5 Changes between 1.0.2f and 1.1.0 [xx XXX xxxx]
7 *) Always DPURIFY. Remove the use of uninitialized memory in the
8 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
11 *) Removed many obsolete configuration items, including
12 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
13 MD2_CHAR, MD2_INT, MD2_LONG
16 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
17 [Rich Salz, with advice from Andy Polyakov]
19 *) Configuration and writing out the results from it has changed.
20 Files such as Makefile include/openssl/opensslconf.h and are now
21 produced through general templates, such as Makefile.in and
22 crypto/opensslconf.h.in and some help from the perl module
25 Also, the center of configuration information is no longer
26 Makefile. Instead, Configure produces a perl module in
27 configdata.pm which holds most of the config data (in the hash
28 table %config), the target data that comes from the target
29 configuration in one of the Configurations/*.conf files (in
33 *) To clarify their intended purposes, the Configure options
34 --prefix and --openssldir change their semantics, and become more
35 straightforward and less interdependent.
37 --prefix shall be used exclusively to give the location INSTALLTOP
38 where programs, scripts, libraries, include files and manuals are
39 going to be installed. The default is now /usr/local.
41 --openssldir shall be used exclusively to give the default
42 location OPENSSLDIR where certificates, private keys, CRLs are
43 managed. This is also where the default openssl.cnf gets
45 If the directory given with this option is a relative path, the
46 values of both the --prefix value and the --openssldir value will
47 be combined to become OPENSSLDIR.
48 The default for --openssldir is INSTALLTOP/ssl.
50 Anyone who uses --openssldir to specify where OpenSSL is to be
51 installed MUST change to use --prefix instead.
54 *) The GOST engine was out of date and therefore it has been removed. An up
55 to date GOST engine is now being maintained in an external repository.
56 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
57 support for GOST ciphersuites (these are only activated if a GOST engine
61 *) EGD is no longer supported by default; use enable-egd when
63 [Ben Kaduk and Rich Salz]
65 *) The distribution now has Makefile.in files, which are used to
66 create Makefile's when Configure is run. *Configure must be run
67 before trying to build now.*
70 *) The return value for SSL_CIPHER_description() for error conditions
74 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
76 Obtaining and performing DNSSEC validation of TLSA records is
77 the application's responsibility. The application provides
78 the TLSA records of its choice to OpenSSL, and these are then
79 used to authenticate the peer.
81 The TLSA records need not even come from DNS. They can, for
82 example, be used to implement local end-entity certificate or
83 trust-anchor "pinning", where the "pin" data takes the form
84 of TLSA records, which can augment or replace verification
85 based on the usual WebPKI public certification authorities.
88 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
89 continues to support deprecated interfaces in default builds.
90 However, applications are strongly advised to compile their
91 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
92 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
93 or the 1.1.0 releases.
95 In environments in which all applications have been ported to
96 not use any deprecated interfaces OpenSSL's Configure script
97 should be used with the --api=1.1.0 option to entirely remove
98 support for the deprecated features from the library and
99 unconditionally disable them in the installed headers.
100 Essentially the same effect can be achieved with the "no-deprecated"
101 argument to Configure, except that this will always restrict
102 the build to just the latest API, rather than a fixed API
105 As applications are ported to future revisions of the API,
106 they should update their compile-time OPENSSL_API_COMPAT define
107 accordingly, but in most cases should be able to continue to
108 compile with later releases.
110 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
111 0x10000000L and 0x00908000L, respectively. However those
112 versions did not support the OPENSSL_API_COMPAT feature, and
113 so applications are not typically tested for explicit support
114 of just the undeprecated features of either release.
117 *) Add support for setting the minimum and maximum supported protocol.
118 It can bet set via the SSL_set_min_proto_version() and
119 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
120 MaxProtcol. It's recommended to use the new APIs to disable
121 protocols instead of disabling individual protocols using
122 SSL_set_options() or SSL_CONF's Protocol. This change also
123 removes support for disabling TLS 1.2 in the OpenSSL TLS
124 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
127 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
130 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
131 and integrates ECDSA and ECDH functionality into EC. Implementations can
132 now redirect key generation and no longer need to convert to or from
135 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
136 include the ec.h header file instead.
139 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
140 ciphers who are no longer supported and drops support the ephemeral RSA key
141 exchange. The LOW ciphers currently doesn't have any ciphers in it.
144 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
145 opaque. For HMAC_CTX, the following constructors and destructors
148 HMAC_CTX *HMAC_CTX_new(void);
149 void HMAC_CTX_free(HMAC_CTX *ctx);
151 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
152 destroy such methods has been added. See EVP_MD_meth_new(3) and
153 EVP_CIPHER_meth_new(3) for documentation.
156 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
157 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
158 EVP_MD_CTX_reset() should be called instead to reinitialise
159 an already created structure.
160 2) For consistency with the majority of our object creators and
161 destructors, EVP_MD_CTX_(create|destroy) were renamed to
162 EVP_MD_CTX_(new|free). The old names are retained as macros
163 for deprecated builds.
166 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
167 cryptographic operations to be performed asynchronously as long as an
168 asynchronous capable engine is used. See the ASYNC_start_job() man page for
169 further details. Libssl has also had this capability integrated with the
170 introduction of the new mode SSL_MODE_ASYNC and associated error
171 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
172 pages. This work was developed in partnership with Intel Corp.
175 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
176 always enabled now. If you want to disable the support you should
177 exclude it using the list of supported ciphers. This also means that the
178 "-no_ecdhe" option has been removed from s_server.
181 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
182 SSL_{CTX_}set1_curves() which can set a list.
185 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
186 curve you want to support using SSL_{CTX_}set1_curves().
189 *) State machine rewrite. The state machine code has been significantly
190 refactored in order to remove much duplication of code and solve issues
191 with the old code (see ssl/statem/README for further details). This change
192 does have some associated API changes. Notably the SSL_state() function
193 has been removed and replaced by SSL_get_state which now returns an
194 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
195 altogether. The previous handshake states defined in ssl.h and ssl3.h have
199 *) All instances of the string "ssleay" in the public API were replaced
200 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
201 Some error codes related to internal RSA_eay API's were renamed.
204 *) The demo files in crypto/threads were moved to demo/threads.
207 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
209 [Matt Caswell, Rich Salz]
211 *) New ASN.1 embed macro.
213 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
214 structure is not allocated: it is part of the parent. That is instead of
222 This reduces memory fragmentation and make it impossible to accidentally
223 set a mandatory field to NULL.
225 This currently only works for some fields specifically a SEQUENCE, CHOICE,
226 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
227 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
231 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
234 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
235 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
236 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
237 DES and RC4 ciphersuites.
240 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
241 This changes the decoding behaviour for some invalid messages,
242 though the change is mostly in the more lenient direction, and
243 legacy behaviour is preserved as much as possible.
246 *) Fix no-stdio build.
247 [ David Woodhouse <David.Woodhouse@intel.com> and also
248 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
250 *) New testing framework
251 The testing framework has been largely rewritten and is now using
252 perl and the perl modules Test::Harness and an extended variant of
253 Test::More called OpenSSL::Test to do its work. All test scripts in
254 test/ have been rewritten into test recipes, and all direct calls to
255 executables in test/Makefile have become individual recipes using the
256 simplified testing OpenSSL::Test::Simple.
258 For documentation on our testing modules, do:
260 perldoc test/testlib/OpenSSL/Test/Simple.pm
261 perldoc test/testlib/OpenSSL/Test.pm
265 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
266 are used; the latter aborts on memory leaks (usually checked on exit).
267 Some undocumented "set malloc, etc., hooks" functions were removed
268 and others were changed. All are now documented.
271 *) In DSA_generate_parameters_ex, if the provided seed is too short,
273 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
275 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
276 from RFC4279, RFC4785, RFC5487, RFC5489.
278 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
279 original RSA_PSK patch.
282 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
283 era flag was never set throughout the codebase (only read). Also removed
284 SSL3_FLAGS_POP_BUFFER which was only used if
285 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
288 *) Changed the default name options in the "ca", "crl", "req" and "x509"
289 to be "oneline" instead of "compat".
292 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
293 not aware of clients that still exhibit this bug, and the workaround
294 hasn't been working properly for a while.
297 *) The return type of BIO_number_read() and BIO_number_written() as well as
298 the corresponding num_read and num_write members in the BIO structure has
299 changed from unsigned long to uint64_t. On platforms where an unsigned
300 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
304 *) Given the pervasive nature of TLS extensions it is inadvisable to run
305 OpenSSL without support for them. It also means that maintaining
306 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
307 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
310 *) Removed support for the two export grade static DH ciphersuites
311 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
312 were newly added (along with a number of other static DH ciphersuites) to
313 1.0.2. However the two export ones have *never* worked since they were
314 introduced. It seems strange in any case to be adding new export
315 ciphersuites, and given "logjam" it also does not seem correct to fix them.
318 *) Version negotiation has been rewritten. In particular SSLv23_method(),
319 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
320 and turned into macros which simply call the new preferred function names
321 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
322 should use the new names instead. Also as part of this change the ssl23.h
323 header file has been removed.
326 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
327 code and the associated standard is no longer considered fit-for-purpose.
330 *) RT2547 was closed. When generating a private key, try to make the
331 output file readable only by the owner. This behavior change might
332 be noticeable when interacting with other software.
334 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
338 *) Added HTTP GET support to the ocsp command.
341 *) Changed default digest for the dgst and enc commands from MD5 to
345 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
348 *) Added support for TLS extended master secret from
349 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
350 initial patch which was a great help during development.
353 *) All libssl internal structures have been removed from the public header
354 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
355 now redundant). Users should not attempt to access internal structures
356 directly. Instead they should use the provided API functions.
359 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
360 Access to deprecated functions can be re-enabled by running config with
361 "enable-deprecated". In addition applications wishing to use deprecated
362 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
363 will, by default, disable some transitive includes that previously existed
364 in the header files (e.g. ec.h will no longer, by default, include bn.h)
367 *) Added support for OCB mode. OpenSSL has been granted a patent license
368 compatible with the OpenSSL license for use of OCB. Details are available
369 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
370 for OCB can be removed by calling config with no-ocb.
373 *) SSLv2 support has been removed. It still supports receiving a SSLv2
374 compatible client hello.
377 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
378 done while fixing the error code for the key-too-small case.
379 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
381 *) CA.sh has been removmed; use CA.pl instead.
384 *) Removed old DES API.
387 *) Remove various unsupported platforms:
393 Sinix/ReliantUNIX RM400
398 16-bit platforms such as WIN16
401 *) Clean up OPENSSL_NO_xxx #define's
402 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
403 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
404 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
405 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
406 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
407 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
408 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
409 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
410 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
411 Remove MS_STATIC; it's a relic from platforms <32 bits.
414 *) Cleaned up dead code
415 Remove all but one '#ifdef undef' which is to be looked at.
418 *) Clean up calling of xxx_free routines.
419 Just like free(), fix most of the xxx_free routines to accept
420 NULL. Remove the non-null checks from callers. Save much code.
423 *) Add secure heap for storage of private keys (when possible).
424 Add BIO_s_secmem(), CBIGNUM, etc.
425 Contributed by Akamai Technologies under our Corporate CLA.
428 *) Experimental support for a new, fast, unbiased prime candidate generator,
429 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
430 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
432 *) New output format NSS in the sess_id command line tool. This allows
433 exporting the session id and the master key in NSS keylog format.
434 [Martin Kaiser <martin@kaiser.cx>]
436 *) Harmonize version and its documentation. -f flag is used to display
438 [mancha <mancha1@zoho.com>]
440 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
441 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
442 [mancha <mancha1@zoho.com>]
444 *) Fix some double frees. These are not thought to be exploitable.
445 [mancha <mancha1@zoho.com>]
447 *) A missing bounds check in the handling of the TLS heartbeat extension
448 can be used to reveal up to 64k of memory to a connected client or
451 Thanks for Neel Mehta of Google Security for discovering this bug and to
452 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
453 preparing the fix (CVE-2014-0160)
454 [Adam Langley, Bodo Moeller]
456 *) Fix for the attack described in the paper "Recovering OpenSSL
457 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
458 by Yuval Yarom and Naomi Benger. Details can be obtained from:
459 http://eprint.iacr.org/2014/140
461 Thanks to Yuval Yarom and Naomi Benger for discovering this
462 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
463 [Yuval Yarom and Naomi Benger]
465 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
466 this fixes a limitation in previous versions of OpenSSL.
469 *) Experimental encrypt-then-mac support.
471 Experimental support for encrypt then mac from
472 draft-gutmann-tls-encrypt-then-mac-02.txt
474 To enable it set the appropriate extension number (0x42 for the test
475 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
477 For non-compliant peers (i.e. just about everything) this should have no
480 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
484 *) Add EVP support for key wrapping algorithms, to avoid problems with
485 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
486 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
487 algorithms and include tests cases.
490 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
494 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
495 MGF1 digest and OAEP label.
498 *) Make openssl verify return errors.
499 [Chris Palmer <palmer@google.com> and Ben Laurie]
501 *) New function ASN1_TIME_diff to calculate the difference between two
502 ASN1_TIME structures or one structure and the current time.
505 *) Update fips_test_suite to support multiple command line options. New
506 test to induce all self test errors in sequence and check expected
510 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
511 sign or verify all in one operation.
514 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
515 test programs and fips_test_suite. Includes functionality to parse
516 the minimal script output of fipsalgest.pl directly.
519 *) Add authorisation parameter to FIPS_module_mode_set().
522 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
525 *) Use separate DRBG fields for internal and external flags. New function
526 FIPS_drbg_health_check() to perform on demand health checking. Add
527 generation tests to fips_test_suite with reduced health check interval to
528 demonstrate periodic health checking. Add "nodh" option to
529 fips_test_suite to skip very slow DH test.
532 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
536 *) More extensive health check for DRBG checking many more failure modes.
537 New function FIPS_selftest_drbg_all() to handle every possible DRBG
538 combination: call this in fips_test_suite.
541 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
542 and POST to handle Dual EC cases.
545 *) Add support for canonical generation of DSA parameter 'g'. See
548 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
549 POST to handle HMAC cases.
552 *) Add functions FIPS_module_version() and FIPS_module_version_text()
553 to return numerical and string versions of the FIPS module number.
556 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
557 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
558 outside the validated module in the FIPS capable OpenSSL.
561 *) Minor change to DRBG entropy callback semantics. In some cases
562 there is no multiple of the block length between min_len and
563 max_len. Allow the callback to return more than max_len bytes
564 of entropy but discard any extra: it is the callback's responsibility
565 to ensure that the extra data discarded does not impact the
566 requested amount of entropy.
569 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
570 information in FIPS186-3, SP800-57 and SP800-131A.
573 *) CCM support via EVP. Interface is very similar to GCM case except we
574 must supply all data in one chunk (i.e. no update, final) and the
575 message length must be supplied if AAD is used. Add algorithm test
579 *) Initial version of POST overhaul. Add POST callback to allow the status
580 of POST to be monitored and/or failures induced. Modify fips_test_suite
581 to use callback. Always run all selftests even if one fails.
584 *) XTS support including algorithm test driver in the fips_gcmtest program.
585 Note: this does increase the maximum key length from 32 to 64 bytes but
586 there should be no binary compatibility issues as existing applications
587 will never use XTS mode.
590 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
591 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
592 performs algorithm blocking for unapproved PRNG types. Also do not
593 set PRNG type in FIPS_mode_set(): leave this to the application.
594 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
595 the standard OpenSSL PRNG: set additional data to a date time vector.
598 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
599 This shouldn't present any incompatibility problems because applications
600 shouldn't be using these directly and any that are will need to rethink
601 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
604 *) Extensive self tests and health checking required by SP800-90 DRBG.
605 Remove strength parameter from FIPS_drbg_instantiate and always
606 instantiate at maximum supported strength.
609 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
612 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
615 *) New function DH_compute_key_padded() to compute a DH key and pad with
616 leading zeroes if needed: this complies with SP800-56A et al.
619 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
620 anything, incomplete, subject to change and largely untested at present.
623 *) Modify fipscanisteronly build option to only build the necessary object
624 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
627 *) Add experimental option FIPSSYMS to give all symbols in
628 fipscanister.o and FIPS or fips prefix. This will avoid
629 conflicts with future versions of OpenSSL. Add perl script
630 util/fipsas.pl to preprocess assembly language source files
631 and rename any affected symbols.
634 *) Add selftest checks and algorithm block of non-fips algorithms in
635 FIPS mode. Remove DES2 from selftests.
638 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
639 return internal method without any ENGINE dependencies. Add new
640 tiny fips sign and verify functions.
643 *) New build option no-ec2m to disable characteristic 2 code.
646 *) New build option "fipscanisteronly". This only builds fipscanister.o
647 and (currently) associated fips utilities. Uses the file Makefile.fips
648 instead of Makefile.org as the prototype.
651 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
652 Update fips_gcmtest to use IV generator.
655 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
656 setting output buffer to NULL. The *Final function must be
657 called although it will not retrieve any additional data. The tag
658 can be set or retrieved with a ctrl. The IV length is by default 12
659 bytes (96 bits) but can be set to an alternative value. If the IV
660 length exceeds the maximum IV length (currently 16 bytes) it cannot be
664 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
665 underlying do_cipher function handles all cipher semantics itself
666 including padding and finalisation. This is useful if (for example)
667 an ENGINE cipher handles block padding itself. The behaviour of
668 do_cipher is subtly changed if this flag is set: the return value
669 is the number of characters written to the output buffer (zero is
670 no longer an error code) or a negative error code. Also if the
671 input buffer is NULL and length 0 finalisation should be performed.
674 *) If a candidate issuer certificate is already part of the constructed
675 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
678 *) Improve forward-security support: add functions
680 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
681 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
683 for use by SSL/TLS servers; the callback function will be called whenever a
684 new session is created, and gets to decide whether the session may be
685 cached to make it resumable (return 0) or not (return 1). (As by the
686 SSL/TLS protocol specifications, the session_id sent by the server will be
687 empty to indicate that the session is not resumable; also, the server will
688 not generate RFC 4507 (RFC 5077) session tickets.)
690 A simple reasonable callback implementation is to return is_forward_secure.
691 This parameter will be set to 1 or 0 depending on the ciphersuite selected
692 by the SSL/TLS server library, indicating whether it can provide forward
694 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
696 *) New -verify_name option in command line utilities to set verification
700 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
701 Add CMAC pkey methods.
704 *) Experimental renegotiation in s_server -www mode. If the client
705 browses /reneg connection is renegotiated. If /renegcert it is
706 renegotiated requesting a certificate.
709 *) Add an "external" session cache for debugging purposes to s_server. This
710 should help trace issues which normally are only apparent in deployed
711 multi-process servers.
714 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
715 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
716 BIO_set_cipher() and some obscure PEM functions were changed so they
717 can now return an error. The RAND changes required a change to the
718 RAND_METHOD structure.
721 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
722 a gcc attribute to warn if the result of a function is ignored. This
723 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
724 whose return value is often ignored.
727 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
728 *) DH small subgroups
730 Historically OpenSSL only ever generated DH parameters based on "safe"
731 primes. More recently (in version 1.0.2) support was provided for
732 generating X9.42 style parameter files such as those required for RFC 5114
733 support. The primes used in such files may not be "safe". Where an
734 application is using DH configured with parameters based on primes that are
735 not "safe" then an attacker could use this fact to find a peer's private
736 DH exponent. This attack requires that the attacker complete multiple
737 handshakes in which the peer uses the same private DH exponent. For example
738 this could be used to discover a TLS server's private DH exponent if it's
739 reusing the private DH exponent or it's using a static DH ciphersuite.
741 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
742 TLS. It is not on by default. If the option is not set then the server
743 reuses the same private DH exponent for the life of the server process and
744 would be vulnerable to this attack. It is believed that many popular
745 applications do set this option and would therefore not be at risk.
747 The fix for this issue adds an additional check where a "q" parameter is
748 available (as is the case in X9.42 based parameters). This detects the
749 only known attack, and is the only possible defense for static DH
750 ciphersuites. This could have some performance impact.
752 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
753 default and cannot be disabled. This could have some performance impact.
755 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
759 *) SSLv2 doesn't block disabled ciphers
761 A malicious client can negotiate SSLv2 ciphers that have been disabled on
762 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
763 been disabled, provided that the SSLv2 protocol was not also disabled via
766 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
767 and Sebastian Schinzel.
771 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
773 *) BN_mod_exp may produce incorrect results on x86_64
775 There is a carry propagating bug in the x86_64 Montgomery squaring
776 procedure. No EC algorithms are affected. Analysis suggests that attacks
777 against RSA and DSA as a result of this defect would be very difficult to
778 perform and are not believed likely. Attacks against DH are considered just
779 feasible (although very difficult) because most of the work necessary to
780 deduce information about a private key may be performed offline. The amount
781 of resources required for such an attack would be very significant and
782 likely only accessible to a limited number of attackers. An attacker would
783 additionally need online access to an unpatched system using the target
784 private key in a scenario with persistent DH parameters and a private
785 key that is shared between multiple clients. For example this can occur by
786 default in OpenSSL DHE based SSL/TLS ciphersuites.
788 This issue was reported to OpenSSL by Hanno Böck.
792 *) Certificate verify crash with missing PSS parameter
794 The signature verification routines will crash with a NULL pointer
795 dereference if presented with an ASN.1 signature using the RSA PSS
796 algorithm and absent mask generation function parameter. Since these
797 routines are used to verify certificate signature algorithms this can be
798 used to crash any certificate verification operation and exploited in a
799 DoS attack. Any application which performs certificate verification is
800 vulnerable including OpenSSL clients and servers which enable client
803 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
807 *) X509_ATTRIBUTE memory leak
809 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
810 memory. This structure is used by the PKCS#7 and CMS routines so any
811 application which reads PKCS#7 or CMS data from untrusted sources is
812 affected. SSL/TLS is not affected.
814 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
819 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
820 This changes the decoding behaviour for some invalid messages,
821 though the change is mostly in the more lenient direction, and
822 legacy behaviour is preserved as much as possible.
825 *) In DSA_generate_parameters_ex, if the provided seed is too short,
827 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
829 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
831 *) Alternate chains certificate forgery
833 During certificate verfification, OpenSSL will attempt to find an
834 alternative certificate chain if the first attempt to build such a chain
835 fails. An error in the implementation of this logic can mean that an
836 attacker could cause certain checks on untrusted certificates to be
837 bypassed, such as the CA flag, enabling them to use a valid leaf
838 certificate to act as a CA and "issue" an invalid certificate.
840 This issue was reported to OpenSSL by Adam Langley/David Benjamin
844 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
846 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
847 incompatibility in the handling of HMAC. The previous ABI has now been
851 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
853 *) Malformed ECParameters causes infinite loop
855 When processing an ECParameters structure OpenSSL enters an infinite loop
856 if the curve specified is over a specially malformed binary polynomial
859 This can be used to perform denial of service against any
860 system which processes public keys, certificate requests or
861 certificates. This includes TLS clients and TLS servers with
862 client authentication enabled.
864 This issue was reported to OpenSSL by Joseph Barr-Pixton.
868 *) Exploitable out-of-bounds read in X509_cmp_time
870 X509_cmp_time does not properly check the length of the ASN1_TIME
871 string and can read a few bytes out of bounds. In addition,
872 X509_cmp_time accepts an arbitrary number of fractional seconds in the
875 An attacker can use this to craft malformed certificates and CRLs of
876 various sizes and potentially cause a segmentation fault, resulting in
877 a DoS on applications that verify certificates or CRLs. TLS clients
878 that verify CRLs are affected. TLS clients and servers with client
879 authentication enabled may be affected if they use custom verification
882 This issue was reported to OpenSSL by Robert Swiecki (Google), and
883 independently by Hanno Böck.
887 *) PKCS7 crash with missing EnvelopedContent
889 The PKCS#7 parsing code does not handle missing inner EncryptedContent
890 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
891 with missing content and trigger a NULL pointer dereference on parsing.
893 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
894 structures from untrusted sources are affected. OpenSSL clients and
895 servers are not affected.
897 This issue was reported to OpenSSL by Michal Zalewski (Google).
901 *) CMS verify infinite loop with unknown hash function
903 When verifying a signedData message the CMS code can enter an infinite loop
904 if presented with an unknown hash function OID. This can be used to perform
905 denial of service against any system which verifies signedData messages using
907 This issue was reported to OpenSSL by Johannes Bauer.
911 *) Race condition handling NewSessionTicket
913 If a NewSessionTicket is received by a multi-threaded client when attempting to
914 reuse a previous ticket then a race condition can occur potentially leading to
915 a double free of the ticket data.
919 *) Only support 256-bit or stronger elliptic curves with the
920 'ecdh_auto' setting (server) or by default (client). Of supported
921 curves, prefer P-256 (both).
924 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
926 *) ClientHello sigalgs DoS fix
928 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
929 invalid signature algorithms extension a NULL pointer dereference will
930 occur. This can be exploited in a DoS attack against the server.
932 This issue was was reported to OpenSSL by David Ramos of Stanford
935 [Stephen Henson and Matt Caswell]
937 *) Multiblock corrupted pointer fix
939 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
940 feature only applies on 64 bit x86 architecture platforms that support AES
941 NI instructions. A defect in the implementation of "multiblock" can cause
942 OpenSSL's internal write buffer to become incorrectly set to NULL when
943 using non-blocking IO. Typically, when the user application is using a
944 socket BIO for writing, this will only result in a failed connection.
945 However if some other BIO is used then it is likely that a segmentation
946 fault will be triggered, thus enabling a potential DoS attack.
948 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
952 *) Segmentation fault in DTLSv1_listen fix
954 The DTLSv1_listen function is intended to be stateless and processes the
955 initial ClientHello from many peers. It is common for user code to loop
956 over the call to DTLSv1_listen until a valid ClientHello is received with
957 an associated cookie. A defect in the implementation of DTLSv1_listen means
958 that state is preserved in the SSL object from one invocation to the next
959 that can lead to a segmentation fault. Errors processing the initial
960 ClientHello can trigger this scenario. An example of such an error could be
961 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
964 This issue was reported to OpenSSL by Per Allansson.
968 *) Segmentation fault in ASN1_TYPE_cmp fix
970 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
971 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
972 certificate signature algorithm consistency this can be used to crash any
973 certificate verification operation and exploited in a DoS attack. Any
974 application which performs certificate verification is vulnerable including
975 OpenSSL clients and servers which enable client authentication.
979 *) Segmentation fault for invalid PSS parameters fix
981 The signature verification routines will crash with a NULL pointer
982 dereference if presented with an ASN.1 signature using the RSA PSS
983 algorithm and invalid parameters. Since these routines are used to verify
984 certificate signature algorithms this can be used to crash any
985 certificate verification operation and exploited in a DoS attack. Any
986 application which performs certificate verification is vulnerable including
987 OpenSSL clients and servers which enable client authentication.
989 This issue was was reported to OpenSSL by Brian Carpenter.
993 *) ASN.1 structure reuse memory corruption fix
995 Reusing a structure in ASN.1 parsing may allow an attacker to cause
996 memory corruption via an invalid write. Such reuse is and has been
997 strongly discouraged and is believed to be rare.
999 Applications that parse structures containing CHOICE or ANY DEFINED BY
1000 components may be affected. Certificate parsing (d2i_X509 and related
1001 functions) are however not affected. OpenSSL clients and servers are
1006 *) PKCS7 NULL pointer dereferences fix
1008 The PKCS#7 parsing code does not handle missing outer ContentInfo
1009 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1010 missing content and trigger a NULL pointer dereference on parsing.
1012 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1013 otherwise parse PKCS#7 structures from untrusted sources are
1014 affected. OpenSSL clients and servers are not affected.
1016 This issue was reported to OpenSSL by Michal Zalewski (Google).
1020 *) DoS via reachable assert in SSLv2 servers fix
1022 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1023 servers that both support SSLv2 and enable export cipher suites by sending
1024 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1026 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1027 (OpenSSL development team).
1031 *) Empty CKE with client auth and DHE fix
1033 If client auth is used then a server can seg fault in the event of a DHE
1034 ciphersuite being selected and a zero length ClientKeyExchange message
1035 being sent by the client. This could be exploited in a DoS attack.
1039 *) Handshake with unseeded PRNG fix
1041 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1042 with an unseeded PRNG. The conditions are:
1043 - The client is on a platform where the PRNG has not been seeded
1044 automatically, and the user has not seeded manually
1045 - A protocol specific client method version has been used (i.e. not
1046 SSL_client_methodv23)
1047 - A ciphersuite is used that does not require additional random data from
1048 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1050 If the handshake succeeds then the client random that has been used will
1051 have been generated from a PRNG with insufficient entropy and therefore the
1052 output may be predictable.
1054 For example using the following command with an unseeded openssl will
1055 succeed on an unpatched platform:
1057 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1061 *) Use After Free following d2i_ECPrivatekey error fix
1063 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1064 could cause a use after free condition. This, in turn, could cause a double
1065 free in several private key parsing functions (such as d2i_PrivateKey
1066 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1067 for applications that receive EC private keys from untrusted
1068 sources. This scenario is considered rare.
1070 This issue was discovered by the BoringSSL project and fixed in their
1075 *) X509_to_X509_REQ NULL pointer deref fix
1077 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1078 the certificate key is invalid. This function is rarely used in practice.
1080 This issue was discovered by Brian Carpenter.
1084 *) Removed the export ciphers from the DEFAULT ciphers
1087 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1089 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1090 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1091 So far those who have to target multiple plaforms would compromise
1092 and argue that binary targeting say ARMv5 would still execute on
1093 ARMv8. "Universal" build resolves this compromise by providing
1094 near-optimal performance even on newer platforms.
1097 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1098 (other platforms pending).
1099 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1101 *) Add support for the SignedCertificateTimestampList certificate and
1102 OCSP response extensions from RFC6962.
1105 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1106 for corner cases. (Certain input points at infinity could lead to
1107 bogus results, with non-infinity inputs mapped to infinity too.)
1110 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1111 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1112 common cases are optimized and there still is room for further
1113 improvements. Vector Permutation AES for Altivec is also added.
1116 *) Add support for little-endian ppc64 Linux target.
1117 [Marcelo Cerri (IBM)]
1119 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1120 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1121 are optimized and there still is room for further improvements.
1122 Both 32- and 64-bit modes are supported.
1123 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1125 *) Improved ARMv7 NEON support.
1128 *) Support for SPARC Architecture 2011 crypto extensions, first
1129 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1130 SHA256/512, MD5, GHASH and modular exponentiation.
1131 [Andy Polyakov, David Miller]
1133 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1135 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1137 *) Support for new and upcoming Intel processors, including AVX2,
1138 BMI and SHA ISA extensions. This includes additional "stitched"
1139 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1142 This work was sponsored by Intel Corp.
1145 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1146 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1147 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1150 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1151 this fixes a limiation in previous versions of OpenSSL.
1154 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1155 MGF1 digest and OAEP label.
1158 *) Add EVP support for key wrapping algorithms, to avoid problems with
1159 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1160 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1161 algorithms and include tests cases.
1164 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1166 [Douglas E. Engert, Steve Henson]
1168 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1169 difference in days and seconds between two tm or ASN1_TIME structures.
1172 *) Add -rev test option to s_server to just reverse order of characters
1173 received by client and send back to server. Also prints an abbreviated
1174 summary of the connection parameters.
1177 *) New option -brief for s_client and s_server to print out a brief summary
1178 of connection parameters.
1181 *) Add callbacks for arbitrary TLS extensions.
1182 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1184 *) New option -crl_download in several openssl utilities to download CRLs
1185 from CRLDP extension in certificates.
1188 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1191 *) New function X509_CRL_diff to generate a delta CRL from the difference
1192 of two full CRLs. Add support to "crl" utility.
1195 *) New functions to set lookup_crls function and to retrieve
1196 X509_STORE from X509_STORE_CTX.
1199 *) Print out deprecated issuer and subject unique ID fields in
1203 *) Extend OCSP I/O functions so they can be used for simple general purpose
1204 HTTP as well as OCSP. New wrapper function which can be used to download
1205 CRLs using the OCSP API.
1208 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1211 *) SSL_CONF* functions. These provide a common framework for application
1212 configuration using configuration files or command lines.
1215 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1216 message callback and prints the results. Needs compile time option
1217 "enable-ssl-trace". New options to s_client and s_server to enable
1221 *) New ctrl and macro to retrieve supported points extensions.
1222 Print out extension in s_server and s_client.
1225 *) New functions to retrieve certificate signature and signature
1229 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1233 *) New Suite B modes for TLS code. These use and enforce the requirements
1234 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1235 only use Suite B curves. The Suite B modes can be set by using the
1236 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1239 *) New chain verification flags for Suite B levels of security. Check
1240 algorithms are acceptable when flags are set in X509_verify_cert.
1243 *) Make tls1_check_chain return a set of flags indicating checks passed
1244 by a certificate chain. Add additional tests to handle client
1245 certificates: checks for matching certificate type and issuer name
1249 *) If an attempt is made to use a signature algorithm not in the peer
1250 preference list abort the handshake. If client has no suitable
1251 signature algorithms in response to a certificate request do not
1252 use the certificate.
1255 *) If server EC tmp key is not in client preference list abort handshake.
1258 *) Add support for certificate stores in CERT structure. This makes it
1259 possible to have different stores per SSL structure or one store in
1260 the parent SSL_CTX. Include distint stores for certificate chain
1261 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1262 to build and store a certificate chain in CERT structure: returing
1263 an error if the chain cannot be built: this will allow applications
1264 to test if a chain is correctly configured.
1266 Note: if the CERT based stores are not set then the parent SSL_CTX
1267 store is used to retain compatibility with existing behaviour.
1271 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1272 mask based on the current session, check mask when sending client
1273 hello and checking the requested ciphersuite.
1276 *) New ctrls to retrieve and set certificate types in a certificate
1277 request message. Print out received values in s_client. If certificate
1278 types is not set with custom values set sensible values based on
1279 supported signature algorithms.
1282 *) Support for distinct client and server supported signature algorithms.
1285 *) Add certificate callback. If set this is called whenever a certificate
1286 is required by client or server. An application can decide which
1287 certificate chain to present based on arbitrary criteria: for example
1288 supported signature algorithms. Add very simple example to s_server.
1289 This fixes many of the problems and restrictions of the existing client
1290 certificate callback: for example you can now clear an existing
1291 certificate and specify the whole chain.
1294 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1295 the certificate can be used for (if anything). Set valid_flags field
1296 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1297 to have similar checks in it.
1299 Add new "cert_flags" field to CERT structure and include a "strict mode".
1300 This enforces some TLS certificate requirements (such as only permitting
1301 certificate signature algorithms contained in the supported algorithms
1302 extension) which some implementations ignore: this option should be used
1303 with caution as it could cause interoperability issues.
1306 *) Update and tidy signature algorithm extension processing. Work out
1307 shared signature algorithms based on preferences and peer algorithms
1308 and print them out in s_client and s_server. Abort handshake if no
1309 shared signature algorithms.
1312 *) Add new functions to allow customised supported signature algorithms
1313 for SSL and SSL_CTX structures. Add options to s_client and s_server
1317 *) New function SSL_certs_clear() to delete all references to certificates
1318 from an SSL structure. Before this once a certificate had been added
1319 it couldn't be removed.
1322 *) Integrate hostname, email address and IP address checking with certificate
1323 verification. New verify options supporting checking in opensl utility.
1326 *) Fixes and wildcard matching support to hostname and email checking
1327 functions. Add manual page.
1328 [Florian Weimer (Red Hat Product Security Team)]
1330 *) New functions to check a hostname email or IP address against a
1331 certificate. Add options x509 utility to print results of checks against
1335 *) Fix OCSP checking.
1336 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1338 *) Initial experimental support for explicitly trusted non-root CAs.
1339 OpenSSL still tries to build a complete chain to a root but if an
1340 intermediate CA has a trust setting included that is used. The first
1341 setting is used: whether to trust (e.g., -addtrust option to the x509
1345 *) Add -trusted_first option which attempts to find certificates in the
1346 trusted store even if an untrusted chain is also supplied.
1349 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1350 platform support for Linux and Android.
1353 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1356 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1357 When in FIPS mode the approved implementations are used as normal,
1358 when not in FIPS mode the internal unapproved versions are used instead.
1359 This means that the FIPS capable OpenSSL isn't forced to use the
1360 (often lower performance) FIPS implementations outside FIPS mode.
1363 *) Transparently support X9.42 DH parameters when calling
1364 PEM_read_bio_DHparameters. This means existing applications can handle
1365 the new parameter format automatically.
1368 *) Initial experimental support for X9.42 DH parameter format: mainly
1369 to support use of 'q' parameter for RFC5114 parameters.
1372 *) Add DH parameters from RFC5114 including test data to dhtest.
1375 *) Support for automatic EC temporary key parameter selection. If enabled
1376 the most preferred EC parameters are automatically used instead of
1377 hardcoded fixed parameters. Now a server just has to call:
1378 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1379 support ECDH and use the most appropriate parameters.
1382 *) Enhance and tidy EC curve and point format TLS extension code. Use
1383 static structures instead of allocation if default values are used.
1384 New ctrls to set curves we wish to support and to retrieve shared curves.
1385 Print out shared curves in s_server. New options to s_server and s_client
1386 to set list of supported curves.
1389 *) New ctrls to retrieve supported signature algorithms and
1390 supported curve values as an array of NIDs. Extend openssl utility
1391 to print out received values.
1394 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1395 between NIDs and the more common NIST names such as "P-256". Enhance
1396 ecparam utility and ECC method to recognise the NIST names for curves.
1399 *) Enhance SSL/TLS certificate chain handling to support different
1400 chains for each certificate instead of one chain in the parent SSL_CTX.
1403 *) Support for fixed DH ciphersuite client authentication: where both
1404 server and client use DH certificates with common parameters.
1407 *) Support for fixed DH ciphersuites: those requiring DH server
1411 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1413 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1414 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1415 X509_CINF_get_signature were reverted post internal team review.
1417 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1419 *) Build fixes for the Windows and OpenVMS platforms
1420 [Matt Caswell and Richard Levitte]
1422 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1424 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1425 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1426 dereference. This could lead to a Denial Of Service attack. Thanks to
1427 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1431 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1432 dtls1_buffer_record function under certain conditions. In particular this
1433 could occur if an attacker sent repeated DTLS records with the same
1434 sequence number but for the next epoch. The memory leak could be exploited
1435 by an attacker in a Denial of Service attack through memory exhaustion.
1436 Thanks to Chris Mueller for reporting this issue.
1440 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1441 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1442 method would be set to NULL which could later result in a NULL pointer
1443 dereference. Thanks to Frank Schmirler for reporting this issue.
1447 *) Abort handshake if server key exchange message is omitted for ephemeral
1450 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1451 reporting this issue.
1455 *) Remove non-export ephemeral RSA code on client and server. This code
1456 violated the TLS standard by allowing the use of temporary RSA keys in
1457 non-export ciphersuites and could be used by a server to effectively
1458 downgrade the RSA key length used to a value smaller than the server
1459 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1460 INRIA or reporting this issue.
1464 *) Fixed issue where DH client certificates are accepted without verification.
1465 An OpenSSL server will accept a DH certificate for client authentication
1466 without the certificate verify message. This effectively allows a client to
1467 authenticate without the use of a private key. This only affects servers
1468 which trust a client certificate authority which issues certificates
1469 containing DH keys: these are extremely rare and hardly ever encountered.
1470 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1475 *) Ensure that the session ID context of an SSL is updated when its
1476 SSL_CTX is updated via SSL_set_SSL_CTX.
1478 The session ID context is typically set from the parent SSL_CTX,
1479 and can vary with the CTX.
1482 *) Fix various certificate fingerprint issues.
1484 By using non-DER or invalid encodings outside the signed portion of a
1485 certificate the fingerprint can be changed without breaking the signature.
1486 Although no details of the signed portion of the certificate can be changed
1487 this can cause problems with some applications: e.g. those using the
1488 certificate fingerprint for blacklists.
1490 1. Reject signatures with non zero unused bits.
1492 If the BIT STRING containing the signature has non zero unused bits reject
1493 the signature. All current signature algorithms require zero unused bits.
1495 2. Check certificate algorithm consistency.
