5 Changes between 1.0.2f and 1.1.0 [xx XXX xxxx]
7 *) Rename REF_CHECK to REF_DEBUG.
10 *) New "unified" build system
12 The "unified" build system is aimed to be a common system for all
13 platforms we support. With it comes new support for VMS.
15 This system builds supports building in a differnt directory tree
16 than the source tree. It produces one Makefile (for unix family
17 or lookalikes), or one descrip.mms (for VMS).
19 The source of information to make the Makefile / descrip.mms is
20 small files called 'build.info', holding the necessary
21 information for each directory with source to compile, and a
22 template in Configurations, like unix-Makefile.tmpl or
25 We rely heavily on the perl module Text::Template.
28 *) Added support for auto-initialisation and de-initialisation of the library.
29 OpenSSL no longer requires explicit init or deinit routines to be called,
30 except in certain circumstances. See the OPENSSL_init_crypto() and
31 OPENSSL_init_ssl() man pages for further information.
34 *) The arguments to the DTLSv1_listen function have changed. Specifically the
35 "peer" argument is now expected to be a BIO_ADDR object.
37 *) Rewrite of BIO networking library. The BIO library lacked consistent
38 support of IPv6, and adding it required some more extensive
39 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
40 which hold all types of addresses and chains of address information.
41 It also introduces a new API, with functions like BIO_socket,
42 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
43 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
44 have been adapted accordingly.
47 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
51 *) CRIME protection: disable compression by default, even if OpenSSL is
52 compiled with zlib enabled. Applications can still enable compression
53 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
54 using the SSL_CONF library to configure compression.
57 *) The signature of the session callback configured with
58 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
59 was explicitly marked as 'const unsigned char*' instead of
63 *) Always DPURIFY. Remove the use of uninitialized memory in the
64 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
67 *) Removed many obsolete configuration items, including
68 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
69 MD2_CHAR, MD2_INT, MD2_LONG
72 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
73 [Rich Salz, with advice from Andy Polyakov]
75 *) Many BN internals have been moved to an internal header file.
76 [Rich Salz with help from Andy Polyakov]
78 *) Configuration and writing out the results from it has changed.
79 Files such as Makefile include/openssl/opensslconf.h and are now
80 produced through general templates, such as Makefile.in and
81 crypto/opensslconf.h.in and some help from the perl module
84 Also, the center of configuration information is no longer
85 Makefile. Instead, Configure produces a perl module in
86 configdata.pm which holds most of the config data (in the hash
87 table %config), the target data that comes from the target
88 configuration in one of the Configurations/*.conf files (in
92 *) To clarify their intended purposes, the Configure options
93 --prefix and --openssldir change their semantics, and become more
94 straightforward and less interdependent.
96 --prefix shall be used exclusively to give the location INSTALLTOP
97 where programs, scripts, libraries, include files and manuals are
98 going to be installed. The default is now /usr/local.
100 --openssldir shall be used exclusively to give the default
101 location OPENSSLDIR where certificates, private keys, CRLs are
102 managed. This is also where the default openssl.cnf gets
104 If the directory given with this option is a relative path, the
105 values of both the --prefix value and the --openssldir value will
106 be combined to become OPENSSLDIR.
107 The default for --openssldir is INSTALLTOP/ssl.
109 Anyone who uses --openssldir to specify where OpenSSL is to be
110 installed MUST change to use --prefix instead.
113 *) The GOST engine was out of date and therefore it has been removed. An up
114 to date GOST engine is now being maintained in an external repository.
115 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
116 support for GOST ciphersuites (these are only activated if a GOST engine
120 *) EGD is no longer supported by default; use enable-egd when
122 [Ben Kaduk and Rich Salz]
124 *) The distribution now has Makefile.in files, which are used to
125 create Makefile's when Configure is run. *Configure must be run
126 before trying to build now.*
129 *) The return value for SSL_CIPHER_description() for error conditions
133 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
135 Obtaining and performing DNSSEC validation of TLSA records is
136 the application's responsibility. The application provides
137 the TLSA records of its choice to OpenSSL, and these are then
138 used to authenticate the peer.
140 The TLSA records need not even come from DNS. They can, for
141 example, be used to implement local end-entity certificate or
142 trust-anchor "pinning", where the "pin" data takes the form
143 of TLSA records, which can augment or replace verification
144 based on the usual WebPKI public certification authorities.
147 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
148 continues to support deprecated interfaces in default builds.
149 However, applications are strongly advised to compile their
150 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
151 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
152 or the 1.1.0 releases.
154 In environments in which all applications have been ported to
155 not use any deprecated interfaces OpenSSL's Configure script
156 should be used with the --api=1.1.0 option to entirely remove
157 support for the deprecated features from the library and
158 unconditionally disable them in the installed headers.
159 Essentially the same effect can be achieved with the "no-deprecated"
160 argument to Configure, except that this will always restrict
161 the build to just the latest API, rather than a fixed API
164 As applications are ported to future revisions of the API,
165 they should update their compile-time OPENSSL_API_COMPAT define
166 accordingly, but in most cases should be able to continue to
167 compile with later releases.
169 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
170 0x10000000L and 0x00908000L, respectively. However those
171 versions did not support the OPENSSL_API_COMPAT feature, and
172 so applications are not typically tested for explicit support
173 of just the undeprecated features of either release.
176 *) Add support for setting the minimum and maximum supported protocol.
177 It can bet set via the SSL_set_min_proto_version() and
178 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
179 MaxProtcol. It's recommended to use the new APIs to disable
180 protocols instead of disabling individual protocols using
181 SSL_set_options() or SSL_CONF's Protocol. This change also
182 removes support for disabling TLS 1.2 in the OpenSSL TLS
183 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
186 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
189 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
190 and integrates ECDSA and ECDH functionality into EC. Implementations can
191 now redirect key generation and no longer need to convert to or from
194 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
195 include the ec.h header file instead.
198 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
199 ciphers who are no longer supported and drops support the ephemeral RSA key
200 exchange. The LOW ciphers currently doesn't have any ciphers in it.
203 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
204 opaque. For HMAC_CTX, the following constructors and destructors
207 HMAC_CTX *HMAC_CTX_new(void);
208 void HMAC_CTX_free(HMAC_CTX *ctx);
210 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
211 destroy such methods has been added. See EVP_MD_meth_new(3) and
212 EVP_CIPHER_meth_new(3) for documentation.
215 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
216 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
217 EVP_MD_CTX_reset() should be called instead to reinitialise
218 an already created structure.
219 2) For consistency with the majority of our object creators and
220 destructors, EVP_MD_CTX_(create|destroy) were renamed to
221 EVP_MD_CTX_(new|free). The old names are retained as macros
222 for deprecated builds.
225 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
226 cryptographic operations to be performed asynchronously as long as an
227 asynchronous capable engine is used. See the ASYNC_start_job() man page for
228 further details. Libssl has also had this capability integrated with the
229 introduction of the new mode SSL_MODE_ASYNC and associated error
230 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
231 pages. This work was developed in partnership with Intel Corp.
234 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
235 always enabled now. If you want to disable the support you should
236 exclude it using the list of supported ciphers. This also means that the
237 "-no_ecdhe" option has been removed from s_server.
240 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
241 SSL_{CTX_}set1_curves() which can set a list.
244 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
245 curve you want to support using SSL_{CTX_}set1_curves().
248 *) State machine rewrite. The state machine code has been significantly
249 refactored in order to remove much duplication of code and solve issues
250 with the old code (see ssl/statem/README for further details). This change
251 does have some associated API changes. Notably the SSL_state() function
252 has been removed and replaced by SSL_get_state which now returns an
253 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
254 altogether. The previous handshake states defined in ssl.h and ssl3.h have
258 *) All instances of the string "ssleay" in the public API were replaced
259 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
260 Some error codes related to internal RSA_eay API's were renamed.
263 *) The demo files in crypto/threads were moved to demo/threads.
266 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
268 [Matt Caswell, Rich Salz]
270 *) New ASN.1 embed macro.
272 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
273 structure is not allocated: it is part of the parent. That is instead of
281 This reduces memory fragmentation and make it impossible to accidentally
282 set a mandatory field to NULL.
284 This currently only works for some fields specifically a SEQUENCE, CHOICE,
285 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
286 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
290 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
293 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
294 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
295 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
296 DES and RC4 ciphersuites.
299 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
300 This changes the decoding behaviour for some invalid messages,
301 though the change is mostly in the more lenient direction, and
302 legacy behaviour is preserved as much as possible.
305 *) Fix no-stdio build.
306 [ David Woodhouse <David.Woodhouse@intel.com> and also
307 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
309 *) New testing framework
310 The testing framework has been largely rewritten and is now using
311 perl and the perl modules Test::Harness and an extended variant of
312 Test::More called OpenSSL::Test to do its work. All test scripts in
313 test/ have been rewritten into test recipes, and all direct calls to
314 executables in test/Makefile have become individual recipes using the
315 simplified testing OpenSSL::Test::Simple.
317 For documentation on our testing modules, do:
319 perldoc test/testlib/OpenSSL/Test/Simple.pm
320 perldoc test/testlib/OpenSSL/Test.pm
324 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
325 are used; the latter aborts on memory leaks (usually checked on exit).
326 Some undocumented "set malloc, etc., hooks" functions were removed
327 and others were changed. All are now documented.
330 *) In DSA_generate_parameters_ex, if the provided seed is too short,
332 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
334 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
335 from RFC4279, RFC4785, RFC5487, RFC5489.
337 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
338 original RSA_PSK patch.
341 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
342 era flag was never set throughout the codebase (only read). Also removed
343 SSL3_FLAGS_POP_BUFFER which was only used if
344 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
347 *) Changed the default name options in the "ca", "crl", "req" and "x509"
348 to be "oneline" instead of "compat".
351 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
352 not aware of clients that still exhibit this bug, and the workaround
353 hasn't been working properly for a while.
356 *) The return type of BIO_number_read() and BIO_number_written() as well as
357 the corresponding num_read and num_write members in the BIO structure has
358 changed from unsigned long to uint64_t. On platforms where an unsigned
359 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
363 *) Given the pervasive nature of TLS extensions it is inadvisable to run
364 OpenSSL without support for them. It also means that maintaining
365 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
366 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
369 *) Removed support for the two export grade static DH ciphersuites
370 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
371 were newly added (along with a number of other static DH ciphersuites) to
372 1.0.2. However the two export ones have *never* worked since they were
373 introduced. It seems strange in any case to be adding new export
374 ciphersuites, and given "logjam" it also does not seem correct to fix them.
377 *) Version negotiation has been rewritten. In particular SSLv23_method(),
378 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
379 and turned into macros which simply call the new preferred function names
380 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
381 should use the new names instead. Also as part of this change the ssl23.h
382 header file has been removed.
385 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
386 code and the associated standard is no longer considered fit-for-purpose.
389 *) RT2547 was closed. When generating a private key, try to make the
390 output file readable only by the owner. This behavior change might
391 be noticeable when interacting with other software.
393 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
397 *) Added HTTP GET support to the ocsp command.
400 *) Changed default digest for the dgst and enc commands from MD5 to
404 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
407 *) Added support for TLS extended master secret from
408 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
409 initial patch which was a great help during development.
412 *) All libssl internal structures have been removed from the public header
413 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
414 now redundant). Users should not attempt to access internal structures
415 directly. Instead they should use the provided API functions.
418 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
419 Access to deprecated functions can be re-enabled by running config with
420 "enable-deprecated". In addition applications wishing to use deprecated
421 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
422 will, by default, disable some transitive includes that previously existed
423 in the header files (e.g. ec.h will no longer, by default, include bn.h)
426 *) Added support for OCB mode. OpenSSL has been granted a patent license
427 compatible with the OpenSSL license for use of OCB. Details are available
428 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
429 for OCB can be removed by calling config with no-ocb.
432 *) SSLv2 support has been removed. It still supports receiving a SSLv2
433 compatible client hello.
436 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
437 done while fixing the error code for the key-too-small case.
438 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
440 *) CA.sh has been removmed; use CA.pl instead.
443 *) Removed old DES API.
446 *) Remove various unsupported platforms:
452 Sinix/ReliantUNIX RM400
457 16-bit platforms such as WIN16
460 *) Clean up OPENSSL_NO_xxx #define's
461 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
462 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
463 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
464 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
465 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
466 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
467 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
468 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
469 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
470 Remove MS_STATIC; it's a relic from platforms <32 bits.
473 *) Cleaned up dead code
474 Remove all but one '#ifdef undef' which is to be looked at.
477 *) Clean up calling of xxx_free routines.
478 Just like free(), fix most of the xxx_free routines to accept
479 NULL. Remove the non-null checks from callers. Save much code.
482 *) Add secure heap for storage of private keys (when possible).
483 Add BIO_s_secmem(), CBIGNUM, etc.
484 Contributed by Akamai Technologies under our Corporate CLA.
487 *) Experimental support for a new, fast, unbiased prime candidate generator,
488 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
489 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
491 *) New output format NSS in the sess_id command line tool. This allows
492 exporting the session id and the master key in NSS keylog format.
493 [Martin Kaiser <martin@kaiser.cx>]
495 *) Harmonize version and its documentation. -f flag is used to display
497 [mancha <mancha1@zoho.com>]
499 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
500 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
501 [mancha <mancha1@zoho.com>]
503 *) Fix some double frees. These are not thought to be exploitable.
504 [mancha <mancha1@zoho.com>]
506 *) A missing bounds check in the handling of the TLS heartbeat extension
507 can be used to reveal up to 64k of memory to a connected client or
510 Thanks for Neel Mehta of Google Security for discovering this bug and to
511 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
512 preparing the fix (CVE-2014-0160)
513 [Adam Langley, Bodo Moeller]
515 *) Fix for the attack described in the paper "Recovering OpenSSL
516 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
517 by Yuval Yarom and Naomi Benger. Details can be obtained from:
518 http://eprint.iacr.org/2014/140
520 Thanks to Yuval Yarom and Naomi Benger for discovering this
521 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
522 [Yuval Yarom and Naomi Benger]
524 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
525 this fixes a limitation in previous versions of OpenSSL.
528 *) Experimental encrypt-then-mac support.
530 Experimental support for encrypt then mac from
531 draft-gutmann-tls-encrypt-then-mac-02.txt
533 To enable it set the appropriate extension number (0x42 for the test
534 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
536 For non-compliant peers (i.e. just about everything) this should have no
539 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
543 *) Add EVP support for key wrapping algorithms, to avoid problems with
544 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
545 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
546 algorithms and include tests cases.
549 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
553 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
554 MGF1 digest and OAEP label.
557 *) Make openssl verify return errors.
558 [Chris Palmer <palmer@google.com> and Ben Laurie]
560 *) New function ASN1_TIME_diff to calculate the difference between two
561 ASN1_TIME structures or one structure and the current time.
564 *) Update fips_test_suite to support multiple command line options. New
565 test to induce all self test errors in sequence and check expected
569 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
570 sign or verify all in one operation.
573 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
574 test programs and fips_test_suite. Includes functionality to parse
575 the minimal script output of fipsalgest.pl directly.
578 *) Add authorisation parameter to FIPS_module_mode_set().
581 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
584 *) Use separate DRBG fields for internal and external flags. New function
585 FIPS_drbg_health_check() to perform on demand health checking. Add
586 generation tests to fips_test_suite with reduced health check interval to
587 demonstrate periodic health checking. Add "nodh" option to
588 fips_test_suite to skip very slow DH test.
591 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
595 *) More extensive health check for DRBG checking many more failure modes.
596 New function FIPS_selftest_drbg_all() to handle every possible DRBG
597 combination: call this in fips_test_suite.
600 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
601 and POST to handle Dual EC cases.
604 *) Add support for canonical generation of DSA parameter 'g'. See
607 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
608 POST to handle HMAC cases.
611 *) Add functions FIPS_module_version() and FIPS_module_version_text()
612 to return numerical and string versions of the FIPS module number.
615 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
616 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
617 outside the validated module in the FIPS capable OpenSSL.
620 *) Minor change to DRBG entropy callback semantics. In some cases
621 there is no multiple of the block length between min_len and
622 max_len. Allow the callback to return more than max_len bytes
623 of entropy but discard any extra: it is the callback's responsibility
624 to ensure that the extra data discarded does not impact the
625 requested amount of entropy.
628 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
629 information in FIPS186-3, SP800-57 and SP800-131A.
632 *) CCM support via EVP. Interface is very similar to GCM case except we
633 must supply all data in one chunk (i.e. no update, final) and the
634 message length must be supplied if AAD is used. Add algorithm test
638 *) Initial version of POST overhaul. Add POST callback to allow the status
639 of POST to be monitored and/or failures induced. Modify fips_test_suite
640 to use callback. Always run all selftests even if one fails.
643 *) XTS support including algorithm test driver in the fips_gcmtest program.
644 Note: this does increase the maximum key length from 32 to 64 bytes but
645 there should be no binary compatibility issues as existing applications
646 will never use XTS mode.
649 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
650 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
651 performs algorithm blocking for unapproved PRNG types. Also do not
652 set PRNG type in FIPS_mode_set(): leave this to the application.
653 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
654 the standard OpenSSL PRNG: set additional data to a date time vector.
657 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
658 This shouldn't present any incompatibility problems because applications
659 shouldn't be using these directly and any that are will need to rethink
660 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
663 *) Extensive self tests and health checking required by SP800-90 DRBG.
664 Remove strength parameter from FIPS_drbg_instantiate and always
665 instantiate at maximum supported strength.
668 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
671 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
674 *) New function DH_compute_key_padded() to compute a DH key and pad with
675 leading zeroes if needed: this complies with SP800-56A et al.
678 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
679 anything, incomplete, subject to change and largely untested at present.
682 *) Modify fipscanisteronly build option to only build the necessary object
683 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
686 *) Add experimental option FIPSSYMS to give all symbols in
687 fipscanister.o and FIPS or fips prefix. This will avoid
688 conflicts with future versions of OpenSSL. Add perl script
689 util/fipsas.pl to preprocess assembly language source files
690 and rename any affected symbols.
693 *) Add selftest checks and algorithm block of non-fips algorithms in
694 FIPS mode. Remove DES2 from selftests.
697 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
698 return internal method without any ENGINE dependencies. Add new
699 tiny fips sign and verify functions.
702 *) New build option no-ec2m to disable characteristic 2 code.
705 *) New build option "fipscanisteronly". This only builds fipscanister.o
706 and (currently) associated fips utilities. Uses the file Makefile.fips
707 instead of Makefile.org as the prototype.
710 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
711 Update fips_gcmtest to use IV generator.
714 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
715 setting output buffer to NULL. The *Final function must be
716 called although it will not retrieve any additional data. The tag
717 can be set or retrieved with a ctrl. The IV length is by default 12
718 bytes (96 bits) but can be set to an alternative value. If the IV
719 length exceeds the maximum IV length (currently 16 bytes) it cannot be
723 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
724 underlying do_cipher function handles all cipher semantics itself
725 including padding and finalisation. This is useful if (for example)
726 an ENGINE cipher handles block padding itself. The behaviour of
727 do_cipher is subtly changed if this flag is set: the return value
728 is the number of characters written to the output buffer (zero is
729 no longer an error code) or a negative error code. Also if the
730 input buffer is NULL and length 0 finalisation should be performed.
733 *) If a candidate issuer certificate is already part of the constructed
734 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
737 *) Improve forward-security support: add functions
739 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
740 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
742 for use by SSL/TLS servers; the callback function will be called whenever a
743 new session is created, and gets to decide whether the session may be
744 cached to make it resumable (return 0) or not (return 1). (As by the
745 SSL/TLS protocol specifications, the session_id sent by the server will be
746 empty to indicate that the session is not resumable; also, the server will
747 not generate RFC 4507 (RFC 5077) session tickets.)
749 A simple reasonable callback implementation is to return is_forward_secure.
750 This parameter will be set to 1 or 0 depending on the ciphersuite selected
751 by the SSL/TLS server library, indicating whether it can provide forward
753 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
755 *) New -verify_name option in command line utilities to set verification
759 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
760 Add CMAC pkey methods.
763 *) Experimental renegotiation in s_server -www mode. If the client
764 browses /reneg connection is renegotiated. If /renegcert it is
765 renegotiated requesting a certificate.
768 *) Add an "external" session cache for debugging purposes to s_server. This
769 should help trace issues which normally are only apparent in deployed
770 multi-process servers.
773 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
774 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
775 BIO_set_cipher() and some obscure PEM functions were changed so they
776 can now return an error. The RAND changes required a change to the
777 RAND_METHOD structure.
780 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
781 a gcc attribute to warn if the result of a function is ignored. This
782 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
783 whose return value is often ignored.
786 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
787 *) DH small subgroups
789 Historically OpenSSL only ever generated DH parameters based on "safe"
790 primes. More recently (in version 1.0.2) support was provided for
791 generating X9.42 style parameter files such as those required for RFC 5114
792 support. The primes used in such files may not be "safe". Where an
793 application is using DH configured with parameters based on primes that are
794 not "safe" then an attacker could use this fact to find a peer's private
795 DH exponent. This attack requires that the attacker complete multiple
796 handshakes in which the peer uses the same private DH exponent. For example
797 this could be used to discover a TLS server's private DH exponent if it's
798 reusing the private DH exponent or it's using a static DH ciphersuite.
800 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
801 TLS. It is not on by default. If the option is not set then the server
802 reuses the same private DH exponent for the life of the server process and
803 would be vulnerable to this attack. It is believed that many popular
804 applications do set this option and would therefore not be at risk.
806 The fix for this issue adds an additional check where a "q" parameter is
807 available (as is the case in X9.42 based parameters). This detects the
808 only known attack, and is the only possible defense for static DH
809 ciphersuites. This could have some performance impact.
811 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
812 default and cannot be disabled. This could have some performance impact.
814 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
818 *) SSLv2 doesn't block disabled ciphers
820 A malicious client can negotiate SSLv2 ciphers that have been disabled on
821 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
822 been disabled, provided that the SSLv2 protocol was not also disabled via
825 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
826 and Sebastian Schinzel.
830 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
832 *) BN_mod_exp may produce incorrect results on x86_64
834 There is a carry propagating bug in the x86_64 Montgomery squaring
835 procedure. No EC algorithms are affected. Analysis suggests that attacks
836 against RSA and DSA as a result of this defect would be very difficult to
837 perform and are not believed likely. Attacks against DH are considered just
838 feasible (although very difficult) because most of the work necessary to
839 deduce information about a private key may be performed offline. The amount
840 of resources required for such an attack would be very significant and
841 likely only accessible to a limited number of attackers. An attacker would
842 additionally need online access to an unpatched system using the target
843 private key in a scenario with persistent DH parameters and a private
844 key that is shared between multiple clients. For example this can occur by
845 default in OpenSSL DHE based SSL/TLS ciphersuites.
847 This issue was reported to OpenSSL by Hanno Böck.
851 *) Certificate verify crash with missing PSS parameter
853 The signature verification routines will crash with a NULL pointer
854 dereference if presented with an ASN.1 signature using the RSA PSS
855 algorithm and absent mask generation function parameter. Since these
856 routines are used to verify certificate signature algorithms this can be
857 used to crash any certificate verification operation and exploited in a
858 DoS attack. Any application which performs certificate verification is
859 vulnerable including OpenSSL clients and servers which enable client
862 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
866 *) X509_ATTRIBUTE memory leak
868 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
869 memory. This structure is used by the PKCS#7 and CMS routines so any
870 application which reads PKCS#7 or CMS data from untrusted sources is
871 affected. SSL/TLS is not affected.
873 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
878 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
879 This changes the decoding behaviour for some invalid messages,
880 though the change is mostly in the more lenient direction, and
881 legacy behaviour is preserved as much as possible.
884 *) In DSA_generate_parameters_ex, if the provided seed is too short,
886 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
888 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
890 *) Alternate chains certificate forgery
892 During certificate verfification, OpenSSL will attempt to find an
893 alternative certificate chain if the first attempt to build such a chain
894 fails. An error in the implementation of this logic can mean that an
895 attacker could cause certain checks on untrusted certificates to be
896 bypassed, such as the CA flag, enabling them to use a valid leaf
897 certificate to act as a CA and "issue" an invalid certificate.
899 This issue was reported to OpenSSL by Adam Langley/David Benjamin
903 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
905 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
906 incompatibility in the handling of HMAC. The previous ABI has now been
910 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
912 *) Malformed ECParameters causes infinite loop
914 When processing an ECParameters structure OpenSSL enters an infinite loop
915 if the curve specified is over a specially malformed binary polynomial
918 This can be used to perform denial of service against any
919 system which processes public keys, certificate requests or
920 certificates. This includes TLS clients and TLS servers with
921 client authentication enabled.
923 This issue was reported to OpenSSL by Joseph Barr-Pixton.
927 *) Exploitable out-of-bounds read in X509_cmp_time
929 X509_cmp_time does not properly check the length of the ASN1_TIME
930 string and can read a few bytes out of bounds. In addition,
931 X509_cmp_time accepts an arbitrary number of fractional seconds in the
934 An attacker can use this to craft malformed certificates and CRLs of
935 various sizes and potentially cause a segmentation fault, resulting in
936 a DoS on applications that verify certificates or CRLs. TLS clients
937 that verify CRLs are affected. TLS clients and servers with client
938 authentication enabled may be affected if they use custom verification
941 This issue was reported to OpenSSL by Robert Swiecki (Google), and
942 independently by Hanno Böck.
946 *) PKCS7 crash with missing EnvelopedContent
948 The PKCS#7 parsing code does not handle missing inner EncryptedContent
949 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
950 with missing content and trigger a NULL pointer dereference on parsing.
952 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
953 structures from untrusted sources are affected. OpenSSL clients and
954 servers are not affected.
956 This issue was reported to OpenSSL by Michal Zalewski (Google).
960 *) CMS verify infinite loop with unknown hash function
962 When verifying a signedData message the CMS code can enter an infinite loop
963 if presented with an unknown hash function OID. This can be used to perform
964 denial of service against any system which verifies signedData messages using
966 This issue was reported to OpenSSL by Johannes Bauer.
970 *) Race condition handling NewSessionTicket
972 If a NewSessionTicket is received by a multi-threaded client when attempting to
973 reuse a previous ticket then a race condition can occur potentially leading to
974 a double free of the ticket data.
978 *) Only support 256-bit or stronger elliptic curves with the
979 'ecdh_auto' setting (server) or by default (client). Of supported
980 curves, prefer P-256 (both).
983 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
985 *) ClientHello sigalgs DoS fix
987 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
988 invalid signature algorithms extension a NULL pointer dereference will
989 occur. This can be exploited in a DoS attack against the server.
991 This issue was was reported to OpenSSL by David Ramos of Stanford
994 [Stephen Henson and Matt Caswell]
996 *) Multiblock corrupted pointer fix
998 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
999 feature only applies on 64 bit x86 architecture platforms that support AES
1000 NI instructions. A defect in the implementation of "multiblock" can cause
1001 OpenSSL's internal write buffer to become incorrectly set to NULL when
1002 using non-blocking IO. Typically, when the user application is using a
1003 socket BIO for writing, this will only result in a failed connection.
1004 However if some other BIO is used then it is likely that a segmentation
1005 fault will be triggered, thus enabling a potential DoS attack.
1007 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1011 *) Segmentation fault in DTLSv1_listen fix
1013 The DTLSv1_listen function is intended to be stateless and processes the
1014 initial ClientHello from many peers. It is common for user code to loop
1015 over the call to DTLSv1_listen until a valid ClientHello is received with
1016 an associated cookie. A defect in the implementation of DTLSv1_listen means
1017 that state is preserved in the SSL object from one invocation to the next
1018 that can lead to a segmentation fault. Errors processing the initial
1019 ClientHello can trigger this scenario. An example of such an error could be
1020 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1023 This issue was reported to OpenSSL by Per Allansson.
1027 *) Segmentation fault in ASN1_TYPE_cmp fix
1029 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1030 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1031 certificate signature algorithm consistency this can be used to crash any
1032 certificate verification operation and exploited in a DoS attack. Any
1033 application which performs certificate verification is vulnerable including
1034 OpenSSL clients and servers which enable client authentication.
1038 *) Segmentation fault for invalid PSS parameters fix
1040 The signature verification routines will crash with a NULL pointer
1041 dereference if presented with an ASN.1 signature using the RSA PSS
1042 algorithm and invalid parameters. Since these routines are used to verify
1043 certificate signature algorithms this can be used to crash any
1044 certificate verification operation and exploited in a DoS attack. Any
1045 application which performs certificate verification is vulnerable including
1046 OpenSSL clients and servers which enable client authentication.
1048 This issue was was reported to OpenSSL by Brian Carpenter.
1052 *) ASN.1 structure reuse memory corruption fix
1054 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1055 memory corruption via an invalid write. Such reuse is and has been
1056 strongly discouraged and is believed to be rare.
1058 Applications that parse structures containing CHOICE or ANY DEFINED BY
1059 components may be affected. Certificate parsing (d2i_X509 and related
1060 functions) are however not affected. OpenSSL clients and servers are
1065 *) PKCS7 NULL pointer dereferences fix
1067 The PKCS#7 parsing code does not handle missing outer ContentInfo
1068 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1069 missing content and trigger a NULL pointer dereference on parsing.
1071 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1072 otherwise parse PKCS#7 structures from untrusted sources are
1073 affected. OpenSSL clients and servers are not affected.
1075 This issue was reported to OpenSSL by Michal Zalewski (Google).
1079 *) DoS via reachable assert in SSLv2 servers fix
1081 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1082 servers that both support SSLv2 and enable export cipher suites by sending
1083 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1085 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1086 (OpenSSL development team).
1090 *) Empty CKE with client auth and DHE fix
1092 If client auth is used then a server can seg fault in the event of a DHE
1093 ciphersuite being selected and a zero length ClientKeyExchange message
1094 being sent by the client. This could be exploited in a DoS attack.
1098 *) Handshake with unseeded PRNG fix
1100 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1101 with an unseeded PRNG. The conditions are:
1102 - The client is on a platform where the PRNG has not been seeded
1103 automatically, and the user has not seeded manually
1104 - A protocol specific client method version has been used (i.e. not
1105 SSL_client_methodv23)
1106 - A ciphersuite is used that does not require additional random data from
1107 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1109 If the handshake succeeds then the client random that has been used will
1110 have been generated from a PRNG with insufficient entropy and therefore the
1111 output may be predictable.
1113 For example using the following command with an unseeded openssl will
1114 succeed on an unpatched platform:
1116 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1120 *) Use After Free following d2i_ECPrivatekey error fix
1122 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1123 could cause a use after free condition. This, in turn, could cause a double
1124 free in several private key parsing functions (such as d2i_PrivateKey
1125 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1126 for applications that receive EC private keys from untrusted
1127 sources. This scenario is considered rare.
1129 This issue was discovered by the BoringSSL project and fixed in their
1134 *) X509_to_X509_REQ NULL pointer deref fix
1136 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1137 the certificate key is invalid. This function is rarely used in practice.
1139 This issue was discovered by Brian Carpenter.
1143 *) Removed the export ciphers from the DEFAULT ciphers
1146 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1148 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1149 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1150 So far those who have to target multiple plaforms would compromise
1151 and argue that binary targeting say ARMv5 would still execute on
1152 ARMv8. "Universal" build resolves this compromise by providing
1153 near-optimal performance even on newer platforms.
1156 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1157 (other platforms pending).
1158 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1160 *) Add support for the SignedCertificateTimestampList certificate and
1161 OCSP response extensions from RFC6962.
1164 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1165 for corner cases. (Certain input points at infinity could lead to
1166 bogus results, with non-infinity inputs mapped to infinity too.)
1169 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1170 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1171 common cases are optimized and there still is room for further
1172 improvements. Vector Permutation AES for Altivec is also added.
1175 *) Add support for little-endian ppc64 Linux target.
1176 [Marcelo Cerri (IBM)]
1178 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1179 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1180 are optimized and there still is room for further improvements.
1181 Both 32- and 64-bit modes are supported.
1182 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1184 *) Improved ARMv7 NEON support.
1187 *) Support for SPARC Architecture 2011 crypto extensions, first
1188 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1189 SHA256/512, MD5, GHASH and modular exponentiation.
1190 [Andy Polyakov, David Miller]
1192 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1194 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1196 *) Support for new and upcoming Intel processors, including AVX2,
1197 BMI and SHA ISA extensions. This includes additional "stitched"
1198 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1201 This work was sponsored by Intel Corp.
1204 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1205 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1206 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1209 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1210 this fixes a limiation in previous versions of OpenSSL.
1213 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1214 MGF1 digest and OAEP label.
1217 *) Add EVP support for key wrapping algorithms, to avoid problems with
1218 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1219 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1220 algorithms and include tests cases.
1223 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1225 [Douglas E. Engert, Steve Henson]
1227 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1228 difference in days and seconds between two tm or ASN1_TIME structures.
1231 *) Add -rev test option to s_server to just reverse order of characters
1232 received by client and send back to server. Also prints an abbreviated
1233 summary of the connection parameters.
1236 *) New option -brief for s_client and s_server to print out a brief summary
1237 of connection parameters.
1240 *) Add callbacks for arbitrary TLS extensions.
1241 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1243 *) New option -crl_download in several openssl utilities to download CRLs
1244 from CRLDP extension in certificates.
1247 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1250 *) New function X509_CRL_diff to generate a delta CRL from the difference
1251 of two full CRLs. Add support to "crl" utility.
1254 *) New functions to set lookup_crls function and to retrieve
1255 X509_STORE from X509_STORE_CTX.
1258 *) Print out deprecated issuer and subject unique ID fields in
1262 *) Extend OCSP I/O functions so they can be used for simple general purpose
1263 HTTP as well as OCSP. New wrapper function which can be used to download
1264 CRLs using the OCSP API.
1267 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1270 *) SSL_CONF* functions. These provide a common framework for application
1271 configuration using configuration files or command lines.
1274 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1275 message callback and prints the results. Needs compile time option
1276 "enable-ssl-trace". New options to s_client and s_server to enable
1280 *) New ctrl and macro to retrieve supported points extensions.
1281 Print out extension in s_server and s_client.
1284 *) New functions to retrieve certificate signature and signature
1288 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1292 *) New Suite B modes for TLS code. These use and enforce the requirements
1293 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1294 only use Suite B curves. The Suite B modes can be set by using the
1295 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1298 *) New chain verification flags for Suite B levels of security. Check
1299 algorithms are acceptable when flags are set in X509_verify_cert.
1302 *) Make tls1_check_chain return a set of flags indicating checks passed
1303 by a certificate chain. Add additional tests to handle client
1304 certificates: checks for matching certificate type and issuer name
1308 *) If an attempt is made to use a signature algorithm not in the peer
1309 preference list abort the handshake. If client has no suitable
1310 signature algorithms in response to a certificate request do not
1311 use the certificate.
1314 *) If server EC tmp key is not in client preference list abort handshake.
1317 *) Add support for certificate stores in CERT structure. This makes it
1318 possible to have different stores per SSL structure or one store in
1319 the parent SSL_CTX. Include distint stores for certificate chain
1320 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1321 to build and store a certificate chain in CERT structure: returing
1322 an error if the chain cannot be built: this will allow applications
1323 to test if a chain is correctly configured.
1325 Note: if the CERT based stores are not set then the parent SSL_CTX
1326 store is used to retain compatibility with existing behaviour.
1330 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1331 mask based on the current session, check mask when sending client
1332 hello and checking the requested ciphersuite.
1335 *) New ctrls to retrieve and set certificate types in a certificate
1336 request message. Print out received values in s_client. If certificate
1337 types is not set with custom values set sensible values based on
1338 supported signature algorithms.
1341 *) Support for distinct client and server supported signature algorithms.
1344 *) Add certificate callback. If set this is called whenever a certificate
1345 is required by client or server. An application can decide which
1346 certificate chain to present based on arbitrary criteria: for example
1347 supported signature algorithms. Add very simple example to s_server.
1348 This fixes many of the problems and restrictions of the existing client
1349 certificate callback: for example you can now clear an existing
1350 certificate and specify the whole chain.
1353 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1354 the certificate can be used for (if anything). Set valid_flags field
1355 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1356 to have similar checks in it.
1358 Add new "cert_flags" field to CERT structure and include a "strict mode".
1359 This enforces some TLS certificate requirements (such as only permitting
1360 certificate signature algorithms contained in the supported algorithms
1361 extension) which some implementations ignore: this option should be used
1362 with caution as it could cause interoperability issues.
1365 *) Update and tidy signature algorithm extension processing. Work out
1366 shared signature algorithms based on preferences and peer algorithms
1367 and print them out in s_client and s_server. Abort handshake if no
1368 shared signature algorithms.
1371 *) Add new functions to allow customised supported signature algorithms
1372 for SSL and SSL_CTX structures. Add options to s_client and s_server
1376 *) New function SSL_certs_clear() to delete all references to certificates
1377 from an SSL structure. Before this once a certificate had been added
1378 it couldn't be removed.
1381 *) Integrate hostname, email address and IP address checking with certificate
1382 verification. New verify options supporting checking in opensl utility.
1385 *) Fixes and wildcard matching support to hostname and email checking
1386 functions. Add manual page.
1387 [Florian Weimer (Red Hat Product Security Team)]
1389 *) New functions to check a hostname email or IP address against a
1390 certificate. Add options x509 utility to print results of checks against
1394 *) Fix OCSP checking.
1395 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1397 *) Initial experimental support for explicitly trusted non-root CAs.
1398 OpenSSL still tries to build a complete chain to a root but if an
1399 intermediate CA has a trust setting included that is used. The first
1400 setting is used: whether to trust (e.g., -addtrust option to the x509
1404 *) Add -trusted_first option which attempts to find certificates in the
1405 trusted store even if an untrusted chain is also supplied.
1408 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1409 platform support for Linux and Android.
1412 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1415 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1416 When in FIPS mode the approved implementations are used as normal,
1417 when not in FIPS mode the internal unapproved versions are used instead.
1418 This means that the FIPS capable OpenSSL isn't forced to use the
1419 (often lower performance) FIPS implementations outside FIPS mode.
1422 *) Transparently support X9.42 DH parameters when calling
1423 PEM_read_bio_DHparameters. This means existing applications can handle
1424 the new parameter format automatically.
1427 *) Initial experimental support for X9.42 DH parameter format: mainly
1428 to support use of 'q' parameter for RFC5114 parameters.
1431 *) Add DH parameters from RFC5114 including test data to dhtest.
1434 *) Support for automatic EC temporary key parameter selection. If enabled
1435 the most preferred EC parameters are automatically used instead of
1436 hardcoded fixed parameters. Now a server just has to call:
1437 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1438 support ECDH and use the most appropriate parameters.
1441 *) Enhance and tidy EC curve and point format TLS extension code. Use
1442 static structures instead of allocation if default values are used.
1443 New ctrls to set curves we wish to support and to retrieve shared curves.
1444 Print out shared curves in s_server. New options to s_server and s_client
1445 to set list of supported curves.
1448 *) New ctrls to retrieve supported signature algorithms and
1449 supported curve values as an array of NIDs. Extend openssl utility
1450 to print out received values.
1453 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1454 between NIDs and the more common NIST names such as "P-256". Enhance
1455 ecparam utility and ECC method to recognise the NIST names for curves.
1458 *) Enhance SSL/TLS certificate chain handling to support different
1459 chains for each certificate instead of one chain in the parent SSL_CTX.
1462 *) Support for fixed DH ciphersuite client authentication: where both
1463 server and client use DH certificates with common parameters.
1466 *) Support for fixed DH ciphersuites: those requiring DH server
1470 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1472 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1473 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1474 X509_CINF_get_signature were reverted post internal team review.
1476 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1478 *) Build fixes for the Windows and OpenVMS platforms
1479 [Matt Caswell and Richard Levitte]
1481 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1483 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1484 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1485 dereference. This could lead to a Denial Of Service attack. Thanks to
1486 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1490 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1491 dtls1_buffer_record function under certain conditions. In particular this
1492 could occur if an attacker sent repeated DTLS records with the same
1493 sequence number but for the next epoch. The memory leak could be exploited
1494 by an attacker in a Denial of Service attack through memory exhaustion.
1495 Thanks to Chris Mueller for reporting this issue.
1499 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1500 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1501 method would be set to NULL which could later result in a NULL pointer
1502 dereference. Thanks to Frank Schmirler for reporting this issue.
1506 *) Abort handshake if server key exchange message is omitted for ephemeral
1509 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1510 reporting this issue.
1514 *) Remove non-export ephemeral RSA code on client and server. This code
1515 violated the TLS standard by allowing the use of temporary RSA keys in
1516 non-export ciphersuites and could be used by a server to effectively
1517 downgrade the RSA key length used to a value smaller than the server
1518 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1519 INRIA or reporting this issue.
1523 *) Fixed issue where DH client certificates are accepted without verification.
