seed the PRNG (previously an explicit byte count was required).
9503 [Ulf Möller, Bodo Möller]
9505 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9506 used (char *) instead of (void *) and had casts all over the place.
9509 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9512 *) Retain source code compatibility for BN_prime_checks macro:
9513 BN_is_prime(..., BN_prime_checks, ...) now uses
9514 BN_prime_checks_for_size to determine the appropriate number of
9515 Rabin-Miller iterations.
9518 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9519 DH_CHECK_P_NOT_SAFE_PRIME.
9520 (Check if this is true? OpenPGP calls them "strong".)
9523 *) Merge the functionality of "dh" and "gendh" programs into a new program
9524 "dhparam". The old programs are retained for now but will handle DH keys
9525 (instead of parameters) in future.
9528 *) Make the ciphers, s_server and s_client programs check the return values
9529 when a new cipher list is set.
9532 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9533 ciphers. Before when the 56bit ciphers were enabled the sorting was
9536 The syntax for the cipher sorting has been extended to support sorting by
9537 cipher-strength (using the strength_bits hard coded in the tables).
9538 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9540 Fix a bug in the cipher-command parser: when supplying a cipher command
9541 string with an "undefined" symbol (neither command nor alphanumeric
9542 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9543 an error is flagged.
9545 Due to the strength-sorting extension, the code of the
9546 ssl_create_cipher_list() function was completely rearranged. I hope that
9547 the readability was also increased :-)
9548 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9550 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9551 for the first serial number and places 2 in the serial number file. This
9552 avoids problems when the root CA is created with serial number zero and
9553 the first user certificate has the same issuer name and serial number
9557 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9558 the new code. Add documentation for this stuff.
9561 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9562 X509_*() to X509at_*() on the grounds that they don't handle X509
9563 structures and behave in an analogous way to the X509v3 functions:
9564 they shouldn't be called directly but wrapper functions should be used
9567 So we also now have some wrapper functions that call the X509at functions
9568 when passed certificate requests. (TO DO: similar things can be done with
9569 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9570 things. Some of these need some d2i or i2d and print functionality
9571 because they handle more complex structures.)
9574 *) Add missing #ifndefs that caused missing symbols when building libssl
9575 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9576 NO_RSA in ssl/s2*.c.
9577 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9579 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9580 has a return value which indicates the quality of the random data
9581 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9582 error queue. New function RAND_pseudo_bytes() generates output that is
9583 guaranteed to be unique but not unpredictable. RAND_add is like
9584 RAND_seed, but takes an extra argument for an entropy estimate
9585 (RAND_seed always assumes full entropy).
9588 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9589 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9590 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9591 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9592 false-positive rate of at most 2^-80 for random input.
9595 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9598 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9599 in the 0.9.5 release), this returns the chain
9600 from an X509_CTX structure with a dup of the stack and all
9601 the X509 reference counts upped: so the stack will exist
9602 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9605 Also make SSL_SESSION_print() print out the verify return
9609 *) Add manpage for the pkcs12 command. Also change the default
9610 behaviour so MAC iteration counts are used unless the new
9611 -nomaciter option is used. This improves file security and
9612 only older versions of MSIE (4.0 for example) need it.
9615 *) Honor the no-xxx Configure options when creating .DEF files.
9618 *) Add PKCS#10 attributes to field table: challengePassword,
9619 unstructuredName and unstructuredAddress. These are taken from
9620 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9621 international characters are used.
9623 More changes to X509_ATTRIBUTE code: allow the setting of types
9624 based on strings. Remove the 'loc' parameter when adding
9625 attributes because these will be a SET OF encoding which is sorted
9629 *) Initial changes to the 'req' utility to allow request generation
9630 automation. This will allow an application to just generate a template
9631 file containing all the field values and have req construct the
9634 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9635 used all over the place including certificate requests and PKCS#7
9636 structures. They are currently handled manually where necessary with
9637 some primitive wrappers for PKCS#7. The new functions behave in a
9638 manner analogous to the X509 extension functions: they allow
9639 attributes to be looked up by NID and added.
9641 Later something similar to the X509V3 code would be desirable to
9642 automatically handle the encoding, decoding and printing of the
9643 more complex types. The string types like challengePassword can
9644 be handled by the string table functions.
9646 Also modified the multi byte string table handling. Now there is
9647 a 'global mask' which masks out certain types. The table itself
9648 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9649 is useful when for example there is only one permissible type
9650 (as in countryName) and using the mask might result in no valid
9654 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9655 SSL_get_peer_finished to allow applications to obtain the latest
9656 Finished messages sent to the peer or expected from the peer,
9657 respectively. (SSL_get_peer_finished is usually the Finished message
9658 actually received from the peer, otherwise the protocol will be aborted.)
9660 As the Finished message are message digests of the complete handshake
9661 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9662 be used for external authentication procedures when the authentication
9663 provided by SSL/TLS is not desired or is not enough.
9666 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9667 the host supports BWX extension and if Compaq C is present on the
9668 $PATH. Just exploiting of the BWX extension results in 20-30%
9669 performance kick for some algorithms, e.g. DES and RC4 to mention
9670 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9674 *) Add support for MS "fast SGC". This is arguably a violation of the
9675 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9676 weak crypto and after checking the certificate is SGC a second one
9677 with strong crypto. MS SGC stops the first handshake after receiving
9678 the server certificate message and sends a second client hello. Since
9679 a server will typically do all the time consuming operations before
9680 expecting any further messages from the client (server key exchange
9681 is the most expensive) there is little difference between the two.
9683 To get OpenSSL to support MS SGC we have to permit a second client
9684 hello message after we have sent server done. In addition we have to
9685 reset the MAC if we do get this second client hello.
9688 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9689 if a DER encoded private key is RSA or DSA traditional format. Changed
9690 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9691 format DER encoded private key. Newer code should use PKCS#8 format which
9692 has the key type encoded in the ASN1 structure. Added DER private key
9693 support to pkcs8 application.
9696 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9697 ciphersuites has been selected (as required by the SSL 3/TLS 1
9698 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9699 is set, we interpret this as a request to violate the specification
9700 (the worst that can happen is a handshake failure, and 'correct'
9701 behaviour would result in a handshake failure anyway).
9704 *) In SSL_CTX_add_session, take into account that there might be multiple
9705 SSL_SESSION structures with the same session ID (e.g. when two threads
9706 concurrently obtain them from an external cache).
9707 The internal cache can handle only one SSL_SESSION with a given ID,
9708 so if there's a conflict, we now throw out the old one to achieve
9712 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9713 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9714 some routines that use cipher OIDs: some ciphers do not have OIDs
9715 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9719 *) Simplify the trust setting structure and code. Now we just have
9720 two sequences of OIDs for trusted and rejected settings. These will
9721 typically have values the same as the extended key usage extension
9722 and any application specific purposes.
9724 The trust checking code now has a default behaviour: it will just
9725 check for an object with the same NID as the passed id. Functions can
9726 be provided to override either the default behaviour or the behaviour
9727 for a given id. SSL client, server and email already have functions
9728 in place for compatibility: they check the NID and also return "trusted"
9729 if the certificate is self signed.
9732 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9733 traditional format into an EVP_PKEY structure.
9736 *) Add a password callback function PEM_cb() which either prompts for
9737 a password if usr_data is NULL or otherwise assumes it is a null
9738 terminated password. Allow passwords to be passed on command line
9739 environment or config files in a few more utilities.
9742 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9743 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9744 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9745 Update documentation.
9748 *) Support for ASN1 "NULL" type. This could be handled before by using
9749 ASN1_TYPE but there wasn't any function that would try to read a NULL
9750 and produce an error if it couldn't. For compatibility we also have
9751 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9752 don't allocate anything because they don't need to.
9755 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9757 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9759 *) Rebuild of the memory allocation routines used by OpenSSL code and
9760 possibly others as well. The purpose is to make an interface that
9761 provide hooks so anyone can build a separate set of allocation and
9762 deallocation routines to be used by OpenSSL, for example memory
9763 pool implementations, or something else, which was previously hard
9764 since Malloc(), Realloc() and Free() were defined as macros having
9765 the values malloc, realloc and free, respectively (except for Win32
9766 compilations). The same is provided for memory debugging code.
9767 OpenSSL already comes with functionality to find memory leaks, but
9768 this gives people a chance to debug other memory problems.