1497 Check the AlgorithmIdentifier inside TBS matches the one in the
1498 certificate signature. NB: this will result in signature failure
1499 errors for some broken certificates.
1501 Thanks to Konrad Kraszewski from Google for reporting this issue.
1503 3. Check DSA/ECDSA signatures use DER.
1505 Reencode DSA/ECDSA signatures and compare with the original received
1506 signature. Return an error if there is a mismatch.
1508 This will reject various cases including garbage after signature
1509 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1510 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1511 (negative or with leading zeroes).
1513 Further analysis was conducted and fixes were developed by Stephen Henson
1514 of the OpenSSL core team.
1519 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1520 results on some platforms, including x86_64. This bug occurs at random
1521 with a very low probability, and is not known to be exploitable in any
1522 way, though its exact impact is difficult to determine. Thanks to Pieter
1523 Wuille (Blockstream) who reported this issue and also suggested an initial
1524 fix. Further analysis was conducted by the OpenSSL development team and
1525 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1526 the OpenSSL core team.
1530 *) Do not resume sessions on the server if the negotiated protocol
1531 version does not match the session's version. Resuming with a different
1532 version, while not strictly forbidden by the RFC, is of questionable
1533 sanity and breaks all known clients.
1534 [David Benjamin, Emilia Käsper]
1536 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1537 early CCS messages during renegotiation. (Note that because
1538 renegotiation is encrypted, this early CCS was not exploitable.)
1541 *) Tighten client-side session ticket handling during renegotiation:
1542 ensure that the client only accepts a session ticket if the server sends
1543 the extension anew in the ServerHello. Previously, a TLS client would
1544 reuse the old extension state and thus accept a session ticket if one was
1545 announced in the initial ServerHello.
1547 Similarly, ensure that the client requires a session ticket if one
1548 was advertised in the ServerHello. Previously, a TLS client would
1549 ignore a missing NewSessionTicket message.
1552 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1554 *) SRTP Memory Leak.
1556 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1557 sends a carefully crafted handshake message, to cause OpenSSL to fail
1558 to free up to 64k of memory causing a memory leak. This could be
1559 exploited in a Denial Of Service attack. This issue affects OpenSSL
1560 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1561 whether SRTP is used or configured. Implementations of OpenSSL that
1562 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1564 The fix was developed by the OpenSSL team.
1568 *) Session Ticket Memory Leak.
1570 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1571 integrity of that ticket is first verified. In the event of a session
1572 ticket integrity check failing, OpenSSL will fail to free memory
1573 causing a memory leak. By sending a large number of invalid session
1574 tickets an attacker could exploit this issue in a Denial Of Service
1579 *) Build option no-ssl3 is incomplete.
1581 When OpenSSL is configured with "no-ssl3" as a build option, servers
1582 could accept and complete a SSL 3.0 handshake, and clients could be
1583 configured to send them.
1585 [Akamai and the OpenSSL team]
1587 *) Add support for TLS_FALLBACK_SCSV.
1588 Client applications doing fallback retries should call
1589 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1591 [Adam Langley, Bodo Moeller]
1593 *) Add additional DigestInfo checks.
1595 Reencode DigestInto in DER and check against the original when
1596 verifying RSA signature: this will reject any improperly encoded
1597 DigestInfo structures.
1599 Note: this is a precautionary measure and no attacks are currently known.
1603 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1605 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1606 SRP code can be overrun an internal buffer. Add sanity check that
1607 g, A, B < N to SRP code.
1609 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1610 Group for discovering this issue.
1614 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1615 TLS 1.0 instead of higher protocol versions when the ClientHello message
1616 is badly fragmented. This allows a man-in-the-middle attacker to force a
1617 downgrade to TLS 1.0 even if both the server and the client support a
1618 higher protocol version, by modifying the client's TLS records.
1620 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1621 researching this issue.
1625 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1626 to a denial of service attack. A malicious server can crash the client
1627 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1628 ciphersuite and sending carefully crafted handshake messages.
1630 Thanks to Felix Gröbert (Google) for discovering and researching this
1635 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1636 to leak memory. This can be exploited through a Denial of Service attack.
1637 Thanks to Adam Langley for discovering and researching this issue.
1641 *) An attacker can force openssl to consume large amounts of memory whilst
1642 processing DTLS handshake messages. This can be exploited through a
1643 Denial of Service attack.
1644 Thanks to Adam Langley for discovering and researching this issue.
1648 *) An attacker can force an error condition which causes openssl to crash
1649 whilst processing DTLS packets due to memory being freed twice. This
1650 can be exploited through a Denial of Service attack.
1651 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1656 *) If a multithreaded client connects to a malicious server using a resumed
1657 session and the server sends an ec point format extension it could write
1658 up to 255 bytes to freed memory.
1660 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1665 *) A malicious server can crash an OpenSSL client with a null pointer
1666 dereference (read) by specifying an SRP ciphersuite even though it was not
1667 properly negotiated with the client. This can be exploited through a
1668 Denial of Service attack.
1670 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1671 discovering and researching this issue.
1675 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1676 X509_name_oneline, X509_name_print_ex et al. to leak some information
1677 from the stack. Applications may be affected if they echo pretty printing
1678 output to the attacker.
1680 Thanks to Ivan Fratric (Google) for discovering this issue.
1682 [Emilia Käsper, and Steve Henson]
1684 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1685 for corner cases. (Certain input points at infinity could lead to
1686 bogus results, with non-infinity inputs mapped to infinity too.)
1689 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1691 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1692 handshake can force the use of weak keying material in OpenSSL
1693 SSL/TLS clients and servers.
1695 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1696 researching this issue. (CVE-2014-0224)
1697 [KIKUCHI Masashi, Steve Henson]
1699 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1700 OpenSSL DTLS client the code can be made to recurse eventually crashing
1703 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1705 [Imre Rad, Steve Henson]
1707 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1708 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1709 client or server. This is potentially exploitable to run arbitrary
1710 code on a vulnerable client or server.
1712 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1713 [Jüri Aedla, Steve Henson]
1715 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1716 are subject to a denial of service attack.
1718 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1719 this issue. (CVE-2014-3470)
1720 [Felix Gröbert, Ivan Fratric, Steve Henson]
1722 *) Harmonize version and its documentation. -f flag is used to display
1724 [mancha <mancha1@zoho.com>]
1726 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1727 in i2d_ECPrivateKey.
1728 [mancha <mancha1@zoho.com>]
1730 *) Fix some double frees. These are not thought to be exploitable.
1731 [mancha <mancha1@zoho.com>]
1733 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1735 *) A missing bounds check in the handling of the TLS heartbeat extension
1736 can be used to reveal up to 64k of memory to a connected client or
1739 Thanks for Neel Mehta of Google Security for discovering this bug and to
1740 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1741 preparing the fix (CVE-2014-0160)
1742 [Adam Langley, Bodo Moeller]
1744 *) Fix for the attack described in the paper "Recovering OpenSSL
1745 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1746 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1747 http://eprint.iacr.org/2014/140
1749 Thanks to Yuval Yarom and Naomi Benger for discovering this
1750 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1751 [Yuval Yarom and Naomi Benger]
1753 *) TLS pad extension: draft-agl-tls-padding-03
1755 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1756 TLS client Hello record length value would otherwise be > 255 and
1757 less that 512 pad with a dummy extension containing zeroes so it
1758 is at least 512 bytes long.
1760 [Adam Langley, Steve Henson]
1762 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1764 *) Fix for TLS record tampering bug. A carefully crafted invalid
1765 handshake could crash OpenSSL with a NULL pointer exception.
1766 Thanks to Anton Johansson for reporting this issues.
1769 *) Keep original DTLS digest and encryption contexts in retransmission
1770 structures so we can use the previous session parameters if they need
1771 to be resent. (CVE-2013-6450)
1774 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1775 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1776 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1777 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1778 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1779 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1780 [Rob Stradling, Adam Langley]
1782 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1784 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1785 supporting platforms or when small records were transferred.
1786 [Andy Polyakov, Steve Henson]
1788 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1790 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1792 This addresses the flaw in CBC record processing discovered by
1793 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1794 at: http://www.isg.rhul.ac.uk/tls/
1796 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1797 Security Group at Royal Holloway, University of London
1798 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1799 Emilia Käsper for the initial patch.
1801 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1803 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1804 ciphersuites which can be exploited in a denial of service attack.
1805 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1806 and detecting this bug and to Wolfgang Ettlinger
1807 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1811 *) Return an error when checking OCSP signatures when key is NULL.
1812 This fixes a DoS attack. (CVE-2013-0166)
1815 *) Make openssl verify return errors.
1816 [Chris Palmer <palmer@google.com> and Ben Laurie]
1818 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1819 the right response is stapled. Also change SSL_get_certificate()
1820 so it returns the certificate actually sent.
1821 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1822 [Rob Stradling <rob.stradling@comodo.com>]
1824 *) Fix possible deadlock when decoding public keys.
1827 *) Don't use TLS 1.0 record version number in initial client hello
1831 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1833 *) Sanity check record length before skipping explicit IV in TLS
1834 1.2, 1.1 and DTLS to fix DoS attack.
1836 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1837 fuzzing as a service testing platform.
1841 *) Initialise tkeylen properly when encrypting CMS messages.
1842 Thanks to Solar Designer of Openwall for reporting this issue.
1845 *) In FIPS mode don't try to use composite ciphers as they are not
1849 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1851 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1852 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1853 mean any application compiled against OpenSSL 1.0.0 headers setting
1854 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1855 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1856 0x10000000L Any application which was previously compiled against
1857 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1858 will need to be recompiled as a result. Letting be results in
1859 inability to disable specifically TLS 1.1 and in client context,
1860 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1863 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1864 disable just protocol X, but all protocols above X *if* there are
1865 protocols *below* X still enabled. In more practical terms it means
1866 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1867 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1868 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1872 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1874 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1875 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1876 in CRYPTO_realloc_clean.
1878 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1879 issue and to Adam Langley <agl@chromium.org> for fixing it.
1881 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1883 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1886 *) Workarounds for some broken servers that "hang" if a client hello
1887 record length exceeds 255 bytes.
1889 1. Do not use record version number > TLS 1.0 in initial client
1890 hello: some (but not all) hanging servers will now work.
1891 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1892 the number of ciphers sent in the client hello. This should be
1893 set to an even number, such as 50, for example by passing:
1894 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1895 Most broken servers should now work.
1896 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1897 TLS 1.2 client support entirely.
1900 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1903 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1905 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1906 STRING form instead of a DigestInfo.
1909 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1910 and the RSA_sign/RSA_verify functions. This was made more apparent when
1911 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1912 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1913 the correct format in RSA_verify so both forms transparently work.
1916 *) Some servers which support TLS 1.0 can choke if we initially indicate
1917 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1918 encrypted premaster secret. As a workaround use the maximum permitted
1919 client version in client hello, this should keep such servers happy
1920 and still work with previous versions of OpenSSL.
1923 *) Add support for TLS/DTLS heartbeats.
1924 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1926 *) Add support for SCTP.
1927 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1929 *) Improved PRNG seeding for VOS.
1930 [Paul Green <Paul.Green@stratus.com>]
1932 *) Extensive assembler packs updates, most notably:
1934 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1935 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1936 - x86_64: bit-sliced AES implementation;
1937 - ARM: NEON support, contemporary platforms optimizations;
1938 - s390x: z196 support;
1939 - *: GHASH and GF(2^m) multiplication implementations;
1943 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1944 (removal of unnecessary code)
1945 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1947 *) Add TLS key material exporter from RFC 5705.
1950 *) Add DTLS-SRTP negotiation from RFC 5764.
1953 *) Add Next Protocol Negotiation,
1954 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1955 disabled with a no-npn flag to config or Configure. Code donated
1957 [Adam Langley <agl@google.com> and Ben Laurie]
1959 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1960 NIST-P256, NIST-P521, with constant-time single point multiplication on
1961 typical inputs. Compiler support for the nonstandard type __uint128_t is
1962 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1963 Code made available under Apache License version 2.0.
1965 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1966 line to include this in your build of OpenSSL, and run "make depend" (or
1967 "make update"). This enables the following EC_METHODs:
1969 EC_GFp_nistp224_method()
1970 EC_GFp_nistp256_method()
1971 EC_GFp_nistp521_method()
1973 EC_GROUP_new_by_curve_name() will automatically use these (while
1974 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1976 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1978 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1979 all platforms. Move ssize_t definition from e_os.h to the public
1980 header file e_os2.h as it now appears in public header file cms.h
1983 *) New -sigopt option to the ca, req and x509 utilities. Additional
1984 signature parameters can be passed using this option and in
1988 *) Add RSA PSS signing function. This will generate and set the
1989 appropriate AlgorithmIdentifiers for PSS based on those in the
1990 corresponding EVP_MD_CTX structure. No application support yet.
1993 *) Support for companion algorithm specific ASN1 signing routines.
1994 New function ASN1_item_sign_ctx() signs a pre-initialised
1995 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1996 the appropriate parameters.
1999 *) Add new algorithm specific ASN1 verification initialisation function
2000 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2001 handling will be the same no matter what EVP_PKEY_METHOD is used.
2002 Add a PSS handler to support verification of PSS signatures: checked
2003 against a number of sample certificates.
2006 *) Add signature printing for PSS. Add PSS OIDs.
2007 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2009 *) Add algorithm specific signature printing. An individual ASN1 method
2010 can now print out signatures instead of the standard hex dump.
2012 More complex signatures (e.g. PSS) can print out more meaningful
2013 information. Include DSA version that prints out the signature
2017 *) Password based recipient info support for CMS library: implementing
2021 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2022 neatly separates the code into cipher and PBE sections and is required
2023 for some algorithms that split PBES2 into separate pieces (such as
2024 password based CMS).
2027 *) Session-handling fixes:
2028 - Fix handling of connections that are resuming with a session ID,
2029 but also support Session Tickets.
2030 - Fix a bug that suppressed issuing of a new ticket if the client
2031 presented a ticket with an expired session.
2032 - Try to set the ticket lifetime hint to something reasonable.
2033 - Make tickets shorter by excluding irrelevant information.
2034 - On the client side, don't ignore renewed tickets.
2035 [Adam Langley, Bodo Moeller (Google)]
2037 *) Fix PSK session representation.
2040 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2042 This work was sponsored by Intel.
2045 *) Add GCM support to TLS library. Some custom code is needed to split
2046 the IV between the fixed (from PRF) and explicit (from TLS record)
2047 portions. This adds all GCM ciphersuites supported by RFC5288 and
2048 RFC5289. Generalise some AES* cipherstrings to include GCM and
2049 add a special AESGCM string for GCM only.
2052 *) Expand range of ctrls for AES GCM. Permit setting invocation
2053 field on decrypt and retrieval of invocation field only on encrypt.
2056 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2057 As required by RFC5289 these ciphersuites cannot be used if for
2058 versions of TLS earlier than 1.2.
2061 *) For FIPS capable OpenSSL interpret a NULL default public key method
2062 as unset and return the appropriate default but do *not* set the default.
2063 This means we can return the appropriate method in applications that
2064 switch between FIPS and non-FIPS modes.
2067 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2068 ENGINE is used then we cannot handle that in the FIPS module so we
2069 keep original code iff non-FIPS operations are allowed.
2072 *) Add -attime option to openssl utilities.
2073 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2075 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2078 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2079 FIPS EC methods unconditionally for now.
2082 *) New build option no-ec2m to disable characteristic 2 code.
2085 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2086 all cases can be covered as some introduce binary incompatibilities.
2089 *) Redirect RSA operations to FIPS module including keygen,
2090 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2093 *) Add similar low level API blocking to ciphers.
2096 *) Low level digest APIs are not approved in FIPS mode: any attempt
2097 to use these will cause a fatal error. Applications that *really* want
2098 to use them can use the private_* version instead.
2101 *) Redirect cipher operations to FIPS module for FIPS builds.
2104 *) Redirect digest operations to FIPS module for FIPS builds.
2107 *) Update build system to add "fips" flag which will link in fipscanister.o
2108 for static and shared library builds embedding a signature if needed.
2111 *) Output TLS supported curves in preference order instead of numerical
2112 order. This is currently hardcoded for the highest order curves first.
2113 This should be configurable so applications can judge speed vs strength.
2116 *) Add TLS v1.2 server support for client authentication.
2119 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2123 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2124 FIPS modules versions.
2127 *) Add TLS v1.2 client side support for client authentication. Keep cache
2128 of handshake records longer as we don't know the hash algorithm to use
2129 until after the certificate request message is received.
2132 *) Initial TLS v1.2 client support. Add a default signature algorithms
2133 extension including all the algorithms we support. Parse new signature
2134 format in client key exchange. Relax some ECC signing restrictions for
2135 TLS v1.2 as indicated in RFC5246.
2138 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2139 to new signature format when needed using client digest preference.
2140 All server ciphersuites should now work correctly in TLS v1.2. No client
2141 support yet and no support for client certificates.
2144 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2145 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2146 ciphersuites. At present only RSA key exchange ciphersuites work with
2147 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2148 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2149 and version checking.
2152 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2153 with this defined it will not be affected by any changes to ssl internal
2154 structures. Add several utility functions to allow openssl application
2155 to work with OPENSSL_NO_SSL_INTERN defined.
2159 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2161 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2164 *) Permit abbreviated handshakes when renegotiating using the function
2165 SSL_renegotiate_abbreviated().
2166 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2168 *) Add call to ENGINE_register_all_complete() to
2169 ENGINE_load_builtin_engines(), so some implementations get used
2170 automatically instead of needing explicit application support.
2173 *) Add support for TLS key exporter as described in RFC5705.
2174 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2176 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2177 a few changes are required:
2179 Add SSL_OP_NO_TLSv1_1 flag.
2180 Add TLSv1_1 methods.
2181 Update version checking logic to handle version 1.1.
2182 Add explicit IV handling (ported from DTLS code).
2183 Add command line options to s_client/s_server.
2186 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2188 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2189 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2190 content decryption and always return the same error. Note: this attack
2191 needs on average 2^20 messages so it only affects automated senders. The
2192 old behaviour can be reenabled in the CMS code by setting the
2193 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2194 an MMA defence is not necessary.
2195 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2196 this issue. (CVE-2012-0884)
2199 *) Fix CVE-2011-4619: make sure we really are receiving a
2200 client hello before rejecting multiple SGC restarts. Thanks to
2201 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2204 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2206 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2207 Thanks to Antonio Martin, Enterprise Secure Access Research and
2208 Development, Cisco Systems, Inc. for discovering this bug and
2209 preparing a fix. (CVE-2012-0050)
2212 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2214 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2215 of the Vaudenay padding oracle attack on CBC mode encryption
2216 which enables an efficient plaintext recovery attack against
2217 the OpenSSL implementation of DTLS. Their attack exploits timing
2218 differences arising during decryption processing. A research
2219 paper describing this attack can be found at:
2220 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2221 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2222 Security Group at Royal Holloway, University of London
2223 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2224 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2225 for preparing the fix. (CVE-2011-4108)
2226 [Robin Seggelmann, Michael Tuexen]
2228 *) Clear bytes used for block padding of SSL 3.0 records.
2230 [Adam Langley (Google)]
2232 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2233 Kadianakis <desnacked@gmail.com> for discovering this issue and
2234 Adam Langley for preparing the fix. (CVE-2011-4619)
2235 [Adam Langley (Google)]
2237 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2238 [Andrey Kulikov <amdeich@gmail.com>]
2240 *) Prevent malformed RFC3779 data triggering an assertion failure.
2241 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2242 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2243 [Rob Austein <sra@hactrn.net>]
2245 *) Improved PRNG seeding for VOS.
2246 [Paul Green <Paul.Green@stratus.com>]
2248 *) Fix ssl_ciph.c set-up race.
2249 [Adam Langley (Google)]
2251 *) Fix spurious failures in ecdsatest.c.
2252 [Emilia Käsper (Google)]
2254 *) Fix the BIO_f_buffer() implementation (which was mixing different
2255 interpretations of the '..._len' fields).
2256 [Adam Langley (Google)]
2258 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2259 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2260 threads won't reuse the same blinding coefficients.
2262 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2263 lock to call BN_BLINDING_invert_ex, and avoids one use of
2264 BN_BLINDING_update for each BN_BLINDING structure (previously,
2265 the last update always remained unused).
2266 [Emilia Käsper (Google)]
2268 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2269 [Bob Buckholz (Google)]
2271 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2273 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2274 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2275 [Kaspar Brand <ossl@velox.ch>]
2277 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2278 for multi-threaded use of ECDH. (CVE-2011-3210)
2279 [Adam Langley (Google)]
2281 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2284 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2285 signature public key algorithm by using OID xref utilities instead.
2286 Before this you could only use some ECC ciphersuites with SHA1 only.
2289 *) Add protection against ECDSA timing attacks as mentioned in the paper
2290 by Billy Bob Brumley and Nicola Tuveri, see:
2292 http://eprint.iacr.org/2011/232.pdf
2294 [Billy Bob Brumley and Nicola Tuveri]
2296 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2298 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2299 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2301 *) Fix bug in string printing code: if *any* escaping is enabled we must
2302 escape the escape character (backslash) or the resulting string is
2306 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2308 *) Disable code workaround for ancient and obsolete Netscape browsers
2309 and servers: an attacker can use it in a ciphersuite downgrade attack.
2310 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2313 *) Fixed J-PAKE implementation error, originally discovered by
2314 Sebastien Martini, further info and confirmation from Stefan
2315 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2318 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2320 *) Fix extension code to avoid race conditions which can result in a buffer
2321 overrun vulnerability: resumed sessions must not be modified as they can
2322 be shared by multiple threads. CVE-2010-3864
2325 *) Fix WIN32 build system to correctly link an ENGINE directory into
2329 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2331 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2333 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2335 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2337 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2338 context. The operation can be customised via the ctrl mechanism in
2339 case ENGINEs want to include additional functionality.
2342 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2345 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2346 output hashes compatible with older versions of OpenSSL.
2347 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2349 *) Fix compression algorithm handling: if resuming a session use the
2350 compression algorithm of the resumed session instead of determining
2351 it from client hello again. Don't allow server to change algorithm.
2354 *) Add load_crls() function to apps tidying load_certs() too. Add option
2355 to verify utility to allow additional CRLs to be included.
2358 *) Update OCSP request code to permit adding custom headers to the request:
2359 some responders need this.
2362 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2364 [Julia Lawall <julia@diku.dk>]
2366 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2367 needlessly dereferenced structures, used obsolete functions and
2368 didn't handle all updated verify codes correctly.
2371 *) Disable MD2 in the default configuration.
2374 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2375 indicate the initial BIO being pushed or popped. This makes it possible
2376 to determine whether the BIO is the one explicitly called or as a result
2377 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2378 it handles reference counts correctly and doesn't zero out the I/O bio
2379 when it is not being explicitly popped. WARNING: applications which
2380 included workarounds for the old buggy behaviour will need to be modified
2381 or they could free up already freed BIOs.
2384 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2385 renaming to all platforms (within the 0.9.8 branch, this was
2386 done conditionally on Netware platforms to avoid a name clash).
2387 [Guenter <lists@gknw.net>]
2389 *) Add ECDHE and PSK support to DTLS.
2390 [Michael Tuexen <tuexen@fh-muenster.de>]
2392 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2396 *) Add "missing" function EVP_MD_flags() (without this the only way to
2397 retrieve a digest flags is by accessing the structure directly. Update
2398 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2399 or cipher is registered as in the "from" argument. Print out all
2400 registered digests in the dgst usage message instead of manually
2401 attempting to work them out.
2404 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2405 this allows the use of compression and extensions. Change default cipher
2406 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2407 by default unless an application cipher string requests it.
2410 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2411 key ids to find matching certificates and keys but some PKCS#12 files
2412 don't follow the (somewhat unwritten) rules and this strategy fails.
2413 Now just gather all certificates together and the first private key
2414 then look for the first certificate that matches the key.
2417 *) Support use of registered digest and cipher names for dgst and cipher
2418 commands instead of having to add each one as a special case. So now
2425 openssl dgst -sha256 foo
2427 and this works for ENGINE based algorithms too.
2431 *) Update Gost ENGINE to support parameter files.
2432 [Victor B. Wagner <vitus@cryptocom.ru>]
2434 *) Support GeneralizedTime in ca utility.
2435 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2437 *) Enhance the hash format used for certificate directory links. The new
2438 form uses the canonical encoding (meaning equivalent names will work
2439 even if they aren't identical) and uses SHA1 instead of MD5. This form
2440 is incompatible with the older format and as a result c_rehash should
2441 be used to rebuild symbolic links.
2444 *) Make PKCS#8 the default write format for private keys, replacing the
2445 traditional format. This form is standardised, more secure and doesn't
2446 include an implicit MD5 dependency.
2449 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2450 committed to OpenSSL should pass this lot as a minimum.
2453 *) Add session ticket override functionality for use by EAP-FAST.
2454 [Jouni Malinen <j@w1.fi>]
2456 *) Modify HMAC functions to return a value. Since these can be implemented
2457 in an ENGINE errors can occur.
2460 *) Type-checked OBJ_bsearch_ex.
2463 *) Type-checked OBJ_bsearch. Also some constification necessitated
2464 by type-checking. Still to come: TXT_DB, bsearch(?),
2465 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2469 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2470 seconds to a tm structure directly, instead of going through OS
2471 specific date routines. This avoids any issues with OS routines such
2472 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2473 and X509_time_adj_ex() to cover the extended range. The existing
2474 X509_time_adj() is still usable and will no longer have any date issues.
2477 *) Delta CRL support. New use deltas option which will attempt to locate
2478 and search any appropriate delta CRLs available.
2480 This work was sponsored by Google.
2483 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2484 code and add additional score elements. Validate alternate CRL paths
2485 as part of the CRL checking and indicate a new error "CRL path validation
2486 error" in this case. Applications wanting additional details can use
2487 the verify callback and check the new "parent" field. If this is not
2488 NULL CRL path validation is taking place. Existing applications wont
2489 see this because it requires extended CRL support which is off by
2492 This work was sponsored by Google.
2495 *) Support for freshest CRL extension.
2497 This work was sponsored by Google.
2500 *) Initial indirect CRL support. Currently only supported in the CRLs
2501 passed directly and not via lookup. Process certificate issuer
2502 CRL entry extension and lookup CRL entries by bother issuer name
2503 and serial number. Check and process CRL issuer entry in IDP extension.
2505 This work was sponsored by Google.
2508 *) Add support for distinct certificate and CRL paths. The CRL issuer
2509 certificate is validated separately in this case. Only enabled if
2510 an extended CRL support flag is set: this flag will enable additional
2511 CRL functionality in future.
2513 This work was sponsored by Google.
2516 *) Add support for policy mappings extension.
2518 This work was sponsored by Google.
2521 *) Fixes to pathlength constraint, self issued certificate handling,
2522 policy processing to align with RFC3280 and PKITS tests.
2524 This work was sponsored by Google.
2527 *) Support for name constraints certificate extension. DN, email, DNS
2528 and URI types are currently supported.
2530 This work was sponsored by Google.
2533 *) To cater for systems that provide a pointer-based thread ID rather
2534 than numeric, deprecate the current numeric thread ID mechanism and
2535 replace it with a structure and associated callback type. This
2536 mechanism allows a numeric "hash" to be extracted from a thread ID in
2537 either case, and on platforms where pointers are larger than 'long',
2538 mixing is done to help ensure the numeric 'hash' is usable even if it
2539 can't be guaranteed unique. The default mechanism is to use "&errno"
2540 as a pointer-based thread ID to distinguish between threads.
2542 Applications that want to provide their own thread IDs should now use
2543 CRYPTO_THREADID_set_callback() to register a callback that will call
2544 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2546 Note that ERR_remove_state() is now deprecated, because it is tied
2547 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2548 to free the current thread's error state should be replaced by
2549 ERR_remove_thread_state(NULL).
2551 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2552 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2553 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2554 application was previously providing a numeric thread callback that
2555 was inappropriate for distinguishing threads, then uniqueness might
2556 have been obtained with &errno that happened immediately in the
2557 intermediate development versions of OpenSSL; this is no longer the
2558 case, the numeric thread callback will now override the automatic use
2560 [Geoff Thorpe, with help from Bodo Moeller]
2562 *) Initial support for different CRL issuing certificates. This covers a
2563 simple case where the self issued certificates in the chain exist and
2564 the real CRL issuer is higher in the existing chain.
2566 This work was sponsored by Google.
2569 *) Removed effectively defunct crypto/store from the build.
2572 *) Revamp of STACK to provide stronger type-checking. Still to come:
2573 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2574 ASN1_STRING, CONF_VALUE.
2577 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2578 RAM on SSL connections. This option can save about 34k per idle SSL.
2581 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2582 STACK, TXT_DB, bsearch, qsort.
2585 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2586 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2587 support for data, signedData, compressedData, digestedData and
2588 encryptedData, envelopedData types included. Scripts to check against
2589 RFC4134 examples draft and interop and consistency checks of many
2590 content types and variants.
2593 *) Add options to enc utility to support use of zlib compression BIO.
2596 *) Extend mk1mf to support importing of options and assembly language
2597 files from Configure script, currently only included in VC-WIN32.
2598 The assembly language rules can now optionally generate the source
2599 files from the associated perl scripts.
2602 *) Implement remaining functionality needed to support GOST ciphersuites.
2603 Interop testing has been performed using CryptoPro implementations.
2604 [Victor B. Wagner <vitus@cryptocom.ru>]
2606 *) s390x assembler pack.
2609 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2613 *) Implement Opaque PRF Input TLS extension as specified in
2614 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2615 official specification yet and no extension type assignment by
2616 IANA exists, this extension (for now) will have to be explicitly
2617 enabled when building OpenSSL by providing the extension number
2618 to use. For example, specify an option
2620 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2622 to the "config" or "Configure" script to enable the extension,
2623 assuming extension number 0x9527 (which is a completely arbitrary
2624 and unofficial assignment based on the MD5 hash of the Internet
2625 Draft). Note that by doing so, you potentially lose
2626 interoperability with other TLS implementations since these might
2627 be using the same extension number for other purposes.
2629 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2630 opaque PRF input value to use in the handshake. This will create
2631 an interal copy of the length-'len' string at 'src', and will
2632 return non-zero for success.
2634 To get more control and flexibility, provide a callback function
2637 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2638 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2642 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2645 Callback function 'cb' will be called in handshakes, and is
2646 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2647 Argument 'arg' is for application purposes (the value as given to
2648 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2649 be provided to the callback function). The callback function
2650 has to return non-zero to report success: usually 1 to use opaque
2651 PRF input just if possible, or 2 to enforce use of the opaque PRF
2652 input. In the latter case, the library will abort the handshake
2653 if opaque PRF input is not successfully negotiated.
2655 Arguments 'peerinput' and 'len' given to the callback function
2656 will always be NULL and 0 in the case of a client. A server will
2657 see the client's opaque PRF input through these variables if
2658 available (NULL and 0 otherwise). Note that if the server
2659 provides an opaque PRF input, the length must be the same as the
2660 length of the client's opaque PRF input.
2662 Note that the callback function will only be called when creating
2663 a new session (session resumption can resume whatever was
2664 previously negotiated), and will not be called in SSL 2.0
2665 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2666 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2667 for applications that need to enforce opaque PRF input.
2671 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2674 [Victor B. Wagner <vitus@cryptocom.ru>]
2676 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2677 RFC4507bis. The encrypted ticket format is an encrypted encoded
2678 SSL_SESSION structure, that way new session features are automatically
2681 If a client application caches session in an SSL_SESSION structure
2682 support is transparent because tickets are now stored in the encoded
2685 The SSL_CTX structure automatically generates keys for ticket
2686 protection in servers so again support should be possible
2687 with no application modification.
2689 If a client or server wishes to disable RFC4507 support then the option
2690 SSL_OP_NO_TICKET can be set.
2692 Add a TLS extension debugging callback to allow the contents of any client
2693 or server extensions to be examined.
2695 This work was sponsored by Google.
2698 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2699 OpenSSL should now compile cleanly on gcc 4.2
2700 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2702 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2703 support including streaming MAC support: this is required for GOST
2704 ciphersuite support.
2705 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2707 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2708 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2709 to output in BER and PEM format.
2712 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2713 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2714 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2715 ENGINE support for HMAC keys which are unextractable. New -mac and
2716 -macopt options to dgst utility.
2719 *) New option -sigopt to dgst utility. Update dgst to use
2720 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2721 alternative signing paramaters such as X9.31 or PSS in the dgst
2725 *) Change ssl_cipher_apply_rule(), the internal function that does
2726 the work each time a ciphersuite string requests enabling
2727 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2728 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2729 the order of disabled ciphersuites such that those ciphersuites
2730 that most recently went from enabled to disabled not only stay
2731 in order with respect to each other, but also have higher priority
2732 than other disabled ciphersuites the next time ciphersuites are
2735 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2736 the same ciphersuites as with "HIGH" alone, but in a specific
2737 order where the PSK ciphersuites come first (since they are the
2738 most recently disabled ciphersuites when "HIGH" is parsed).
2740 Also, change ssl_create_cipher_list() (using this new
2741 funcionality) such that between otherwise identical
2742 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2746 *) Change ssl_create_cipher_list() so that it automatically
2747 arranges the ciphersuites in reasonable order before starting
2748 to process the rule string. Thus, the definition for "DEFAULT"
2749 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2750 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2751 This makes it much easier to arrive at a reasonable default order
2752 in applications for which anonymous ciphers are OK (meaning
2753 that you can't actually use DEFAULT).
2754 [Bodo Moeller; suggested by Victor Duchovni]
2756 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2757 processing) into multiple integers instead of setting
2758 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2759 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2760 (These masks as well as the individual bit definitions are hidden
2761 away into the non-exported interface ssl/ssl_locl.h, so this
2762 change to the definition of the SSL_CIPHER structure shouldn't
2763 affect applications.) This give us more bits for each of these
2764 categories, so there is no longer a need to coagulate AES128 and
2765 AES256 into a single algorithm bit, and to coagulate Camellia128
2766 and Camellia256 into a single algorithm bit, which has led to all
2769 Thus, among other things, the kludge introduced in 0.9.7m and
2770 0.9.8e for masking out AES256 independently of AES128 or masking
2771 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2773 With the change, we also introduce new ciphersuite aliases that
2774 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2778 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2779 Use the leftmost N bytes of the signature input if the input is
2780 larger than the prime q (with N being the size in bytes of q).
2783 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2784 it yet and it is largely untested.
2787 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2790 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2791 some compilers (gcc 4.2 and later) reject their use. Safestack is
2792 reimplemented. Update ASN1 to avoid use of legacy functions.
2795 *) Win32/64 targets are linked with Winsock2.
2798 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2799 to external functions. This can be used to increase CRL handling
2800 efficiency especially when CRLs are very large by (for example) storing
2801 the CRL revoked certificates in a database.
2804 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2805 new CRLs added to a directory can be used. New command line option
2806 -verify_return_error to s_client and s_server. This causes real errors
2807 to be returned by the verify callback instead of carrying on no matter
2808 what. This reflects the way a "real world" verify callback would behave.
2811 *) GOST engine, supporting several GOST algorithms and public key formats.
2812 Kindly donated by Cryptocom.
2815 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2816 partitioned by DP are handled but no indirect CRL or reason partitioning
2817 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2818 selected via a scoring technique which handles IDP and AKID in CRLs.
2821 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2822 will ultimately be used for all verify operations: this will remove the
2823 X509_STORE dependency on certificate verification and allow alternative
2824 lookup methods. X509_STORE based implementations of these two callbacks.
2827 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2828 Modify get_crl() to find a valid (unexpired) CRL if possible.
2831 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2832 this would be called X509_CRL_cmp() but that name is already used by
2833 a function that just compares CRL issuer names. Cache several CRL
2834 extensions in X509_CRL structure and cache CRLDP in X509.
2837 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2838 this maps equivalent X509_NAME structures into a consistent structure.
2839 Name comparison can then be performed rapidly using memcmp().
2842 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2846 *) Allow digests to supply their own micalg string for S/MIME type using
2847 the ctrl EVP_MD_CTRL_MICALG.
2850 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2851 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2852 ctrl. It can then customise the structure before and/or after signing
2856 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2857 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2858 to free up any added signature OIDs.
2861 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2862 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2863 digest and cipher tables. New options added to openssl utility:
2864 list-message-digest-algorithms and list-cipher-algorithms.
2867 *) Change the array representation of binary polynomials: the list
2868 of degrees of non-zero coefficients is now terminated with -1.
2869 Previously it was terminated with 0, which was also part of the
2870 value; thus, the array representation was not applicable to
2871 polynomials where t^0 has coefficient zero. This change makes
2872 the array representation useful in a more general context.
2875 *) Various modifications and fixes to SSL/TLS cipher string
2876 handling. For ECC, the code now distinguishes between fixed ECDH
2877 with RSA certificates on the one hand and with ECDSA certificates
2878 on the other hand, since these are separate ciphersuites. The
2879 unused code for Fortezza ciphersuites has been removed.
2881 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2882 (not "ECDHE"). For consistency with the code for DH
2883 certificates, use of ECDH certificates is now considered ECDH
2884 authentication, not RSA or ECDSA authentication (the latter is
2885 merely the CA's signing algorithm and not actively used in the
2888 The temporary ciphersuite alias "ECCdraft" is no longer
2889 available, and ECC ciphersuites are no longer excluded from "ALL"
2890 and "DEFAULT". The following aliases now exist for RFC 4492
2891 ciphersuites, most of these by analogy with the DH case:
2893 kECDHr - ECDH cert, signed with RSA
2894 kECDHe - ECDH cert, signed with ECDSA
2895 kECDH - ECDH cert (signed with either RSA or ECDSA)
2896 kEECDH - ephemeral ECDH
2897 ECDH - ECDH cert or ephemeral ECDH
2903 AECDH - anonymous ECDH
2904 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2908 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2909 Use correct micalg parameters depending on digest(s) in signed message.
2912 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2913 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2916 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2917 an engine to register a method. Add ENGINE lookups for methods and
2918 functional reference processing.
2921 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2922 EVP_{Sign,Verify}* which allow an application to customise the signature
2926 *) New -resign option to smime utility. This adds one or more signers
2927 to an existing PKCS#7 signedData structure. Also -md option to use an
2928 alternative message digest algorithm for signing.
2931 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2932 create PKCS7 structures containing multiple signers. Update smime
2933 application to support multiple signers.
2936 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2940 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2941 Reorganize PBE internals to lookup from a static table using NIDs,
2942 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2943 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2944 PRF which will be automatically used with PBES2.
2947 *) Replace the algorithm specific calls to generate keys in "req" with the
2951 *) Update PKCS#7 enveloped data routines to use new API. This is now
2952 supported by any public key method supporting the encrypt operation. A
2953 ctrl is added to allow the public key algorithm to examine or modify
2954 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2958 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2959 a default digest type to use. In most cases this will be SHA1 but some
2960 algorithms (such as GOST) need to specify an alternative digest. The
2961 return value indicates how strong the preference is 1 means optional and
2962 2 is mandatory (that is it is the only supported type). Modify
2963 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2964 use the default md. Update openssl utilities to use the default digest
2965 type for signing if it is not explicitly indicated.
2968 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2969 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2970 signing method from the key type. This effectively removes the link
2971 between digests and public key types.
2974 *) Add an OID cross reference table and utility functions. Its purpose is to
2975 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2976 rsaEncryption. This will allow some of the algorithm specific hackery
2977 needed to use the correct OID to be removed.
2980 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2981 structures for PKCS7_sign(). They are now set up by the relevant public
2985 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2988 *) Add support for key derivation (agreement) in the API, DH method and
2992 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2993 public and private key formats. As a side effect these add additional
2994 command line functionality not previously available: DSA signatures can be
2995 generated and verified using pkeyutl and DH key support and generation in
3000 [Oliver Tappe <zooey@hirschkaefer.de>]
3002 *) New make target "install_html_docs" installs HTML renditions of the
3004 [Oliver Tappe <zooey@hirschkaefer.de>]
3006 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3007 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3008 support key and parameter generation and add initial key generation
3009 functionality for RSA.
3012 *) Add functions for main EVP_PKEY_method operations. The undocumented
3013 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3014 EVP_PKEY_{encrypt,decrypt}_old.