1524 An OpenSSL server will accept a DH certificate for client authentication
1525 without the certificate verify message. This effectively allows a client to
1526 authenticate without the use of a private key. This only affects servers
1527 which trust a client certificate authority which issues certificates
1528 containing DH keys: these are extremely rare and hardly ever encountered.
1529 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1534 *) Ensure that the session ID context of an SSL is updated when its
1535 SSL_CTX is updated via SSL_set_SSL_CTX.
1537 The session ID context is typically set from the parent SSL_CTX,
1538 and can vary with the CTX.
1541 *) Fix various certificate fingerprint issues.
1543 By using non-DER or invalid encodings outside the signed portion of a
1544 certificate the fingerprint can be changed without breaking the signature.
1545 Although no details of the signed portion of the certificate can be changed
1546 this can cause problems with some applications: e.g. those using the
1547 certificate fingerprint for blacklists.
1549 1. Reject signatures with non zero unused bits.
1551 If the BIT STRING containing the signature has non zero unused bits reject
1552 the signature. All current signature algorithms require zero unused bits.
1554 2. Check certificate algorithm consistency.
1556 Check the AlgorithmIdentifier inside TBS matches the one in the
1557 certificate signature. NB: this will result in signature failure
1558 errors for some broken certificates.
1560 Thanks to Konrad Kraszewski from Google for reporting this issue.
1562 3. Check DSA/ECDSA signatures use DER.
1564 Reencode DSA/ECDSA signatures and compare with the original received
1565 signature. Return an error if there is a mismatch.
1567 This will reject various cases including garbage after signature
1568 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1569 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1570 (negative or with leading zeroes).
1572 Further analysis was conducted and fixes were developed by Stephen Henson
1573 of the OpenSSL core team.
1578 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1579 results on some platforms, including x86_64. This bug occurs at random
1580 with a very low probability, and is not known to be exploitable in any
1581 way, though its exact impact is difficult to determine. Thanks to Pieter
1582 Wuille (Blockstream) who reported this issue and also suggested an initial
1583 fix. Further analysis was conducted by the OpenSSL development team and
1584 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1585 the OpenSSL core team.
1589 *) Do not resume sessions on the server if the negotiated protocol
1590 version does not match the session's version. Resuming with a different
1591 version, while not strictly forbidden by the RFC, is of questionable
1592 sanity and breaks all known clients.
1593 [David Benjamin, Emilia Käsper]
1595 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1596 early CCS messages during renegotiation. (Note that because
1597 renegotiation is encrypted, this early CCS was not exploitable.)
1600 *) Tighten client-side session ticket handling during renegotiation:
1601 ensure that the client only accepts a session ticket if the server sends
1602 the extension anew in the ServerHello. Previously, a TLS client would
1603 reuse the old extension state and thus accept a session ticket if one was
1604 announced in the initial ServerHello.
1606 Similarly, ensure that the client requires a session ticket if one
1607 was advertised in the ServerHello. Previously, a TLS client would
1608 ignore a missing NewSessionTicket message.
1611 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1613 *) SRTP Memory Leak.
1615 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1616 sends a carefully crafted handshake message, to cause OpenSSL to fail
1617 to free up to 64k of memory causing a memory leak. This could be
1618 exploited in a Denial Of Service attack. This issue affects OpenSSL
1619 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1620 whether SRTP is used or configured. Implementations of OpenSSL that
1621 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1623 The fix was developed by the OpenSSL team.
1627 *) Session Ticket Memory Leak.
1629 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1630 integrity of that ticket is first verified. In the event of a session
1631 ticket integrity check failing, OpenSSL will fail to free memory
1632 causing a memory leak. By sending a large number of invalid session
1633 tickets an attacker could exploit this issue in a Denial Of Service
1638 *) Build option no-ssl3 is incomplete.
1640 When OpenSSL is configured with "no-ssl3" as a build option, servers
1641 could accept and complete a SSL 3.0 handshake, and clients could be
1642 configured to send them.
1644 [Akamai and the OpenSSL team]
1646 *) Add support for TLS_FALLBACK_SCSV.
1647 Client applications doing fallback retries should call
1648 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1650 [Adam Langley, Bodo Moeller]
1652 *) Add additional DigestInfo checks.
1654 Reencode DigestInto in DER and check against the original when
1655 verifying RSA signature: this will reject any improperly encoded
1656 DigestInfo structures.
1658 Note: this is a precautionary measure and no attacks are currently known.
1662 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1664 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1665 SRP code can be overrun an internal buffer. Add sanity check that
1666 g, A, B < N to SRP code.
1668 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1669 Group for discovering this issue.
1673 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1674 TLS 1.0 instead of higher protocol versions when the ClientHello message
1675 is badly fragmented. This allows a man-in-the-middle attacker to force a
1676 downgrade to TLS 1.0 even if both the server and the client support a
1677 higher protocol version, by modifying the client's TLS records.
1679 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1680 researching this issue.
1684 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1685 to a denial of service attack. A malicious server can crash the client
1686 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1687 ciphersuite and sending carefully crafted handshake messages.
1689 Thanks to Felix Gröbert (Google) for discovering and researching this
1694 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1695 to leak memory. This can be exploited through a Denial of Service attack.
1696 Thanks to Adam Langley for discovering and researching this issue.
1700 *) An attacker can force openssl to consume large amounts of memory whilst
1701 processing DTLS handshake messages. This can be exploited through a
1702 Denial of Service attack.
1703 Thanks to Adam Langley for discovering and researching this issue.
1707 *) An attacker can force an error condition which causes openssl to crash
1708 whilst processing DTLS packets due to memory being freed twice. This
1709 can be exploited through a Denial of Service attack.
1710 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1715 *) If a multithreaded client connects to a malicious server using a resumed
1716 session and the server sends an ec point format extension it could write
1717 up to 255 bytes to freed memory.
1719 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1724 *) A malicious server can crash an OpenSSL client with a null pointer
1725 dereference (read) by specifying an SRP ciphersuite even though it was not
1726 properly negotiated with the client. This can be exploited through a
1727 Denial of Service attack.
1729 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1730 discovering and researching this issue.
1734 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1735 X509_name_oneline, X509_name_print_ex et al. to leak some information
1736 from the stack. Applications may be affected if they echo pretty printing
1737 output to the attacker.
1739 Thanks to Ivan Fratric (Google) for discovering this issue.
1741 [Emilia Käsper, and Steve Henson]
1743 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1744 for corner cases. (Certain input points at infinity could lead to
1745 bogus results, with non-infinity inputs mapped to infinity too.)
1748 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1750 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1751 handshake can force the use of weak keying material in OpenSSL
1752 SSL/TLS clients and servers.
1754 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1755 researching this issue. (CVE-2014-0224)
1756 [KIKUCHI Masashi, Steve Henson]
1758 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1759 OpenSSL DTLS client the code can be made to recurse eventually crashing
1762 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1764 [Imre Rad, Steve Henson]
1766 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1767 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1768 client or server. This is potentially exploitable to run arbitrary
1769 code on a vulnerable client or server.
1771 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1772 [Jüri Aedla, Steve Henson]
1774 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1775 are subject to a denial of service attack.
1777 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1778 this issue. (CVE-2014-3470)
1779 [Felix Gröbert, Ivan Fratric, Steve Henson]
1781 *) Harmonize version and its documentation. -f flag is used to display
1783 [mancha <mancha1@zoho.com>]
1785 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1786 in i2d_ECPrivateKey.
1787 [mancha <mancha1@zoho.com>]
1789 *) Fix some double frees. These are not thought to be exploitable.
1790 [mancha <mancha1@zoho.com>]
1792 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1794 *) A missing bounds check in the handling of the TLS heartbeat extension
1795 can be used to reveal up to 64k of memory to a connected client or
1798 Thanks for Neel Mehta of Google Security for discovering this bug and to
1799 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1800 preparing the fix (CVE-2014-0160)
1801 [Adam Langley, Bodo Moeller]
1803 *) Fix for the attack described in the paper "Recovering OpenSSL
1804 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1805 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1806 http://eprint.iacr.org/2014/140
1808 Thanks to Yuval Yarom and Naomi Benger for discovering this
1809 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1810 [Yuval Yarom and Naomi Benger]
1812 *) TLS pad extension: draft-agl-tls-padding-03
1814 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1815 TLS client Hello record length value would otherwise be > 255 and
1816 less that 512 pad with a dummy extension containing zeroes so it
1817 is at least 512 bytes long.
1819 [Adam Langley, Steve Henson]
1821 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1823 *) Fix for TLS record tampering bug. A carefully crafted invalid
1824 handshake could crash OpenSSL with a NULL pointer exception.
1825 Thanks to Anton Johansson for reporting this issues.
1828 *) Keep original DTLS digest and encryption contexts in retransmission
1829 structures so we can use the previous session parameters if they need
1830 to be resent. (CVE-2013-6450)
1833 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1834 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1835 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1836 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1837 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1838 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1839 [Rob Stradling, Adam Langley]
1841 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1843 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1844 supporting platforms or when small records were transferred.
1845 [Andy Polyakov, Steve Henson]
1847 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1849 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1851 This addresses the flaw in CBC record processing discovered by
1852 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1853 at: http://www.isg.rhul.ac.uk/tls/
1855 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1856 Security Group at Royal Holloway, University of London
1857 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1858 Emilia Käsper for the initial patch.
1860 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1862 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1863 ciphersuites which can be exploited in a denial of service attack.
1864 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1865 and detecting this bug and to Wolfgang Ettlinger
1866 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1870 *) Return an error when checking OCSP signatures when key is NULL.
1871 This fixes a DoS attack. (CVE-2013-0166)
1874 *) Make openssl verify return errors.
1875 [Chris Palmer <palmer@google.com> and Ben Laurie]
1877 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1878 the right response is stapled. Also change SSL_get_certificate()
1879 so it returns the certificate actually sent.
1880 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1881 [Rob Stradling <rob.stradling@comodo.com>]
1883 *) Fix possible deadlock when decoding public keys.
1886 *) Don't use TLS 1.0 record version number in initial client hello
1890 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1892 *) Sanity check record length before skipping explicit IV in TLS
1893 1.2, 1.1 and DTLS to fix DoS attack.
1895 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1896 fuzzing as a service testing platform.
1900 *) Initialise tkeylen properly when encrypting CMS messages.
1901 Thanks to Solar Designer of Openwall for reporting this issue.
1904 *) In FIPS mode don't try to use composite ciphers as they are not
1908 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1910 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1911 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1912 mean any application compiled against OpenSSL 1.0.0 headers setting
1913 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1914 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1915 0x10000000L Any application which was previously compiled against
1916 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1917 will need to be recompiled as a result. Letting be results in
1918 inability to disable specifically TLS 1.1 and in client context,
1919 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1922 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1923 disable just protocol X, but all protocols above X *if* there are
1924 protocols *below* X still enabled. In more practical terms it means
1925 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1926 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1927 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1931 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1933 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1934 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1935 in CRYPTO_realloc_clean.
1937 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1938 issue and to Adam Langley <agl@chromium.org> for fixing it.
1940 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1942 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1945 *) Workarounds for some broken servers that "hang" if a client hello
1946 record length exceeds 255 bytes.
1948 1. Do not use record version number > TLS 1.0 in initial client
1949 hello: some (but not all) hanging servers will now work.
1950 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1951 the number of ciphers sent in the client hello. This should be
1952 set to an even number, such as 50, for example by passing:
1953 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1954 Most broken servers should now work.
1955 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1956 TLS 1.2 client support entirely.
1959 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1962 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1964 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1965 STRING form instead of a DigestInfo.
1968 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1969 and the RSA_sign/RSA_verify functions. This was made more apparent when
1970 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1971 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1972 the correct format in RSA_verify so both forms transparently work.
1975 *) Some servers which support TLS 1.0 can choke if we initially indicate
1976 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1977 encrypted premaster secret. As a workaround use the maximum permitted
1978 client version in client hello, this should keep such servers happy
1979 and still work with previous versions of OpenSSL.
1982 *) Add support for TLS/DTLS heartbeats.
1983 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1985 *) Add support for SCTP.
1986 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1988 *) Improved PRNG seeding for VOS.
1989 [Paul Green <Paul.Green@stratus.com>]
1991 *) Extensive assembler packs updates, most notably:
1993 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1994 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1995 - x86_64: bit-sliced AES implementation;
1996 - ARM: NEON support, contemporary platforms optimizations;
1997 - s390x: z196 support;
1998 - *: GHASH and GF(2^m) multiplication implementations;
2002 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2003 (removal of unnecessary code)
2004 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2006 *) Add TLS key material exporter from RFC 5705.
2009 *) Add DTLS-SRTP negotiation from RFC 5764.
2012 *) Add Next Protocol Negotiation,
2013 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2014 disabled with a no-npn flag to config or Configure. Code donated
2016 [Adam Langley <agl@google.com> and Ben Laurie]
2018 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2019 NIST-P256, NIST-P521, with constant-time single point multiplication on
2020 typical inputs. Compiler support for the nonstandard type __uint128_t is
2021 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2022 Code made available under Apache License version 2.0.
2024 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2025 line to include this in your build of OpenSSL, and run "make depend" (or
2026 "make update"). This enables the following EC_METHODs:
2028 EC_GFp_nistp224_method()
2029 EC_GFp_nistp256_method()
2030 EC_GFp_nistp521_method()
2032 EC_GROUP_new_by_curve_name() will automatically use these (while
2033 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2035 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2037 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2038 all platforms. Move ssize_t definition from e_os.h to the public
2039 header file e_os2.h as it now appears in public header file cms.h
2042 *) New -sigopt option to the ca, req and x509 utilities. Additional
2043 signature parameters can be passed using this option and in
2047 *) Add RSA PSS signing function. This will generate and set the
2048 appropriate AlgorithmIdentifiers for PSS based on those in the
2049 corresponding EVP_MD_CTX structure. No application support yet.
2052 *) Support for companion algorithm specific ASN1 signing routines.
2053 New function ASN1_item_sign_ctx() signs a pre-initialised
2054 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2055 the appropriate parameters.
2058 *) Add new algorithm specific ASN1 verification initialisation function
2059 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2060 handling will be the same no matter what EVP_PKEY_METHOD is used.
2061 Add a PSS handler to support verification of PSS signatures: checked
2062 against a number of sample certificates.
2065 *) Add signature printing for PSS. Add PSS OIDs.
2066 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2068 *) Add algorithm specific signature printing. An individual ASN1 method
2069 can now print out signatures instead of the standard hex dump.
2071 More complex signatures (e.g. PSS) can print out more meaningful
2072 information. Include DSA version that prints out the signature
2076 *) Password based recipient info support for CMS library: implementing
2080 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2081 neatly separates the code into cipher and PBE sections and is required
2082 for some algorithms that split PBES2 into separate pieces (such as
2083 password based CMS).
2086 *) Session-handling fixes:
2087 - Fix handling of connections that are resuming with a session ID,
2088 but also support Session Tickets.
2089 - Fix a bug that suppressed issuing of a new ticket if the client
2090 presented a ticket with an expired session.
2091 - Try to set the ticket lifetime hint to something reasonable.
2092 - Make tickets shorter by excluding irrelevant information.
2093 - On the client side, don't ignore renewed tickets.
2094 [Adam Langley, Bodo Moeller (Google)]
2096 *) Fix PSK session representation.
2099 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2101 This work was sponsored by Intel.
2104 *) Add GCM support to TLS library. Some custom code is needed to split
2105 the IV between the fixed (from PRF) and explicit (from TLS record)
2106 portions. This adds all GCM ciphersuites supported by RFC5288 and
2107 RFC5289. Generalise some AES* cipherstrings to include GCM and
2108 add a special AESGCM string for GCM only.
2111 *) Expand range of ctrls for AES GCM. Permit setting invocation
2112 field on decrypt and retrieval of invocation field only on encrypt.
2115 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2116 As required by RFC5289 these ciphersuites cannot be used if for
2117 versions of TLS earlier than 1.2.
2120 *) For FIPS capable OpenSSL interpret a NULL default public key method
2121 as unset and return the appropriate default but do *not* set the default.
2122 This means we can return the appropriate method in applications that
2123 switch between FIPS and non-FIPS modes.
2126 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2127 ENGINE is used then we cannot handle that in the FIPS module so we
2128 keep original code iff non-FIPS operations are allowed.
2131 *) Add -attime option to openssl utilities.
2132 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2134 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2137 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2138 FIPS EC methods unconditionally for now.
2141 *) New build option no-ec2m to disable characteristic 2 code.
2144 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2145 all cases can be covered as some introduce binary incompatibilities.
2148 *) Redirect RSA operations to FIPS module including keygen,
2149 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2152 *) Add similar low level API blocking to ciphers.
2155 *) Low level digest APIs are not approved in FIPS mode: any attempt
2156 to use these will cause a fatal error. Applications that *really* want
2157 to use them can use the private_* version instead.
2160 *) Redirect cipher operations to FIPS module for FIPS builds.
2163 *) Redirect digest operations to FIPS module for FIPS builds.
2166 *) Update build system to add "fips" flag which will link in fipscanister.o
2167 for static and shared library builds embedding a signature if needed.
2170 *) Output TLS supported curves in preference order instead of numerical
2171 order. This is currently hardcoded for the highest order curves first.
2172 This should be configurable so applications can judge speed vs strength.
2175 *) Add TLS v1.2 server support for client authentication.
2178 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2182 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2183 FIPS modules versions.
2186 *) Add TLS v1.2 client side support for client authentication. Keep cache
2187 of handshake records longer as we don't know the hash algorithm to use
2188 until after the certificate request message is received.
2191 *) Initial TLS v1.2 client support. Add a default signature algorithms
2192 extension including all the algorithms we support. Parse new signature
2193 format in client key exchange. Relax some ECC signing restrictions for
2194 TLS v1.2 as indicated in RFC5246.
2197 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2198 to new signature format when needed using client digest preference.
2199 All server ciphersuites should now work correctly in TLS v1.2. No client
2200 support yet and no support for client certificates.
2203 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2204 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2205 ciphersuites. At present only RSA key exchange ciphersuites work with
2206 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2207 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2208 and version checking.
2211 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2212 with this defined it will not be affected by any changes to ssl internal
2213 structures. Add several utility functions to allow openssl application
2214 to work with OPENSSL_NO_SSL_INTERN defined.
2218 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2220 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2223 *) Permit abbreviated handshakes when renegotiating using the function
2224 SSL_renegotiate_abbreviated().
2225 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2227 *) Add call to ENGINE_register_all_complete() to
2228 ENGINE_load_builtin_engines(), so some implementations get used
2229 automatically instead of needing explicit application support.
2232 *) Add support for TLS key exporter as described in RFC5705.
2233 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2235 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2236 a few changes are required:
2238 Add SSL_OP_NO_TLSv1_1 flag.
2239 Add TLSv1_1 methods.
2240 Update version checking logic to handle version 1.1.
2241 Add explicit IV handling (ported from DTLS code).
2242 Add command line options to s_client/s_server.
2245 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2247 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2248 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2249 content decryption and always return the same error. Note: this attack
2250 needs on average 2^20 messages so it only affects automated senders. The
2251 old behaviour can be reenabled in the CMS code by setting the
2252 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2253 an MMA defence is not necessary.
2254 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2255 this issue. (CVE-2012-0884)
2258 *) Fix CVE-2011-4619: make sure we really are receiving a
2259 client hello before rejecting multiple SGC restarts. Thanks to
2260 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2263 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2265 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2266 Thanks to Antonio Martin, Enterprise Secure Access Research and
2267 Development, Cisco Systems, Inc. for discovering this bug and
2268 preparing a fix. (CVE-2012-0050)
2271 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2273 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2274 of the Vaudenay padding oracle attack on CBC mode encryption
2275 which enables an efficient plaintext recovery attack against
2276 the OpenSSL implementation of DTLS. Their attack exploits timing
2277 differences arising during decryption processing. A research
2278 paper describing this attack can be found at:
2279 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2280 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2281 Security Group at Royal Holloway, University of London
2282 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2283 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2284 for preparing the fix. (CVE-2011-4108)
2285 [Robin Seggelmann, Michael Tuexen]
2287 *) Clear bytes used for block padding of SSL 3.0 records.
2289 [Adam Langley (Google)]
2291 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2292 Kadianakis <desnacked@gmail.com> for discovering this issue and
2293 Adam Langley for preparing the fix. (CVE-2011-4619)
2294 [Adam Langley (Google)]
2296 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2297 [Andrey Kulikov <amdeich@gmail.com>]
2299 *) Prevent malformed RFC3779 data triggering an assertion failure.
2300 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2301 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2302 [Rob Austein <sra@hactrn.net>]
2304 *) Improved PRNG seeding for VOS.
2305 [Paul Green <Paul.Green@stratus.com>]
2307 *) Fix ssl_ciph.c set-up race.
2308 [Adam Langley (Google)]
2310 *) Fix spurious failures in ecdsatest.c.
2311 [Emilia Käsper (Google)]
2313 *) Fix the BIO_f_buffer() implementation (which was mixing different
2314 interpretations of the '..._len' fields).
2315 [Adam Langley (Google)]
2317 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2318 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2319 threads won't reuse the same blinding coefficients.
2321 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2322 lock to call BN_BLINDING_invert_ex, and avoids one use of
2323 BN_BLINDING_update for each BN_BLINDING structure (previously,
2324 the last update always remained unused).
2325 [Emilia Käsper (Google)]
2327 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2328 [Bob Buckholz (Google)]
2330 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2332 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2333 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2334 [Kaspar Brand <ossl@velox.ch>]
2336 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2337 for multi-threaded use of ECDH. (CVE-2011-3210)
2338 [Adam Langley (Google)]
2340 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2343 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2344 signature public key algorithm by using OID xref utilities instead.
2345 Before this you could only use some ECC ciphersuites with SHA1 only.
2348 *) Add protection against ECDSA timing attacks as mentioned in the paper
2349 by Billy Bob Brumley and Nicola Tuveri, see:
2351 http://eprint.iacr.org/2011/232.pdf
2353 [Billy Bob Brumley and Nicola Tuveri]
2355 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2357 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2358 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2360 *) Fix bug in string printing code: if *any* escaping is enabled we must
2361 escape the escape character (backslash) or the resulting string is
2365 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2367 *) Disable code workaround for ancient and obsolete Netscape browsers
2368 and servers: an attacker can use it in a ciphersuite downgrade attack.
2369 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2372 *) Fixed J-PAKE implementation error, originally discovered by
2373 Sebastien Martini, further info and confirmation from Stefan
2374 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2377 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2379 *) Fix extension code to avoid race conditions which can result in a buffer
2380 overrun vulnerability: resumed sessions must not be modified as they can
2381 be shared by multiple threads. CVE-2010-3864
2384 *) Fix WIN32 build system to correctly link an ENGINE directory into
2388 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2390 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2392 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2394 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2396 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2397 context. The operation can be customised via the ctrl mechanism in
2398 case ENGINEs want to include additional functionality.
2401 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2404 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2405 output hashes compatible with older versions of OpenSSL.
2406 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2408 *) Fix compression algorithm handling: if resuming a session use the
2409 compression algorithm of the resumed session instead of determining
2410 it from client hello again. Don't allow server to change algorithm.
2413 *) Add load_crls() function to apps tidying load_certs() too. Add option
2414 to verify utility to allow additional CRLs to be included.
2417 *) Update OCSP request code to permit adding custom headers to the request:
2418 some responders need this.
2421 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2423 [Julia Lawall <julia@diku.dk>]
2425 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2426 needlessly dereferenced structures, used obsolete functions and
2427 didn't handle all updated verify codes correctly.
2430 *) Disable MD2 in the default configuration.
2433 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2434 indicate the initial BIO being pushed or popped. This makes it possible
2435 to determine whether the BIO is the one explicitly called or as a result
2436 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2437 it handles reference counts correctly and doesn't zero out the I/O bio
2438 when it is not being explicitly popped. WARNING: applications which
2439 included workarounds for the old buggy behaviour will need to be modified
2440 or they could free up already freed BIOs.
2443 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2444 renaming to all platforms (within the 0.9.8 branch, this was
2445 done conditionally on Netware platforms to avoid a name clash).
2446 [Guenter <lists@gknw.net>]
2448 *) Add ECDHE and PSK support to DTLS.
2449 [Michael Tuexen <tuexen@fh-muenster.de>]
2451 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2455 *) Add "missing" function EVP_MD_flags() (without this the only way to
2456 retrieve a digest flags is by accessing the structure directly. Update
2457 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2458 or cipher is registered as in the "from" argument. Print out all
2459 registered digests in the dgst usage message instead of manually
2460 attempting to work them out.
2463 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2464 this allows the use of compression and extensions. Change default cipher
2465 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2466 by default unless an application cipher string requests it.
2469 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2470 key ids to find matching certificates and keys but some PKCS#12 files
2471 don't follow the (somewhat unwritten) rules and this strategy fails.
2472 Now just gather all certificates together and the first private key
2473 then look for the first certificate that matches the key.
2476 *) Support use of registered digest and cipher names for dgst and cipher
2477 commands instead of having to add each one as a special case. So now
2484 openssl dgst -sha256 foo
2486 and this works for ENGINE based algorithms too.
2490 *) Update Gost ENGINE to support parameter files.
2491 [Victor B. Wagner <vitus@cryptocom.ru>]
2493 *) Support GeneralizedTime in ca utility.
2494 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2496 *) Enhance the hash format used for certificate directory links. The new
2497 form uses the canonical encoding (meaning equivalent names will work
2498 even if they aren't identical) and uses SHA1 instead of MD5. This form
2499 is incompatible with the older format and as a result c_rehash should
2500 be used to rebuild symbolic links.
2503 *) Make PKCS#8 the default write format for private keys, replacing the
2504 traditional format. This form is standardised, more secure and doesn't
2505 include an implicit MD5 dependency.
2508 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2509 committed to OpenSSL should pass this lot as a minimum.
2512 *) Add session ticket override functionality for use by EAP-FAST.
2513 [Jouni Malinen <j@w1.fi>]
2515 *) Modify HMAC functions to return a value. Since these can be implemented
2516 in an ENGINE errors can occur.
2519 *) Type-checked OBJ_bsearch_ex.
2522 *) Type-checked OBJ_bsearch. Also some constification necessitated
2523 by type-checking. Still to come: TXT_DB, bsearch(?),
2524 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2528 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2529 seconds to a tm structure directly, instead of going through OS
2530 specific date routines. This avoids any issues with OS routines such
2531 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2532 and X509_time_adj_ex() to cover the extended range. The existing
2533 X509_time_adj() is still usable and will no longer have any date issues.
2536 *) Delta CRL support. New use deltas option which will attempt to locate
2537 and search any appropriate delta CRLs available.
2539 This work was sponsored by Google.
2542 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2543 code and add additional score elements. Validate alternate CRL paths
2544 as part of the CRL checking and indicate a new error "CRL path validation
2545 error" in this case. Applications wanting additional details can use
2546 the verify callback and check the new "parent" field. If this is not
2547 NULL CRL path validation is taking place. Existing applications wont
2548 see this because it requires extended CRL support which is off by
2551 This work was sponsored by Google.
2554 *) Support for freshest CRL extension.
2556 This work was sponsored by Google.
2559 *) Initial indirect CRL support. Currently only supported in the CRLs
2560 passed directly and not via lookup. Process certificate issuer
2561 CRL entry extension and lookup CRL entries by bother issuer name
2562 and serial number. Check and process CRL issuer entry in IDP extension.
2564 This work was sponsored by Google.
2567 *) Add support for distinct certificate and CRL paths. The CRL issuer
2568 certificate is validated separately in this case. Only enabled if
2569 an extended CRL support flag is set: this flag will enable additional
2570 CRL functionality in future.
2572 This work was sponsored by Google.
2575 *) Add support for policy mappings extension.
2577 This work was sponsored by Google.
2580 *) Fixes to pathlength constraint, self issued certificate handling,
2581 policy processing to align with RFC3280 and PKITS tests.
2583 This work was sponsored by Google.
2586 *) Support for name constraints certificate extension. DN, email, DNS
2587 and URI types are currently supported.
2589 This work was sponsored by Google.
2592 *) To cater for systems that provide a pointer-based thread ID rather
2593 than numeric, deprecate the current numeric thread ID mechanism and
2594 replace it with a structure and associated callback type. This
2595 mechanism allows a numeric "hash" to be extracted from a thread ID in
2596 either case, and on platforms where pointers are larger than 'long',
2597 mixing is done to help ensure the numeric 'hash' is usable even if it
2598 can't be guaranteed unique. The default mechanism is to use "&errno"
2599 as a pointer-based thread ID to distinguish between threads.
2601 Applications that want to provide their own thread IDs should now use
2602 CRYPTO_THREADID_set_callback() to register a callback that will call
2603 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2605 Note that ERR_remove_state() is now deprecated, because it is tied
2606 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2607 to free the current thread's error state should be replaced by
2608 ERR_remove_thread_state(NULL).
2610 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2611 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2612 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2613 application was previously providing a numeric thread callback that
2614 was inappropriate for distinguishing threads, then uniqueness might
2615 have been obtained with &errno that happened immediately in the
2616 intermediate development versions of OpenSSL; this is no longer the
2617 case, the numeric thread callback will now override the automatic use
2619 [Geoff Thorpe, with help from Bodo Moeller]
2621 *) Initial support for different CRL issuing certificates. This covers a
2622 simple case where the self issued certificates in the chain exist and
2623 the real CRL issuer is higher in the existing chain.
2625 This work was sponsored by Google.
2628 *) Removed effectively defunct crypto/store from the build.
2631 *) Revamp of STACK to provide stronger type-checking. Still to come:
2632 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2633 ASN1_STRING, CONF_VALUE.
2636 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2637 RAM on SSL connections. This option can save about 34k per idle SSL.
2640 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2641 STACK, TXT_DB, bsearch, qsort.
2644 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2645 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2646 support for data, signedData, compressedData, digestedData and
2647 encryptedData, envelopedData types included. Scripts to check against
2648 RFC4134 examples draft and interop and consistency checks of many
2649 content types and variants.
2652 *) Add options to enc utility to support use of zlib compression BIO.
2655 *) Extend mk1mf to support importing of options and assembly language
2656 files from Configure script, currently only included in VC-WIN32.
2657 The assembly language rules can now optionally generate the source
2658 files from the associated perl scripts.
2661 *) Implement remaining functionality needed to support GOST ciphersuites.
2662 Interop testing has been performed using CryptoPro implementations.
2663 [Victor B. Wagner <vitus@cryptocom.ru>]
2665 *) s390x assembler pack.
2668 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2672 *) Implement Opaque PRF Input TLS extension as specified in
2673 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2674 official specification yet and no extension type assignment by
2675 IANA exists, this extension (for now) will have to be explicitly
2676 enabled when building OpenSSL by providing the extension number
2677 to use. For example, specify an option
2679 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2681 to the "config" or "Configure" script to enable the extension,
2682 assuming extension number 0x9527 (which is a completely arbitrary
2683 and unofficial assignment based on the MD5 hash of the Internet
2684 Draft). Note that by doing so, you potentially lose
2685 interoperability with other TLS implementations since these might
2686 be using the same extension number for other purposes.
2688 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2689 opaque PRF input value to use in the handshake. This will create
2690 an interal copy of the length-'len' string at 'src', and will
2691 return non-zero for success.
2693 To get more control and flexibility, provide a callback function
2696 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2697 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2701 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2704 Callback function 'cb' will be called in handshakes, and is
2705 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2706 Argument 'arg' is for application purposes (the value as given to
2707 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2708 be provided to the callback function). The callback function
2709 has to return non-zero to report success: usually 1 to use opaque
2710 PRF input just if possible, or 2 to enforce use of the opaque PRF
2711 input. In the latter case, the library will abort the handshake
2712 if opaque PRF input is not successfully negotiated.
2714 Arguments 'peerinput' and 'len' given to the callback function
2715 will always be NULL and 0 in the case of a client. A server will
2716 see the client's opaque PRF input through these variables if
2717 available (NULL and 0 otherwise). Note that if the server
2718 provides an opaque PRF input, the length must be the same as the
2719 length of the client's opaque PRF input.
2721 Note that the callback function will only be called when creating
2722 a new session (session resumption can resume whatever was
2723 previously negotiated), and will not be called in SSL 2.0
2724 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2725 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2726 for applications that need to enforce opaque PRF input.
2730 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2733 [Victor B. Wagner <vitus@cryptocom.ru>]
2735 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2736 RFC4507bis. The encrypted ticket format is an encrypted encoded
2737 SSL_SESSION structure, that way new session features are automatically
2740 If a client application caches session in an SSL_SESSION structure
2741 support is transparent because tickets are now stored in the encoded
2744 The SSL_CTX structure automatically generates keys for ticket
2745 protection in servers so again support should be possible
2746 with no application modification.
2748 If a client or server wishes to disable RFC4507 support then the option
2749 SSL_OP_NO_TICKET can be set.
2751 Add a TLS extension debugging callback to allow the contents of any client
2752 or server extensions to be examined.
2754 This work was sponsored by Google.
2757 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2758 OpenSSL should now compile cleanly on gcc 4.2
2759 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2761 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2762 support including streaming MAC support: this is required for GOST
2763 ciphersuite support.
2764 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2766 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2767 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2768 to output in BER and PEM format.
2771 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2772 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2773 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2774 ENGINE support for HMAC keys which are unextractable. New -mac and
2775 -macopt options to dgst utility.
2778 *) New option -sigopt to dgst utility. Update dgst to use
2779 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2780 alternative signing parameters such as X9.31 or PSS in the dgst
2784 *) Change ssl_cipher_apply_rule(), the internal function that does
2785 the work each time a ciphersuite string requests enabling
2786 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2787 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2788 the order of disabled ciphersuites such that those ciphersuites
2789 that most recently went from enabled to disabled not only stay
2790 in order with respect to each other, but also have higher priority
2791 than other disabled ciphersuites the next time ciphersuites are
2794 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2795 the same ciphersuites as with "HIGH" alone, but in a specific
2796 order where the PSK ciphersuites come first (since they are the
2797 most recently disabled ciphersuites when "HIGH" is parsed).
2799 Also, change ssl_create_cipher_list() (using this new
2800 funcionality) such that between otherwise identical
2801 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2805 *) Change ssl_create_cipher_list() so that it automatically
2806 arranges the ciphersuites in reasonable order before starting
2807 to process the rule string. Thus, the definition for "DEFAULT"
2808 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2809 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2810 This makes it much easier to arrive at a reasonable default order
2811 in applications for which anonymous ciphers are OK (meaning
2812 that you can't actually use DEFAULT).
2813 [Bodo Moeller; suggested by Victor Duchovni]
2815 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2816 processing) into multiple integers instead of setting
2817 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2818 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2819 (These masks as well as the individual bit definitions are hidden
2820 away into the non-exported interface ssl/ssl_locl.h, so this
2821 change to the definition of the SSL_CIPHER structure shouldn't
2822 affect applications.) This give us more bits for each of these
2823 categories, so there is no longer a need to coagulate AES128 and
2824 AES256 into a single algorithm bit, and to coagulate Camellia128
2825 and Camellia256 into a single algorithm bit, which has led to all
2828 Thus, among other things, the kludge introduced in 0.9.7m and
2829 0.9.8e for masking out AES256 independently of AES128 or masking
2830 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2832 With the change, we also introduce new ciphersuite aliases that
2833 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2837 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2838 Use the leftmost N bytes of the signature input if the input is
2839 larger than the prime q (with N being the size in bytes of q).
2842 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2843 it yet and it is largely untested.
2846 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2849 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2850 some compilers (gcc 4.2 and later) reject their use. Safestack is
2851 reimplemented. Update ASN1 to avoid use of legacy functions.
2854 *) Win32/64 targets are linked with Winsock2.
2857 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2858 to external functions. This can be used to increase CRL handling
2859 efficiency especially when CRLs are very large by (for example) storing
2860 the CRL revoked certificates in a database.
2863 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2864 new CRLs added to a directory can be used. New command line option
2865 -verify_return_error to s_client and s_server. This causes real errors
2866 to be returned by the verify callback instead of carrying on no matter
2867 what. This reflects the way a "real world" verify callback would behave.
2870 *) GOST engine, supporting several GOST algorithms and public key formats.
2871 Kindly donated by Cryptocom.
2874 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2875 partitioned by DP are handled but no indirect CRL or reason partitioning
2876 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2877 selected via a scoring technique which handles IDP and AKID in CRLs.
2880 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2881 will ultimately be used for all verify operations: this will remove the
2882 X509_STORE dependency on certificate verification and allow alternative
2883 lookup methods. X509_STORE based implementations of these two callbacks.
2886 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2887 Modify get_crl() to find a valid (unexpired) CRL if possible.
2890 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2891 this would be called X509_CRL_cmp() but that name is already used by
2892 a function that just compares CRL issuer names. Cache several CRL
2893 extensions in X509_CRL structure and cache CRLDP in X509.
2896 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2897 this maps equivalent X509_NAME structures into a consistent structure.
2898 Name comparison can then be performed rapidly using memcmp().
2901 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2905 *) Allow digests to supply their own micalg string for S/MIME type using
2906 the ctrl EVP_MD_CTRL_MICALG.
2909 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2910 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2911 ctrl. It can then customise the structure before and/or after signing
2915 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2916 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2917 to free up any added signature OIDs.
2920 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2921 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2922 digest and cipher tables. New options added to openssl utility:
2923 list-message-digest-algorithms and list-cipher-algorithms.
2926 *) Change the array representation of binary polynomials: the list
2927 of degrees of non-zero coefficients is now terminated with -1.
2928 Previously it was terminated with 0, which was also part of the
2929 value; thus, the array representation was not applicable to
2930 polynomials where t^0 has coefficient zero. This change makes
2931 the array representation useful in a more general context.
2934 *) Various modifications and fixes to SSL/TLS cipher string
2935 handling. For ECC, the code now distinguishes between fixed ECDH
2936 with RSA certificates on the one hand and with ECDSA certificates
2937 on the other hand, since these are separate ciphersuites. The
2938 unused code for Fortezza ciphersuites has been removed.
2940 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2941 (not "ECDHE"). For consistency with the code for DH
2942 certificates, use of ECDH certificates is now considered ECDH
2943 authentication, not RSA or ECDSA authentication (the latter is
2944 merely the CA's signing algorithm and not actively used in the
2947 The temporary ciphersuite alias "ECCdraft" is no longer
2948 available, and ECC ciphersuites are no longer excluded from "ALL"
2949 and "DEFAULT". The following aliases now exist for RFC 4492
2950 ciphersuites, most of these by analogy with the DH case:
2952 kECDHr - ECDH cert, signed with RSA
2953 kECDHe - ECDH cert, signed with ECDSA
2954 kECDH - ECDH cert (signed with either RSA or ECDSA)
2955 kEECDH - ephemeral ECDH
2956 ECDH - ECDH cert or ephemeral ECDH
2962 AECDH - anonymous ECDH
2963 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2967 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2968 Use correct micalg parameters depending on digest(s) in signed message.
2971 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2972 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2975 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2976 an engine to register a method. Add ENGINE lookups for methods and
2977 functional reference processing.
2980 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2981 EVP_{Sign,Verify}* which allow an application to customise the signature
2985 *) New -resign option to smime utility. This adds one or more signers
2986 to an existing PKCS#7 signedData structure. Also -md option to use an
2987 alternative message digest algorithm for signing.
2990 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2991 create PKCS7 structures containing multiple signers. Update smime
2992 application to support multiple signers.
2995 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2999 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3000 Reorganize PBE internals to lookup from a static table using NIDs,
3001 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3002 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3003 PRF which will be automatically used with PBES2.
3006 *) Replace the algorithm specific calls to generate keys in "req" with the
3010 *) Update PKCS#7 enveloped data routines to use new API. This is now
3011 supported by any public key method supporting the encrypt operation. A
3012 ctrl is added to allow the public key algorithm to examine or modify
3013 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3017 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3018 a default digest type to use. In most cases this will be SHA1 but some
3019 algorithms (such as GOST) need to specify an alternative digest. The
3020 return value indicates how strong the preference is 1 means optional and
3021 2 is mandatory (that is it is the only supported type). Modify
3022 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3023 use the default md. Update openssl utilities to use the default digest
3024 type for signing if it is not explicitly indicated.
3027 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3028 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3029 signing method from the key type. This effectively removes the link
3030 between digests and public key types.
3033 *) Add an OID cross reference table and utility functions. Its purpose is to
3034 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3035 rsaEncryption. This will allow some of the algorithm specific hackery
3036 needed to use the correct OID to be removed.
3039 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3040 structures for PKCS7_sign(). They are now set up by the relevant public
3044 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3047 *) Add support for key derivation (agreement) in the API, DH method and
3051 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3052 public and private key formats. As a side effect these add additional
3053 command line functionality not previously available: DSA signatures can be
3054 generated and verified using pkeyutl and DH key support and generation in
3059 [Oliver Tappe <zooey@hirschkaefer.de>]
3061 *) New make target "install_html_docs" installs HTML renditions of the
3063 [Oliver Tappe <zooey@hirschkaefer.de>]
3065 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3066 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3067 support key and parameter generation and add initial key generation
3068 functionality for RSA.