9770 With these changes, a new set of functions and macros have appeared:
9772 CRYPTO_set_mem_debug_functions() [F]
9773 CRYPTO_get_mem_debug_functions() [F]
9774 CRYPTO_dbg_set_options() [F]
9775 CRYPTO_dbg_get_options() [F]
9776 CRYPTO_malloc_debug_init() [M]
9778 The memory debug functions are NULL by default, unless the library
9779 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9780 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9781 gives the standard debugging functions that come with OpenSSL) or
9782 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9783 provided by the library user) must be used. When the standard
9784 debugging functions are used, CRYPTO_dbg_set_options can be used to
9785 request additional information:
9786 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9787 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9789 Also, things like CRYPTO_set_mem_functions will always give the
9790 expected result (the new set of functions is used for allocation
9791 and deallocation) at all times, regardless of platform and compiler
9794 To finish it up, some functions that were never use in any other
9795 way than through macros have a new API and new semantic:
9798 CRYPTO_dbg_realloc()
9801 All macros of value have retained their old syntax.
9802 [Richard Levitte and Bodo Moeller]
9804 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9805 ordering of SMIMECapabilities wasn't in "strength order" and there
9806 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9810 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9811 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9812 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9814 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9815 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9816 functionality to handle multipart/signed properly) and a utility
9817 called 'smime' to call all this stuff. This is based on code I
9818 originally wrote for Celo who have kindly allowed it to be
9819 included in OpenSSL.
9822 *) Add variants des_set_key_checked and des_set_key_unchecked of
9823 des_set_key (aka des_key_sched). Global variable des_check_key
9824 decides which of these is called by des_set_key; this way
9825 des_check_key behaves as it always did, but applications and
9826 the library itself, which was buggy for des_check_key == 1,
9827 have a cleaner way to pick the version they need.
9830 *) New function PKCS12_newpass() which changes the password of a
9834 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9835 dynamic mix. In both cases the ids can be used as an index into the
9836 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9837 functions so they accept a list of the field values and the
9838 application doesn't need to directly manipulate the X509_TRUST
9842 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9846 *) Modify the way the V3 extension code looks up extensions. This now
9847 works in a similar way to the object code: we have some "standard"
9848 extensions in a static table which is searched with OBJ_bsearch()
9849 and the application can add dynamic ones if needed. The file
9850 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9851 updated whenever a new extension is added to the core code and kept
9852 in ext_nid order. There is a simple program 'tabtest.c' which checks
9853 this. New extensions are not added too often so this file can readily
9854 be maintained manually.
9856 There are two big advantages in doing things this way. The extensions
9857 can be looked up immediately and no longer need to be "added" using
9858 X509V3_add_standard_extensions(): this function now does nothing.
9859 [Side note: I get *lots* of email saying the extension code doesn't
9860 work because people forget to call this function]
9861 Also no dynamic allocation is done unless new extensions are added:
9862 so if we don't add custom extensions there is no need to call
9863 X509V3_EXT_cleanup().
9866 *) Modify enc utility's salting as follows: make salting the default. Add a
9867 magic header, so unsalted files fail gracefully instead of just decrypting
9868 to garbage. This is because not salting is a big security hole, so people
9869 should be discouraged from doing it.
9872 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9873 digest to be passed on the command line but it only used this
9874 parameter when signing a certificate. Modified so all relevant
9875 operations are affected by the digest parameter including the
9876 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9877 DSA key was used because it didn't fix the digest.
9880 *) Initial certificate chain verify code. Currently tests the untrusted
9881 certificates for consistency with the verify purpose (which is set
9882 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9884 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9885 this is because it will reject chains with invalid extensions whereas
9886 every previous version of OpenSSL and SSLeay made no checks at all.
9888 Trust code: checks the root CA for the relevant trust settings. Trust
9889 settings have an initial value consistent with the verify purpose: e.g.
9890 if the verify purpose is for SSL client use it expects the CA to be
9891 trusted for SSL client use. However the default value can be changed to
9892 permit custom trust settings: one example of this would be to only trust
9893 certificates from a specific "secure" set of CAs.
9895 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9896 which should be used for version portability: especially since the
9897 verify structure is likely to change more often now.
9899 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9900 to set them. If not set then assume SSL clients will verify SSL servers
9903 Two new options to the verify program: -untrusted allows a set of
9904 untrusted certificates to be passed in and -purpose which sets the
9905 intended purpose of the certificate. If a purpose is set then the
9906 new chain verify code is used to check extension consistency.
9909 *) Support for the authority information access extension.
9912 *) Modify RSA and DSA PEM read routines to transparently handle
9913 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9914 public keys in a format compatible with certificate
9915 SubjectPublicKeyInfo structures. Unfortunately there were already
9916 functions called *_PublicKey_* which used various odd formats so
9917 these are retained for compatibility: however the DSA variants were
9918 never in a public release so they have been deleted. Changed dsa/rsa
9919 utilities to handle the new format: note no releases ever handled public
9920 keys so we should be OK.
9922 The primary motivation for this change is to avoid the same fiasco
9923 that dogs private keys: there are several incompatible private key
9924 formats some of which are standard and some OpenSSL specific and
9925 require various evil hacks to allow partial transparent handling and
9926 even then it doesn't work with DER formats. Given the option anything
9927 other than PKCS#8 should be dumped: but the other formats have to
9928 stay in the name of compatibility.
9930 With public keys and the benefit of hindsight one standard format
9931 is used which works with EVP_PKEY, RSA or DSA structures: though
9932 it clearly returns an error if you try to read the wrong kind of key.
9934 Added a -pubkey option to the 'x509' utility to output the public key.
9935 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9936 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9937 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9938 that do the same as the EVP_PKEY_assign_*() except they up the
9939 reference count of the added key (they don't "swallow" the
9943 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9944 CRLs would fail if the file contained no certificates or no CRLs:
9945 added a new function to read in both types and return the number
9946 read: this means that if none are read it will be an error. The
9947 DER versions of the certificate and CRL reader would always fail
9948 because it isn't possible to mix certificates and CRLs in DER format
9949 without choking one or the other routine. Changed this to just read
9950 a certificate: this is the best we can do. Also modified the code
9951 in apps/verify.c to take notice of return codes: it was previously
9952 attempting to read in certificates from NULL pointers and ignoring
9953 any errors: this is one reason why the cert and CRL reader seemed
9954 to work. It doesn't check return codes from the default certificate
9955 routines: these may well fail if the certificates aren't installed.
9958 *) Code to support otherName option in GeneralName.
9961 *) First update to verify code. Change the verify utility
9962 so it warns if it is passed a self signed certificate:
9963 for consistency with the normal behaviour. X509_verify
9964 has been modified to it will now verify a self signed
9965 certificate if *exactly* the same certificate appears
9966 in the store: it was previously impossible to trust a
9967 single self signed certificate. This means that:
9968 openssl verify ss.pem
9969 now gives a warning about a self signed certificate but
9970 openssl verify -CAfile ss.pem ss.pem
9974 *) For servers, store verify_result in SSL_SESSION data structure
9975 (and add it to external session representation).
9976 This is needed when client certificate verifications fails,
9977 but an application-provided verification callback (set by
9978 SSL_CTX_set_cert_verify_callback) allows accepting the session
9979 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9980 but returns 1): When the session is reused, we have to set
9981 ssl->verify_result to the appropriate error code to avoid
9983 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9985 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9986 case in PKCS7_dataInit() where the signed PKCS7 structure
9987 didn't contain any existing data because it was being created.
9988 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9990 *) Add a salt to the key derivation routines in enc.c. This
9991 forms the first 8 bytes of the encrypted file. Also add a
9992 -S option to allow a salt to be input on the command line.
9995 *) New function X509_cmp(). Oddly enough there wasn't a function
9996 to compare two certificates. We do this by working out the SHA1
9997 hash and comparing that. X509_cmp() will be needed by the trust
10001 *) SSL_get1_session() is like SSL_get_session(), but increments
10002 the reference count in the SSL_SESSION returned.
10003 [Geoff Thorpe <geoff@eu.c2.net>]
10005 *) Fix for 'req': it was adding a null to request attributes.
10006 Also change the X509_LOOKUP and X509_INFO code to handle
10007 certificate auxiliary information.
10010 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10014 *) Add the possibility to add extra information to the memory leak
10015 detecting output, to form tracebacks, showing from where each
10016 allocation was originated: CRYPTO_push_info("constant string") adds
10017 the string plus current file name and line number to a per-thread
10018 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10019 is like calling CYRPTO_pop_info() until the stack is empty.
10020 Also updated memory leak detection code to be multi-thread-safe.
10023 *) Add options -text and -noout to pkcs7 utility and delete the
10024 encryption options which never did anything. Update docs.
10027 *) Add options to some of the utilities to allow the pass phrase
10028 to be included on either the command line (not recommended on
10029 OSes like Unix) or read from the environment. Update the
10030 manpages and fix a few bugs.
10033 *) Add a few manpages for some of the openssl commands.