3017 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3018 key API, doesn't do much yet.
3021 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3022 public key algorithms. New option to openssl utility:
3023 "list-public-key-algorithms" to print out info.
3026 *) Implement the Supported Elliptic Curves Extension for
3027 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3030 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3031 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3034 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3035 utilities such as rsa, dsa, dsaparam etc except they process any key
3039 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3040 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3041 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3045 *) Initial support for pluggable public key ASN1.
3046 De-spaghettify the public key ASN1 handling. Move public and private
3047 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3048 algorithm specific handling to a single module within the relevant
3049 algorithm directory. Add functions to allow (near) opaque processing
3050 of public and private key structures.
3053 *) Implement the Supported Point Formats Extension for
3054 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3057 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3058 for the psk identity [hint] and the psk callback functions to the
3059 SSL_SESSION, SSL and SSL_CTX structure.
3062 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3066 SSL_CTX_use_psk_identity_hint
3067 SSL_get_psk_identity_hint
3068 SSL_get_psk_identity
3069 SSL_use_psk_identity_hint
3071 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3073 *) Add RFC 3161 compliant time stamp request creation, response generation
3074 and response verification functionality.
3075 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3077 *) Add initial support for TLS extensions, specifically for the server_name
3078 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3079 have new members for a host name. The SSL data structure has an
3080 additional member SSL_CTX *initial_ctx so that new sessions can be
3081 stored in that context to allow for session resumption, even after the
3082 SSL has been switched to a new SSL_CTX in reaction to a client's
3083 server_name extension.
3085 New functions (subject to change):
3087 SSL_get_servername()
3088 SSL_get_servername_type()
3091 New CTRL codes and macros (subject to change):
3093 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3094 - SSL_CTX_set_tlsext_servername_callback()
3095 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3096 - SSL_CTX_set_tlsext_servername_arg()
3097 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3099 openssl s_client has a new '-servername ...' option.
3101 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3102 '-key2 ...', '-servername_fatal' (subject to change). This allows
3103 testing the HostName extension for a specific single host name ('-cert'
3104 and '-key' remain fallbacks for handshakes without HostName
3105 negotiation). If the unrecognized_name alert has to be sent, this by
3106 default is a warning; it becomes fatal with the '-servername_fatal'
3109 [Peter Sylvester, Remy Allais, Christophe Renou]
3111 *) Whirlpool hash implementation is added.
3114 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3115 bn(64,32). Because of instruction set limitations it doesn't have
3116 any negative impact on performance. This was done mostly in order
3117 to make it possible to share assembler modules, such as bn_mul_mont
3118 implementations, between 32- and 64-bit builds without hassle.
3121 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3122 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3126 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3127 dedicated Montgomery multiplication procedure, is introduced.
3128 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3129 "64-bit" performance on certain 32-bit targets.
3132 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3133 in SSL structures. New SSL ctrl to set maximum send fragment size.
3134 Save memory by seeting the I/O buffer sizes dynamically instead of
3135 using the maximum available value.
3138 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3139 in addition to the text details.
3142 *) Very, very preliminary EXPERIMENTAL support for printing of general
3143 ASN1 structures. This currently produces rather ugly output and doesn't
3144 handle several customised structures at all.
3147 *) Integrated support for PVK file format and some related formats such
3148 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3149 these in the 'rsa' and 'dsa' utilities.
3152 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3155 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3156 place for the (very old) "NETSCAPE" format certificates which are now
3157 handled using new ASN1 code equivalents.
3160 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3161 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3162 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3165 *) Modify CRL distribution points extension code to print out previously
3166 unsupported fields. Enhance extension setting code to allow setting of
3170 *) Add print and set support for Issuing Distribution Point CRL extension.
3173 *) Change 'Configure' script to enable Camellia by default.
3176 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3178 *) When rejecting SSL/TLS records due to an incorrect version number, never
3179 update s->server with a new major version number. As of
3180 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3181 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3182 the previous behavior could result in a read attempt at NULL when
3183 receiving specific incorrect SSL/TLS records once record payload
3184 protection is active. (CVE-2010-0740)
3185 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3187 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3188 could be crashed if the relevant tables were not present (e.g. chrooted).
3189 [Tomas Hoger <thoger@redhat.com>]
3191 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3193 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3194 [Martin Olsson, Neel Mehta]
3196 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3197 accommodate for stack sorting, always a write lock!).
3200 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3201 excessive delays in the RAND_poll(): over a minute. As a workaround
3202 include a time check in the inner Heap32Next loop too.
3205 *) The code that handled flushing of data in SSL/TLS originally used the
3206 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3207 the problem outlined in PR#1949. The fix suggested there however can
3208 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3209 of Apache). So instead simplify the code to flush unconditionally.
3210 This should be fine since flushing with no data to flush is a no op.
3213 *) Handle TLS versions 2.0 and later properly and correctly use the
3214 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3215 off ancient servers have a habit of sticking around for a while...
3218 *) Modify compression code so it frees up structures without using the
3219 ex_data callbacks. This works around a problem where some applications
3220 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3221 restarting) then use compression (e.g. SSL with compression) later.
3222 This results in significant per-connection memory leaks and
3223 has caused some security issues including CVE-2008-1678 and
3227 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3228 change when encrypting or decrypting.
3231 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3232 connect and renegotiate with servers which do not support RI.
3233 Until RI is more widely deployed this option is enabled by default.
3236 *) Add "missing" ssl ctrls to clear options and mode.
3239 *) If client attempts to renegotiate and doesn't support RI respond with
3240 a no_renegotiation alert as required by RFC5746. Some renegotiating
3241 TLS clients will continue a connection gracefully when they receive
3242 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3243 waiting for a server hello which it will never receive. Now we treat a
3244 received no_renegotiation alert as a fatal error. This is because
3245 applications requesting a renegotiation might well expect it to succeed
3246 and would have no code in place to handle the server denying it so the
3247 only safe thing to do is to terminate the connection.
3250 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3251 peer supports secure renegotiation and 0 otherwise. Print out peer
3252 renegotiation support in s_client/s_server.
3255 *) Replace the highly broken and deprecated SPKAC certification method with
3256 the updated NID creation version. This should correctly handle UTF8.
3259 *) Implement RFC5746. Re-enable renegotiation but require the extension
3260 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3261 turns out to be a bad idea. It has been replaced by
3262 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3263 SSL_CTX_set_options(). This is really not recommended unless you
3264 know what you are doing.
3265 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3267 *) Fixes to stateless session resumption handling. Use initial_ctx when
3268 issuing and attempting to decrypt tickets in case it has changed during
3269 servername handling. Use a non-zero length session ID when attempting
3270 stateless session resumption: this makes it possible to determine if
3271 a resumption has occurred immediately after receiving server hello
3272 (several places in OpenSSL subtly assume this) instead of later in
3276 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3277 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3278 fixes for a few places where the return code is not checked
3280 [Julia Lawall <julia@diku.dk>]
3282 *) Add --strict-warnings option to Configure script to include devteam
3283 warnings in other configurations.
3286 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3287 makes it possible to install openssl libraries in locations which
3288 have names other than "lib", for example "/usr/lib64" which some
3290 [Steve Henson, based on patch from Jeremy Utley]
3292 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3293 X690 8.9.12 and can produce some misleading textual output of OIDs.
3294 [Steve Henson, reported by Dan Kaminsky]
3296 *) Delete MD2 from algorithm tables. This follows the recommendation in
3297 several standards that it is not used in new applications due to
3298 several cryptographic weaknesses. For binary compatibility reasons
3299 the MD2 API is still compiled in by default.
3302 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3306 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3307 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3309 [Guenter <lists@gknw.net>]
3311 *) Fix the server certificate chain building code to use X509_verify_cert(),
3312 it used to have an ad-hoc builder which was unable to cope with anything
3313 other than a simple chain.
3314 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3316 *) Don't check self signed certificate signatures in X509_verify_cert()
3317 by default (a flag can override this): it just wastes time without
3318 adding any security. As a useful side effect self signed root CAs
3319 with non-FIPS digests are now usable in FIPS mode.
3322 *) In dtls1_process_out_of_seq_message() the check if the current message
3323 is already buffered was missing. For every new message was memory
3324 allocated, allowing an attacker to perform an denial of service attack
3325 with sending out of seq handshake messages until there is no memory
3326 left. Additionally every future messege was buffered, even if the
3327 sequence number made no sense and would be part of another handshake.
3328 So only messages with sequence numbers less than 10 in advance will be
3329 buffered. (CVE-2009-1378)
3330 [Robin Seggelmann, discovered by Daniel Mentz]
3332 *) Records are buffered if they arrive with a future epoch to be
3333 processed after finishing the corresponding handshake. There is
3334 currently no limitation to this buffer allowing an attacker to perform
3335 a DOS attack with sending records with future epochs until there is no
3336 memory left. This patch adds the pqueue_size() function to determine
3337 the size of a buffer and limits the record buffer to 100 entries.
3339 [Robin Seggelmann, discovered by Daniel Mentz]
3341 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3342 parent structure is freed. (CVE-2009-1379)
3345 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3346 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3349 [Ilya O. <vrghost@gmail.com>]
3351 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3353 *) Disable renegotiation completely - this fixes a severe security
3354 problem (CVE-2009-3555) at the cost of breaking all
3355 renegotiation. Renegotiation can be re-enabled by setting
3356 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3357 run-time. This is really not recommended unless you know what
3361 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3363 *) Don't set val to NULL when freeing up structures, it is freed up by
3364 underlying code. If sizeof(void *) > sizeof(long) this can result in
3365 zeroing past the valid field. (CVE-2009-0789)
3366 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3368 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3369 checked correctly. This would allow some invalid signed attributes to
3370 appear to verify correctly. (CVE-2009-0591)
3371 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3373 *) Reject UniversalString and BMPString types with invalid lengths. This
3374 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3375 a legal length. (CVE-2009-0590)
3378 *) Set S/MIME signing as the default purpose rather than setting it
3379 unconditionally. This allows applications to override it at the store
3383 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3384 to handle some structures.
3387 *) Improve efficiency of mem_gets: don't search whole buffer each time
3389 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3391 *) New -hex option for openssl rand.
3394 *) Print out UTF8String and NumericString when parsing ASN1.
3397 *) Support NumericString type for name components.
3400 *) Allow CC in the environment to override the automatically chosen
3401 compiler. Note that nothing is done to ensure flags work with the
3405 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3407 *) Properly check EVP_VerifyFinal() and similar return values
3409 [Ben Laurie, Bodo Moeller, Google Security Team]
3411 *) Enable TLS extensions by default.
3414 *) Allow the CHIL engine to be loaded, whether the application is
3415 multithreaded or not. (This does not release the developer from the
3416 obligation to set up the dynamic locking callbacks.)
3417 [Sander Temme <sander@temme.net>]
3419 *) Use correct exit code if there is an error in dgst command.
3420 [Steve Henson; problem pointed out by Roland Dirlewanger]
3422 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3423 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3426 *) Add experimental JPAKE support, including demo authentication in
3427 s_client and s_server.
3430 *) Set the comparison function in v3_addr_canonize().
3431 [Rob Austein <sra@hactrn.net>]
3433 *) Add support for XMPP STARTTLS in s_client.
3434 [Philip Paeps <philip@freebsd.org>]
3436 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3437 to ensure that even with this option, only ciphersuites in the
3438 server's preference list will be accepted. (Note that the option
3439 applies only when resuming a session, so the earlier behavior was
3440 just about the algorithm choice for symmetric cryptography.)
3443 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3445 *) Fix NULL pointer dereference if a DTLS server received
3446 ChangeCipherSpec as first record (CVE-2009-1386).
3449 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3450 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3453 *) The fix in 0.9.8c that supposedly got rid of unsafe
3454 double-checked locking was incomplete for RSA blinding,
3455 addressing just one layer of what turns out to have been
3456 doubly unsafe triple-checked locking.
3458 So now fix this for real by retiring the MONT_HELPER macro
3459 in crypto/rsa/rsa_eay.c.
3461 [Bodo Moeller; problem pointed out by Marius Schilder]
3463 *) Various precautionary measures:
3465 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3467 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3468 (NB: This would require knowledge of the secret session ticket key
3469 to exploit, in which case you'd be SOL either way.)
3471 - Change bn_nist.c so that it will properly handle input BIGNUMs
3472 outside the expected range.
3474 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3477 [Neel Mehta, Bodo Moeller]
3479 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3480 the load fails. Useful for distros.
3481 [Ben Laurie and the FreeBSD team]
3483 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3486 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3489 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3491 This work was sponsored by Logica.
3494 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3495 keystores. Support for SSL/TLS client authentication too.
3496 Not compiled unless enable-capieng specified to Configure.
3498 This work was sponsored by Logica.
3501 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3502 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3503 attribute creation routines such as certifcate requests and PKCS#12
3507 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3509 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3510 handshake which could lead to a cilent crash as found using the
3511 Codenomicon TLS test suite (CVE-2008-1672)
3512 [Steve Henson, Mark Cox]
3514 *) Fix double free in TLS server name extensions which could lead to
3515 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3518 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3520 Clear the error queue to ensure that error entries left from
3521 older function calls do not interfere with the correct operation.
3522 [Lutz Jaenicke, Erik de Castro Lopo]
3524 *) Remove root CA certificates of commercial CAs:
3526 The OpenSSL project does not recommend any specific CA and does not
3527 have any policy with respect to including or excluding any CA.
3528 Therefore it does not make any sense to ship an arbitrary selection
3529 of root CA certificates with the OpenSSL software.
3532 *) RSA OAEP patches to fix two separate invalid memory reads.
3533 The first one involves inputs when 'lzero' is greater than
3534 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3535 before the beginning of from). The second one involves inputs where
3536 the 'db' section contains nothing but zeroes (there is a one-byte
3537 invalid read after the end of 'db').
3538 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3540 *) Partial backport from 0.9.9-dev:
3542 Introduce bn_mul_mont (dedicated Montgomery multiplication
3543 procedure) as a candidate for BIGNUM assembler implementation.
3544 While 0.9.9-dev uses assembler for various architectures, only
3545 x86_64 is available by default here in the 0.9.8 branch, and
3546 32-bit x86 is available through a compile-time setting.
3548 To try the 32-bit x86 assembler implementation, use Configure
3549 option "enable-montasm" (which exists only for this backport).
3551 As "enable-montasm" for 32-bit x86 disclaims code stability
3552 anyway, in this constellation we activate additional code
3553 backported from 0.9.9-dev for further performance improvements,
3554 namely BN_from_montgomery_word. (To enable this otherwise,
3555 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3557 [Andy Polyakov (backport partially by Bodo Moeller)]
3559 *) Add TLS session ticket callback. This allows an application to set
3560 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3561 values. This is useful for key rollover for example where several key
3562 sets may exist with different names.
3565 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3566 This was broken until now in 0.9.8 releases, such that the only way
3567 a registered ENGINE could be used (assuming it initialises
3568 successfully on the host) was to explicitly set it as the default
3569 for the relevant algorithms. This is in contradiction with 0.9.7
3570 behaviour and the documentation. With this fix, when an ENGINE is
3571 registered into a given algorithm's table of implementations, the
3572 'uptodate' flag is reset so that auto-discovery will be used next
3573 time a new context for that algorithm attempts to select an
3575 [Ian Lister (tweaked by Geoff Thorpe)]
3577 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3578 implemention in the following ways:
3580 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3583 Lack of BER streaming support means one pass streaming processing is
3584 only supported if data is detached: setting the streaming flag is
3585 ignored for embedded content.
3587 CMS support is disabled by default and must be explicitly enabled
3588 with the enable-cms configuration option.
3591 *) Update the GMP engine glue to do direct copies between BIGNUM and
3592 mpz_t when openssl and GMP use the same limb size. Otherwise the
3593 existing "conversion via a text string export" trick is still used.
3594 [Paul Sheer <paulsheer@gmail.com>]
3596 *) Zlib compression BIO. This is a filter BIO which compressed and
3597 uncompresses any data passed through it.
3600 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3601 RFC3394 compatible AES key wrapping.
3604 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3605 sets string data without copying. X509_ALGOR_set0() and
3606 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3607 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3608 from an X509_ATTRIBUTE structure optionally checking it occurs only
3609 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3613 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3614 to get the expected BN_FLG_CONSTTIME behavior.
3615 [Bodo Moeller (Google)]
3619 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3620 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3621 - added some more tests to do_tests.pl
3622 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3623 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3624 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3625 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3626 - various changes to netware.pl to enable gcc-cross builds on Win32
3628 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3629 - various changes to fix missing prototype warnings
3630 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3631 - added AES, WHIRLPOOL and CPUID assembler code to build files
3632 - added missing AES assembler make rules to mk1mf.pl
3633 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3634 [Guenter Knauf <eflash@gmx.net>]
3636 *) Implement certificate status request TLS extension defined in RFC3546.
3637 A client can set the appropriate parameters and receive the encoded
3638 OCSP response via a callback. A server can query the supplied parameters
3639 and set the encoded OCSP response in the callback. Add simplified examples
3640 to s_client and s_server.
3643 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3645 *) Fix various bugs:
3646 + Binary incompatibility of ssl_ctx_st structure
3647 + DTLS interoperation with non-compliant servers
3648 + Don't call get_session_cb() without proposed session
3649 + Fix ia64 assembler code
3650 [Andy Polyakov, Steve Henson]
3652 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3654 *) DTLS Handshake overhaul. There were longstanding issues with
3655 OpenSSL DTLS implementation, which were making it impossible for
3656 RFC 4347 compliant client to communicate with OpenSSL server.
3657 Unfortunately just fixing these incompatibilities would "cut off"
3658 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3659 server keeps tolerating non RFC compliant syntax. The opposite is
3660 not true, 0.9.8f client can not communicate with earlier server.
3661 This update even addresses CVE-2007-4995.
3664 *) Changes to avoid need for function casts in OpenSSL: some compilers
3665 (gcc 4.2 and later) reject their use.
3666 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3669 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3670 RFC4507bis. The encrypted ticket format is an encrypted encoded
3671 SSL_SESSION structure, that way new session features are automatically
3674 If a client application caches session in an SSL_SESSION structure
3675 support is transparent because tickets are now stored in the encoded
3678 The SSL_CTX structure automatically generates keys for ticket
3679 protection in servers so again support should be possible
3680 with no application modification.
3682 If a client or server wishes to disable RFC4507 support then the option
3683 SSL_OP_NO_TICKET can be set.
3685 Add a TLS extension debugging callback to allow the contents of any client
3686 or server extensions to be examined.
3688 This work was sponsored by Google.
3691 *) Add initial support for TLS extensions, specifically for the server_name
3692 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3693 have new members for a host name. The SSL data structure has an
3694 additional member SSL_CTX *initial_ctx so that new sessions can be
3695 stored in that context to allow for session resumption, even after the
3696 SSL has been switched to a new SSL_CTX in reaction to a client's
3697 server_name extension.
3699 New functions (subject to change):
3701 SSL_get_servername()
3702 SSL_get_servername_type()
3705 New CTRL codes and macros (subject to change):
3707 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3708 - SSL_CTX_set_tlsext_servername_callback()
3709 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3710 - SSL_CTX_set_tlsext_servername_arg()
3711 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3713 openssl s_client has a new '-servername ...' option.
3715 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3716 '-key2 ...', '-servername_fatal' (subject to change). This allows
3717 testing the HostName extension for a specific single host name ('-cert'
3718 and '-key' remain fallbacks for handshakes without HostName
3719 negotiation). If the unrecognized_name alert has to be sent, this by
3720 default is a warning; it becomes fatal with the '-servername_fatal'
3723 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3725 *) Add AES and SSE2 assembly language support to VC++ build.
3728 *) Mitigate attack on final subtraction in Montgomery reduction.
3731 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3732 (which previously caused an internal error).
3735 *) Squeeze another 10% out of IGE mode when in != out.
3738 *) AES IGE mode speedup.
3739 [Dean Gaudet (Google)]
3741 *) Add the Korean symmetric 128-bit cipher SEED (see
3742 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3743 add SEED ciphersuites from RFC 4162:
3745 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3746 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3747 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3748 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3750 To minimize changes between patchlevels in the OpenSSL 0.9.8
3751 series, SEED remains excluded from compilation unless OpenSSL
3752 is configured with 'enable-seed'.
3753 [KISA, Bodo Moeller]
3755 *) Mitigate branch prediction attacks, which can be practical if a
3756 single processor is shared, allowing a spy process to extract
3757 information. For detailed background information, see
3758 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3759 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3760 and Necessary Software Countermeasures"). The core of the change
3761 are new versions BN_div_no_branch() and
3762 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3763 respectively, which are slower, but avoid the security-relevant
3764 conditional branches. These are automatically called by BN_div()
3765 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3766 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3767 remove a conditional branch.
3769 BN_FLG_CONSTTIME is the new name for the previous
3770 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3771 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3772 in the exponent causes BN_mod_exp_mont() to use the alternative
3773 implementation in BN_mod_exp_mont_consttime().) The old name
3774 remains as a deprecated alias.
3776 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3777 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3778 constant-time implementations for more than just exponentiation.
3779 Here too the old name is kept as a deprecated alias.
3781 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3782 the BN_BLINDING structure gets an independent copy of the
3783 modulus. This means that the previous "BIGNUM *m" argument to
3784 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3785 essentially becomes "const BIGNUM *m", although we can't actually
3786 change this in the header file before 0.9.9. It allows
3787 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3788 enable BN_FLG_CONSTTIME.
3790 [Matthew D Wood (Intel Corp)]
3792 *) In the SSL/TLS server implementation, be strict about session ID
3793 context matching (which matters if an application uses a single
3794 external cache for different purposes). Previously,
3795 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3796 set. This did ensure strict client verification, but meant that,
3797 with applications using a single external cache for quite
3798 different requirements, clients could circumvent ciphersuite
3799 restrictions for a given session ID context by starting a session
3800 in a different context.
3803 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3804 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3805 authentication-only ciphersuites.
3808 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3809 not complete and could lead to a possible single byte overflow
3810 (CVE-2007-5135) [Ben Laurie]
3812 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3814 *) Since AES128 and AES256 (and similarly Camellia128 and
3815 Camellia256) share a single mask bit in the logic of
3816 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3817 kludge to work properly if AES128 is available and AES256 isn't
3818 (or if Camellia128 is available and Camellia256 isn't).
3821 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3822 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3823 When a point or a seed is encoded in a BIT STRING, we need to
3824 prevent the removal of trailing zero bits to get the proper DER
3825 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3826 of a NamedBitList, for which trailing 0 bits need to be removed.)
3829 *) Have SSL/TLS server implementation tolerate "mismatched" record
3830 protocol version while receiving ClientHello even if the
3831 ClientHello is fragmented. (The server can't insist on the
3832 particular protocol version it has chosen before the ServerHello
3833 message has informed the client about his choice.)
3836 *) Add RFC 3779 support.
3837 [Rob Austein for ARIN, Ben Laurie]
3839 *) Load error codes if they are not already present instead of using a
3840 static variable. This allows them to be cleanly unloaded and reloaded.
3841 Improve header file function name parsing.
3844 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3845 or CAPABILITY handshake as required by RFCs.
3848 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3850 *) Introduce limits to prevent malicious keys being able to
3851 cause a denial of service. (CVE-2006-2940)
3852 [Steve Henson, Bodo Moeller]
3854 *) Fix ASN.1 parsing of certain invalid structures that can result
3855 in a denial of service. (CVE-2006-2937) [Steve Henson]
3857 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3858 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3860 *) Fix SSL client code which could crash if connecting to a
3861 malicious SSLv2 server. (CVE-2006-4343)
3862 [Tavis Ormandy and Will Drewry, Google Security Team]
3864 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3865 match only those. Before that, "AES256-SHA" would be interpreted
3866 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3867 the same strength classification in 0.9.7h) as we currently only
3868 have a single AES bit in the ciphersuite description bitmap.
3869 That change, however, also applied to ciphersuite strings such as
3870 "RC4-MD5" that intentionally matched multiple ciphersuites --
3871 namely, SSL 2.0 ciphersuites in addition to the more common ones
3872 from SSL 3.0/TLS 1.0.
3874 So we change the selection algorithm again: Naming an explicit
3875 ciphersuite selects this one ciphersuite, and any other similar
3876 ciphersuite (same bitmap) from *other* protocol versions.
3877 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3878 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3880 Since SSL 2.0 does not have any ciphersuites for which the
3881 128/256 bit distinction would be relevant, this works for now.
3882 The proper fix will be to use different bits for AES128 and
3883 AES256, which would have avoided the problems from the beginning;
3884 however, bits are scarce, so we can only do this in a new release
3885 (not just a patchlevel) when we can change the SSL_CIPHER
3886 definition to split the single 'unsigned long mask' bitmap into
3887 multiple values to extend the available space.
3891 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3893 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3894 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3896 *) Add AES IGE and biIGE modes.
3899 *) Change the Unix randomness entropy gathering to use poll() when
3900 possible instead of select(), since the latter has some
3901 undesirable limitations.
3902 [Darryl Miles via Richard Levitte and Bodo Moeller]
3904 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3905 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3906 cannot be implicitly activated as part of, e.g., the "AES" alias.
3907 However, please upgrade to OpenSSL 0.9.9[-dev] for
3908 non-experimental use of the ECC ciphersuites to get TLS extension
3909 support, which is required for curve and point format negotiation
3910 to avoid potential handshake problems.
3913 *) Disable rogue ciphersuites:
3915 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3916 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3917 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3919 The latter two were purportedly from
3920 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3923 Also deactivate the remaining ciphersuites from
3924 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3925 unofficial, and the ID has long expired.
3928 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3929 dual-core machines) and other potential thread-safety issues.
3932 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3933 versions), which is now available for royalty-free use
3934 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3935 Also, add Camellia TLS ciphersuites from RFC 4132.
3937 To minimize changes between patchlevels in the OpenSSL 0.9.8
3938 series, Camellia remains excluded from compilation unless OpenSSL
3939 is configured with 'enable-camellia'.
3942 *) Disable the padding bug check when compression is in use. The padding
3943 bug check assumes the first packet is of even length, this is not
3944 necessarily true if compresssion is enabled and can result in false
3945 positives causing handshake failure. The actual bug test is ancient
3946 code so it is hoped that implementations will either have fixed it by
3947 now or any which still have the bug do not support compression.
3950 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3952 *) When applying a cipher rule check to see if string match is an explicit
3953 cipher suite and only match that one cipher suite if it is.
3956 *) Link in manifests for VC++ if needed.
3957 [Austin Ziegler <halostatue@gmail.com>]
3959 *) Update support for ECC-based TLS ciphersuites according to
3960 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3961 TLS extensions, which are supported starting with the 0.9.9
3962 branch, not in the OpenSSL 0.9.8 branch).
3965 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3966 opaque EVP_CIPHER_CTX handling.
3969 *) Fixes and enhancements to zlib compression code. We now only use
3970 "zlib1.dll" and use the default __cdecl calling convention on Win32
3971 to conform with the standards mentioned here:
3972 http://www.zlib.net/DLL_FAQ.txt
3973 Static zlib linking now works on Windows and the new --with-zlib-include
3974 --with-zlib-lib options to Configure can be used to supply the location
3975 of the headers and library. Gracefully handle case where zlib library
3979 *) Several fixes and enhancements to the OID generation code. The old code
3980 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3981 handle numbers larger than ULONG_MAX, truncated printing and had a
3982 non standard OBJ_obj2txt() behaviour.
3985 *) Add support for building of engines under engine/ as shared libraries
3986 under VC++ build system.
3989 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3990 Hopefully, we will not see any false combination of paths any more.
3993 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3995 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3996 (part of SSL_OP_ALL). This option used to disable the
3997 countermeasure against man-in-the-middle protocol-version
3998 rollback in the SSL 2.0 server implementation, which is a bad
3999 idea. (CVE-2005-2969)
4001 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4002 for Information Security, National Institute of Advanced Industrial
4003 Science and Technology [AIST], Japan)]
4005 *) Add two function to clear and return the verify parameter flags.
4008 *) Keep cipherlists sorted in the source instead of sorting them at
4009 runtime, thus removing the need for a lock.
4012 *) Avoid some small subgroup attacks in Diffie-Hellman.
4013 [Nick Mathewson and Ben Laurie]
4015 *) Add functions for well-known primes.
4018 *) Extended Windows CE support.
4019 [Satoshi Nakamura and Andy Polyakov]
4021 *) Initialize SSL_METHOD structures at compile time instead of during
4022 runtime, thus removing the need for a lock.
4025 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4026 attempting to decrypt each encrypted key in turn. Add support to
4030 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4032 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4035 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4038 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4039 key into the same file any more.
4042 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4045 *) Add -utf8 command line and config file option to 'ca'.
4046 [Stefan <stf@udoma.org]
4048 *) Removed the macro des_crypt(), as it seems to conflict with some
4049 libraries. Use DES_crypt().
4052 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4053 involves renaming the source and generated shared-libs for
4054 both. The engines will accept the corrected or legacy ids
4055 ('ncipher' and '4758_cca' respectively) when binding. NB,
4056 this only applies when building 'shared'.
4057 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4059 *) Add attribute functions to EVP_PKEY structure. Modify
4060 PKCS12_create() to recognize a CSP name attribute and
4061 use it. Make -CSP option work again in pkcs12 utility.
4064 *) Add new functionality to the bn blinding code:
4065 - automatic re-creation of the BN_BLINDING parameters after
4066 a fixed number of uses (currently 32)
4067 - add new function for parameter creation
4068 - introduce flags to control the update behaviour of the
4069 BN_BLINDING parameters
4070 - hide BN_BLINDING structure
4071 Add a second BN_BLINDING slot to the RSA structure to improve
4072 performance when a single RSA object is shared among several
4076 *) Add support for DTLS.
4077 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4079 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4080 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4083 *) Remove buggy and incomplete DH cert support from
4084 ssl/ssl_rsa.c and ssl/s3_both.c
4087 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4088 the apps/openssl applications.
4091 *) Compile clean with "-Wall -Wmissing-prototypes
4092 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4093 DEBUG_SAFESTACK must also be set.
4096 *) Change ./Configure so that certain algorithms can be disabled by default.
4097 The new counterpiece to "no-xxx" is "enable-xxx".
4099 The patented RC5 and MDC2 algorithms will now be disabled unless
4100 "enable-rc5" and "enable-mdc2", respectively, are specified.
4102 (IDEA remains enabled despite being patented. This is because IDEA
4103 is frequently required for interoperability, and there is no license
4104 fee for non-commercial use. As before, "no-idea" can be used to
4105 avoid this algorithm.)
4109 *) Add processing of proxy certificates (see RFC 3820). This work was
4110 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4111 EGEE (Enabling Grids for E-science in Europe).
4114 *) RC4 performance overhaul on modern architectures/implementations, such
4115 as Intel P4, IA-64 and AMD64.
4118 *) New utility extract-section.pl. This can be used specify an alternative
4119 section number in a pod file instead of having to treat each file as
4120 a separate case in Makefile. This can be done by adding two lines to the
4123 =for comment openssl_section:XXX
4125 The blank line is mandatory.
4129 *) New arguments -certform, -keyform and -pass for s_client and s_server
4130 to allow alternative format key and certificate files and passphrase
4134 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4135 update associated structures and add various utility functions.
4137 Add new policy related verify parameters, include policy checking in
4138 standard verify code. Enhance 'smime' application with extra parameters
4139 to support policy checking and print out.
4142 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4143 Nehemiah processors. These extensions support AES encryption in hardware
4144 as well as RNG (though RNG support is currently disabled).
4145 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4147 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4150 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4151 [Andy Polyakov and a number of other people]
4153 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4154 implementation contributed by IBM.
4155 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4157 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4158 exponent rather than 'unsigned long'. There is a corresponding change to
4159 the new 'rsa_keygen' element of the RSA_METHOD structure.
4160 [Jelte Jansen, Geoff Thorpe]
4162 *) Functionality for creating the initial serial number file is now
4163 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4165 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4166 number file to 1, which is bound to cause problems. To avoid
4167 the problems while respecting compatibility between different 0.9.7
4168 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4169 CA.pl for serial number initialization. With the new release 0.9.8,
4170 we can fix the problem directly in the 'ca' utility.)
4173 *) Reduced header interdepencies by declaring more opaque objects in
4174 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4175 give fewer recursive includes, which could break lazy source code - so
4176 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4177 developers should define this symbol when building and using openssl to
4178 ensure they track the recommended behaviour, interfaces, [etc], but
4179 backwards-compatible behaviour prevails when this isn't defined.
4182 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4185 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4186 This will generate a random key of the appropriate length based on the
4187 cipher context. The EVP_CIPHER can provide its own random key generation
4188 routine to support keys of a specific form. This is used in the des and
4189 3des routines to generate a key of the correct parity. Update S/MIME
4190 code to use new functions and hence generate correct parity DES keys.
4191 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4192 valid (weak or incorrect parity).
4195 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4196 as looking them up. This is useful when the verified structure may contain
4197 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4198 present unless the new PKCS7_NO_CRL flag is asserted.
4201 *) Extend ASN1 oid configuration module. It now additionally accepts the
4204 shortName = some long name, 1.2.3.4
4207 *) Reimplemented the BN_CTX implementation. There is now no more static
4208 limitation on the number of variables it can handle nor the depth of the
4209 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4210 information can now expand as required, and rather than having a single
4211 static array of bignums, BN_CTX now uses a linked-list of such arrays
4212 allowing it to expand on demand whilst maintaining the usefulness of
4213 BN_CTX's "bundling".
4216 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4217 to allow all RSA operations to function using a single BN_CTX.
4220 *) Preliminary support for certificate policy evaluation and checking. This
4221 is initially intended to pass the tests outlined in "Conformance Testing
4222 of Relying Party Client Certificate Path Processing Logic" v1.07.
4225 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4226 remained unused and not that useful. A variety of other little bignum
4227 tweaks and fixes have also been made continuing on from the audit (see
4231 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4232 associated ASN1, EVP and SSL functions and old ASN1 macros.
4235 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4236 and this should never fail. So the return value from the use of
4237 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4238 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4241 *) BN_CTX_get() should return zero-valued bignums, providing the same
4242 initialised value as BN_new().
4243 [Geoff Thorpe, suggested by Ulf Möller]
4245 *) Support for inhibitAnyPolicy certificate extension.
4248 *) An audit of the BIGNUM code is underway, for which debugging code is
4249 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4250 is considered valid when processing BIGNUMs, and causes execution to
4251 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4252 further steps are taken to deliberately pollute unused data in BIGNUM
4253 structures to try and expose faulty code further on. For now, openssl will
4254 (in its default mode of operation) continue to tolerate the inconsistent
4255 forms that it has tolerated in the past, but authors and packagers should
4256 consider trying openssl and their own applications when compiled with
4257 these debugging symbols defined. It will help highlight potential bugs in
4258 their own code, and will improve the test coverage for OpenSSL itself. At
4259 some point, these tighter rules will become openssl's default to improve
4260 maintainability, though the assert()s and other overheads will remain only
4261 in debugging configurations. See bn.h for more details.
4262 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4264 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4265 that can only be obtained through BN_CTX_new() (which implicitly
4266 initialises it). The presence of this function only made it possible
4267 to overwrite an existing structure (and cause memory leaks).
4270 *) Because of the callback-based approach for implementing LHASH as a
4271 template type, lh_insert() adds opaque objects to hash-tables and
4272 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4273 to clean up those corresponding objects before destroying the hash table
4274 (and losing the object pointers). So some over-zealous constifications in
4275 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4276 objects as "const" and the lh_doall[_arg] callback wrappers are not
4277 prototyped to have "const" restrictions on the object pointers they are
4278 given (and so aren't required to cast them away any more).
4281 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4282 (speed) prefers to use its own implementation. The two implementations
4283 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4284 its object type properly exposed (MS_TM) instead of casting to/from "char
4285 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4286 aren't necessarily the greatest nomenclatures - but this is what was used
4287 internally to the implementation so I've used that for now.
4290 *) Ensure that deprecated functions do not get compiled when
4291 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4292 the self-tests were still using deprecated key-generation functions so
4293 these have been updated also.
4296 *) Reorganise PKCS#7 code to separate the digest location functionality
4297 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4298 New function PKCS7_set_digest() to set the digest type for PKCS#7
4299 digestedData type. Add additional code to correctly generate the
4300 digestedData type and add support for this type in PKCS7 initialization
4304 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4305 structure of type "other".
4308 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4309 sure the loop does correctly stop and breaking ("division by zero")
4310 modulus operations are not performed. The (pre-generated) prime
4311 table crypto/bn/bn_prime.h was already correct, but it could not be
4312 re-generated on some platforms because of the "division by zero"
4313 situation in the script.
4314 [Ralf S. Engelschall]
4316 *) Update support for ECC-based TLS ciphersuites according to
4317 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4318 SHA-1 now is only used for "small" curves (where the
4319 representation of a field element takes up to 24 bytes); for
4320 larger curves, the field element resulting from ECDH is directly
4321 used as premaster secret.
4322 [Douglas Stebila (Sun Microsystems Laboratories)]
4324 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4325 curve secp160r1 to the tests.
4326 [Douglas Stebila (Sun Microsystems Laboratories)]
4328 *) Add the possibility to load symbols globally with DSO.
4329 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4331 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4332 control of the error stack.
4335 *) Add support for STORE in ENGINE.
4338 *) Add the STORE type. The intention is to provide a common interface
4339 to certificate and key stores, be they simple file-based stores, or
4340 HSM-type store, or LDAP stores, or...
4341 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4344 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4345 pass a list of arguments to any function as well as provide a way
4346 for a function to pass data back to the caller.
4349 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4350 works like BUF_strdup() but can be used to duplicate a portion of
4351 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4355 *) Add the function sk_find_ex() which works like sk_find(), but will
4356 return an index to an element even if an exact match couldn't be
4357 found. The index is guaranteed to point at the element where the
4358 searched-for key would be inserted to preserve sorting order.
4361 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4362 takes an extra flags argument for optional functionality. Currently,
4363 the following flags are defined:
4365 OBJ_BSEARCH_VALUE_ON_NOMATCH
4366 This one gets OBJ_bsearch_ex() to return a pointer to the first
4367 element where the comparing function returns a negative or zero
4370 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4371 This one gets OBJ_bsearch_ex() to return a pointer to the first
4372 element where the comparing function returns zero. This is useful
4373 if there are more than one element where the comparing function
4377 *) Make it possible to create self-signed certificates with 'openssl ca'
4378 in such a way that the self-signed certificate becomes part of the
4379 CA database and uses the same mechanisms for serial number generation
4380 as all other certificate signing. The new flag '-selfsign' enables
4381 this functionality. Adapt CA.sh and CA.pl.in.
4384 *) Add functionality to check the public key of a certificate request
4385 against a given private. This is useful to check that a certificate
4386 request can be signed by that key (self-signing).
4389 *) Make it possible to have multiple active certificates with the same
4390 subject in the CA index file. This is done only if the keyword
4391 'unique_subject' is set to 'no' in the main CA section (default
4392 if 'CA_default') of the configuration file. The value is saved
4393 with the database itself in a separate index attribute file,
4394 named like the index file with '.attr' appended to the name.
4397 *) Generate muti valued AVAs using '+' notation in config files for
4401 *) Support for nameConstraints certificate extension.
4404 *) Support for policyConstraints certificate extension.
4407 *) Support for policyMappings certificate extension.
4410 *) Make sure the default DSA_METHOD implementation only uses its
4411 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4412 and change its own handlers to be NULL so as to remove unnecessary
4413 indirection. This lets alternative implementations fallback to the
4414 default implementation more easily.
4417 *) Support for directoryName in GeneralName related extensions
4421 *) Make it possible to link applications using Makefile.shared.
4422 Make that possible even when linking against static libraries!
4425 *) Support for single pass processing for S/MIME signing. This now
4426 means that S/MIME signing can be done from a pipe, in addition
4427 cleartext signing (multipart/signed type) is effectively streaming
4428 and the signed data does not need to be all held in memory.
4430 This is done with a new flag PKCS7_STREAM. When this flag is set
4431 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4432 is done after the data is output (and digests calculated) in
4433 SMIME_write_PKCS7().
4436 *) Add full support for -rpath/-R, both in shared libraries and
4437 applications, at least on the platforms where it's known how
4441 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4442 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4443 will now compute a table of multiples of the generator that
4444 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4445 faster (notably in the case of a single point multiplication,
4446 scalar * generator).