3071 *) Add functions for main EVP_PKEY_method operations. The undocumented
3072 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3073 EVP_PKEY_{encrypt,decrypt}_old.
3076 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3077 key API, doesn't do much yet.
3080 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3081 public key algorithms. New option to openssl utility:
3082 "list-public-key-algorithms" to print out info.
3085 *) Implement the Supported Elliptic Curves Extension for
3086 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3089 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3090 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3093 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3094 utilities such as rsa, dsa, dsaparam etc except they process any key
3098 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3099 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3100 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3104 *) Initial support for pluggable public key ASN1.
3105 De-spaghettify the public key ASN1 handling. Move public and private
3106 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3107 algorithm specific handling to a single module within the relevant
3108 algorithm directory. Add functions to allow (near) opaque processing
3109 of public and private key structures.
3112 *) Implement the Supported Point Formats Extension for
3113 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3116 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3117 for the psk identity [hint] and the psk callback functions to the
3118 SSL_SESSION, SSL and SSL_CTX structure.
3121 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3125 SSL_CTX_use_psk_identity_hint
3126 SSL_get_psk_identity_hint
3127 SSL_get_psk_identity
3128 SSL_use_psk_identity_hint
3130 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3132 *) Add RFC 3161 compliant time stamp request creation, response generation
3133 and response verification functionality.
3134 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3136 *) Add initial support for TLS extensions, specifically for the server_name
3137 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3138 have new members for a host name. The SSL data structure has an
3139 additional member SSL_CTX *initial_ctx so that new sessions can be
3140 stored in that context to allow for session resumption, even after the
3141 SSL has been switched to a new SSL_CTX in reaction to a client's
3142 server_name extension.
3144 New functions (subject to change):
3146 SSL_get_servername()
3147 SSL_get_servername_type()
3150 New CTRL codes and macros (subject to change):
3152 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3153 - SSL_CTX_set_tlsext_servername_callback()
3154 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3155 - SSL_CTX_set_tlsext_servername_arg()
3156 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3158 openssl s_client has a new '-servername ...' option.
3160 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3161 '-key2 ...', '-servername_fatal' (subject to change). This allows
3162 testing the HostName extension for a specific single host name ('-cert'
3163 and '-key' remain fallbacks for handshakes without HostName
3164 negotiation). If the unrecognized_name alert has to be sent, this by
3165 default is a warning; it becomes fatal with the '-servername_fatal'
3168 [Peter Sylvester, Remy Allais, Christophe Renou]
3170 *) Whirlpool hash implementation is added.
3173 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3174 bn(64,32). Because of instruction set limitations it doesn't have
3175 any negative impact on performance. This was done mostly in order
3176 to make it possible to share assembler modules, such as bn_mul_mont
3177 implementations, between 32- and 64-bit builds without hassle.
3180 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3181 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3185 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3186 dedicated Montgomery multiplication procedure, is introduced.
3187 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3188 "64-bit" performance on certain 32-bit targets.
3191 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3192 in SSL structures. New SSL ctrl to set maximum send fragment size.
3193 Save memory by seeting the I/O buffer sizes dynamically instead of
3194 using the maximum available value.
3197 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3198 in addition to the text details.
3201 *) Very, very preliminary EXPERIMENTAL support for printing of general
3202 ASN1 structures. This currently produces rather ugly output and doesn't
3203 handle several customised structures at all.
3206 *) Integrated support for PVK file format and some related formats such
3207 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3208 these in the 'rsa' and 'dsa' utilities.
3211 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3214 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3215 place for the (very old) "NETSCAPE" format certificates which are now
3216 handled using new ASN1 code equivalents.
3219 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3220 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3221 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3224 *) Modify CRL distribution points extension code to print out previously
3225 unsupported fields. Enhance extension setting code to allow setting of
3229 *) Add print and set support for Issuing Distribution Point CRL extension.
3232 *) Change 'Configure' script to enable Camellia by default.
3235 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3237 *) When rejecting SSL/TLS records due to an incorrect version number, never
3238 update s->server with a new major version number. As of
3239 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3240 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3241 the previous behavior could result in a read attempt at NULL when
3242 receiving specific incorrect SSL/TLS records once record payload
3243 protection is active. (CVE-2010-0740)
3244 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3246 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3247 could be crashed if the relevant tables were not present (e.g. chrooted).
3248 [Tomas Hoger <thoger@redhat.com>]
3250 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3252 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3253 [Martin Olsson, Neel Mehta]
3255 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3256 accommodate for stack sorting, always a write lock!).
3259 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3260 excessive delays in the RAND_poll(): over a minute. As a workaround
3261 include a time check in the inner Heap32Next loop too.
3264 *) The code that handled flushing of data in SSL/TLS originally used the
3265 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3266 the problem outlined in PR#1949. The fix suggested there however can
3267 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3268 of Apache). So instead simplify the code to flush unconditionally.
3269 This should be fine since flushing with no data to flush is a no op.
3272 *) Handle TLS versions 2.0 and later properly and correctly use the
3273 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3274 off ancient servers have a habit of sticking around for a while...
3277 *) Modify compression code so it frees up structures without using the
3278 ex_data callbacks. This works around a problem where some applications
3279 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3280 restarting) then use compression (e.g. SSL with compression) later.
3281 This results in significant per-connection memory leaks and
3282 has caused some security issues including CVE-2008-1678 and
3286 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3287 change when encrypting or decrypting.
3290 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3291 connect and renegotiate with servers which do not support RI.
3292 Until RI is more widely deployed this option is enabled by default.
3295 *) Add "missing" ssl ctrls to clear options and mode.
3298 *) If client attempts to renegotiate and doesn't support RI respond with
3299 a no_renegotiation alert as required by RFC5746. Some renegotiating
3300 TLS clients will continue a connection gracefully when they receive
3301 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3302 waiting for a server hello which it will never receive. Now we treat a
3303 received no_renegotiation alert as a fatal error. This is because
3304 applications requesting a renegotiation might well expect it to succeed
3305 and would have no code in place to handle the server denying it so the
3306 only safe thing to do is to terminate the connection.
3309 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3310 peer supports secure renegotiation and 0 otherwise. Print out peer
3311 renegotiation support in s_client/s_server.
3314 *) Replace the highly broken and deprecated SPKAC certification method with
3315 the updated NID creation version. This should correctly handle UTF8.
3318 *) Implement RFC5746. Re-enable renegotiation but require the extension
3319 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3320 turns out to be a bad idea. It has been replaced by
3321 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3322 SSL_CTX_set_options(). This is really not recommended unless you
3323 know what you are doing.
3324 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3326 *) Fixes to stateless session resumption handling. Use initial_ctx when
3327 issuing and attempting to decrypt tickets in case it has changed during
3328 servername handling. Use a non-zero length session ID when attempting
3329 stateless session resumption: this makes it possible to determine if
3330 a resumption has occurred immediately after receiving server hello
3331 (several places in OpenSSL subtly assume this) instead of later in
3335 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3336 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3337 fixes for a few places where the return code is not checked
3339 [Julia Lawall <julia@diku.dk>]
3341 *) Add --strict-warnings option to Configure script to include devteam
3342 warnings in other configurations.
3345 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3346 makes it possible to install openssl libraries in locations which
3347 have names other than "lib", for example "/usr/lib64" which some
3349 [Steve Henson, based on patch from Jeremy Utley]
3351 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3352 X690 8.9.12 and can produce some misleading textual output of OIDs.
3353 [Steve Henson, reported by Dan Kaminsky]
3355 *) Delete MD2 from algorithm tables. This follows the recommendation in
3356 several standards that it is not used in new applications due to
3357 several cryptographic weaknesses. For binary compatibility reasons
3358 the MD2 API is still compiled in by default.
3361 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3365 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3366 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3368 [Guenter <lists@gknw.net>]
3370 *) Fix the server certificate chain building code to use X509_verify_cert(),
3371 it used to have an ad-hoc builder which was unable to cope with anything
3372 other than a simple chain.
3373 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3375 *) Don't check self signed certificate signatures in X509_verify_cert()
3376 by default (a flag can override this): it just wastes time without
3377 adding any security. As a useful side effect self signed root CAs
3378 with non-FIPS digests are now usable in FIPS mode.
3381 *) In dtls1_process_out_of_seq_message() the check if the current message
3382 is already buffered was missing. For every new message was memory
3383 allocated, allowing an attacker to perform an denial of service attack
3384 with sending out of seq handshake messages until there is no memory
3385 left. Additionally every future messege was buffered, even if the
3386 sequence number made no sense and would be part of another handshake.
3387 So only messages with sequence numbers less than 10 in advance will be
3388 buffered. (CVE-2009-1378)
3389 [Robin Seggelmann, discovered by Daniel Mentz]
3391 *) Records are buffered if they arrive with a future epoch to be
3392 processed after finishing the corresponding handshake. There is
3393 currently no limitation to this buffer allowing an attacker to perform
3394 a DOS attack with sending records with future epochs until there is no
3395 memory left. This patch adds the pqueue_size() function to determine
3396 the size of a buffer and limits the record buffer to 100 entries.
3398 [Robin Seggelmann, discovered by Daniel Mentz]
3400 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3401 parent structure is freed. (CVE-2009-1379)
3404 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3405 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3408 [Ilya O. <vrghost@gmail.com>]
3410 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3412 *) Disable renegotiation completely - this fixes a severe security
3413 problem (CVE-2009-3555) at the cost of breaking all
3414 renegotiation. Renegotiation can be re-enabled by setting
3415 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3416 run-time. This is really not recommended unless you know what
3420 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3422 *) Don't set val to NULL when freeing up structures, it is freed up by
3423 underlying code. If sizeof(void *) > sizeof(long) this can result in
3424 zeroing past the valid field. (CVE-2009-0789)
3425 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3427 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3428 checked correctly. This would allow some invalid signed attributes to
3429 appear to verify correctly. (CVE-2009-0591)
3430 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3432 *) Reject UniversalString and BMPString types with invalid lengths. This
3433 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3434 a legal length. (CVE-2009-0590)
3437 *) Set S/MIME signing as the default purpose rather than setting it
3438 unconditionally. This allows applications to override it at the store
3442 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3443 to handle some structures.
3446 *) Improve efficiency of mem_gets: don't search whole buffer each time
3448 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3450 *) New -hex option for openssl rand.
3453 *) Print out UTF8String and NumericString when parsing ASN1.
3456 *) Support NumericString type for name components.
3459 *) Allow CC in the environment to override the automatically chosen
3460 compiler. Note that nothing is done to ensure flags work with the
3464 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3466 *) Properly check EVP_VerifyFinal() and similar return values
3468 [Ben Laurie, Bodo Moeller, Google Security Team]
3470 *) Enable TLS extensions by default.
3473 *) Allow the CHIL engine to be loaded, whether the application is
3474 multithreaded or not. (This does not release the developer from the
3475 obligation to set up the dynamic locking callbacks.)
3476 [Sander Temme <sander@temme.net>]
3478 *) Use correct exit code if there is an error in dgst command.
3479 [Steve Henson; problem pointed out by Roland Dirlewanger]
3481 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3482 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3485 *) Add experimental JPAKE support, including demo authentication in
3486 s_client and s_server.
3489 *) Set the comparison function in v3_addr_canonize().
3490 [Rob Austein <sra@hactrn.net>]
3492 *) Add support for XMPP STARTTLS in s_client.
3493 [Philip Paeps <philip@freebsd.org>]
3495 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3496 to ensure that even with this option, only ciphersuites in the
3497 server's preference list will be accepted. (Note that the option
3498 applies only when resuming a session, so the earlier behavior was
3499 just about the algorithm choice for symmetric cryptography.)
3502 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3504 *) Fix NULL pointer dereference if a DTLS server received
3505 ChangeCipherSpec as first record (CVE-2009-1386).
3508 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3509 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3512 *) The fix in 0.9.8c that supposedly got rid of unsafe
3513 double-checked locking was incomplete for RSA blinding,
3514 addressing just one layer of what turns out to have been
3515 doubly unsafe triple-checked locking.
3517 So now fix this for real by retiring the MONT_HELPER macro
3518 in crypto/rsa/rsa_eay.c.
3520 [Bodo Moeller; problem pointed out by Marius Schilder]
3522 *) Various precautionary measures:
3524 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3526 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3527 (NB: This would require knowledge of the secret session ticket key
3528 to exploit, in which case you'd be SOL either way.)
3530 - Change bn_nist.c so that it will properly handle input BIGNUMs
3531 outside the expected range.
3533 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3536 [Neel Mehta, Bodo Moeller]
3538 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3539 the load fails. Useful for distros.
3540 [Ben Laurie and the FreeBSD team]
3542 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3545 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3548 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3550 This work was sponsored by Logica.
3553 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3554 keystores. Support for SSL/TLS client authentication too.
3555 Not compiled unless enable-capieng specified to Configure.
3557 This work was sponsored by Logica.
3560 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3561 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3562 attribute creation routines such as certifcate requests and PKCS#12
3566 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3568 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3569 handshake which could lead to a cilent crash as found using the
3570 Codenomicon TLS test suite (CVE-2008-1672)
3571 [Steve Henson, Mark Cox]
3573 *) Fix double free in TLS server name extensions which could lead to
3574 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3577 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3579 Clear the error queue to ensure that error entries left from
3580 older function calls do not interfere with the correct operation.
3581 [Lutz Jaenicke, Erik de Castro Lopo]
3583 *) Remove root CA certificates of commercial CAs:
3585 The OpenSSL project does not recommend any specific CA and does not
3586 have any policy with respect to including or excluding any CA.
3587 Therefore it does not make any sense to ship an arbitrary selection
3588 of root CA certificates with the OpenSSL software.
3591 *) RSA OAEP patches to fix two separate invalid memory reads.
3592 The first one involves inputs when 'lzero' is greater than
3593 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3594 before the beginning of from). The second one involves inputs where
3595 the 'db' section contains nothing but zeroes (there is a one-byte
3596 invalid read after the end of 'db').
3597 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3599 *) Partial backport from 0.9.9-dev:
3601 Introduce bn_mul_mont (dedicated Montgomery multiplication
3602 procedure) as a candidate for BIGNUM assembler implementation.
3603 While 0.9.9-dev uses assembler for various architectures, only
3604 x86_64 is available by default here in the 0.9.8 branch, and
3605 32-bit x86 is available through a compile-time setting.
3607 To try the 32-bit x86 assembler implementation, use Configure
3608 option "enable-montasm" (which exists only for this backport).
3610 As "enable-montasm" for 32-bit x86 disclaims code stability
3611 anyway, in this constellation we activate additional code
3612 backported from 0.9.9-dev for further performance improvements,
3613 namely BN_from_montgomery_word. (To enable this otherwise,
3614 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3616 [Andy Polyakov (backport partially by Bodo Moeller)]
3618 *) Add TLS session ticket callback. This allows an application to set
3619 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3620 values. This is useful for key rollover for example where several key
3621 sets may exist with different names.
3624 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3625 This was broken until now in 0.9.8 releases, such that the only way
3626 a registered ENGINE could be used (assuming it initialises
3627 successfully on the host) was to explicitly set it as the default
3628 for the relevant algorithms. This is in contradiction with 0.9.7
3629 behaviour and the documentation. With this fix, when an ENGINE is
3630 registered into a given algorithm's table of implementations, the
3631 'uptodate' flag is reset so that auto-discovery will be used next
3632 time a new context for that algorithm attempts to select an
3634 [Ian Lister (tweaked by Geoff Thorpe)]
3636 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3637 implemention in the following ways:
3639 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3642 Lack of BER streaming support means one pass streaming processing is
3643 only supported if data is detached: setting the streaming flag is
3644 ignored for embedded content.
3646 CMS support is disabled by default and must be explicitly enabled
3647 with the enable-cms configuration option.
3650 *) Update the GMP engine glue to do direct copies between BIGNUM and
3651 mpz_t when openssl and GMP use the same limb size. Otherwise the
3652 existing "conversion via a text string export" trick is still used.
3653 [Paul Sheer <paulsheer@gmail.com>]
3655 *) Zlib compression BIO. This is a filter BIO which compressed and
3656 uncompresses any data passed through it.
3659 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3660 RFC3394 compatible AES key wrapping.
3663 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3664 sets string data without copying. X509_ALGOR_set0() and
3665 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3666 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3667 from an X509_ATTRIBUTE structure optionally checking it occurs only
3668 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3672 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3673 to get the expected BN_FLG_CONSTTIME behavior.
3674 [Bodo Moeller (Google)]
3678 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3679 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3680 - added some more tests to do_tests.pl
3681 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3682 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3683 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3684 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3685 - various changes to netware.pl to enable gcc-cross builds on Win32
3687 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3688 - various changes to fix missing prototype warnings
3689 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3690 - added AES, WHIRLPOOL and CPUID assembler code to build files
3691 - added missing AES assembler make rules to mk1mf.pl
3692 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3693 [Guenter Knauf <eflash@gmx.net>]
3695 *) Implement certificate status request TLS extension defined in RFC3546.
3696 A client can set the appropriate parameters and receive the encoded
3697 OCSP response via a callback. A server can query the supplied parameters
3698 and set the encoded OCSP response in the callback. Add simplified examples
3699 to s_client and s_server.
3702 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3704 *) Fix various bugs:
3705 + Binary incompatibility of ssl_ctx_st structure
3706 + DTLS interoperation with non-compliant servers
3707 + Don't call get_session_cb() without proposed session
3708 + Fix ia64 assembler code
3709 [Andy Polyakov, Steve Henson]
3711 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3713 *) DTLS Handshake overhaul. There were longstanding issues with
3714 OpenSSL DTLS implementation, which were making it impossible for
3715 RFC 4347 compliant client to communicate with OpenSSL server.
3716 Unfortunately just fixing these incompatibilities would "cut off"
3717 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3718 server keeps tolerating non RFC compliant syntax. The opposite is
3719 not true, 0.9.8f client can not communicate with earlier server.
3720 This update even addresses CVE-2007-4995.
3723 *) Changes to avoid need for function casts in OpenSSL: some compilers
3724 (gcc 4.2 and later) reject their use.
3725 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3728 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3729 RFC4507bis. The encrypted ticket format is an encrypted encoded
3730 SSL_SESSION structure, that way new session features are automatically
3733 If a client application caches session in an SSL_SESSION structure
3734 support is transparent because tickets are now stored in the encoded
3737 The SSL_CTX structure automatically generates keys for ticket
3738 protection in servers so again support should be possible
3739 with no application modification.
3741 If a client or server wishes to disable RFC4507 support then the option
3742 SSL_OP_NO_TICKET can be set.
3744 Add a TLS extension debugging callback to allow the contents of any client
3745 or server extensions to be examined.
3747 This work was sponsored by Google.
3750 *) Add initial support for TLS extensions, specifically for the server_name
3751 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3752 have new members for a host name. The SSL data structure has an
3753 additional member SSL_CTX *initial_ctx so that new sessions can be
3754 stored in that context to allow for session resumption, even after the
3755 SSL has been switched to a new SSL_CTX in reaction to a client's
3756 server_name extension.
3758 New functions (subject to change):
3760 SSL_get_servername()
3761 SSL_get_servername_type()
3764 New CTRL codes and macros (subject to change):
3766 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3767 - SSL_CTX_set_tlsext_servername_callback()
3768 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3769 - SSL_CTX_set_tlsext_servername_arg()
3770 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3772 openssl s_client has a new '-servername ...' option.
3774 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3775 '-key2 ...', '-servername_fatal' (subject to change). This allows
3776 testing the HostName extension for a specific single host name ('-cert'
3777 and '-key' remain fallbacks for handshakes without HostName
3778 negotiation). If the unrecognized_name alert has to be sent, this by
3779 default is a warning; it becomes fatal with the '-servername_fatal'
3782 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3784 *) Add AES and SSE2 assembly language support to VC++ build.
3787 *) Mitigate attack on final subtraction in Montgomery reduction.
3790 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3791 (which previously caused an internal error).
3794 *) Squeeze another 10% out of IGE mode when in != out.
3797 *) AES IGE mode speedup.
3798 [Dean Gaudet (Google)]
3800 *) Add the Korean symmetric 128-bit cipher SEED (see
3801 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3802 add SEED ciphersuites from RFC 4162:
3804 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3805 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3806 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3807 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3809 To minimize changes between patchlevels in the OpenSSL 0.9.8
3810 series, SEED remains excluded from compilation unless OpenSSL
3811 is configured with 'enable-seed'.
3812 [KISA, Bodo Moeller]
3814 *) Mitigate branch prediction attacks, which can be practical if a
3815 single processor is shared, allowing a spy process to extract
3816 information. For detailed background information, see
3817 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3818 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3819 and Necessary Software Countermeasures"). The core of the change
3820 are new versions BN_div_no_branch() and
3821 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3822 respectively, which are slower, but avoid the security-relevant
3823 conditional branches. These are automatically called by BN_div()
3824 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3825 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3826 remove a conditional branch.
3828 BN_FLG_CONSTTIME is the new name for the previous
3829 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3830 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3831 in the exponent causes BN_mod_exp_mont() to use the alternative
3832 implementation in BN_mod_exp_mont_consttime().) The old name
3833 remains as a deprecated alias.
3835 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3836 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3837 constant-time implementations for more than just exponentiation.
3838 Here too the old name is kept as a deprecated alias.
3840 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3841 the BN_BLINDING structure gets an independent copy of the
3842 modulus. This means that the previous "BIGNUM *m" argument to
3843 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3844 essentially becomes "const BIGNUM *m", although we can't actually
3845 change this in the header file before 0.9.9. It allows
3846 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3847 enable BN_FLG_CONSTTIME.
3849 [Matthew D Wood (Intel Corp)]
3851 *) In the SSL/TLS server implementation, be strict about session ID
3852 context matching (which matters if an application uses a single
3853 external cache for different purposes). Previously,
3854 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3855 set. This did ensure strict client verification, but meant that,
3856 with applications using a single external cache for quite
3857 different requirements, clients could circumvent ciphersuite
3858 restrictions for a given session ID context by starting a session
3859 in a different context.
3862 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3863 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3864 authentication-only ciphersuites.
3867 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3868 not complete and could lead to a possible single byte overflow
3869 (CVE-2007-5135) [Ben Laurie]
3871 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3873 *) Since AES128 and AES256 (and similarly Camellia128 and
3874 Camellia256) share a single mask bit in the logic of
3875 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3876 kludge to work properly if AES128 is available and AES256 isn't
3877 (or if Camellia128 is available and Camellia256 isn't).
3880 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3881 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3882 When a point or a seed is encoded in a BIT STRING, we need to
3883 prevent the removal of trailing zero bits to get the proper DER
3884 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3885 of a NamedBitList, for which trailing 0 bits need to be removed.)
3888 *) Have SSL/TLS server implementation tolerate "mismatched" record
3889 protocol version while receiving ClientHello even if the
3890 ClientHello is fragmented. (The server can't insist on the
3891 particular protocol version it has chosen before the ServerHello
3892 message has informed the client about his choice.)
3895 *) Add RFC 3779 support.
3896 [Rob Austein for ARIN, Ben Laurie]
3898 *) Load error codes if they are not already present instead of using a
3899 static variable. This allows them to be cleanly unloaded and reloaded.
3900 Improve header file function name parsing.
3903 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3904 or CAPABILITY handshake as required by RFCs.
3907 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3909 *) Introduce limits to prevent malicious keys being able to
3910 cause a denial of service. (CVE-2006-2940)
3911 [Steve Henson, Bodo Moeller]
3913 *) Fix ASN.1 parsing of certain invalid structures that can result
3914 in a denial of service. (CVE-2006-2937) [Steve Henson]
3916 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3917 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3919 *) Fix SSL client code which could crash if connecting to a
3920 malicious SSLv2 server. (CVE-2006-4343)
3921 [Tavis Ormandy and Will Drewry, Google Security Team]
3923 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3924 match only those. Before that, "AES256-SHA" would be interpreted
3925 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3926 the same strength classification in 0.9.7h) as we currently only
3927 have a single AES bit in the ciphersuite description bitmap.
3928 That change, however, also applied to ciphersuite strings such as
3929 "RC4-MD5" that intentionally matched multiple ciphersuites --
3930 namely, SSL 2.0 ciphersuites in addition to the more common ones
3931 from SSL 3.0/TLS 1.0.
3933 So we change the selection algorithm again: Naming an explicit
3934 ciphersuite selects this one ciphersuite, and any other similar
3935 ciphersuite (same bitmap) from *other* protocol versions.
3936 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3937 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3939 Since SSL 2.0 does not have any ciphersuites for which the
3940 128/256 bit distinction would be relevant, this works for now.
3941 The proper fix will be to use different bits for AES128 and
3942 AES256, which would have avoided the problems from the beginning;
3943 however, bits are scarce, so we can only do this in a new release
3944 (not just a patchlevel) when we can change the SSL_CIPHER
3945 definition to split the single 'unsigned long mask' bitmap into
3946 multiple values to extend the available space.
3950 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3952 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3953 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3955 *) Add AES IGE and biIGE modes.
3958 *) Change the Unix randomness entropy gathering to use poll() when
3959 possible instead of select(), since the latter has some
3960 undesirable limitations.
3961 [Darryl Miles via Richard Levitte and Bodo Moeller]
3963 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3964 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3965 cannot be implicitly activated as part of, e.g., the "AES" alias.
3966 However, please upgrade to OpenSSL 0.9.9[-dev] for
3967 non-experimental use of the ECC ciphersuites to get TLS extension
3968 support, which is required for curve and point format negotiation
3969 to avoid potential handshake problems.
3972 *) Disable rogue ciphersuites:
3974 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3975 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3976 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3978 The latter two were purportedly from
3979 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3982 Also deactivate the remaining ciphersuites from
3983 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3984 unofficial, and the ID has long expired.
3987 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
3988 dual-core machines) and other potential thread-safety issues.
3991 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3992 versions), which is now available for royalty-free use
3993 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3994 Also, add Camellia TLS ciphersuites from RFC 4132.
3996 To minimize changes between patchlevels in the OpenSSL 0.9.8
3997 series, Camellia remains excluded from compilation unless OpenSSL
3998 is configured with 'enable-camellia'.
4001 *) Disable the padding bug check when compression is in use. The padding
4002 bug check assumes the first packet is of even length, this is not
4003 necessarily true if compresssion is enabled and can result in false
4004 positives causing handshake failure. The actual bug test is ancient
4005 code so it is hoped that implementations will either have fixed it by
4006 now or any which still have the bug do not support compression.
4009 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4011 *) When applying a cipher rule check to see if string match is an explicit
4012 cipher suite and only match that one cipher suite if it is.
4015 *) Link in manifests for VC++ if needed.
4016 [Austin Ziegler <halostatue@gmail.com>]
4018 *) Update support for ECC-based TLS ciphersuites according to
4019 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4020 TLS extensions, which are supported starting with the 0.9.9
4021 branch, not in the OpenSSL 0.9.8 branch).
4024 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4025 opaque EVP_CIPHER_CTX handling.
4028 *) Fixes and enhancements to zlib compression code. We now only use
4029 "zlib1.dll" and use the default __cdecl calling convention on Win32
4030 to conform with the standards mentioned here:
4031 http://www.zlib.net/DLL_FAQ.txt
4032 Static zlib linking now works on Windows and the new --with-zlib-include
4033 --with-zlib-lib options to Configure can be used to supply the location
4034 of the headers and library. Gracefully handle case where zlib library
4038 *) Several fixes and enhancements to the OID generation code. The old code
4039 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4040 handle numbers larger than ULONG_MAX, truncated printing and had a
4041 non standard OBJ_obj2txt() behaviour.
4044 *) Add support for building of engines under engine/ as shared libraries
4045 under VC++ build system.
4048 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4049 Hopefully, we will not see any false combination of paths any more.
4052 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4054 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4055 (part of SSL_OP_ALL). This option used to disable the
4056 countermeasure against man-in-the-middle protocol-version
4057 rollback in the SSL 2.0 server implementation, which is a bad
4058 idea. (CVE-2005-2969)
4060 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4061 for Information Security, National Institute of Advanced Industrial
4062 Science and Technology [AIST], Japan)]
4064 *) Add two function to clear and return the verify parameter flags.
4067 *) Keep cipherlists sorted in the source instead of sorting them at
4068 runtime, thus removing the need for a lock.
4071 *) Avoid some small subgroup attacks in Diffie-Hellman.
4072 [Nick Mathewson and Ben Laurie]
4074 *) Add functions for well-known primes.
4077 *) Extended Windows CE support.
4078 [Satoshi Nakamura and Andy Polyakov]
4080 *) Initialize SSL_METHOD structures at compile time instead of during
4081 runtime, thus removing the need for a lock.
4084 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4085 attempting to decrypt each encrypted key in turn. Add support to
4089 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4091 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4094 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4097 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4098 key into the same file any more.
4101 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4104 *) Add -utf8 command line and config file option to 'ca'.
4105 [Stefan <stf@udoma.org]
4107 *) Removed the macro des_crypt(), as it seems to conflict with some
4108 libraries. Use DES_crypt().
4111 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4112 involves renaming the source and generated shared-libs for
4113 both. The engines will accept the corrected or legacy ids
4114 ('ncipher' and '4758_cca' respectively) when binding. NB,
4115 this only applies when building 'shared'.
4116 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4118 *) Add attribute functions to EVP_PKEY structure. Modify
4119 PKCS12_create() to recognize a CSP name attribute and
4120 use it. Make -CSP option work again in pkcs12 utility.
4123 *) Add new functionality to the bn blinding code:
4124 - automatic re-creation of the BN_BLINDING parameters after
4125 a fixed number of uses (currently 32)
4126 - add new function for parameter creation
4127 - introduce flags to control the update behaviour of the
4128 BN_BLINDING parameters
4129 - hide BN_BLINDING structure
4130 Add a second BN_BLINDING slot to the RSA structure to improve
4131 performance when a single RSA object is shared among several
4135 *) Add support for DTLS.
4136 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4138 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4139 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4142 *) Remove buggy and incomplete DH cert support from
4143 ssl/ssl_rsa.c and ssl/s3_both.c
4146 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4147 the apps/openssl applications.
4150 *) Compile clean with "-Wall -Wmissing-prototypes
4151 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4152 DEBUG_SAFESTACK must also be set.
4155 *) Change ./Configure so that certain algorithms can be disabled by default.
4156 The new counterpiece to "no-xxx" is "enable-xxx".
4158 The patented RC5 and MDC2 algorithms will now be disabled unless
4159 "enable-rc5" and "enable-mdc2", respectively, are specified.
4161 (IDEA remains enabled despite being patented. This is because IDEA
4162 is frequently required for interoperability, and there is no license
4163 fee for non-commercial use. As before, "no-idea" can be used to
4164 avoid this algorithm.)
4168 *) Add processing of proxy certificates (see RFC 3820). This work was
4169 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4170 EGEE (Enabling Grids for E-science in Europe).
4173 *) RC4 performance overhaul on modern architectures/implementations, such
4174 as Intel P4, IA-64 and AMD64.
4177 *) New utility extract-section.pl. This can be used specify an alternative
4178 section number in a pod file instead of having to treat each file as
4179 a separate case in Makefile. This can be done by adding two lines to the
4182 =for comment openssl_section:XXX
4184 The blank line is mandatory.
4188 *) New arguments -certform, -keyform and -pass for s_client and s_server
4189 to allow alternative format key and certificate files and passphrase
4193 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4194 update associated structures and add various utility functions.
4196 Add new policy related verify parameters, include policy checking in
4197 standard verify code. Enhance 'smime' application with extra parameters
4198 to support policy checking and print out.
4201 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4202 Nehemiah processors. These extensions support AES encryption in hardware
4203 as well as RNG (though RNG support is currently disabled).
4204 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4206 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4209 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4210 [Andy Polyakov and a number of other people]
4212 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4213 implementation contributed by IBM.
4214 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4216 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4217 exponent rather than 'unsigned long'. There is a corresponding change to
4218 the new 'rsa_keygen' element of the RSA_METHOD structure.
4219 [Jelte Jansen, Geoff Thorpe]
4221 *) Functionality for creating the initial serial number file is now
4222 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4224 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4225 number file to 1, which is bound to cause problems. To avoid
4226 the problems while respecting compatibility between different 0.9.7
4227 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4228 CA.pl for serial number initialization. With the new release 0.9.8,
4229 we can fix the problem directly in the 'ca' utility.)
4232 *) Reduced header interdepencies by declaring more opaque objects in
4233 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4234 give fewer recursive includes, which could break lazy source code - so
4235 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4236 developers should define this symbol when building and using openssl to
4237 ensure they track the recommended behaviour, interfaces, [etc], but
4238 backwards-compatible behaviour prevails when this isn't defined.
4241 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4244 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4245 This will generate a random key of the appropriate length based on the
4246 cipher context. The EVP_CIPHER can provide its own random key generation
4247 routine to support keys of a specific form. This is used in the des and
4248 3des routines to generate a key of the correct parity. Update S/MIME
4249 code to use new functions and hence generate correct parity DES keys.
4250 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4251 valid (weak or incorrect parity).
4254 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4255 as looking them up. This is useful when the verified structure may contain
4256 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4257 present unless the new PKCS7_NO_CRL flag is asserted.
4260 *) Extend ASN1 oid configuration module. It now additionally accepts the
4263 shortName = some long name, 1.2.3.4
4266 *) Reimplemented the BN_CTX implementation. There is now no more static
4267 limitation on the number of variables it can handle nor the depth of the
4268 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4269 information can now expand as required, and rather than having a single
4270 static array of bignums, BN_CTX now uses a linked-list of such arrays
4271 allowing it to expand on demand whilst maintaining the usefulness of
4272 BN_CTX's "bundling".
4275 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4276 to allow all RSA operations to function using a single BN_CTX.
4279 *) Preliminary support for certificate policy evaluation and checking. This
4280 is initially intended to pass the tests outlined in "Conformance Testing
4281 of Relying Party Client Certificate Path Processing Logic" v1.07.
4284 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4285 remained unused and not that useful. A variety of other little bignum
4286 tweaks and fixes have also been made continuing on from the audit (see
4290 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4291 associated ASN1, EVP and SSL functions and old ASN1 macros.
4294 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4295 and this should never fail. So the return value from the use of
4296 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4297 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4300 *) BN_CTX_get() should return zero-valued bignums, providing the same
4301 initialised value as BN_new().
4302 [Geoff Thorpe, suggested by Ulf Möller]
4304 *) Support for inhibitAnyPolicy certificate extension.
4307 *) An audit of the BIGNUM code is underway, for which debugging code is
4308 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4309 is considered valid when processing BIGNUMs, and causes execution to
4310 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4311 further steps are taken to deliberately pollute unused data in BIGNUM
4312 structures to try and expose faulty code further on. For now, openssl will
4313 (in its default mode of operation) continue to tolerate the inconsistent
4314 forms that it has tolerated in the past, but authors and packagers should
4315 consider trying openssl and their own applications when compiled with
4316 these debugging symbols defined. It will help highlight potential bugs in
4317 their own code, and will improve the test coverage for OpenSSL itself. At
4318 some point, these tighter rules will become openssl's default to improve
4319 maintainability, though the assert()s and other overheads will remain only
4320 in debugging configurations. See bn.h for more details.
4321 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4323 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4324 that can only be obtained through BN_CTX_new() (which implicitly
4325 initialises it). The presence of this function only made it possible
4326 to overwrite an existing structure (and cause memory leaks).
4329 *) Because of the callback-based approach for implementing LHASH as a
4330 template type, lh_insert() adds opaque objects to hash-tables and
4331 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4332 to clean up those corresponding objects before destroying the hash table
4333 (and losing the object pointers). So some over-zealous constifications in
4334 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4335 objects as "const" and the lh_doall[_arg] callback wrappers are not
4336 prototyped to have "const" restrictions on the object pointers they are
4337 given (and so aren't required to cast them away any more).
4340 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4341 (speed) prefers to use its own implementation. The two implementations
4342 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4343 its object type properly exposed (MS_TM) instead of casting to/from "char
4344 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4345 aren't necessarily the greatest nomenclatures - but this is what was used
4346 internally to the implementation so I've used that for now.
4349 *) Ensure that deprecated functions do not get compiled when
4350 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4351 the self-tests were still using deprecated key-generation functions so
4352 these have been updated also.
4355 *) Reorganise PKCS#7 code to separate the digest location functionality
4356 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4357 New function PKCS7_set_digest() to set the digest type for PKCS#7
4358 digestedData type. Add additional code to correctly generate the
4359 digestedData type and add support for this type in PKCS7 initialization
4363 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4364 structure of type "other".
4367 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4368 sure the loop does correctly stop and breaking ("division by zero")
4369 modulus operations are not performed. The (pre-generated) prime
4370 table crypto/bn/bn_prime.h was already correct, but it could not be
4371 re-generated on some platforms because of the "division by zero"
4372 situation in the script.
4373 [Ralf S. Engelschall]
4375 *) Update support for ECC-based TLS ciphersuites according to
4376 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4377 SHA-1 now is only used for "small" curves (where the
4378 representation of a field element takes up to 24 bytes); for
4379 larger curves, the field element resulting from ECDH is directly
4380 used as premaster secret.
4381 [Douglas Stebila (Sun Microsystems Laboratories)]
4383 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4384 curve secp160r1 to the tests.
4385 [Douglas Stebila (Sun Microsystems Laboratories)]
4387 *) Add the possibility to load symbols globally with DSO.
4388 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4390 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4391 control of the error stack.
4394 *) Add support for STORE in ENGINE.
4397 *) Add the STORE type. The intention is to provide a common interface
4398 to certificate and key stores, be they simple file-based stores, or
4399 HSM-type store, or LDAP stores, or...
4400 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4403 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4404 pass a list of arguments to any function as well as provide a way
4405 for a function to pass data back to the caller.
4408 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4409 works like BUF_strdup() but can be used to duplicate a portion of
4410 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4414 *) Add the function sk_find_ex() which works like sk_find(), but will
4415 return an index to an element even if an exact match couldn't be
4416 found. The index is guaranteed to point at the element where the
4417 searched-for key would be inserted to preserve sorting order.
4420 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4421 takes an extra flags argument for optional functionality. Currently,
4422 the following flags are defined:
4424 OBJ_BSEARCH_VALUE_ON_NOMATCH
4425 This one gets OBJ_bsearch_ex() to return a pointer to the first
4426 element where the comparing function returns a negative or zero
4429 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4430 This one gets OBJ_bsearch_ex() to return a pointer to the first
4431 element where the comparing function returns zero. This is useful
4432 if there are more than one element where the comparing function
4436 *) Make it possible to create self-signed certificates with 'openssl ca'
4437 in such a way that the self-signed certificate becomes part of the
4438 CA database and uses the same mechanisms for serial number generation
4439 as all other certificate signing. The new flag '-selfsign' enables
4440 this functionality. Adapt CA.sh and CA.pl.in.
4443 *) Add functionality to check the public key of a certificate request
4444 against a given private. This is useful to check that a certificate
4445 request can be signed by that key (self-signing).
4448 *) Make it possible to have multiple active certificates with the same
4449 subject in the CA index file. This is done only if the keyword
4450 'unique_subject' is set to 'no' in the main CA section (default
4451 if 'CA_default') of the configuration file. The value is saved
4452 with the database itself in a separate index attribute file,
4453 named like the index file with '.attr' appended to the name.
4456 *) Generate muti valued AVAs using '+' notation in config files for
4460 *) Support for nameConstraints certificate extension.
4463 *) Support for policyConstraints certificate extension.
4466 *) Support for policyMappings certificate extension.
4469 *) Make sure the default DSA_METHOD implementation only uses its
4470 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4471 and change its own handlers to be NULL so as to remove unnecessary
4472 indirection. This lets alternative implementations fallback to the
4473 default implementation more easily.
4476 *) Support for directoryName in GeneralName related extensions
4480 *) Make it possible to link applications using Makefile.shared.
4481 Make that possible even when linking against static libraries!
4484 *) Support for single pass processing for S/MIME signing. This now
4485 means that S/MIME signing can be done from a pipe, in addition
4486 cleartext signing (multipart/signed type) is effectively streaming
4487 and the signed data does not need to be all held in memory.
4489 This is done with a new flag PKCS7_STREAM. When this flag is set
4490 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4491 is done after the data is output (and digests calculated) in
4492 SMIME_write_PKCS7().
4495 *) Add full support for -rpath/-R, both in shared libraries and
4496 applications, at least on the platforms where it's known how
4500 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4501 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4502 will now compute a table of multiples of the generator that
4503 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4504 faster (notably in the case of a single point multiplication,
4505 scalar * generator).