10036 *) Fix the -revoke option in ca. It was freeing up memory twice,
10037 leaking and not finding already revoked certificates.
10040 *) Extensive changes to support certificate auxiliary information.
10041 This involves the use of X509_CERT_AUX structure and X509_AUX
10042 functions. An X509_AUX function such as PEM_read_X509_AUX()
10043 can still read in a certificate file in the usual way but it
10044 will also read in any additional "auxiliary information". By
10045 doing things this way a fair degree of compatibility can be
10046 retained: existing certificates can have this information added
10047 using the new 'x509' options.
10049 Current auxiliary information includes an "alias" and some trust
10050 settings. The trust settings will ultimately be used in enhanced
10051 certificate chain verification routines: currently a certificate
10052 can only be trusted if it is self signed and then it is trusted
10056 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10057 The problem was that one of the replacement routines had not been working
10058 since SSLeay releases. For now the offending routine has been replaced
10059 with non-optimised assembler. Even so, this now gives around 95%
10060 performance improvement for 1024 bit RSA signs.
10063 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10064 handling. Most clients have the effective key size in bits equal to
10065 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10066 A few however don't do this and instead use the size of the decrypted key
10067 to determine the RC2 key length and the AlgorithmIdentifier to determine
10068 the effective key length. In this case the effective key length can still
10069 be 40 bits but the key length can be 168 bits for example. This is fixed
10070 by manually forcing an RC2 key into the EVP_PKEY structure because the
10071 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10072 the key length and effective key length are equal.
10075 *) Add a bunch of functions that should simplify the creation of
10076 X509_NAME structures. Now you should be able to do:
10077 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10078 and have it automatically work out the correct field type and fill in
10079 the structures. The more adventurous can try:
10080 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10081 and it will (hopefully) work out the correct multibyte encoding.
10084 *) Change the 'req' utility to use the new field handling and multibyte
10085 copy routines. Before the DN field creation was handled in an ad hoc
10086 way in req, ca, and x509 which was rather broken and didn't support
10087 BMPStrings or UTF8Strings. Since some software doesn't implement
10088 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10089 using the dirstring_type option. See the new comment in the default
10090 openssl.cnf for more info.
10093 *) Make crypto/rand/md_rand.c more robust:
10094 - Assure unique random numbers after fork().
10095 - Make sure that concurrent threads access the global counter and
10096 md serializably so that we never lose entropy in them
10097 or use exactly the same state in multiple threads.
10098 Access to the large state is not always serializable because
10099 the additional locking could be a performance killer, and
10100 md should be large enough anyway.
10103 *) New file apps/app_rand.c with commonly needed functionality
10104 for handling the random seed file.
10106 Use the random seed file in some applications that previously did not:
10108 dsaparam -genkey (which also ignored its '-rand' option),
10111 x509 (when signing).
10112 Except on systems with /dev/urandom, it is crucial to have a random
10113 seed file at least for key creation, DSA signing, and for DH exchanges;
10114 for RSA signatures we could do without one.
10116 gendh and gendsa (unlike genrsa) used to read only the first byte
10117 of each file listed in the '-rand' option. The function as previously
10118 found in genrsa is now in app_rand.c and is used by all programs
10119 that support '-rand'.
10122 *) In RAND_write_file, use mode 0600 for creating files;
10123 don't just chmod when it may be too late.
10126 *) Report an error from X509_STORE_load_locations
10127 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10130 *) New function ASN1_mbstring_copy() this copies a string in either
10131 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10132 into an ASN1_STRING type. A mask of permissible types is passed
10133 and it chooses the "minimal" type to use or an error if not type
10137 *) Add function equivalents to the various macros in asn1.h. The old
10138 macros are retained with an M_ prefix. Code inside the library can
10139 use the M_ macros. External code (including the openssl utility)
10140 should *NOT* in order to be "shared library friendly".
10143 *) Add various functions that can check a certificate's extensions
10144 to see if it usable for various purposes such as SSL client,
10145 server or S/MIME and CAs of these types. This is currently
10146 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10147 verification. Also added a -purpose flag to x509 utility to
10148 print out all the purposes.
10151 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10155 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10156 for, obtain and decode and extension and obtain its critical flag.
10157 This allows all the necessary extension code to be handled in a
10158 single function call.
10161 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10162 platforms. See crypto/rc4/rc4_enc.c for further details.
10165 *) New -noout option to asn1parse. This causes no output to be produced
10166 its main use is when combined with -strparse and -out to extract data
10167 from a file (which may not be in ASN.1 format).
10170 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10171 when producing the local key id.
10172 [Richard Levitte <levitte@stacken.kth.se>]
10174 *) New option -dhparam in s_server. This allows a DH parameter file to be
10175 stated explicitly. If it is not stated then it tries the first server
10176 certificate file. The previous behaviour hard coded the filename
10180 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10181 a public key to be input or output. For example:
10182 openssl rsa -in key.pem -pubout -out pubkey.pem
10183 Also added necessary DSA public key functions to handle this.
10186 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10187 in the message. This was handled by allowing
10188 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10189 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10191 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10192 to the end of the strings whereas this didn't. This would cause problems
10193 if strings read with d2i_ASN1_bytes() were later modified.
10194 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10196 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10197 data and it contains EOF it will end up returning an error. This is
10198 caused by input 46 bytes long. The cause is due to the way base64
10199 BIOs find the start of base64 encoded data. They do this by trying a
10200 trial decode on each line until they find one that works. When they
10201 do a flag is set and it starts again knowing it can pass all the
10202 data directly through the decoder. Unfortunately it doesn't reset
10203 the context it uses. This means that if EOF is reached an attempt
10204 is made to pass two EOFs through the context and this causes the
10205 resulting error. This can also cause other problems as well. As is
10206 usual with these problems it takes *ages* to find and the fix is
10207 trivial: move one line.
10208 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10210 *) Ugly workaround to get s_client and s_server working under Windows. The
10211 old code wouldn't work because it needed to select() on sockets and the
10212 tty (for keypresses and to see if data could be written). Win32 only
10213 supports select() on sockets so we select() with a 1s timeout on the
10214 sockets and then see if any characters are waiting to be read, if none
10215 are present then we retry, we also assume we can always write data to
10216 the tty. This isn't nice because the code then blocks until we've
10217 received a complete line of data and it is effectively polling the
10218 keyboard at 1s intervals: however it's quite a bit better than not
10219 working at all :-) A dedicated Windows application might handle this
10220 with an event loop for example.
10223 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10224 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10225 will be called when RSA_sign() and RSA_verify() are used. This is useful
10226 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10227 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10228 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10229 This necessitated the support of an extra signature type NID_md5_sha1
10230 for SSL signatures and modifications to the SSL library to use it instead
10231 of calling RSA_public_decrypt() and RSA_private_encrypt().
10234 *) Add new -verify -CAfile and -CApath options to the crl program, these
10235 will lookup a CRL issuers certificate and verify the signature in a
10236 similar way to the verify program. Tidy up the crl program so it
10237 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10238 less strict. It will now permit CRL extensions even if it is not
10239 a V2 CRL: this will allow it to tolerate some broken CRLs.
10242 *) Initialize all non-automatic variables each time one of the openssl
10243 sub-programs is started (this is necessary as they may be started
10244 multiple times from the "OpenSSL>" prompt).
10245 [Lennart Bang, Bodo Moeller]
10247 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10248 removing all other RSA functionality (this is what NO_RSA does). This
10249 is so (for example) those in the US can disable those operations covered
10250 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10254 *) Non-copying interface to BIO pairs.
10255 (still largely untested)
10258 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10259 ASCII string. This was handled independently in various places before.
10262 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10263 UTF8 strings a character at a time.
10266 *) Use client_version from client hello to select the protocol
10267 (s23_srvr.c) and for RSA client key exchange verification
10268 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10271 *) Add various utility functions to handle SPKACs, these were previously
10272 handled by poking round in the structure internals. Added new function
10273 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10274 print, verify and generate SPKACs. Based on an original idea from
10275 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10278 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10281 *) Allow the config file extension section to be overwritten on the
10282 command line. Based on an original idea from Massimiliano Pala
10283 <madwolf@comune.modena.it>. The new option is called -extensions
10284 and can be applied to ca, req and x509. Also -reqexts to override
10285 the request extensions in req and -crlexts to override the crl extensions
10289 *) Add new feature to the SPKAC handling in ca. Now you can include
10290 the same field multiple times by preceding it by "XXXX." for example:
10293 this is the same syntax as used in the req config file.
10296 *) Allow certificate extensions to be added to certificate requests. These
10297 are specified in a 'req_extensions' option of the req section of the
10298 config file. They can be printed out with the -text option to req but
10299 are otherwise ignored at present.