4447 [Nils Larsch, Bodo Moeller]
4449 *) IPv6 support for certificate extensions. The various extensions
4450 which use the IP:a.b.c.d can now take IPv6 addresses using the
4451 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4455 *) Added an ENGINE that implements RSA by performing private key
4456 exponentiations with the GMP library. The conversions to and from
4457 GMP's mpz_t format aren't optimised nor are any montgomery forms
4458 cached, and on x86 it appears OpenSSL's own performance has caught up.
4459 However there are likely to be other architectures where GMP could
4460 provide a boost. This ENGINE is not built in by default, but it can be
4461 specified at Configure time and should be accompanied by the necessary
4462 linker additions, eg;
4463 ./config -DOPENSSL_USE_GMP -lgmp
4466 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4467 testing availability of engines with "-t" - the old behaviour is
4468 produced by increasing the feature's verbosity with "-tt".
4471 *) ECDSA routines: under certain error conditions uninitialized BN objects
4472 could be freed. Solution: make sure initialization is performed early
4473 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4477 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4478 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4479 software implementations. For DSA and DH, parameter generation can
4480 also be overriden by providing the appropriate method callbacks.
4483 *) Change the "progress" mechanism used in key-generation and
4484 primality testing to functions that take a new BN_GENCB pointer in
4485 place of callback/argument pairs. The new API functions have "_ex"
4486 postfixes and the older functions are reimplemented as wrappers for
4487 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4488 declarations of the old functions to help (graceful) attempts to
4489 migrate to the new functions. Also, the new key-generation API
4490 functions operate on a caller-supplied key-structure and return
4491 success/failure rather than returning a key or NULL - this is to
4492 help make "keygen" another member function of RSA_METHOD etc.
4494 Example for using the new callback interface:
4496 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4500 BN_GENCB_set(&my_cb, my_callback, my_arg);
4502 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4503 /* For the meaning of a, b in calls to my_callback(), see the
4504 * documentation of the function that calls the callback.
4505 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4506 * my_callback should return 1 if it wants BN_is_prime_ex()
4507 * to continue, or 0 to stop.
4512 *) Change the ZLIB compression method to be stateful, and make it
4513 available to TLS with the number defined in
4514 draft-ietf-tls-compression-04.txt.
4517 *) Add the ASN.1 structures and functions for CertificatePair, which
4518 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4520 CertificatePair ::= SEQUENCE {
4521 forward [0] Certificate OPTIONAL,
4522 reverse [1] Certificate OPTIONAL,
4523 -- at least one of the pair shall be present -- }
4525 Also implement the PEM functions to read and write certificate
4526 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4528 This needed to be defined, mostly for the sake of the LDAP
4529 attribute crossCertificatePair, but may prove useful elsewhere as
4533 *) Make it possible to inhibit symlinking of shared libraries in
4534 Makefile.shared, for Cygwin's sake.
4537 *) Extend the BIGNUM API by creating a function
4538 void BN_set_negative(BIGNUM *a, int neg);
4539 and a macro that behave like
4540 int BN_is_negative(const BIGNUM *a);
4542 to avoid the need to access 'a->neg' directly in applications.
4545 *) Implement fast modular reduction for pseudo-Mersenne primes
4546 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4547 EC_GROUP_new_curve_GFp() will now automatically use this
4549 [Nils Larsch <nla@trustcenter.de>]
4551 *) Add new lock type (CRYPTO_LOCK_BN).
4554 *) Change the ENGINE framework to automatically load engines
4555 dynamically from specific directories unless they could be
4556 found to already be built in or loaded. Move all the
4557 current engines except for the cryptodev one to a new
4559 The engines in engines/ are built as shared libraries if
4560 the "shared" options was given to ./Configure or ./config.
4561 Otherwise, they are inserted in libcrypto.a.
4562 /usr/local/ssl/engines is the default directory for dynamic
4563 engines, but that can be overriden at configure time through
4564 the usual use of --prefix and/or --openssldir, and at run
4565 time with the environment variable OPENSSL_ENGINES.
4566 [Geoff Thorpe and Richard Levitte]
4568 *) Add Makefile.shared, a helper makefile to build shared
4569 libraries. Adapt Makefile.org.
4572 *) Add version info to Win32 DLLs.
4573 [Peter 'Luna' Runestig" <peter@runestig.com>]
4575 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4576 can be added using this API to created arbitrary PKCS#12
4577 files while avoiding the low level API.
4579 New options to PKCS12_create(), key or cert can be NULL and
4580 will then be omitted from the output file. The encryption
4581 algorithm NIDs can be set to -1 for no encryption, the mac
4582 iteration count can be set to 0 to omit the mac.
4584 Enhance pkcs12 utility by making the -nokeys and -nocerts
4585 options work when creating a PKCS#12 file. New option -nomac
4586 to omit the mac, NONE can be set for an encryption algorithm.
4587 New code is modified to use the enhanced PKCS12_create()
4588 instead of the low level API.
4591 *) Extend ASN1 encoder to support indefinite length constructed
4592 encoding. This can output sequences tags and octet strings in
4593 this form. Modify pk7_asn1.c to support indefinite length
4594 encoding. This is experimental and needs additional code to
4595 be useful, such as an ASN1 bio and some enhanced streaming
4598 Extend template encode functionality so that tagging is passed
4599 down to the template encoder.
4602 *) Let 'openssl req' fail if an argument to '-newkey' is not
4603 recognized instead of using RSA as a default.
4606 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4607 As these are not official, they are not included in "ALL";
4608 the "ECCdraft" ciphersuite group alias can be used to select them.
4609 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4611 *) Add ECDH engine support.
4612 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4614 *) Add ECDH in new directory crypto/ecdh/.
4615 [Douglas Stebila (Sun Microsystems Laboratories)]
4617 *) Let BN_rand_range() abort with an error after 100 iterations
4618 without success (which indicates a broken PRNG).
4621 *) Change BN_mod_sqrt() so that it verifies that the input value
4622 is really the square of the return value. (Previously,
4623 BN_mod_sqrt would show GIGO behaviour.)
4626 *) Add named elliptic curves over binary fields from X9.62, SECG,
4627 and WAP/WTLS; add OIDs that were still missing.
4629 [Sheueling Chang Shantz and Douglas Stebila
4630 (Sun Microsystems Laboratories)]
4632 *) Extend the EC library for elliptic curves over binary fields
4633 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4636 EC_GF2m_simple_method
4640 EC_GROUP_new_curve_GF2m
4641 EC_GROUP_set_curve_GF2m
4642 EC_GROUP_get_curve_GF2m
4643 EC_POINT_set_affine_coordinates_GF2m
4644 EC_POINT_get_affine_coordinates_GF2m
4645 EC_POINT_set_compressed_coordinates_GF2m
4647 Point compression for binary fields is disabled by default for
4648 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4651 As binary polynomials are represented as BIGNUMs, various members
4652 of the EC_GROUP and EC_POINT data structures can be shared
4653 between the implementations for prime fields and binary fields;
4654 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4655 are essentially identical to their ..._GFp counterparts.
4656 (For simplicity, the '..._GFp' prefix has been dropped from
4657 various internal method names.)
4659 An internal 'field_div' method (similar to 'field_mul' and
4660 'field_sqr') has been added; this is used only for binary fields.
4662 [Sheueling Chang Shantz and Douglas Stebila
4663 (Sun Microsystems Laboratories)]
4665 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4666 through methods ('mul', 'precompute_mult').
4668 The generic implementations (now internally called 'ec_wNAF_mul'
4669 and 'ec_wNAF_precomputed_mult') remain the default if these
4670 methods are undefined.
4672 [Sheueling Chang Shantz and Douglas Stebila
4673 (Sun Microsystems Laboratories)]
4675 *) New function EC_GROUP_get_degree, which is defined through
4676 EC_METHOD. For curves over prime fields, this returns the bit
4677 length of the modulus.
4679 [Sheueling Chang Shantz and Douglas Stebila
4680 (Sun Microsystems Laboratories)]
4682 *) New functions EC_GROUP_dup, EC_POINT_dup.
4683 (These simply call ..._new and ..._copy).
4685 [Sheueling Chang Shantz and Douglas Stebila
4686 (Sun Microsystems Laboratories)]
4688 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4689 Polynomials are represented as BIGNUMs (where the sign bit is not
4690 used) in the following functions [macros]:
4693 BN_GF2m_sub [= BN_GF2m_add]
4694 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4695 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4696 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4698 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4699 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4700 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4701 BN_GF2m_cmp [= BN_ucmp]
4703 (Note that only the 'mod' functions are actually for fields GF(2^m).
4704 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4706 For some functions, an the irreducible polynomial defining a
4707 field can be given as an 'unsigned int[]' with strictly
4708 decreasing elements giving the indices of those bits that are set;
4709 i.e., p[] represents the polynomial
4710 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4712 p[0] > p[1] > ... > p[k] = 0.
4713 This applies to the following functions:
4718 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4719 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4721 BN_GF2m_mod_sqrt_arr
4722 BN_GF2m_mod_solve_quad_arr
4726 Conversion can be performed by the following functions:
4731 bntest.c has additional tests for binary polynomial arithmetic.
4733 Two implementations for BN_GF2m_mod_div() are available.
4734 The default algorithm simply uses BN_GF2m_mod_inv() and
4735 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4736 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4737 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4739 [Sheueling Chang Shantz and Douglas Stebila
4740 (Sun Microsystems Laboratories)]
4742 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4743 functionality is disabled at compile-time.
4744 [Douglas Stebila <douglas.stebila@sun.com>]
4746 *) Change default behaviour of 'openssl asn1parse' so that more
4747 information is visible when viewing, e.g., a certificate:
4749 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4750 mode the content of non-printable OCTET STRINGs is output in a
4751 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4752 avoid the appearance of a printable string.
4753 [Nils Larsch <nla@trustcenter.de>]
4755 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4757 EC_GROUP_set_asn1_flag()
4758 EC_GROUP_get_asn1_flag()
4759 EC_GROUP_set_point_conversion_form()
4760 EC_GROUP_get_point_conversion_form()
4761 These control ASN1 encoding details:
4762 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4763 has been set to OPENSSL_EC_NAMED_CURVE.
4764 - Points are encoded in uncompressed form by default; options for
4765 asn1_for are as for point2oct, namely
4766 POINT_CONVERSION_COMPRESSED
4767 POINT_CONVERSION_UNCOMPRESSED
4768 POINT_CONVERSION_HYBRID
4770 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4773 EC_GROUP_get0_seed()
4774 EC_GROUP_get_seed_len()
4775 This is used only for ASN1 purposes (so far).
4776 [Nils Larsch <nla@trustcenter.de>]
4778 *) Add 'field_type' member to EC_METHOD, which holds the NID
4779 of the appropriate field type OID. The new function
4780 EC_METHOD_get_field_type() returns this value.
4781 [Nils Larsch <nla@trustcenter.de>]
4786 EC_POINT_point2hex()
4787 EC_POINT_hex2point()
4788 providing useful interfaces to EC_POINT_point2oct() and
4789 EC_POINT_oct2point().
4790 [Nils Larsch <nla@trustcenter.de>]
4792 *) Change internals of the EC library so that the functions
4793 EC_GROUP_set_generator()
4794 EC_GROUP_get_generator()
4795 EC_GROUP_get_order()
4796 EC_GROUP_get_cofactor()
4797 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4798 to methods, which would lead to unnecessary code duplication when
4799 adding different types of curves.
4800 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4802 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4803 arithmetic, and such that modified wNAFs are generated
4804 (which avoid length expansion in many cases).
4807 *) Add a function EC_GROUP_check_discriminant() (defined via
4808 EC_METHOD) that verifies that the curve discriminant is non-zero.
4810 Add a function EC_GROUP_check() that makes some sanity tests
4811 on a EC_GROUP, its generator and order. This includes
4812 EC_GROUP_check_discriminant().
4813 [Nils Larsch <nla@trustcenter.de>]
4815 *) Add ECDSA in new directory crypto/ecdsa/.
4817 Add applications 'openssl ecparam' and 'openssl ecdsa'
4818 (these are based on 'openssl dsaparam' and 'openssl dsa').
4820 ECDSA support is also included in various other files across the
4821 library. Most notably,
4822 - 'openssl req' now has a '-newkey ecdsa:file' option;
4823 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4824 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4825 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4826 them suitable for ECDSA where domain parameters must be
4827 extracted before the specific public key;
4828 - ECDSA engine support has been added.
4829 [Nils Larsch <nla@trustcenter.de>]
4831 *) Include some named elliptic curves, and add OIDs from X9.62,
4832 SECG, and WAP/WTLS. Each curve can be obtained from the new
4834 EC_GROUP_new_by_curve_name(),
4835 and the list of available named curves can be obtained with
4836 EC_get_builtin_curves().
4837 Also add a 'curve_name' member to EC_GROUP objects, which can be
4839 EC_GROUP_set_curve_name()
4840 EC_GROUP_get_curve_name()
4841 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4843 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4844 was actually never needed) and in BN_mul(). The removal in BN_mul()
4845 required a small change in bn_mul_part_recursive() and the addition
4846 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4847 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4848 bn_sub_words() and bn_add_words() except they take arrays with
4852 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4854 *) Cleanse PEM buffers before freeing them since they may contain
4856 [Benjamin Bennett <ben@psc.edu>]
4858 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4859 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4860 authentication-only ciphersuites.
4863 *) Since AES128 and AES256 share a single mask bit in the logic of
4864 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4865 kludge to work properly if AES128 is available and AES256 isn't.
4868 *) Expand security boundary to match 1.1.1 module.
4871 *) Remove redundant features: hash file source, editing of test vectors
4872 modify fipsld to use external fips_premain.c signature.
4875 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4876 run algorithm test programs.
4879 *) Make algorithm test programs more tolerant of whitespace.
4882 *) Have SSL/TLS server implementation tolerate "mismatched" record
4883 protocol version while receiving ClientHello even if the
4884 ClientHello is fragmented. (The server can't insist on the
4885 particular protocol version it has chosen before the ServerHello
4886 message has informed the client about his choice.)
4889 *) Load error codes if they are not already present instead of using a
4890 static variable. This allows them to be cleanly unloaded and reloaded.
4893 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4895 *) Introduce limits to prevent malicious keys being able to
4896 cause a denial of service. (CVE-2006-2940)
4897 [Steve Henson, Bodo Moeller]
4899 *) Fix ASN.1 parsing of certain invalid structures that can result
4900 in a denial of service. (CVE-2006-2937) [Steve Henson]
4902 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4903 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4905 *) Fix SSL client code which could crash if connecting to a
4906 malicious SSLv2 server. (CVE-2006-4343)
4907 [Tavis Ormandy and Will Drewry, Google Security Team]
4909 *) Change ciphersuite string processing so that an explicit
4910 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4911 will no longer include "AES128-SHA"), and any other similar
4912 ciphersuite (same bitmap) from *other* protocol versions (so that
4913 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4914 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4915 changes from 0.9.8b and 0.9.8d.
4918 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4920 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4921 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4923 *) Change the Unix randomness entropy gathering to use poll() when
4924 possible instead of select(), since the latter has some
4925 undesirable limitations.
4926 [Darryl Miles via Richard Levitte and Bodo Moeller]
4928 *) Disable rogue ciphersuites:
4930 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4931 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4932 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4934 The latter two were purportedly from
4935 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4938 Also deactive the remaining ciphersuites from
4939 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4940 unofficial, and the ID has long expired.
4943 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4944 dual-core machines) and other potential thread-safety issues.
4947 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4949 *) Adapt fipsld and the build system to link against the validated FIPS
4950 module in FIPS mode.
4953 *) Fixes for VC++ 2005 build under Windows.
4956 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4957 from a Windows bash shell such as MSYS. It is autodetected from the
4958 "config" script when run from a VC++ environment. Modify standard VC++
4959 build to use fipscanister.o from the GNU make build.
4962 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4964 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4965 The value now differs depending on if you build for FIPS or not.
4966 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4967 safely run with a non-FIPSed libcrypto, as it may crash because of
4968 the difference induced by this change.
4971 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4973 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4974 (part of SSL_OP_ALL). This option used to disable the
4975 countermeasure against man-in-the-middle protocol-version
4976 rollback in the SSL 2.0 server implementation, which is a bad
4977 idea. (CVE-2005-2969)
4979 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4980 for Information Security, National Institute of Advanced Industrial
4981 Science and Technology [AIST], Japan)]
4983 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4984 mainly for FIPS compliance and not fully integrated at this stage.
4987 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4988 the exponentiation using a fixed-length exponent. (Otherwise,
4989 the information leaked through timing could expose the secret key
4990 after many signatures; cf. Bleichenbacher's attack on DSA with
4994 *) Make a new fixed-window mod_exp implementation the default for
4995 RSA, DSA, and DH private-key operations so that the sequence of
4996 squares and multiplies and the memory access pattern are
4997 independent of the particular secret key. This will mitigate
4998 cache-timing and potential related attacks.
5000 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5001 and this is automatically used by BN_mod_exp_mont() if the new flag
5002 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5003 will use this BN flag for private exponents unless the flag
5004 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5005 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5007 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5009 *) Change the client implementation for SSLv23_method() and
5010 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5011 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5012 (Previously, the SSL 2.0 backwards compatible Client Hello
5013 message format would be used even with SSL_OP_NO_SSLv2.)
5016 *) Add support for smime-type MIME parameter in S/MIME messages which some
5020 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5021 a threadsafe manner. Modify rsa code to use new function and add calls
5022 to dsa and dh code (which had race conditions before).
5025 *) Include the fixed error library code in the C error file definitions
5026 instead of fixing them up at runtime. This keeps the error code
5027 structures constant.
5030 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5032 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5035 *) Fixes for newer kerberos headers. NB: the casts are needed because
5036 the 'length' field is signed on one version and unsigned on another
5037 with no (?) obvious way to tell the difference, without these VC++
5038 complains. Also the "definition" of FAR (blank) is no longer included
5039 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5040 some needed definitions.
5043 *) Undo Cygwin change.
5046 *) Added support for proxy certificates according to RFC 3820.
5047 Because they may be a security thread to unaware applications,
5048 they must be explicitely allowed in run-time. See
5049 docs/HOWTO/proxy_certificates.txt for further information.
5052 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5054 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5055 server and client random values. Previously
5056 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5057 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5059 This change has negligible security impact because:
5061 1. Server and client random values still have 24 bytes of pseudo random
5064 2. Server and client random values are sent in the clear in the initial
5067 3. The master secret is derived using the premaster secret (48 bytes in
5068 size for static RSA ciphersuites) as well as client server and random
5071 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5074 [Stephen Henson, reported by UK NISCC]
5076 *) Use Windows randomness collection on Cygwin.
5079 *) Fix hang in EGD/PRNGD query when communication socket is closed
5080 prematurely by EGD/PRNGD.
5081 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5083 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5086 *) Back-port of selected performance improvements from development
5087 branch, as well as improved support for PowerPC platforms.
5090 *) Add lots of checks for memory allocation failure, error codes to indicate
5091 failure and freeing up memory if a failure occurs.
5092 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5094 *) Add new -passin argument to dgst.
5097 *) Perform some character comparisons of different types in X509_NAME_cmp:
5098 this is needed for some certificates that reencode DNs into UTF8Strings
5099 (in violation of RFC3280) and can't or wont issue name rollover
5103 *) Make an explicit check during certificate validation to see that
5104 the CA setting in each certificate on the chain is correct. As a
5105 side effect always do the following basic checks on extensions,
5106 not just when there's an associated purpose to the check:
5108 - if there is an unhandled critical extension (unless the user
5109 has chosen to ignore this fault)
5110 - if the path length has been exceeded (if one is set at all)
5111 - that certain extensions fit the associated purpose (if one has
5115 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5117 *) Avoid a race condition when CRLs are checked in a multi threaded
5118 environment. This would happen due to the reordering of the revoked
5119 entries during signature checking and serial number lookup. Now the
5120 encoding is cached and the serial number sort performed under a lock.
5121 Add new STACK function sk_is_sorted().
5124 *) Add Delta CRL to the extension code.
5127 *) Various fixes to s3_pkt.c so alerts are sent properly.
5128 [David Holmes <d.holmes@f5.com>]
5130 *) Reduce the chances of duplicate issuer name and serial numbers (in
5131 violation of RFC3280) using the OpenSSL certificate creation utilities.
5132 This is done by creating a random 64 bit value for the initial serial
5133 number when a serial number file is created or when a self signed
5134 certificate is created using 'openssl req -x509'. The initial serial
5135 number file is created using 'openssl x509 -next_serial' in CA.pl
5136 rather than being initialized to 1.
5139 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5141 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5142 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5143 [Joe Orton, Steve Henson]
5145 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5147 [Joe Orton, Steve Henson]
5149 *) Make it possible to have multiple active certificates with the same
5150 subject in the CA index file. This is done only if the keyword
5151 'unique_subject' is set to 'no' in the main CA section (default
5152 if 'CA_default') of the configuration file. The value is saved
5153 with the database itself in a separate index attribute file,
5154 named like the index file with '.attr' appended to the name.
5157 *) X509 verify fixes. Disable broken certificate workarounds when
5158 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5159 keyUsage extension present. Don't accept CRLs with unhandled critical
5160 extensions: since verify currently doesn't process CRL extensions this
5161 rejects a CRL with *any* critical extensions. Add new verify error codes
5165 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5166 A clarification of RFC2560 will require the use of OCTET STRINGs and
5167 some implementations cannot handle the current raw format. Since OpenSSL
5168 copies and compares OCSP nonces as opaque blobs without any attempt at
5169 parsing them this should not create any compatibility issues.
5172 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5173 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5174 this HMAC (and other) operations are several times slower than OpenSSL
5178 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5179 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5181 *) Use the correct content when signing type "other".
5184 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5186 *) Fix various bugs revealed by running the NISCC test suite:
5188 Stop out of bounds reads in the ASN1 code when presented with
5189 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5191 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5193 If verify callback ignores invalid public key errors don't try to check
5194 certificate signature with the NULL public key.
5198 *) New -ignore_err option in ocsp application to stop the server
5199 exiting on the first error in a request.
5202 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5203 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5207 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5208 extra data after the compression methods not only for TLS 1.0
5209 but also for SSL 3.0 (as required by the specification).
5210 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5212 *) Change X509_certificate_type() to mark the key as exported/exportable
5213 when it's 512 *bits* long, not 512 bytes.
5216 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5217 blocks during encryption.
5220 *) Various fixes to base64 BIO and non blocking I/O. On write
5221 flushes were not handled properly if the BIO retried. On read
5222 data was not being buffered properly and had various logic bugs.
5223 This also affects blocking I/O when the data being decoded is a
5227 *) Various S/MIME bugfixes and compatibility changes:
5228 output correct application/pkcs7 MIME type if
5229 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5230 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5231 of files as .eml work). Correctly handle very long lines in MIME
5235 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5237 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5238 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5239 a protocol version number mismatch like a decryption error
5240 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5243 *) Turn on RSA blinding by default in the default implementation
5244 to avoid a timing attack. Applications that don't want it can call
5245 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5246 They would be ill-advised to do so in most cases.
5247 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5249 *) Change RSA blinding code so that it works when the PRNG is not
5250 seeded (in this case, the secret RSA exponent is abused as
5251 an unpredictable seed -- if it is not unpredictable, there
5252 is no point in blinding anyway). Make RSA blinding thread-safe
5253 by remembering the creator's thread ID in rsa->blinding and
5254 having all other threads use local one-time blinding factors
5255 (this requires more computation than sharing rsa->blinding, but
5256 avoids excessive locking; and if an RSA object is not shared
5257 between threads, blinding will still be very fast).
5260 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5261 ENGINE as defaults for all supported algorithms irrespective of
5262 the 'flags' parameter. 'flags' is now honoured, so applications
5263 should make sure they are passing it correctly.
5266 *) Target "mingw" now allows native Windows code to be generated in
5267 the Cygwin environment as well as with the MinGW compiler.
5270 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5272 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5273 via timing by performing a MAC computation even if incorrect
5274 block cipher padding has been found. This is a countermeasure
5275 against active attacks where the attacker has to distinguish
5276 between bad padding and a MAC verification error. (CVE-2003-0078)
5278 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5279 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5280 Martin Vuagnoux (EPFL, Ilion)]
5282 *) Make the no-err option work as intended. The intention with no-err
5283 is not to have the whole error stack handling routines removed from
5284 libcrypto, it's only intended to remove all the function name and
5285 reason texts, thereby removing some of the footprint that may not
5286 be interesting if those errors aren't displayed anyway.
5288 NOTE: it's still possible for any application or module to have it's
5289 own set of error texts inserted. The routines are there, just not
5290 used by default when no-err is given.
5293 *) Add support for FreeBSD on IA64.
5294 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5296 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5297 Kerberos function mit_des_cbc_cksum(). Before this change,
5298 the value returned by DES_cbc_cksum() was like the one from
5299 mit_des_cbc_cksum(), except the bytes were swapped.
5300 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5302 *) Allow an application to disable the automatic SSL chain building.
5303 Before this a rather primitive chain build was always performed in
5304 ssl3_output_cert_chain(): an application had no way to send the
5305 correct chain if the automatic operation produced an incorrect result.
5307 Now the chain builder is disabled if either:
5309 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5311 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5313 The reasoning behind this is that an application would not want the
5314 auto chain building to take place if extra chain certificates are
5315 present and it might also want a means of sending no additional
5316 certificates (for example the chain has two certificates and the
5320 *) Add the possibility to build without the ENGINE framework.
5321 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5323 *) Under Win32 gmtime() can return NULL: check return value in
5324 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5327 *) DSA routines: under certain error conditions uninitialized BN objects
5328 could be freed. Solution: make sure initialization is performed early
5329 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5330 Nils Larsch <nla@trustcenter.de> via PR#459)
5333 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5334 checked on reconnect on the client side, therefore session resumption
5335 could still fail with a "ssl session id is different" error. This
5336 behaviour is masked when SSL_OP_ALL is used due to
5337 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5338 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5339 followup to PR #377.
5342 *) IA-32 assembler support enhancements: unified ELF targets, support
5343 for SCO/Caldera platforms, fix for Cygwin shared build.
5346 *) Add support for FreeBSD on sparc64. As a consequence, support for
5347 FreeBSD on non-x86 processors is separate from x86 processors on
5348 the config script, much like the NetBSD support.
5349 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5351 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5353 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5356 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5357 code (06) was taken as the first octet of the session ID and the last
5358 octet was ignored consequently. As a result SSLv2 client side session
5359 caching could not have worked due to the session ID mismatch between
5361 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5365 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5366 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5370 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5371 seems that in spite of existing for more than a year, many application
5372 author have done nothing to provide the necessary callbacks, which
5373 means that this particular engine will not work properly anywhere.
5374 This is a very unfortunate situation which forces us, in the name
5375 of usability, to give the hw_ncipher.c a static lock, which is part
5377 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5378 appear in 0.9.8 or later. We EXPECT application authors to have
5379 dealt properly with this when 0.9.8 is released (unless we actually
5380 make such changes in the libcrypto locking code that changes will
5381 have to be made anyway).
5384 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5385 octets have been read, EOF or an error occurs. Without this change
5386 some truncated ASN1 structures will not produce an error.
5389 *) Disable Heimdal support, since it hasn't been fully implemented.
5390 Still give the possibility to force the use of Heimdal, but with
5391 warnings and a request that patches get sent to openssl-dev.
5394 *) Add the VC-CE target, introduce the WINCE sysname, and add
5395 INSTALL.WCE and appropriate conditionals to make it build.
5396 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5398 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5399 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5400 edit numbers of the version.
5401 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5403 *) Introduce safe string copy and catenation functions
5404 (BUF_strlcpy() and BUF_strlcat()).
5405 [Ben Laurie (CHATS) and Richard Levitte]
5407 *) Avoid using fixed-size buffers for one-line DNs.
5408 [Ben Laurie (CHATS)]
5410 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5411 resizing buffers containing secrets, and use where appropriate.
5412 [Ben Laurie (CHATS)]
5414 *) Avoid using fixed size buffers for configuration file location.
5415 [Ben Laurie (CHATS)]
5417 *) Avoid filename truncation for various CA files.
5418 [Ben Laurie (CHATS)]
5420 *) Use sizeof in preference to magic numbers.
5421 [Ben Laurie (CHATS)]
5423 *) Avoid filename truncation in cert requests.
5424 [Ben Laurie (CHATS)]
5426 *) Add assertions to check for (supposedly impossible) buffer
5428 [Ben Laurie (CHATS)]
5430 *) Don't cache truncated DNS entries in the local cache (this could
5431 potentially lead to a spoofing attack).
5432 [Ben Laurie (CHATS)]
5434 *) Fix various buffers to be large enough for hex/decimal
5435 representations in a platform independent manner.
5436 [Ben Laurie (CHATS)]
5438 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5439 resizing buffers containing secrets, and use where appropriate.
5440 [Ben Laurie (CHATS)]
5442 *) Add BIO_indent() to avoid much slightly worrying code to do
5444 [Ben Laurie (CHATS)]
5446 *) Convert sprintf()/BIO_puts() to BIO_printf().
5447 [Ben Laurie (CHATS)]
5449 *) buffer_gets() could terminate with the buffer only half
5451 [Ben Laurie (CHATS)]
5453 *) Add assertions to prevent user-supplied crypto functions from
5454 overflowing internal buffers by having large block sizes, etc.
5455 [Ben Laurie (CHATS)]
5457 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5459 [Ben Laurie (CHATS)]
5461 *) Eliminate unused copy of key in RC4.
5462 [Ben Laurie (CHATS)]
5464 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5465 [Ben Laurie (CHATS)]
5467 *) Fix off-by-one error in EGD path.
5468 [Ben Laurie (CHATS)]
5470 *) If RANDFILE path is too long, ignore instead of truncating.
5471 [Ben Laurie (CHATS)]
5473 *) Eliminate unused and incorrectly sized X.509 structure
5475 [Ben Laurie (CHATS)]
5477 *) Eliminate unused and dangerous function knumber().
5478 [Ben Laurie (CHATS)]
5480 *) Eliminate unused and dangerous structure, KSSL_ERR.
5481 [Ben Laurie (CHATS)]
5483 *) Protect against overlong session ID context length in an encoded
5484 session object. Since these are local, this does not appear to be
5486 [Ben Laurie (CHATS)]
5488 *) Change from security patch (see 0.9.6e below) that did not affect
5489 the 0.9.6 release series:
5491 Remote buffer overflow in SSL3 protocol - an attacker could
5492 supply an oversized master key in Kerberos-enabled versions.
5494 [Ben Laurie (CHATS)]
5496 *) Change the SSL kerb5 codes to match RFC 2712.
5499 *) Make -nameopt work fully for req and add -reqopt switch.
5500 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5502 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5503 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5505 *) Make sure tests can be performed even if the corresponding algorithms
5506 have been removed entirely. This was also the last step to make
5507 OpenSSL compilable with DJGPP under all reasonable conditions.
5508 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5510 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5511 to allow version independent disabling of normally unselected ciphers,
5512 which may be activated as a side-effect of selecting a single cipher.
5514 (E.g., cipher list string "RSA" enables ciphersuites that are left
5515 out of "ALL" because they do not provide symmetric encryption.
5516 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5517 [Lutz Jaenicke, Bodo Moeller]
5519 *) Add appropriate support for separate platform-dependent build
5520 directories. The recommended way to make a platform-dependent
5521 build directory is the following (tested on Linux), maybe with
5524 # Place yourself outside of the OpenSSL source tree. In
5525 # this example, the environment variable OPENSSL_SOURCE
5526 # is assumed to contain the absolute OpenSSL source directory.
5527 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5528 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5529 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5530 mkdir -p `dirname $F`
5531 ln -s $OPENSSL_SOURCE/$F $F
5534 To be absolutely sure not to disturb the source tree, a "make clean"
5535 is a good thing. If it isn't successful, don't worry about it,
5536 it probably means the source directory is very clean.
5539 *) Make sure any ENGINE control commands make local copies of string
5540 pointers passed to them whenever necessary. Otherwise it is possible
5541 the caller may have overwritten (or deallocated) the original string
5542 data when a later ENGINE operation tries to use the stored values.
5543 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5545 *) Improve diagnostics in file reading and command-line digests.
5546 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5548 *) Add AES modes CFB and OFB to the object database. Correct an
5549 error in AES-CFB decryption.
5552 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5553 allows existing EVP_CIPHER_CTX structures to be reused after
5554 calling EVP_*Final(). This behaviour is used by encryption
5555 BIOs and some applications. This has the side effect that
5556 applications must explicitly clean up cipher contexts with
5557 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5560 *) Check the values of dna and dnb in bn_mul_recursive before calling
5561 bn_mul_comba (a non zero value means the a or b arrays do not contain
5562 n2 elements) and fallback to bn_mul_normal if either is not zero.
5565 *) Fix escaping of non-ASCII characters when using the -subj option
5566 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5569 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5570 form for "surname", serialNumber has no short form.
5571 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5572 therefore remove "mail" short name for "internet 7".
5573 The OID for unique identifiers in X509 certificates is
5574 x500UniqueIdentifier, not uniqueIdentifier.
5575 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5578 *) Add an "init" command to the ENGINE config module and auto initialize
5579 ENGINEs. Without any "init" command the ENGINE will be initialized
5580 after all ctrl commands have been executed on it. If init=1 the
5581 ENGINE is initailized at that point (ctrls before that point are run
5582 on the uninitialized ENGINE and after on the initialized one). If
5583 init=0 then the ENGINE will not be iniatialized at all.
5586 *) Fix the 'app_verify_callback' interface so that the user-defined
5587 argument is actually passed to the callback: In the
5588 SSL_CTX_set_cert_verify_callback() prototype, the callback
5589 declaration has been changed from
5592 int (*cb)(X509_STORE_CTX *,void *);
5593 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5594 i=s->ctx->app_verify_callback(&ctx)
5595 has been changed into
5596 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5598 To update applications using SSL_CTX_set_cert_verify_callback(),
5599 a dummy argument can be added to their callback functions.
5600 [D. K. Smetters <smetters@parc.xerox.com>]
5602 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5603 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5605 *) Add and OPENSSL_LOAD_CONF define which will cause
5606 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5607 This allows older applications to transparently support certain
5608 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5609 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5610 load the config file and OPENSSL_add_all_algorithms_conf() which will
5611 always load it have also been added.
5614 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5615 Adjust NIDs and EVP layer.
5616 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5618 *) Config modules support in openssl utility.
5620 Most commands now load modules from the config file,
5621 though in a few (such as version) this isn't done
5622 because it couldn't be used for anything.
5624 In the case of ca and req the config file used is
5625 the same as the utility itself: that is the -config
5626 command line option can be used to specify an
5630 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5631 use "openssl_conf" if filename is NULL use default openssl config file.
5634 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5635 config section name. Add a new flag to tolerate a missing config file
5636 and move code to CONF_modules_load_file().
5639 *) Support for crypto accelerator cards from Accelerated Encryption
5640 Processing, www.aep.ie. (Use engine 'aep')
5641 The support was copied from 0.9.6c [engine] and adapted/corrected
5642 to work with the new engine framework.
5643 [AEP Inc. and Richard Levitte]
5645 *) Support for SureWare crypto accelerator cards from Baltimore
5646 Technologies. (Use engine 'sureware')
5647 The support was copied from 0.9.6c [engine] and adapted
5648 to work with the new engine framework.
5651 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5652 make the newer ENGINE framework commands for the CHIL engine work.
5653 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5655 *) Make it possible to produce shared libraries on ReliantUNIX.
5656 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5658 *) Add the configuration target debug-linux-ppro.
5659 Make 'openssl rsa' use the general key loading routines
5660 implemented in apps.c, and make those routines able to
5661 handle the key format FORMAT_NETSCAPE and the variant
5663 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5665 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5666 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5668 *) Add -keyform to rsautl, and document -engine.
5669 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5671 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5672 BIO_R_NO_SUCH_FILE error code rather than the generic
5673 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5676 *) Add new functions
5678 ERR_peek_last_error_line
5679 ERR_peek_last_error_line_data.
5680 These are similar to
5683 ERR_peek_error_line_data,
5684 but report on the latest error recorded rather than the first one
5685 still in the error queue.
5686 [Ben Laurie, Bodo Moeller]
5688 *) default_algorithms option in ENGINE config module. This allows things
5690 default_algorithms = ALL
5691 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5694 *) Preliminary ENGINE config module.
5697 *) New experimental application configuration code.
5700 *) Change the AES code to follow the same name structure as all other
5701 symmetric ciphers, and behave the same way. Move everything to
5702 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5703 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5705 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5706 [Ben Laurie and Theo de Raadt]
5708 *) Add option to output public keys in req command.
5709 [Massimiliano Pala madwolf@openca.org]
5711 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5712 (up to about 10% better than before for P-192 and P-224).
5715 *) New functions/macros
5717 SSL_CTX_set_msg_callback(ctx, cb)
5718 SSL_CTX_set_msg_callback_arg(ctx, arg)
5719 SSL_set_msg_callback(ssl, cb)
5720 SSL_set_msg_callback_arg(ssl, arg)
5722 to request calling a callback function
5724 void cb(int write_p, int version, int content_type,
5725 const void *buf, size_t len, SSL *ssl, void *arg)
5727 whenever a protocol message has been completely received
5728 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5729 protocol version according to which the SSL library interprets
5730 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5731 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5732 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5733 specification (change_cipher_spec(20), alert(21), handshake(22)).
5734 'buf' and 'len' point to the actual message, 'ssl' to the
5735 SSL object, and 'arg' is the application-defined value set by
5736 SSL[_CTX]_set_msg_callback_arg().
5738 'openssl s_client' and 'openssl s_server' have new '-msg' options
5739 to enable a callback that displays all protocol messages.
5742 *) Change the shared library support so shared libraries are built as
5743 soon as the corresponding static library is finished, and thereby get
5744 openssl and the test programs linked against the shared library.
5745 This still only happens when the keyword "shard" has been given to
5746 the configuration scripts.
5748 NOTE: shared library support is still an experimental thing, and
5749 backward binary compatibility is still not guaranteed.
5750 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5752 *) Add support for Subject Information Access extension.
5753 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5755 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5756 additional bytes when new memory had to be allocated, not just
5757 when reusing an existing buffer.
5760 *) New command line and configuration option 'utf8' for the req command.
5761 This allows field values to be specified as UTF8 strings.
5764 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5765 runs for the former and machine-readable output for the latter.
5768 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5769 of the e-mail address in the DN (i.e., it will go into a certificate
5770 extension only). The new configuration file option 'email_in_dn = no'
5771 has the same effect.
5772 [Massimiliano Pala madwolf@openca.org]
5774 *) Change all functions with names starting with des_ to be starting
5775 with DES_ instead. Add wrappers that are compatible with libdes,
5776 but are named _ossl_old_des_*. Finally, add macros that map the
5777 des_* symbols to the corresponding _ossl_old_des_* if libdes
5778 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5779 desired, the des_* symbols will be mapped to DES_*, with one
5782 Since we provide two compatibility mappings, the user needs to
5783 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5784 compatibility is desired. The default (i.e., when that macro
5785 isn't defined) is OpenSSL 0.9.6c compatibility.
5787 There are also macros that enable and disable the support of old
5788 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5789 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5790 are defined, the default will apply: to support the old des routines.
5792 In either case, one must include openssl/des.h to get the correct
5793 definitions. Do not try to just include openssl/des_old.h, that
5796 NOTE: This is a major break of an old API into a new one. Software
5797 authors are encouraged to switch to the DES_ style functions. Some
5798 time in the future, des_old.h and the libdes compatibility functions
5799 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5800 default), and then completely removed.
5803 *) Test for certificates which contain unsupported critical extensions.
5804 If such a certificate is found during a verify operation it is
5805 rejected by default: this behaviour can be overridden by either
5806 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5807 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5808 X509_supported_extension() has also been added which returns 1 if a
5809 particular extension is supported.
5812 *) Modify the behaviour of EVP cipher functions in similar way to digests
5813 to retain compatibility with existing code.