4506 [Nils Larsch, Bodo Moeller]
4508 *) IPv6 support for certificate extensions. The various extensions
4509 which use the IP:a.b.c.d can now take IPv6 addresses using the
4510 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4514 *) Added an ENGINE that implements RSA by performing private key
4515 exponentiations with the GMP library. The conversions to and from
4516 GMP's mpz_t format aren't optimised nor are any montgomery forms
4517 cached, and on x86 it appears OpenSSL's own performance has caught up.
4518 However there are likely to be other architectures where GMP could
4519 provide a boost. This ENGINE is not built in by default, but it can be
4520 specified at Configure time and should be accompanied by the necessary
4521 linker additions, eg;
4522 ./config -DOPENSSL_USE_GMP -lgmp
4525 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4526 testing availability of engines with "-t" - the old behaviour is
4527 produced by increasing the feature's verbosity with "-tt".
4530 *) ECDSA routines: under certain error conditions uninitialized BN objects
4531 could be freed. Solution: make sure initialization is performed early
4532 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4536 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4537 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4538 software implementations. For DSA and DH, parameter generation can
4539 also be overriden by providing the appropriate method callbacks.
4542 *) Change the "progress" mechanism used in key-generation and
4543 primality testing to functions that take a new BN_GENCB pointer in
4544 place of callback/argument pairs. The new API functions have "_ex"
4545 postfixes and the older functions are reimplemented as wrappers for
4546 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4547 declarations of the old functions to help (graceful) attempts to
4548 migrate to the new functions. Also, the new key-generation API
4549 functions operate on a caller-supplied key-structure and return
4550 success/failure rather than returning a key or NULL - this is to
4551 help make "keygen" another member function of RSA_METHOD etc.
4553 Example for using the new callback interface:
4555 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4559 BN_GENCB_set(&my_cb, my_callback, my_arg);
4561 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4562 /* For the meaning of a, b in calls to my_callback(), see the
4563 * documentation of the function that calls the callback.
4564 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4565 * my_callback should return 1 if it wants BN_is_prime_ex()
4566 * to continue, or 0 to stop.
4571 *) Change the ZLIB compression method to be stateful, and make it
4572 available to TLS with the number defined in
4573 draft-ietf-tls-compression-04.txt.
4576 *) Add the ASN.1 structures and functions for CertificatePair, which
4577 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4579 CertificatePair ::= SEQUENCE {
4580 forward [0] Certificate OPTIONAL,
4581 reverse [1] Certificate OPTIONAL,
4582 -- at least one of the pair shall be present -- }
4584 Also implement the PEM functions to read and write certificate
4585 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4587 This needed to be defined, mostly for the sake of the LDAP
4588 attribute crossCertificatePair, but may prove useful elsewhere as
4592 *) Make it possible to inhibit symlinking of shared libraries in
4593 Makefile.shared, for Cygwin's sake.
4596 *) Extend the BIGNUM API by creating a function
4597 void BN_set_negative(BIGNUM *a, int neg);
4598 and a macro that behave like
4599 int BN_is_negative(const BIGNUM *a);
4601 to avoid the need to access 'a->neg' directly in applications.
4604 *) Implement fast modular reduction for pseudo-Mersenne primes
4605 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4606 EC_GROUP_new_curve_GFp() will now automatically use this
4608 [Nils Larsch <nla@trustcenter.de>]
4610 *) Add new lock type (CRYPTO_LOCK_BN).
4613 *) Change the ENGINE framework to automatically load engines
4614 dynamically from specific directories unless they could be
4615 found to already be built in or loaded. Move all the
4616 current engines except for the cryptodev one to a new
4618 The engines in engines/ are built as shared libraries if
4619 the "shared" options was given to ./Configure or ./config.
4620 Otherwise, they are inserted in libcrypto.a.
4621 /usr/local/ssl/engines is the default directory for dynamic
4622 engines, but that can be overriden at configure time through
4623 the usual use of --prefix and/or --openssldir, and at run
4624 time with the environment variable OPENSSL_ENGINES.
4625 [Geoff Thorpe and Richard Levitte]
4627 *) Add Makefile.shared, a helper makefile to build shared
4628 libraries. Adapt Makefile.org.
4631 *) Add version info to Win32 DLLs.
4632 [Peter 'Luna' Runestig" <peter@runestig.com>]
4634 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4635 can be added using this API to created arbitrary PKCS#12
4636 files while avoiding the low level API.
4638 New options to PKCS12_create(), key or cert can be NULL and
4639 will then be omitted from the output file. The encryption
4640 algorithm NIDs can be set to -1 for no encryption, the mac
4641 iteration count can be set to 0 to omit the mac.
4643 Enhance pkcs12 utility by making the -nokeys and -nocerts
4644 options work when creating a PKCS#12 file. New option -nomac
4645 to omit the mac, NONE can be set for an encryption algorithm.
4646 New code is modified to use the enhanced PKCS12_create()
4647 instead of the low level API.
4650 *) Extend ASN1 encoder to support indefinite length constructed
4651 encoding. This can output sequences tags and octet strings in
4652 this form. Modify pk7_asn1.c to support indefinite length
4653 encoding. This is experimental and needs additional code to
4654 be useful, such as an ASN1 bio and some enhanced streaming
4657 Extend template encode functionality so that tagging is passed
4658 down to the template encoder.
4661 *) Let 'openssl req' fail if an argument to '-newkey' is not
4662 recognized instead of using RSA as a default.
4665 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4666 As these are not official, they are not included in "ALL";
4667 the "ECCdraft" ciphersuite group alias can be used to select them.
4668 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4670 *) Add ECDH engine support.
4671 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4673 *) Add ECDH in new directory crypto/ecdh/.
4674 [Douglas Stebila (Sun Microsystems Laboratories)]
4676 *) Let BN_rand_range() abort with an error after 100 iterations
4677 without success (which indicates a broken PRNG).
4680 *) Change BN_mod_sqrt() so that it verifies that the input value
4681 is really the square of the return value. (Previously,
4682 BN_mod_sqrt would show GIGO behaviour.)
4685 *) Add named elliptic curves over binary fields from X9.62, SECG,
4686 and WAP/WTLS; add OIDs that were still missing.
4688 [Sheueling Chang Shantz and Douglas Stebila
4689 (Sun Microsystems Laboratories)]
4691 *) Extend the EC library for elliptic curves over binary fields
4692 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4695 EC_GF2m_simple_method
4699 EC_GROUP_new_curve_GF2m
4700 EC_GROUP_set_curve_GF2m
4701 EC_GROUP_get_curve_GF2m
4702 EC_POINT_set_affine_coordinates_GF2m
4703 EC_POINT_get_affine_coordinates_GF2m
4704 EC_POINT_set_compressed_coordinates_GF2m
4706 Point compression for binary fields is disabled by default for
4707 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4710 As binary polynomials are represented as BIGNUMs, various members
4711 of the EC_GROUP and EC_POINT data structures can be shared
4712 between the implementations for prime fields and binary fields;
4713 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4714 are essentially identical to their ..._GFp counterparts.
4715 (For simplicity, the '..._GFp' prefix has been dropped from
4716 various internal method names.)
4718 An internal 'field_div' method (similar to 'field_mul' and
4719 'field_sqr') has been added; this is used only for binary fields.
4721 [Sheueling Chang Shantz and Douglas Stebila
4722 (Sun Microsystems Laboratories)]
4724 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4725 through methods ('mul', 'precompute_mult').
4727 The generic implementations (now internally called 'ec_wNAF_mul'
4728 and 'ec_wNAF_precomputed_mult') remain the default if these
4729 methods are undefined.
4731 [Sheueling Chang Shantz and Douglas Stebila
4732 (Sun Microsystems Laboratories)]
4734 *) New function EC_GROUP_get_degree, which is defined through
4735 EC_METHOD. For curves over prime fields, this returns the bit
4736 length of the modulus.
4738 [Sheueling Chang Shantz and Douglas Stebila
4739 (Sun Microsystems Laboratories)]
4741 *) New functions EC_GROUP_dup, EC_POINT_dup.
4742 (These simply call ..._new and ..._copy).
4744 [Sheueling Chang Shantz and Douglas Stebila
4745 (Sun Microsystems Laboratories)]
4747 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4748 Polynomials are represented as BIGNUMs (where the sign bit is not
4749 used) in the following functions [macros]:
4752 BN_GF2m_sub [= BN_GF2m_add]
4753 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4754 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4755 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4757 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4758 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4759 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4760 BN_GF2m_cmp [= BN_ucmp]
4762 (Note that only the 'mod' functions are actually for fields GF(2^m).
4763 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4765 For some functions, an the irreducible polynomial defining a
4766 field can be given as an 'unsigned int[]' with strictly
4767 decreasing elements giving the indices of those bits that are set;
4768 i.e., p[] represents the polynomial
4769 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4771 p[0] > p[1] > ... > p[k] = 0.
4772 This applies to the following functions:
4777 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4778 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4780 BN_GF2m_mod_sqrt_arr
4781 BN_GF2m_mod_solve_quad_arr
4785 Conversion can be performed by the following functions:
4790 bntest.c has additional tests for binary polynomial arithmetic.
4792 Two implementations for BN_GF2m_mod_div() are available.
4793 The default algorithm simply uses BN_GF2m_mod_inv() and
4794 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4795 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4796 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4798 [Sheueling Chang Shantz and Douglas Stebila
4799 (Sun Microsystems Laboratories)]
4801 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4802 functionality is disabled at compile-time.
4803 [Douglas Stebila <douglas.stebila@sun.com>]
4805 *) Change default behaviour of 'openssl asn1parse' so that more
4806 information is visible when viewing, e.g., a certificate:
4808 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4809 mode the content of non-printable OCTET STRINGs is output in a
4810 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4811 avoid the appearance of a printable string.
4812 [Nils Larsch <nla@trustcenter.de>]
4814 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4816 EC_GROUP_set_asn1_flag()
4817 EC_GROUP_get_asn1_flag()
4818 EC_GROUP_set_point_conversion_form()
4819 EC_GROUP_get_point_conversion_form()
4820 These control ASN1 encoding details:
4821 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4822 has been set to OPENSSL_EC_NAMED_CURVE.
4823 - Points are encoded in uncompressed form by default; options for
4824 asn1_for are as for point2oct, namely
4825 POINT_CONVERSION_COMPRESSED
4826 POINT_CONVERSION_UNCOMPRESSED
4827 POINT_CONVERSION_HYBRID
4829 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4832 EC_GROUP_get0_seed()
4833 EC_GROUP_get_seed_len()
4834 This is used only for ASN1 purposes (so far).
4835 [Nils Larsch <nla@trustcenter.de>]
4837 *) Add 'field_type' member to EC_METHOD, which holds the NID
4838 of the appropriate field type OID. The new function
4839 EC_METHOD_get_field_type() returns this value.
4840 [Nils Larsch <nla@trustcenter.de>]
4845 EC_POINT_point2hex()
4846 EC_POINT_hex2point()
4847 providing useful interfaces to EC_POINT_point2oct() and
4848 EC_POINT_oct2point().
4849 [Nils Larsch <nla@trustcenter.de>]
4851 *) Change internals of the EC library so that the functions
4852 EC_GROUP_set_generator()
4853 EC_GROUP_get_generator()
4854 EC_GROUP_get_order()
4855 EC_GROUP_get_cofactor()
4856 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4857 to methods, which would lead to unnecessary code duplication when
4858 adding different types of curves.
4859 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4861 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4862 arithmetic, and such that modified wNAFs are generated
4863 (which avoid length expansion in many cases).
4866 *) Add a function EC_GROUP_check_discriminant() (defined via
4867 EC_METHOD) that verifies that the curve discriminant is non-zero.
4869 Add a function EC_GROUP_check() that makes some sanity tests
4870 on a EC_GROUP, its generator and order. This includes
4871 EC_GROUP_check_discriminant().
4872 [Nils Larsch <nla@trustcenter.de>]
4874 *) Add ECDSA in new directory crypto/ecdsa/.
4876 Add applications 'openssl ecparam' and 'openssl ecdsa'
4877 (these are based on 'openssl dsaparam' and 'openssl dsa').
4879 ECDSA support is also included in various other files across the
4880 library. Most notably,
4881 - 'openssl req' now has a '-newkey ecdsa:file' option;
4882 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4883 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4884 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4885 them suitable for ECDSA where domain parameters must be
4886 extracted before the specific public key;
4887 - ECDSA engine support has been added.
4888 [Nils Larsch <nla@trustcenter.de>]
4890 *) Include some named elliptic curves, and add OIDs from X9.62,
4891 SECG, and WAP/WTLS. Each curve can be obtained from the new
4893 EC_GROUP_new_by_curve_name(),
4894 and the list of available named curves can be obtained with
4895 EC_get_builtin_curves().
4896 Also add a 'curve_name' member to EC_GROUP objects, which can be
4898 EC_GROUP_set_curve_name()
4899 EC_GROUP_get_curve_name()
4900 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4902 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4903 was actually never needed) and in BN_mul(). The removal in BN_mul()
4904 required a small change in bn_mul_part_recursive() and the addition
4905 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4906 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4907 bn_sub_words() and bn_add_words() except they take arrays with
4911 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4913 *) Cleanse PEM buffers before freeing them since they may contain
4915 [Benjamin Bennett <ben@psc.edu>]
4917 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4918 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4919 authentication-only ciphersuites.
4922 *) Since AES128 and AES256 share a single mask bit in the logic of
4923 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4924 kludge to work properly if AES128 is available and AES256 isn't.
4927 *) Expand security boundary to match 1.1.1 module.
4930 *) Remove redundant features: hash file source, editing of test vectors
4931 modify fipsld to use external fips_premain.c signature.
4934 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4935 run algorithm test programs.
4938 *) Make algorithm test programs more tolerant of whitespace.
4941 *) Have SSL/TLS server implementation tolerate "mismatched" record
4942 protocol version while receiving ClientHello even if the
4943 ClientHello is fragmented. (The server can't insist on the
4944 particular protocol version it has chosen before the ServerHello
4945 message has informed the client about his choice.)
4948 *) Load error codes if they are not already present instead of using a
4949 static variable. This allows them to be cleanly unloaded and reloaded.
4952 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4954 *) Introduce limits to prevent malicious keys being able to
4955 cause a denial of service. (CVE-2006-2940)
4956 [Steve Henson, Bodo Moeller]
4958 *) Fix ASN.1 parsing of certain invalid structures that can result
4959 in a denial of service. (CVE-2006-2937) [Steve Henson]
4961 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4962 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4964 *) Fix SSL client code which could crash if connecting to a
4965 malicious SSLv2 server. (CVE-2006-4343)
4966 [Tavis Ormandy and Will Drewry, Google Security Team]
4968 *) Change ciphersuite string processing so that an explicit
4969 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4970 will no longer include "AES128-SHA"), and any other similar
4971 ciphersuite (same bitmap) from *other* protocol versions (so that
4972 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4973 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4974 changes from 0.9.8b and 0.9.8d.
4977 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4979 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4980 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4982 *) Change the Unix randomness entropy gathering to use poll() when
4983 possible instead of select(), since the latter has some
4984 undesirable limitations.
4985 [Darryl Miles via Richard Levitte and Bodo Moeller]
4987 *) Disable rogue ciphersuites:
4989 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4990 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4991 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4993 The latter two were purportedly from
4994 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4997 Also deactive the remaining ciphersuites from
4998 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4999 unofficial, and the ID has long expired.
5002 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5003 dual-core machines) and other potential thread-safety issues.
5006 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5008 *) Adapt fipsld and the build system to link against the validated FIPS
5009 module in FIPS mode.
5012 *) Fixes for VC++ 2005 build under Windows.
5015 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5016 from a Windows bash shell such as MSYS. It is autodetected from the
5017 "config" script when run from a VC++ environment. Modify standard VC++
5018 build to use fipscanister.o from the GNU make build.
5021 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5023 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5024 The value now differs depending on if you build for FIPS or not.
5025 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5026 safely run with a non-FIPSed libcrypto, as it may crash because of
5027 the difference induced by this change.
5030 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5032 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5033 (part of SSL_OP_ALL). This option used to disable the
5034 countermeasure against man-in-the-middle protocol-version
5035 rollback in the SSL 2.0 server implementation, which is a bad
5036 idea. (CVE-2005-2969)
5038 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5039 for Information Security, National Institute of Advanced Industrial
5040 Science and Technology [AIST], Japan)]
5042 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5043 mainly for FIPS compliance and not fully integrated at this stage.
5046 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5047 the exponentiation using a fixed-length exponent. (Otherwise,
5048 the information leaked through timing could expose the secret key
5049 after many signatures; cf. Bleichenbacher's attack on DSA with
5053 *) Make a new fixed-window mod_exp implementation the default for
5054 RSA, DSA, and DH private-key operations so that the sequence of
5055 squares and multiplies and the memory access pattern are
5056 independent of the particular secret key. This will mitigate
5057 cache-timing and potential related attacks.
5059 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5060 and this is automatically used by BN_mod_exp_mont() if the new flag
5061 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5062 will use this BN flag for private exponents unless the flag
5063 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5064 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5066 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5068 *) Change the client implementation for SSLv23_method() and
5069 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5070 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5071 (Previously, the SSL 2.0 backwards compatible Client Hello
5072 message format would be used even with SSL_OP_NO_SSLv2.)
5075 *) Add support for smime-type MIME parameter in S/MIME messages which some
5079 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5080 a threadsafe manner. Modify rsa code to use new function and add calls
5081 to dsa and dh code (which had race conditions before).
5084 *) Include the fixed error library code in the C error file definitions
5085 instead of fixing them up at runtime. This keeps the error code
5086 structures constant.
5089 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5091 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5094 *) Fixes for newer kerberos headers. NB: the casts are needed because
5095 the 'length' field is signed on one version and unsigned on another
5096 with no (?) obvious way to tell the difference, without these VC++
5097 complains. Also the "definition" of FAR (blank) is no longer included
5098 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5099 some needed definitions.
5102 *) Undo Cygwin change.
5105 *) Added support for proxy certificates according to RFC 3820.
5106 Because they may be a security thread to unaware applications,
5107 they must be explicitly allowed in run-time. See
5108 docs/HOWTO/proxy_certificates.txt for further information.
5111 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5113 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5114 server and client random values. Previously
5115 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5116 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5118 This change has negligible security impact because:
5120 1. Server and client random values still have 24 bytes of pseudo random
5123 2. Server and client random values are sent in the clear in the initial
5126 3. The master secret is derived using the premaster secret (48 bytes in
5127 size for static RSA ciphersuites) as well as client server and random
5130 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5133 [Stephen Henson, reported by UK NISCC]
5135 *) Use Windows randomness collection on Cygwin.
5138 *) Fix hang in EGD/PRNGD query when communication socket is closed
5139 prematurely by EGD/PRNGD.
5140 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5142 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5145 *) Back-port of selected performance improvements from development
5146 branch, as well as improved support for PowerPC platforms.
5149 *) Add lots of checks for memory allocation failure, error codes to indicate
5150 failure and freeing up memory if a failure occurs.
5151 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5153 *) Add new -passin argument to dgst.
5156 *) Perform some character comparisons of different types in X509_NAME_cmp:
5157 this is needed for some certificates that reencode DNs into UTF8Strings
5158 (in violation of RFC3280) and can't or wont issue name rollover
5162 *) Make an explicit check during certificate validation to see that
5163 the CA setting in each certificate on the chain is correct. As a
5164 side effect always do the following basic checks on extensions,
5165 not just when there's an associated purpose to the check:
5167 - if there is an unhandled critical extension (unless the user
5168 has chosen to ignore this fault)
5169 - if the path length has been exceeded (if one is set at all)
5170 - that certain extensions fit the associated purpose (if one has
5174 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5176 *) Avoid a race condition when CRLs are checked in a multi threaded
5177 environment. This would happen due to the reordering of the revoked
5178 entries during signature checking and serial number lookup. Now the
5179 encoding is cached and the serial number sort performed under a lock.
5180 Add new STACK function sk_is_sorted().
5183 *) Add Delta CRL to the extension code.
5186 *) Various fixes to s3_pkt.c so alerts are sent properly.
5187 [David Holmes <d.holmes@f5.com>]
5189 *) Reduce the chances of duplicate issuer name and serial numbers (in
5190 violation of RFC3280) using the OpenSSL certificate creation utilities.
5191 This is done by creating a random 64 bit value for the initial serial
5192 number when a serial number file is created or when a self signed
5193 certificate is created using 'openssl req -x509'. The initial serial
5194 number file is created using 'openssl x509 -next_serial' in CA.pl
5195 rather than being initialized to 1.
5198 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5200 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5201 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5202 [Joe Orton, Steve Henson]
5204 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5206 [Joe Orton, Steve Henson]
5208 *) Make it possible to have multiple active certificates with the same
5209 subject in the CA index file. This is done only if the keyword
5210 'unique_subject' is set to 'no' in the main CA section (default
5211 if 'CA_default') of the configuration file. The value is saved
5212 with the database itself in a separate index attribute file,
5213 named like the index file with '.attr' appended to the name.
5216 *) X509 verify fixes. Disable broken certificate workarounds when
5217 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5218 keyUsage extension present. Don't accept CRLs with unhandled critical
5219 extensions: since verify currently doesn't process CRL extensions this
5220 rejects a CRL with *any* critical extensions. Add new verify error codes
5224 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5225 A clarification of RFC2560 will require the use of OCTET STRINGs and
5226 some implementations cannot handle the current raw format. Since OpenSSL
5227 copies and compares OCSP nonces as opaque blobs without any attempt at
5228 parsing them this should not create any compatibility issues.
5231 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5232 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5233 this HMAC (and other) operations are several times slower than OpenSSL
5237 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5238 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5240 *) Use the correct content when signing type "other".
5243 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5245 *) Fix various bugs revealed by running the NISCC test suite:
5247 Stop out of bounds reads in the ASN1 code when presented with
5248 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5250 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5252 If verify callback ignores invalid public key errors don't try to check
5253 certificate signature with the NULL public key.
5257 *) New -ignore_err option in ocsp application to stop the server
5258 exiting on the first error in a request.
5261 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5262 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5266 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5267 extra data after the compression methods not only for TLS 1.0
5268 but also for SSL 3.0 (as required by the specification).
5269 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5271 *) Change X509_certificate_type() to mark the key as exported/exportable
5272 when it's 512 *bits* long, not 512 bytes.
5275 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5276 blocks during encryption.
5279 *) Various fixes to base64 BIO and non blocking I/O. On write
5280 flushes were not handled properly if the BIO retried. On read
5281 data was not being buffered properly and had various logic bugs.
5282 This also affects blocking I/O when the data being decoded is a
5286 *) Various S/MIME bugfixes and compatibility changes:
5287 output correct application/pkcs7 MIME type if
5288 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5289 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5290 of files as .eml work). Correctly handle very long lines in MIME
5294 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5296 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5297 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5298 a protocol version number mismatch like a decryption error
5299 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5302 *) Turn on RSA blinding by default in the default implementation
5303 to avoid a timing attack. Applications that don't want it can call
5304 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5305 They would be ill-advised to do so in most cases.
5306 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5308 *) Change RSA blinding code so that it works when the PRNG is not
5309 seeded (in this case, the secret RSA exponent is abused as
5310 an unpredictable seed -- if it is not unpredictable, there
5311 is no point in blinding anyway). Make RSA blinding thread-safe
5312 by remembering the creator's thread ID in rsa->blinding and
5313 having all other threads use local one-time blinding factors
5314 (this requires more computation than sharing rsa->blinding, but
5315 avoids excessive locking; and if an RSA object is not shared
5316 between threads, blinding will still be very fast).
5319 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5320 ENGINE as defaults for all supported algorithms irrespective of
5321 the 'flags' parameter. 'flags' is now honoured, so applications
5322 should make sure they are passing it correctly.
5325 *) Target "mingw" now allows native Windows code to be generated in
5326 the Cygwin environment as well as with the MinGW compiler.
5329 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5331 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5332 via timing by performing a MAC computation even if incorrect
5333 block cipher padding has been found. This is a countermeasure
5334 against active attacks where the attacker has to distinguish
5335 between bad padding and a MAC verification error. (CVE-2003-0078)
5337 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5338 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5339 Martin Vuagnoux (EPFL, Ilion)]
5341 *) Make the no-err option work as intended. The intention with no-err
5342 is not to have the whole error stack handling routines removed from
5343 libcrypto, it's only intended to remove all the function name and
5344 reason texts, thereby removing some of the footprint that may not
5345 be interesting if those errors aren't displayed anyway.
5347 NOTE: it's still possible for any application or module to have it's
5348 own set of error texts inserted. The routines are there, just not
5349 used by default when no-err is given.
5352 *) Add support for FreeBSD on IA64.
5353 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5355 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5356 Kerberos function mit_des_cbc_cksum(). Before this change,
5357 the value returned by DES_cbc_cksum() was like the one from
5358 mit_des_cbc_cksum(), except the bytes were swapped.
5359 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5361 *) Allow an application to disable the automatic SSL chain building.
5362 Before this a rather primitive chain build was always performed in
5363 ssl3_output_cert_chain(): an application had no way to send the
5364 correct chain if the automatic operation produced an incorrect result.
5366 Now the chain builder is disabled if either:
5368 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5370 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5372 The reasoning behind this is that an application would not want the
5373 auto chain building to take place if extra chain certificates are
5374 present and it might also want a means of sending no additional
5375 certificates (for example the chain has two certificates and the
5379 *) Add the possibility to build without the ENGINE framework.
5380 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5382 *) Under Win32 gmtime() can return NULL: check return value in
5383 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5386 *) DSA routines: under certain error conditions uninitialized BN objects
5387 could be freed. Solution: make sure initialization is performed early
5388 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5389 Nils Larsch <nla@trustcenter.de> via PR#459)
5392 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5393 checked on reconnect on the client side, therefore session resumption
5394 could still fail with a "ssl session id is different" error. This
5395 behaviour is masked when SSL_OP_ALL is used due to
5396 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5397 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5398 followup to PR #377.
5401 *) IA-32 assembler support enhancements: unified ELF targets, support
5402 for SCO/Caldera platforms, fix for Cygwin shared build.
5405 *) Add support for FreeBSD on sparc64. As a consequence, support for
5406 FreeBSD on non-x86 processors is separate from x86 processors on
5407 the config script, much like the NetBSD support.
5408 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5410 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5412 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5415 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5416 code (06) was taken as the first octet of the session ID and the last
5417 octet was ignored consequently. As a result SSLv2 client side session
5418 caching could not have worked due to the session ID mismatch between
5420 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5424 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5425 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5429 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5430 seems that in spite of existing for more than a year, many application
5431 author have done nothing to provide the necessary callbacks, which
5432 means that this particular engine will not work properly anywhere.
5433 This is a very unfortunate situation which forces us, in the name
5434 of usability, to give the hw_ncipher.c a static lock, which is part
5436 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5437 appear in 0.9.8 or later. We EXPECT application authors to have
5438 dealt properly with this when 0.9.8 is released (unless we actually
5439 make such changes in the libcrypto locking code that changes will
5440 have to be made anyway).
5443 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5444 octets have been read, EOF or an error occurs. Without this change
5445 some truncated ASN1 structures will not produce an error.
5448 *) Disable Heimdal support, since it hasn't been fully implemented.
5449 Still give the possibility to force the use of Heimdal, but with
5450 warnings and a request that patches get sent to openssl-dev.
5453 *) Add the VC-CE target, introduce the WINCE sysname, and add
5454 INSTALL.WCE and appropriate conditionals to make it build.
5455 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5457 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5458 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5459 edit numbers of the version.
5460 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5462 *) Introduce safe string copy and catenation functions
5463 (BUF_strlcpy() and BUF_strlcat()).
5464 [Ben Laurie (CHATS) and Richard Levitte]
5466 *) Avoid using fixed-size buffers for one-line DNs.
5467 [Ben Laurie (CHATS)]
5469 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5470 resizing buffers containing secrets, and use where appropriate.
5471 [Ben Laurie (CHATS)]
5473 *) Avoid using fixed size buffers for configuration file location.
5474 [Ben Laurie (CHATS)]
5476 *) Avoid filename truncation for various CA files.
5477 [Ben Laurie (CHATS)]
5479 *) Use sizeof in preference to magic numbers.
5480 [Ben Laurie (CHATS)]
5482 *) Avoid filename truncation in cert requests.
5483 [Ben Laurie (CHATS)]
5485 *) Add assertions to check for (supposedly impossible) buffer
5487 [Ben Laurie (CHATS)]
5489 *) Don't cache truncated DNS entries in the local cache (this could
5490 potentially lead to a spoofing attack).
5491 [Ben Laurie (CHATS)]
5493 *) Fix various buffers to be large enough for hex/decimal
5494 representations in a platform independent manner.
5495 [Ben Laurie (CHATS)]
5497 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5498 resizing buffers containing secrets, and use where appropriate.
5499 [Ben Laurie (CHATS)]
5501 *) Add BIO_indent() to avoid much slightly worrying code to do
5503 [Ben Laurie (CHATS)]
5505 *) Convert sprintf()/BIO_puts() to BIO_printf().
5506 [Ben Laurie (CHATS)]
5508 *) buffer_gets() could terminate with the buffer only half
5510 [Ben Laurie (CHATS)]
5512 *) Add assertions to prevent user-supplied crypto functions from
5513 overflowing internal buffers by having large block sizes, etc.
5514 [Ben Laurie (CHATS)]
5516 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5518 [Ben Laurie (CHATS)]
5520 *) Eliminate unused copy of key in RC4.
5521 [Ben Laurie (CHATS)]
5523 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5524 [Ben Laurie (CHATS)]
5526 *) Fix off-by-one error in EGD path.
5527 [Ben Laurie (CHATS)]
5529 *) If RANDFILE path is too long, ignore instead of truncating.
5530 [Ben Laurie (CHATS)]
5532 *) Eliminate unused and incorrectly sized X.509 structure
5534 [Ben Laurie (CHATS)]
5536 *) Eliminate unused and dangerous function knumber().
5537 [Ben Laurie (CHATS)]
5539 *) Eliminate unused and dangerous structure, KSSL_ERR.
5540 [Ben Laurie (CHATS)]
5542 *) Protect against overlong session ID context length in an encoded
5543 session object. Since these are local, this does not appear to be
5545 [Ben Laurie (CHATS)]
5547 *) Change from security patch (see 0.9.6e below) that did not affect
5548 the 0.9.6 release series:
5550 Remote buffer overflow in SSL3 protocol - an attacker could
5551 supply an oversized master key in Kerberos-enabled versions.
5553 [Ben Laurie (CHATS)]
5555 *) Change the SSL kerb5 codes to match RFC 2712.
5558 *) Make -nameopt work fully for req and add -reqopt switch.
5559 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5561 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5562 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5564 *) Make sure tests can be performed even if the corresponding algorithms
5565 have been removed entirely. This was also the last step to make
5566 OpenSSL compilable with DJGPP under all reasonable conditions.
5567 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5569 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5570 to allow version independent disabling of normally unselected ciphers,
5571 which may be activated as a side-effect of selecting a single cipher.
5573 (E.g., cipher list string "RSA" enables ciphersuites that are left
5574 out of "ALL" because they do not provide symmetric encryption.
5575 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5576 [Lutz Jaenicke, Bodo Moeller]
5578 *) Add appropriate support for separate platform-dependent build
5579 directories. The recommended way to make a platform-dependent
5580 build directory is the following (tested on Linux), maybe with
5583 # Place yourself outside of the OpenSSL source tree. In
5584 # this example, the environment variable OPENSSL_SOURCE
5585 # is assumed to contain the absolute OpenSSL source directory.
5586 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5587 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5588 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5589 mkdir -p `dirname $F`
5590 ln -s $OPENSSL_SOURCE/$F $F
5593 To be absolutely sure not to disturb the source tree, a "make clean"
5594 is a good thing. If it isn't successful, don't worry about it,
5595 it probably means the source directory is very clean.
5598 *) Make sure any ENGINE control commands make local copies of string
5599 pointers passed to them whenever necessary. Otherwise it is possible
5600 the caller may have overwritten (or deallocated) the original string
5601 data when a later ENGINE operation tries to use the stored values.
5602 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5604 *) Improve diagnostics in file reading and command-line digests.
5605 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5607 *) Add AES modes CFB and OFB to the object database. Correct an
5608 error in AES-CFB decryption.
5611 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5612 allows existing EVP_CIPHER_CTX structures to be reused after
5613 calling EVP_*Final(). This behaviour is used by encryption
5614 BIOs and some applications. This has the side effect that
5615 applications must explicitly clean up cipher contexts with
5616 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5619 *) Check the values of dna and dnb in bn_mul_recursive before calling
5620 bn_mul_comba (a non zero value means the a or b arrays do not contain
5621 n2 elements) and fallback to bn_mul_normal if either is not zero.
5624 *) Fix escaping of non-ASCII characters when using the -subj option
5625 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5628 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5629 form for "surname", serialNumber has no short form.
5630 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5631 therefore remove "mail" short name for "internet 7".
5632 The OID for unique identifiers in X509 certificates is
5633 x500UniqueIdentifier, not uniqueIdentifier.
5634 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5637 *) Add an "init" command to the ENGINE config module and auto initialize
5638 ENGINEs. Without any "init" command the ENGINE will be initialized
5639 after all ctrl commands have been executed on it. If init=1 the
5640 ENGINE is initailized at that point (ctrls before that point are run
5641 on the uninitialized ENGINE and after on the initialized one). If
5642 init=0 then the ENGINE will not be iniatialized at all.
5645 *) Fix the 'app_verify_callback' interface so that the user-defined
5646 argument is actually passed to the callback: In the
5647 SSL_CTX_set_cert_verify_callback() prototype, the callback
5648 declaration has been changed from
5651 int (*cb)(X509_STORE_CTX *,void *);
5652 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5653 i=s->ctx->app_verify_callback(&ctx)
5654 has been changed into
5655 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5657 To update applications using SSL_CTX_set_cert_verify_callback(),
5658 a dummy argument can be added to their callback functions.
5659 [D. K. Smetters <smetters@parc.xerox.com>]
5661 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5662 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5664 *) Add and OPENSSL_LOAD_CONF define which will cause
5665 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5666 This allows older applications to transparently support certain
5667 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5668 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5669 load the config file and OPENSSL_add_all_algorithms_conf() which will
5670 always load it have also been added.
5673 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5674 Adjust NIDs and EVP layer.
5675 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5677 *) Config modules support in openssl utility.
5679 Most commands now load modules from the config file,
5680 though in a few (such as version) this isn't done
5681 because it couldn't be used for anything.
5683 In the case of ca and req the config file used is
5684 the same as the utility itself: that is the -config
5685 command line option can be used to specify an
5689 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5690 use "openssl_conf" if filename is NULL use default openssl config file.
5693 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5694 config section name. Add a new flag to tolerate a missing config file
5695 and move code to CONF_modules_load_file().
5698 *) Support for crypto accelerator cards from Accelerated Encryption
5699 Processing, www.aep.ie. (Use engine 'aep')
5700 The support was copied from 0.9.6c [engine] and adapted/corrected
5701 to work with the new engine framework.
5702 [AEP Inc. and Richard Levitte]
5704 *) Support for SureWare crypto accelerator cards from Baltimore
5705 Technologies. (Use engine 'sureware')
5706 The support was copied from 0.9.6c [engine] and adapted
5707 to work with the new engine framework.
5710 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5711 make the newer ENGINE framework commands for the CHIL engine work.
5712 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5714 *) Make it possible to produce shared libraries on ReliantUNIX.
5715 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5717 *) Add the configuration target debug-linux-ppro.
5718 Make 'openssl rsa' use the general key loading routines
5719 implemented in apps.c, and make those routines able to
5720 handle the key format FORMAT_NETSCAPE and the variant
5722 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5724 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5725 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5727 *) Add -keyform to rsautl, and document -engine.
5728 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5730 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5731 BIO_R_NO_SUCH_FILE error code rather than the generic
5732 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5735 *) Add new functions
5737 ERR_peek_last_error_line
5738 ERR_peek_last_error_line_data.
5739 These are similar to
5742 ERR_peek_error_line_data,
5743 but report on the latest error recorded rather than the first one
5744 still in the error queue.
5745 [Ben Laurie, Bodo Moeller]
5747 *) default_algorithms option in ENGINE config module. This allows things
5749 default_algorithms = ALL
5750 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5753 *) Preliminary ENGINE config module.
5756 *) New experimental application configuration code.
5759 *) Change the AES code to follow the same name structure as all other
5760 symmetric ciphers, and behave the same way. Move everything to
5761 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5762 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5764 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5765 [Ben Laurie and Theo de Raadt]
5767 *) Add option to output public keys in req command.
5768 [Massimiliano Pala madwolf@openca.org]
5770 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5771 (up to about 10% better than before for P-192 and P-224).
5774 *) New functions/macros
5776 SSL_CTX_set_msg_callback(ctx, cb)
5777 SSL_CTX_set_msg_callback_arg(ctx, arg)
5778 SSL_set_msg_callback(ssl, cb)
5779 SSL_set_msg_callback_arg(ssl, arg)
5781 to request calling a callback function
5783 void cb(int write_p, int version, int content_type,
5784 const void *buf, size_t len, SSL *ssl, void *arg)
5786 whenever a protocol message has been completely received
5787 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5788 protocol version according to which the SSL library interprets
5789 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5790 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5791 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5792 specification (change_cipher_spec(20), alert(21), handshake(22)).
5793 'buf' and 'len' point to the actual message, 'ssl' to the
5794 SSL object, and 'arg' is the application-defined value set by
5795 SSL[_CTX]_set_msg_callback_arg().
5797 'openssl s_client' and 'openssl s_server' have new '-msg' options
5798 to enable a callback that displays all protocol messages.
5801 *) Change the shared library support so shared libraries are built as
5802 soon as the corresponding static library is finished, and thereby get
5803 openssl and the test programs linked against the shared library.
5804 This still only happens when the keyword "shard" has been given to
5805 the configuration scripts.
5807 NOTE: shared library support is still an experimental thing, and
5808 backward binary compatibility is still not guaranteed.
5809 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5811 *) Add support for Subject Information Access extension.
5812 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5814 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5815 additional bytes when new memory had to be allocated, not just
5816 when reusing an existing buffer.
5819 *) New command line and configuration option 'utf8' for the req command.
5820 This allows field values to be specified as UTF8 strings.
5823 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5824 runs for the former and machine-readable output for the latter.
5827 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5828 of the e-mail address in the DN (i.e., it will go into a certificate
5829 extension only). The new configuration file option 'email_in_dn = no'
5830 has the same effect.
5831 [Massimiliano Pala madwolf@openca.org]
5833 *) Change all functions with names starting with des_ to be starting
5834 with DES_ instead. Add wrappers that are compatible with libdes,
5835 but are named _ossl_old_des_*. Finally, add macros that map the
5836 des_* symbols to the corresponding _ossl_old_des_* if libdes
5837 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5838 desired, the des_* symbols will be mapped to DES_*, with one
5841 Since we provide two compatibility mappings, the user needs to
5842 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5843 compatibility is desired. The default (i.e., when that macro
5844 isn't defined) is OpenSSL 0.9.6c compatibility.
5846 There are also macros that enable and disable the support of old
5847 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5848 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5849 are defined, the default will apply: to support the old des routines.
5851 In either case, one must include openssl/des.h to get the correct
5852 definitions. Do not try to just include openssl/des_old.h, that
5855 NOTE: This is a major break of an old API into a new one. Software
5856 authors are encouraged to switch to the DES_ style functions. Some
5857 time in the future, des_old.h and the libdes compatibility functions
5858 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5859 default), and then completely removed.
5862 *) Test for certificates which contain unsupported critical extensions.
5863 If such a certificate is found during a verify operation it is
5864 rejected by default: this behaviour can be overridden by either
5865 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5866 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5867 X509_supported_extension() has also been added which returns 1 if a
5868 particular extension is supported.
5871 *) Modify the behaviour of EVP cipher functions in similar way to digests
5872 to retain compatibility with existing code.
5875 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5876 compatibility with existing code. In particular the 'ctx' parameter does
5877 not have to be to be initialized before the call to EVP_DigestInit() and
5878 it is tidied up after a call to EVP_DigestFinal(). New function
5879 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5880 EVP_MD_CTX_copy() changed to not require the destination to be
5881 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5882 requires the destination to be valid.
5884 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5885 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5888 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5889 so that complete 'Handshake' protocol structures are kept in memory
5890 instead of overwriting 'msg_type' and 'length' with 'body' data.