10302 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10303 data read consists of only the final block it would not decrypted because
10304 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10305 A misplaced 'break' also meant the decrypted final block might not be
10306 copied until the next read.
10309 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10310 a few extra parameters to the DH structure: these will be useful if
10311 for example we want the value of 'q' or implement X9.42 DH.
10314 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10315 provides hooks that allow the default DSA functions or functions on a
10316 "per key" basis to be replaced. This allows hardware acceleration and
10317 hardware key storage to be handled without major modification to the
10318 library. Also added low level modexp hooks and CRYPTO_EX structure and
10319 associated functions.
10322 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10323 as "read only": it can't be written to and the buffer it points to will
10324 not be freed. Reading from a read only BIO is much more efficient than
10325 a normal memory BIO. This was added because there are several times when
10326 an area of memory needs to be read from a BIO. The previous method was
10327 to create a memory BIO and write the data to it, this results in two
10328 copies of the data and an O(n^2) reading algorithm. There is a new
10329 function BIO_new_mem_buf() which creates a read only memory BIO from
10330 an area of memory. Also modified the PKCS#7 routines to use read only
10334 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10335 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10336 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10337 but a retry condition occurred while trying to read the rest.
10340 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10341 NID_pkcs7_encrypted by default: this was wrong since this should almost
10342 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10343 the encrypted data type: this is a more sensible place to put it and it
10344 allows the PKCS#12 code to be tidied up that duplicated this
10348 *) Changed obj_dat.pl script so it takes its input and output files on
10349 the command line. This should avoid shell escape redirection problems
10353 *) Initial support for certificate extension requests, these are included
10354 in things like Xenroll certificate requests. Included functions to allow
10355 extensions to be obtained and added.
10358 *) -crlf option to s_client and s_server for sending newlines as
10359 CRLF (as required by many protocols).
10362 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10364 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10365 [Ralf S. Engelschall]
10367 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10368 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10370 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10374 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10375 DH parameters/keys (q is lost during that conversion, but the resulting
10376 DH parameters contain its length).
10378 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10379 much faster than DH_generate_parameters (which creates parameters
10380 where p = 2*q + 1), and also the smaller q makes DH computations
10381 much more efficient (160-bit exponentiation instead of 1024-bit
10382 exponentiation); so this provides a convenient way to support DHE
10383 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10384 utter importance to use
10385 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10387 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10388 when such DH parameters are used, because otherwise small subgroup
10389 attacks may become possible!
10392 *) Avoid memory leak in i2d_DHparams.
10395 *) Allow the -k option to be used more than once in the enc program:
10396 this allows the same encrypted message to be read by multiple recipients.
10399 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10400 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10401 it will always use the numerical form of the OID, even if it has a short
10405 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10406 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10407 otherwise bn_mod_exp was called. In the case of hardware keys for example
10408 no private key components need be present and it might store extra data
10409 in the RSA structure, which cannot be accessed from bn_mod_exp.
10410 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10411 private key operations.
10414 *) Added support for SPARC Linux.
10417 *) pem_password_cb function type incompatibly changed from
10418 typedef int pem_password_cb(char *buf, int size, int rwflag);
10420 ....(char *buf, int size, int rwflag, void *userdata);
10421 so that applications can pass data to their callbacks:
10422 The PEM[_ASN1]_{read,write}... functions and macros now take an
10423 additional void * argument, which is just handed through whenever
10424 the password callback is called.
10425 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10427 New function SSL_CTX_set_default_passwd_cb_userdata.
10429 Compatibility note: As many C implementations push function arguments
10430 onto the stack in reverse order, the new library version is likely to
10431 interoperate with programs that have been compiled with the old
10432 pem_password_cb definition (PEM_whatever takes some data that
10433 happens to be on the stack as its last argument, and the callback
10434 just ignores this garbage); but there is no guarantee whatsoever that
10437 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10438 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10439 problems not only on Windows, but also on some Unix platforms.
10440 To avoid problematic command lines, these definitions are now in an
10441 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10442 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10445 *) MIPS III/IV assembler module is reimplemented.
10448 *) More DES library cleanups: remove references to srand/rand and
10449 delete an unused file.
10452 *) Add support for the the free Netwide assembler (NASM) under Win32,
10453 since not many people have MASM (ml) and it can be hard to obtain.
10454 This is currently experimental but it seems to work OK and pass all
10455 the tests. Check out INSTALL.W32 for info.
10458 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10459 without temporary keys kept an extra copy of the server key,
10460 and connections with temporary keys did not free everything in case
10464 *) New function RSA_check_key and new openssl rsa option -check
10465 for verifying the consistency of RSA keys.
10466 [Ulf Moeller, Bodo Moeller]
10468 *) Various changes to make Win32 compile work:
10469 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10470 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10471 comparison" warnings.
10472 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10475 *) Add a debugging option to PKCS#5 v2 key generation function: when
10476 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10477 derived keys are printed to stderr.
10480 *) Copy the flags in ASN1_STRING_dup().
10481 [Roman E. Pavlov <pre@mo.msk.ru>]
10483 *) The x509 application mishandled signing requests containing DSA
10484 keys when the signing key was also DSA and the parameters didn't match.
10486 It was supposed to omit the parameters when they matched the signing key:
10487 the verifying software was then supposed to automatically use the CA's
10488 parameters if they were absent from the end user certificate.
10490 Omitting parameters is no longer recommended. The test was also
10491 the wrong way round! This was probably due to unusual behaviour in
10492 EVP_cmp_parameters() which returns 1 if the parameters match.
10493 This meant that parameters were omitted when they *didn't* match and
10494 the certificate was useless. Certificates signed with 'ca' didn't have
10496 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10498 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10499 The interface is as follows:
10500 Applications can use
10501 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10502 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10503 "off" is now the default.
10504 The library internally uses
10505 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10506 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10507 to disable memory-checking temporarily.
10509 Some inconsistent states that previously were possible (and were
10510 even the default) are now avoided.
10512 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10513 with each memory chunk allocated; this is occasionally more helpful
10514 than just having a counter.
10516 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10518 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10522 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10523 which largely parallels "options", but is for changing API behaviour,
10524 whereas "options" are about protocol behaviour.
10525 Initial "mode" flags are:
10527 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10528 a single record has been written.
10529 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10530 retries use the same buffer location.
10531 (But all of the contents must be
10535 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10538 *) Fix problems with no-hmac etc.
10539 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10541 *) New functions RSA_get_default_method(), RSA_set_method() and
10542 RSA_get_method(). These allows replacement of RSA_METHODs without having
10543 to mess around with the internals of an RSA structure.
10546 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10547 Also really enable memory leak checks in openssl.c and in some
10549 [Chad C. Mulligan, Bodo Moeller]
10551 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10552 up the length of negative integers. This has now been simplified to just
10553 store the length when it is first determined and use it later, rather
10554 than trying to keep track of where data is copied and updating it to
10556 [Steve Henson, reported by Brien Wheeler
10557 <bwheeler@authentica-security.com>]
10559 *) Add a new function PKCS7_signatureVerify. This allows the verification
10560 of a PKCS#7 signature but with the signing certificate passed to the
10561 function itself. This contrasts with PKCS7_dataVerify which assumes the
10562 certificate is present in the PKCS#7 structure. This isn't always the
10563 case: certificates can be omitted from a PKCS#7 structure and be
10564 distributed by "out of band" means (such as a certificate database).
10567 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10568 function prototypes in pem.h, also change util/mkdef.pl to add the
10569 necessary function names.
10572 *) mk1mf.pl (used by Windows builds) did not properly read the
10573 options set by Configure in the top level Makefile, and Configure
10574 was not even able to write more than one option correctly.
10575 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10578 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10579 file to be loaded from a BIO or FILE pointer. The BIO version will
10580 for example allow memory BIOs to contain config info.
10583 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10584 Whoever hopes to achieve shared-library compatibility across versions
10585 must use this, not the compile-time macro.
10586 (Exercise 0.9.4: Which is the minimum library version required by
10588 Note: All this applies only to multi-threaded programs, others don't
10592 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10593 through a BIO pair triggered the default case, i.e.
10594 SSLerr(...,SSL_R_UNKNOWN_STATE).
10597 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10598 can use the SSL library even if none of the specific BIOs is
10602 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10603 for the encoded length.
10604 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10606 *) Add initial documentation of the X509V3 functions.
10609 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10610 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10611 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10612 secure PKCS#8 private key format with a high iteration count.
10615 *) Fix determination of Perl interpreter: A perl or perl5
10616 _directory_ in $PATH was also accepted as the interpreter.