5816 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5817 compatibility with existing code. In particular the 'ctx' parameter does
5818 not have to be to be initialized before the call to EVP_DigestInit() and
5819 it is tidied up after a call to EVP_DigestFinal(). New function
5820 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5821 EVP_MD_CTX_copy() changed to not require the destination to be
5822 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5823 requires the destination to be valid.
5825 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5826 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5829 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5830 so that complete 'Handshake' protocol structures are kept in memory
5831 instead of overwriting 'msg_type' and 'length' with 'body' data.
5834 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5835 [Massimo Santin via Richard Levitte]
5837 *) Major restructuring to the underlying ENGINE code. This includes
5838 reduction of linker bloat, separation of pure "ENGINE" manipulation
5839 (initialisation, etc) from functionality dealing with implementations
5840 of specific crypto iterfaces. This change also introduces integrated
5841 support for symmetric ciphers and digest implementations - so ENGINEs
5842 can now accelerate these by providing EVP_CIPHER and EVP_MD
5843 implementations of their own. This is detailed in crypto/engine/README
5844 as it couldn't be adequately described here. However, there are a few
5845 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5846 were changed in the original introduction of ENGINE code have now
5847 reverted back - the hooking from this code to ENGINE is now a good
5848 deal more passive and at run-time, operations deal directly with
5849 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5850 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5851 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5852 they were not being used by the framework as there is no concept of a
5853 BIGNUM_METHOD and they could not be generalised to the new
5854 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5855 ENGINE_cpy() has been removed as it cannot be consistently defined in
5859 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5862 *) Change mkdef.pl to sort symbols that get the same entry number,
5863 and make sure the automatically generated functions ERR_load_*
5864 become part of libeay.num as well.
5867 *) New function SSL_renegotiate_pending(). This returns true once
5868 renegotiation has been requested (either SSL_renegotiate() call
5869 or HelloRequest/ClientHello received from the peer) and becomes
5870 false once a handshake has been completed.
5871 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5872 sends a HelloRequest, but does not ensure that a handshake takes
5873 place. SSL_renegotiate_pending() is useful for checking if the
5874 client has followed the request.)
5877 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5878 By default, clients may request session resumption even during
5879 renegotiation (if session ID contexts permit); with this option,
5880 session resumption is possible only in the first handshake.
5882 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5883 more bits available for options that should not be part of
5884 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5887 *) Add some demos for certificate and certificate request creation.
5890 *) Make maximum certificate chain size accepted from the peer application
5891 settable (SSL*_get/set_max_cert_list()), as proposed by
5892 "Douglas E. Engert" <deengert@anl.gov>.
5895 *) Add support for shared libraries for Unixware-7
5896 (Boyd Lynn Gerber <gerberb@zenez.com>).
5899 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5900 be done prior to destruction. Use this to unload error strings from
5901 ENGINEs that load their own error strings. NB: This adds two new API
5902 functions to "get" and "set" this destroy handler in an ENGINE.
5905 *) Alter all existing ENGINE implementations (except "openssl" and
5906 "openbsd") to dynamically instantiate their own error strings. This
5907 makes them more flexible to be built both as statically-linked ENGINEs
5908 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5909 Also, add stub code to each that makes building them as self-contained
5910 shared-libraries easier (see README.ENGINE).
5913 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5914 implementations into applications that are completely implemented in
5915 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5916 commands that can be used to configure what shared-library to load and
5917 to control aspects of the way it is handled. Also, made an update to
5918 the README.ENGINE file that brings its information up-to-date and
5919 provides some information and instructions on the "dynamic" ENGINE
5920 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5923 *) Make it possible to unload ranges of ERR strings with a new
5924 "ERR_unload_strings" function.
5927 *) Add a copy() function to EVP_MD.
5930 *) Make EVP_MD routines take a context pointer instead of just the
5931 md_data void pointer.
5934 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5935 that the digest can only process a single chunk of data
5936 (typically because it is provided by a piece of
5937 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5938 is only going to provide a single chunk of data, and hence the
5939 framework needn't accumulate the data for oneshot drivers.
5942 *) As with "ERR", make it possible to replace the underlying "ex_data"
5943 functions. This change also alters the storage and management of global
5944 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5945 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5946 index counters. The API functions that use this state have been changed
5947 to take a "class_index" rather than pointers to the class's local STACK
5948 and counter, and there is now an API function to dynamically create new
5949 classes. This centralisation allows us to (a) plug a lot of the
5950 thread-safety problems that existed, and (b) makes it possible to clean
5951 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5952 such data would previously have always leaked in application code and
5953 workarounds were in place to make the memory debugging turn a blind eye
5954 to it. Application code that doesn't use this new function will still
5955 leak as before, but their memory debugging output will announce it now
5956 rather than letting it slide.
5958 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5959 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5960 has a return value to indicate success or failure.
5963 *) Make it possible to replace the underlying "ERR" functions such that the
5964 global state (2 LHASH tables and 2 locks) is only used by the "default"
5965 implementation. This change also adds two functions to "get" and "set"
5966 the implementation prior to it being automatically set the first time
5967 any other ERR function takes place. Ie. an application can call "get",
5968 pass the return value to a module it has just loaded, and that module
5969 can call its own "set" function using that value. This means the
5970 module's "ERR" operations will use (and modify) the error state in the
5971 application and not in its own statically linked copy of OpenSSL code.
5974 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5975 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5976 the operation, and provides a more encapsulated way for external code
5977 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5978 to use these functions rather than manually incrementing the counts.
5980 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5983 *) Add EVP test program.
5986 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5989 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5990 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5991 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5992 These allow a CRL to be built without having to access X509_CRL fields
5993 directly. Modify 'ca' application to use new functions.
5996 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5997 bug workarounds. Rollback attack detection is a security feature.
5998 The problem will only arise on OpenSSL servers when TLSv1 is not
5999 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6000 Software authors not wanting to support TLSv1 will have special reasons
6001 for their choice and can explicitly enable this option.
6002 [Bodo Moeller, Lutz Jaenicke]
6004 *) Rationalise EVP so it can be extended: don't include a union of
6005 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6006 (similar to those existing for EVP_CIPHER_CTX).
6011 EVP_MD_CTX_init(&md); /* new function call */
6012 EVP_DigestInit(&md, EVP_sha1());
6013 EVP_DigestUpdate(&md, in, len);
6014 EVP_DigestFinal(&md, out, NULL);
6015 EVP_MD_CTX_cleanup(&md); /* new function call */
6019 *) Make DES key schedule conform to the usual scheme, as well as
6020 correcting its structure. This means that calls to DES functions
6021 now have to pass a pointer to a des_key_schedule instead of a
6022 plain des_key_schedule (which was actually always a pointer
6025 des_key_schedule ks;
6027 des_set_key_checked(..., &ks);
6028 des_ncbc_encrypt(..., &ks, ...);
6030 (Note that a later change renames 'des_...' into 'DES_...'.)
6033 *) Initial reduction of linker bloat: the use of some functions, such as
6034 PEM causes large amounts of unused functions to be linked in due to
6035 poor organisation. For example pem_all.c contains every PEM function
6036 which has a knock on effect of linking in large amounts of (unused)
6037 ASN1 code. Grouping together similar functions and splitting unrelated
6038 functions prevents this.
6041 *) Cleanup of EVP macros.
6044 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6045 correct _ecb suffix.
6048 *) Add initial OCSP responder support to ocsp application. The
6049 revocation information is handled using the text based index
6050 use by the ca application. The responder can either handle
6051 requests generated internally, supplied in files (for example
6052 via a CGI script) or using an internal minimal server.
6055 *) Add configuration choices to get zlib compression for TLS.
6058 *) Changes to Kerberos SSL for RFC 2712 compliance:
6059 1. Implemented real KerberosWrapper, instead of just using
6060 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6061 2. Implemented optional authenticator field of KerberosWrapper.
6063 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6064 and authenticator structs; see crypto/krb5/.
6066 Generalized Kerberos calls to support multiple Kerberos libraries.
6067 [Vern Staats <staatsvr@asc.hpc.mil>,
6068 Jeffrey Altman <jaltman@columbia.edu>
6069 via Richard Levitte]
6071 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6072 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6073 values for each of the key sizes rather than having just
6074 parameters (and 'speed' generating keys each time).
6077 *) Speed up EVP routines.
6080 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6081 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6082 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6083 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6085 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6086 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6087 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6090 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6092 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6095 *) Added the OS2-EMX target.
6096 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6098 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6099 to support NCONF routines in extension code. New function CONF_set_nconf()
6100 to allow functions which take an NCONF to also handle the old LHASH
6101 structure: this means that the old CONF compatible routines can be
6102 retained (in particular wrt extensions) without having to duplicate the
6103 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6106 *) Enhance the general user interface with mechanisms for inner control
6107 and with possibilities to have yes/no kind of prompts.
6110 *) Change all calls to low level digest routines in the library and
6111 applications to use EVP. Add missing calls to HMAC_cleanup() and
6112 don't assume HMAC_CTX can be copied using memcpy().
6113 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6115 *) Add the possibility to control engines through control names but with
6116 arbitrary arguments instead of just a string.
6117 Change the key loaders to take a UI_METHOD instead of a callback
6118 function pointer. NOTE: this breaks binary compatibility with earlier
6119 versions of OpenSSL [engine].
6120 Adapt the nCipher code for these new conditions and add a card insertion
6124 *) Enhance the general user interface with mechanisms to better support
6125 dialog box interfaces, application-defined prompts, the possibility
6126 to use defaults (for example default passwords from somewhere else)
6127 and interrupts/cancellations.
6130 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6131 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6134 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6135 tidy up some unnecessarily weird code in 'sk_new()').
6136 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6138 *) Change the key loading routines for ENGINEs to use the same kind
6139 callback (pem_password_cb) as all other routines that need this
6143 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6144 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6145 than this minimum value is recommended.
6148 *) New random seeder for OpenVMS, using the system process statistics
6149 that are easily reachable.
6152 *) Windows apparently can't transparently handle global
6153 variables defined in DLLs. Initialisations such as:
6155 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6157 wont compile. This is used by the any applications that need to
6158 declare their own ASN1 modules. This was fixed by adding the option
6159 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6160 needed for static libraries under Win32.
6163 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6164 setting of purpose and trust fields. New X509_STORE trust and
6165 purpose functions and tidy up setting in other SSL functions.
6168 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6169 structure. These are inherited by X509_STORE_CTX when it is
6170 initialised. This allows various defaults to be set in the
6171 X509_STORE structure (such as flags for CRL checking and custom
6172 purpose or trust settings) for functions which only use X509_STORE_CTX
6173 internally such as S/MIME.
6175 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6176 trust settings if they are not set in X509_STORE. This allows X509_STORE
6177 purposes and trust (in S/MIME for example) to override any set by default.
6179 Add command line options for CRL checking to smime, s_client and s_server
6183 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6184 are set then the CRL is looked up in the X509_STORE structure and
6185 its validity and signature checked, then if the certificate is found
6186 in the CRL the verify fails with a revoked error.
6188 Various new CRL related callbacks added to X509_STORE_CTX structure.
6190 Command line options added to 'verify' application to support this.
6192 This needs some additional work, such as being able to handle multiple
6193 CRLs with different times, extension based lookup (rather than just
6194 by subject name) and ultimately more complete V2 CRL extension
6198 *) Add a general user interface API (crypto/ui/). This is designed
6199 to replace things like des_read_password and friends (backward
6200 compatibility functions using this new API are provided).
6201 The purpose is to remove prompting functions from the DES code
6202 section as well as provide for prompting through dialog boxes in
6203 a window system and the like.
6206 *) Add "ex_data" support to ENGINE so implementations can add state at a
6207 per-structure level rather than having to store it globally.
6210 *) Make it possible for ENGINE structures to be copied when retrieved by
6211 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6212 This causes the "original" ENGINE structure to act like a template,
6213 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6214 operational state can be localised to each ENGINE structure, despite the
6215 fact they all share the same "methods". New ENGINE structures returned in
6216 this case have no functional references and the return value is the single
6217 structural reference. This matches the single structural reference returned
6218 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6222 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6223 needs to match any other type at all we need to manually clear the
6227 *) Changes to the "openssl engine" utility to include;
6228 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6229 about an ENGINE's available control commands.
6230 - executing control commands from command line arguments using the
6231 '-pre' and '-post' switches. '-post' is only used if '-t' is
6232 specified and the ENGINE is successfully initialised. The syntax for
6233 the individual commands are colon-separated, for example;
6234 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6237 *) New dynamic control command support for ENGINEs. ENGINEs can now
6238 declare their own commands (numbers), names (strings), descriptions,
6239 and input types for run-time discovery by calling applications. A
6240 subset of these commands are implicitly classed as "executable"
6241 depending on their input type, and only these can be invoked through
6242 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6243 can be based on user input, config files, etc). The distinction is
6244 that "executable" commands cannot return anything other than a boolean
6245 result and can only support numeric or string input, whereas some
6246 discoverable commands may only be for direct use through
6247 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6248 pointers, or other custom uses. The "executable" commands are to
6249 support parameterisations of ENGINE behaviour that can be
6250 unambiguously defined by ENGINEs and used consistently across any
6251 OpenSSL-based application. Commands have been added to all the
6252 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6253 control over shared-library paths without source code alterations.
6256 *) Changed all ENGINE implementations to dynamically allocate their
6257 ENGINEs rather than declaring them statically. Apart from this being
6258 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6259 this also allows the implementations to compile without using the
6260 internal engine_int.h header.
6263 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6264 'const' value. Any code that should be able to modify a RAND_METHOD
6265 should already have non-const pointers to it (ie. they should only
6266 modify their own ones).
6269 *) Made a variety of little tweaks to the ENGINE code.
6270 - "atalla" and "ubsec" string definitions were moved from header files
6271 to C code. "nuron" string definitions were placed in variables
6272 rather than hard-coded - allowing parameterisation of these values
6273 later on via ctrl() commands.
6274 - Removed unused "#if 0"'d code.
6275 - Fixed engine list iteration code so it uses ENGINE_free() to release
6276 structural references.
6277 - Constified the RAND_METHOD element of ENGINE structures.
6278 - Constified various get/set functions as appropriate and added
6279 missing functions (including a catch-all ENGINE_cpy that duplicates
6280 all ENGINE values onto a new ENGINE except reference counts/state).
6281 - Removed NULL parameter checks in get/set functions. Setting a method
6282 or function to NULL is a way of cancelling out a previously set
6283 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6284 and doesn't justify the extra error symbols and code.
6285 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6286 flags from engine_int.h to engine.h.
6287 - Changed prototypes for ENGINE handler functions (init(), finish(),
6288 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6291 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6292 to the algorithm using long division. The binary algorithm can be
6293 used only if the modulus is odd. On 32-bit systems, it is faster
6294 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6295 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6296 up to 450 bits. In 64-bit environments, the binary algorithm
6297 appears to be advantageous for much longer moduli; here we use it
6298 for moduli up to 2048 bits.
6301 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6302 could not support the combine flag in choice fields.
6305 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6306 extensions from a certificate request to the certificate.
6309 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6310 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6311 file: this allows the display of the certificate about to be
6312 signed to be customised, to allow certain fields to be included
6313 or excluded and extension details. The old system didn't display
6314 multicharacter strings properly, omitted fields not in the policy
6315 and couldn't display additional details such as extensions.
6318 *) Function EC_POINTs_mul for multiple scalar multiplication
6319 of an arbitrary number of elliptic curve points
6320 \sum scalars[i]*points[i],
6321 optionally including the generator defined for the EC_GROUP:
6322 scalar*generator + \sum scalars[i]*points[i].
6324 EC_POINT_mul is a simple wrapper function for the typical case
6325 that the point list has just one item (besides the optional
6329 *) First EC_METHODs for curves over GF(p):
6331 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6332 operations and provides various method functions that can also
6333 operate with faster implementations of modular arithmetic.
6335 EC_GFp_mont_method() reuses most functions that are part of
6336 EC_GFp_simple_method, but uses Montgomery arithmetic.
6338 [Bodo Moeller; point addition and point doubling
6339 implementation directly derived from source code provided by
6340 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6342 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6343 crypto/ec/ec_lib.c):
6345 Curves are EC_GROUP objects (with an optional group generator)
6346 based on EC_METHODs that are built into the library.
6348 Points are EC_POINT objects based on EC_GROUP objects.
6350 Most of the framework would be able to handle curves over arbitrary
6351 finite fields, but as there are no obvious types for fields other
6352 than GF(p), some functions are limited to that for now.
6355 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6356 that the file contains a complete HTTP response.
6359 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6360 change the def and num file printf format specifier from "%-40sXXX"
6361 to "%-39s XXX". The latter will always guarantee a space after the
6362 field while the former will cause them to run together if the field
6363 is 40 of more characters long.
6366 *) Constify the cipher and digest 'method' functions and structures
6367 and modify related functions to take constant EVP_MD and EVP_CIPHER
6371 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6372 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6375 *) Modify EVP_Digest*() routines so they now return values. Although the
6376 internal software routines can never fail additional hardware versions
6380 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6382 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6383 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6386 ERR_R_NESTED_ASN1_ERROR
6388 ERR_R_MISSING_ASN1_EOS
6389 were 4 .. 9, conflicting with
6390 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6392 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6393 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6395 Add new error code 'ERR_R_INTERNAL_ERROR'.
6398 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6402 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6403 sets the subject name for a new request or supersedes the
6404 subject name in a given request. Formats that can be parsed are
6405 'CN=Some Name, OU=myOU, C=IT'
6407 'CN=Some Name/OU=myOU/C=IT'.
6409 Add options '-batch' and '-verbose' to 'openssl req'.
6410 [Massimiliano Pala <madwolf@hackmasters.net>]
6412 *) Introduce the possibility to access global variables through
6413 functions on platform were that's the best way to handle exporting
6414 global variables in shared libraries. To enable this functionality,
6415 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6416 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6417 is normally done by Configure or something similar).
6419 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6420 in the source file (foo.c) like this:
6422 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6423 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6425 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6426 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6428 OPENSSL_DECLARE_GLOBAL(int,foo);
6429 #define foo OPENSSL_GLOBAL_REF(foo)
6430 OPENSSL_DECLARE_GLOBAL(double,bar);
6431 #define bar OPENSSL_GLOBAL_REF(bar)
6433 The #defines are very important, and therefore so is including the
6434 header file everywhere where the defined globals are used.
6436 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6437 of ASN.1 items, but that structure is a bit different.
6439 The largest change is in util/mkdef.pl which has been enhanced with
6440 better and easier to understand logic to choose which symbols should
6441 go into the Windows .def files as well as a number of fixes and code
6442 cleanup (among others, algorithm keywords are now sorted
6443 lexicographically to avoid constant rewrites).
6446 *) In BN_div() keep a copy of the sign of 'num' before writing the
6447 result to 'rm' because if rm==num the value will be overwritten
6448 and produce the wrong result if 'num' is negative: this caused
6449 problems with BN_mod() and BN_nnmod().
6452 *) Function OCSP_request_verify(). This checks the signature on an
6453 OCSP request and verifies the signer certificate. The signer
6454 certificate is just checked for a generic purpose and OCSP request
6458 *) Add OCSP_check_validity() function to check the validity of OCSP
6459 responses. OCSP responses are prepared in real time and may only
6460 be a few seconds old. Simply checking that the current time lies
6461 between thisUpdate and nextUpdate max reject otherwise valid responses
6462 caused by either OCSP responder or client clock inaccuracy. Instead
6463 we allow thisUpdate and nextUpdate to fall within a certain period of
6464 the current time. The age of the response can also optionally be
6465 checked. Two new options -validity_period and -status_age added to
6469 *) If signature or public key algorithm is unrecognized print out its
6470 OID rather that just UNKNOWN.
6473 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6474 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6475 ID to be generated from the issuer certificate alone which can then be
6476 passed to OCSP_id_issuer_cmp().
6479 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6480 ASN1 modules to export functions returning ASN1_ITEM pointers
6481 instead of the ASN1_ITEM structures themselves. This adds several
6482 new macros which allow the underlying ASN1 function/structure to
6483 be accessed transparently. As a result code should not use ASN1_ITEM
6484 references directly (such as &X509_it) but instead use the relevant
6485 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6486 use of the new ASN1 code on platforms where exporting structures
6487 is problematical (for example in shared libraries) but exporting
6488 functions returning pointers to structures is not.
6491 *) Add support for overriding the generation of SSL/TLS session IDs.
6492 These callbacks can be registered either in an SSL_CTX or per SSL.
6493 The purpose of this is to allow applications to control, if they wish,
6494 the arbitrary values chosen for use as session IDs, particularly as it
6495 can be useful for session caching in multiple-server environments. A
6496 command-line switch for testing this (and any client code that wishes
6497 to use such a feature) has been added to "s_server".
6498 [Geoff Thorpe, Lutz Jaenicke]
6500 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6501 of the form '#if defined(...) || defined(...) || ...' and
6502 '#if !defined(...) && !defined(...) && ...'. This also avoids
6503 the growing number of special cases it was previously handling.
6506 *) Make all configuration macros available for application by making
6507 sure they are available in opensslconf.h, by giving them names starting
6508 with "OPENSSL_" to avoid conflicts with other packages and by making
6509 sure e_os2.h will cover all platform-specific cases together with
6511 Additionally, it is now possible to define configuration/platform-
6512 specific names (called "system identities"). In the C code, these
6513 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6514 macro with the name beginning with "OPENSSL_SYS_", which is determined
6515 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6519 *) New option -set_serial to 'req' and 'x509' this allows the serial
6520 number to use to be specified on the command line. Previously self
6521 signed certificates were hard coded with serial number 0 and the
6522 CA options of 'x509' had to use a serial number in a file which was
6526 *) New options to 'ca' utility to support V2 CRL entry extensions.
6527 Currently CRL reason, invalidity date and hold instruction are
6528 supported. Add new CRL extensions to V3 code and some new objects.
6531 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6532 disable standard block padding (aka PKCS#5 padding) in the EVP
6533 API, which was previously mandatory. This means that the data is
6534 not padded in any way and so the total length much be a multiple
6535 of the block size, otherwise an error occurs.
6538 *) Initial (incomplete) OCSP SSL support.
6541 *) New function OCSP_parse_url(). This splits up a URL into its host,
6542 port and path components: primarily to parse OCSP URLs. New -url
6543 option to ocsp utility.
6546 *) New nonce behavior. The return value of OCSP_check_nonce() now
6547 reflects the various checks performed. Applications can decide
6548 whether to tolerate certain situations such as an absent nonce
6549 in a response when one was present in a request: the ocsp application
6550 just prints out a warning. New function OCSP_add1_basic_nonce()
6551 this is to allow responders to include a nonce in a response even if
6552 the request is nonce-less.
6555 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6556 skipped when using openssl x509 multiple times on a single input file,
6557 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6560 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6561 set string type: to handle setting ASN1_TIME structures. Fix ca
6562 utility to correctly initialize revocation date of CRLs.
6565 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6566 the clients preferred ciphersuites and rather use its own preferences.
6567 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6568 Internet Explorer by ensuring unchanged hash method during stepup.
6569 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6572 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6573 to aes and add a new 'exist' option to print out symbols that don't
6577 *) Additional options to ocsp utility to allow flags to be set and
6578 additional certificates supplied.
6581 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6582 OCSP client a number of certificate to only verify the response
6586 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6587 handle the new API. Currently only ECB, CBC modes supported. Add new
6590 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6591 Encryption Standard (AES) Ciphersuites for Transport Layer
6592 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6593 not enabled by default and were not part of the "ALL" ciphersuite
6594 alias because they were not yet official; they could be
6595 explicitly requested by specifying the "AESdraft" ciphersuite
6596 group alias. In the final release of OpenSSL 0.9.7, the group
6597 alias is called "AES" and is part of "ALL".)
6598 [Ben Laurie, Steve Henson, Bodo Moeller]
6600 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6601 request to response.
6604 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6605 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6606 extract information from a certificate request. OCSP_response_create()
6607 creates a response and optionally adds a basic response structure.
6608 OCSP_basic_add1_status() adds a complete single response to a basic
6609 response and returns the OCSP_SINGLERESP structure just added (to allow
6610 extensions to be included for example). OCSP_basic_add1_cert() adds a
6611 certificate to a basic response and OCSP_basic_sign() signs a basic
6612 response with various flags. New helper functions ASN1_TIME_check()
6613 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6614 (converts ASN1_TIME to GeneralizedTime).
6617 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6618 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6619 structure from a certificate. X509_pubkey_digest() digests the public_key
6620 contents: this is used in various key identifiers.
6623 *) Make sk_sort() tolerate a NULL argument.
6624 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6626 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6627 passed by the function are trusted implicitly. If any of them signed the
6628 response then it is assumed to be valid and is not verified.
6631 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6632 to data. This was previously part of the PKCS7 ASN1 code. This
6633 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6634 [Steve Henson, reported by Kenneth R. Robinette
6635 <support@securenetterm.com>]
6637 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6638 routines: without these tracing memory leaks is very painful.
6639 Fix leaks in PKCS12 and PKCS7 routines.
6642 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6643 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6644 effectively meant GeneralizedTime would never be used. Now it
6645 is initialised to -1 but X509_time_adj() now has to check the value
6646 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6647 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6648 [Steve Henson, reported by Kenneth R. Robinette
6649 <support@securenetterm.com>]
6651 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6652 result in a zero length in the ASN1_INTEGER structure which was
6653 not consistent with the structure when d2i_ASN1_INTEGER() was used
6654 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6655 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6656 where it did not print out a minus for negative ASN1_INTEGER.
6659 *) Add summary printout to ocsp utility. The various functions which
6660 convert status values to strings have been renamed to:
6661 OCSP_response_status_str(), OCSP_cert_status_str() and
6662 OCSP_crl_reason_str() and are no longer static. New options
6663 to verify nonce values and to disable verification. OCSP response
6664 printout format cleaned up.
6667 *) Add additional OCSP certificate checks. These are those specified
6668 in RFC2560. This consists of two separate checks: the CA of the
6669 certificate being checked must either be the OCSP signer certificate
6670 or the issuer of the OCSP signer certificate. In the latter case the
6671 OCSP signer certificate must contain the OCSP signing extended key
6672 usage. This check is performed by attempting to match the OCSP
6673 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6674 in the OCSP_CERTID structures of the response.
6677 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6678 and related routines. This uses the standard OpenSSL certificate
6679 verify routines to perform initial checks (just CA validity) and
6680 to obtain the certificate chain. Then additional checks will be
6681 performed on the chain. Currently the root CA is checked to see
6682 if it is explicitly trusted for OCSP signing. This is used to set
6683 a root CA as a global signing root: that is any certificate that
6684 chains to that CA is an acceptable OCSP signing certificate.
6687 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6688 extensions from a separate configuration file.
6689 As when reading extensions from the main configuration file,
6690 the '-extensions ...' option may be used for specifying the
6692 [Massimiliano Pala <madwolf@comune.modena.it>]
6694 *) New OCSP utility. Allows OCSP requests to be generated or
6695 read. The request can be sent to a responder and the output
6696 parsed, outputed or printed in text form. Not complete yet:
6697 still needs to check the OCSP response validity.
6700 *) New subcommands for 'openssl ca':
6701 'openssl ca -status <serial>' prints the status of the cert with
6702 the given serial number (according to the index file).
6703 'openssl ca -updatedb' updates the expiry status of certificates
6705 [Massimiliano Pala <madwolf@comune.modena.it>]
6707 *) New '-newreq-nodes' command option to CA.pl. This is like
6708 '-newreq', but calls 'openssl req' with the '-nodes' option
6709 so that the resulting key is not encrypted.
6710 [Damien Miller <djm@mindrot.org>]
6712 *) New configuration for the GNU Hurd.
6713 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6715 *) Initial code to implement OCSP basic response verify. This
6716 is currently incomplete. Currently just finds the signer's
6717 certificate and verifies the signature on the response.
6720 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6721 value of OPENSSLDIR. This is available via the new '-d' option
6722 to 'openssl version', and is also included in 'openssl version -a'.
6725 *) Allowing defining memory allocation callbacks that will be given
6726 file name and line number information in additional arguments
6727 (a const char* and an int). The basic functionality remains, as
6728 well as the original possibility to just replace malloc(),
6729 realloc() and free() by functions that do not know about these
6730 additional arguments. To register and find out the current
6731 settings for extended allocation functions, the following
6732 functions are provided:
6734 CRYPTO_set_mem_ex_functions
6735 CRYPTO_set_locked_mem_ex_functions
6736 CRYPTO_get_mem_ex_functions
6737 CRYPTO_get_locked_mem_ex_functions
6739 These work the same way as CRYPTO_set_mem_functions and friends.
6740 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6741 extended allocation function is enabled.
6742 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6743 a conventional allocation function is enabled.
6744 [Richard Levitte, Bodo Moeller]
6746 *) Finish off removing the remaining LHASH function pointer casts.
6747 There should no longer be any prototype-casting required when using
6748 the LHASH abstraction, and any casts that remain are "bugs". See
6749 the callback types and macros at the head of lhash.h for details
6750 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6753 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6754 If /dev/[u]random devices are not available or do not return enough
6755 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6757 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6758 /etc/entropy will be queried once each in this sequence, quering stops
6759 when enough entropy was collected without querying more sockets.
6762 *) Change the Unix RAND_poll() variant to be able to poll several
6763 random devices, as specified by DEVRANDOM, until a sufficient amount
6764 of data has been collected. We spend at most 10 ms on each file
6765 (select timeout) and read in non-blocking mode. DEVRANDOM now
6766 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6767 (previously it was just the string "/dev/urandom"), so on typical
6768 platforms the 10 ms delay will never occur.
6769 Also separate out the Unix variant to its own file, rand_unix.c.
6770 For VMS, there's a currently-empty rand_vms.c.
6773 *) Move OCSP client related routines to ocsp_cl.c. These
6774 provide utility functions which an application needing
6775 to issue a request to an OCSP responder and analyse the
6776 response will typically need: as opposed to those which an
6777 OCSP responder itself would need which will be added later.
6779 OCSP_request_sign() signs an OCSP request with an API similar
6780 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6781 response. OCSP_response_get1_basic() extracts basic response
6782 from response. OCSP_resp_find_status(): finds and extracts status
6783 information from an OCSP_CERTID structure (which will be created
6784 when the request structure is built). These are built from lower
6785 level functions which work on OCSP_SINGLERESP structures but
6786 wont normally be used unless the application wishes to examine
6787 extensions in the OCSP response for example.
6789 Replace nonce routines with a pair of functions.
6790 OCSP_request_add1_nonce() adds a nonce value and optionally
6791 generates a random value. OCSP_check_nonce() checks the
6792 validity of the nonce in an OCSP response.
6795 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6796 This doesn't copy the supplied OCSP_CERTID and avoids the
6797 need to free up the newly created id. Change return type
6798 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6799 This can then be used to add extensions to the request.
6800 Deleted OCSP_request_new(), since most of its functionality
6801 is now in OCSP_REQUEST_new() (and the case insensitive name
6802 clash) apart from the ability to set the request name which
6803 will be added elsewhere.
6806 *) Update OCSP API. Remove obsolete extensions argument from
6807 various functions. Extensions are now handled using the new
6808 OCSP extension code. New simple OCSP HTTP function which
6809 can be used to send requests and parse the response.
6812 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6813 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6814 uses the special reorder version of SET OF to sort the attributes
6815 and reorder them to match the encoded order. This resolves a long
6816 standing problem: a verify on a PKCS7 structure just after signing
6817 it used to fail because the attribute order did not match the
6818 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6819 it uses the received order. This is necessary to tolerate some broken
6820 software that does not order SET OF. This is handled by encoding
6821 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6822 to produce the required SET OF.
6825 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6826 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6827 files to get correct declarations of the ASN.1 item variables.
6830 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6831 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6832 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6833 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6834 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6835 ASN1_ITEM and no wrapper functions.
6838 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6839 replace the old function pointer based I/O routines. Change most of
6840 the *_d2i_bio() and *_d2i_fp() functions to use these.
6843 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6844 lines, recognice more "algorithms" that can be deselected, and make
6845 it complain about algorithm deselection that isn't recognised.
6848 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6849 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6850 to use new functions. Add NO_ASN1_OLD which can be set to remove
6851 some old style ASN1 functions: this can be used to determine if old
6852 code will still work when these eventually go away.
6855 *) New extension functions for OCSP structures, these follow the
6856 same conventions as certificates and CRLs.
6859 *) New function X509V3_add1_i2d(). This automatically encodes and
6860 adds an extension. Its behaviour can be customised with various
6861 flags to append, replace or delete. Various wrappers added for
6862 certifcates and CRLs.
6865 *) Fix to avoid calling the underlying ASN1 print routine when
6866 an extension cannot be parsed. Correct a typo in the
6867 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6870 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6871 entries for variables.
6874 *) Add functionality to apps/openssl.c for detecting locking
6875 problems: As the program is single-threaded, all we have
6876 to do is register a locking callback using an array for
6877 storing which locks are currently held by the program.
6880 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6881 SSL_get_ex_data_X509_STORE_idx(), which is used in
6882 ssl_verify_cert_chain() and thus can be called at any time
6883 during TLS/SSL handshakes so that thread-safety is essential.
6884 Unfortunately, the ex_data design is not at all suited
6885 for multi-threaded use, so it probably should be abolished.
6888 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6889 [Broadcom, tweaked and integrated by Geoff Thorpe]
6891 *) Move common extension printing code to new function
6892 X509V3_print_extensions(). Reorganise OCSP print routines and
6893 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6896 *) New function X509_signature_print() to remove duplication in some
6900 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6901 set (this was treated exactly the same as SET OF previously). This
6902 is used to reorder the STACK representing the structure to match the
6903 encoding. This will be used to get round a problem where a PKCS7
6904 structure which was signed could not be verified because the STACK
6905 order did not reflect the encoded order.
6908 *) Reimplement the OCSP ASN1 module using the new code.
6911 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6912 for its ASN1 operations. The old style function pointers still exist
6913 for now but they will eventually go away.
6916 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6917 completely replaces the old ASN1 functionality with a table driven
6918 encoder and decoder which interprets an ASN1_ITEM structure describing
6919 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6920 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6921 has also been converted to the new form.
6924 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6925 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6926 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6927 for negative moduli.
6930 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6931 of not touching the result's sign bit.
6934 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6938 *) Changed the LHASH code to use prototypes for callbacks, and created
6939 macros to declare and implement thin (optionally static) functions
6940 that provide type-safety and avoid function pointer casting for the
6941 type-specific callbacks.
6944 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6946 [Veers Staats <staatsvr@asc.hpc.mil>,
6947 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6949 *) Reformat the FAQ so the different questions and answers can be divided
6950 in sections depending on the subject.
6953 *) Have the zlib compression code load ZLIB.DLL dynamically under
6957 *) New function BN_mod_sqrt for computing square roots modulo a prime
6958 (using the probabilistic Tonelli-Shanks algorithm unless
6959 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6960 be handled deterministically).
6961 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6963 *) Make BN_mod_inverse faster by explicitly handling small quotients
6964 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6965 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6968 *) New function BN_kronecker.
6971 *) Fix BN_gcd so that it works on negative inputs; the result is
6972 positive unless both parameters are zero.
6973 Previously something reasonably close to an infinite loop was
6974 possible because numbers could be growing instead of shrinking
6975 in the implementation of Euclid's algorithm.
6978 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6979 sign of the number in question.
6981 Fix BN_is_word(a,w) to work correctly for w == 0.
6983 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6984 because its test if the absolute value of 'a' equals 'w'.
6985 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6986 it exists mostly for use in the implementations of BN_is_zero(),
6987 BN_is_one(), and BN_is_word().
6990 *) New function BN_swap.
6993 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6994 the exponentiation functions are more likely to produce reasonable
6995 results on negative inputs.
6998 *) Change BN_mod_mul so that the result is always non-negative.
6999 Previously, it could be negative if one of the factors was negative;
7000 I don't think anyone really wanted that behaviour.
7003 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7004 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7005 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7006 and add new functions:
7015 BN_mod_lshift1_quick
7019 These functions always generate non-negative results.
7021 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7022 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7024 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7025 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7026 be reduced modulo m.
7027 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7030 The following entry accidentally appeared in the CHANGES file
7031 distributed with OpenSSL 0.9.7. The modifications described in
7032 it do *not* apply to OpenSSL 0.9.7.
7034 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7035 was actually never needed) and in BN_mul(). The removal in BN_mul()
7036 required a small change in bn_mul_part_recursive() and the addition
7037 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7038 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7039 bn_sub_words() and bn_add_words() except they take arrays with
7044 *) In 'openssl passwd', verify passwords read from the terminal
7045 unless the '-salt' option is used (which usually means that
7046 verification would just waste user's time since the resulting
7047 hash is going to be compared with some given password hash)
7048 or the new '-noverify' option is used.
7050 This is an incompatible change, but it does not affect
7051 non-interactive use of 'openssl passwd' (passwords on the command
7052 line, '-stdin' option, '-in ...' option) and thus should not
7056 *) Remove all references to RSAref, since there's no more need for it.
7059 *) Make DSO load along a path given through an environment variable
7060 (SHLIB_PATH) with shl_load().
7063 *) Constify the ENGINE code as a result of BIGNUM constification.
7064 Also constify the RSA code and most things related to it. In a
7065 few places, most notable in the depth of the ASN.1 code, ugly
7066 casts back to non-const were required (to be solved at a later
7070 *) Make it so the openssl application has all engines loaded by default.
7073 *) Constify the BIGNUM routines a little more.
7076 *) Add the following functions:
7078 ENGINE_load_cswift()
7080 ENGINE_load_atalla()
7082 ENGINE_load_builtin_engines()
7084 That way, an application can itself choose if external engines that
7085 are built-in in OpenSSL shall ever be used or not. The benefit is
7086 that applications won't have to be linked with libdl or other dso
7087 libraries unless it's really needed.
7089 Changed 'openssl engine' to load all engines on demand.
7090 Changed the engine header files to avoid the duplication of some
7091 declarations (they differed!).
7094 *) 'openssl engine' can now list capabilities.
7097 *) Better error reporting in 'openssl engine'.
7100 *) Never call load_dh_param(NULL) in s_server.
7103 *) Add engine application. It can currently list engines by name and
7104 identity, and test if they are actually available.
7107 *) Improve RPM specification file by forcing symbolic linking and making
7108 sure the installed documentation is also owned by root.root.
7109 [Damien Miller <djm@mindrot.org>]
7111 *) Give the OpenSSL applications more possibilities to make use of
7112 keys (public as well as private) handled by engines.
7115 *) Add OCSP code that comes from CertCo.
7118 *) Add VMS support for the Rijndael code.
7121 *) Added untested support for Nuron crypto accelerator.
7124 *) Add support for external cryptographic devices. This code was
7125 previously distributed separately as the "engine" branch.
7126 [Geoff Thorpe, Richard Levitte]
7128 *) Rework the filename-translation in the DSO code. It is now possible to
7129 have far greater control over how a "name" is turned into a filename
7130 depending on the operating environment and any oddities about the
7131 different shared library filenames on each system.
7134 *) Support threads on FreeBSD-elf in Configure.
7137 *) Fix for SHA1 assembly problem with MASM: it produces
7138 warnings about corrupt line number information when assembling
7139 with debugging information. This is caused by the overlapping
7141 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7144 NCONF_get_number() has no error checking at all. As a replacement,
7145 NCONF_get_number_e() is defined (_e for "error checking") and is
7146 promoted strongly. The old NCONF_get_number is kept around for
7147 binary backward compatibility.
7148 Make it possible for methods to load from something other than a BIO,
7149 by providing a function pointer that is given a name instead of a BIO.
7150 For example, this could be used to load configuration data from an
7154 *) Fix for non blocking accept BIOs. Added new I/O special reason
7155 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7156 with non blocking I/O was not possible because no retry code was
7157 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7161 *) Added the beginnings of Rijndael support.
7164 *) Fix for bug in DirectoryString mask setting. Add support for
7165 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7166 to allow certificate printing to more controllable, additional
7167 'certopt' option to 'x509' to allow new printing options to be
7171 *) Clean old EAY MD5 hack from e_os.h.