5893 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5894 [Massimo Santin via Richard Levitte]
5896 *) Major restructuring to the underlying ENGINE code. This includes
5897 reduction of linker bloat, separation of pure "ENGINE" manipulation
5898 (initialisation, etc) from functionality dealing with implementations
5899 of specific crypto iterfaces. This change also introduces integrated
5900 support for symmetric ciphers and digest implementations - so ENGINEs
5901 can now accelerate these by providing EVP_CIPHER and EVP_MD
5902 implementations of their own. This is detailed in crypto/engine/README
5903 as it couldn't be adequately described here. However, there are a few
5904 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5905 were changed in the original introduction of ENGINE code have now
5906 reverted back - the hooking from this code to ENGINE is now a good
5907 deal more passive and at run-time, operations deal directly with
5908 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5909 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5910 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5911 they were not being used by the framework as there is no concept of a
5912 BIGNUM_METHOD and they could not be generalised to the new
5913 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5914 ENGINE_cpy() has been removed as it cannot be consistently defined in
5918 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5921 *) Change mkdef.pl to sort symbols that get the same entry number,
5922 and make sure the automatically generated functions ERR_load_*
5923 become part of libeay.num as well.
5926 *) New function SSL_renegotiate_pending(). This returns true once
5927 renegotiation has been requested (either SSL_renegotiate() call
5928 or HelloRequest/ClientHello received from the peer) and becomes
5929 false once a handshake has been completed.
5930 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5931 sends a HelloRequest, but does not ensure that a handshake takes
5932 place. SSL_renegotiate_pending() is useful for checking if the
5933 client has followed the request.)
5936 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5937 By default, clients may request session resumption even during
5938 renegotiation (if session ID contexts permit); with this option,
5939 session resumption is possible only in the first handshake.
5941 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5942 more bits available for options that should not be part of
5943 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5946 *) Add some demos for certificate and certificate request creation.
5949 *) Make maximum certificate chain size accepted from the peer application
5950 settable (SSL*_get/set_max_cert_list()), as proposed by
5951 "Douglas E. Engert" <deengert@anl.gov>.
5954 *) Add support for shared libraries for Unixware-7
5955 (Boyd Lynn Gerber <gerberb@zenez.com>).
5958 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5959 be done prior to destruction. Use this to unload error strings from
5960 ENGINEs that load their own error strings. NB: This adds two new API
5961 functions to "get" and "set" this destroy handler in an ENGINE.
5964 *) Alter all existing ENGINE implementations (except "openssl" and
5965 "openbsd") to dynamically instantiate their own error strings. This
5966 makes them more flexible to be built both as statically-linked ENGINEs
5967 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5968 Also, add stub code to each that makes building them as self-contained
5969 shared-libraries easier (see README.ENGINE).
5972 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5973 implementations into applications that are completely implemented in
5974 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5975 commands that can be used to configure what shared-library to load and
5976 to control aspects of the way it is handled. Also, made an update to
5977 the README.ENGINE file that brings its information up-to-date and
5978 provides some information and instructions on the "dynamic" ENGINE
5979 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5982 *) Make it possible to unload ranges of ERR strings with a new
5983 "ERR_unload_strings" function.
5986 *) Add a copy() function to EVP_MD.
5989 *) Make EVP_MD routines take a context pointer instead of just the
5990 md_data void pointer.
5993 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5994 that the digest can only process a single chunk of data
5995 (typically because it is provided by a piece of
5996 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5997 is only going to provide a single chunk of data, and hence the
5998 framework needn't accumulate the data for oneshot drivers.
6001 *) As with "ERR", make it possible to replace the underlying "ex_data"
6002 functions. This change also alters the storage and management of global
6003 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6004 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6005 index counters. The API functions that use this state have been changed
6006 to take a "class_index" rather than pointers to the class's local STACK
6007 and counter, and there is now an API function to dynamically create new
6008 classes. This centralisation allows us to (a) plug a lot of the
6009 thread-safety problems that existed, and (b) makes it possible to clean
6010 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6011 such data would previously have always leaked in application code and
6012 workarounds were in place to make the memory debugging turn a blind eye
6013 to it. Application code that doesn't use this new function will still
6014 leak as before, but their memory debugging output will announce it now
6015 rather than letting it slide.
6017 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6018 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6019 has a return value to indicate success or failure.
6022 *) Make it possible to replace the underlying "ERR" functions such that the
6023 global state (2 LHASH tables and 2 locks) is only used by the "default"
6024 implementation. This change also adds two functions to "get" and "set"
6025 the implementation prior to it being automatically set the first time
6026 any other ERR function takes place. Ie. an application can call "get",
6027 pass the return value to a module it has just loaded, and that module
6028 can call its own "set" function using that value. This means the
6029 module's "ERR" operations will use (and modify) the error state in the
6030 application and not in its own statically linked copy of OpenSSL code.
6033 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6034 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6035 the operation, and provides a more encapsulated way for external code
6036 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6037 to use these functions rather than manually incrementing the counts.
6039 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6042 *) Add EVP test program.
6045 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6048 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6049 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6050 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6051 These allow a CRL to be built without having to access X509_CRL fields
6052 directly. Modify 'ca' application to use new functions.
6055 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6056 bug workarounds. Rollback attack detection is a security feature.
6057 The problem will only arise on OpenSSL servers when TLSv1 is not
6058 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6059 Software authors not wanting to support TLSv1 will have special reasons
6060 for their choice and can explicitly enable this option.
6061 [Bodo Moeller, Lutz Jaenicke]
6063 *) Rationalise EVP so it can be extended: don't include a union of
6064 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6065 (similar to those existing for EVP_CIPHER_CTX).
6070 EVP_MD_CTX_init(&md); /* new function call */
6071 EVP_DigestInit(&md, EVP_sha1());
6072 EVP_DigestUpdate(&md, in, len);
6073 EVP_DigestFinal(&md, out, NULL);
6074 EVP_MD_CTX_cleanup(&md); /* new function call */
6078 *) Make DES key schedule conform to the usual scheme, as well as
6079 correcting its structure. This means that calls to DES functions
6080 now have to pass a pointer to a des_key_schedule instead of a
6081 plain des_key_schedule (which was actually always a pointer
6084 des_key_schedule ks;
6086 des_set_key_checked(..., &ks);
6087 des_ncbc_encrypt(..., &ks, ...);
6089 (Note that a later change renames 'des_...' into 'DES_...'.)
6092 *) Initial reduction of linker bloat: the use of some functions, such as
6093 PEM causes large amounts of unused functions to be linked in due to
6094 poor organisation. For example pem_all.c contains every PEM function
6095 which has a knock on effect of linking in large amounts of (unused)
6096 ASN1 code. Grouping together similar functions and splitting unrelated
6097 functions prevents this.
6100 *) Cleanup of EVP macros.
6103 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6104 correct _ecb suffix.
6107 *) Add initial OCSP responder support to ocsp application. The
6108 revocation information is handled using the text based index
6109 use by the ca application. The responder can either handle
6110 requests generated internally, supplied in files (for example
6111 via a CGI script) or using an internal minimal server.
6114 *) Add configuration choices to get zlib compression for TLS.
6117 *) Changes to Kerberos SSL for RFC 2712 compliance:
6118 1. Implemented real KerberosWrapper, instead of just using
6119 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6120 2. Implemented optional authenticator field of KerberosWrapper.
6122 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6123 and authenticator structs; see crypto/krb5/.
6125 Generalized Kerberos calls to support multiple Kerberos libraries.
6126 [Vern Staats <staatsvr@asc.hpc.mil>,
6127 Jeffrey Altman <jaltman@columbia.edu>
6128 via Richard Levitte]
6130 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6131 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6132 values for each of the key sizes rather than having just
6133 parameters (and 'speed' generating keys each time).
6136 *) Speed up EVP routines.
6139 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6140 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6141 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6142 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6144 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6145 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6146 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6149 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6151 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6154 *) Added the OS2-EMX target.
6155 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6157 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6158 to support NCONF routines in extension code. New function CONF_set_nconf()
6159 to allow functions which take an NCONF to also handle the old LHASH
6160 structure: this means that the old CONF compatible routines can be
6161 retained (in particular wrt extensions) without having to duplicate the
6162 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6165 *) Enhance the general user interface with mechanisms for inner control
6166 and with possibilities to have yes/no kind of prompts.
6169 *) Change all calls to low level digest routines in the library and
6170 applications to use EVP. Add missing calls to HMAC_cleanup() and
6171 don't assume HMAC_CTX can be copied using memcpy().
6172 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6174 *) Add the possibility to control engines through control names but with
6175 arbitrary arguments instead of just a string.
6176 Change the key loaders to take a UI_METHOD instead of a callback
6177 function pointer. NOTE: this breaks binary compatibility with earlier
6178 versions of OpenSSL [engine].
6179 Adapt the nCipher code for these new conditions and add a card insertion
6183 *) Enhance the general user interface with mechanisms to better support
6184 dialog box interfaces, application-defined prompts, the possibility
6185 to use defaults (for example default passwords from somewhere else)
6186 and interrupts/cancellations.
6189 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6190 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6193 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6194 tidy up some unnecessarily weird code in 'sk_new()').
6195 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6197 *) Change the key loading routines for ENGINEs to use the same kind
6198 callback (pem_password_cb) as all other routines that need this
6202 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6203 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6204 than this minimum value is recommended.
6207 *) New random seeder for OpenVMS, using the system process statistics
6208 that are easily reachable.
6211 *) Windows apparently can't transparently handle global
6212 variables defined in DLLs. Initialisations such as:
6214 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6216 wont compile. This is used by the any applications that need to
6217 declare their own ASN1 modules. This was fixed by adding the option
6218 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6219 needed for static libraries under Win32.
6222 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6223 setting of purpose and trust fields. New X509_STORE trust and
6224 purpose functions and tidy up setting in other SSL functions.
6227 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6228 structure. These are inherited by X509_STORE_CTX when it is
6229 initialised. This allows various defaults to be set in the
6230 X509_STORE structure (such as flags for CRL checking and custom
6231 purpose or trust settings) for functions which only use X509_STORE_CTX
6232 internally such as S/MIME.
6234 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6235 trust settings if they are not set in X509_STORE. This allows X509_STORE
6236 purposes and trust (in S/MIME for example) to override any set by default.
6238 Add command line options for CRL checking to smime, s_client and s_server
6242 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6243 are set then the CRL is looked up in the X509_STORE structure and
6244 its validity and signature checked, then if the certificate is found
6245 in the CRL the verify fails with a revoked error.
6247 Various new CRL related callbacks added to X509_STORE_CTX structure.
6249 Command line options added to 'verify' application to support this.
6251 This needs some additional work, such as being able to handle multiple
6252 CRLs with different times, extension based lookup (rather than just
6253 by subject name) and ultimately more complete V2 CRL extension
6257 *) Add a general user interface API (crypto/ui/). This is designed
6258 to replace things like des_read_password and friends (backward
6259 compatibility functions using this new API are provided).
6260 The purpose is to remove prompting functions from the DES code
6261 section as well as provide for prompting through dialog boxes in
6262 a window system and the like.
6265 *) Add "ex_data" support to ENGINE so implementations can add state at a
6266 per-structure level rather than having to store it globally.
6269 *) Make it possible for ENGINE structures to be copied when retrieved by
6270 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6271 This causes the "original" ENGINE structure to act like a template,
6272 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6273 operational state can be localised to each ENGINE structure, despite the
6274 fact they all share the same "methods". New ENGINE structures returned in
6275 this case have no functional references and the return value is the single
6276 structural reference. This matches the single structural reference returned
6277 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6281 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6282 needs to match any other type at all we need to manually clear the
6286 *) Changes to the "openssl engine" utility to include;
6287 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6288 about an ENGINE's available control commands.
6289 - executing control commands from command line arguments using the
6290 '-pre' and '-post' switches. '-post' is only used if '-t' is
6291 specified and the ENGINE is successfully initialised. The syntax for
6292 the individual commands are colon-separated, for example;
6293 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6296 *) New dynamic control command support for ENGINEs. ENGINEs can now
6297 declare their own commands (numbers), names (strings), descriptions,
6298 and input types for run-time discovery by calling applications. A
6299 subset of these commands are implicitly classed as "executable"
6300 depending on their input type, and only these can be invoked through
6301 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6302 can be based on user input, config files, etc). The distinction is
6303 that "executable" commands cannot return anything other than a boolean
6304 result and can only support numeric or string input, whereas some
6305 discoverable commands may only be for direct use through
6306 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6307 pointers, or other custom uses. The "executable" commands are to
6308 support parameterisations of ENGINE behaviour that can be
6309 unambiguously defined by ENGINEs and used consistently across any
6310 OpenSSL-based application. Commands have been added to all the
6311 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6312 control over shared-library paths without source code alterations.
6315 *) Changed all ENGINE implementations to dynamically allocate their
6316 ENGINEs rather than declaring them statically. Apart from this being
6317 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6318 this also allows the implementations to compile without using the
6319 internal engine_int.h header.
6322 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6323 'const' value. Any code that should be able to modify a RAND_METHOD
6324 should already have non-const pointers to it (ie. they should only
6325 modify their own ones).
6328 *) Made a variety of little tweaks to the ENGINE code.
6329 - "atalla" and "ubsec" string definitions were moved from header files
6330 to C code. "nuron" string definitions were placed in variables
6331 rather than hard-coded - allowing parameterisation of these values
6332 later on via ctrl() commands.
6333 - Removed unused "#if 0"'d code.
6334 - Fixed engine list iteration code so it uses ENGINE_free() to release
6335 structural references.
6336 - Constified the RAND_METHOD element of ENGINE structures.
6337 - Constified various get/set functions as appropriate and added
6338 missing functions (including a catch-all ENGINE_cpy that duplicates
6339 all ENGINE values onto a new ENGINE except reference counts/state).
6340 - Removed NULL parameter checks in get/set functions. Setting a method
6341 or function to NULL is a way of cancelling out a previously set
6342 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6343 and doesn't justify the extra error symbols and code.
6344 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6345 flags from engine_int.h to engine.h.
6346 - Changed prototypes for ENGINE handler functions (init(), finish(),
6347 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6350 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6351 to the algorithm using long division. The binary algorithm can be
6352 used only if the modulus is odd. On 32-bit systems, it is faster
6353 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6354 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6355 up to 450 bits. In 64-bit environments, the binary algorithm
6356 appears to be advantageous for much longer moduli; here we use it
6357 for moduli up to 2048 bits.
6360 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6361 could not support the combine flag in choice fields.
6364 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6365 extensions from a certificate request to the certificate.
6368 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6369 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6370 file: this allows the display of the certificate about to be
6371 signed to be customised, to allow certain fields to be included
6372 or excluded and extension details. The old system didn't display
6373 multicharacter strings properly, omitted fields not in the policy
6374 and couldn't display additional details such as extensions.
6377 *) Function EC_POINTs_mul for multiple scalar multiplication
6378 of an arbitrary number of elliptic curve points
6379 \sum scalars[i]*points[i],
6380 optionally including the generator defined for the EC_GROUP:
6381 scalar*generator + \sum scalars[i]*points[i].
6383 EC_POINT_mul is a simple wrapper function for the typical case
6384 that the point list has just one item (besides the optional
6388 *) First EC_METHODs for curves over GF(p):
6390 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6391 operations and provides various method functions that can also
6392 operate with faster implementations of modular arithmetic.
6394 EC_GFp_mont_method() reuses most functions that are part of
6395 EC_GFp_simple_method, but uses Montgomery arithmetic.
6397 [Bodo Moeller; point addition and point doubling
6398 implementation directly derived from source code provided by
6399 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6401 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6402 crypto/ec/ec_lib.c):
6404 Curves are EC_GROUP objects (with an optional group generator)
6405 based on EC_METHODs that are built into the library.
6407 Points are EC_POINT objects based on EC_GROUP objects.
6409 Most of the framework would be able to handle curves over arbitrary
6410 finite fields, but as there are no obvious types for fields other
6411 than GF(p), some functions are limited to that for now.
6414 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6415 that the file contains a complete HTTP response.
6418 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6419 change the def and num file printf format specifier from "%-40sXXX"
6420 to "%-39s XXX". The latter will always guarantee a space after the
6421 field while the former will cause them to run together if the field
6422 is 40 of more characters long.
6425 *) Constify the cipher and digest 'method' functions and structures
6426 and modify related functions to take constant EVP_MD and EVP_CIPHER
6430 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6431 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6434 *) Modify EVP_Digest*() routines so they now return values. Although the
6435 internal software routines can never fail additional hardware versions
6439 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6441 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6442 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6445 ERR_R_NESTED_ASN1_ERROR
6447 ERR_R_MISSING_ASN1_EOS
6448 were 4 .. 9, conflicting with
6449 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6451 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6452 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6454 Add new error code 'ERR_R_INTERNAL_ERROR'.
6457 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6461 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6462 sets the subject name for a new request or supersedes the
6463 subject name in a given request. Formats that can be parsed are
6464 'CN=Some Name, OU=myOU, C=IT'
6466 'CN=Some Name/OU=myOU/C=IT'.
6468 Add options '-batch' and '-verbose' to 'openssl req'.
6469 [Massimiliano Pala <madwolf@hackmasters.net>]
6471 *) Introduce the possibility to access global variables through
6472 functions on platform were that's the best way to handle exporting
6473 global variables in shared libraries. To enable this functionality,
6474 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6475 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6476 is normally done by Configure or something similar).
6478 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6479 in the source file (foo.c) like this:
6481 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6482 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6484 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6485 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6487 OPENSSL_DECLARE_GLOBAL(int,foo);
6488 #define foo OPENSSL_GLOBAL_REF(foo)
6489 OPENSSL_DECLARE_GLOBAL(double,bar);
6490 #define bar OPENSSL_GLOBAL_REF(bar)
6492 The #defines are very important, and therefore so is including the
6493 header file everywhere where the defined globals are used.
6495 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6496 of ASN.1 items, but that structure is a bit different.
6498 The largest change is in util/mkdef.pl which has been enhanced with
6499 better and easier to understand logic to choose which symbols should
6500 go into the Windows .def files as well as a number of fixes and code
6501 cleanup (among others, algorithm keywords are now sorted
6502 lexicographically to avoid constant rewrites).
6505 *) In BN_div() keep a copy of the sign of 'num' before writing the
6506 result to 'rm' because if rm==num the value will be overwritten
6507 and produce the wrong result if 'num' is negative: this caused
6508 problems with BN_mod() and BN_nnmod().
6511 *) Function OCSP_request_verify(). This checks the signature on an
6512 OCSP request and verifies the signer certificate. The signer
6513 certificate is just checked for a generic purpose and OCSP request
6517 *) Add OCSP_check_validity() function to check the validity of OCSP
6518 responses. OCSP responses are prepared in real time and may only
6519 be a few seconds old. Simply checking that the current time lies
6520 between thisUpdate and nextUpdate max reject otherwise valid responses
6521 caused by either OCSP responder or client clock inaccuracy. Instead
6522 we allow thisUpdate and nextUpdate to fall within a certain period of
6523 the current time. The age of the response can also optionally be
6524 checked. Two new options -validity_period and -status_age added to
6528 *) If signature or public key algorithm is unrecognized print out its
6529 OID rather that just UNKNOWN.
6532 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6533 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6534 ID to be generated from the issuer certificate alone which can then be
6535 passed to OCSP_id_issuer_cmp().
6538 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6539 ASN1 modules to export functions returning ASN1_ITEM pointers
6540 instead of the ASN1_ITEM structures themselves. This adds several
6541 new macros which allow the underlying ASN1 function/structure to
6542 be accessed transparently. As a result code should not use ASN1_ITEM
6543 references directly (such as &X509_it) but instead use the relevant
6544 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6545 use of the new ASN1 code on platforms where exporting structures
6546 is problematical (for example in shared libraries) but exporting
6547 functions returning pointers to structures is not.
6550 *) Add support for overriding the generation of SSL/TLS session IDs.
6551 These callbacks can be registered either in an SSL_CTX or per SSL.
6552 The purpose of this is to allow applications to control, if they wish,
6553 the arbitrary values chosen for use as session IDs, particularly as it
6554 can be useful for session caching in multiple-server environments. A
6555 command-line switch for testing this (and any client code that wishes
6556 to use such a feature) has been added to "s_server".
6557 [Geoff Thorpe, Lutz Jaenicke]
6559 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6560 of the form '#if defined(...) || defined(...) || ...' and
6561 '#if !defined(...) && !defined(...) && ...'. This also avoids
6562 the growing number of special cases it was previously handling.
6565 *) Make all configuration macros available for application by making
6566 sure they are available in opensslconf.h, by giving them names starting
6567 with "OPENSSL_" to avoid conflicts with other packages and by making
6568 sure e_os2.h will cover all platform-specific cases together with
6570 Additionally, it is now possible to define configuration/platform-
6571 specific names (called "system identities"). In the C code, these
6572 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6573 macro with the name beginning with "OPENSSL_SYS_", which is determined
6574 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6578 *) New option -set_serial to 'req' and 'x509' this allows the serial
6579 number to use to be specified on the command line. Previously self
6580 signed certificates were hard coded with serial number 0 and the
6581 CA options of 'x509' had to use a serial number in a file which was
6585 *) New options to 'ca' utility to support V2 CRL entry extensions.
6586 Currently CRL reason, invalidity date and hold instruction are
6587 supported. Add new CRL extensions to V3 code and some new objects.
6590 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6591 disable standard block padding (aka PKCS#5 padding) in the EVP
6592 API, which was previously mandatory. This means that the data is
6593 not padded in any way and so the total length much be a multiple
6594 of the block size, otherwise an error occurs.
6597 *) Initial (incomplete) OCSP SSL support.
6600 *) New function OCSP_parse_url(). This splits up a URL into its host,
6601 port and path components: primarily to parse OCSP URLs. New -url
6602 option to ocsp utility.
6605 *) New nonce behavior. The return value of OCSP_check_nonce() now
6606 reflects the various checks performed. Applications can decide
6607 whether to tolerate certain situations such as an absent nonce
6608 in a response when one was present in a request: the ocsp application
6609 just prints out a warning. New function OCSP_add1_basic_nonce()
6610 this is to allow responders to include a nonce in a response even if
6611 the request is nonce-less.
6614 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6615 skipped when using openssl x509 multiple times on a single input file,
6616 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6619 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6620 set string type: to handle setting ASN1_TIME structures. Fix ca
6621 utility to correctly initialize revocation date of CRLs.
6624 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6625 the clients preferred ciphersuites and rather use its own preferences.
6626 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6627 Internet Explorer by ensuring unchanged hash method during stepup.
6628 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6631 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6632 to aes and add a new 'exist' option to print out symbols that don't
6636 *) Additional options to ocsp utility to allow flags to be set and
6637 additional certificates supplied.
6640 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6641 OCSP client a number of certificate to only verify the response
6645 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6646 handle the new API. Currently only ECB, CBC modes supported. Add new
6649 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6650 Encryption Standard (AES) Ciphersuites for Transport Layer
6651 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6652 not enabled by default and were not part of the "ALL" ciphersuite
6653 alias because they were not yet official; they could be
6654 explicitly requested by specifying the "AESdraft" ciphersuite
6655 group alias. In the final release of OpenSSL 0.9.7, the group
6656 alias is called "AES" and is part of "ALL".)
6657 [Ben Laurie, Steve Henson, Bodo Moeller]
6659 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6660 request to response.
6663 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6664 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6665 extract information from a certificate request. OCSP_response_create()
6666 creates a response and optionally adds a basic response structure.
6667 OCSP_basic_add1_status() adds a complete single response to a basic
6668 response and returns the OCSP_SINGLERESP structure just added (to allow
6669 extensions to be included for example). OCSP_basic_add1_cert() adds a
6670 certificate to a basic response and OCSP_basic_sign() signs a basic
6671 response with various flags. New helper functions ASN1_TIME_check()
6672 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6673 (converts ASN1_TIME to GeneralizedTime).
6676 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6677 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6678 structure from a certificate. X509_pubkey_digest() digests the public_key
6679 contents: this is used in various key identifiers.
6682 *) Make sk_sort() tolerate a NULL argument.
6683 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6685 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6686 passed by the function are trusted implicitly. If any of them signed the
6687 response then it is assumed to be valid and is not verified.
6690 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6691 to data. This was previously part of the PKCS7 ASN1 code. This
6692 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6693 [Steve Henson, reported by Kenneth R. Robinette
6694 <support@securenetterm.com>]
6696 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6697 routines: without these tracing memory leaks is very painful.
6698 Fix leaks in PKCS12 and PKCS7 routines.
6701 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6702 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6703 effectively meant GeneralizedTime would never be used. Now it
6704 is initialised to -1 but X509_time_adj() now has to check the value
6705 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6706 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6707 [Steve Henson, reported by Kenneth R. Robinette
6708 <support@securenetterm.com>]
6710 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6711 result in a zero length in the ASN1_INTEGER structure which was
6712 not consistent with the structure when d2i_ASN1_INTEGER() was used
6713 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6714 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6715 where it did not print out a minus for negative ASN1_INTEGER.
6718 *) Add summary printout to ocsp utility. The various functions which
6719 convert status values to strings have been renamed to:
6720 OCSP_response_status_str(), OCSP_cert_status_str() and
6721 OCSP_crl_reason_str() and are no longer static. New options
6722 to verify nonce values and to disable verification. OCSP response
6723 printout format cleaned up.
6726 *) Add additional OCSP certificate checks. These are those specified
6727 in RFC2560. This consists of two separate checks: the CA of the
6728 certificate being checked must either be the OCSP signer certificate
6729 or the issuer of the OCSP signer certificate. In the latter case the
6730 OCSP signer certificate must contain the OCSP signing extended key
6731 usage. This check is performed by attempting to match the OCSP
6732 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6733 in the OCSP_CERTID structures of the response.
6736 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6737 and related routines. This uses the standard OpenSSL certificate
6738 verify routines to perform initial checks (just CA validity) and
6739 to obtain the certificate chain. Then additional checks will be
6740 performed on the chain. Currently the root CA is checked to see
6741 if it is explicitly trusted for OCSP signing. This is used to set
6742 a root CA as a global signing root: that is any certificate that
6743 chains to that CA is an acceptable OCSP signing certificate.
6746 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6747 extensions from a separate configuration file.
6748 As when reading extensions from the main configuration file,
6749 the '-extensions ...' option may be used for specifying the
6751 [Massimiliano Pala <madwolf@comune.modena.it>]
6753 *) New OCSP utility. Allows OCSP requests to be generated or
6754 read. The request can be sent to a responder and the output
6755 parsed, outputed or printed in text form. Not complete yet:
6756 still needs to check the OCSP response validity.
6759 *) New subcommands for 'openssl ca':
6760 'openssl ca -status <serial>' prints the status of the cert with
6761 the given serial number (according to the index file).
6762 'openssl ca -updatedb' updates the expiry status of certificates
6764 [Massimiliano Pala <madwolf@comune.modena.it>]
6766 *) New '-newreq-nodes' command option to CA.pl. This is like
6767 '-newreq', but calls 'openssl req' with the '-nodes' option
6768 so that the resulting key is not encrypted.
6769 [Damien Miller <djm@mindrot.org>]
6771 *) New configuration for the GNU Hurd.
6772 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6774 *) Initial code to implement OCSP basic response verify. This
6775 is currently incomplete. Currently just finds the signer's
6776 certificate and verifies the signature on the response.
6779 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6780 value of OPENSSLDIR. This is available via the new '-d' option
6781 to 'openssl version', and is also included in 'openssl version -a'.
6784 *) Allowing defining memory allocation callbacks that will be given
6785 file name and line number information in additional arguments
6786 (a const char* and an int). The basic functionality remains, as
6787 well as the original possibility to just replace malloc(),
6788 realloc() and free() by functions that do not know about these
6789 additional arguments. To register and find out the current
6790 settings for extended allocation functions, the following
6791 functions are provided:
6793 CRYPTO_set_mem_ex_functions
6794 CRYPTO_set_locked_mem_ex_functions
6795 CRYPTO_get_mem_ex_functions
6796 CRYPTO_get_locked_mem_ex_functions
6798 These work the same way as CRYPTO_set_mem_functions and friends.
6799 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6800 extended allocation function is enabled.
6801 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6802 a conventional allocation function is enabled.
6803 [Richard Levitte, Bodo Moeller]
6805 *) Finish off removing the remaining LHASH function pointer casts.
6806 There should no longer be any prototype-casting required when using
6807 the LHASH abstraction, and any casts that remain are "bugs". See
6808 the callback types and macros at the head of lhash.h for details
6809 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6812 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6813 If /dev/[u]random devices are not available or do not return enough
6814 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6816 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6817 /etc/entropy will be queried once each in this sequence, quering stops
6818 when enough entropy was collected without querying more sockets.
6821 *) Change the Unix RAND_poll() variant to be able to poll several
6822 random devices, as specified by DEVRANDOM, until a sufficient amount
6823 of data has been collected. We spend at most 10 ms on each file
6824 (select timeout) and read in non-blocking mode. DEVRANDOM now
6825 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6826 (previously it was just the string "/dev/urandom"), so on typical
6827 platforms the 10 ms delay will never occur.
6828 Also separate out the Unix variant to its own file, rand_unix.c.
6829 For VMS, there's a currently-empty rand_vms.c.
6832 *) Move OCSP client related routines to ocsp_cl.c. These
6833 provide utility functions which an application needing
6834 to issue a request to an OCSP responder and analyse the
6835 response will typically need: as opposed to those which an
6836 OCSP responder itself would need which will be added later.
6838 OCSP_request_sign() signs an OCSP request with an API similar
6839 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6840 response. OCSP_response_get1_basic() extracts basic response
6841 from response. OCSP_resp_find_status(): finds and extracts status
6842 information from an OCSP_CERTID structure (which will be created
6843 when the request structure is built). These are built from lower
6844 level functions which work on OCSP_SINGLERESP structures but
6845 wont normally be used unless the application wishes to examine
6846 extensions in the OCSP response for example.
6848 Replace nonce routines with a pair of functions.
6849 OCSP_request_add1_nonce() adds a nonce value and optionally
6850 generates a random value. OCSP_check_nonce() checks the
6851 validity of the nonce in an OCSP response.
6854 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6855 This doesn't copy the supplied OCSP_CERTID and avoids the
6856 need to free up the newly created id. Change return type
6857 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6858 This can then be used to add extensions to the request.
6859 Deleted OCSP_request_new(), since most of its functionality
6860 is now in OCSP_REQUEST_new() (and the case insensitive name
6861 clash) apart from the ability to set the request name which
6862 will be added elsewhere.
6865 *) Update OCSP API. Remove obsolete extensions argument from
6866 various functions. Extensions are now handled using the new
6867 OCSP extension code. New simple OCSP HTTP function which
6868 can be used to send requests and parse the response.
6871 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6872 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6873 uses the special reorder version of SET OF to sort the attributes
6874 and reorder them to match the encoded order. This resolves a long
6875 standing problem: a verify on a PKCS7 structure just after signing
6876 it used to fail because the attribute order did not match the
6877 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6878 it uses the received order. This is necessary to tolerate some broken
6879 software that does not order SET OF. This is handled by encoding
6880 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6881 to produce the required SET OF.
6884 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6885 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6886 files to get correct declarations of the ASN.1 item variables.
6889 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6890 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6891 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6892 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6893 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6894 ASN1_ITEM and no wrapper functions.
6897 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6898 replace the old function pointer based I/O routines. Change most of
6899 the *_d2i_bio() and *_d2i_fp() functions to use these.
6902 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6903 lines, recognice more "algorithms" that can be deselected, and make
6904 it complain about algorithm deselection that isn't recognised.
6907 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6908 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6909 to use new functions. Add NO_ASN1_OLD which can be set to remove
6910 some old style ASN1 functions: this can be used to determine if old
6911 code will still work when these eventually go away.
6914 *) New extension functions for OCSP structures, these follow the
6915 same conventions as certificates and CRLs.
6918 *) New function X509V3_add1_i2d(). This automatically encodes and
6919 adds an extension. Its behaviour can be customised with various
6920 flags to append, replace or delete. Various wrappers added for
6921 certifcates and CRLs.
6924 *) Fix to avoid calling the underlying ASN1 print routine when
6925 an extension cannot be parsed. Correct a typo in the
6926 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6929 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6930 entries for variables.
6933 *) Add functionality to apps/openssl.c for detecting locking
6934 problems: As the program is single-threaded, all we have
6935 to do is register a locking callback using an array for
6936 storing which locks are currently held by the program.
6939 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6940 SSL_get_ex_data_X509_STORE_idx(), which is used in
6941 ssl_verify_cert_chain() and thus can be called at any time
6942 during TLS/SSL handshakes so that thread-safety is essential.
6943 Unfortunately, the ex_data design is not at all suited
6944 for multi-threaded use, so it probably should be abolished.
6947 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6948 [Broadcom, tweaked and integrated by Geoff Thorpe]
6950 *) Move common extension printing code to new function
6951 X509V3_print_extensions(). Reorganise OCSP print routines and
6952 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6955 *) New function X509_signature_print() to remove duplication in some
6959 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6960 set (this was treated exactly the same as SET OF previously). This
6961 is used to reorder the STACK representing the structure to match the
6962 encoding. This will be used to get round a problem where a PKCS7
6963 structure which was signed could not be verified because the STACK
6964 order did not reflect the encoded order.
6967 *) Reimplement the OCSP ASN1 module using the new code.
6970 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6971 for its ASN1 operations. The old style function pointers still exist
6972 for now but they will eventually go away.
6975 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6976 completely replaces the old ASN1 functionality with a table driven
6977 encoder and decoder which interprets an ASN1_ITEM structure describing
6978 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6979 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6980 has also been converted to the new form.
6983 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6984 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6985 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6986 for negative moduli.
6989 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6990 of not touching the result's sign bit.
6993 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6997 *) Changed the LHASH code to use prototypes for callbacks, and created
6998 macros to declare and implement thin (optionally static) functions
6999 that provide type-safety and avoid function pointer casting for the
7000 type-specific callbacks.
7003 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7005 [Veers Staats <staatsvr@asc.hpc.mil>,
7006 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7008 *) Reformat the FAQ so the different questions and answers can be divided
7009 in sections depending on the subject.
7012 *) Have the zlib compression code load ZLIB.DLL dynamically under
7016 *) New function BN_mod_sqrt for computing square roots modulo a prime
7017 (using the probabilistic Tonelli-Shanks algorithm unless
7018 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7019 be handled deterministically).
7020 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7022 *) Make BN_mod_inverse faster by explicitly handling small quotients
7023 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7024 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7027 *) New function BN_kronecker.
7030 *) Fix BN_gcd so that it works on negative inputs; the result is
7031 positive unless both parameters are zero.
7032 Previously something reasonably close to an infinite loop was
7033 possible because numbers could be growing instead of shrinking
7034 in the implementation of Euclid's algorithm.
7037 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7038 sign of the number in question.
7040 Fix BN_is_word(a,w) to work correctly for w == 0.
7042 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7043 because its test if the absolute value of 'a' equals 'w'.
7044 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7045 it exists mostly for use in the implementations of BN_is_zero(),
7046 BN_is_one(), and BN_is_word().
7049 *) New function BN_swap.
7052 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7053 the exponentiation functions are more likely to produce reasonable
7054 results on negative inputs.
7057 *) Change BN_mod_mul so that the result is always non-negative.
7058 Previously, it could be negative if one of the factors was negative;
7059 I don't think anyone really wanted that behaviour.
7062 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7063 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7064 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7065 and add new functions:
7074 BN_mod_lshift1_quick
7078 These functions always generate non-negative results.
7080 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7081 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7083 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7084 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7085 be reduced modulo m.
7086 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7089 The following entry accidentally appeared in the CHANGES file
7090 distributed with OpenSSL 0.9.7. The modifications described in
7091 it do *not* apply to OpenSSL 0.9.7.
7093 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7094 was actually never needed) and in BN_mul(). The removal in BN_mul()
7095 required a small change in bn_mul_part_recursive() and the addition
7096 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7097 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7098 bn_sub_words() and bn_add_words() except they take arrays with
7103 *) In 'openssl passwd', verify passwords read from the terminal
7104 unless the '-salt' option is used (which usually means that
7105 verification would just waste user's time since the resulting
7106 hash is going to be compared with some given password hash)
7107 or the new '-noverify' option is used.
7109 This is an incompatible change, but it does not affect
7110 non-interactive use of 'openssl passwd' (passwords on the command
7111 line, '-stdin' option, '-in ...' option) and thus should not
7115 *) Remove all references to RSAref, since there's no more need for it.
7118 *) Make DSO load along a path given through an environment variable
7119 (SHLIB_PATH) with shl_load().
7122 *) Constify the ENGINE code as a result of BIGNUM constification.
7123 Also constify the RSA code and most things related to it. In a
7124 few places, most notable in the depth of the ASN.1 code, ugly
7125 casts back to non-const were required (to be solved at a later
7129 *) Make it so the openssl application has all engines loaded by default.
7132 *) Constify the BIGNUM routines a little more.
7135 *) Add the following functions:
7137 ENGINE_load_cswift()
7139 ENGINE_load_atalla()
7141 ENGINE_load_builtin_engines()
7143 That way, an application can itself choose if external engines that
7144 are built-in in OpenSSL shall ever be used or not. The benefit is
7145 that applications won't have to be linked with libdl or other dso
7146 libraries unless it's really needed.
7148 Changed 'openssl engine' to load all engines on demand.
7149 Changed the engine header files to avoid the duplication of some
7150 declarations (they differed!).
7153 *) 'openssl engine' can now list capabilities.
7156 *) Better error reporting in 'openssl engine'.
7159 *) Never call load_dh_param(NULL) in s_server.
7162 *) Add engine application. It can currently list engines by name and
7163 identity, and test if they are actually available.
7166 *) Improve RPM specification file by forcing symbolic linking and making
7167 sure the installed documentation is also owned by root.root.
7168 [Damien Miller <djm@mindrot.org>]
7170 *) Give the OpenSSL applications more possibilities to make use of
7171 keys (public as well as private) handled by engines.
7174 *) Add OCSP code that comes from CertCo.
7177 *) Add VMS support for the Rijndael code.
7180 *) Added untested support for Nuron crypto accelerator.
7183 *) Add support for external cryptographic devices. This code was
7184 previously distributed separately as the "engine" branch.
7185 [Geoff Thorpe, Richard Levitte]
7187 *) Rework the filename-translation in the DSO code. It is now possible to
7188 have far greater control over how a "name" is turned into a filename
7189 depending on the operating environment and any oddities about the
7190 different shared library filenames on each system.
7193 *) Support threads on FreeBSD-elf in Configure.
7196 *) Fix for SHA1 assembly problem with MASM: it produces
7197 warnings about corrupt line number information when assembling
7198 with debugging information. This is caused by the overlapping
7200 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7203 NCONF_get_number() has no error checking at all. As a replacement,
7204 NCONF_get_number_e() is defined (_e for "error checking") and is
7205 promoted strongly. The old NCONF_get_number is kept around for
7206 binary backward compatibility.
7207 Make it possible for methods to load from something other than a BIO,
7208 by providing a function pointer that is given a name instead of a BIO.
7209 For example, this could be used to load configuration data from an
7213 *) Fix for non blocking accept BIOs. Added new I/O special reason
7214 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7215 with non blocking I/O was not possible because no retry code was
7216 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7220 *) Added the beginnings of Rijndael support.
7223 *) Fix for bug in DirectoryString mask setting. Add support for
7224 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7225 to allow certificate printing to more controllable, additional
7226 'certopt' option to 'x509' to allow new printing options to be
7230 *) Clean old EAY MD5 hack from e_os.h.
7233 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7235 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7236 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7237 [Joe Orton, Steve Henson]
7239 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7241 *) Fix additional bug revealed by the NISCC test suite:
7243 Stop bug triggering large recursion when presented with
7244 certain ASN.1 tags (CVE-2003-0851)
7247 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7249 *) Fix various bugs revealed by running the NISCC test suite:
7251 Stop out of bounds reads in the ASN1 code when presented with
7252 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7254 If verify callback ignores invalid public key errors don't try to check
7255 certificate signature with the NULL public key.
7259 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7260 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7264 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7265 extra data after the compression methods not only for TLS 1.0
7266 but also for SSL 3.0 (as required by the specification).
7267 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7269 *) Change X509_certificate_type() to mark the key as exported/exportable
7270 when it's 512 *bits* long, not 512 bytes.
7273 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7275 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7276 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7277 a protocol version number mismatch like a decryption error
7278 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7281 *) Turn on RSA blinding by default in the default implementation
7282 to avoid a timing attack. Applications that don't want it can call
7283 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7284 They would be ill-advised to do so in most cases.
7285 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7287 *) Change RSA blinding code so that it works when the PRNG is not
7288 seeded (in this case, the secret RSA exponent is abused as
7289 an unpredictable seed -- if it is not unpredictable, there
7290 is no point in blinding anyway). Make RSA blinding thread-safe
7291 by remembering the creator's thread ID in rsa->blinding and
7292 having all other threads use local one-time blinding factors
7293 (this requires more computation than sharing rsa->blinding, but
7294 avoids excessive locking; and if an RSA object is not shared
7295 between threads, blinding will still be very fast).