10617 [Ralf S. Engelschall]
10619 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10620 wrong with it but it was very old and did things like calling
10621 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10622 unusual formatting.
10625 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10626 to use the new extension code.
10629 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10630 with macros. This should make it easier to change their form, add extra
10631 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10635 *) Add to configuration table a new entry that can specify an alternative
10636 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10637 according to Mark Crispin <MRC@Panda.COM>.
10641 *) DES CBC did not update the IV. Weird.
10644 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10645 Changing the behaviour of the former might break existing programs --
10646 where IV updating is needed, des_ncbc_encrypt can be used.
10649 *) When bntest is run from "make test" it drives bc to check its
10650 calculations, as well as internally checking them. If an internal check
10651 fails, it needs to cause bc to give a non-zero result or make test carries
10652 on without noticing the failure. Fixed.
10655 *) DES library cleanups.
10658 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10659 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10660 ciphers. NOTE: although the key derivation function has been verified
10661 against some published test vectors it has not been extensively tested
10662 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10666 *) Instead of "mkdir -p", which is not fully portable, use new
10667 Perl script "util/mkdir-p.pl".
10670 *) Rewrite the way password based encryption (PBE) is handled. It used to
10671 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10672 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10673 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10674 the 'parameter' field of the AlgorithmIdentifier is passed to the
10675 underlying key generation function so it must do its own ASN1 parsing.
10676 This has also changed the EVP_PBE_CipherInit() function which now has a
10677 'parameter' argument instead of literal salt and iteration count values
10678 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10681 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10682 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10683 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10684 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10685 value was just used as a "magic string" and not used directly its
10686 value doesn't matter.
10689 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10693 *) "linux-sparc64" configuration (ultrapenguin).
10694 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10695 "linux-sparc" configuration.
10696 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10698 *) config now generates no-xxx options for missing ciphers.
10701 *) Support the EBCDIC character set (work in progress).
10702 File ebcdic.c not yet included because it has a different license.
10703 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10705 *) Support BS2000/OSD-POSIX.
10706 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10708 *) Make callbacks for key generation use void * instead of char *.
10711 *) Make S/MIME samples compile (not yet tested).
10714 *) Additional typesafe stacks.
10717 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10721 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10723 *) New configuration variant "sco5-gcc".
10725 *) Updated some demos.
10726 [Sean O Riordain, Wade Scholine]
10728 *) Add missing BIO_free at exit of pkcs12 application.
10731 *) Fix memory leak in conf.c.
10734 *) Updates for Win32 to assembler version of MD5.
10737 *) Set #! path to perl in apps/der_chop to where we found it
10738 instead of using a fixed path.
10741 *) SHA library changes for irix64-mips4-cc.
10744 *) Improvements for VMS support.
10748 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10750 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10751 This also avoids the problems with SC4.2 and unpatched SC5.
10752 [Andy Polyakov <appro@fy.chalmers.se>]
10754 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10755 These are required because of the typesafe stack would otherwise break
10756 existing code. If old code used a structure member which used to be STACK
10757 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10758 sk_num or sk_value it would produce an error because the num, data members
10759 are not present in STACK_OF. Now it just produces a warning. sk_set
10760 replaces the old method of assigning a value to sk_value
10761 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10762 that does this will no longer work (and should use sk_set instead) but
10763 this could be regarded as a "questionable" behaviour anyway.
10766 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10767 correctly handle encrypted S/MIME data.
10770 *) Change type of various DES function arguments from des_cblock
10771 (which means, in function argument declarations, pointer to char)
10772 to des_cblock * (meaning pointer to array with 8 char elements),
10773 which allows the compiler to do more typechecking; it was like
10774 that back in SSLeay, but with lots of ugly casts.
10776 Introduce new type const_des_cblock.
10779 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10780 problems: find RecipientInfo structure that matches recipient certificate
10781 and initialise the ASN1 structures properly based on passed cipher.
10784 *) Belatedly make the BN tests actually check the results.
10787 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10788 to and from BNs: it was completely broken. New compilation option
10789 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10790 key elements as negative integers.
10793 *) Reorganize and speed up MD5.
10794 [Andy Polyakov <appro@fy.chalmers.se>]
10797 [Richard Levitte <richard@levitte.org>]
10799 *) New option -out to asn1parse to allow the parsed structure to be
10800 output to a file. This is most useful when combined with the -strparse
10801 option to examine the output of things like OCTET STRINGS.
10804 *) Make SSL library a little more fool-proof by not requiring any longer
10805 that SSL_set_{accept,connect}_state be called before
10806 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10807 in many applications because usually everything *appeared* to work as
10808 intended anyway -- now it really works as intended).
10811 *) Move openssl.cnf out of lib/.
10814 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10815 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10816 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10817 [Ralf S. Engelschall]
10819 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10820 handle PKCS#7 enveloped data properly.
10821 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10823 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10824 copying pointers. The cert_st handling is changed by this in
10825 various ways (and thus what used to be known as ctx->default_cert
10826 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10827 any longer when s->cert does not give us what we need).
10828 ssl_cert_instantiate becomes obsolete by this change.
10829 As soon as we've got the new code right (possibly it already is?),
10830 we have solved a couple of bugs of the earlier code where s->cert
10831 was used as if it could not have been shared with other SSL structures.
10833 Note that using the SSL API in certain dirty ways now will result
10834 in different behaviour than observed with earlier library versions:
10835 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10836 does not influence s as it used to.
10838 In order to clean up things more thoroughly, inside SSL_SESSION
10839 we don't use CERT any longer, but a new structure SESS_CERT
10840 that holds per-session data (if available); currently, this is
10841 the peer's certificate chain and, for clients, the server's certificate
10842 and temporary key. CERT holds only those values that can have
10843 meaningful defaults in an SSL_CTX.
10846 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10847 from the internal representation. Various PKCS#7 fixes: remove some
10848 evil casts and set the enc_dig_alg field properly based on the signing
10852 *) Allow PKCS#12 password to be set from the command line or the
10853 environment. Let 'ca' get its config file name from the environment
10854 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10858 *) Allow certificate policies extension to use an IA5STRING for the
10859 organization field. This is contrary to the PKIX definition but
10860 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10864 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10865 without disallowing inline assembler and the like for non-pedantic builds.
10868 *) Support Borland C++ builder.
10869 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10871 *) Support Mingw32.
10874 *) SHA-1 cleanups and performance enhancements.
10875 [Andy Polyakov <appro@fy.chalmers.se>]
10877 *) Sparc v8plus assembler for the bignum library.
10878 [Andy Polyakov <appro@fy.chalmers.se>]
10880 *) Accept any -xxx and +xxx compiler options in Configure.
10883 *) Update HPUX configuration.
10886 *) Add missing sk_<type>_unshift() function to safestack.h
10887 [Ralf S. Engelschall]
10889 *) New function SSL_CTX_use_certificate_chain_file that sets the
10890 "extra_cert"s in addition to the certificate. (This makes sense
10891 only for "PEM" format files, as chains as a whole are not
10895 *) Support verify_depth from the SSL API.
10896 x509_vfy.c had what can be considered an off-by-one-error:
10897 Its depth (which was not part of the external interface)
10898 was actually counting the number of certificates in a chain;
10899 now it really counts the depth.
10902 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10903 instead of X509err, which often resulted in confusing error
10904 messages since the error codes are not globally unique
10905 (e.g. an alleged error in ssl3_accept when a certificate
10906 didn't match the private key).
10908 *) New function SSL_CTX_set_session_id_context that allows to set a default
10909 value (so that you don't need SSL_set_session_id_context for each
10910 connection using the SSL_CTX).
10913 *) OAEP decoding bug fix.
10916 *) Support INSTALL_PREFIX for package builders, as proposed by
10920 *) New Configure options "threads" and "no-threads". For systems
10921 where the proper compiler options are known (currently Solaris
10922 and Linux), "threads" is the default.
10925 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10928 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10929 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10930 such as /usr/local/bin.
10933 *) "make linux-shared" to build shared libraries.
10934 [Niels Poppe <niels@netbox.org>]
10936 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10939 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10940 extension adding in x509 utility.
10943 *) Remove NOPROTO sections and error code comments.
10946 *) Partial rewrite of the DEF file generator to now parse the ANSI
10950 *) New Configure options --prefix=DIR and --openssldir=DIR.
10953 *) Complete rewrite of the error code script(s). It is all now handled
10954 by one script at the top level which handles error code gathering,
10955 header rewriting and C source file generation. It should be much better
10956 than the old method: it now uses a modified version of Ulf's parser to
10957 read the ANSI prototypes in all header files (thus the old K&R definitions
10958 aren't needed for error creation any more) and do a better job of
10959 translating function codes into names. The old 'ASN1 error code imbedded
10960 in a comment' is no longer necessary and it doesn't use .err files which
10961 have now been deleted. Also the error code call doesn't have to appear all
10962 on one line (which resulted in some large lines...).