7174 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7176 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7177 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7178 [Joe Orton, Steve Henson]
7180 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7182 *) Fix additional bug revealed by the NISCC test suite:
7184 Stop bug triggering large recursion when presented with
7185 certain ASN.1 tags (CVE-2003-0851)
7188 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7190 *) Fix various bugs revealed by running the NISCC test suite:
7192 Stop out of bounds reads in the ASN1 code when presented with
7193 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7195 If verify callback ignores invalid public key errors don't try to check
7196 certificate signature with the NULL public key.
7200 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7201 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7205 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7206 extra data after the compression methods not only for TLS 1.0
7207 but also for SSL 3.0 (as required by the specification).
7208 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7210 *) Change X509_certificate_type() to mark the key as exported/exportable
7211 when it's 512 *bits* long, not 512 bytes.
7214 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7216 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7217 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7218 a protocol version number mismatch like a decryption error
7219 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7222 *) Turn on RSA blinding by default in the default implementation
7223 to avoid a timing attack. Applications that don't want it can call
7224 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7225 They would be ill-advised to do so in most cases.
7226 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7228 *) Change RSA blinding code so that it works when the PRNG is not
7229 seeded (in this case, the secret RSA exponent is abused as
7230 an unpredictable seed -- if it is not unpredictable, there
7231 is no point in blinding anyway). Make RSA blinding thread-safe
7232 by remembering the creator's thread ID in rsa->blinding and
7233 having all other threads use local one-time blinding factors
7234 (this requires more computation than sharing rsa->blinding, but
7235 avoids excessive locking; and if an RSA object is not shared
7236 between threads, blinding will still be very fast).
7239 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7241 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7242 via timing by performing a MAC computation even if incorrrect
7243 block cipher padding has been found. This is a countermeasure
7244 against active attacks where the attacker has to distinguish
7245 between bad padding and a MAC verification error. (CVE-2003-0078)
7247 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7248 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7249 Martin Vuagnoux (EPFL, Ilion)]
7251 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7253 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7254 memory from it's contents. This is done with a counter that will
7255 place alternating values in each byte. This can be used to solve
7256 two issues: 1) the removal of calls to memset() by highly optimizing
7257 compilers, and 2) cleansing with other values than 0, since those can
7258 be read through on certain media, for example a swap space on disk.
7261 *) Bugfix: client side session caching did not work with external caching,
7262 because the session->cipher setting was not restored when reloading
7263 from the external cache. This problem was masked, when
7264 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7265 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7268 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7269 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7270 [Zeev Lieber <zeev-l@yahoo.com>]
7272 *) Undo an undocumented change introduced in 0.9.6e which caused
7273 repeated calls to OpenSSL_add_all_ciphers() and
7274 OpenSSL_add_all_digests() to be ignored, even after calling
7278 *) Change the default configuration reader to deal with last line not
7279 being properly terminated.
7282 *) Change X509_NAME_cmp() so it applies the special rules on handling
7283 DN values that are of type PrintableString, as well as RDNs of type
7284 emailAddress where the value has the type ia5String.
7285 [stefank@valicert.com via Richard Levitte]
7287 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7288 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7289 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7290 the bitwise-OR of the two for use by the majority of applications
7291 wanting this behaviour, and update the docs. The documented
7292 behaviour and actual behaviour were inconsistent and had been
7293 changing anyway, so this is more a bug-fix than a behavioural
7295 [Geoff Thorpe, diagnosed by Nadav Har'El]
7297 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7298 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7301 *) Fix initialization code race conditions in
7302 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7303 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7304 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7305 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7306 ssl2_get_cipher_by_char(),
7307 ssl3_get_cipher_by_char().
7308 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7310 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7311 the cached sessions are flushed, as the remove_cb() might use ex_data
7312 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7313 (see [openssl.org #212]).
7314 [Geoff Thorpe, Lutz Jaenicke]
7316 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7317 length, instead of the encoding length to d2i_ASN1_OBJECT.
7320 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7322 *) [In 0.9.6g-engine release:]
7323 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7324 [Lynn Gazis <lgazis@rainbow.com>]
7326 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7328 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7329 and get fix the header length calculation.
7330 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7331 Alon Kantor <alonk@checkpoint.com> (and others),
7334 *) Use proper error handling instead of 'assertions' in buffer
7335 overflow checks added in 0.9.6e. This prevents DoS (the
7336 assertions could call abort()).
7337 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7339 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7341 *) Add various sanity checks to asn1_get_length() to reject
7342 the ASN1 length bytes if they exceed sizeof(long), will appear
7343 negative or the content length exceeds the length of the
7345 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7347 *) Fix cipher selection routines: ciphers without encryption had no flags
7348 for the cipher strength set and where therefore not handled correctly
7349 by the selection routines (PR #130).
7352 *) Fix EVP_dsa_sha macro.
7356 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7357 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7358 that was added in OpenSSL 0.9.6d.
7360 As the countermeasure turned out to be incompatible with some
7361 broken SSL implementations, the new option is part of SSL_OP_ALL.
7362 SSL_OP_ALL is usually employed when compatibility with weird SSL
7363 implementations is desired (e.g. '-bugs' option to 's_client' and
7364 's_server'), so the new option is automatically set in many
7368 *) Changes in security patch:
7370 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7371 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7372 Air Force Materiel Command, USAF, under agreement number
7375 *) Add various sanity checks to asn1_get_length() to reject
7376 the ASN1 length bytes if they exceed sizeof(long), will appear
7377 negative or the content length exceeds the length of the
7378 supplied buffer. (CVE-2002-0659)
7379 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7381 *) Assertions for various potential buffer overflows, not known to
7383 [Ben Laurie (CHATS)]
7385 *) Various temporary buffers to hold ASCII versions of integers were
7386 too small for 64 bit platforms. (CVE-2002-0655)
7387 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7389 *) Remote buffer overflow in SSL3 protocol - an attacker could
7390 supply an oversized session ID to a client. (CVE-2002-0656)
7391 [Ben Laurie (CHATS)]
7393 *) Remote buffer overflow in SSL2 protocol - an attacker could
7394 supply an oversized client master key. (CVE-2002-0656)
7395 [Ben Laurie (CHATS)]
7397 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7399 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7400 encoded as NULL) with id-dsa-with-sha1.
7401 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7403 *) Check various X509_...() return values in apps/req.c.
7404 [Nils Larsch <nla@trustcenter.de>]
7406 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7407 an end-of-file condition would erronously be flagged, when the CRLF
7408 was just at the end of a processed block. The bug was discovered when
7409 processing data through a buffering memory BIO handing the data to a
7410 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7411 <ptsekov@syntrex.com> and Nedelcho Stanev.
7414 *) Implement a countermeasure against a vulnerability recently found
7415 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7416 before application data chunks to avoid the use of known IVs
7417 with data potentially chosen by the attacker.
7420 *) Fix length checks in ssl3_get_client_hello().
7423 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7424 to prevent ssl3_read_internal() from incorrectly assuming that
7425 ssl3_read_bytes() found application data while handshake
7426 processing was enabled when in fact s->s3->in_read_app_data was
7427 merely automatically cleared during the initial handshake.
7428 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7430 *) Fix object definitions for Private and Enterprise: they were not
7431 recognized in their shortname (=lowercase) representation. Extend
7432 obj_dat.pl to issue an error when using undefined keywords instead
7433 of silently ignoring the problem (Svenning Sorensen
7434 <sss@sss.dnsalias.net>).
7437 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7438 generators, i.e. generators other than 2 and 5. (Previously, the
7439 code did not properly initialise the 'add' and 'rem' values to
7440 BN_generate_prime().)
7442 In the new general case, we do not insist that 'generator' is
7443 actually a primitive root: This requirement is rather pointless;
7444 a generator of the order-q subgroup is just as good, if not
7448 *) Map new X509 verification errors to alerts. Discovered and submitted by
7449 Tom Wu <tom@arcot.com>.
7452 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7453 returning non-zero before the data has been completely received
7454 when using non-blocking I/O.
7455 [Bodo Moeller; problem pointed out by John Hughes]
7457 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7458 [Ben Laurie, Lutz Jaenicke]
7460 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7461 Yoram Zahavi <YoramZ@gilian.com>).
7464 *) Add information about CygWin 1.3 and on, and preserve proper
7465 configuration for the versions before that.
7466 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7468 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7469 check whether we deal with a copy of a session and do not delete from
7470 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7471 <izhar@checkpoint.com>.
7474 *) Do not store session data into the internal session cache, if it
7475 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7476 flag is set). Proposed by Aslam <aslam@funk.com>.
7479 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7483 *) [In 0.9.6d-engine release:]
7484 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7485 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7487 *) Add the configuration target linux-s390x.
7488 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7490 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7491 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7492 variable as an indication that a ClientHello message has been
7493 received. As the flag value will be lost between multiple
7494 invocations of ssl3_accept when using non-blocking I/O, the
7495 function may not be aware that a handshake has actually taken
7496 place, thus preventing a new session from being added to the
7499 To avoid this problem, we now set s->new_session to 2 instead of
7500 using a local variable.
7501 [Lutz Jaenicke, Bodo Moeller]
7503 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7504 if the SSL_R_LENGTH_MISMATCH error is detected.
7505 [Geoff Thorpe, Bodo Moeller]
7507 *) New 'shared_ldflag' column in Configure platform table.
7510 *) Fix EVP_CIPHER_mode macro.
7511 ["Dan S. Camper" <dan@bti.net>]
7513 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7514 type, we must throw them away by setting rr->length to 0.
7515 [D P Chang <dpc@qualys.com>]
7517 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7519 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7520 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7521 worked incorrectly for those cases where range = 10..._2 and
7522 3*range is two bits longer than range.)
7525 *) Only add signing time to PKCS7 structures if it is not already
7529 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7530 OBJ_ld_ce should be OBJ_id_ce.
7531 Also some ip-pda OIDs in crypto/objects/objects.txt were
7532 incorrect (cf. RFC 3039).
7533 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7535 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7536 returns early because it has nothing to do.
7537 [Andy Schneider <andy.schneider@bjss.co.uk>]
7539 *) [In 0.9.6c-engine release:]
7540 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7541 [Andy Schneider <andy.schneider@bjss.co.uk>]
7543 *) [In 0.9.6c-engine release:]
7544 Add support for Cryptographic Appliance's keyserver technology.
7545 (Use engine 'keyclient')
7546 [Cryptographic Appliances and Geoff Thorpe]
7548 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7549 is called via tools/c89.sh because arguments have to be
7550 rearranged (all '-L' options must appear before the first object
7552 [Richard Shapiro <rshapiro@abinitio.com>]
7554 *) [In 0.9.6c-engine release:]
7555 Add support for Broadcom crypto accelerator cards, backported
7557 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7559 *) [In 0.9.6c-engine release:]
7560 Add support for SureWare crypto accelerator cards from
7561 Baltimore Technologies. (Use engine 'sureware')
7562 [Baltimore Technologies and Mark Cox]
7564 *) [In 0.9.6c-engine release:]
7565 Add support for crypto accelerator cards from Accelerated
7566 Encryption Processing, www.aep.ie. (Use engine 'aep')
7567 [AEP Inc. and Mark Cox]
7569 *) Add a configuration entry for gcc on UnixWare.
7570 [Gary Benson <gbenson@redhat.com>]
7572 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7573 messages are stored in a single piece (fixed-length part and
7574 variable-length part combined) and fix various bugs found on the way.
7577 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7578 instead. BIO_gethostbyname() does not know what timeouts are
7579 appropriate, so entries would stay in cache even when they have
7581 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7583 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7584 faced with a pathologically small ClientHello fragment that does
7585 not contain client_version: Instead of aborting with an error,
7586 simply choose the highest available protocol version (i.e.,
7587 TLS 1.0 unless it is disabled). In practice, ClientHello
7588 messages are never sent like this, but this change gives us
7589 strictly correct behaviour at least for TLS.
7592 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7593 never resets s->method to s->ctx->method when called from within
7594 one of the SSL handshake functions.
7595 [Bodo Moeller; problem pointed out by Niko Baric]
7597 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7598 (sent using the client's version number) if client_version is
7599 smaller than the protocol version in use. Also change
7600 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7601 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7602 the client will at least see that alert.
7605 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7609 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7610 client receives HelloRequest while in a handshake.
7611 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7613 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7614 should end in 'break', not 'goto end' which circumvents various
7615 cleanups done in state SSL_ST_OK. But session related stuff
7616 must be disabled for SSL_ST_OK in the case that we just sent a
7619 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7620 before just sending a HelloRequest.
7621 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7623 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7624 reveal whether illegal block cipher padding was found or a MAC
7625 verification error occured. (Neither SSLerr() codes nor alerts
7626 are directly visible to potential attackers, but the information
7627 may leak via logfiles.)
7629 Similar changes are not required for the SSL 2.0 implementation
7630 because the number of padding bytes is sent in clear for SSL 2.0,
7631 and the extra bytes are just ignored. However ssl/s2_pkt.c
7632 failed to verify that the purported number of padding bytes is in
7636 *) Add OpenUNIX-8 support including shared libraries
7637 (Boyd Lynn Gerber <gerberb@zenez.com>).
7640 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7641 'wristwatch attack' using huge encoding parameters (cf.
7642 James H. Manger's CRYPTO 2001 paper). Note that the
7643 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7644 encoding parameters and hence was not vulnerable.
7647 *) BN_sqr() bug fix.
7648 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7650 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7651 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7652 followed by modular reduction.
7653 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7655 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7656 equivalent based on BN_pseudo_rand() instead of BN_rand().
7659 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7660 This function was broken, as the check for a new client hello message
7661 to handle SGC did not allow these large messages.
7662 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7665 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7668 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7669 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7672 *) Rework the configuration and shared library support for Tru64 Unix.
7673 The configuration part makes use of modern compiler features and
7674 still retains old compiler behavior for those that run older versions
7675 of the OS. The shared library support part includes a variant that
7676 uses the RPATH feature, and is available through the special
7677 configuration target "alpha-cc-rpath", which will never be selected
7679 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7681 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7682 with the same message size as in ssl3_get_certificate_request().
7683 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7684 messages might inadvertently be reject as too long.
7685 [Petr Lampa <lampa@fee.vutbr.cz>]
7687 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7690 *) Modified SSL library such that the verify_callback that has been set
7691 specificly for an SSL object with SSL_set_verify() is actually being
7692 used. Before the change, a verify_callback set with this function was
7693 ignored and the verify_callback() set in the SSL_CTX at the time of
7694 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7695 to allow the necessary settings.
7698 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7699 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7700 done automatically (in contradiction to the requirements of the C
7701 standard). This made problems when used from OpenSSH.
7704 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7705 dh->length and always used
7707 BN_rand_range(priv_key, dh->p).
7709 BN_rand_range() is not necessary for Diffie-Hellman, and this
7710 specific range makes Diffie-Hellman unnecessarily inefficient if
7711 dh->length (recommended exponent length) is much smaller than the
7712 length of dh->p. We could use BN_rand_range() if the order of
7713 the subgroup was stored in the DH structure, but we only have
7718 BN_rand(priv_key, l, ...)
7720 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7726 RSA_eay_public_encrypt
7727 RSA_eay_private_decrypt
7728 RSA_eay_private_encrypt (signing)
7729 RSA_eay_public_decrypt (signature verification)
7731 (default implementations for RSA_public_encrypt,
7732 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7733 always reject numbers >= n.
7736 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7737 to synchronize access to 'locking_thread'. This is necessary on
7738 systems where access to 'locking_thread' (an 'unsigned long'
7739 variable) is not atomic.
7742 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7743 *before* setting the 'crypto_lock_rand' flag. The previous code had
7744 a race condition if 0 is a valid thread ID.
7745 [Travis Vitek <vitek@roguewave.com>]
7747 *) Add support for shared libraries under Irix.
7748 [Albert Chin-A-Young <china@thewrittenword.com>]
7750 *) Add configuration option to build on Linux on both big-endian and
7752 [Ralf Baechle <ralf@uni-koblenz.de>]
7754 *) Add the possibility to create shared libraries on HP-UX.
7757 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7759 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7760 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7761 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7762 PRNG state recovery was possible based on the output of
7763 one PRNG request appropriately sized to gain knowledge on
7764 'md' followed by enough consecutive 1-byte PRNG requests
7765 to traverse all of 'state'.
7767 1. When updating 'md_local' (the current thread's copy of 'md')
7768 during PRNG output generation, hash all of the previous
7769 'md_local' value, not just the half used for PRNG output.
7771 2. Make the number of bytes from 'state' included into the hash
7772 independent from the number of PRNG bytes requested.
7774 The first measure alone would be sufficient to avoid
7775 Markku-Juhani's attack. (Actually it had never occurred
7776 to me that the half of 'md_local' used for chaining was the
7777 half from which PRNG output bytes were taken -- I had always
7778 assumed that the secret half would be used.) The second
7779 measure makes sure that additional data from 'state' is never
7780 mixed into 'md_local' in small portions; this heuristically
7781 further strengthens the PRNG.
7784 *) Fix crypto/bn/asm/mips3.s.
7787 *) When only the key is given to "enc", the IV is undefined. Print out
7788 an error message in this case.
7791 *) Handle special case when X509_NAME is empty in X509 printing routines.
7794 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7795 positive and less than q.
7798 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7799 used: it isn't thread safe and the add_lock_callback should handle
7801 [Paul Rose <Paul.Rose@bridge.com>]
7803 *) Verify that incoming data obeys the block size in
7804 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7808 [Ulf Möller, Bodo Möller]
7810 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7811 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7812 when fixing the server behaviour for backwards-compatible 'client
7813 hello' messages. (Note that the attack is impractical against
7814 SSL 3.0 and TLS 1.0 anyway because length and version checking
7815 means that the probability of guessing a valid ciphertext is
7816 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7819 Before 0.9.5, the countermeasure (hide the error by generating a
7820 random 'decryption result') did not work properly because
7821 ERR_clear_error() was missing, meaning that SSL_get_error() would
7822 detect the supposedly ignored error.
7824 Both problems are now fixed.
7827 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7828 (previously it was 1024).
7831 *) Fix for compatibility mode trust settings: ignore trust settings
7832 unless some valid trust or reject settings are present.
7835 *) Fix for blowfish EVP: its a variable length cipher.
7838 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7839 parameters in DSA public key structures and return an error in the
7840 DSA routines if parameters are absent.
7843 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7844 in the current directory if neither $RANDFILE nor $HOME was set.
7845 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7846 caused some confusion to Windows users who haven't defined $HOME.
7847 Thus RAND_file_name() is changed again: e_os.h can define a
7848 DEFAULT_HOME, which will be used if $HOME is not set.
7849 For Windows, we use "C:"; on other platforms, we still require
7850 environment variables.
7852 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7853 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7854 having multiple threads call RAND_poll() concurrently.
7857 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7858 combination of a flag and a thread ID variable.
7859 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7860 flag), *other* threads can enter ssleay_add_bytes without obeying
7861 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7862 that they do not hold after the first thread unsets add_do_not_lock).
7865 *) Change bctest again: '-x' expressions are not available in all
7869 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7871 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7872 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7874 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7875 the default extension for executables, if any. Also, make the perl
7876 scripts that use symlink() to test if it really exists and use "cp"
7877 if it doesn't. All this made OpenSSL compilable and installable in
7881 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7882 If SEQUENCE is length is indefinite just set c->slen to the total
7883 amount of data available.
7884 [Steve Henson, reported by shige@FreeBSD.org]
7885 [This change does not apply to 0.9.7.]
7887 *) Change bctest to avoid here-documents inside command substitution
7888 (workaround for FreeBSD /bin/sh bug).
7889 For compatibility with Ultrix, avoid shell functions (introduced
7890 in the bctest version that searches along $PATH).
7893 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7894 with des_encrypt() defined on some operating systems, like Solaris
7898 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7899 On the Importance of Eliminating Errors in Cryptographic
7900 Computations, J. Cryptology 14 (2001) 2, 101-119,
7901 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7904 *) MIPS assembler BIGNUM division bug fix.
7907 *) Disabled incorrect Alpha assembler code.
7910 *) Fix PKCS#7 decode routines so they correctly update the length
7911 after reading an EOC for the EXPLICIT tag.
7913 [This change does not apply to 0.9.7.]
7915 *) Fix bug in PKCS#12 key generation routines. This was triggered
7916 if a 3DES key was generated with a 0 initial byte. Include
7917 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7918 (but broken) behaviour.
7921 *) Enhance bctest to search for a working bc along $PATH and print
7923 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7925 *) Fix memory leaks in err.c: free err_data string if necessary;
7926 don't write to the wrong index in ERR_set_error_data.
7929 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7933 *) Replace rdtsc with _emit statements for VC++ version 5.
7934 [Jeremy Cooper <jeremy@baymoo.org>]
7936 *) Make it possible to reuse SSLv2 sessions.
7939 *) In copy_email() check for >= 0 as a return value for
7940 X509_NAME_get_index_by_NID() since 0 is a valid index.
7941 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7943 *) Avoid coredump with unsupported or invalid public keys by checking if
7944 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7945 PKCS7_verify() fails with non detached data.
7948 *) Don't use getenv in library functions when run as setuid/setgid.
7949 New function OPENSSL_issetugid().
7952 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7953 due to incorrect handling of multi-threading:
7955 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7957 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7959 3. Count how many times MemCheck_off() has been called so that
7960 nested use can be treated correctly. This also avoids
7961 inband-signalling in the previous code (which relied on the
7962 assumption that thread ID 0 is impossible).
7965 *) Add "-rand" option also to s_client and s_server.
7968 *) Fix CPU detection on Irix 6.x.
7969 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7970 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7972 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7975 [This change does not apply to 0.9.7.]
7977 *) Use the cached encoding of an X509_NAME structure rather than
7978 copying it. This is apparently the reason for the libsafe "errors"
7979 but the code is actually correct.
7982 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7983 Bleichenbacher's DSA attack.
7984 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7985 to be set and top=0 forces the highest bit to be set; top=-1 is new
7986 and leaves the highest bit random.
7987 [Ulf Moeller, Bodo Moeller]
7989 *) In the NCONF_...-based implementations for CONF_... queries
7990 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7991 a temporary CONF structure with the data component set to NULL
7992 (which gives segmentation faults in lh_retrieve).
7993 Instead, use NULL for the CONF pointer in CONF_get_string and
7994 CONF_get_number (which may use environment variables) and directly
7995 return NULL from CONF_get_section.
7998 *) Fix potential buffer overrun for EBCDIC.
8001 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8002 keyUsage if basicConstraints absent for a CA.
8005 *) Make SMIME_write_PKCS7() write mail header values with a format that
8006 is more generally accepted (no spaces before the semicolon), since
8007 some programs can't parse those values properly otherwise. Also make
8008 sure BIO's that break lines after each write do not create invalid
8012 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8013 macros previously used would not encode an empty SEQUENCE OF
8014 and break the signature.
8016 [This change does not apply to 0.9.7.]
8018 *) Zero the premaster secret after deriving the master secret in
8022 *) Add some EVP_add_digest_alias registrations (as found in
8023 OpenSSL_add_all_digests()) to SSL_library_init()
8024 aka OpenSSL_add_ssl_algorithms(). This provides improved
8025 compatibility with peers using X.509 certificates
8026 with unconventional AlgorithmIdentifier OIDs.
8029 *) Fix for Irix with NO_ASM.
8030 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8032 *) ./config script fixes.
8033 [Ulf Moeller, Richard Levitte]
8035 *) Fix 'openssl passwd -1'.
8038 *) Change PKCS12_key_gen_asc() so it can cope with non null
8039 terminated strings whose length is passed in the passlen
8040 parameter, for example from PEM callbacks. This was done
8041 by adding an extra length parameter to asc2uni().
8042 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8044 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8045 call failed, free the DSA structure.
8048 *) Fix to uni2asc() to cope with zero length Unicode strings.
8049 These are present in some PKCS#12 files.
8052 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8053 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8054 when writing a 32767 byte record.
8055 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8057 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8058 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8060 (RSA objects have a reference count access to which is protected
8061 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8062 so they are meant to be shared between threads.)
8063 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8064 "Reddie, Steven" <Steven.Reddie@ca.com>]
8066 *) Fix a deadlock in CRYPTO_mem_leaks().
8069 *) Use better test patterns in bntest.
8072 *) rand_win.c fix for Borland C.
8075 *) BN_rshift bugfix for n == 0.
8078 *) Add a 'bctest' script that checks for some known 'bc' bugs
8079 so that 'make test' does not abort just because 'bc' is broken.
8082 *) Store verify_result within SSL_SESSION also for client side to
8083 avoid potential security hole. (Re-used sessions on the client side
8084 always resulted in verify_result==X509_V_OK, not using the original
8085 result of the server certificate verification.)
8088 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8089 SSL3_RT_APPLICATION_DATA, return 0.
8090 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8094 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8095 releases, have been re-implemented by renaming the previous
8096 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8097 and ssl3_read_internal, respectively, and adding 'peek' parameters
8098 to them. The new ssl[23]_{read,peek} functions are calls to
8099 ssl[23]_read_internal with the 'peek' flag set appropriately.
8100 A 'peek' parameter has also been added to ssl3_read_bytes, which
8101 does the actual work for ssl3_read_internal.
8104 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8105 the method-specific "init()" handler. Also clean up ex_data after
8106 calling the method-specific "finish()" handler. Previously, this was
8107 happening the other way round.
8110 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8111 The previous value, 12, was not always sufficient for BN_mod_exp().
8114 *) Make sure that shared libraries get the internal name engine with
8115 the full version number and not just 0. This should mark the
8116 shared libraries as not backward compatible. Of course, this should
8117 be changed again when we can guarantee backward binary compatibility.
8120 *) Fix typo in get_cert_by_subject() in by_dir.c
8121 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8123 *) Rework the system to generate shared libraries:
8125 - Make note of the expected extension for the shared libraries and
8126 if there is a need for symbolic links from for example libcrypto.so.0
8127 to libcrypto.so.0.9.7. There is extended info in Configure for
8130 - Make as few rebuilds of the shared libraries as possible.
8132 - Still avoid linking the OpenSSL programs with the shared libraries.
8134 - When installing, install the shared libraries separately from the
8138 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8140 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8141 and not in SSL_clear because the latter is also used by the
8142 accept/connect functions; previously, the settings made by
8143 SSL_set_read_ahead would be lost during the handshake.
8144 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8146 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8147 Previously, it would create entries for disabled algorithms no
8151 *) Added several new manual pages for SSL_* function.
8154 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8156 *) In ssl23_get_client_hello, generate an error message when faced
8157 with an initial SSL 3.0/TLS record that is too small to contain the
8158 first two bytes of the ClientHello message, i.e. client_version.
8159 (Note that this is a pathologic case that probably has never happened
8160 in real life.) The previous approach was to use the version number
8161 from the record header as a substitute; but our protocol choice
8162 should not depend on that one because it is not authenticated
8163 by the Finished messages.
8166 *) More robust randomness gathering functions for Windows.
8167 [Jeffrey Altman <jaltman@columbia.edu>]
8169 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8170 not set then we don't setup the error code for issuer check errors
8171 to avoid possibly overwriting other errors which the callback does
8172 handle. If an application does set the flag then we assume it knows
8173 what it is doing and can handle the new informational codes
8177 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8178 a general "ANY" type, as such it should be able to decode anything
8179 including tagged types. However it didn't check the class so it would
8180 wrongly interpret tagged types in the same way as their universal
8181 counterpart and unknown types were just rejected. Changed so that the
8182 tagged and unknown types are handled in the same way as a SEQUENCE:
8183 that is the encoding is stored intact. There is also a new type
8184 "V_ASN1_OTHER" which is used when the class is not universal, in this
8185 case we have no idea what the actual type is so we just lump them all
8189 *) On VMS, stdout may very well lead to a file that is written to
8190 in a record-oriented fashion. That means that every write() will
8191 write a separate record, which will be read separately by the
8192 programs trying to read from it. This can be very confusing.
8194 The solution is to put a BIO filter in the way that will buffer
8195 text until a linefeed is reached, and then write everything a
8196 line at a time, so every record written will be an actual line,
8197 not chunks of lines and not (usually doesn't happen, but I've
8198 seen it once) several lines in one record. BIO_f_linebuffer() is
8201 Currently, it's a VMS-only method, because that's where it has
8202 been tested well enough.
8205 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8206 it can return incorrect results.
8207 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8208 but it was in 0.9.6-beta[12].)
8211 *) Disable the check for content being present when verifying detached
8212 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8213 include zero length content when signing messages.
8216 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8217 BIO_ctrl (for BIO pairs).
8220 *) Add DSO method for VMS.
8223 *) Bug fix: Montgomery multiplication could produce results with the
8227 *) Add RPM specification openssl.spec and modify it to build three
8228 packages. The default package contains applications, application
8229 documentation and run-time libraries. The devel package contains
8230 include files, static libraries and function documentation. The
8231 doc package contains the contents of the doc directory. The original
8232 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8235 *) Add a large number of documentation files for many SSL routines.
8236 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8238 *) Add a configuration entry for Sony News 4.
8239 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8241 *) Don't set the two most significant bits to one when generating a
8242 random number < q in the DSA library.
8245 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8246 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8247 the underlying transport is blocking) if a handshake took place.
8248 (The default behaviour is needed by applications such as s_client
8249 and s_server that use select() to determine when to use SSL_read;
8250 but for applications that know in advance when to expect data, it
8251 just makes things more complicated.)
8254 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8258 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8259 work better on such systems.
8260 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8262 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8263 Update PKCS12_parse() so it copies the friendlyName and the
8264 keyid to the certificates aux info.
8267 *) Fix bug in PKCS7_verify() which caused an infinite loop
8268 if there was more than one signature.
8269 [Sven Uszpelkat <su@celocom.de>]
8271 *) Major change in util/mkdef.pl to include extra information
8272 about each symbol, as well as presenting variables as well
8273 as functions. This change means that there's n more need
8274 to rebuild the .num files when some algorithms are excluded.
8277 *) Allow the verify time to be set by an application,
8278 rather than always using the current time.
8281 *) Phase 2 verify code reorganisation. The certificate
8282 verify code now looks up an issuer certificate by a
8283 number of criteria: subject name, authority key id
8284 and key usage. It also verifies self signed certificates
8285 by the same criteria. The main comparison function is
8286 X509_check_issued() which performs these checks.
8288 Lot of changes were necessary in order to support this
8289 without completely rewriting the lookup code.
8291 Authority and subject key identifier are now cached.
8293 The LHASH 'certs' is X509_STORE has now been replaced
8294 by a STACK_OF(X509_OBJECT). This is mainly because an
8295 LHASH can't store or retrieve multiple objects with
8296 the same hash value.
8298 As a result various functions (which were all internal
8299 use only) have changed to handle the new X509_STORE
8300 structure. This will break anything that messed round
8301 with X509_STORE internally.
8303 The functions X509_STORE_add_cert() now checks for an
8304 exact match, rather than just subject name.
8306 The X509_STORE API doesn't directly support the retrieval
8307 of multiple certificates matching a given criteria, however
8308 this can be worked round by performing a lookup first
8309 (which will fill the cache with candidate certificates)
8310 and then examining the cache for matches. This is probably
8311 the best we can do without throwing out X509_LOOKUP
8312 entirely (maybe later...).
8314 The X509_VERIFY_CTX structure has been enhanced considerably.
8316 All certificate lookup operations now go via a get_issuer()
8317 callback. Although this currently uses an X509_STORE it
8318 can be replaced by custom lookups. This is a simple way
8319 to bypass the X509_STORE hackery necessary to make this
8320 work and makes it possible to use more efficient techniques
8321 in future. A very simple version which uses a simple
8322 STACK for its trusted certificate store is also provided
8323 using X509_STORE_CTX_trusted_stack().
8325 The verify_cb() and verify() callbacks now have equivalents
8326 in the X509_STORE_CTX structure.
8328 X509_STORE_CTX also has a 'flags' field which can be used
8329 to customise the verify behaviour.
8332 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8333 excludes S/MIME capabilities.
8336 *) When a certificate request is read in keep a copy of the
8337 original encoding of the signed data and use it when outputing
8338 again. Signatures then use the original encoding rather than
8339 a decoded, encoded version which may cause problems if the
8340 request is improperly encoded.
8343 *) For consistency with other BIO_puts implementations, call
8344 buffer_write(b, ...) directly in buffer_puts instead of calling
8347 In BIO_puts, increment b->num_write as in BIO_write.
8348 [Peter.Sylvester@EdelWeb.fr]
8350 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8351 BN_zero, we may not return a BIGNUM with an array consisting of
8355 *) Avoid calling abort() from within the library when problems are
8356 detected, except if preprocessor symbols have been defined
8357 (such as REF_CHECK, BN_DEBUG etc.).
8360 *) New openssl application 'rsautl'. This utility can be
8361 used for low level RSA operations. DER public key
8362 BIO/fp routines also added.
8365 *) New Configure entry and patches for compiling on QNX 4.
8366 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8368 *) A demo state-machine implementation was sponsored by
8369 Nuron (http://www.nuron.com/) and is now available in
8370 demos/state_machine.
8373 *) New options added to the 'dgst' utility for signature
8374 generation and verification.
8377 *) Unrecognized PKCS#7 content types are now handled via a
8378 catch all ASN1_TYPE structure. This allows unsupported
8379 types to be stored as a "blob" and an application can
8380 encode and decode it manually.
8383 *) Fix various signed/unsigned issues to make a_strex.c
8385 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8387 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8388 length if passed a buffer. ASN1_INTEGER_to_BN failed
8389 if passed a NULL BN and its argument was negative.
8390 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8392 *) Modification to PKCS#7 encoding routines to output definite
8393 length encoding. Since currently the whole structures are in
8394 memory there's not real point in using indefinite length
8395 constructed encoding. However if OpenSSL is compiled with
8396 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8399 *) Added BIO_vprintf() and BIO_vsnprintf().
8402 *) Added more prefixes to parse for in the the strings written
8403 through a logging bio, to cover all the levels that are available
8404 through syslog. The prefixes are now:
8406 PANIC, EMERG, EMR => LOG_EMERG
8407 ALERT, ALR => LOG_ALERT
8408 CRIT, CRI => LOG_CRIT
8409 ERROR, ERR => LOG_ERR
8410 WARNING, WARN, WAR => LOG_WARNING
8411 NOTICE, NOTE, NOT => LOG_NOTICE
8412 INFO, INF => LOG_INFO
8413 DEBUG, DBG => LOG_DEBUG
8415 and as before, if none of those prefixes are present at the
8416 beginning of the string, LOG_ERR is chosen.
8418 On Win32, the LOG_* levels are mapped according to this:
8420 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8421 LOG_WARNING => EVENTLOG_WARNING_TYPE
8422 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8426 *) Made it possible to reconfigure with just the configuration
8427 argument "reconf" or "reconfigure". The command line arguments
8428 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8429 and are retrieved from there when reconfiguring.
8433 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8435 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8438 *) The obj_dat.pl script was messing up the sorting of object
8439 names. The reason was that it compared the quoted version
8440 of strings as a result "OCSP" > "OCSP Signing" because
8441 " > SPACE. Changed script to store unquoted versions of
8442 names and add quotes on output. It was also omitting some
8443 names from the lookup table if they were given a default
8444 value (that is if SN is missing it is given the same
8445 value as LN and vice versa), these are now added on the
8446 grounds that if an object has a name we should be able to
8447 look it up. Finally added warning output when duplicate
8448 short or long names are found.
8451 *) Changes needed for Tandem NSK.
8452 [Scott Uroff <scott@xypro.com>]
8454 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8455 RSA_padding_check_SSLv23(), special padding was never detected
8456 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8457 version rollback attacks was not effective.
8459 In s23_clnt.c, don't use special rollback-attack detection padding
8460 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8461 client; similarly, in s23_srvr.c, don't do the rollback check if
8462 SSL 2.0 is the only protocol enabled in the server.
8465 *) Make it possible to get hexdumps of unprintable data with 'openssl
8466 asn1parse'. By implication, the functions ASN1_parse_dump() and
8467 BIO_dump_indent() are added.
8470 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8471 these print out strings and name structures based on various
8472 flags including RFC2253 support and proper handling of
8473 multibyte characters. Added options to the 'x509' utility
8474 to allow the various flags to be set.
8477 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8478 Also change the functions X509_cmp_current_time() and
8479 X509_gmtime_adj() work with an ASN1_TIME structure,
8480 this will enable certificates using GeneralizedTime in validity
8481 dates to be checked.
8484 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8485 negative public key encodings) on by default,
8486 NO_NEG_PUBKEY_BUG can be set to disable it.
8489 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8490 content octets. An i2c_ASN1_OBJECT is unnecessary because
8491 the encoding can be trivially obtained from the structure.
8494 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8495 not read locks (CRYPTO_r_[un]lock).
8498 *) A first attempt at creating official support for shared
8499 libraries through configuration. I've kept it so the
8500 default is static libraries only, and the OpenSSL programs
8501 are always statically linked for now, but there are
8502 preparations for dynamic linking in place.
8503 This has been tested on Linux and Tru64.
8506 *) Randomness polling function for Win9x, as described in:
8507 Peter Gutmann, Software Generation of Practically Strong
8511 *) Fix so PRNG is seeded in req if using an already existing
8515 *) New options to smime application. -inform and -outform
8516 allow alternative formats for the S/MIME message including
8517 PEM and DER. The -content option allows the content to be
8518 specified separately. This should allow things like Netscape
8519 form signing output easier to verify.
8522 *) Fix the ASN1 encoding of tags using the 'long form'.
8525 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8526 STRING types. These convert content octets to and from the
8527 underlying type. The actual tag and length octets are
8528 already assumed to have been read in and checked. These
8529 are needed because all other string types have virtually
8530 identical handling apart from the tag. By having versions
8531 of the ASN1 functions that just operate on content octets
8532 IMPLICIT tagging can be handled properly. It also allows
8533 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8534 and ASN1_INTEGER are identical apart from the tag.
8537 *) Change the handling of OID objects as follows:
8539 - New object identifiers are inserted in objects.txt, following
8540 the syntax given in objects.README.
8541 - objects.pl is used to process obj_mac.num and create a new
8543 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8546 This is currently kind of a hack, and the perl code in objects.pl
8547 isn't very elegant, but it works as I intended. The simplest way
8548 to check that it worked correctly is to look in obj_dat.h and
8549 check the array nid_objs and make sure the objects haven't moved
8550 around (this is important!). Additions are OK, as well as
8551 consistent name changes.
8554 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8557 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8558 The given file adds to whatever has already been seeded into the
8559 random pool through the RANDFILE configuration file option or
8560 environment variable, or the default random state file.
8563 *) mkstack.pl now sorts each macro group into lexical order.
8564 Previously the output order depended on the order the files
8565 appeared in the directory, resulting in needless rewriting
8569 *) Patches to make OpenSSL compile under Win32 again. Mostly
8570 work arounds for the VC++ problem that it treats func() as
8571 func(void). Also stripped out the parts of mkdef.pl that
8572 added extra typesafe functions: these no longer exist.
8575 *) Reorganisation of the stack code. The macros are now all
8576 collected in safestack.h . Each macro is defined in terms of
8577 a "stack macro" of the form SKM_<name>(type, a, b). The
8578 DEBUG_SAFESTACK is now handled in terms of function casts,
8579 this has the advantage of retaining type safety without the
8580 use of additional functions. If DEBUG_SAFESTACK is not defined
8581 then the non typesafe macros are used instead. Also modified the
8582 mkstack.pl script to handle the new form. Needs testing to see
8583 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8584 the default if no major problems. Similar behaviour for ASN1_SET_OF
8585 and PKCS12_STACK_OF.
8588 *) When some versions of IIS use the 'NET' form of private key the
8589 key derivation algorithm is different. Normally MD5(password) is
8590 used as a 128 bit RC4 key. In the modified case
8591 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8592 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8593 as the old Netscape_RSA functions except they have an additional
8594 'sgckey' parameter which uses the modified algorithm. Also added
8595 an -sgckey command line option to the rsa utility. Thanks to
8596 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8597 algorithm to openssl-dev.
8600 *) The evp_local.h macros were using 'c.##kname' which resulted in
8601 invalid expansion on some systems (SCO 5.0.5 for example).