7298 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7300 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7301 via timing by performing a MAC computation even if incorrrect
7302 block cipher padding has been found. This is a countermeasure
7303 against active attacks where the attacker has to distinguish
7304 between bad padding and a MAC verification error. (CVE-2003-0078)
7306 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7307 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7308 Martin Vuagnoux (EPFL, Ilion)]
7310 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7312 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7313 memory from it's contents. This is done with a counter that will
7314 place alternating values in each byte. This can be used to solve
7315 two issues: 1) the removal of calls to memset() by highly optimizing
7316 compilers, and 2) cleansing with other values than 0, since those can
7317 be read through on certain media, for example a swap space on disk.
7320 *) Bugfix: client side session caching did not work with external caching,
7321 because the session->cipher setting was not restored when reloading
7322 from the external cache. This problem was masked, when
7323 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7324 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7327 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7328 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7329 [Zeev Lieber <zeev-l@yahoo.com>]
7331 *) Undo an undocumented change introduced in 0.9.6e which caused
7332 repeated calls to OpenSSL_add_all_ciphers() and
7333 OpenSSL_add_all_digests() to be ignored, even after calling
7337 *) Change the default configuration reader to deal with last line not
7338 being properly terminated.
7341 *) Change X509_NAME_cmp() so it applies the special rules on handling
7342 DN values that are of type PrintableString, as well as RDNs of type
7343 emailAddress where the value has the type ia5String.
7344 [stefank@valicert.com via Richard Levitte]
7346 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7347 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7348 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7349 the bitwise-OR of the two for use by the majority of applications
7350 wanting this behaviour, and update the docs. The documented
7351 behaviour and actual behaviour were inconsistent and had been
7352 changing anyway, so this is more a bug-fix than a behavioural
7354 [Geoff Thorpe, diagnosed by Nadav Har'El]
7356 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7357 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7360 *) Fix initialization code race conditions in
7361 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7362 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7363 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7364 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7365 ssl2_get_cipher_by_char(),
7366 ssl3_get_cipher_by_char().
7367 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7369 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7370 the cached sessions are flushed, as the remove_cb() might use ex_data
7371 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7372 (see [openssl.org #212]).
7373 [Geoff Thorpe, Lutz Jaenicke]
7375 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7376 length, instead of the encoding length to d2i_ASN1_OBJECT.
7379 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7381 *) [In 0.9.6g-engine release:]
7382 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7383 [Lynn Gazis <lgazis@rainbow.com>]
7385 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7387 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7388 and get fix the header length calculation.
7389 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7390 Alon Kantor <alonk@checkpoint.com> (and others),
7393 *) Use proper error handling instead of 'assertions' in buffer
7394 overflow checks added in 0.9.6e. This prevents DoS (the
7395 assertions could call abort()).
7396 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7398 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7400 *) Add various sanity checks to asn1_get_length() to reject
7401 the ASN1 length bytes if they exceed sizeof(long), will appear
7402 negative or the content length exceeds the length of the
7404 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7406 *) Fix cipher selection routines: ciphers without encryption had no flags
7407 for the cipher strength set and where therefore not handled correctly
7408 by the selection routines (PR #130).
7411 *) Fix EVP_dsa_sha macro.
7415 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7416 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7417 that was added in OpenSSL 0.9.6d.
7419 As the countermeasure turned out to be incompatible with some
7420 broken SSL implementations, the new option is part of SSL_OP_ALL.
7421 SSL_OP_ALL is usually employed when compatibility with weird SSL
7422 implementations is desired (e.g. '-bugs' option to 's_client' and
7423 's_server'), so the new option is automatically set in many
7427 *) Changes in security patch:
7429 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7430 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7431 Air Force Materiel Command, USAF, under agreement number
7434 *) Add various sanity checks to asn1_get_length() to reject
7435 the ASN1 length bytes if they exceed sizeof(long), will appear
7436 negative or the content length exceeds the length of the
7437 supplied buffer. (CVE-2002-0659)
7438 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7440 *) Assertions for various potential buffer overflows, not known to
7442 [Ben Laurie (CHATS)]
7444 *) Various temporary buffers to hold ASCII versions of integers were
7445 too small for 64 bit platforms. (CVE-2002-0655)
7446 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7448 *) Remote buffer overflow in SSL3 protocol - an attacker could
7449 supply an oversized session ID to a client. (CVE-2002-0656)
7450 [Ben Laurie (CHATS)]
7452 *) Remote buffer overflow in SSL2 protocol - an attacker could
7453 supply an oversized client master key. (CVE-2002-0656)
7454 [Ben Laurie (CHATS)]
7456 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7458 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7459 encoded as NULL) with id-dsa-with-sha1.
7460 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7462 *) Check various X509_...() return values in apps/req.c.
7463 [Nils Larsch <nla@trustcenter.de>]
7465 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7466 an end-of-file condition would erronously be flagged, when the CRLF
7467 was just at the end of a processed block. The bug was discovered when
7468 processing data through a buffering memory BIO handing the data to a
7469 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7470 <ptsekov@syntrex.com> and Nedelcho Stanev.
7473 *) Implement a countermeasure against a vulnerability recently found
7474 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7475 before application data chunks to avoid the use of known IVs
7476 with data potentially chosen by the attacker.
7479 *) Fix length checks in ssl3_get_client_hello().
7482 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7483 to prevent ssl3_read_internal() from incorrectly assuming that
7484 ssl3_read_bytes() found application data while handshake
7485 processing was enabled when in fact s->s3->in_read_app_data was
7486 merely automatically cleared during the initial handshake.
7487 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7489 *) Fix object definitions for Private and Enterprise: they were not
7490 recognized in their shortname (=lowercase) representation. Extend
7491 obj_dat.pl to issue an error when using undefined keywords instead
7492 of silently ignoring the problem (Svenning Sorensen
7493 <sss@sss.dnsalias.net>).
7496 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7497 generators, i.e. generators other than 2 and 5. (Previously, the
7498 code did not properly initialise the 'add' and 'rem' values to
7499 BN_generate_prime().)
7501 In the new general case, we do not insist that 'generator' is
7502 actually a primitive root: This requirement is rather pointless;
7503 a generator of the order-q subgroup is just as good, if not
7507 *) Map new X509 verification errors to alerts. Discovered and submitted by
7508 Tom Wu <tom@arcot.com>.
7511 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7512 returning non-zero before the data has been completely received
7513 when using non-blocking I/O.
7514 [Bodo Moeller; problem pointed out by John Hughes]
7516 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7517 [Ben Laurie, Lutz Jaenicke]
7519 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7520 Yoram Zahavi <YoramZ@gilian.com>).
7523 *) Add information about CygWin 1.3 and on, and preserve proper
7524 configuration for the versions before that.
7525 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7527 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7528 check whether we deal with a copy of a session and do not delete from
7529 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7530 <izhar@checkpoint.com>.
7533 *) Do not store session data into the internal session cache, if it
7534 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7535 flag is set). Proposed by Aslam <aslam@funk.com>.
7538 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7542 *) [In 0.9.6d-engine release:]
7543 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7544 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7546 *) Add the configuration target linux-s390x.
7547 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7549 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7550 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7551 variable as an indication that a ClientHello message has been
7552 received. As the flag value will be lost between multiple
7553 invocations of ssl3_accept when using non-blocking I/O, the
7554 function may not be aware that a handshake has actually taken
7555 place, thus preventing a new session from being added to the
7558 To avoid this problem, we now set s->new_session to 2 instead of
7559 using a local variable.
7560 [Lutz Jaenicke, Bodo Moeller]
7562 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7563 if the SSL_R_LENGTH_MISMATCH error is detected.
7564 [Geoff Thorpe, Bodo Moeller]
7566 *) New 'shared_ldflag' column in Configure platform table.
7569 *) Fix EVP_CIPHER_mode macro.
7570 ["Dan S. Camper" <dan@bti.net>]
7572 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7573 type, we must throw them away by setting rr->length to 0.
7574 [D P Chang <dpc@qualys.com>]
7576 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7578 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7579 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7580 worked incorrectly for those cases where range = 10..._2 and
7581 3*range is two bits longer than range.)
7584 *) Only add signing time to PKCS7 structures if it is not already
7588 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7589 OBJ_ld_ce should be OBJ_id_ce.
7590 Also some ip-pda OIDs in crypto/objects/objects.txt were
7591 incorrect (cf. RFC 3039).
7592 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7594 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7595 returns early because it has nothing to do.
7596 [Andy Schneider <andy.schneider@bjss.co.uk>]
7598 *) [In 0.9.6c-engine release:]
7599 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7600 [Andy Schneider <andy.schneider@bjss.co.uk>]
7602 *) [In 0.9.6c-engine release:]
7603 Add support for Cryptographic Appliance's keyserver technology.
7604 (Use engine 'keyclient')
7605 [Cryptographic Appliances and Geoff Thorpe]
7607 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7608 is called via tools/c89.sh because arguments have to be
7609 rearranged (all '-L' options must appear before the first object
7611 [Richard Shapiro <rshapiro@abinitio.com>]
7613 *) [In 0.9.6c-engine release:]
7614 Add support for Broadcom crypto accelerator cards, backported
7616 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7618 *) [In 0.9.6c-engine release:]
7619 Add support for SureWare crypto accelerator cards from
7620 Baltimore Technologies. (Use engine 'sureware')
7621 [Baltimore Technologies and Mark Cox]
7623 *) [In 0.9.6c-engine release:]
7624 Add support for crypto accelerator cards from Accelerated
7625 Encryption Processing, www.aep.ie. (Use engine 'aep')
7626 [AEP Inc. and Mark Cox]
7628 *) Add a configuration entry for gcc on UnixWare.
7629 [Gary Benson <gbenson@redhat.com>]
7631 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7632 messages are stored in a single piece (fixed-length part and
7633 variable-length part combined) and fix various bugs found on the way.
7636 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7637 instead. BIO_gethostbyname() does not know what timeouts are
7638 appropriate, so entries would stay in cache even when they have
7640 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7642 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7643 faced with a pathologically small ClientHello fragment that does
7644 not contain client_version: Instead of aborting with an error,
7645 simply choose the highest available protocol version (i.e.,
7646 TLS 1.0 unless it is disabled). In practice, ClientHello
7647 messages are never sent like this, but this change gives us
7648 strictly correct behaviour at least for TLS.
7651 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7652 never resets s->method to s->ctx->method when called from within
7653 one of the SSL handshake functions.
7654 [Bodo Moeller; problem pointed out by Niko Baric]
7656 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7657 (sent using the client's version number) if client_version is
7658 smaller than the protocol version in use. Also change
7659 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7660 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7661 the client will at least see that alert.
7664 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7668 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7669 client receives HelloRequest while in a handshake.
7670 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7672 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7673 should end in 'break', not 'goto end' which circumvents various
7674 cleanups done in state SSL_ST_OK. But session related stuff
7675 must be disabled for SSL_ST_OK in the case that we just sent a
7678 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7679 before just sending a HelloRequest.
7680 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7682 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7683 reveal whether illegal block cipher padding was found or a MAC
7684 verification error occurred. (Neither SSLerr() codes nor alerts
7685 are directly visible to potential attackers, but the information
7686 may leak via logfiles.)
7688 Similar changes are not required for the SSL 2.0 implementation
7689 because the number of padding bytes is sent in clear for SSL 2.0,
7690 and the extra bytes are just ignored. However ssl/s2_pkt.c
7691 failed to verify that the purported number of padding bytes is in
7695 *) Add OpenUNIX-8 support including shared libraries
7696 (Boyd Lynn Gerber <gerberb@zenez.com>).
7699 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7700 'wristwatch attack' using huge encoding parameters (cf.
7701 James H. Manger's CRYPTO 2001 paper). Note that the
7702 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7703 encoding parameters and hence was not vulnerable.
7706 *) BN_sqr() bug fix.
7707 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7709 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7710 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7711 followed by modular reduction.
7712 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7714 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7715 equivalent based on BN_pseudo_rand() instead of BN_rand().
7718 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7719 This function was broken, as the check for a new client hello message
7720 to handle SGC did not allow these large messages.
7721 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7724 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7727 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7728 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7731 *) Rework the configuration and shared library support for Tru64 Unix.
7732 The configuration part makes use of modern compiler features and
7733 still retains old compiler behavior for those that run older versions
7734 of the OS. The shared library support part includes a variant that
7735 uses the RPATH feature, and is available through the special
7736 configuration target "alpha-cc-rpath", which will never be selected
7738 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7740 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7741 with the same message size as in ssl3_get_certificate_request().
7742 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7743 messages might inadvertently be reject as too long.
7744 [Petr Lampa <lampa@fee.vutbr.cz>]
7746 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7749 *) Modified SSL library such that the verify_callback that has been set
7750 specificly for an SSL object with SSL_set_verify() is actually being
7751 used. Before the change, a verify_callback set with this function was
7752 ignored and the verify_callback() set in the SSL_CTX at the time of
7753 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7754 to allow the necessary settings.
7757 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7758 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7759 done automatically (in contradiction to the requirements of the C
7760 standard). This made problems when used from OpenSSH.
7763 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7764 dh->length and always used
7766 BN_rand_range(priv_key, dh->p).
7768 BN_rand_range() is not necessary for Diffie-Hellman, and this
7769 specific range makes Diffie-Hellman unnecessarily inefficient if
7770 dh->length (recommended exponent length) is much smaller than the
7771 length of dh->p. We could use BN_rand_range() if the order of
7772 the subgroup was stored in the DH structure, but we only have
7777 BN_rand(priv_key, l, ...)
7779 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7785 RSA_eay_public_encrypt
7786 RSA_eay_private_decrypt
7787 RSA_eay_private_encrypt (signing)
7788 RSA_eay_public_decrypt (signature verification)
7790 (default implementations for RSA_public_encrypt,
7791 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7792 always reject numbers >= n.
7795 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7796 to synchronize access to 'locking_thread'. This is necessary on
7797 systems where access to 'locking_thread' (an 'unsigned long'
7798 variable) is not atomic.
7801 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7802 *before* setting the 'crypto_lock_rand' flag. The previous code had
7803 a race condition if 0 is a valid thread ID.
7804 [Travis Vitek <vitek@roguewave.com>]
7806 *) Add support for shared libraries under Irix.
7807 [Albert Chin-A-Young <china@thewrittenword.com>]
7809 *) Add configuration option to build on Linux on both big-endian and
7811 [Ralf Baechle <ralf@uni-koblenz.de>]
7813 *) Add the possibility to create shared libraries on HP-UX.
7816 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7818 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7819 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7820 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7821 PRNG state recovery was possible based on the output of
7822 one PRNG request appropriately sized to gain knowledge on
7823 'md' followed by enough consecutive 1-byte PRNG requests
7824 to traverse all of 'state'.
7826 1. When updating 'md_local' (the current thread's copy of 'md')
7827 during PRNG output generation, hash all of the previous
7828 'md_local' value, not just the half used for PRNG output.
7830 2. Make the number of bytes from 'state' included into the hash
7831 independent from the number of PRNG bytes requested.
7833 The first measure alone would be sufficient to avoid
7834 Markku-Juhani's attack. (Actually it had never occurred
7835 to me that the half of 'md_local' used for chaining was the
7836 half from which PRNG output bytes were taken -- I had always
7837 assumed that the secret half would be used.) The second
7838 measure makes sure that additional data from 'state' is never
7839 mixed into 'md_local' in small portions; this heuristically
7840 further strengthens the PRNG.
7843 *) Fix crypto/bn/asm/mips3.s.
7846 *) When only the key is given to "enc", the IV is undefined. Print out
7847 an error message in this case.
7850 *) Handle special case when X509_NAME is empty in X509 printing routines.
7853 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7854 positive and less than q.
7857 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7858 used: it isn't thread safe and the add_lock_callback should handle
7860 [Paul Rose <Paul.Rose@bridge.com>]
7862 *) Verify that incoming data obeys the block size in
7863 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7867 [Ulf Möller, Bodo Möller]
7869 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7870 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7871 when fixing the server behaviour for backwards-compatible 'client
7872 hello' messages. (Note that the attack is impractical against
7873 SSL 3.0 and TLS 1.0 anyway because length and version checking
7874 means that the probability of guessing a valid ciphertext is
7875 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7878 Before 0.9.5, the countermeasure (hide the error by generating a
7879 random 'decryption result') did not work properly because
7880 ERR_clear_error() was missing, meaning that SSL_get_error() would
7881 detect the supposedly ignored error.
7883 Both problems are now fixed.
7886 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7887 (previously it was 1024).
7890 *) Fix for compatibility mode trust settings: ignore trust settings
7891 unless some valid trust or reject settings are present.
7894 *) Fix for blowfish EVP: its a variable length cipher.
7897 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7898 parameters in DSA public key structures and return an error in the
7899 DSA routines if parameters are absent.
7902 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7903 in the current directory if neither $RANDFILE nor $HOME was set.
7904 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7905 caused some confusion to Windows users who haven't defined $HOME.
7906 Thus RAND_file_name() is changed again: e_os.h can define a
7907 DEFAULT_HOME, which will be used if $HOME is not set.
7908 For Windows, we use "C:"; on other platforms, we still require
7909 environment variables.
7911 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7912 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7913 having multiple threads call RAND_poll() concurrently.
7916 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7917 combination of a flag and a thread ID variable.
7918 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7919 flag), *other* threads can enter ssleay_add_bytes without obeying
7920 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7921 that they do not hold after the first thread unsets add_do_not_lock).
7924 *) Change bctest again: '-x' expressions are not available in all
7928 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7930 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7931 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7933 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7934 the default extension for executables, if any. Also, make the perl
7935 scripts that use symlink() to test if it really exists and use "cp"
7936 if it doesn't. All this made OpenSSL compilable and installable in
7940 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7941 If SEQUENCE is length is indefinite just set c->slen to the total
7942 amount of data available.
7943 [Steve Henson, reported by shige@FreeBSD.org]
7944 [This change does not apply to 0.9.7.]
7946 *) Change bctest to avoid here-documents inside command substitution
7947 (workaround for FreeBSD /bin/sh bug).
7948 For compatibility with Ultrix, avoid shell functions (introduced
7949 in the bctest version that searches along $PATH).
7952 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7953 with des_encrypt() defined on some operating systems, like Solaris
7957 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7958 On the Importance of Eliminating Errors in Cryptographic
7959 Computations, J. Cryptology 14 (2001) 2, 101-119,
7960 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7963 *) MIPS assembler BIGNUM division bug fix.
7966 *) Disabled incorrect Alpha assembler code.
7969 *) Fix PKCS#7 decode routines so they correctly update the length
7970 after reading an EOC for the EXPLICIT tag.
7972 [This change does not apply to 0.9.7.]
7974 *) Fix bug in PKCS#12 key generation routines. This was triggered
7975 if a 3DES key was generated with a 0 initial byte. Include
7976 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7977 (but broken) behaviour.
7980 *) Enhance bctest to search for a working bc along $PATH and print
7982 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7984 *) Fix memory leaks in err.c: free err_data string if necessary;
7985 don't write to the wrong index in ERR_set_error_data.
7988 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7992 *) Replace rdtsc with _emit statements for VC++ version 5.
7993 [Jeremy Cooper <jeremy@baymoo.org>]
7995 *) Make it possible to reuse SSLv2 sessions.
7998 *) In copy_email() check for >= 0 as a return value for
7999 X509_NAME_get_index_by_NID() since 0 is a valid index.
8000 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8002 *) Avoid coredump with unsupported or invalid public keys by checking if
8003 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8004 PKCS7_verify() fails with non detached data.
8007 *) Don't use getenv in library functions when run as setuid/setgid.
8008 New function OPENSSL_issetugid().
8011 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8012 due to incorrect handling of multi-threading:
8014 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8016 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8018 3. Count how many times MemCheck_off() has been called so that
8019 nested use can be treated correctly. This also avoids
8020 inband-signalling in the previous code (which relied on the
8021 assumption that thread ID 0 is impossible).
8024 *) Add "-rand" option also to s_client and s_server.
8027 *) Fix CPU detection on Irix 6.x.
8028 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8029 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8031 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8034 [This change does not apply to 0.9.7.]
8036 *) Use the cached encoding of an X509_NAME structure rather than
8037 copying it. This is apparently the reason for the libsafe "errors"
8038 but the code is actually correct.
8041 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8042 Bleichenbacher's DSA attack.
8043 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8044 to be set and top=0 forces the highest bit to be set; top=-1 is new
8045 and leaves the highest bit random.
8046 [Ulf Moeller, Bodo Moeller]
8048 *) In the NCONF_...-based implementations for CONF_... queries
8049 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8050 a temporary CONF structure with the data component set to NULL
8051 (which gives segmentation faults in lh_retrieve).
8052 Instead, use NULL for the CONF pointer in CONF_get_string and
8053 CONF_get_number (which may use environment variables) and directly
8054 return NULL from CONF_get_section.
8057 *) Fix potential buffer overrun for EBCDIC.
8060 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8061 keyUsage if basicConstraints absent for a CA.
8064 *) Make SMIME_write_PKCS7() write mail header values with a format that
8065 is more generally accepted (no spaces before the semicolon), since
8066 some programs can't parse those values properly otherwise. Also make
8067 sure BIO's that break lines after each write do not create invalid
8071 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8072 macros previously used would not encode an empty SEQUENCE OF
8073 and break the signature.
8075 [This change does not apply to 0.9.7.]
8077 *) Zero the premaster secret after deriving the master secret in
8081 *) Add some EVP_add_digest_alias registrations (as found in
8082 OpenSSL_add_all_digests()) to SSL_library_init()
8083 aka OpenSSL_add_ssl_algorithms(). This provides improved
8084 compatibility with peers using X.509 certificates
8085 with unconventional AlgorithmIdentifier OIDs.
8088 *) Fix for Irix with NO_ASM.
8089 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8091 *) ./config script fixes.
8092 [Ulf Moeller, Richard Levitte]
8094 *) Fix 'openssl passwd -1'.
8097 *) Change PKCS12_key_gen_asc() so it can cope with non null
8098 terminated strings whose length is passed in the passlen
8099 parameter, for example from PEM callbacks. This was done
8100 by adding an extra length parameter to asc2uni().
8101 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8103 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8104 call failed, free the DSA structure.
8107 *) Fix to uni2asc() to cope with zero length Unicode strings.
8108 These are present in some PKCS#12 files.
8111 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8112 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8113 when writing a 32767 byte record.
8114 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8116 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8117 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8119 (RSA objects have a reference count access to which is protected
8120 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8121 so they are meant to be shared between threads.)
8122 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8123 "Reddie, Steven" <Steven.Reddie@ca.com>]
8125 *) Fix a deadlock in CRYPTO_mem_leaks().
8128 *) Use better test patterns in bntest.
8131 *) rand_win.c fix for Borland C.
8134 *) BN_rshift bugfix for n == 0.
8137 *) Add a 'bctest' script that checks for some known 'bc' bugs
8138 so that 'make test' does not abort just because 'bc' is broken.
8141 *) Store verify_result within SSL_SESSION also for client side to
8142 avoid potential security hole. (Re-used sessions on the client side
8143 always resulted in verify_result==X509_V_OK, not using the original
8144 result of the server certificate verification.)
8147 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8148 SSL3_RT_APPLICATION_DATA, return 0.
8149 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8153 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8154 releases, have been re-implemented by renaming the previous
8155 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8156 and ssl3_read_internal, respectively, and adding 'peek' parameters
8157 to them. The new ssl[23]_{read,peek} functions are calls to
8158 ssl[23]_read_internal with the 'peek' flag set appropriately.
8159 A 'peek' parameter has also been added to ssl3_read_bytes, which
8160 does the actual work for ssl3_read_internal.
8163 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8164 the method-specific "init()" handler. Also clean up ex_data after
8165 calling the method-specific "finish()" handler. Previously, this was
8166 happening the other way round.
8169 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8170 The previous value, 12, was not always sufficient for BN_mod_exp().
8173 *) Make sure that shared libraries get the internal name engine with
8174 the full version number and not just 0. This should mark the
8175 shared libraries as not backward compatible. Of course, this should
8176 be changed again when we can guarantee backward binary compatibility.
8179 *) Fix typo in get_cert_by_subject() in by_dir.c
8180 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8182 *) Rework the system to generate shared libraries:
8184 - Make note of the expected extension for the shared libraries and
8185 if there is a need for symbolic links from for example libcrypto.so.0
8186 to libcrypto.so.0.9.7. There is extended info in Configure for
8189 - Make as few rebuilds of the shared libraries as possible.
8191 - Still avoid linking the OpenSSL programs with the shared libraries.
8193 - When installing, install the shared libraries separately from the
8197 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8199 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8200 and not in SSL_clear because the latter is also used by the
8201 accept/connect functions; previously, the settings made by
8202 SSL_set_read_ahead would be lost during the handshake.
8203 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8205 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8206 Previously, it would create entries for disabled algorithms no
8210 *) Added several new manual pages for SSL_* function.
8213 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8215 *) In ssl23_get_client_hello, generate an error message when faced
8216 with an initial SSL 3.0/TLS record that is too small to contain the
8217 first two bytes of the ClientHello message, i.e. client_version.
8218 (Note that this is a pathologic case that probably has never happened
8219 in real life.) The previous approach was to use the version number
8220 from the record header as a substitute; but our protocol choice
8221 should not depend on that one because it is not authenticated
8222 by the Finished messages.
8225 *) More robust randomness gathering functions for Windows.
8226 [Jeffrey Altman <jaltman@columbia.edu>]
8228 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8229 not set then we don't setup the error code for issuer check errors
8230 to avoid possibly overwriting other errors which the callback does
8231 handle. If an application does set the flag then we assume it knows
8232 what it is doing and can handle the new informational codes
8236 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8237 a general "ANY" type, as such it should be able to decode anything
8238 including tagged types. However it didn't check the class so it would
8239 wrongly interpret tagged types in the same way as their universal
8240 counterpart and unknown types were just rejected. Changed so that the
8241 tagged and unknown types are handled in the same way as a SEQUENCE:
8242 that is the encoding is stored intact. There is also a new type
8243 "V_ASN1_OTHER" which is used when the class is not universal, in this
8244 case we have no idea what the actual type is so we just lump them all
8248 *) On VMS, stdout may very well lead to a file that is written to
8249 in a record-oriented fashion. That means that every write() will
8250 write a separate record, which will be read separately by the
8251 programs trying to read from it. This can be very confusing.
8253 The solution is to put a BIO filter in the way that will buffer
8254 text until a linefeed is reached, and then write everything a
8255 line at a time, so every record written will be an actual line,
8256 not chunks of lines and not (usually doesn't happen, but I've
8257 seen it once) several lines in one record. BIO_f_linebuffer() is
8260 Currently, it's a VMS-only method, because that's where it has
8261 been tested well enough.
8264 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8265 it can return incorrect results.
8266 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8267 but it was in 0.9.6-beta[12].)
8270 *) Disable the check for content being present when verifying detached
8271 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8272 include zero length content when signing messages.
8275 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8276 BIO_ctrl (for BIO pairs).
8279 *) Add DSO method for VMS.
8282 *) Bug fix: Montgomery multiplication could produce results with the
8286 *) Add RPM specification openssl.spec and modify it to build three
8287 packages. The default package contains applications, application
8288 documentation and run-time libraries. The devel package contains
8289 include files, static libraries and function documentation. The
8290 doc package contains the contents of the doc directory. The original
8291 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8294 *) Add a large number of documentation files for many SSL routines.
8295 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8297 *) Add a configuration entry for Sony News 4.
8298 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8300 *) Don't set the two most significant bits to one when generating a
8301 random number < q in the DSA library.
8304 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8305 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8306 the underlying transport is blocking) if a handshake took place.
8307 (The default behaviour is needed by applications such as s_client
8308 and s_server that use select() to determine when to use SSL_read;
8309 but for applications that know in advance when to expect data, it
8310 just makes things more complicated.)
8313 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8317 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8318 work better on such systems.
8319 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8321 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8322 Update PKCS12_parse() so it copies the friendlyName and the
8323 keyid to the certificates aux info.
8326 *) Fix bug in PKCS7_verify() which caused an infinite loop
8327 if there was more than one signature.
8328 [Sven Uszpelkat <su@celocom.de>]
8330 *) Major change in util/mkdef.pl to include extra information
8331 about each symbol, as well as presenting variables as well
8332 as functions. This change means that there's n more need
8333 to rebuild the .num files when some algorithms are excluded.
8336 *) Allow the verify time to be set by an application,
8337 rather than always using the current time.
8340 *) Phase 2 verify code reorganisation. The certificate
8341 verify code now looks up an issuer certificate by a
8342 number of criteria: subject name, authority key id
8343 and key usage. It also verifies self signed certificates
8344 by the same criteria. The main comparison function is
8345 X509_check_issued() which performs these checks.
8347 Lot of changes were necessary in order to support this
8348 without completely rewriting the lookup code.
8350 Authority and subject key identifier are now cached.
8352 The LHASH 'certs' is X509_STORE has now been replaced
8353 by a STACK_OF(X509_OBJECT). This is mainly because an
8354 LHASH can't store or retrieve multiple objects with
8355 the same hash value.
8357 As a result various functions (which were all internal
8358 use only) have changed to handle the new X509_STORE
8359 structure. This will break anything that messed round
8360 with X509_STORE internally.
8362 The functions X509_STORE_add_cert() now checks for an
8363 exact match, rather than just subject name.
8365 The X509_STORE API doesn't directly support the retrieval
8366 of multiple certificates matching a given criteria, however
8367 this can be worked round by performing a lookup first
8368 (which will fill the cache with candidate certificates)
8369 and then examining the cache for matches. This is probably
8370 the best we can do without throwing out X509_LOOKUP
8371 entirely (maybe later...).
8373 The X509_VERIFY_CTX structure has been enhanced considerably.
8375 All certificate lookup operations now go via a get_issuer()
8376 callback. Although this currently uses an X509_STORE it
8377 can be replaced by custom lookups. This is a simple way
8378 to bypass the X509_STORE hackery necessary to make this
8379 work and makes it possible to use more efficient techniques
8380 in future. A very simple version which uses a simple
8381 STACK for its trusted certificate store is also provided
8382 using X509_STORE_CTX_trusted_stack().
8384 The verify_cb() and verify() callbacks now have equivalents
8385 in the X509_STORE_CTX structure.
8387 X509_STORE_CTX also has a 'flags' field which can be used
8388 to customise the verify behaviour.
8391 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8392 excludes S/MIME capabilities.
8395 *) When a certificate request is read in keep a copy of the
8396 original encoding of the signed data and use it when outputing
8397 again. Signatures then use the original encoding rather than
8398 a decoded, encoded version which may cause problems if the
8399 request is improperly encoded.
8402 *) For consistency with other BIO_puts implementations, call
8403 buffer_write(b, ...) directly in buffer_puts instead of calling
8406 In BIO_puts, increment b->num_write as in BIO_write.
8407 [Peter.Sylvester@EdelWeb.fr]
8409 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8410 BN_zero, we may not return a BIGNUM with an array consisting of
8414 *) Avoid calling abort() from within the library when problems are
8415 detected, except if preprocessor symbols have been defined
8416 (such as REF_CHECK, BN_DEBUG etc.).
8419 *) New openssl application 'rsautl'. This utility can be
8420 used for low level RSA operations. DER public key
8421 BIO/fp routines also added.
8424 *) New Configure entry and patches for compiling on QNX 4.
8425 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8427 *) A demo state-machine implementation was sponsored by
8428 Nuron (http://www.nuron.com/) and is now available in
8429 demos/state_machine.
8432 *) New options added to the 'dgst' utility for signature
8433 generation and verification.
8436 *) Unrecognized PKCS#7 content types are now handled via a
8437 catch all ASN1_TYPE structure. This allows unsupported
8438 types to be stored as a "blob" and an application can
8439 encode and decode it manually.
8442 *) Fix various signed/unsigned issues to make a_strex.c
8444 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8446 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8447 length if passed a buffer. ASN1_INTEGER_to_BN failed
8448 if passed a NULL BN and its argument was negative.
8449 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8451 *) Modification to PKCS#7 encoding routines to output definite
8452 length encoding. Since currently the whole structures are in
8453 memory there's not real point in using indefinite length
8454 constructed encoding. However if OpenSSL is compiled with
8455 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8458 *) Added BIO_vprintf() and BIO_vsnprintf().
8461 *) Added more prefixes to parse for in the the strings written
8462 through a logging bio, to cover all the levels that are available
8463 through syslog. The prefixes are now:
8465 PANIC, EMERG, EMR => LOG_EMERG
8466 ALERT, ALR => LOG_ALERT
8467 CRIT, CRI => LOG_CRIT
8468 ERROR, ERR => LOG_ERR
8469 WARNING, WARN, WAR => LOG_WARNING
8470 NOTICE, NOTE, NOT => LOG_NOTICE
8471 INFO, INF => LOG_INFO
8472 DEBUG, DBG => LOG_DEBUG
8474 and as before, if none of those prefixes are present at the
8475 beginning of the string, LOG_ERR is chosen.
8477 On Win32, the LOG_* levels are mapped according to this:
8479 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8480 LOG_WARNING => EVENTLOG_WARNING_TYPE
8481 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8485 *) Made it possible to reconfigure with just the configuration
8486 argument "reconf" or "reconfigure". The command line arguments
8487 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8488 and are retrieved from there when reconfiguring.
8492 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8494 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8497 *) The obj_dat.pl script was messing up the sorting of object
8498 names. The reason was that it compared the quoted version
8499 of strings as a result "OCSP" > "OCSP Signing" because
8500 " > SPACE. Changed script to store unquoted versions of
8501 names and add quotes on output. It was also omitting some
8502 names from the lookup table if they were given a default
8503 value (that is if SN is missing it is given the same
8504 value as LN and vice versa), these are now added on the
8505 grounds that if an object has a name we should be able to
8506 look it up. Finally added warning output when duplicate
8507 short or long names are found.
8510 *) Changes needed for Tandem NSK.
8511 [Scott Uroff <scott@xypro.com>]
8513 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8514 RSA_padding_check_SSLv23(), special padding was never detected
8515 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8516 version rollback attacks was not effective.
8518 In s23_clnt.c, don't use special rollback-attack detection padding
8519 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8520 client; similarly, in s23_srvr.c, don't do the rollback check if
8521 SSL 2.0 is the only protocol enabled in the server.
8524 *) Make it possible to get hexdumps of unprintable data with 'openssl
8525 asn1parse'. By implication, the functions ASN1_parse_dump() and
8526 BIO_dump_indent() are added.
8529 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8530 these print out strings and name structures based on various
8531 flags including RFC2253 support and proper handling of
8532 multibyte characters. Added options to the 'x509' utility
8533 to allow the various flags to be set.
8536 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8537 Also change the functions X509_cmp_current_time() and
8538 X509_gmtime_adj() work with an ASN1_TIME structure,
8539 this will enable certificates using GeneralizedTime in validity
8540 dates to be checked.
8543 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8544 negative public key encodings) on by default,
8545 NO_NEG_PUBKEY_BUG can be set to disable it.
8548 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8549 content octets. An i2c_ASN1_OBJECT is unnecessary because
8550 the encoding can be trivially obtained from the structure.
8553 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8554 not read locks (CRYPTO_r_[un]lock).
8557 *) A first attempt at creating official support for shared
8558 libraries through configuration. I've kept it so the
8559 default is static libraries only, and the OpenSSL programs
8560 are always statically linked for now, but there are
8561 preparations for dynamic linking in place.
8562 This has been tested on Linux and Tru64.
8565 *) Randomness polling function for Win9x, as described in:
8566 Peter Gutmann, Software Generation of Practically Strong
8570 *) Fix so PRNG is seeded in req if using an already existing
8574 *) New options to smime application. -inform and -outform
8575 allow alternative formats for the S/MIME message including
8576 PEM and DER. The -content option allows the content to be
8577 specified separately. This should allow things like Netscape
8578 form signing output easier to verify.
8581 *) Fix the ASN1 encoding of tags using the 'long form'.
8584 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8585 STRING types. These convert content octets to and from the
8586 underlying type. The actual tag and length octets are
8587 already assumed to have been read in and checked. These
8588 are needed because all other string types have virtually
8589 identical handling apart from the tag. By having versions
8590 of the ASN1 functions that just operate on content octets
8591 IMPLICIT tagging can be handled properly. It also allows
8592 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8593 and ASN1_INTEGER are identical apart from the tag.
8596 *) Change the handling of OID objects as follows:
8598 - New object identifiers are inserted in objects.txt, following
8599 the syntax given in objects.README.
8600 - objects.pl is used to process obj_mac.num and create a new
8602 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8605 This is currently kind of a hack, and the perl code in objects.pl
8606 isn't very elegant, but it works as I intended. The simplest way
8607 to check that it worked correctly is to look in obj_dat.h and
8608 check the array nid_objs and make sure the objects haven't moved
8609 around (this is important!). Additions are OK, as well as
8610 consistent name changes.
8613 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8616 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8617 The given file adds to whatever has already been seeded into the
8618 random pool through the RANDFILE configuration file option or
8619 environment variable, or the default random state file.
8622 *) mkstack.pl now sorts each macro group into lexical order.
8623 Previously the output order depended on the order the files
8624 appeared in the directory, resulting in needless rewriting
8628 *) Patches to make OpenSSL compile under Win32 again. Mostly
8629 work arounds for the VC++ problem that it treats func() as
8630 func(void). Also stripped out the parts of mkdef.pl that
8631 added extra typesafe functions: these no longer exist.
8634 *) Reorganisation of the stack code. The macros are now all
8635 collected in safestack.h . Each macro is defined in terms of
8636 a "stack macro" of the form SKM_<name>(type, a, b). The
8637 DEBUG_SAFESTACK is now handled in terms of function casts,
8638 this has the advantage of retaining type safety without the
8639 use of additional functions. If DEBUG_SAFESTACK is not defined
8640 then the non typesafe macros are used instead. Also modified the
8641 mkstack.pl script to handle the new form. Needs testing to see
8642 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8643 the default if no major problems. Similar behaviour for ASN1_SET_OF
8644 and PKCS12_STACK_OF.
8647 *) When some versions of IIS use the 'NET' form of private key the
8648 key derivation algorithm is different. Normally MD5(password) is
8649 used as a 128 bit RC4 key. In the modified case
8650 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8651 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8652 as the old Netscape_RSA functions except they have an additional
8653 'sgckey' parameter which uses the modified algorithm. Also added
8654 an -sgckey command line option to the rsa utility. Thanks to
8655 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8656 algorithm to openssl-dev.
8659 *) The evp_local.h macros were using 'c.##kname' which resulted in
8660 invalid expansion on some systems (SCO 5.0.5 for example).
8661 Corrected to 'c.kname'.
8662 [Phillip Porch <root@theporch.com>]
8664 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8665 a STACK of email addresses from a certificate or request, these look
8666 in the subject name and the subject alternative name extensions and
8667 omit any duplicate addresses.
8670 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8671 This makes DSA verification about 2 % faster.
8674 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8675 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8676 plus overhead for 1024 bit moduli).
8677 This makes exponentiations about 0.5 % faster for 1024 bit
8678 exponents (as measured by "openssl speed rsa2048").
8681 *) Rename memory handling macros to avoid conflicts with other
8683 Malloc => OPENSSL_malloc
8684 Malloc_locked => OPENSSL_malloc_locked
8685 Realloc => OPENSSL_realloc
8686 Free => OPENSSL_free
8689 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8690 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8693 *) CygWin32 support.
8694 [John Jarvie <jjarvie@newsguy.com>]
8696 *) The type-safe stack code has been rejigged. It is now only compiled
8697 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8698 by default all type-specific stack functions are "#define"d back to
8699 standard stack functions. This results in more streamlined output
8700 but retains the type-safety checking possibilities of the original
8704 *) The STACK code has been cleaned up, and certain type declarations
8705 that didn't make a lot of sense have been brought in line. This has
8706 also involved a cleanup of sorts in safestack.h to more correctly
8707 map type-safe stack functions onto their plain stack counterparts.
8708 This work has also resulted in a variety of "const"ifications of
8709 lots of the code, especially "_cmp" operations which should normally
8710 be prototyped with "const" parameters anyway.
8713 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8714 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8715 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8716 where all of 'md' is used each time the PRNG is used, but 'state'
8717 is used only indexed by a cyclic counter. As entropy may not be
8718 well distributed from the beginning, 'md' is important as a
8719 chaining variable. However, the output function chains only half
8720 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8721 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8722 in all of 'state' being rewritten, with the new values depending
8723 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8726 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8727 the handshake is continued after ssl_verify_cert_chain();
8728 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8729 can lead to 'unexplainable' connection aborts later.
8730 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8732 *) Major EVP API cipher revision.
8733 Add hooks for extra EVP features. This allows various cipher
8734 parameters to be set in the EVP interface. Support added for variable
8735 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8736 setting of RC2 and RC5 parameters.
8738 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8741 Remove lots of duplicated code from the EVP library. For example *every*
8742 cipher init() function handles the 'iv' in the same way according to the
8743 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8744 for CFB and OFB modes they zero ctx->num.