10965 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10968 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10969 0 (which usually indicates a closed connection), but continue reading.
10972 *) Fix some race conditions.
10975 *) Add support for CRL distribution points extension. Add Certificate
10976 Policies and CRL distribution points documentation.
10979 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10982 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10983 8 of keying material. Merlin has also confirmed interop with this fix
10984 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10985 [Merlin Hughes <merlin@baltimore.ie>]
10987 *) Fix lots of warnings.
10988 [Richard Levitte <levitte@stacken.kth.se>]
10990 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10991 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10992 [Richard Levitte <levitte@stacken.kth.se>]
10994 *) Fix problems with sizeof(long) == 8.
10995 [Andy Polyakov <appro@fy.chalmers.se>]
10997 *) Change functions to ANSI C.
11000 *) Fix typos in error codes.
11001 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11003 *) Remove defunct assembler files from Configure.
11006 *) SPARC v8 assembler BIGNUM implementation.
11007 [Andy Polyakov <appro@fy.chalmers.se>]
11009 *) Support for Certificate Policies extension: both print and set.
11010 Various additions to support the r2i method this uses.
11013 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11014 return a const string when you are expecting an allocated buffer.
11017 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11018 types DirectoryString and DisplayText.
11021 *) Add code to allow r2i extensions to access the configuration database,
11022 add an LHASH database driver and add several ctx helper functions.
11025 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11026 fail when they extended the size of a BIGNUM.
11029 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11030 support typesafe stack.
11033 *) Fix typo in SSL_[gs]et_options().
11034 [Nils Frostberg <nils@medcom.se>]
11036 *) Delete various functions and files that belonged to the (now obsolete)
11037 old X509V3 handling code.
11040 *) New Configure option "rsaref".
11043 *) Don't auto-generate pem.h.
11046 *) Introduce type-safe ASN.1 SETs.
11049 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11050 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11052 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11053 that links with OpenSSL (well at least cause lots of warnings), but fear
11054 not: the conversion is trivial, and it eliminates loads of evil casts. A
11055 few STACKed things have been converted already. Feel free to convert more.
11056 In the fullness of time, I'll do away with the STACK type altogether.
11059 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11060 specified in <certfile> by updating the entry in the index.txt file.
11061 This way one no longer has to edit the index.txt file manually for
11062 revoking a certificate. The -revoke option does the gory details now.
11063 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11065 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11066 `-text' option at all and this way the `-noout -text' combination was
11067 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11068 [Ralf S. Engelschall]
11070 *) Make sure a corresponding plain text error message exists for the
11071 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11072 verify callback function determined that a certificate was revoked.
11073 [Ralf S. Engelschall]
11075 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11076 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11077 all available cipers including rc5, which was forgotten until now.
11078 In order to let the testing shell script know which algorithms
11079 are available, a new (up to now undocumented) command
11080 "openssl list-cipher-commands" is used.
11083 *) Bugfix: s_client occasionally would sleep in select() when
11084 it should have checked SSL_pending() first.
11087 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11088 the raw DSA values prior to ASN.1 encoding.
11091 *) Tweaks to Configure
11092 [Niels Poppe <niels@netbox.org>]
11094 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11098 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11101 *) New config option to avoid instructions that are illegal on the 80386.
11102 The default code is faster, but requires at least a 486.
11105 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11106 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11107 same as SSL2_VERSION anyway.
11110 *) New "-showcerts" option for s_client.
11113 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11114 application. Various cleanups and fixes.
11117 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11118 modify error routines to work internally. Add error codes and PBE init
11119 to library startup routines.
11122 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11123 packing functions to asn1 and evp. Changed function names and error
11124 codes along the way.
11127 *) PKCS12 integration: and so it begins... First of several patches to
11128 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11129 objects to objects.h
11132 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11133 and display support for Thawte strong extranet extension.
11136 *) Add LinuxPPC support.
11137 [Jeff Dubrule <igor@pobox.org>]
11139 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11140 bn_div_words in alpha.s.
11141 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11143 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11144 OAEP isn't supported when OpenSSL is built with RSAref.
11145 [Ulf Moeller <ulf@fitug.de>]
11147 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11148 so they no longer are missing under -DNOPROTO.
11149 [Soren S. Jorvang <soren@t.dk>]
11152 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11154 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11155 doesn't work when the session is reused. Coming soon!
11158 *) Fix a security hole, that allows sessions to be reused in the wrong
11159 context thus bypassing client cert protection! All software that uses
11160 client certs and session caches in multiple contexts NEEDS PATCHING to
11161 allow session reuse! A fuller solution is in the works.
11162 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11164 *) Some more source tree cleanups (removed obsolete files
11165 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11166 permission on "config" script to be executable) and a fix for the INSTALL
11168 [Ulf Moeller <ulf@fitug.de>]
11170 *) Remove some legacy and erroneous uses of malloc, free instead of
11172 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11174 *) Make rsa_oaep_test return non-zero on error.
11175 [Ulf Moeller <ulf@fitug.de>]
11177 *) Add support for native Solaris shared libraries. Configure
11178 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11179 if someone would make that last step automatic.
11180 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11182 *) ctx_size was not built with the right compiler during "make links". Fixed.
11185 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11186 except NULL ciphers". This means the default cipher list will no longer
11187 enable NULL ciphers. They need to be specifically enabled e.g. with
11188 the string "DEFAULT:eNULL".
11191 *) Fix to RSA private encryption routines: if p < q then it would
11192 occasionally produce an invalid result. This will only happen with
11193 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11196 *) Be less restrictive and allow also `perl util/perlpath.pl
11197 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11198 because this way one can also use an interpreter named `perl5' (which is
11199 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11200 installed as `perl').
11201 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11203 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11204 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11206 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11207 advapi32.lib to Win32 build and change the pem test comparison
11208 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11209 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11210 and crypto/des/ede_cbcm_enc.c.
11213 *) DES quad checksum was broken on big-endian architectures. Fixed.
11216 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11217 Win32 test batch file so it (might) work again. The Win32 test batch file
11218 is horrible: I feel ill....
11221 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11222 in e_os.h. Audit of header files to check ANSI and non ANSI
11223 sections: 10 functions were absent from non ANSI section and not exported
11224 from Windows DLLs. Fixed up libeay.num for new functions.
11227 *) Make `openssl version' output lines consistent.
11228 [Ralf S. Engelschall]
11230 *) Fix Win32 symbol export lists for BIO functions: Added
11231 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11232 to ms/libeay{16,32}.def.
11233 [Ralf S. Engelschall]
11235 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11236 fine under Unix and passes some trivial tests I've now added. But the
11237 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11238 added to make sure no one expects that this stuff really works in the
11239 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11240 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11242 [Ralf S. Engelschall]
11244 *) Fix the generation of two part addresses in perl.
11245 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11247 *) Add config entry for Linux on MIPS.
11248 [John Tobey <jtobey@channel1.com>]
11250 *) Make links whenever Configure is run, unless we are on Windoze.
11253 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11254 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11258 *) Add a useful kludge to allow package maintainers to specify compiler and
11259 other platforms details on the command line without having to patch the
11260 Configure script everytime: One now can use ``perl Configure
11261 <id>:<details>'', i.e. platform ids are allowed to have details appended
11262 to them (separated by colons). This is treated as there would be a static
11263 pre-configured entry in Configure's %table under key <id> with value
11264 <details> and ``perl Configure <id>'' is called. So, when you want to
11265 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11266 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11267 now, which overrides the FreeBSD-elf entry on-the-fly.
11268 [Ralf S. Engelschall]
11270 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11273 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11274 on the `perl Configure ...' command line. This way one can compile
11275 OpenSSL libraries with Position Independent Code (PIC) which is needed
11276 for linking it into DSOs.
11277 [Ralf S. Engelschall]
11279 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11283 *) Cleaned up the LICENSE document: The official contact for any license
11284 questions now is the OpenSSL core team under openssl-core@openssl.org.
11285 And add a paragraph about the dual-license situation to make sure people
11286 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11287 to the OpenSSL toolkit.
11288 [Ralf S. Engelschall]
11290 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11291 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11292 Additionally cleaned up the `make links' target: Remove unnecessary
11293 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11294 to speed processing and no longer clutter the display with confusing
11295 stuff. Instead only the actually done links are displayed.