8602 Corrected to 'c.kname'.
8603 [Phillip Porch <root@theporch.com>]
8605 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8606 a STACK of email addresses from a certificate or request, these look
8607 in the subject name and the subject alternative name extensions and
8608 omit any duplicate addresses.
8611 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8612 This makes DSA verification about 2 % faster.
8615 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8616 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8617 plus overhead for 1024 bit moduli).
8618 This makes exponentiations about 0.5 % faster for 1024 bit
8619 exponents (as measured by "openssl speed rsa2048").
8622 *) Rename memory handling macros to avoid conflicts with other
8624 Malloc => OPENSSL_malloc
8625 Malloc_locked => OPENSSL_malloc_locked
8626 Realloc => OPENSSL_realloc
8627 Free => OPENSSL_free
8630 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8631 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8634 *) CygWin32 support.
8635 [John Jarvie <jjarvie@newsguy.com>]
8637 *) The type-safe stack code has been rejigged. It is now only compiled
8638 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8639 by default all type-specific stack functions are "#define"d back to
8640 standard stack functions. This results in more streamlined output
8641 but retains the type-safety checking possibilities of the original
8645 *) The STACK code has been cleaned up, and certain type declarations
8646 that didn't make a lot of sense have been brought in line. This has
8647 also involved a cleanup of sorts in safestack.h to more correctly
8648 map type-safe stack functions onto their plain stack counterparts.
8649 This work has also resulted in a variety of "const"ifications of
8650 lots of the code, especially "_cmp" operations which should normally
8651 be prototyped with "const" parameters anyway.
8654 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8655 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8656 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8657 where all of 'md' is used each time the PRNG is used, but 'state'
8658 is used only indexed by a cyclic counter. As entropy may not be
8659 well distributed from the beginning, 'md' is important as a
8660 chaining variable. However, the output function chains only half
8661 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8662 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8663 in all of 'state' being rewritten, with the new values depending
8664 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8667 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8668 the handshake is continued after ssl_verify_cert_chain();
8669 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8670 can lead to 'unexplainable' connection aborts later.
8671 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8673 *) Major EVP API cipher revision.
8674 Add hooks for extra EVP features. This allows various cipher
8675 parameters to be set in the EVP interface. Support added for variable
8676 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8677 setting of RC2 and RC5 parameters.
8679 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8682 Remove lots of duplicated code from the EVP library. For example *every*
8683 cipher init() function handles the 'iv' in the same way according to the
8684 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8685 for CFB and OFB modes they zero ctx->num.
8687 New functionality allows removal of S/MIME code RC2 hack.
8689 Most of the routines have the same form and so can be declared in terms
8692 By shifting this to the top level EVP_CipherInit() it can be removed from
8693 all individual ciphers. If the cipher wants to handle IVs or keys
8694 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8697 Change lots of functions like EVP_EncryptUpdate() to now return a
8698 value: although software versions of the algorithms cannot fail
8699 any installed hardware versions can.
8702 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8703 this option is set, tolerate broken clients that send the negotiated
8704 protocol version number instead of the requested protocol version
8708 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8709 i.e. non-zero for export ciphersuites, zero otherwise.
8710 Previous versions had this flag inverted, inconsistent with
8711 rsa_tmp_cb (..._TMP_RSA_CB).
8712 [Bodo Moeller; problem reported by Amit Chopra]
8714 *) Add missing DSA library text string. Work around for some IIS
8715 key files with invalid SEQUENCE encoding.
8718 *) Add a document (doc/standards.txt) that list all kinds of standards
8719 and so on that are implemented in OpenSSL.
8722 *) Enhance c_rehash script. Old version would mishandle certificates
8723 with the same subject name hash and wouldn't handle CRLs at all.
8724 Added -fingerprint option to crl utility, to support new c_rehash
8728 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8731 *) Fix for SSL server purpose checking. Server checking was
8732 rejecting certificates which had extended key usage present
8733 but no ssl client purpose.
8734 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8736 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8737 is a little unclear about how a blank password is handled.
8738 Since the password in encoded as a BMPString with terminating
8739 double NULL a zero length password would end up as just the
8740 double NULL. However no password at all is different and is
8741 handled differently in the PKCS#12 key generation code. NS
8742 treats a blank password as zero length. MSIE treats it as no
8743 password on export: but it will try both on import. We now do
8744 the same: PKCS12_parse() tries zero length and no password if
8745 the password is set to "" or NULL (NULL is now a valid password:
8746 it wasn't before) as does the pkcs12 application.
8749 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8750 perror when PEM_read_bio_X509_REQ fails, the error message must
8751 be obtained from the error queue.
8754 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8755 it in ERR_remove_state if appropriate, and change ERR_get_state
8756 accordingly to avoid race conditions (this is necessary because
8757 thread_hash is no longer constant once set).
8760 *) Bugfix for linux-elf makefile.one.
8763 *) RSA_get_default_method() will now cause a default
8764 RSA_METHOD to be chosen if one doesn't exist already.
8765 Previously this was only set during a call to RSA_new()
8766 or RSA_new_method(NULL) meaning it was possible for
8767 RSA_get_default_method() to return NULL.
8770 *) Added native name translation to the existing DSO code
8771 that will convert (if the flag to do so is set) filenames
8772 that are sufficiently small and have no path information
8773 into a canonical native form. Eg. "blah" converted to
8774 "libblah.so" or "blah.dll" etc.
8777 *) New function ERR_error_string_n(e, buf, len) which is like
8778 ERR_error_string(e, buf), but writes at most 'len' bytes
8779 including the 0 terminator. For ERR_error_string_n, 'buf'
8781 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8783 *) CONF library reworked to become more general. A new CONF
8784 configuration file reader "class" is implemented as well as a
8785 new functions (NCONF_*, for "New CONF") to handle it. The now
8786 old CONF_* functions are still there, but are reimplemented to
8787 work in terms of the new functions. Also, a set of functions
8788 to handle the internal storage of the configuration data is
8789 provided to make it easier to write new configuration file
8790 reader "classes" (I can definitely see something reading a
8791 configuration file in XML format, for example), called _CONF_*,
8792 or "the configuration storage API"...
8794 The new configuration file reading functions are:
8796 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8797 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8799 NCONF_default, NCONF_WIN32
8801 NCONF_dump_fp, NCONF_dump_bio
8803 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8804 NCONF_new creates a new CONF object. This works in the same way
8805 as other interfaces in OpenSSL, like the BIO interface.
8806 NCONF_dump_* dump the internal storage of the configuration file,
8807 which is useful for debugging. All other functions take the same
8808 arguments as the old CONF_* functions wth the exception of the
8809 first that must be a `CONF *' instead of a `LHASH *'.
8811 To make it easer to use the new classes with the old CONF_* functions,
8812 the function CONF_set_default_method is provided.
8815 *) Add '-tls1' option to 'openssl ciphers', which was already
8816 mentioned in the documentation but had not been implemented.
8817 (This option is not yet really useful because even the additional
8818 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8821 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8822 OpenSSL-based applications) load shared libraries and bind to
8823 them in a portable way.
8824 [Geoff Thorpe, with contributions from Richard Levitte]
8826 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8828 *) Make sure _lrotl and _lrotr are only used with MSVC.
8830 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8831 (the default implementation of RAND_status).
8833 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8834 to '-clrext' (= clear extensions), as intended and documented.
8835 [Bodo Moeller; inconsistency pointed out by Michael Attili
8838 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8839 was larger than the MD block size.
8840 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8842 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8843 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8844 using the passed key: if the passed key was a private key the result
8845 of X509_print(), for example, would be to print out all the private key
8849 *) des_quad_cksum() byte order bug fix.
8850 [Ulf Möller, using the problem description in krb4-0.9.7, where
8851 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8853 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8855 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8857 *) For easily testing in shell scripts whether some command
8858 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8859 returns with exit code 0 iff no command of the given name is available.
8860 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8861 the output goes to stdout and nothing is printed to stderr.
8862 Additional arguments are always ignored.
8864 Since for each cipher there is a command of the same name,
8865 the 'no-cipher' compilation switches can be tested this way.
8867 ('openssl no-XXX' is not able to detect pseudo-commands such
8868 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8871 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8874 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8875 is set; it will be thrown away anyway because each handshake creates
8877 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8878 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8879 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8880 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8883 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8884 'Q' and 'R' lose their special meanings (quit/renegotiate).
8885 This is part of what -quiet does; unlike -quiet, -ign_eof
8886 does not suppress any output.
8889 *) Add compatibility options to the purpose and trust code. The
8890 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8891 accepts a certificate or CA, this was the previous behaviour,
8892 with all the associated security issues.
8894 X509_TRUST_COMPAT is the old trust behaviour: only and
8895 automatically trust self signed roots in certificate store. A
8896 new trust setting X509_TRUST_DEFAULT is used to specify that
8897 a purpose has no associated trust setting and it should instead
8898 use the value in the default purpose.
8901 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8902 and fix a memory leak.
8905 *) In util/mkerr.pl (which implements 'make errors'), preserve
8906 reason strings from the previous version of the .c file, as
8907 the default to have only downcase letters (and digits) in
8908 automatically generated reasons codes is not always appropriate.
8911 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8912 using strerror. Previously, ERR_reason_error_string() returned
8913 library names as reason strings for SYSerr; but SYSerr is a special
8914 case where small numbers are errno values, not library numbers.
8917 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8918 converts DSA parameters into DH parameters. (When creating parameters,
8919 DSA_generate_parameters is used.)
8922 *) Include 'length' (recommended exponent length) in C code generated
8923 by 'openssl dhparam -C'.
8926 *) The second argument to set_label in perlasm was already being used
8927 so couldn't be used as a "file scope" flag. Moved to third argument
8931 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8932 instead of RAND_bytes for encryption IVs and salts.
8935 *) Include RAND_status() into RAND_METHOD instead of implementing
8936 it only for md_rand.c Otherwise replacing the PRNG by calling
8937 RAND_set_rand_method would be impossible.
8940 *) Don't let DSA_generate_key() enter an infinite loop if the random
8941 number generation fails.
8944 *) New 'rand' application for creating pseudo-random output.
8947 *) Added configuration support for Linux/IA64
8948 [Rolf Haberrecker <rolf@suse.de>]
8950 *) Assembler module support for Mingw32.
8953 *) Shared library support for HPUX (in shlib/).
8954 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8956 *) Shared library support for Solaris gcc.
8957 [Lutz Behnke <behnke@trustcenter.de>]
8959 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8961 *) PKCS7_encrypt() was adding text MIME headers twice because they
8962 were added manually and by SMIME_crlf_copy().
8965 *) In bntest.c don't call BN_rand with zero bits argument.
8966 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8968 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8969 case was implemented. This caused BN_div_recp() to fail occasionally.
8972 *) Add an optional second argument to the set_label() in the perl
8973 assembly language builder. If this argument exists and is set
8974 to 1 it signals that the assembler should use a symbol whose
8975 scope is the entire file, not just the current function. This
8976 is needed with MASM which uses the format label:: for this scope.
8977 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8979 *) Change the ASN1 types so they are typedefs by default. Before
8980 almost all types were #define'd to ASN1_STRING which was causing
8981 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8985 *) Change names of new functions to the new get1/get0 naming
8986 convention: After 'get1', the caller owns a reference count
8987 and has to call ..._free; 'get0' returns a pointer to some
8988 data structure without incrementing reference counters.
8989 (Some of the existing 'get' functions increment a reference
8990 counter, some don't.)
8991 Similarly, 'set1' and 'add1' functions increase reference
8992 counters or duplicate objects.
8995 *) Allow for the possibility of temp RSA key generation failure:
8996 the code used to assume it always worked and crashed on failure.
8999 *) Fix potential buffer overrun problem in BIO_printf().
9000 [Ulf Möller, using public domain code by Patrick Powell; problem
9001 pointed out by David Sacerdote <das33@cornell.edu>]
9003 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9004 RAND_egd() and RAND_status(). In the command line application,
9005 the EGD socket can be specified like a seed file using RANDFILE
9009 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9010 Some CAs (e.g. Verisign) distribute certificates in this form.
9013 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9014 list to exclude them. This means that no special compilation option
9015 is needed to use anonymous DH: it just needs to be included in the
9019 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9020 EVP_MD_type. The old functionality is available in a new macro called
9021 EVP_MD_md(). Change code that uses it and update docs.
9024 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9025 where the 'void *' argument is replaced by a function pointer argument.
9026 Previously 'void *' was abused to point to functions, which works on
9027 many platforms, but is not correct. As these functions are usually
9028 called by macros defined in OpenSSL header files, most source code
9029 should work without changes.
9032 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9033 sections with information on -D... compiler switches used for
9034 compiling the library so that applications can see them. To enable
9035 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9036 must be defined. E.g.,
9037 #define OPENSSL_ALGORITHM_DEFINES
9038 #include <openssl/opensslconf.h>
9039 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9040 [Richard Levitte, Ulf and Bodo Möller]
9042 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9046 *) Change the 'other' type in certificate aux info to a STACK_OF
9047 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9048 the required ASN1 format: arbitrary types determined by an OID.
9051 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9052 argument to 'req'. This is not because the function is newer or
9053 better than others it just uses the work 'NEW' in the certificate
9054 request header lines. Some software needs this.
9057 *) Reorganise password command line arguments: now passwords can be
9058 obtained from various sources. Delete the PEM_cb function and make
9059 it the default behaviour: i.e. if the callback is NULL and the
9060 usrdata argument is not NULL interpret it as a null terminated pass
9061 phrase. If usrdata and the callback are NULL then the pass phrase
9062 is prompted for as usual.
9065 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9066 the support is automatically enabled. The resulting binaries will
9067 autodetect the card and use it if present.
9068 [Ben Laurie and Compaq Inc.]
9070 *) Work around for Netscape hang bug. This sends certificate request
9071 and server done in one record. Since this is perfectly legal in the
9072 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9073 the bugs/SSLv3 entry for more info.
9076 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9079 *) Add -rand argument to smime and pkcs12 applications and read/write
9083 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9086 *) Add command line password options to the remaining applications.
9089 *) Bug fix for BN_div_recp() for numerators with an even number of
9093 *) More tests in bntest.c, and changed test_bn output.
9096 *) ./config recognizes MacOS X now.
9099 *) Bug fix for BN_div() when the first words of num and divsor are
9100 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9103 *) Add support for various broken PKCS#8 formats, and command line
9104 options to produce them.
9107 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9108 get temporary BIGNUMs from a BN_CTX.
9111 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9115 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9116 include a #define from the old name to the new. The original intent
9117 was that statically linked binaries could for example just call
9118 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9119 link with digests. This never worked because SSLeay_add_all_digests()
9120 and SSLeay_add_all_ciphers() were in the same source file so calling
9121 one would link with the other. They are now in separate source files.
9124 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9127 *) Use a less unusual form of the Miller-Rabin primality test (it used
9128 a binary algorithm for exponentiation integrated into the Miller-Rabin
9129 loop, our standard modexp algorithms are faster).
9132 *) Support for the EBCDIC character set completed.
9133 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9135 *) Source code cleanups: use const where appropriate, eliminate casts,
9136 use void * instead of char * in lhash.
9139 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9140 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9141 this the server could overwrite ephemeral keys that the client
9145 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9146 using 50 iterations of the Rabin-Miller test.
9148 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9149 iterations of the Rabin-Miller test as required by the appendix
9150 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9151 As BN_is_prime_fasttest includes trial division, DSA parameter
9152 generation becomes much faster.
9154 This implies a change for the callback functions in DSA_is_prime
9155 and DSA_generate_parameters: The callback function is called once
9156 for each positive witness in the Rabin-Miller test, not just
9157 occasionally in the inner loop; and the parameters to the
9158 callback function now provide an iteration count for the outer
9159 loop rather than for the current invocation of the inner loop.
9160 DSA_generate_parameters additionally can call the callback
9161 function with an 'iteration count' of -1, meaning that a
9162 candidate has passed the trial division test (when q is generated
9163 from an application-provided seed, trial division is skipped).
9166 *) New function BN_is_prime_fasttest that optionally does trial
9167 division before starting the Rabin-Miller test and has
9168 an additional BN_CTX * argument (whereas BN_is_prime always
9169 has to allocate at least one BN_CTX).
9170 'callback(1, -1, cb_arg)' is called when a number has passed the
9171 trial division stage.
9174 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9178 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9181 *) New function BN_pseudo_rand().
9184 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9185 bignum version of BN_from_montgomery() with the working code from
9186 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9190 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9191 made it impossible to use the same SSL_SESSION data structure in
9192 SSL2 clients in multiple threads.
9195 *) The return value of RAND_load_file() no longer counts bytes obtained
9196 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9197 to seed the PRNG (previously an explicit byte count was required).
9198 [Ulf Möller, Bodo Möller]
9200 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9201 used (char *) instead of (void *) and had casts all over the place.
9204 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9207 *) Retain source code compatibility for BN_prime_checks macro:
9208 BN_is_prime(..., BN_prime_checks, ...) now uses
9209 BN_prime_checks_for_size to determine the appropriate number of
9210 Rabin-Miller iterations.
9213 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9214 DH_CHECK_P_NOT_SAFE_PRIME.
9215 (Check if this is true? OpenPGP calls them "strong".)
9218 *) Merge the functionality of "dh" and "gendh" programs into a new program
9219 "dhparam". The old programs are retained for now but will handle DH keys
9220 (instead of parameters) in future.
9223 *) Make the ciphers, s_server and s_client programs check the return values
9224 when a new cipher list is set.
9227 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9228 ciphers. Before when the 56bit ciphers were enabled the sorting was
9231 The syntax for the cipher sorting has been extended to support sorting by
9232 cipher-strength (using the strength_bits hard coded in the tables).
9233 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9235 Fix a bug in the cipher-command parser: when supplying a cipher command
9236 string with an "undefined" symbol (neither command nor alphanumeric
9237 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9238 an error is flagged.
9240 Due to the strength-sorting extension, the code of the
9241 ssl_create_cipher_list() function was completely rearranged. I hope that
9242 the readability was also increased :-)
9243 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9245 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9246 for the first serial number and places 2 in the serial number file. This
9247 avoids problems when the root CA is created with serial number zero and
9248 the first user certificate has the same issuer name and serial number
9252 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9253 the new code. Add documentation for this stuff.
9256 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9257 X509_*() to X509at_*() on the grounds that they don't handle X509
9258 structures and behave in an analogous way to the X509v3 functions:
9259 they shouldn't be called directly but wrapper functions should be used
9262 So we also now have some wrapper functions that call the X509at functions
9263 when passed certificate requests. (TO DO: similar things can be done with
9264 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9265 things. Some of these need some d2i or i2d and print functionality
9266 because they handle more complex structures.)
9269 *) Add missing #ifndefs that caused missing symbols when building libssl
9270 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9271 NO_RSA in ssl/s2*.c.
9272 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9274 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9275 has a return value which indicates the quality of the random data
9276 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9277 error queue. New function RAND_pseudo_bytes() generates output that is
9278 guaranteed to be unique but not unpredictable. RAND_add is like
9279 RAND_seed, but takes an extra argument for an entropy estimate
9280 (RAND_seed always assumes full entropy).
9283 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9284 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9285 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9286 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9287 false-positive rate of at most 2^-80 for random input.
9290 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9293 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9294 in the 0.9.5 release), this returns the chain
9295 from an X509_CTX structure with a dup of the stack and all
9296 the X509 reference counts upped: so the stack will exist
9297 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9300 Also make SSL_SESSION_print() print out the verify return
9304 *) Add manpage for the pkcs12 command. Also change the default
9305 behaviour so MAC iteration counts are used unless the new
9306 -nomaciter option is used. This improves file security and
9307 only older versions of MSIE (4.0 for example) need it.
9310 *) Honor the no-xxx Configure options when creating .DEF files.
9313 *) Add PKCS#10 attributes to field table: challengePassword,
9314 unstructuredName and unstructuredAddress. These are taken from
9315 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9316 international characters are used.
9318 More changes to X509_ATTRIBUTE code: allow the setting of types
9319 based on strings. Remove the 'loc' parameter when adding
9320 attributes because these will be a SET OF encoding which is sorted
9324 *) Initial changes to the 'req' utility to allow request generation
9325 automation. This will allow an application to just generate a template
9326 file containing all the field values and have req construct the
9329 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9330 used all over the place including certificate requests and PKCS#7
9331 structures. They are currently handled manually where necessary with
9332 some primitive wrappers for PKCS#7. The new functions behave in a
9333 manner analogous to the X509 extension functions: they allow
9334 attributes to be looked up by NID and added.
9336 Later something similar to the X509V3 code would be desirable to
9337 automatically handle the encoding, decoding and printing of the
9338 more complex types. The string types like challengePassword can
9339 be handled by the string table functions.
9341 Also modified the multi byte string table handling. Now there is
9342 a 'global mask' which masks out certain types. The table itself
9343 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9344 is useful when for example there is only one permissible type
9345 (as in countryName) and using the mask might result in no valid
9349 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9350 SSL_get_peer_finished to allow applications to obtain the latest
9351 Finished messages sent to the peer or expected from the peer,
9352 respectively. (SSL_get_peer_finished is usually the Finished message
9353 actually received from the peer, otherwise the protocol will be aborted.)
9355 As the Finished message are message digests of the complete handshake
9356 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9357 be used for external authentication procedures when the authentication
9358 provided by SSL/TLS is not desired or is not enough.
9361 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9362 the host supports BWX extension and if Compaq C is present on the
9363 $PATH. Just exploiting of the BWX extension results in 20-30%
9364 performance kick for some algorithms, e.g. DES and RC4 to mention
9365 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9369 *) Add support for MS "fast SGC". This is arguably a violation of the
9370 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9371 weak crypto and after checking the certificate is SGC a second one
9372 with strong crypto. MS SGC stops the first handshake after receiving
9373 the server certificate message and sends a second client hello. Since
9374 a server will typically do all the time consuming operations before
9375 expecting any further messages from the client (server key exchange
9376 is the most expensive) there is little difference between the two.
9378 To get OpenSSL to support MS SGC we have to permit a second client
9379 hello message after we have sent server done. In addition we have to
9380 reset the MAC if we do get this second client hello.
9383 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9384 if a DER encoded private key is RSA or DSA traditional format. Changed
9385 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9386 format DER encoded private key. Newer code should use PKCS#8 format which
9387 has the key type encoded in the ASN1 structure. Added DER private key
9388 support to pkcs8 application.
9391 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9392 ciphersuites has been selected (as required by the SSL 3/TLS 1
9393 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9394 is set, we interpret this as a request to violate the specification
9395 (the worst that can happen is a handshake failure, and 'correct'
9396 behaviour would result in a handshake failure anyway).
9399 *) In SSL_CTX_add_session, take into account that there might be multiple
9400 SSL_SESSION structures with the same session ID (e.g. when two threads
9401 concurrently obtain them from an external cache).
9402 The internal cache can handle only one SSL_SESSION with a given ID,
9403 so if there's a conflict, we now throw out the old one to achieve
9407 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9408 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9409 some routines that use cipher OIDs: some ciphers do not have OIDs
9410 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9414 *) Simplify the trust setting structure and code. Now we just have
9415 two sequences of OIDs for trusted and rejected settings. These will
9416 typically have values the same as the extended key usage extension
9417 and any application specific purposes.
9419 The trust checking code now has a default behaviour: it will just
9420 check for an object with the same NID as the passed id. Functions can
9421 be provided to override either the default behaviour or the behaviour
9422 for a given id. SSL client, server and email already have functions
9423 in place for compatibility: they check the NID and also return "trusted"
9424 if the certificate is self signed.
9427 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9428 traditional format into an EVP_PKEY structure.
9431 *) Add a password callback function PEM_cb() which either prompts for
9432 a password if usr_data is NULL or otherwise assumes it is a null
9433 terminated password. Allow passwords to be passed on command line
9434 environment or config files in a few more utilities.
9437 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9438 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9439 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9440 Update documentation.
9443 *) Support for ASN1 "NULL" type. This could be handled before by using
9444 ASN1_TYPE but there wasn't any function that would try to read a NULL
9445 and produce an error if it couldn't. For compatibility we also have
9446 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9447 don't allocate anything because they don't need to.
9450 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9452 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9454 *) Rebuild of the memory allocation routines used by OpenSSL code and
9455 possibly others as well. The purpose is to make an interface that
9456 provide hooks so anyone can build a separate set of allocation and
9457 deallocation routines to be used by OpenSSL, for example memory
9458 pool implementations, or something else, which was previously hard
9459 since Malloc(), Realloc() and Free() were defined as macros having
9460 the values malloc, realloc and free, respectively (except for Win32
9461 compilations). The same is provided for memory debugging code.
9462 OpenSSL already comes with functionality to find memory leaks, but
9463 this gives people a chance to debug other memory problems.
9465 With these changes, a new set of functions and macros have appeared:
9467 CRYPTO_set_mem_debug_functions() [F]
9468 CRYPTO_get_mem_debug_functions() [F]
9469 CRYPTO_dbg_set_options() [F]
9470 CRYPTO_dbg_get_options() [F]
9471 CRYPTO_malloc_debug_init() [M]
9473 The memory debug functions are NULL by default, unless the library
9474 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9475 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9476 gives the standard debugging functions that come with OpenSSL) or
9477 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9478 provided by the library user) must be used. When the standard
9479 debugging functions are used, CRYPTO_dbg_set_options can be used to
9480 request additional information:
9481 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9482 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9484 Also, things like CRYPTO_set_mem_functions will always give the
9485 expected result (the new set of functions is used for allocation
9486 and deallocation) at all times, regardless of platform and compiler
9489 To finish it up, some functions that were never use in any other
9490 way than through macros have a new API and new semantic:
9493 CRYPTO_dbg_realloc()
9496 All macros of value have retained their old syntax.
9497 [Richard Levitte and Bodo Moeller]
9499 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9500 ordering of SMIMECapabilities wasn't in "strength order" and there
9501 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9505 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9506 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9507 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9509 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9510 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9511 functionality to handle multipart/signed properly) and a utility
9512 called 'smime' to call all this stuff. This is based on code I
9513 originally wrote for Celo who have kindly allowed it to be
9514 included in OpenSSL.
9517 *) Add variants des_set_key_checked and des_set_key_unchecked of
9518 des_set_key (aka des_key_sched). Global variable des_check_key
9519 decides which of these is called by des_set_key; this way
9520 des_check_key behaves as it always did, but applications and
9521 the library itself, which was buggy for des_check_key == 1,
9522 have a cleaner way to pick the version they need.
9525 *) New function PKCS12_newpass() which changes the password of a
9529 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9530 dynamic mix. In both cases the ids can be used as an index into the
9531 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9532 functions so they accept a list of the field values and the
9533 application doesn't need to directly manipulate the X509_TRUST
9537 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9541 *) Modify the way the V3 extension code looks up extensions. This now
9542 works in a similar way to the object code: we have some "standard"
9543 extensions in a static table which is searched with OBJ_bsearch()
9544 and the application can add dynamic ones if needed. The file
9545 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9546 updated whenever a new extension is added to the core code and kept
9547 in ext_nid order. There is a simple program 'tabtest.c' which checks
9548 this. New extensions are not added too often so this file can readily
9549 be maintained manually.
9551 There are two big advantages in doing things this way. The extensions
9552 can be looked up immediately and no longer need to be "added" using
9553 X509V3_add_standard_extensions(): this function now does nothing.
9554 [Side note: I get *lots* of email saying the extension code doesn't
9555 work because people forget to call this function]
9556 Also no dynamic allocation is done unless new extensions are added:
9557 so if we don't add custom extensions there is no need to call
9558 X509V3_EXT_cleanup().
9561 *) Modify enc utility's salting as follows: make salting the default. Add a
9562 magic header, so unsalted files fail gracefully instead of just decrypting
9563 to garbage. This is because not salting is a big security hole, so people
9564 should be discouraged from doing it.
9567 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9568 digest to be passed on the command line but it only used this
9569 parameter when signing a certificate. Modified so all relevant
9570 operations are affected by the digest parameter including the
9571 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9572 DSA key was used because it didn't fix the digest.
9575 *) Initial certificate chain verify code. Currently tests the untrusted
9576 certificates for consistency with the verify purpose (which is set
9577 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9579 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9580 this is because it will reject chains with invalid extensions whereas
9581 every previous version of OpenSSL and SSLeay made no checks at all.
9583 Trust code: checks the root CA for the relevant trust settings. Trust
9584 settings have an initial value consistent with the verify purpose: e.g.
9585 if the verify purpose is for SSL client use it expects the CA to be
9586 trusted for SSL client use. However the default value can be changed to
9587 permit custom trust settings: one example of this would be to only trust
9588 certificates from a specific "secure" set of CAs.
9590 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9591 which should be used for version portability: especially since the
9592 verify structure is likely to change more often now.
9594 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9595 to set them. If not set then assume SSL clients will verify SSL servers
9598 Two new options to the verify program: -untrusted allows a set of
9599 untrusted certificates to be passed in and -purpose which sets the
9600 intended purpose of the certificate. If a purpose is set then the
9601 new chain verify code is used to check extension consistency.
9604 *) Support for the authority information access extension.
9607 *) Modify RSA and DSA PEM read routines to transparently handle
9608 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9609 public keys in a format compatible with certificate
9610 SubjectPublicKeyInfo structures. Unfortunately there were already
9611 functions called *_PublicKey_* which used various odd formats so
9612 these are retained for compatibility: however the DSA variants were
9613 never in a public release so they have been deleted. Changed dsa/rsa
9614 utilities to handle the new format: note no releases ever handled public
9615 keys so we should be OK.
9617 The primary motivation for this change is to avoid the same fiasco
9618 that dogs private keys: there are several incompatible private key
9619 formats some of which are standard and some OpenSSL specific and
9620 require various evil hacks to allow partial transparent handling and
9621 even then it doesn't work with DER formats. Given the option anything
9622 other than PKCS#8 should be dumped: but the other formats have to
9623 stay in the name of compatibility.
9625 With public keys and the benefit of hindsight one standard format
9626 is used which works with EVP_PKEY, RSA or DSA structures: though
9627 it clearly returns an error if you try to read the wrong kind of key.
9629 Added a -pubkey option to the 'x509' utility to output the public key.
9630 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9631 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9632 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9633 that do the same as the EVP_PKEY_assign_*() except they up the
9634 reference count of the added key (they don't "swallow" the
9638 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9639 CRLs would fail if the file contained no certificates or no CRLs:
9640 added a new function to read in both types and return the number
9641 read: this means that if none are read it will be an error. The
9642 DER versions of the certificate and CRL reader would always fail
9643 because it isn't possible to mix certificates and CRLs in DER format
9644 without choking one or the other routine. Changed this to just read
9645 a certificate: this is the best we can do. Also modified the code
9646 in apps/verify.c to take notice of return codes: it was previously
9647 attempting to read in certificates from NULL pointers and ignoring
9648 any errors: this is one reason why the cert and CRL reader seemed
9649 to work. It doesn't check return codes from the default certificate
9650 routines: these may well fail if the certificates aren't installed.
9653 *) Code to support otherName option in GeneralName.
9656 *) First update to verify code. Change the verify utility
9657 so it warns if it is passed a self signed certificate:
9658 for consistency with the normal behaviour. X509_verify
9659 has been modified to it will now verify a self signed
9660 certificate if *exactly* the same certificate appears
9661 in the store: it was previously impossible to trust a
9662 single self signed certificate. This means that:
9663 openssl verify ss.pem
9664 now gives a warning about a self signed certificate but
9665 openssl verify -CAfile ss.pem ss.pem
9669 *) For servers, store verify_result in SSL_SESSION data structure
9670 (and add it to external session representation).
9671 This is needed when client certificate verifications fails,
9672 but an application-provided verification callback (set by
9673 SSL_CTX_set_cert_verify_callback) allows accepting the session
9674 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9675 but returns 1): When the session is reused, we have to set
9676 ssl->verify_result to the appropriate error code to avoid
9678 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9680 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9681 case in PKCS7_dataInit() where the signed PKCS7 structure
9682 didn't contain any existing data because it was being created.
9683 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9685 *) Add a salt to the key derivation routines in enc.c. This
9686 forms the first 8 bytes of the encrypted file. Also add a
9687 -S option to allow a salt to be input on the command line.
9690 *) New function X509_cmp(). Oddly enough there wasn't a function
9691 to compare two certificates. We do this by working out the SHA1
9692 hash and comparing that. X509_cmp() will be needed by the trust
9696 *) SSL_get1_session() is like SSL_get_session(), but increments
9697 the reference count in the SSL_SESSION returned.
9698 [Geoff Thorpe <geoff@eu.c2.net>]
9700 *) Fix for 'req': it was adding a null to request attributes.
9701 Also change the X509_LOOKUP and X509_INFO code to handle
9702 certificate auxiliary information.
9705 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9709 *) Add the possibility to add extra information to the memory leak
9710 detecting output, to form tracebacks, showing from where each
9711 allocation was originated: CRYPTO_push_info("constant string") adds
9712 the string plus current file name and line number to a per-thread
9713 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9714 is like calling CYRPTO_pop_info() until the stack is empty.
9715 Also updated memory leak detection code to be multi-thread-safe.
9718 *) Add options -text and -noout to pkcs7 utility and delete the
9719 encryption options which never did anything. Update docs.
9722 *) Add options to some of the utilities to allow the pass phrase
9723 to be included on either the command line (not recommended on
9724 OSes like Unix) or read from the environment. Update the
9725 manpages and fix a few bugs.
9728 *) Add a few manpages for some of the openssl commands.
9731 *) Fix the -revoke option in ca. It was freeing up memory twice,
9732 leaking and not finding already revoked certificates.
9735 *) Extensive changes to support certificate auxiliary information.
9736 This involves the use of X509_CERT_AUX structure and X509_AUX
9737 functions. An X509_AUX function such as PEM_read_X509_AUX()
9738 can still read in a certificate file in the usual way but it
9739 will also read in any additional "auxiliary information". By
9740 doing things this way a fair degree of compatibility can be
9741 retained: existing certificates can have this information added
9742 using the new 'x509' options.
9744 Current auxiliary information includes an "alias" and some trust
9745 settings. The trust settings will ultimately be used in enhanced
9746 certificate chain verification routines: currently a certificate
9747 can only be trusted if it is self signed and then it is trusted
9751 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9752 The problem was that one of the replacement routines had not been working
9753 since SSLeay releases. For now the offending routine has been replaced
9754 with non-optimised assembler. Even so, this now gives around 95%
9755 performance improvement for 1024 bit RSA signs.
9758 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9759 handling. Most clients have the effective key size in bits equal to
9760 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9761 A few however don't do this and instead use the size of the decrypted key
9762 to determine the RC2 key length and the AlgorithmIdentifier to determine
9763 the effective key length. In this case the effective key length can still
9764 be 40 bits but the key length can be 168 bits for example. This is fixed
9765 by manually forcing an RC2 key into the EVP_PKEY structure because the
9766 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9767 the key length and effective key length are equal.
9770 *) Add a bunch of functions that should simplify the creation of
9771 X509_NAME structures. Now you should be able to do:
9772 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9773 and have it automatically work out the correct field type and fill in
9774 the structures. The more adventurous can try:
9775 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9776 and it will (hopefully) work out the correct multibyte encoding.
9779 *) Change the 'req' utility to use the new field handling and multibyte
9780 copy routines. Before the DN field creation was handled in an ad hoc
9781 way in req, ca, and x509 which was rather broken and didn't support
9782 BMPStrings or UTF8Strings. Since some software doesn't implement
9783 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9784 using the dirstring_type option. See the new comment in the default
9785 openssl.cnf for more info.
9788 *) Make crypto/rand/md_rand.c more robust:
9789 - Assure unique random numbers after fork().
9790 - Make sure that concurrent threads access the global counter and
9791 md serializably so that we never lose entropy in them
9792 or use exactly the same state in multiple threads.
9793 Access to the large state is not always serializable because
9794 the additional locking could be a performance killer, and
9795 md should be large enough anyway.
9798 *) New file apps/app_rand.c with commonly needed functionality
9799 for handling the random seed file.
9801 Use the random seed file in some applications that previously did not:
9803 dsaparam -genkey (which also ignored its '-rand' option),
9806 x509 (when signing).
9807 Except on systems with /dev/urandom, it is crucial to have a random
9808 seed file at least for key creation, DSA signing, and for DH exchanges;
9809 for RSA signatures we could do without one.
9811 gendh and gendsa (unlike genrsa) used to read only the first byte
9812 of each file listed in the '-rand' option. The function as previously
9813 found in genrsa is now in app_rand.c and is used by all programs
9814 that support '-rand'.
9817 *) In RAND_write_file, use mode 0600 for creating files;
9818 don't just chmod when it may be too late.
9821 *) Report an error from X509_STORE_load_locations
9822 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9825 *) New function ASN1_mbstring_copy() this copies a string in either
9826 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9827 into an ASN1_STRING type. A mask of permissible types is passed
9828 and it chooses the "minimal" type to use or an error if not type
9832 *) Add function equivalents to the various macros in asn1.h. The old
9833 macros are retained with an M_ prefix. Code inside the library can
9834 use the M_ macros. External code (including the openssl utility)
9835 should *NOT* in order to be "shared library friendly".
9838 *) Add various functions that can check a certificate's extensions
9839 to see if it usable for various purposes such as SSL client,
9840 server or S/MIME and CAs of these types. This is currently
9841 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9842 verification. Also added a -purpose flag to x509 utility to
9843 print out all the purposes.
9846 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9850 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9851 for, obtain and decode and extension and obtain its critical flag.
9852 This allows all the necessary extension code to be handled in a
9853 single function call.
9856 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9857 platforms. See crypto/rc4/rc4_enc.c for further details.
9860 *) New -noout option to asn1parse. This causes no output to be produced
9861 its main use is when combined with -strparse and -out to extract data
9862 from a file (which may not be in ASN.1 format).
9865 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9866 when producing the local key id.
9867 [Richard Levitte <levitte@stacken.kth.se>]
9869 *) New option -dhparam in s_server. This allows a DH parameter file to be
9870 stated explicitly. If it is not stated then it tries the first server
9871 certificate file. The previous behaviour hard coded the filename
9875 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9876 a public key to be input or output. For example:
9877 openssl rsa -in key.pem -pubout -out pubkey.pem
9878 Also added necessary DSA public key functions to handle this.
9881 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9882 in the message. This was handled by allowing
9883 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9884 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9886 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9887 to the end of the strings whereas this didn't. This would cause problems
9888 if strings read with d2i_ASN1_bytes() were later modified.
9889 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9891 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9892 data and it contains EOF it will end up returning an error. This is
9893 caused by input 46 bytes long. The cause is due to the way base64
9894 BIOs find the start of base64 encoded data. They do this by trying a
9895 trial decode on each line until they find one that works. When they
9896 do a flag is set and it starts again knowing it can pass all the
9897 data directly through the decoder. Unfortunately it doesn't reset
9898 the context it uses. This means that if EOF is reached an attempt
9899 is made to pass two EOFs through the context and this causes the
9900 resulting error. This can also cause other problems as well. As is
9901 usual with these problems it takes *ages* to find and the fix is
9902 trivial: move one line.
9903 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9905 *) Ugly workaround to get s_client and s_server working under Windows. The
9906 old code wouldn't work because it needed to select() on sockets and the
9907 tty (for keypresses and to see if data could be written). Win32 only
9908 supports select() on sockets so we select() with a 1s timeout on the
9909 sockets and then see if any characters are waiting to be read, if none
9910 are present then we retry, we also assume we can always write data to
9911 the tty. This isn't nice because the code then blocks until we've
9912 received a complete line of data and it is effectively polling the
9913 keyboard at 1s intervals: however it's quite a bit better than not
9914 working at all :-) A dedicated Windows application might handle this
9915 with an event loop for example.
9918 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9919 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9920 will be called when RSA_sign() and RSA_verify() are used. This is useful
9921 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9922 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9923 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9924 This necessitated the support of an extra signature type NID_md5_sha1
9925 for SSL signatures and modifications to the SSL library to use it instead
9926 of calling RSA_public_decrypt() and RSA_private_encrypt().
9929 *) Add new -verify -CAfile and -CApath options to the crl program, these
9930 will lookup a CRL issuers certificate and verify the signature in a
9931 similar way to the verify program. Tidy up the crl program so it
9932 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9933 less strict. It will now permit CRL extensions even if it is not
9934 a V2 CRL: this will allow it to tolerate some broken CRLs.