8746 New functionality allows removal of S/MIME code RC2 hack.
8748 Most of the routines have the same form and so can be declared in terms
8751 By shifting this to the top level EVP_CipherInit() it can be removed from
8752 all individual ciphers. If the cipher wants to handle IVs or keys
8753 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8756 Change lots of functions like EVP_EncryptUpdate() to now return a
8757 value: although software versions of the algorithms cannot fail
8758 any installed hardware versions can.
8761 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8762 this option is set, tolerate broken clients that send the negotiated
8763 protocol version number instead of the requested protocol version
8767 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8768 i.e. non-zero for export ciphersuites, zero otherwise.
8769 Previous versions had this flag inverted, inconsistent with
8770 rsa_tmp_cb (..._TMP_RSA_CB).
8771 [Bodo Moeller; problem reported by Amit Chopra]
8773 *) Add missing DSA library text string. Work around for some IIS
8774 key files with invalid SEQUENCE encoding.
8777 *) Add a document (doc/standards.txt) that list all kinds of standards
8778 and so on that are implemented in OpenSSL.
8781 *) Enhance c_rehash script. Old version would mishandle certificates
8782 with the same subject name hash and wouldn't handle CRLs at all.
8783 Added -fingerprint option to crl utility, to support new c_rehash
8787 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8790 *) Fix for SSL server purpose checking. Server checking was
8791 rejecting certificates which had extended key usage present
8792 but no ssl client purpose.
8793 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8795 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8796 is a little unclear about how a blank password is handled.
8797 Since the password in encoded as a BMPString with terminating
8798 double NULL a zero length password would end up as just the
8799 double NULL. However no password at all is different and is
8800 handled differently in the PKCS#12 key generation code. NS
8801 treats a blank password as zero length. MSIE treats it as no
8802 password on export: but it will try both on import. We now do
8803 the same: PKCS12_parse() tries zero length and no password if
8804 the password is set to "" or NULL (NULL is now a valid password:
8805 it wasn't before) as does the pkcs12 application.
8808 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8809 perror when PEM_read_bio_X509_REQ fails, the error message must
8810 be obtained from the error queue.
8813 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8814 it in ERR_remove_state if appropriate, and change ERR_get_state
8815 accordingly to avoid race conditions (this is necessary because
8816 thread_hash is no longer constant once set).
8819 *) Bugfix for linux-elf makefile.one.
8822 *) RSA_get_default_method() will now cause a default
8823 RSA_METHOD to be chosen if one doesn't exist already.
8824 Previously this was only set during a call to RSA_new()
8825 or RSA_new_method(NULL) meaning it was possible for
8826 RSA_get_default_method() to return NULL.
8829 *) Added native name translation to the existing DSO code
8830 that will convert (if the flag to do so is set) filenames
8831 that are sufficiently small and have no path information
8832 into a canonical native form. Eg. "blah" converted to
8833 "libblah.so" or "blah.dll" etc.
8836 *) New function ERR_error_string_n(e, buf, len) which is like
8837 ERR_error_string(e, buf), but writes at most 'len' bytes
8838 including the 0 terminator. For ERR_error_string_n, 'buf'
8840 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8842 *) CONF library reworked to become more general. A new CONF
8843 configuration file reader "class" is implemented as well as a
8844 new functions (NCONF_*, for "New CONF") to handle it. The now
8845 old CONF_* functions are still there, but are reimplemented to
8846 work in terms of the new functions. Also, a set of functions
8847 to handle the internal storage of the configuration data is
8848 provided to make it easier to write new configuration file
8849 reader "classes" (I can definitely see something reading a
8850 configuration file in XML format, for example), called _CONF_*,
8851 or "the configuration storage API"...
8853 The new configuration file reading functions are:
8855 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8856 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8858 NCONF_default, NCONF_WIN32
8860 NCONF_dump_fp, NCONF_dump_bio
8862 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8863 NCONF_new creates a new CONF object. This works in the same way
8864 as other interfaces in OpenSSL, like the BIO interface.
8865 NCONF_dump_* dump the internal storage of the configuration file,
8866 which is useful for debugging. All other functions take the same
8867 arguments as the old CONF_* functions wth the exception of the
8868 first that must be a `CONF *' instead of a `LHASH *'.
8870 To make it easer to use the new classes with the old CONF_* functions,
8871 the function CONF_set_default_method is provided.
8874 *) Add '-tls1' option to 'openssl ciphers', which was already
8875 mentioned in the documentation but had not been implemented.
8876 (This option is not yet really useful because even the additional
8877 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8880 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8881 OpenSSL-based applications) load shared libraries and bind to
8882 them in a portable way.
8883 [Geoff Thorpe, with contributions from Richard Levitte]
8885 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8887 *) Make sure _lrotl and _lrotr are only used with MSVC.
8889 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8890 (the default implementation of RAND_status).
8892 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8893 to '-clrext' (= clear extensions), as intended and documented.
8894 [Bodo Moeller; inconsistency pointed out by Michael Attili
8897 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8898 was larger than the MD block size.
8899 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8901 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8902 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8903 using the passed key: if the passed key was a private key the result
8904 of X509_print(), for example, would be to print out all the private key
8908 *) des_quad_cksum() byte order bug fix.
8909 [Ulf Möller, using the problem description in krb4-0.9.7, where
8910 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8912 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8914 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8916 *) For easily testing in shell scripts whether some command
8917 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8918 returns with exit code 0 iff no command of the given name is available.
8919 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8920 the output goes to stdout and nothing is printed to stderr.
8921 Additional arguments are always ignored.
8923 Since for each cipher there is a command of the same name,
8924 the 'no-cipher' compilation switches can be tested this way.
8926 ('openssl no-XXX' is not able to detect pseudo-commands such
8927 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8930 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8933 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8934 is set; it will be thrown away anyway because each handshake creates
8936 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8937 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8938 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8939 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8942 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8943 'Q' and 'R' lose their special meanings (quit/renegotiate).
8944 This is part of what -quiet does; unlike -quiet, -ign_eof
8945 does not suppress any output.
8948 *) Add compatibility options to the purpose and trust code. The
8949 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8950 accepts a certificate or CA, this was the previous behaviour,
8951 with all the associated security issues.
8953 X509_TRUST_COMPAT is the old trust behaviour: only and
8954 automatically trust self signed roots in certificate store. A
8955 new trust setting X509_TRUST_DEFAULT is used to specify that
8956 a purpose has no associated trust setting and it should instead
8957 use the value in the default purpose.
8960 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8961 and fix a memory leak.
8964 *) In util/mkerr.pl (which implements 'make errors'), preserve
8965 reason strings from the previous version of the .c file, as
8966 the default to have only downcase letters (and digits) in
8967 automatically generated reasons codes is not always appropriate.
8970 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8971 using strerror. Previously, ERR_reason_error_string() returned
8972 library names as reason strings for SYSerr; but SYSerr is a special
8973 case where small numbers are errno values, not library numbers.
8976 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8977 converts DSA parameters into DH parameters. (When creating parameters,
8978 DSA_generate_parameters is used.)
8981 *) Include 'length' (recommended exponent length) in C code generated
8982 by 'openssl dhparam -C'.
8985 *) The second argument to set_label in perlasm was already being used
8986 so couldn't be used as a "file scope" flag. Moved to third argument
8990 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8991 instead of RAND_bytes for encryption IVs and salts.
8994 *) Include RAND_status() into RAND_METHOD instead of implementing
8995 it only for md_rand.c Otherwise replacing the PRNG by calling
8996 RAND_set_rand_method would be impossible.
8999 *) Don't let DSA_generate_key() enter an infinite loop if the random
9000 number generation fails.
9003 *) New 'rand' application for creating pseudo-random output.
9006 *) Added configuration support for Linux/IA64
9007 [Rolf Haberrecker <rolf@suse.de>]
9009 *) Assembler module support for Mingw32.
9012 *) Shared library support for HPUX (in shlib/).
9013 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9015 *) Shared library support for Solaris gcc.
9016 [Lutz Behnke <behnke@trustcenter.de>]
9018 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9020 *) PKCS7_encrypt() was adding text MIME headers twice because they
9021 were added manually and by SMIME_crlf_copy().
9024 *) In bntest.c don't call BN_rand with zero bits argument.
9025 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9027 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9028 case was implemented. This caused BN_div_recp() to fail occasionally.
9031 *) Add an optional second argument to the set_label() in the perl
9032 assembly language builder. If this argument exists and is set
9033 to 1 it signals that the assembler should use a symbol whose
9034 scope is the entire file, not just the current function. This
9035 is needed with MASM which uses the format label:: for this scope.
9036 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9038 *) Change the ASN1 types so they are typedefs by default. Before
9039 almost all types were #define'd to ASN1_STRING which was causing
9040 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9044 *) Change names of new functions to the new get1/get0 naming
9045 convention: After 'get1', the caller owns a reference count
9046 and has to call ..._free; 'get0' returns a pointer to some
9047 data structure without incrementing reference counters.
9048 (Some of the existing 'get' functions increment a reference
9049 counter, some don't.)
9050 Similarly, 'set1' and 'add1' functions increase reference
9051 counters or duplicate objects.
9054 *) Allow for the possibility of temp RSA key generation failure:
9055 the code used to assume it always worked and crashed on failure.
9058 *) Fix potential buffer overrun problem in BIO_printf().
9059 [Ulf Möller, using public domain code by Patrick Powell; problem
9060 pointed out by David Sacerdote <das33@cornell.edu>]
9062 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9063 RAND_egd() and RAND_status(). In the command line application,
9064 the EGD socket can be specified like a seed file using RANDFILE
9068 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9069 Some CAs (e.g. Verisign) distribute certificates in this form.
9072 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9073 list to exclude them. This means that no special compilation option
9074 is needed to use anonymous DH: it just needs to be included in the
9078 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9079 EVP_MD_type. The old functionality is available in a new macro called
9080 EVP_MD_md(). Change code that uses it and update docs.
9083 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9084 where the 'void *' argument is replaced by a function pointer argument.
9085 Previously 'void *' was abused to point to functions, which works on
9086 many platforms, but is not correct. As these functions are usually
9087 called by macros defined in OpenSSL header files, most source code
9088 should work without changes.
9091 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9092 sections with information on -D... compiler switches used for
9093 compiling the library so that applications can see them. To enable
9094 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9095 must be defined. E.g.,
9096 #define OPENSSL_ALGORITHM_DEFINES
9097 #include <openssl/opensslconf.h>
9098 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9099 [Richard Levitte, Ulf and Bodo Möller]
9101 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9105 *) Change the 'other' type in certificate aux info to a STACK_OF
9106 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9107 the required ASN1 format: arbitrary types determined by an OID.
9110 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9111 argument to 'req'. This is not because the function is newer or
9112 better than others it just uses the work 'NEW' in the certificate
9113 request header lines. Some software needs this.
9116 *) Reorganise password command line arguments: now passwords can be
9117 obtained from various sources. Delete the PEM_cb function and make
9118 it the default behaviour: i.e. if the callback is NULL and the
9119 usrdata argument is not NULL interpret it as a null terminated pass
9120 phrase. If usrdata and the callback are NULL then the pass phrase
9121 is prompted for as usual.
9124 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9125 the support is automatically enabled. The resulting binaries will
9126 autodetect the card and use it if present.
9127 [Ben Laurie and Compaq Inc.]
9129 *) Work around for Netscape hang bug. This sends certificate request
9130 and server done in one record. Since this is perfectly legal in the
9131 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9132 the bugs/SSLv3 entry for more info.
9135 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9138 *) Add -rand argument to smime and pkcs12 applications and read/write
9142 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9145 *) Add command line password options to the remaining applications.
9148 *) Bug fix for BN_div_recp() for numerators with an even number of
9152 *) More tests in bntest.c, and changed test_bn output.
9155 *) ./config recognizes MacOS X now.
9158 *) Bug fix for BN_div() when the first words of num and divsor are
9159 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9162 *) Add support for various broken PKCS#8 formats, and command line
9163 options to produce them.
9166 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9167 get temporary BIGNUMs from a BN_CTX.
9170 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9174 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9175 include a #define from the old name to the new. The original intent
9176 was that statically linked binaries could for example just call
9177 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9178 link with digests. This never worked because SSLeay_add_all_digests()
9179 and SSLeay_add_all_ciphers() were in the same source file so calling
9180 one would link with the other. They are now in separate source files.
9183 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9186 *) Use a less unusual form of the Miller-Rabin primality test (it used
9187 a binary algorithm for exponentiation integrated into the Miller-Rabin
9188 loop, our standard modexp algorithms are faster).
9191 *) Support for the EBCDIC character set completed.
9192 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9194 *) Source code cleanups: use const where appropriate, eliminate casts,
9195 use void * instead of char * in lhash.
9198 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9199 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9200 this the server could overwrite ephemeral keys that the client
9204 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9205 using 50 iterations of the Rabin-Miller test.
9207 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9208 iterations of the Rabin-Miller test as required by the appendix
9209 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9210 As BN_is_prime_fasttest includes trial division, DSA parameter
9211 generation becomes much faster.
9213 This implies a change for the callback functions in DSA_is_prime
9214 and DSA_generate_parameters: The callback function is called once
9215 for each positive witness in the Rabin-Miller test, not just
9216 occasionally in the inner loop; and the parameters to the
9217 callback function now provide an iteration count for the outer
9218 loop rather than for the current invocation of the inner loop.
9219 DSA_generate_parameters additionally can call the callback
9220 function with an 'iteration count' of -1, meaning that a
9221 candidate has passed the trial division test (when q is generated
9222 from an application-provided seed, trial division is skipped).
9225 *) New function BN_is_prime_fasttest that optionally does trial
9226 division before starting the Rabin-Miller test and has
9227 an additional BN_CTX * argument (whereas BN_is_prime always
9228 has to allocate at least one BN_CTX).
9229 'callback(1, -1, cb_arg)' is called when a number has passed the
9230 trial division stage.
9233 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9237 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9240 *) New function BN_pseudo_rand().
9243 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9244 bignum version of BN_from_montgomery() with the working code from
9245 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9249 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9250 made it impossible to use the same SSL_SESSION data structure in
9251 SSL2 clients in multiple threads.
9254 *) The return value of RAND_load_file() no longer counts bytes obtained
9255 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9256 to seed the PRNG (previously an explicit byte count was required).
9257 [Ulf Möller, Bodo Möller]
9259 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9260 used (char *) instead of (void *) and had casts all over the place.
9263 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9266 *) Retain source code compatibility for BN_prime_checks macro:
9267 BN_is_prime(..., BN_prime_checks, ...) now uses
9268 BN_prime_checks_for_size to determine the appropriate number of
9269 Rabin-Miller iterations.
9272 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9273 DH_CHECK_P_NOT_SAFE_PRIME.
9274 (Check if this is true? OpenPGP calls them "strong".)
9277 *) Merge the functionality of "dh" and "gendh" programs into a new program
9278 "dhparam". The old programs are retained for now but will handle DH keys
9279 (instead of parameters) in future.
9282 *) Make the ciphers, s_server and s_client programs check the return values
9283 when a new cipher list is set.
9286 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9287 ciphers. Before when the 56bit ciphers were enabled the sorting was
9290 The syntax for the cipher sorting has been extended to support sorting by
9291 cipher-strength (using the strength_bits hard coded in the tables).
9292 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9294 Fix a bug in the cipher-command parser: when supplying a cipher command
9295 string with an "undefined" symbol (neither command nor alphanumeric
9296 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9297 an error is flagged.
9299 Due to the strength-sorting extension, the code of the
9300 ssl_create_cipher_list() function was completely rearranged. I hope that
9301 the readability was also increased :-)
9302 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9304 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9305 for the first serial number and places 2 in the serial number file. This
9306 avoids problems when the root CA is created with serial number zero and
9307 the first user certificate has the same issuer name and serial number
9311 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9312 the new code. Add documentation for this stuff.
9315 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9316 X509_*() to X509at_*() on the grounds that they don't handle X509
9317 structures and behave in an analogous way to the X509v3 functions:
9318 they shouldn't be called directly but wrapper functions should be used
9321 So we also now have some wrapper functions that call the X509at functions
9322 when passed certificate requests. (TO DO: similar things can be done with
9323 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9324 things. Some of these need some d2i or i2d and print functionality
9325 because they handle more complex structures.)
9328 *) Add missing #ifndefs that caused missing symbols when building libssl
9329 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9330 NO_RSA in ssl/s2*.c.
9331 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9333 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9334 has a return value which indicates the quality of the random data
9335 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9336 error queue. New function RAND_pseudo_bytes() generates output that is
9337 guaranteed to be unique but not unpredictable. RAND_add is like
9338 RAND_seed, but takes an extra argument for an entropy estimate
9339 (RAND_seed always assumes full entropy).
9342 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9343 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9344 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9345 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9346 false-positive rate of at most 2^-80 for random input.
9349 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9352 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9353 in the 0.9.5 release), this returns the chain
9354 from an X509_CTX structure with a dup of the stack and all
9355 the X509 reference counts upped: so the stack will exist
9356 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9359 Also make SSL_SESSION_print() print out the verify return
9363 *) Add manpage for the pkcs12 command. Also change the default
9364 behaviour so MAC iteration counts are used unless the new
9365 -nomaciter option is used. This improves file security and
9366 only older versions of MSIE (4.0 for example) need it.
9369 *) Honor the no-xxx Configure options when creating .DEF files.
9372 *) Add PKCS#10 attributes to field table: challengePassword,
9373 unstructuredName and unstructuredAddress. These are taken from
9374 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9375 international characters are used.
9377 More changes to X509_ATTRIBUTE code: allow the setting of types
9378 based on strings. Remove the 'loc' parameter when adding
9379 attributes because these will be a SET OF encoding which is sorted
9383 *) Initial changes to the 'req' utility to allow request generation
9384 automation. This will allow an application to just generate a template
9385 file containing all the field values and have req construct the
9388 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9389 used all over the place including certificate requests and PKCS#7
9390 structures. They are currently handled manually where necessary with
9391 some primitive wrappers for PKCS#7. The new functions behave in a
9392 manner analogous to the X509 extension functions: they allow
9393 attributes to be looked up by NID and added.
9395 Later something similar to the X509V3 code would be desirable to
9396 automatically handle the encoding, decoding and printing of the
9397 more complex types. The string types like challengePassword can
9398 be handled by the string table functions.
9400 Also modified the multi byte string table handling. Now there is
9401 a 'global mask' which masks out certain types. The table itself
9402 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9403 is useful when for example there is only one permissible type
9404 (as in countryName) and using the mask might result in no valid
9408 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9409 SSL_get_peer_finished to allow applications to obtain the latest
9410 Finished messages sent to the peer or expected from the peer,
9411 respectively. (SSL_get_peer_finished is usually the Finished message
9412 actually received from the peer, otherwise the protocol will be aborted.)
9414 As the Finished message are message digests of the complete handshake
9415 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9416 be used for external authentication procedures when the authentication
9417 provided by SSL/TLS is not desired or is not enough.
9420 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9421 the host supports BWX extension and if Compaq C is present on the
9422 $PATH. Just exploiting of the BWX extension results in 20-30%
9423 performance kick for some algorithms, e.g. DES and RC4 to mention
9424 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9428 *) Add support for MS "fast SGC". This is arguably a violation of the
9429 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9430 weak crypto and after checking the certificate is SGC a second one
9431 with strong crypto. MS SGC stops the first handshake after receiving
9432 the server certificate message and sends a second client hello. Since
9433 a server will typically do all the time consuming operations before
9434 expecting any further messages from the client (server key exchange
9435 is the most expensive) there is little difference between the two.
9437 To get OpenSSL to support MS SGC we have to permit a second client
9438 hello message after we have sent server done. In addition we have to
9439 reset the MAC if we do get this second client hello.
9442 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9443 if a DER encoded private key is RSA or DSA traditional format. Changed
9444 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9445 format DER encoded private key. Newer code should use PKCS#8 format which
9446 has the key type encoded in the ASN1 structure. Added DER private key
9447 support to pkcs8 application.
9450 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9451 ciphersuites has been selected (as required by the SSL 3/TLS 1
9452 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9453 is set, we interpret this as a request to violate the specification
9454 (the worst that can happen is a handshake failure, and 'correct'
9455 behaviour would result in a handshake failure anyway).
9458 *) In SSL_CTX_add_session, take into account that there might be multiple
9459 SSL_SESSION structures with the same session ID (e.g. when two threads
9460 concurrently obtain them from an external cache).
9461 The internal cache can handle only one SSL_SESSION with a given ID,
9462 so if there's a conflict, we now throw out the old one to achieve
9466 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9467 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9468 some routines that use cipher OIDs: some ciphers do not have OIDs
9469 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9473 *) Simplify the trust setting structure and code. Now we just have
9474 two sequences of OIDs for trusted and rejected settings. These will
9475 typically have values the same as the extended key usage extension
9476 and any application specific purposes.
9478 The trust checking code now has a default behaviour: it will just
9479 check for an object with the same NID as the passed id. Functions can
9480 be provided to override either the default behaviour or the behaviour
9481 for a given id. SSL client, server and email already have functions
9482 in place for compatibility: they check the NID and also return "trusted"
9483 if the certificate is self signed.
9486 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9487 traditional format into an EVP_PKEY structure.
9490 *) Add a password callback function PEM_cb() which either prompts for
9491 a password if usr_data is NULL or otherwise assumes it is a null
9492 terminated password. Allow passwords to be passed on command line
9493 environment or config files in a few more utilities.
9496 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9497 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9498 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9499 Update documentation.
9502 *) Support for ASN1 "NULL" type. This could be handled before by using
9503 ASN1_TYPE but there wasn't any function that would try to read a NULL
9504 and produce an error if it couldn't. For compatibility we also have
9505 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9506 don't allocate anything because they don't need to.
9509 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9511 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9513 *) Rebuild of the memory allocation routines used by OpenSSL code and
9514 possibly others as well. The purpose is to make an interface that
9515 provide hooks so anyone can build a separate set of allocation and
9516 deallocation routines to be used by OpenSSL, for example memory
9517 pool implementations, or something else, which was previously hard
9518 since Malloc(), Realloc() and Free() were defined as macros having
9519 the values malloc, realloc and free, respectively (except for Win32
9520 compilations). The same is provided for memory debugging code.
9521 OpenSSL already comes with functionality to find memory leaks, but
9522 this gives people a chance to debug other memory problems.
9524 With these changes, a new set of functions and macros have appeared:
9526 CRYPTO_set_mem_debug_functions() [F]
9527 CRYPTO_get_mem_debug_functions() [F]
9528 CRYPTO_dbg_set_options() [F]
9529 CRYPTO_dbg_get_options() [F]
9530 CRYPTO_malloc_debug_init() [M]
9532 The memory debug functions are NULL by default, unless the library
9533 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9534 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9535 gives the standard debugging functions that come with OpenSSL) or
9536 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9537 provided by the library user) must be used. When the standard
9538 debugging functions are used, CRYPTO_dbg_set_options can be used to
9539 request additional information:
9540 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9541 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9543 Also, things like CRYPTO_set_mem_functions will always give the
9544 expected result (the new set of functions is used for allocation
9545 and deallocation) at all times, regardless of platform and compiler
9548 To finish it up, some functions that were never use in any other
9549 way than through macros have a new API and new semantic:
9552 CRYPTO_dbg_realloc()
9555 All macros of value have retained their old syntax.
9556 [Richard Levitte and Bodo Moeller]
9558 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9559 ordering of SMIMECapabilities wasn't in "strength order" and there
9560 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9564 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9565 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9566 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9568 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9569 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9570 functionality to handle multipart/signed properly) and a utility
9571 called 'smime' to call all this stuff. This is based on code I
9572 originally wrote for Celo who have kindly allowed it to be
9573 included in OpenSSL.
9576 *) Add variants des_set_key_checked and des_set_key_unchecked of
9577 des_set_key (aka des_key_sched). Global variable des_check_key
9578 decides which of these is called by des_set_key; this way
9579 des_check_key behaves as it always did, but applications and
9580 the library itself, which was buggy for des_check_key == 1,
9581 have a cleaner way to pick the version they need.
9584 *) New function PKCS12_newpass() which changes the password of a
9588 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9589 dynamic mix. In both cases the ids can be used as an index into the
9590 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9591 functions so they accept a list of the field values and the
9592 application doesn't need to directly manipulate the X509_TRUST
9596 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9600 *) Modify the way the V3 extension code looks up extensions. This now
9601 works in a similar way to the object code: we have some "standard"
9602 extensions in a static table which is searched with OBJ_bsearch()
9603 and the application can add dynamic ones if needed. The file
9604 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9605 updated whenever a new extension is added to the core code and kept
9606 in ext_nid order. There is a simple program 'tabtest.c' which checks
9607 this. New extensions are not added too often so this file can readily
9608 be maintained manually.
9610 There are two big advantages in doing things this way. The extensions
9611 can be looked up immediately and no longer need to be "added" using
9612 X509V3_add_standard_extensions(): this function now does nothing.
9613 [Side note: I get *lots* of email saying the extension code doesn't
9614 work because people forget to call this function]
9615 Also no dynamic allocation is done unless new extensions are added:
9616 so if we don't add custom extensions there is no need to call
9617 X509V3_EXT_cleanup().
9620 *) Modify enc utility's salting as follows: make salting the default. Add a
9621 magic header, so unsalted files fail gracefully instead of just decrypting
9622 to garbage. This is because not salting is a big security hole, so people
9623 should be discouraged from doing it.
9626 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9627 digest to be passed on the command line but it only used this
9628 parameter when signing a certificate. Modified so all relevant
9629 operations are affected by the digest parameter including the
9630 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9631 DSA key was used because it didn't fix the digest.
9634 *) Initial certificate chain verify code. Currently tests the untrusted
9635 certificates for consistency with the verify purpose (which is set
9636 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9638 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9639 this is because it will reject chains with invalid extensions whereas
9640 every previous version of OpenSSL and SSLeay made no checks at all.
9642 Trust code: checks the root CA for the relevant trust settings. Trust
9643 settings have an initial value consistent with the verify purpose: e.g.
9644 if the verify purpose is for SSL client use it expects the CA to be
9645 trusted for SSL client use. However the default value can be changed to
9646 permit custom trust settings: one example of this would be to only trust
9647 certificates from a specific "secure" set of CAs.
9649 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9650 which should be used for version portability: especially since the
9651 verify structure is likely to change more often now.
9653 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9654 to set them. If not set then assume SSL clients will verify SSL servers
9657 Two new options to the verify program: -untrusted allows a set of
9658 untrusted certificates to be passed in and -purpose which sets the
9659 intended purpose of the certificate. If a purpose is set then the
9660 new chain verify code is used to check extension consistency.
9663 *) Support for the authority information access extension.
9666 *) Modify RSA and DSA PEM read routines to transparently handle
9667 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9668 public keys in a format compatible with certificate
9669 SubjectPublicKeyInfo structures. Unfortunately there were already
9670 functions called *_PublicKey_* which used various odd formats so
9671 these are retained for compatibility: however the DSA variants were
9672 never in a public release so they have been deleted. Changed dsa/rsa
9673 utilities to handle the new format: note no releases ever handled public
9674 keys so we should be OK.
9676 The primary motivation for this change is to avoid the same fiasco
9677 that dogs private keys: there are several incompatible private key
9678 formats some of which are standard and some OpenSSL specific and
9679 require various evil hacks to allow partial transparent handling and
9680 even then it doesn't work with DER formats. Given the option anything
9681 other than PKCS#8 should be dumped: but the other formats have to
9682 stay in the name of compatibility.
9684 With public keys and the benefit of hindsight one standard format
9685 is used which works with EVP_PKEY, RSA or DSA structures: though
9686 it clearly returns an error if you try to read the wrong kind of key.
9688 Added a -pubkey option to the 'x509' utility to output the public key.
9689 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9690 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9691 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9692 that do the same as the EVP_PKEY_assign_*() except they up the
9693 reference count of the added key (they don't "swallow" the
9697 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9698 CRLs would fail if the file contained no certificates or no CRLs:
9699 added a new function to read in both types and return the number
9700 read: this means that if none are read it will be an error. The
9701 DER versions of the certificate and CRL reader would always fail
9702 because it isn't possible to mix certificates and CRLs in DER format
9703 without choking one or the other routine. Changed this to just read
9704 a certificate: this is the best we can do. Also modified the code
9705 in apps/verify.c to take notice of return codes: it was previously
9706 attempting to read in certificates from NULL pointers and ignoring
9707 any errors: this is one reason why the cert and CRL reader seemed
9708 to work. It doesn't check return codes from the default certificate
9709 routines: these may well fail if the certificates aren't installed.
9712 *) Code to support otherName option in GeneralName.
9715 *) First update to verify code. Change the verify utility
9716 so it warns if it is passed a self signed certificate:
9717 for consistency with the normal behaviour. X509_verify
9718 has been modified to it will now verify a self signed
9719 certificate if *exactly* the same certificate appears
9720 in the store: it was previously impossible to trust a
9721 single self signed certificate. This means that:
9722 openssl verify ss.pem
9723 now gives a warning about a self signed certificate but
9724 openssl verify -CAfile ss.pem ss.pem
9728 *) For servers, store verify_result in SSL_SESSION data structure
9729 (and add it to external session representation).
9730 This is needed when client certificate verifications fails,
9731 but an application-provided verification callback (set by
9732 SSL_CTX_set_cert_verify_callback) allows accepting the session
9733 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9734 but returns 1): When the session is reused, we have to set
9735 ssl->verify_result to the appropriate error code to avoid
9737 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9739 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9740 case in PKCS7_dataInit() where the signed PKCS7 structure
9741 didn't contain any existing data because it was being created.
9742 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9744 *) Add a salt to the key derivation routines in enc.c. This
9745 forms the first 8 bytes of the encrypted file. Also add a
9746 -S option to allow a salt to be input on the command line.
9749 *) New function X509_cmp(). Oddly enough there wasn't a function
9750 to compare two certificates. We do this by working out the SHA1
9751 hash and comparing that. X509_cmp() will be needed by the trust
9755 *) SSL_get1_session() is like SSL_get_session(), but increments
9756 the reference count in the SSL_SESSION returned.
9757 [Geoff Thorpe <geoff@eu.c2.net>]
9759 *) Fix for 'req': it was adding a null to request attributes.
9760 Also change the X509_LOOKUP and X509_INFO code to handle
9761 certificate auxiliary information.
9764 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9768 *) Add the possibility to add extra information to the memory leak
9769 detecting output, to form tracebacks, showing from where each
9770 allocation was originated: CRYPTO_push_info("constant string") adds
9771 the string plus current file name and line number to a per-thread
9772 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9773 is like calling CYRPTO_pop_info() until the stack is empty.
9774 Also updated memory leak detection code to be multi-thread-safe.
9777 *) Add options -text and -noout to pkcs7 utility and delete the
9778 encryption options which never did anything. Update docs.
9781 *) Add options to some of the utilities to allow the pass phrase
9782 to be included on either the command line (not recommended on
9783 OSes like Unix) or read from the environment. Update the
9784 manpages and fix a few bugs.
9787 *) Add a few manpages for some of the openssl commands.
9790 *) Fix the -revoke option in ca. It was freeing up memory twice,
9791 leaking and not finding already revoked certificates.
9794 *) Extensive changes to support certificate auxiliary information.
9795 This involves the use of X509_CERT_AUX structure and X509_AUX
9796 functions. An X509_AUX function such as PEM_read_X509_AUX()
9797 can still read in a certificate file in the usual way but it
9798 will also read in any additional "auxiliary information". By
9799 doing things this way a fair degree of compatibility can be
9800 retained: existing certificates can have this information added
9801 using the new 'x509' options.
9803 Current auxiliary information includes an "alias" and some trust
9804 settings. The trust settings will ultimately be used in enhanced
9805 certificate chain verification routines: currently a certificate
9806 can only be trusted if it is self signed and then it is trusted
9810 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9811 The problem was that one of the replacement routines had not been working
9812 since SSLeay releases. For now the offending routine has been replaced
9813 with non-optimised assembler. Even so, this now gives around 95%
9814 performance improvement for 1024 bit RSA signs.
9817 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9818 handling. Most clients have the effective key size in bits equal to
9819 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9820 A few however don't do this and instead use the size of the decrypted key
9821 to determine the RC2 key length and the AlgorithmIdentifier to determine
9822 the effective key length. In this case the effective key length can still
9823 be 40 bits but the key length can be 168 bits for example. This is fixed
9824 by manually forcing an RC2 key into the EVP_PKEY structure because the
9825 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9826 the key length and effective key length are equal.
9829 *) Add a bunch of functions that should simplify the creation of
9830 X509_NAME structures. Now you should be able to do:
9831 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9832 and have it automatically work out the correct field type and fill in
9833 the structures. The more adventurous can try:
9834 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9835 and it will (hopefully) work out the correct multibyte encoding.
9838 *) Change the 'req' utility to use the new field handling and multibyte
9839 copy routines. Before the DN field creation was handled in an ad hoc
9840 way in req, ca, and x509 which was rather broken and didn't support
9841 BMPStrings or UTF8Strings. Since some software doesn't implement
9842 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9843 using the dirstring_type option. See the new comment in the default
9844 openssl.cnf for more info.
9847 *) Make crypto/rand/md_rand.c more robust:
9848 - Assure unique random numbers after fork().
9849 - Make sure that concurrent threads access the global counter and
9850 md serializably so that we never lose entropy in them
9851 or use exactly the same state in multiple threads.
9852 Access to the large state is not always serializable because
9853 the additional locking could be a performance killer, and
9854 md should be large enough anyway.
9857 *) New file apps/app_rand.c with commonly needed functionality
9858 for handling the random seed file.
9860 Use the random seed file in some applications that previously did not:
9862 dsaparam -genkey (which also ignored its '-rand' option),
9865 x509 (when signing).
9866 Except on systems with /dev/urandom, it is crucial to have a random
9867 seed file at least for key creation, DSA signing, and for DH exchanges;
9868 for RSA signatures we could do without one.
9870 gendh and gendsa (unlike genrsa) used to read only the first byte
9871 of each file listed in the '-rand' option. The function as previously
9872 found in genrsa is now in app_rand.c and is used by all programs
9873 that support '-rand'.
9876 *) In RAND_write_file, use mode 0600 for creating files;
9877 don't just chmod when it may be too late.
9880 *) Report an error from X509_STORE_load_locations
9881 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9884 *) New function ASN1_mbstring_copy() this copies a string in either
9885 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9886 into an ASN1_STRING type. A mask of permissible types is passed
9887 and it chooses the "minimal" type to use or an error if not type
9891 *) Add function equivalents to the various macros in asn1.h. The old
9892 macros are retained with an M_ prefix. Code inside the library can
9893 use the M_ macros. External code (including the openssl utility)
9894 should *NOT* in order to be "shared library friendly".
9897 *) Add various functions that can check a certificate's extensions
9898 to see if it usable for various purposes such as SSL client,
9899 server or S/MIME and CAs of these types. This is currently
9900 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9901 verification. Also added a -purpose flag to x509 utility to
9902 print out all the purposes.
9905 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9909 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9910 for, obtain and decode and extension and obtain its critical flag.
9911 This allows all the necessary extension code to be handled in a
9912 single function call.
9915 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9916 platforms. See crypto/rc4/rc4_enc.c for further details.
9919 *) New -noout option to asn1parse. This causes no output to be produced
9920 its main use is when combined with -strparse and -out to extract data
9921 from a file (which may not be in ASN.1 format).
9924 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9925 when producing the local key id.
9926 [Richard Levitte <levitte@stacken.kth.se>]
9928 *) New option -dhparam in s_server. This allows a DH parameter file to be
9929 stated explicitly. If it is not stated then it tries the first server
9930 certificate file. The previous behaviour hard coded the filename
9934 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9935 a public key to be input or output. For example:
9936 openssl rsa -in key.pem -pubout -out pubkey.pem
9937 Also added necessary DSA public key functions to handle this.
9940 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9941 in the message. This was handled by allowing
9942 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9943 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9945 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9946 to the end of the strings whereas this didn't. This would cause problems
9947 if strings read with d2i_ASN1_bytes() were later modified.
9948 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9950 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9951 data and it contains EOF it will end up returning an error. This is
9952 caused by input 46 bytes long. The cause is due to the way base64
9953 BIOs find the start of base64 encoded data. They do this by trying a
9954 trial decode on each line until they find one that works. When they
9955 do a flag is set and it starts again knowing it can pass all the
9956 data directly through the decoder. Unfortunately it doesn't reset
9957 the context it uses. This means that if EOF is reached an attempt
9958 is made to pass two EOFs through the context and this causes the
9959 resulting error. This can also cause other problems as well. As is
9960 usual with these problems it takes *ages* to find and the fix is
9961 trivial: move one line.
9962 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9964 *) Ugly workaround to get s_client and s_server working under Windows. The
9965 old code wouldn't work because it needed to select() on sockets and the
9966 tty (for keypresses and to see if data could be written). Win32 only
9967 supports select() on sockets so we select() with a 1s timeout on the
9968 sockets and then see if any characters are waiting to be read, if none
9969 are present then we retry, we also assume we can always write data to
9970 the tty. This isn't nice because the code then blocks until we've
9971 received a complete line of data and it is effectively polling the
9972 keyboard at 1s intervals: however it's quite a bit better than not
9973 working at all :-) A dedicated Windows application might handle this
9974 with an event loop for example.
9977 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9978 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9979 will be called when RSA_sign() and RSA_verify() are used. This is useful
9980 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9981 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9982 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9983 This necessitated the support of an extra signature type NID_md5_sha1
9984 for SSL signatures and modifications to the SSL library to use it instead
9985 of calling RSA_public_decrypt() and RSA_private_encrypt().
9988 *) Add new -verify -CAfile and -CApath options to the crl program, these
9989 will lookup a CRL issuers certificate and verify the signature in a
9990 similar way to the verify program. Tidy up the crl program so it
9991 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9992 less strict. It will now permit CRL extensions even if it is not
9993 a V2 CRL: this will allow it to tolerate some broken CRLs.
9996 *) Initialize all non-automatic variables each time one of the openssl
9997 sub-programs is started (this is necessary as they may be started
9998 multiple times from the "OpenSSL>" prompt).
9999 [Lennart Bang, Bodo Moeller]
10001 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10002 removing all other RSA functionality (this is what NO_RSA does). This
10003 is so (for example) those in the US can disable those operations covered
10004 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10008 *) Non-copying interface to BIO pairs.
10009 (still largely untested)
10012 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10013 ASCII string. This was handled independently in various places before.
10016 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10017 UTF8 strings a character at a time.
10020 *) Use client_version from client hello to select the protocol
10021 (s23_srvr.c) and for RSA client key exchange verification
10022 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10025 *) Add various utility functions to handle SPKACs, these were previously
10026 handled by poking round in the structure internals. Added new function
10027 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10028 print, verify and generate SPKACs. Based on an original idea from
10029 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10032 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10035 *) Allow the config file extension section to be overwritten on the
10036 command line. Based on an original idea from Massimiliano Pala
10037 <madwolf@comune.modena.it>. The new option is called -extensions
10038 and can be applied to ca, req and x509. Also -reqexts to override
10039 the request extensions in req and -crlexts to override the crl extensions
10043 *) Add new feature to the SPKAC handling in ca. Now you can include
10044 the same field multiple times by preceding it by "XXXX." for example:
10047 this is the same syntax as used in the req config file.
10050 *) Allow certificate extensions to be added to certificate requests. These
10051 are specified in a 'req_extensions' option of the req section of the
10052 config file. They can be printed out with the -text option to req but
10053 are otherwise ignored at present.
10056 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10057 data read consists of only the final block it would not decrypted because
10058 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10059 A misplaced 'break' also meant the decrypted final block might not be
10060 copied until the next read.
10063 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10064 a few extra parameters to the DH structure: these will be useful if
10065 for example we want the value of 'q' or implement X9.42 DH.
10068 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10069 provides hooks that allow the default DSA functions or functions on a
10070 "per key" basis to be replaced. This allows hardware acceleration and
10071 hardware key storage to be handled without major modification to the
10072 library. Also added low level modexp hooks and CRYPTO_EX structure and
10073 associated functions.
10076 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10077 as "read only": it can't be written to and the buffer it points to will
10078 not be freed. Reading from a read only BIO is much more efficient than
10079 a normal memory BIO. This was added because there are several times when
10080 an area of memory needs to be read from a BIO. The previous method was
10081 to create a memory BIO and write the data to it, this results in two
10082 copies of the data and an O(n^2) reading algorithm. There is a new
10083 function BIO_new_mem_buf() which creates a read only memory BIO from
10084 an area of memory. Also modified the PKCS#7 routines to use read only
10088 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10089 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10090 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10091 but a retry condition occurred while trying to read the rest.
10094 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10095 NID_pkcs7_encrypted by default: this was wrong since this should almost
10096 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10097 the encrypted data type: this is a more sensible place to put it and it
10098 allows the PKCS#12 code to be tidied up that duplicated this
10102 *) Changed obj_dat.pl script so it takes its input and output files on
10103 the command line. This should avoid shell escape redirection problems
10107 *) Initial support for certificate extension requests, these are included
10108 in things like Xenroll certificate requests. Included functions to allow
10109 extensions to be obtained and added.