11296 [Ralf S. Engelschall]
11298 *) Permit null encryption ciphersuites, used for authentication only. It used
11299 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11300 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11304 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11305 signed attributes when verifying signatures (this would break them),
11306 the detached data encoding was wrong and public keys obtained using
11307 X509_get_pubkey() weren't freed.
11310 *) Add text documentation for the BUFFER functions. Also added a work around
11311 to a Win95 console bug. This was triggered by the password read stuff: the
11312 last character typed gets carried over to the next fread(). If you were
11313 generating a new cert request using 'req' for example then the last
11314 character of the passphrase would be CR which would then enter the first
11318 *) Added the new `Includes OpenSSL Cryptography Software' button as
11319 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11320 button and can be used by applications based on OpenSSL to show the
11321 relationship to the OpenSSL project.
11322 [Ralf S. Engelschall]
11324 *) Remove confusing variables in function signatures in files
11325 ssl/ssl_lib.c and ssl/ssl.h.
11326 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11328 *) Don't install bss_file.c under PREFIX/include/
11329 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11331 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11332 functions that return function pointers and has support for NT specific
11333 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11334 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11335 unsigned to signed types: this was killing the Win32 compile.
11338 *) Add new certificate file to stack functions,
11339 SSL_add_dir_cert_subjects_to_stack() and
11340 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11341 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11342 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11343 This means that Apache-SSL and similar packages don't have to mess around
11344 to add as many CAs as they want to the preferred list.
11347 *) Experiment with doxygen documentation. Currently only partially applied to
11349 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11350 openssl.doxy as the configuration file.
11353 *) Get rid of remaining C++-style comments which strict C compilers hate.
11354 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11356 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11357 compiled in by default: it has problems with large keys.
11360 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11361 DH private keys and/or callback functions which directly correspond to
11362 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11363 is needed for applications which have to configure certificates on a
11364 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11366 For the RSA certificate situation is makes no difference, but
11367 for the DSA certificate situation this fixes the "no shared cipher"
11368 problem where the OpenSSL cipher selection procedure failed because the
11369 temporary keys were not overtaken from the context and the API provided
11370 no way to reconfigure them.
11371 The new functions now let applications reconfigure the stuff and they
11372 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11373 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11374 non-public-API function ssl_cert_instantiate() is used as a helper
11375 function and also to reduce code redundancy inside ssl_rsa.c.
11376 [Ralf S. Engelschall]
11378 *) Move s_server -dcert and -dkey options out of the undocumented feature
11379 area because they are useful for the DSA situation and should be
11380 recognized by the users.
11381 [Ralf S. Engelschall]
11383 *) Fix the cipher decision scheme for export ciphers: the export bits are
11384 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11385 SSL_EXP_MASK. So, the original variable has to be used instead of the
11386 already masked variable.
11387 [Richard Levitte <levitte@stacken.kth.se>]
11389 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11390 [Richard Levitte <levitte@stacken.kth.se>]
11392 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11393 from `int' to `unsigned int' because it's a length and initialized by
11394 EVP_DigestFinal() which expects an `unsigned int *'.
11395 [Richard Levitte <levitte@stacken.kth.se>]
11397 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11398 script. Instead use the usual Shell->Perl transition trick.
11399 [Ralf S. Engelschall]
11401 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11402 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11403 -noout -modulus' as it's already the case for `openssl rsa -noout
11404 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11405 currently the public key is printed (a decision which was already done by
11406 `openssl dsa -modulus' in the past) which serves a similar purpose.
11407 Additionally the NO_RSA no longer completely removes the whole -modulus
11408 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11410 [Ralf S. Engelschall]
11412 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11413 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11414 [Arne Ansper <arne@ats.cyber.ee>]
11416 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11417 to be added. Now both 'req' and 'ca' can use new objects defined in the
11421 *) Add cool BIO that does syslog (or event log on NT).
11422 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11424 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11425 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11426 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11427 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11430 *) Add preliminary config info for new extension code.
11433 *) Make RSA_NO_PADDING really use no padding.
11434 [Ulf Moeller <ulf@fitug.de>]
11436 *) Generate errors when private/public key check is done.
11439 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11440 for some CRL extensions and new objects added.
11443 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11444 key usage extension and fuller support for authority key id.
11447 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11448 padding method for RSA, which is recommended for new applications in PKCS
11449 #1 v2.0 (RFC 2437, October 1998).
11450 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11451 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11452 against Bleichbacher's attack on RSA.
11453 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11456 *) Updates to the new SSL compression code
11457 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11459 *) Fix so that the version number in the master secret, when passed
11460 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11461 (because the server will not accept higher), that the version number
11462 is 0x03,0x01, not 0x03,0x00
11463 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11465 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11466 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11467 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11470 *) Support for RAW extensions where an arbitrary extension can be
11471 created by including its DER encoding. See apps/openssl.cnf for
11475 *) Make sure latest Perl versions don't interpret some generated C array
11476 code as Perl array code in the crypto/err/err_genc.pl script.
11477 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11479 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11480 not many people have the assembler. Various Win32 compilation fixes and
11481 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11482 build instructions.
11485 *) Modify configure script 'Configure' to automatically create crypto/date.h
11486 file under Win32 and also build pem.h from pem.org. New script
11487 util/mkfiles.pl to create the MINFO file on environments that can't do a
11488 'make files': perl util/mkfiles.pl >MINFO should work.
11491 *) Major rework of DES function declarations, in the pursuit of correctness
11492 and purity. As a result, many evil casts evaporated, and some weirdness,
11493 too. You may find this causes warnings in your code. Zapping your evil
11494 casts will probably fix them. Mostly.
11497 *) Fix for a typo in asn1.h. Bug fix to object creation script
11498 obj_dat.pl. It considered a zero in an object definition to mean
11499 "end of object": none of the objects in objects.h have any zeros
11500 so it wasn't spotted.
11501 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11503 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11504 Masking (CBCM). In the absence of test vectors, the best I have been able
11505 to do is check that the decrypt undoes the encrypt, so far. Send me test
11506 vectors if you have them.
11509 *) Correct calculation of key length for export ciphers (too much space was
11510 allocated for null ciphers). This has not been tested!
11513 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11514 message is now correct (it understands "crypto" and "ssl" on its
11515 command line). There is also now an "update" option. This will update
11516 the util/ssleay.num and util/libeay.num files with any new functions.
11518 perl util/mkdef.pl crypto ssl update
11519 it will update them.
11522 *) Overhauled the Perl interface (perl/*):
11523 - ported BN stuff to OpenSSL's different BN library
11524 - made the perl/ source tree CVS-aware
11525 - renamed the package from SSLeay to OpenSSL (the files still contain
11526 their history because I've copied them in the repository)
11527 - removed obsolete files (the test scripts will be replaced
11528 by better Test::Harness variants in the future)
11529 [Ralf S. Engelschall]
11531 *) First cut for a very conservative source tree cleanup:
11532 1. merge various obsolete readme texts into doc/ssleay.txt
11533 where we collect the old documents and readme texts.
11534 2. remove the first part of files where I'm already sure that we no
11535 longer need them because of three reasons: either they are just temporary
11536 files which were left by Eric or they are preserved original files where
11537 I've verified that the diff is also available in the CVS via "cvs diff
11538 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11539 the crypto/md/ stuff).
11540 [Ralf S. Engelschall]
11542 *) More extension code. Incomplete support for subject and issuer alt
11543 name, issuer and authority key id. Change the i2v function parameters
11544 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11545 what that's for :-) Fix to ASN1 macro which messed up
11546 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11549 *) Preliminary support for ENUMERATED type. This is largely copied from the
11553 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11554 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11556 *) Make sure `make rehash' target really finds the `openssl' program.
11557 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11559 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11560 like to hear about it if this slows down other processors.
11563 *) Add CygWin32 platform information to Configure script.
11564 [Alan Batie <batie@aahz.jf.intel.com>]
11566 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11567 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11569 *) New program nseq to manipulate netscape certificate sequences
11572 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11576 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11577 but the BN code had some problems that would cause failures when
11578 doing certificate verification and some other functions.
11579 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11581 *) Add ASN1 and PEM code to support netscape certificate sequences.
11584 *) Add ASN1 and PEM code to support netscape certificate sequences.
11587 *) Add several PKIX and private extended key usage OIDs.
11590 *) Modify the 'ca' program to handle the new extension code. Modify
11591 openssl.cnf for new extension format, add comments.
11594 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11595 and add a sample to openssl.cnf so req -x509 now adds appropriate
11599 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11600 error code, add initial support to X509_print() and x509 application.
11603 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11604 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11605 stuff is currently isolated and isn't even compiled yet.
11608 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11609 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11610 Removed the versions check from X509 routines when loading extensions:
11611 this allows certain broken certificates that don't set the version
11612 properly to be processed.
11615 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11616 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11617 can still be regenerated with "make depend".
11620 *) Spelling mistake in C version of CAST-128.
11621 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11623 *) Changes to the error generation code. The perl script err-code.pl
11624 now reads in the old error codes and retains the old numbers, only
11625 adding new ones if necessary. It also only changes the .err files if new
11626 codes are added. The makefiles have been modified to only insert errors
11627 when needed (to avoid needlessly modifying header files). This is done
11628 by only inserting errors if the .err file is newer than the auto generated
11629 C file. To rebuild all the error codes from scratch (the old behaviour)
11630 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11631 or delete all the .err files.
11634 *) CAST-128 was incorrectly implemented for short keys. The C version has
11635 been fixed, but is untested. The assembler versions are also fixed, but
11636 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11637 to regenerate it if needed.
11638 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11639 Hagino <itojun@kame.net>]
11641 *) File was opened incorrectly in randfile.c.
11642 [Ulf Möller <ulf@fitug.de>]
11644 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11645 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11646 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11647 al: it's just almost always a UTCTime. Note this patch adds new error
11648 codes so do a "make errors" if there are problems.
11651 *) Correct Linux 1 recognition in config.
11652 [Ulf Möller <ulf@fitug.de>]
11654 *) Remove pointless MD5 hash when using DSA keys in ca.
11655 [Anonymous <nobody@replay.com>]
11657 *) Generate an error if given an empty string as a cert directory. Also
11658 generate an error if handed NULL (previously returned 0 to indicate an
11659 error, but didn't set one).
11660 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11662 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11665 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11666 parameters. This was causing a warning which killed off the Win32 compile.
11669 *) Remove C++ style comments from crypto/bn/bn_local.h.
11670 [Neil Costigan <neil.costigan@celocom.com>]
11672 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11673 based on a text string, looking up short and long names and finally
11674 "dot" format. The "dot" format stuff didn't work. Added new function
11675 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11676 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11677 OID is not part of the table.
11680 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11681 X509_LOOKUP_by_alias().
11684 *) Sort openssl functions by name.
11687 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11688 encryption from sample DSA keys (in case anyone is interested the password
11692 *) Make _all_ *_free functions accept a NULL pointer.
11693 [Frans Heymans <fheymans@isaserver.be>]
11695 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11697 [Anonymous <nobody@replay.com>]
11699 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11700 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11702 *) Don't blow it for numeric -newkey arguments to apps/req.
11703 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11705 *) Temp key "for export" tests were wrong in s3_srvr.c.
11706 [Anonymous <nobody@replay.com>]
11708 *) Add prototype for temp key callback functions
11709 SSL_CTX_set_tmp_{rsa,dh}_callback().
11712 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11713 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11716 *) X509_name_add_entry() freed the wrong thing after an error.
11717 [Arne Ansper <arne@ats.cyber.ee>]
11719 *) rsa_eay.c would attempt to free a NULL context.
11720 [Arne Ansper <arne@ats.cyber.ee>]
11722 *) BIO_s_socket() had a broken should_retry() on Windoze.
11723 [Arne Ansper <arne@ats.cyber.ee>]
11725 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11726 [Arne Ansper <arne@ats.cyber.ee>]
11728 *) Make sure the already existing X509_STORE->depth variable is initialized
11729 in X509_STORE_new(), but document the fact that this variable is still
11730 unused in the certificate verification process.
11731 [Ralf S. Engelschall]
11733 *) Fix the various library and apps files to free up pkeys obtained from
11734 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11737 *) Fix reference counting in X509_PUBKEY_get(). This makes
11738 demos/maurice/example2.c work, amongst others, probably.
11739 [Steve Henson and Ben Laurie]
11741 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11742 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11743 are no longer created. This way we have a single and consistent command
11744 line interface `openssl <command>', similar to `cvs <command>'.
11745 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11747 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11748 BIT STRING wrapper always have zero unused bits.
11751 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11754 *) Make the top-level INSTALL documentation easier to understand.
11757 *) Makefiles updated to exit if an error occurs in a sub-directory
11758 make (including if user presses ^C) [Paul Sutton]
11760 *) Make Montgomery context stuff explicit in RSA data structure.
11763 *) Fix build order of pem and err to allow for generated pem.h.
11766 *) Fix renumbering bug in X509_NAME_delete_entry().
11769 *) Enhanced the err-ins.pl script so it makes the error library number
11770 global and can add a library name. This is needed for external ASN1 and
11771 other error libraries.
11774 *) Fixed sk_insert which never worked properly.
11777 *) Fix ASN1 macros so they can handle indefinite length constructed
11778 EXPLICIT tags. Some non standard certificates use these: they can now
11782 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11783 into a single doc/ssleay.txt bundle. This way the information is still
11784 preserved but no longer messes up this directory. Now it's new room for
11785 the new set of documentation files.
11786 [Ralf S. Engelschall]
11788 *) SETs were incorrectly DER encoded. This was a major pain, because they
11789 shared code with SEQUENCEs, which aren't coded the same. This means that
11790 almost everything to do with SETs or SEQUENCEs has either changed name or
11791 number of arguments.
11792 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11794 *) Fix test data to work with the above.
11797 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11798 was already fixed by Eric for 0.9.1 it seems.
11799 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11801 *) Autodetect FreeBSD3.
11804 *) Fix various bugs in Configure. This affects the following platforms:
11808 unixware-2.0-pentium
11812 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11813 before they are needed.
11816 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11820 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11822 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11823 changed SSLeay to OpenSSL in version strings.
11824 [Ralf S. Engelschall]
11826 *) Some fixups to the top-level documents.
11829 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11830 because the symlink to include/ was missing.
11831 [Ralf S. Engelschall]
11833 *) Incorporated the popular no-RSA/DSA-only patches
11834 which allow to compile a RSA-free SSLeay.
11835 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11837 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11838 when "ssleay" is still not found.
11839 [Ralf S. Engelschall]
11841 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11842 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11844 *) Updated the README file.
11845 [Ralf S. Engelschall]
11847 *) Added various .cvsignore files in the CVS repository subdirs
11848 to make a "cvs update" really silent.
11849 [Ralf S. Engelschall]
11851 *) Recompiled the error-definition header files and added
11852 missing symbols to the Win32 linker tables.
11853 [Ralf S. Engelschall]
11855 *) Cleaned up the top-level documents;
11856 o new files: CHANGES and LICENSE
11857 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11858 o merged COPYRIGHT into LICENSE
11859 o removed obsolete TODO file
11860 o renamed MICROSOFT to INSTALL.W32
11861 [Ralf S. Engelschall]
11863 *) Removed dummy files from the 0.9.1b source tree:
11864 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11865 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11866 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11867 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11868 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11869 [Ralf S. Engelschall]
11871 *) Added various platform portability fixes.
11874 *) The Genesis of the OpenSSL rpject:
11875 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11876 Young and Tim J. Hudson created while they were working for C2Net until
11878 [The OpenSSL Project]
11881 Changes between 0.9.0b and 0.9.1b [not released]
11883 *) Updated a few CA certificates under certs/
11886 *) Changed some BIGNUM api stuff.
11889 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11890 DGUX x86, Linux Alpha, etc.
11893 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11894 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11898 *) Add -strparse option to asn1pars program which parses nested
11900 [Dr Stephen Henson <shenson@bigfoot.com>]
11902 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11905 *) DSA fix for "ca" program.
11908 *) Added "-genkey" option to "dsaparam" program.
11911 *) Added RIPE MD160 (rmd160) message digest.
11914 *) Added -a (all) option to "ssleay version" command.
11917 *) Added PLATFORM define which is the id given to Configure.
11920 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11923 *) Extended the ASN.1 parser routines.
11926 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11929 *) Added a BN_CTX to the BN library.
11932 *) Fixed the weak key values in DES library
11935 *) Changed API in EVP library for cipher aliases.
11938 *) Added support for RC2/64bit cipher.
11941 *) Converted the lhash library to the crypto/mem.c functions.
11944 *) Added more recognized ASN.1 object ids.
11947 *) Added more RSA padding checks for SSL/TLS.
11950 *) Added BIO proxy/filter functionality.
11953 *) Added extra_certs to SSL_CTX which can be used
11954 send extra CA certificates to the client in the CA cert chain sending
11955 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11958 *) Now Fortezza is denied in the authentication phase because
11959 this is key exchange mechanism is not supported by SSLeay at all.
11962 *) Additional PKCS1 checks.
11965 *) Support the string "TLSv1" for all TLS v1 ciphers.
11968 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11969 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11972 *) Fixed a few memory leaks.
11975 *) Fixed various code and comment typos.
11978 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11979 bytes sent in the client random.
11980 [Edward Bishop <ebishop@spyglass.com>]