10112 *) -crlf option to s_client and s_server for sending newlines as
10113 CRLF (as required by many protocols).
10116 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10118 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10119 [Ralf S. Engelschall]
10121 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10122 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10124 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10128 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10129 DH parameters/keys (q is lost during that conversion, but the resulting
10130 DH parameters contain its length).
10132 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10133 much faster than DH_generate_parameters (which creates parameters
10134 where p = 2*q + 1), and also the smaller q makes DH computations
10135 much more efficient (160-bit exponentiation instead of 1024-bit
10136 exponentiation); so this provides a convenient way to support DHE
10137 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10138 utter importance to use
10139 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10141 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10142 when such DH parameters are used, because otherwise small subgroup
10143 attacks may become possible!
10146 *) Avoid memory leak in i2d_DHparams.
10149 *) Allow the -k option to be used more than once in the enc program:
10150 this allows the same encrypted message to be read by multiple recipients.
10153 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10154 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10155 it will always use the numerical form of the OID, even if it has a short
10159 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10160 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10161 otherwise bn_mod_exp was called. In the case of hardware keys for example
10162 no private key components need be present and it might store extra data
10163 in the RSA structure, which cannot be accessed from bn_mod_exp.
10164 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10165 private key operations.
10168 *) Added support for SPARC Linux.
10171 *) pem_password_cb function type incompatibly changed from
10172 typedef int pem_password_cb(char *buf, int size, int rwflag);
10174 ....(char *buf, int size, int rwflag, void *userdata);
10175 so that applications can pass data to their callbacks:
10176 The PEM[_ASN1]_{read,write}... functions and macros now take an
10177 additional void * argument, which is just handed through whenever
10178 the password callback is called.
10179 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10181 New function SSL_CTX_set_default_passwd_cb_userdata.
10183 Compatibility note: As many C implementations push function arguments
10184 onto the stack in reverse order, the new library version is likely to
10185 interoperate with programs that have been compiled with the old
10186 pem_password_cb definition (PEM_whatever takes some data that
10187 happens to be on the stack as its last argument, and the callback
10188 just ignores this garbage); but there is no guarantee whatsoever that
10191 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10192 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10193 problems not only on Windows, but also on some Unix platforms.
10194 To avoid problematic command lines, these definitions are now in an
10195 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10196 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10199 *) MIPS III/IV assembler module is reimplemented.
10202 *) More DES library cleanups: remove references to srand/rand and
10203 delete an unused file.
10206 *) Add support for the the free Netwide assembler (NASM) under Win32,
10207 since not many people have MASM (ml) and it can be hard to obtain.
10208 This is currently experimental but it seems to work OK and pass all
10209 the tests. Check out INSTALL.W32 for info.
10212 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10213 without temporary keys kept an extra copy of the server key,
10214 and connections with temporary keys did not free everything in case
10218 *) New function RSA_check_key and new openssl rsa option -check
10219 for verifying the consistency of RSA keys.
10220 [Ulf Moeller, Bodo Moeller]
10222 *) Various changes to make Win32 compile work:
10223 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10224 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10225 comparison" warnings.
10226 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10229 *) Add a debugging option to PKCS#5 v2 key generation function: when
10230 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10231 derived keys are printed to stderr.
10234 *) Copy the flags in ASN1_STRING_dup().
10235 [Roman E. Pavlov <pre@mo.msk.ru>]
10237 *) The x509 application mishandled signing requests containing DSA
10238 keys when the signing key was also DSA and the parameters didn't match.
10240 It was supposed to omit the parameters when they matched the signing key:
10241 the verifying software was then supposed to automatically use the CA's
10242 parameters if they were absent from the end user certificate.
10244 Omitting parameters is no longer recommended. The test was also
10245 the wrong way round! This was probably due to unusual behaviour in
10246 EVP_cmp_parameters() which returns 1 if the parameters match.
10247 This meant that parameters were omitted when they *didn't* match and
10248 the certificate was useless. Certificates signed with 'ca' didn't have
10250 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10252 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10253 The interface is as follows:
10254 Applications can use
10255 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10256 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10257 "off" is now the default.
10258 The library internally uses
10259 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10260 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10261 to disable memory-checking temporarily.
10263 Some inconsistent states that previously were possible (and were
10264 even the default) are now avoided.
10266 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10267 with each memory chunk allocated; this is occasionally more helpful
10268 than just having a counter.
10270 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10272 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10276 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10277 which largely parallels "options", but is for changing API behaviour,
10278 whereas "options" are about protocol behaviour.
10279 Initial "mode" flags are:
10281 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10282 a single record has been written.
10283 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10284 retries use the same buffer location.
10285 (But all of the contents must be
10289 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10292 *) Fix problems with no-hmac etc.
10293 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10295 *) New functions RSA_get_default_method(), RSA_set_method() and
10296 RSA_get_method(). These allows replacement of RSA_METHODs without having
10297 to mess around with the internals of an RSA structure.
10300 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10301 Also really enable memory leak checks in openssl.c and in some
10303 [Chad C. Mulligan, Bodo Moeller]
10305 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10306 up the length of negative integers. This has now been simplified to just
10307 store the length when it is first determined and use it later, rather
10308 than trying to keep track of where data is copied and updating it to
10310 [Steve Henson, reported by Brien Wheeler
10311 <bwheeler@authentica-security.com>]
10313 *) Add a new function PKCS7_signatureVerify. This allows the verification
10314 of a PKCS#7 signature but with the signing certificate passed to the
10315 function itself. This contrasts with PKCS7_dataVerify which assumes the
10316 certificate is present in the PKCS#7 structure. This isn't always the
10317 case: certificates can be omitted from a PKCS#7 structure and be
10318 distributed by "out of band" means (such as a certificate database).
10321 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10322 function prototypes in pem.h, also change util/mkdef.pl to add the
10323 necessary function names.
10326 *) mk1mf.pl (used by Windows builds) did not properly read the
10327 options set by Configure in the top level Makefile, and Configure
10328 was not even able to write more than one option correctly.
10329 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10332 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10333 file to be loaded from a BIO or FILE pointer. The BIO version will
10334 for example allow memory BIOs to contain config info.
10337 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10338 Whoever hopes to achieve shared-library compatibility across versions
10339 must use this, not the compile-time macro.
10340 (Exercise 0.9.4: Which is the minimum library version required by
10342 Note: All this applies only to multi-threaded programs, others don't
10346 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10347 through a BIO pair triggered the default case, i.e.
10348 SSLerr(...,SSL_R_UNKNOWN_STATE).
10351 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10352 can use the SSL library even if none of the specific BIOs is
10356 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10357 for the encoded length.
10358 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10360 *) Add initial documentation of the X509V3 functions.
10363 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10364 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10365 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10366 secure PKCS#8 private key format with a high iteration count.
10369 *) Fix determination of Perl interpreter: A perl or perl5
10370 _directory_ in $PATH was also accepted as the interpreter.
10371 [Ralf S. Engelschall]
10373 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10374 wrong with it but it was very old and did things like calling
10375 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10376 unusual formatting.
10379 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10380 to use the new extension code.
10383 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10384 with macros. This should make it easier to change their form, add extra
10385 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10389 *) Add to configuration table a new entry that can specify an alternative
10390 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10391 according to Mark Crispin <MRC@Panda.COM>.
10395 *) DES CBC did not update the IV. Weird.
10398 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10399 Changing the behaviour of the former might break existing programs --
10400 where IV updating is needed, des_ncbc_encrypt can be used.
10403 *) When bntest is run from "make test" it drives bc to check its
10404 calculations, as well as internally checking them. If an internal check
10405 fails, it needs to cause bc to give a non-zero result or make test carries
10406 on without noticing the failure. Fixed.
10409 *) DES library cleanups.
10412 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10413 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10414 ciphers. NOTE: although the key derivation function has been verified
10415 against some published test vectors it has not been extensively tested
10416 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10420 *) Instead of "mkdir -p", which is not fully portable, use new
10421 Perl script "util/mkdir-p.pl".
10424 *) Rewrite the way password based encryption (PBE) is handled. It used to
10425 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10426 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10427 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10428 the 'parameter' field of the AlgorithmIdentifier is passed to the
10429 underlying key generation function so it must do its own ASN1 parsing.
10430 This has also changed the EVP_PBE_CipherInit() function which now has a
10431 'parameter' argument instead of literal salt and iteration count values
10432 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10435 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10436 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10437 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10438 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10439 value was just used as a "magic string" and not used directly its
10440 value doesn't matter.
10443 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10447 *) "linux-sparc64" configuration (ultrapenguin).
10448 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10449 "linux-sparc" configuration.
10450 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10452 *) config now generates no-xxx options for missing ciphers.
10455 *) Support the EBCDIC character set (work in progress).
10456 File ebcdic.c not yet included because it has a different license.
10457 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10459 *) Support BS2000/OSD-POSIX.
10460 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10462 *) Make callbacks for key generation use void * instead of char *.
10465 *) Make S/MIME samples compile (not yet tested).
10468 *) Additional typesafe stacks.
10471 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10475 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10477 *) New configuration variant "sco5-gcc".
10479 *) Updated some demos.
10480 [Sean O Riordain, Wade Scholine]
10482 *) Add missing BIO_free at exit of pkcs12 application.
10485 *) Fix memory leak in conf.c.
10488 *) Updates for Win32 to assembler version of MD5.
10491 *) Set #! path to perl in apps/der_chop to where we found it
10492 instead of using a fixed path.
10495 *) SHA library changes for irix64-mips4-cc.
10498 *) Improvements for VMS support.
10502 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10504 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10505 This also avoids the problems with SC4.2 and unpatched SC5.
10506 [Andy Polyakov <appro@fy.chalmers.se>]
10508 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10509 These are required because of the typesafe stack would otherwise break
10510 existing code. If old code used a structure member which used to be STACK
10511 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10512 sk_num or sk_value it would produce an error because the num, data members
10513 are not present in STACK_OF. Now it just produces a warning. sk_set
10514 replaces the old method of assigning a value to sk_value
10515 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10516 that does this will no longer work (and should use sk_set instead) but
10517 this could be regarded as a "questionable" behaviour anyway.
10520 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10521 correctly handle encrypted S/MIME data.
10524 *) Change type of various DES function arguments from des_cblock
10525 (which means, in function argument declarations, pointer to char)
10526 to des_cblock * (meaning pointer to array with 8 char elements),
10527 which allows the compiler to do more typechecking; it was like
10528 that back in SSLeay, but with lots of ugly casts.
10530 Introduce new type const_des_cblock.
10533 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10534 problems: find RecipientInfo structure that matches recipient certificate
10535 and initialise the ASN1 structures properly based on passed cipher.
10538 *) Belatedly make the BN tests actually check the results.
10541 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10542 to and from BNs: it was completely broken. New compilation option
10543 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10544 key elements as negative integers.
10547 *) Reorganize and speed up MD5.
10548 [Andy Polyakov <appro@fy.chalmers.se>]
10551 [Richard Levitte <richard@levitte.org>]
10553 *) New option -out to asn1parse to allow the parsed structure to be
10554 output to a file. This is most useful when combined with the -strparse
10555 option to examine the output of things like OCTET STRINGS.
10558 *) Make SSL library a little more fool-proof by not requiring any longer
10559 that SSL_set_{accept,connect}_state be called before
10560 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10561 in many applications because usually everything *appeared* to work as
10562 intended anyway -- now it really works as intended).
10565 *) Move openssl.cnf out of lib/.
10568 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10569 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10570 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10571 [Ralf S. Engelschall]
10573 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10574 handle PKCS#7 enveloped data properly.
10575 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10577 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10578 copying pointers. The cert_st handling is changed by this in
10579 various ways (and thus what used to be known as ctx->default_cert
10580 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10581 any longer when s->cert does not give us what we need).
10582 ssl_cert_instantiate becomes obsolete by this change.
10583 As soon as we've got the new code right (possibly it already is?),
10584 we have solved a couple of bugs of the earlier code where s->cert
10585 was used as if it could not have been shared with other SSL structures.
10587 Note that using the SSL API in certain dirty ways now will result
10588 in different behaviour than observed with earlier library versions:
10589 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10590 does not influence s as it used to.
10592 In order to clean up things more thoroughly, inside SSL_SESSION
10593 we don't use CERT any longer, but a new structure SESS_CERT
10594 that holds per-session data (if available); currently, this is
10595 the peer's certificate chain and, for clients, the server's certificate
10596 and temporary key. CERT holds only those values that can have
10597 meaningful defaults in an SSL_CTX.
10600 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10601 from the internal representation. Various PKCS#7 fixes: remove some
10602 evil casts and set the enc_dig_alg field properly based on the signing
10606 *) Allow PKCS#12 password to be set from the command line or the
10607 environment. Let 'ca' get its config file name from the environment
10608 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10612 *) Allow certificate policies extension to use an IA5STRING for the
10613 organization field. This is contrary to the PKIX definition but
10614 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10618 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10619 without disallowing inline assembler and the like for non-pedantic builds.
10622 *) Support Borland C++ builder.
10623 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10625 *) Support Mingw32.
10628 *) SHA-1 cleanups and performance enhancements.
10629 [Andy Polyakov <appro@fy.chalmers.se>]
10631 *) Sparc v8plus assembler for the bignum library.
10632 [Andy Polyakov <appro@fy.chalmers.se>]
10634 *) Accept any -xxx and +xxx compiler options in Configure.
10637 *) Update HPUX configuration.
10640 *) Add missing sk_<type>_unshift() function to safestack.h
10641 [Ralf S. Engelschall]
10643 *) New function SSL_CTX_use_certificate_chain_file that sets the
10644 "extra_cert"s in addition to the certificate. (This makes sense
10645 only for "PEM" format files, as chains as a whole are not
10649 *) Support verify_depth from the SSL API.
10650 x509_vfy.c had what can be considered an off-by-one-error:
10651 Its depth (which was not part of the external interface)
10652 was actually counting the number of certificates in a chain;
10653 now it really counts the depth.
10656 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10657 instead of X509err, which often resulted in confusing error
10658 messages since the error codes are not globally unique
10659 (e.g. an alleged error in ssl3_accept when a certificate
10660 didn't match the private key).
10662 *) New function SSL_CTX_set_session_id_context that allows to set a default
10663 value (so that you don't need SSL_set_session_id_context for each
10664 connection using the SSL_CTX).
10667 *) OAEP decoding bug fix.
10670 *) Support INSTALL_PREFIX for package builders, as proposed by
10674 *) New Configure options "threads" and "no-threads". For systems
10675 where the proper compiler options are known (currently Solaris
10676 and Linux), "threads" is the default.
10679 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10682 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10683 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10684 such as /usr/local/bin.
10687 *) "make linux-shared" to build shared libraries.
10688 [Niels Poppe <niels@netbox.org>]
10690 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10693 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10694 extension adding in x509 utility.
10697 *) Remove NOPROTO sections and error code comments.
10700 *) Partial rewrite of the DEF file generator to now parse the ANSI
10704 *) New Configure options --prefix=DIR and --openssldir=DIR.
10707 *) Complete rewrite of the error code script(s). It is all now handled
10708 by one script at the top level which handles error code gathering,
10709 header rewriting and C source file generation. It should be much better
10710 than the old method: it now uses a modified version of Ulf's parser to
10711 read the ANSI prototypes in all header files (thus the old K&R definitions
10712 aren't needed for error creation any more) and do a better job of
10713 translating function codes into names. The old 'ASN1 error code imbedded
10714 in a comment' is no longer necessary and it doesn't use .err files which
10715 have now been deleted. Also the error code call doesn't have to appear all
10716 on one line (which resulted in some large lines...).
10719 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10722 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10723 0 (which usually indicates a closed connection), but continue reading.
10726 *) Fix some race conditions.
10729 *) Add support for CRL distribution points extension. Add Certificate
10730 Policies and CRL distribution points documentation.
10733 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10736 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10737 8 of keying material. Merlin has also confirmed interop with this fix
10738 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10739 [Merlin Hughes <merlin@baltimore.ie>]
10741 *) Fix lots of warnings.
10742 [Richard Levitte <levitte@stacken.kth.se>]
10744 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10745 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10746 [Richard Levitte <levitte@stacken.kth.se>]
10748 *) Fix problems with sizeof(long) == 8.
10749 [Andy Polyakov <appro@fy.chalmers.se>]
10751 *) Change functions to ANSI C.
10754 *) Fix typos in error codes.
10755 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10757 *) Remove defunct assembler files from Configure.
10760 *) SPARC v8 assembler BIGNUM implementation.
10761 [Andy Polyakov <appro@fy.chalmers.se>]
10763 *) Support for Certificate Policies extension: both print and set.
10764 Various additions to support the r2i method this uses.
10767 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10768 return a const string when you are expecting an allocated buffer.
10771 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10772 types DirectoryString and DisplayText.
10775 *) Add code to allow r2i extensions to access the configuration database,
10776 add an LHASH database driver and add several ctx helper functions.
10779 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10780 fail when they extended the size of a BIGNUM.
10783 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10784 support typesafe stack.
10787 *) Fix typo in SSL_[gs]et_options().
10788 [Nils Frostberg <nils@medcom.se>]
10790 *) Delete various functions and files that belonged to the (now obsolete)
10791 old X509V3 handling code.
10794 *) New Configure option "rsaref".
10797 *) Don't auto-generate pem.h.
10800 *) Introduce type-safe ASN.1 SETs.
10803 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10804 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10806 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10807 that links with OpenSSL (well at least cause lots of warnings), but fear
10808 not: the conversion is trivial, and it eliminates loads of evil casts. A
10809 few STACKed things have been converted already. Feel free to convert more.
10810 In the fullness of time, I'll do away with the STACK type altogether.
10813 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10814 specified in <certfile> by updating the entry in the index.txt file.
10815 This way one no longer has to edit the index.txt file manually for
10816 revoking a certificate. The -revoke option does the gory details now.
10817 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10819 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10820 `-text' option at all and this way the `-noout -text' combination was
10821 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10822 [Ralf S. Engelschall]
10824 *) Make sure a corresponding plain text error message exists for the
10825 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10826 verify callback function determined that a certificate was revoked.
10827 [Ralf S. Engelschall]
10829 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10830 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10831 all available cipers including rc5, which was forgotten until now.
10832 In order to let the testing shell script know which algorithms
10833 are available, a new (up to now undocumented) command
10834 "openssl list-cipher-commands" is used.
10837 *) Bugfix: s_client occasionally would sleep in select() when
10838 it should have checked SSL_pending() first.
10841 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10842 the raw DSA values prior to ASN.1 encoding.
10845 *) Tweaks to Configure
10846 [Niels Poppe <niels@netbox.org>]
10848 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10852 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10855 *) New config option to avoid instructions that are illegal on the 80386.
10856 The default code is faster, but requires at least a 486.
10859 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10860 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10861 same as SSL2_VERSION anyway.
10864 *) New "-showcerts" option for s_client.
10867 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10868 application. Various cleanups and fixes.
10871 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10872 modify error routines to work internally. Add error codes and PBE init
10873 to library startup routines.
10876 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10877 packing functions to asn1 and evp. Changed function names and error
10878 codes along the way.
10881 *) PKCS12 integration: and so it begins... First of several patches to
10882 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10883 objects to objects.h
10886 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10887 and display support for Thawte strong extranet extension.
10890 *) Add LinuxPPC support.
10891 [Jeff Dubrule <igor@pobox.org>]
10893 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10894 bn_div_words in alpha.s.
10895 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10897 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10898 OAEP isn't supported when OpenSSL is built with RSAref.
10899 [Ulf Moeller <ulf@fitug.de>]
10901 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10902 so they no longer are missing under -DNOPROTO.
10903 [Soren S. Jorvang <soren@t.dk>]
10906 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10908 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10909 doesn't work when the session is reused. Coming soon!
10912 *) Fix a security hole, that allows sessions to be reused in the wrong
10913 context thus bypassing client cert protection! All software that uses
10914 client certs and session caches in multiple contexts NEEDS PATCHING to
10915 allow session reuse! A fuller solution is in the works.
10916 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10918 *) Some more source tree cleanups (removed obsolete files
10919 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10920 permission on "config" script to be executable) and a fix for the INSTALL
10922 [Ulf Moeller <ulf@fitug.de>]
10924 *) Remove some legacy and erroneous uses of malloc, free instead of
10926 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10928 *) Make rsa_oaep_test return non-zero on error.
10929 [Ulf Moeller <ulf@fitug.de>]
10931 *) Add support for native Solaris shared libraries. Configure
10932 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10933 if someone would make that last step automatic.
10934 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10936 *) ctx_size was not built with the right compiler during "make links". Fixed.
10939 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10940 except NULL ciphers". This means the default cipher list will no longer
10941 enable NULL ciphers. They need to be specifically enabled e.g. with
10942 the string "DEFAULT:eNULL".
10945 *) Fix to RSA private encryption routines: if p < q then it would
10946 occasionally produce an invalid result. This will only happen with
10947 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10950 *) Be less restrictive and allow also `perl util/perlpath.pl
10951 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10952 because this way one can also use an interpreter named `perl5' (which is
10953 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10954 installed as `perl').
10955 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10957 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10958 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10960 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10961 advapi32.lib to Win32 build and change the pem test comparison
10962 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10963 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10964 and crypto/des/ede_cbcm_enc.c.
10967 *) DES quad checksum was broken on big-endian architectures. Fixed.
10970 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10971 Win32 test batch file so it (might) work again. The Win32 test batch file
10972 is horrible: I feel ill....
10975 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10976 in e_os.h. Audit of header files to check ANSI and non ANSI
10977 sections: 10 functions were absent from non ANSI section and not exported
10978 from Windows DLLs. Fixed up libeay.num for new functions.
10981 *) Make `openssl version' output lines consistent.
10982 [Ralf S. Engelschall]
10984 *) Fix Win32 symbol export lists for BIO functions: Added
10985 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10986 to ms/libeay{16,32}.def.
10987 [Ralf S. Engelschall]
10989 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10990 fine under Unix and passes some trivial tests I've now added. But the
10991 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10992 added to make sure no one expects that this stuff really works in the
10993 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10994 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10996 [Ralf S. Engelschall]
10998 *) Fix the generation of two part addresses in perl.
10999 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11001 *) Add config entry for Linux on MIPS.
11002 [John Tobey <jtobey@channel1.com>]
11004 *) Make links whenever Configure is run, unless we are on Windoze.
11007 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11008 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11012 *) Add a useful kludge to allow package maintainers to specify compiler and
11013 other platforms details on the command line without having to patch the
11014 Configure script everytime: One now can use ``perl Configure
11015 <id>:<details>'', i.e. platform ids are allowed to have details appended
11016 to them (separated by colons). This is treated as there would be a static
11017 pre-configured entry in Configure's %table under key <id> with value
11018 <details> and ``perl Configure <id>'' is called. So, when you want to
11019 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11020 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11021 now, which overrides the FreeBSD-elf entry on-the-fly.
11022 [Ralf S. Engelschall]
11024 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11027 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11028 on the `perl Configure ...' command line. This way one can compile
11029 OpenSSL libraries with Position Independent Code (PIC) which is needed
11030 for linking it into DSOs.
11031 [Ralf S. Engelschall]
11033 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11037 *) Cleaned up the LICENSE document: The official contact for any license
11038 questions now is the OpenSSL core team under openssl-core@openssl.org.
11039 And add a paragraph about the dual-license situation to make sure people
11040 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11041 to the OpenSSL toolkit.
11042 [Ralf S. Engelschall]
11044 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11045 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11046 Additionally cleaned up the `make links' target: Remove unnecessary
11047 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11048 to speed processing and no longer clutter the display with confusing
11049 stuff. Instead only the actually done links are displayed.
11050 [Ralf S. Engelschall]
11052 *) Permit null encryption ciphersuites, used for authentication only. It used
11053 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11054 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11058 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11059 signed attributes when verifying signatures (this would break them),
11060 the detached data encoding was wrong and public keys obtained using
11061 X509_get_pubkey() weren't freed.
11064 *) Add text documentation for the BUFFER functions. Also added a work around
11065 to a Win95 console bug. This was triggered by the password read stuff: the
11066 last character typed gets carried over to the next fread(). If you were
11067 generating a new cert request using 'req' for example then the last
11068 character of the passphrase would be CR which would then enter the first
11072 *) Added the new `Includes OpenSSL Cryptography Software' button as
11073 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11074 button and can be used by applications based on OpenSSL to show the
11075 relationship to the OpenSSL project.
11076 [Ralf S. Engelschall]
11078 *) Remove confusing variables in function signatures in files
11079 ssl/ssl_lib.c and ssl/ssl.h.
11080 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11082 *) Don't install bss_file.c under PREFIX/include/
11083 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11085 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11086 functions that return function pointers and has support for NT specific
11087 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11088 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11089 unsigned to signed types: this was killing the Win32 compile.
11092 *) Add new certificate file to stack functions,
11093 SSL_add_dir_cert_subjects_to_stack() and
11094 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11095 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11096 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11097 This means that Apache-SSL and similar packages don't have to mess around
11098 to add as many CAs as they want to the preferred list.
11101 *) Experiment with doxygen documentation. Currently only partially applied to
11103 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11104 openssl.doxy as the configuration file.
11107 *) Get rid of remaining C++-style comments which strict C compilers hate.
11108 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11110 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11111 compiled in by default: it has problems with large keys.
11114 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11115 DH private keys and/or callback functions which directly correspond to
11116 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11117 is needed for applications which have to configure certificates on a
11118 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11120 For the RSA certificate situation is makes no difference, but
11121 for the DSA certificate situation this fixes the "no shared cipher"
11122 problem where the OpenSSL cipher selection procedure failed because the
11123 temporary keys were not overtaken from the context and the API provided
11124 no way to reconfigure them.
11125 The new functions now let applications reconfigure the stuff and they
11126 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11127 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11128 non-public-API function ssl_cert_instantiate() is used as a helper
11129 function and also to reduce code redundancy inside ssl_rsa.c.
11130 [Ralf S. Engelschall]
11132 *) Move s_server -dcert and -dkey options out of the undocumented feature
11133 area because they are useful for the DSA situation and should be
11134 recognized by the users.
11135 [Ralf S. Engelschall]
11137 *) Fix the cipher decision scheme for export ciphers: the export bits are
11138 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11139 SSL_EXP_MASK. So, the original variable has to be used instead of the
11140 already masked variable.
11141 [Richard Levitte <levitte@stacken.kth.se>]
11143 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11144 [Richard Levitte <levitte@stacken.kth.se>]
11146 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11147 from `int' to `unsigned int' because it's a length and initialized by
11148 EVP_DigestFinal() which expects an `unsigned int *'.
11149 [Richard Levitte <levitte@stacken.kth.se>]
11151 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11152 script. Instead use the usual Shell->Perl transition trick.
11153 [Ralf S. Engelschall]
11155 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11156 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11157 -noout -modulus' as it's already the case for `openssl rsa -noout
11158 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11159 currently the public key is printed (a decision which was already done by
11160 `openssl dsa -modulus' in the past) which serves a similar purpose.
11161 Additionally the NO_RSA no longer completely removes the whole -modulus
11162 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11164 [Ralf S. Engelschall]
11166 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11167 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11168 [Arne Ansper <arne@ats.cyber.ee>]
11170 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11171 to be added. Now both 'req' and 'ca' can use new objects defined in the
11175 *) Add cool BIO that does syslog (or event log on NT).
11176 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11178 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11179 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11180 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11181 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11184 *) Add preliminary config info for new extension code.
11187 *) Make RSA_NO_PADDING really use no padding.
11188 [Ulf Moeller <ulf@fitug.de>]
11190 *) Generate errors when private/public key check is done.
11193 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11194 for some CRL extensions and new objects added.
11197 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11198 key usage extension and fuller support for authority key id.
11201 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11202 padding method for RSA, which is recommended for new applications in PKCS
11203 #1 v2.0 (RFC 2437, October 1998).
11204 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11205 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11206 against Bleichbacher's attack on RSA.
11207 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11210 *) Updates to the new SSL compression code
11211 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11213 *) Fix so that the version number in the master secret, when passed
11214 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11215 (because the server will not accept higher), that the version number
11216 is 0x03,0x01, not 0x03,0x00
11217 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11219 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11220 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11221 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11224 *) Support for RAW extensions where an arbitrary extension can be
11225 created by including its DER encoding. See apps/openssl.cnf for
11229 *) Make sure latest Perl versions don't interpret some generated C array
11230 code as Perl array code in the crypto/err/err_genc.pl script.
11231 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11233 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11234 not many people have the assembler. Various Win32 compilation fixes and
11235 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11236 build instructions.
11239 *) Modify configure script 'Configure' to automatically create crypto/date.h
11240 file under Win32 and also build pem.h from pem.org. New script
11241 util/mkfiles.pl to create the MINFO file on environments that can't do a
11242 'make files': perl util/mkfiles.pl >MINFO should work.
11245 *) Major rework of DES function declarations, in the pursuit of correctness
11246 and purity. As a result, many evil casts evaporated, and some weirdness,
11247 too. You may find this causes warnings in your code. Zapping your evil
11248 casts will probably fix them. Mostly.
11251 *) Fix for a typo in asn1.h. Bug fix to object creation script
11252 obj_dat.pl. It considered a zero in an object definition to mean
11253 "end of object": none of the objects in objects.h have any zeros
11254 so it wasn't spotted.
11255 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11257 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11258 Masking (CBCM). In the absence of test vectors, the best I have been able
11259 to do is check that the decrypt undoes the encrypt, so far. Send me test
11260 vectors if you have them.
11263 *) Correct calculation of key length for export ciphers (too much space was
11264 allocated for null ciphers). This has not been tested!
11267 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11268 message is now correct (it understands "crypto" and "ssl" on its
11269 command line). There is also now an "update" option. This will update
11270 the util/ssleay.num and util/libeay.num files with any new functions.
11272 perl util/mkdef.pl crypto ssl update
11273 it will update them.
11276 *) Overhauled the Perl interface (perl/*):
11277 - ported BN stuff to OpenSSL's different BN library
11278 - made the perl/ source tree CVS-aware
11279 - renamed the package from SSLeay to OpenSSL (the files still contain
11280 their history because I've copied them in the repository)
11281 - removed obsolete files (the test scripts will be replaced
11282 by better Test::Harness variants in the future)
11283 [Ralf S. Engelschall]
11285 *) First cut for a very conservative source tree cleanup:
11286 1. merge various obsolete readme texts into doc/ssleay.txt
11287 where we collect the old documents and readme texts.
11288 2. remove the first part of files where I'm already sure that we no
11289 longer need them because of three reasons: either they are just temporary
11290 files which were left by Eric or they are preserved original files where
11291 I've verified that the diff is also available in the CVS via "cvs diff
11292 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11293 the crypto/md/ stuff).
11294 [Ralf S. Engelschall]
11296 *) More extension code. Incomplete support for subject and issuer alt
11297 name, issuer and authority key id. Change the i2v function parameters
11298 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11299 what that's for :-) Fix to ASN1 macro which messed up
11300 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11303 *) Preliminary support for ENUMERATED type. This is largely copied from the
11307 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11308 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11310 *) Make sure `make rehash' target really finds the `openssl' program.
11311 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11313 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11314 like to hear about it if this slows down other processors.
11317 *) Add CygWin32 platform information to Configure script.
11318 [Alan Batie <batie@aahz.jf.intel.com>]
11320 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11321 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11323 *) New program nseq to manipulate netscape certificate sequences
11326 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11330 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11331 but the BN code had some problems that would cause failures when
11332 doing certificate verification and some other functions.
11333 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11335 *) Add ASN1 and PEM code to support netscape certificate sequences.
11338 *) Add ASN1 and PEM code to support netscape certificate sequences.
11341 *) Add several PKIX and private extended key usage OIDs.
11344 *) Modify the 'ca' program to handle the new extension code. Modify
11345 openssl.cnf for new extension format, add comments.
11348 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11349 and add a sample to openssl.cnf so req -x509 now adds appropriate
11353 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11354 error code, add initial support to X509_print() and x509 application.
11357 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11358 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11359 stuff is currently isolated and isn't even compiled yet.
11362 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11363 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11364 Removed the versions check from X509 routines when loading extensions:
11365 this allows certain broken certificates that don't set the version
11366 properly to be processed.
11369 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11370 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11371 can still be regenerated with "make depend".
11374 *) Spelling mistake in C version of CAST-128.
11375 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11377 *) Changes to the error generation code. The perl script err-code.pl
11378 now reads in the old error codes and retains the old numbers, only
11379 adding new ones if necessary. It also only changes the .err files if new
11380 codes are added. The makefiles have been modified to only insert errors
11381 when needed (to avoid needlessly modifying header files). This is done
11382 by only inserting errors if the .err file is newer than the auto generated
11383 C file. To rebuild all the error codes from scratch (the old behaviour)
11384 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11385 or delete all the .err files.
11388 *) CAST-128 was incorrectly implemented for short keys. The C version has
11389 been fixed, but is untested. The assembler versions are also fixed, but
11390 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11391 to regenerate it if needed.
11392 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11393 Hagino <itojun@kame.net>]
11395 *) File was opened incorrectly in randfile.c.
11396 [Ulf Möller <ulf@fitug.de>]
11398 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11399 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11400 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11401 al: it's just almost always a UTCTime. Note this patch adds new error
11402 codes so do a "make errors" if there are problems.
11405 *) Correct Linux 1 recognition in config.
11406 [Ulf Möller <ulf@fitug.de>]
11408 *) Remove pointless MD5 hash when using DSA keys in ca.
11409 [Anonymous <nobody@replay.com>]
11411 *) Generate an error if given an empty string as a cert directory. Also
11412 generate an error if handed NULL (previously returned 0 to indicate an
11413 error, but didn't set one).
11414 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11416 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11419 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11420 parameters. This was causing a warning which killed off the Win32 compile.
11423 *) Remove C++ style comments from crypto/bn/bn_local.h.
11424 [Neil Costigan <neil.costigan@celocom.com>]
11426 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11427 based on a text string, looking up short and long names and finally
11428 "dot" format. The "dot" format stuff didn't work. Added new function
11429 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11430 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11431 OID is not part of the table.
11434 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11435 X509_LOOKUP_by_alias().
11438 *) Sort openssl functions by name.
11441 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11442 encryption from sample DSA keys (in case anyone is interested the password
11446 *) Make _all_ *_free functions accept a NULL pointer.
11447 [Frans Heymans <fheymans@isaserver.be>]
11449 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11451 [Anonymous <nobody@replay.com>]
11453 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11454 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11456 *) Don't blow it for numeric -newkey arguments to apps/req.
11457 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11459 *) Temp key "for export" tests were wrong in s3_srvr.c.
11460 [Anonymous <nobody@replay.com>]
11462 *) Add prototype for temp key callback functions
11463 SSL_CTX_set_tmp_{rsa,dh}_callback().
11466 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11467 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11470 *) X509_name_add_entry() freed the wrong thing after an error.
11471 [Arne Ansper <arne@ats.cyber.ee>]
11473 *) rsa_eay.c would attempt to free a NULL context.
11474 [Arne Ansper <arne@ats.cyber.ee>]
11476 *) BIO_s_socket() had a broken should_retry() on Windoze.
11477 [Arne Ansper <arne@ats.cyber.ee>]
11479 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11480 [Arne Ansper <arne@ats.cyber.ee>]
11482 *) Make sure the already existing X509_STORE->depth variable is initialized
11483 in X509_STORE_new(), but document the fact that this variable is still
11484 unused in the certificate verification process.
11485 [Ralf S. Engelschall]
11487 *) Fix the various library and apps files to free up pkeys obtained from
11488 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11491 *) Fix reference counting in X509_PUBKEY_get(). This makes
11492 demos/maurice/example2.c work, amongst others, probably.
11493 [Steve Henson and Ben Laurie]
11495 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11496 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11497 are no longer created. This way we have a single and consistent command
11498 line interface `openssl <command>', similar to `cvs <command>'.
11499 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11501 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11502 BIT STRING wrapper always have zero unused bits.
11505 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11508 *) Make the top-level INSTALL documentation easier to understand.
11511 *) Makefiles updated to exit if an error occurs in a sub-directory
11512 make (including if user presses ^C) [Paul Sutton]
11514 *) Make Montgomery context stuff explicit in RSA data structure.
11517 *) Fix build order of pem and err to allow for generated pem.h.
11520 *) Fix renumbering bug in X509_NAME_delete_entry().
11523 *) Enhanced the err-ins.pl script so it makes the error library number
11524 global and can add a library name. This is needed for external ASN1 and
11525 other error libraries.
11528 *) Fixed sk_insert which never worked properly.
11531 *) Fix ASN1 macros so they can handle indefinite length constructed
11532 EXPLICIT tags. Some non standard certificates use these: they can now
11536 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11537 into a single doc/ssleay.txt bundle. This way the information is still
11538 preserved but no longer messes up this directory. Now it's new room for
11539 the new set of documentation files.
11540 [Ralf S. Engelschall]
11542 *) SETs were incorrectly DER encoded. This was a major pain, because they
11543 shared code with SEQUENCEs, which aren't coded the same. This means that
11544 almost everything to do with SETs or SEQUENCEs has either changed name or
11545 number of arguments.
11546 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11548 *) Fix test data to work with the above.
11551 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11552 was already fixed by Eric for 0.9.1 it seems.
11553 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11555 *) Autodetect FreeBSD3.
11558 *) Fix various bugs in Configure. This affects the following platforms:
11562 unixware-2.0-pentium
11566 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11567 before they are needed.
11570 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11574 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11576 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11577 changed SSLeay to OpenSSL in version strings.
11578 [Ralf S. Engelschall]
11580 *) Some fixups to the top-level documents.
11583 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11584 because the symlink to include/ was missing.
11585 [Ralf S. Engelschall]
11587 *) Incorporated the popular no-RSA/DSA-only patches
11588 which allow to compile a RSA-free SSLeay.
11589 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11591 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11592 when "ssleay" is still not found.
11593 [Ralf S. Engelschall]
11595 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11596 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11598 *) Updated the README file.
11599 [Ralf S. Engelschall]
11601 *) Added various .cvsignore files in the CVS repository subdirs
11602 to make a "cvs update" really silent.
11603 [Ralf S. Engelschall]
11605 *) Recompiled the error-definition header files and added
11606 missing symbols to the Win32 linker tables.
11607 [Ralf S. Engelschall]
11609 *) Cleaned up the top-level documents;
11610 o new files: CHANGES and LICENSE
11611 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11612 o merged COPYRIGHT into LICENSE
11613 o removed obsolete TODO file
11614 o renamed MICROSOFT to INSTALL.W32
11615 [Ralf S. Engelschall]
11617 *) Removed dummy files from the 0.9.1b source tree:
11618 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11619 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11620 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11621 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11622 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11623 [Ralf S. Engelschall]
11625 *) Added various platform portability fixes.
11628 *) The Genesis of the OpenSSL rpject:
11629 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11630 Young and Tim J. Hudson created while they were working for C2Net until
11632 [The OpenSSL Project]
11635 Changes between 0.9.0b and 0.9.1b [not released]
11637 *) Updated a few CA certificates under certs/
11640 *) Changed some BIGNUM api stuff.
11643 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11644 DGUX x86, Linux Alpha, etc.
11647 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11648 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11652 *) Add -strparse option to asn1pars program which parses nested
11654 [Dr Stephen Henson <shenson@bigfoot.com>]
11656 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11659 *) DSA fix for "ca" program.
11662 *) Added "-genkey" option to "dsaparam" program.
11665 *) Added RIPE MD160 (rmd160) message digest.
11668 *) Added -a (all) option to "ssleay version" command.
11671 *) Added PLATFORM define which is the id given to Configure.
11674 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11677 *) Extended the ASN.1 parser routines.
11680 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11683 *) Added a BN_CTX to the BN library.
11686 *) Fixed the weak key values in DES library
11689 *) Changed API in EVP library for cipher aliases.
11692 *) Added support for RC2/64bit cipher.
11695 *) Converted the lhash library to the crypto/mem.c functions.
11698 *) Added more recognized ASN.1 object ids.
11701 *) Added more RSA padding checks for SSL/TLS.
11704 *) Added BIO proxy/filter functionality.
11707 *) Added extra_certs to SSL_CTX which can be used
11708 send extra CA certificates to the client in the CA cert chain sending
11709 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11712 *) Now Fortezza is denied in the authentication phase because
11713 this is key exchange mechanism is not supported by SSLeay at all.
11716 *) Additional PKCS1 checks.
11719 *) Support the string "TLSv1" for all TLS v1 ciphers.
11722 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11723 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11726 *) Fixed a few memory leaks.
11729 *) Fixed various code and comment typos.
11732 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11733 bytes sent in the client random.
11734 [Edward Bishop <ebishop@spyglass.com>]