9937 *) Initialize all non-automatic variables each time one of the openssl
9938 sub-programs is started (this is necessary as they may be started
9939 multiple times from the "OpenSSL>" prompt).
9940 [Lennart Bang, Bodo Moeller]
9942 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9943 removing all other RSA functionality (this is what NO_RSA does). This
9944 is so (for example) those in the US can disable those operations covered
9945 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9949 *) Non-copying interface to BIO pairs.
9950 (still largely untested)
9953 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9954 ASCII string. This was handled independently in various places before.
9957 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9958 UTF8 strings a character at a time.
9961 *) Use client_version from client hello to select the protocol
9962 (s23_srvr.c) and for RSA client key exchange verification
9963 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9966 *) Add various utility functions to handle SPKACs, these were previously
9967 handled by poking round in the structure internals. Added new function
9968 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9969 print, verify and generate SPKACs. Based on an original idea from
9970 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9973 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9976 *) Allow the config file extension section to be overwritten on the
9977 command line. Based on an original idea from Massimiliano Pala
9978 <madwolf@comune.modena.it>. The new option is called -extensions
9979 and can be applied to ca, req and x509. Also -reqexts to override
9980 the request extensions in req and -crlexts to override the crl extensions
9984 *) Add new feature to the SPKAC handling in ca. Now you can include
9985 the same field multiple times by preceding it by "XXXX." for example:
9988 this is the same syntax as used in the req config file.
9991 *) Allow certificate extensions to be added to certificate requests. These
9992 are specified in a 'req_extensions' option of the req section of the
9993 config file. They can be printed out with the -text option to req but
9994 are otherwise ignored at present.
9997 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9998 data read consists of only the final block it would not decrypted because
9999 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10000 A misplaced 'break' also meant the decrypted final block might not be
10001 copied until the next read.
10004 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10005 a few extra parameters to the DH structure: these will be useful if
10006 for example we want the value of 'q' or implement X9.42 DH.
10009 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10010 provides hooks that allow the default DSA functions or functions on a
10011 "per key" basis to be replaced. This allows hardware acceleration and
10012 hardware key storage to be handled without major modification to the
10013 library. Also added low level modexp hooks and CRYPTO_EX structure and
10014 associated functions.
10017 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10018 as "read only": it can't be written to and the buffer it points to will
10019 not be freed. Reading from a read only BIO is much more efficient than
10020 a normal memory BIO. This was added because there are several times when
10021 an area of memory needs to be read from a BIO. The previous method was
10022 to create a memory BIO and write the data to it, this results in two
10023 copies of the data and an O(n^2) reading algorithm. There is a new
10024 function BIO_new_mem_buf() which creates a read only memory BIO from
10025 an area of memory. Also modified the PKCS#7 routines to use read only
10029 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10030 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10031 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10032 but a retry condition occured while trying to read the rest.
10035 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10036 NID_pkcs7_encrypted by default: this was wrong since this should almost
10037 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10038 the encrypted data type: this is a more sensible place to put it and it
10039 allows the PKCS#12 code to be tidied up that duplicated this
10043 *) Changed obj_dat.pl script so it takes its input and output files on
10044 the command line. This should avoid shell escape redirection problems
10048 *) Initial support for certificate extension requests, these are included
10049 in things like Xenroll certificate requests. Included functions to allow
10050 extensions to be obtained and added.
10053 *) -crlf option to s_client and s_server for sending newlines as
10054 CRLF (as required by many protocols).
10057 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10059 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10060 [Ralf S. Engelschall]
10062 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10063 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10065 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10069 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10070 DH parameters/keys (q is lost during that conversion, but the resulting
10071 DH parameters contain its length).
10073 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10074 much faster than DH_generate_parameters (which creates parameters
10075 where p = 2*q + 1), and also the smaller q makes DH computations
10076 much more efficient (160-bit exponentiation instead of 1024-bit
10077 exponentiation); so this provides a convenient way to support DHE
10078 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10079 utter importance to use
10080 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10082 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10083 when such DH parameters are used, because otherwise small subgroup
10084 attacks may become possible!
10087 *) Avoid memory leak in i2d_DHparams.
10090 *) Allow the -k option to be used more than once in the enc program:
10091 this allows the same encrypted message to be read by multiple recipients.
10094 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10095 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10096 it will always use the numerical form of the OID, even if it has a short
10100 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10101 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10102 otherwise bn_mod_exp was called. In the case of hardware keys for example
10103 no private key components need be present and it might store extra data
10104 in the RSA structure, which cannot be accessed from bn_mod_exp.
10105 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10106 private key operations.
10109 *) Added support for SPARC Linux.
10112 *) pem_password_cb function type incompatibly changed from
10113 typedef int pem_password_cb(char *buf, int size, int rwflag);
10115 ....(char *buf, int size, int rwflag, void *userdata);
10116 so that applications can pass data to their callbacks:
10117 The PEM[_ASN1]_{read,write}... functions and macros now take an
10118 additional void * argument, which is just handed through whenever
10119 the password callback is called.
10120 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10122 New function SSL_CTX_set_default_passwd_cb_userdata.
10124 Compatibility note: As many C implementations push function arguments
10125 onto the stack in reverse order, the new library version is likely to
10126 interoperate with programs that have been compiled with the old
10127 pem_password_cb definition (PEM_whatever takes some data that
10128 happens to be on the stack as its last argument, and the callback
10129 just ignores this garbage); but there is no guarantee whatsoever that
10132 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10133 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10134 problems not only on Windows, but also on some Unix platforms.
10135 To avoid problematic command lines, these definitions are now in an
10136 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10137 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10140 *) MIPS III/IV assembler module is reimplemented.
10143 *) More DES library cleanups: remove references to srand/rand and
10144 delete an unused file.
10147 *) Add support for the the free Netwide assembler (NASM) under Win32,
10148 since not many people have MASM (ml) and it can be hard to obtain.
10149 This is currently experimental but it seems to work OK and pass all
10150 the tests. Check out INSTALL.W32 for info.
10153 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10154 without temporary keys kept an extra copy of the server key,
10155 and connections with temporary keys did not free everything in case
10159 *) New function RSA_check_key and new openssl rsa option -check
10160 for verifying the consistency of RSA keys.
10161 [Ulf Moeller, Bodo Moeller]
10163 *) Various changes to make Win32 compile work:
10164 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10165 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10166 comparison" warnings.
10167 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10170 *) Add a debugging option to PKCS#5 v2 key generation function: when
10171 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10172 derived keys are printed to stderr.
10175 *) Copy the flags in ASN1_STRING_dup().
10176 [Roman E. Pavlov <pre@mo.msk.ru>]
10178 *) The x509 application mishandled signing requests containing DSA
10179 keys when the signing key was also DSA and the parameters didn't match.
10181 It was supposed to omit the parameters when they matched the signing key:
10182 the verifying software was then supposed to automatically use the CA's
10183 parameters if they were absent from the end user certificate.
10185 Omitting parameters is no longer recommended. The test was also
10186 the wrong way round! This was probably due to unusual behaviour in
10187 EVP_cmp_parameters() which returns 1 if the parameters match.
10188 This meant that parameters were omitted when they *didn't* match and
10189 the certificate was useless. Certificates signed with 'ca' didn't have
10191 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10193 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10194 The interface is as follows:
10195 Applications can use
10196 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10197 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10198 "off" is now the default.
10199 The library internally uses
10200 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10201 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10202 to disable memory-checking temporarily.
10204 Some inconsistent states that previously were possible (and were
10205 even the default) are now avoided.
10207 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10208 with each memory chunk allocated; this is occasionally more helpful
10209 than just having a counter.
10211 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10213 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10217 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10218 which largely parallels "options", but is for changing API behaviour,
10219 whereas "options" are about protocol behaviour.
10220 Initial "mode" flags are:
10222 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10223 a single record has been written.
10224 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10225 retries use the same buffer location.
10226 (But all of the contents must be
10230 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10233 *) Fix problems with no-hmac etc.
10234 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10236 *) New functions RSA_get_default_method(), RSA_set_method() and
10237 RSA_get_method(). These allows replacement of RSA_METHODs without having
10238 to mess around with the internals of an RSA structure.
10241 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10242 Also really enable memory leak checks in openssl.c and in some
10244 [Chad C. Mulligan, Bodo Moeller]
10246 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10247 up the length of negative integers. This has now been simplified to just
10248 store the length when it is first determined and use it later, rather
10249 than trying to keep track of where data is copied and updating it to
10251 [Steve Henson, reported by Brien Wheeler
10252 <bwheeler@authentica-security.com>]
10254 *) Add a new function PKCS7_signatureVerify. This allows the verification
10255 of a PKCS#7 signature but with the signing certificate passed to the
10256 function itself. This contrasts with PKCS7_dataVerify which assumes the
10257 certificate is present in the PKCS#7 structure. This isn't always the
10258 case: certificates can be omitted from a PKCS#7 structure and be
10259 distributed by "out of band" means (such as a certificate database).
10262 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10263 function prototypes in pem.h, also change util/mkdef.pl to add the
10264 necessary function names.
10267 *) mk1mf.pl (used by Windows builds) did not properly read the
10268 options set by Configure in the top level Makefile, and Configure
10269 was not even able to write more than one option correctly.
10270 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10273 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10274 file to be loaded from a BIO or FILE pointer. The BIO version will
10275 for example allow memory BIOs to contain config info.
10278 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10279 Whoever hopes to achieve shared-library compatibility across versions
10280 must use this, not the compile-time macro.
10281 (Exercise 0.9.4: Which is the minimum library version required by
10283 Note: All this applies only to multi-threaded programs, others don't
10287 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10288 through a BIO pair triggered the default case, i.e.
10289 SSLerr(...,SSL_R_UNKNOWN_STATE).
10292 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10293 can use the SSL library even if none of the specific BIOs is
10297 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10298 for the encoded length.
10299 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10301 *) Add initial documentation of the X509V3 functions.
10304 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10305 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10306 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10307 secure PKCS#8 private key format with a high iteration count.
10310 *) Fix determination of Perl interpreter: A perl or perl5
10311 _directory_ in $PATH was also accepted as the interpreter.
10312 [Ralf S. Engelschall]
10314 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10315 wrong with it but it was very old and did things like calling
10316 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10317 unusual formatting.
10320 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10321 to use the new extension code.
10324 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10325 with macros. This should make it easier to change their form, add extra
10326 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10330 *) Add to configuration table a new entry that can specify an alternative
10331 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10332 according to Mark Crispin <MRC@Panda.COM>.
10336 *) DES CBC did not update the IV. Weird.
10339 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10340 Changing the behaviour of the former might break existing programs --
10341 where IV updating is needed, des_ncbc_encrypt can be used.
10344 *) When bntest is run from "make test" it drives bc to check its
10345 calculations, as well as internally checking them. If an internal check
10346 fails, it needs to cause bc to give a non-zero result or make test carries
10347 on without noticing the failure. Fixed.
10350 *) DES library cleanups.
10353 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10354 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10355 ciphers. NOTE: although the key derivation function has been verified
10356 against some published test vectors it has not been extensively tested
10357 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10361 *) Instead of "mkdir -p", which is not fully portable, use new
10362 Perl script "util/mkdir-p.pl".
10365 *) Rewrite the way password based encryption (PBE) is handled. It used to
10366 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10367 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10368 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10369 the 'parameter' field of the AlgorithmIdentifier is passed to the
10370 underlying key generation function so it must do its own ASN1 parsing.
10371 This has also changed the EVP_PBE_CipherInit() function which now has a
10372 'parameter' argument instead of literal salt and iteration count values
10373 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10376 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10377 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10378 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10379 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10380 value was just used as a "magic string" and not used directly its
10381 value doesn't matter.
10384 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10388 *) "linux-sparc64" configuration (ultrapenguin).
10389 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10390 "linux-sparc" configuration.
10391 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10393 *) config now generates no-xxx options for missing ciphers.
10396 *) Support the EBCDIC character set (work in progress).
10397 File ebcdic.c not yet included because it has a different license.
10398 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10400 *) Support BS2000/OSD-POSIX.
10401 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10403 *) Make callbacks for key generation use void * instead of char *.
10406 *) Make S/MIME samples compile (not yet tested).
10409 *) Additional typesafe stacks.
10412 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10416 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10418 *) New configuration variant "sco5-gcc".
10420 *) Updated some demos.
10421 [Sean O Riordain, Wade Scholine]
10423 *) Add missing BIO_free at exit of pkcs12 application.
10426 *) Fix memory leak in conf.c.
10429 *) Updates for Win32 to assembler version of MD5.
10432 *) Set #! path to perl in apps/der_chop to where we found it
10433 instead of using a fixed path.
10436 *) SHA library changes for irix64-mips4-cc.
10439 *) Improvements for VMS support.
10443 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10445 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10446 This also avoids the problems with SC4.2 and unpatched SC5.
10447 [Andy Polyakov <appro@fy.chalmers.se>]
10449 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10450 These are required because of the typesafe stack would otherwise break
10451 existing code. If old code used a structure member which used to be STACK
10452 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10453 sk_num or sk_value it would produce an error because the num, data members
10454 are not present in STACK_OF. Now it just produces a warning. sk_set
10455 replaces the old method of assigning a value to sk_value
10456 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10457 that does this will no longer work (and should use sk_set instead) but
10458 this could be regarded as a "questionable" behaviour anyway.
10461 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10462 correctly handle encrypted S/MIME data.
10465 *) Change type of various DES function arguments from des_cblock
10466 (which means, in function argument declarations, pointer to char)
10467 to des_cblock * (meaning pointer to array with 8 char elements),
10468 which allows the compiler to do more typechecking; it was like
10469 that back in SSLeay, but with lots of ugly casts.
10471 Introduce new type const_des_cblock.
10474 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10475 problems: find RecipientInfo structure that matches recipient certificate
10476 and initialise the ASN1 structures properly based on passed cipher.
10479 *) Belatedly make the BN tests actually check the results.
10482 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10483 to and from BNs: it was completely broken. New compilation option
10484 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10485 key elements as negative integers.
10488 *) Reorganize and speed up MD5.
10489 [Andy Polyakov <appro@fy.chalmers.se>]
10492 [Richard Levitte <richard@levitte.org>]
10494 *) New option -out to asn1parse to allow the parsed structure to be
10495 output to a file. This is most useful when combined with the -strparse
10496 option to examine the output of things like OCTET STRINGS.
10499 *) Make SSL library a little more fool-proof by not requiring any longer
10500 that SSL_set_{accept,connect}_state be called before
10501 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10502 in many applications because usually everything *appeared* to work as
10503 intended anyway -- now it really works as intended).
10506 *) Move openssl.cnf out of lib/.
10509 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10510 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10511 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10512 [Ralf S. Engelschall]
10514 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10515 handle PKCS#7 enveloped data properly.
10516 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10518 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10519 copying pointers. The cert_st handling is changed by this in
10520 various ways (and thus what used to be known as ctx->default_cert
10521 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10522 any longer when s->cert does not give us what we need).
10523 ssl_cert_instantiate becomes obsolete by this change.
10524 As soon as we've got the new code right (possibly it already is?),
10525 we have solved a couple of bugs of the earlier code where s->cert
10526 was used as if it could not have been shared with other SSL structures.
10528 Note that using the SSL API in certain dirty ways now will result
10529 in different behaviour than observed with earlier library versions:
10530 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10531 does not influence s as it used to.
10533 In order to clean up things more thoroughly, inside SSL_SESSION
10534 we don't use CERT any longer, but a new structure SESS_CERT
10535 that holds per-session data (if available); currently, this is
10536 the peer's certificate chain and, for clients, the server's certificate
10537 and temporary key. CERT holds only those values that can have
10538 meaningful defaults in an SSL_CTX.
10541 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10542 from the internal representation. Various PKCS#7 fixes: remove some
10543 evil casts and set the enc_dig_alg field properly based on the signing
10547 *) Allow PKCS#12 password to be set from the command line or the
10548 environment. Let 'ca' get its config file name from the environment
10549 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10553 *) Allow certificate policies extension to use an IA5STRING for the
10554 organization field. This is contrary to the PKIX definition but
10555 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10559 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10560 without disallowing inline assembler and the like for non-pedantic builds.
10563 *) Support Borland C++ builder.
10564 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10566 *) Support Mingw32.
10569 *) SHA-1 cleanups and performance enhancements.
10570 [Andy Polyakov <appro@fy.chalmers.se>]
10572 *) Sparc v8plus assembler for the bignum library.
10573 [Andy Polyakov <appro@fy.chalmers.se>]
10575 *) Accept any -xxx and +xxx compiler options in Configure.
10578 *) Update HPUX configuration.
10581 *) Add missing sk_<type>_unshift() function to safestack.h
10582 [Ralf S. Engelschall]
10584 *) New function SSL_CTX_use_certificate_chain_file that sets the
10585 "extra_cert"s in addition to the certificate. (This makes sense
10586 only for "PEM" format files, as chains as a whole are not
10590 *) Support verify_depth from the SSL API.
10591 x509_vfy.c had what can be considered an off-by-one-error:
10592 Its depth (which was not part of the external interface)
10593 was actually counting the number of certificates in a chain;
10594 now it really counts the depth.
10597 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10598 instead of X509err, which often resulted in confusing error
10599 messages since the error codes are not globally unique
10600 (e.g. an alleged error in ssl3_accept when a certificate
10601 didn't match the private key).
10603 *) New function SSL_CTX_set_session_id_context that allows to set a default
10604 value (so that you don't need SSL_set_session_id_context for each
10605 connection using the SSL_CTX).
10608 *) OAEP decoding bug fix.
10611 *) Support INSTALL_PREFIX for package builders, as proposed by
10615 *) New Configure options "threads" and "no-threads". For systems
10616 where the proper compiler options are known (currently Solaris
10617 and Linux), "threads" is the default.
10620 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10623 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10624 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10625 such as /usr/local/bin.
10628 *) "make linux-shared" to build shared libraries.
10629 [Niels Poppe <niels@netbox.org>]
10631 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10634 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10635 extension adding in x509 utility.
10638 *) Remove NOPROTO sections and error code comments.
10641 *) Partial rewrite of the DEF file generator to now parse the ANSI
10645 *) New Configure options --prefix=DIR and --openssldir=DIR.
10648 *) Complete rewrite of the error code script(s). It is all now handled
10649 by one script at the top level which handles error code gathering,
10650 header rewriting and C source file generation. It should be much better
10651 than the old method: it now uses a modified version of Ulf's parser to
10652 read the ANSI prototypes in all header files (thus the old K&R definitions
10653 aren't needed for error creation any more) and do a better job of
10654 translating function codes into names. The old 'ASN1 error code imbedded
10655 in a comment' is no longer necessary and it doesn't use .err files which
10656 have now been deleted. Also the error code call doesn't have to appear all
10657 on one line (which resulted in some large lines...).
10660 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10663 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10664 0 (which usually indicates a closed connection), but continue reading.
10667 *) Fix some race conditions.
10670 *) Add support for CRL distribution points extension. Add Certificate
10671 Policies and CRL distribution points documentation.
10674 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10677 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10678 8 of keying material. Merlin has also confirmed interop with this fix
10679 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10680 [Merlin Hughes <merlin@baltimore.ie>]
10682 *) Fix lots of warnings.
10683 [Richard Levitte <levitte@stacken.kth.se>]
10685 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10686 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10687 [Richard Levitte <levitte@stacken.kth.se>]
10689 *) Fix problems with sizeof(long) == 8.
10690 [Andy Polyakov <appro@fy.chalmers.se>]
10692 *) Change functions to ANSI C.
10695 *) Fix typos in error codes.
10696 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10698 *) Remove defunct assembler files from Configure.
10701 *) SPARC v8 assembler BIGNUM implementation.
10702 [Andy Polyakov <appro@fy.chalmers.se>]
10704 *) Support for Certificate Policies extension: both print and set.
10705 Various additions to support the r2i method this uses.
10708 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10709 return a const string when you are expecting an allocated buffer.
10712 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10713 types DirectoryString and DisplayText.
10716 *) Add code to allow r2i extensions to access the configuration database,
10717 add an LHASH database driver and add several ctx helper functions.
10720 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10721 fail when they extended the size of a BIGNUM.
10724 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10725 support typesafe stack.
10728 *) Fix typo in SSL_[gs]et_options().
10729 [Nils Frostberg <nils@medcom.se>]
10731 *) Delete various functions and files that belonged to the (now obsolete)
10732 old X509V3 handling code.
10735 *) New Configure option "rsaref".
10738 *) Don't auto-generate pem.h.
10741 *) Introduce type-safe ASN.1 SETs.
10744 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10745 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10747 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10748 that links with OpenSSL (well at least cause lots of warnings), but fear
10749 not: the conversion is trivial, and it eliminates loads of evil casts. A
10750 few STACKed things have been converted already. Feel free to convert more.
10751 In the fullness of time, I'll do away with the STACK type altogether.
10754 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10755 specified in <certfile> by updating the entry in the index.txt file.
10756 This way one no longer has to edit the index.txt file manually for
10757 revoking a certificate. The -revoke option does the gory details now.
10758 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10760 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10761 `-text' option at all and this way the `-noout -text' combination was
10762 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10763 [Ralf S. Engelschall]
10765 *) Make sure a corresponding plain text error message exists for the
10766 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10767 verify callback function determined that a certificate was revoked.
10768 [Ralf S. Engelschall]
10770 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10771 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10772 all available cipers including rc5, which was forgotten until now.
10773 In order to let the testing shell script know which algorithms
10774 are available, a new (up to now undocumented) command
10775 "openssl list-cipher-commands" is used.
10778 *) Bugfix: s_client occasionally would sleep in select() when
10779 it should have checked SSL_pending() first.
10782 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10783 the raw DSA values prior to ASN.1 encoding.
10786 *) Tweaks to Configure
10787 [Niels Poppe <niels@netbox.org>]
10789 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10793 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10796 *) New config option to avoid instructions that are illegal on the 80386.
10797 The default code is faster, but requires at least a 486.
10800 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10801 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10802 same as SSL2_VERSION anyway.
10805 *) New "-showcerts" option for s_client.
10808 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10809 application. Various cleanups and fixes.
10812 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10813 modify error routines to work internally. Add error codes and PBE init
10814 to library startup routines.
10817 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10818 packing functions to asn1 and evp. Changed function names and error
10819 codes along the way.
10822 *) PKCS12 integration: and so it begins... First of several patches to
10823 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10824 objects to objects.h
10827 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10828 and display support for Thawte strong extranet extension.
10831 *) Add LinuxPPC support.
10832 [Jeff Dubrule <igor@pobox.org>]
10834 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10835 bn_div_words in alpha.s.
10836 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10838 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10839 OAEP isn't supported when OpenSSL is built with RSAref.
10840 [Ulf Moeller <ulf@fitug.de>]
10842 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10843 so they no longer are missing under -DNOPROTO.
10844 [Soren S. Jorvang <soren@t.dk>]
10847 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10849 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10850 doesn't work when the session is reused. Coming soon!
10853 *) Fix a security hole, that allows sessions to be reused in the wrong
10854 context thus bypassing client cert protection! All software that uses
10855 client certs and session caches in multiple contexts NEEDS PATCHING to
10856 allow session reuse! A fuller solution is in the works.
10857 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10859 *) Some more source tree cleanups (removed obsolete files
10860 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10861 permission on "config" script to be executable) and a fix for the INSTALL
10863 [Ulf Moeller <ulf@fitug.de>]
10865 *) Remove some legacy and erroneous uses of malloc, free instead of
10867 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10869 *) Make rsa_oaep_test return non-zero on error.
10870 [Ulf Moeller <ulf@fitug.de>]
10872 *) Add support for native Solaris shared libraries. Configure
10873 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10874 if someone would make that last step automatic.
10875 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10877 *) ctx_size was not built with the right compiler during "make links". Fixed.
10880 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10881 except NULL ciphers". This means the default cipher list will no longer
10882 enable NULL ciphers. They need to be specifically enabled e.g. with
10883 the string "DEFAULT:eNULL".
10886 *) Fix to RSA private encryption routines: if p < q then it would
10887 occasionally produce an invalid result. This will only happen with
10888 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10891 *) Be less restrictive and allow also `perl util/perlpath.pl
10892 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10893 because this way one can also use an interpreter named `perl5' (which is
10894 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10895 installed as `perl').
10896 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10898 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10899 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10901 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10902 advapi32.lib to Win32 build and change the pem test comparison
10903 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10904 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10905 and crypto/des/ede_cbcm_enc.c.
10908 *) DES quad checksum was broken on big-endian architectures. Fixed.
10911 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10912 Win32 test batch file so it (might) work again. The Win32 test batch file
10913 is horrible: I feel ill....
10916 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10917 in e_os.h. Audit of header files to check ANSI and non ANSI
10918 sections: 10 functions were absent from non ANSI section and not exported
10919 from Windows DLLs. Fixed up libeay.num for new functions.
10922 *) Make `openssl version' output lines consistent.
10923 [Ralf S. Engelschall]
10925 *) Fix Win32 symbol export lists for BIO functions: Added
10926 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10927 to ms/libeay{16,32}.def.
10928 [Ralf S. Engelschall]
10930 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10931 fine under Unix and passes some trivial tests I've now added. But the
10932 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10933 added to make sure no one expects that this stuff really works in the
10934 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10935 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10937 [Ralf S. Engelschall]
10939 *) Fix the generation of two part addresses in perl.
10940 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10942 *) Add config entry for Linux on MIPS.
10943 [John Tobey <jtobey@channel1.com>]
10945 *) Make links whenever Configure is run, unless we are on Windoze.
10948 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10949 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10953 *) Add a useful kludge to allow package maintainers to specify compiler and
10954 other platforms details on the command line without having to patch the
10955 Configure script everytime: One now can use ``perl Configure
10956 <id>:<details>'', i.e. platform ids are allowed to have details appended
10957 to them (separated by colons). This is treated as there would be a static
10958 pre-configured entry in Configure's %table under key <id> with value
10959 <details> and ``perl Configure <id>'' is called. So, when you want to
10960 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10961 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10962 now, which overrides the FreeBSD-elf entry on-the-fly.
10963 [Ralf S. Engelschall]
10965 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10968 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10969 on the `perl Configure ...' command line. This way one can compile
10970 OpenSSL libraries with Position Independent Code (PIC) which is needed
10971 for linking it into DSOs.
10972 [Ralf S. Engelschall]
10974 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10978 *) Cleaned up the LICENSE document: The official contact for any license
10979 questions now is the OpenSSL core team under openssl-core@openssl.org.
10980 And add a paragraph about the dual-license situation to make sure people
10981 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10982 to the OpenSSL toolkit.
10983 [Ralf S. Engelschall]
10985 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10986 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10987 Additionally cleaned up the `make links' target: Remove unnecessary
10988 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10989 to speed processing and no longer clutter the display with confusing
10990 stuff. Instead only the actually done links are displayed.
10991 [Ralf S. Engelschall]
10993 *) Permit null encryption ciphersuites, used for authentication only. It used
10994 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10995 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10999 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11000 signed attributes when verifying signatures (this would break them),
11001 the detached data encoding was wrong and public keys obtained using
11002 X509_get_pubkey() weren't freed.
11005 *) Add text documentation for the BUFFER functions. Also added a work around
11006 to a Win95 console bug. This was triggered by the password read stuff: the
11007 last character typed gets carried over to the next fread(). If you were
11008 generating a new cert request using 'req' for example then the last
11009 character of the passphrase would be CR which would then enter the first
11013 *) Added the new `Includes OpenSSL Cryptography Software' button as
11014 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11015 button and can be used by applications based on OpenSSL to show the
11016 relationship to the OpenSSL project.
11017 [Ralf S. Engelschall]
11019 *) Remove confusing variables in function signatures in files
11020 ssl/ssl_lib.c and ssl/ssl.h.
11021 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11023 *) Don't install bss_file.c under PREFIX/include/
11024 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11026 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11027 functions that return function pointers and has support for NT specific
11028 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11029 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11030 unsigned to signed types: this was killing the Win32 compile.
11033 *) Add new certificate file to stack functions,
11034 SSL_add_dir_cert_subjects_to_stack() and
11035 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11036 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11037 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11038 This means that Apache-SSL and similar packages don't have to mess around
11039 to add as many CAs as they want to the preferred list.
11042 *) Experiment with doxygen documentation. Currently only partially applied to
11044 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11045 openssl.doxy as the configuration file.
11048 *) Get rid of remaining C++-style comments which strict C compilers hate.
11049 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11051 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11052 compiled in by default: it has problems with large keys.
11055 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11056 DH private keys and/or callback functions which directly correspond to
11057 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11058 is needed for applications which have to configure certificates on a
11059 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11061 For the RSA certificate situation is makes no difference, but
11062 for the DSA certificate situation this fixes the "no shared cipher"
11063 problem where the OpenSSL cipher selection procedure failed because the
11064 temporary keys were not overtaken from the context and the API provided
11065 no way to reconfigure them.
11066 The new functions now let applications reconfigure the stuff and they
11067 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11068 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11069 non-public-API function ssl_cert_instantiate() is used as a helper
11070 function and also to reduce code redundancy inside ssl_rsa.c.
11071 [Ralf S. Engelschall]
11073 *) Move s_server -dcert and -dkey options out of the undocumented feature
11074 area because they are useful for the DSA situation and should be
11075 recognized by the users.
11076 [Ralf S. Engelschall]
11078 *) Fix the cipher decision scheme for export ciphers: the export bits are
11079 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11080 SSL_EXP_MASK. So, the original variable has to be used instead of the
11081 already masked variable.
11082 [Richard Levitte <levitte@stacken.kth.se>]
11084 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11085 [Richard Levitte <levitte@stacken.kth.se>]
11087 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11088 from `int' to `unsigned int' because it's a length and initialized by
11089 EVP_DigestFinal() which expects an `unsigned int *'.
11090 [Richard Levitte <levitte@stacken.kth.se>]
11092 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11093 script. Instead use the usual Shell->Perl transition trick.
11094 [Ralf S. Engelschall]
11096 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11097 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11098 -noout -modulus' as it's already the case for `openssl rsa -noout
11099 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11100 currently the public key is printed (a decision which was already done by
11101 `openssl dsa -modulus' in the past) which serves a similar purpose.
11102 Additionally the NO_RSA no longer completely removes the whole -modulus
11103 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11105 [Ralf S. Engelschall]
11107 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11108 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11109 [Arne Ansper <arne@ats.cyber.ee>]
11111 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11112 to be added. Now both 'req' and 'ca' can use new objects defined in the
11116 *) Add cool BIO that does syslog (or event log on NT).
11117 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11119 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11120 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11121 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11122 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11125 *) Add preliminary config info for new extension code.
11128 *) Make RSA_NO_PADDING really use no padding.
11129 [Ulf Moeller <ulf@fitug.de>]
11131 *) Generate errors when private/public key check is done.
11134 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11135 for some CRL extensions and new objects added.
11138 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11139 key usage extension and fuller support for authority key id.
11142 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11143 padding method for RSA, which is recommended for new applications in PKCS
11144 #1 v2.0 (RFC 2437, October 1998).
11145 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11146 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11147 against Bleichbacher's attack on RSA.
11148 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11151 *) Updates to the new SSL compression code
11152 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11154 *) Fix so that the version number in the master secret, when passed
11155 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11156 (because the server will not accept higher), that the version number
11157 is 0x03,0x01, not 0x03,0x00
11158 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11160 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11161 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11162 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11165 *) Support for RAW extensions where an arbitrary extension can be
11166 created by including its DER encoding. See apps/openssl.cnf for
11170 *) Make sure latest Perl versions don't interpret some generated C array
11171 code as Perl array code in the crypto/err/err_genc.pl script.
11172 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11174 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11175 not many people have the assembler. Various Win32 compilation fixes and
11176 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11177 build instructions.
11180 *) Modify configure script 'Configure' to automatically create crypto/date.h
11181 file under Win32 and also build pem.h from pem.org. New script
11182 util/mkfiles.pl to create the MINFO file on environments that can't do a
11183 'make files': perl util/mkfiles.pl >MINFO should work.
11186 *) Major rework of DES function declarations, in the pursuit of correctness
11187 and purity. As a result, many evil casts evaporated, and some weirdness,
11188 too. You may find this causes warnings in your code. Zapping your evil
11189 casts will probably fix them. Mostly.
11192 *) Fix for a typo in asn1.h. Bug fix to object creation script
11193 obj_dat.pl. It considered a zero in an object definition to mean
11194 "end of object": none of the objects in objects.h have any zeros
11195 so it wasn't spotted.
11196 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11198 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11199 Masking (CBCM). In the absence of test vectors, the best I have been able
11200 to do is check that the decrypt undoes the encrypt, so far. Send me test
11201 vectors if you have them.
11204 *) Correct calculation of key length for export ciphers (too much space was
11205 allocated for null ciphers). This has not been tested!
11208 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11209 message is now correct (it understands "crypto" and "ssl" on its
11210 command line). There is also now an "update" option. This will update
11211 the util/ssleay.num and util/libeay.num files with any new functions.
11213 perl util/mkdef.pl crypto ssl update
11214 it will update them.
11217 *) Overhauled the Perl interface (perl/*):
11218 - ported BN stuff to OpenSSL's different BN library
11219 - made the perl/ source tree CVS-aware
11220 - renamed the package from SSLeay to OpenSSL (the files still contain
11221 their history because I've copied them in the repository)
11222 - removed obsolete files (the test scripts will be replaced
11223 by better Test::Harness variants in the future)
11224 [Ralf S. Engelschall]
11226 *) First cut for a very conservative source tree cleanup:
11227 1. merge various obsolete readme texts into doc/ssleay.txt
11228 where we collect the old documents and readme texts.
11229 2. remove the first part of files where I'm already sure that we no
11230 longer need them because of three reasons: either they are just temporary
11231 files which were left by Eric or they are preserved original files where
11232 I've verified that the diff is also available in the CVS via "cvs diff
11233 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11234 the crypto/md/ stuff).
11235 [Ralf S. Engelschall]
11237 *) More extension code. Incomplete support for subject and issuer alt
11238 name, issuer and authority key id. Change the i2v function parameters
11239 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11240 what that's for :-) Fix to ASN1 macro which messed up
11241 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11244 *) Preliminary support for ENUMERATED type. This is largely copied from the
11248 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11249 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11251 *) Make sure `make rehash' target really finds the `openssl' program.
11252 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11254 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11255 like to hear about it if this slows down other processors.
11258 *) Add CygWin32 platform information to Configure script.
11259 [Alan Batie <batie@aahz.jf.intel.com>]
11261 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11262 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11264 *) New program nseq to manipulate netscape certificate sequences
11267 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11271 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11272 but the BN code had some problems that would cause failures when
11273 doing certificate verification and some other functions.
11274 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11276 *) Add ASN1 and PEM code to support netscape certificate sequences.
11279 *) Add ASN1 and PEM code to support netscape certificate sequences.
11282 *) Add several PKIX and private extended key usage OIDs.
11285 *) Modify the 'ca' program to handle the new extension code. Modify
11286 openssl.cnf for new extension format, add comments.
11289 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11290 and add a sample to openssl.cnf so req -x509 now adds appropriate
11294 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11295 error code, add initial support to X509_print() and x509 application.
11298 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11299 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11300 stuff is currently isolated and isn't even compiled yet.
11303 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11304 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11305 Removed the versions check from X509 routines when loading extensions:
11306 this allows certain broken certificates that don't set the version
11307 properly to be processed.
11310 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11311 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11312 can still be regenerated with "make depend".
11315 *) Spelling mistake in C version of CAST-128.
11316 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11318 *) Changes to the error generation code. The perl script err-code.pl
11319 now reads in the old error codes and retains the old numbers, only
11320 adding new ones if necessary. It also only changes the .err files if new
11321 codes are added. The makefiles have been modified to only insert errors
11322 when needed (to avoid needlessly modifying header files). This is done
11323 by only inserting errors if the .err file is newer than the auto generated
11324 C file. To rebuild all the error codes from scratch (the old behaviour)
11325 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11326 or delete all the .err files.
11329 *) CAST-128 was incorrectly implemented for short keys. The C version has
11330 been fixed, but is untested. The assembler versions are also fixed, but
11331 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11332 to regenerate it if needed.
11333 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11334 Hagino <itojun@kame.net>]
11336 *) File was opened incorrectly in randfile.c.
11337 [Ulf Möller <ulf@fitug.de>]
11339 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11340 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11341 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11342 al: it's just almost always a UTCTime. Note this patch adds new error
11343 codes so do a "make errors" if there are problems.
11346 *) Correct Linux 1 recognition in config.
11347 [Ulf Möller <ulf@fitug.de>]
11349 *) Remove pointless MD5 hash when using DSA keys in ca.
11350 [Anonymous <nobody@replay.com>]
11352 *) Generate an error if given an empty string as a cert directory. Also
11353 generate an error if handed NULL (previously returned 0 to indicate an
11354 error, but didn't set one).
11355 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11357 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11360 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11361 parameters. This was causing a warning which killed off the Win32 compile.
11364 *) Remove C++ style comments from crypto/bn/bn_local.h.
11365 [Neil Costigan <neil.costigan@celocom.com>]
11367 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11368 based on a text string, looking up short and long names and finally
11369 "dot" format. The "dot" format stuff didn't work. Added new function
11370 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11371 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11372 OID is not part of the table.
11375 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11376 X509_LOOKUP_by_alias().
11379 *) Sort openssl functions by name.
11382 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11383 encryption from sample DSA keys (in case anyone is interested the password
11387 *) Make _all_ *_free functions accept a NULL pointer.
11388 [Frans Heymans <fheymans@isaserver.be>]
11390 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11392 [Anonymous <nobody@replay.com>]
11394 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11395 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11397 *) Don't blow it for numeric -newkey arguments to apps/req.
11398 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11400 *) Temp key "for export" tests were wrong in s3_srvr.c.
11401 [Anonymous <nobody@replay.com>]
11403 *) Add prototype for temp key callback functions
11404 SSL_CTX_set_tmp_{rsa,dh}_callback().
11407 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11408 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11411 *) X509_name_add_entry() freed the wrong thing after an error.
11412 [Arne Ansper <arne@ats.cyber.ee>]
11414 *) rsa_eay.c would attempt to free a NULL context.
11415 [Arne Ansper <arne@ats.cyber.ee>]
11417 *) BIO_s_socket() had a broken should_retry() on Windoze.
11418 [Arne Ansper <arne@ats.cyber.ee>]
11420 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11421 [Arne Ansper <arne@ats.cyber.ee>]
11423 *) Make sure the already existing X509_STORE->depth variable is initialized
11424 in X509_STORE_new(), but document the fact that this variable is still
11425 unused in the certificate verification process.
11426 [Ralf S. Engelschall]
11428 *) Fix the various library and apps files to free up pkeys obtained from
11429 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11432 *) Fix reference counting in X509_PUBKEY_get(). This makes
11433 demos/maurice/example2.c work, amongst others, probably.
11434 [Steve Henson and Ben Laurie]
11436 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11437 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11438 are no longer created. This way we have a single and consistent command
11439 line interface `openssl <command>', similar to `cvs <command>'.
11440 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11442 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11443 BIT STRING wrapper always have zero unused bits.
11446 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11449 *) Make the top-level INSTALL documentation easier to understand.
11452 *) Makefiles updated to exit if an error occurs in a sub-directory
11453 make (including if user presses ^C) [Paul Sutton]
11455 *) Make Montgomery context stuff explicit in RSA data structure.
11458 *) Fix build order of pem and err to allow for generated pem.h.
11461 *) Fix renumbering bug in X509_NAME_delete_entry().
11464 *) Enhanced the err-ins.pl script so it makes the error library number
11465 global and can add a library name. This is needed for external ASN1 and
11466 other error libraries.
11469 *) Fixed sk_insert which never worked properly.
11472 *) Fix ASN1 macros so they can handle indefinite length constructed
11473 EXPLICIT tags. Some non standard certificates use these: they can now
11477 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11478 into a single doc/ssleay.txt bundle. This way the information is still
11479 preserved but no longer messes up this directory. Now it's new room for
11480 the new set of documentation files.
11481 [Ralf S. Engelschall]
11483 *) SETs were incorrectly DER encoded. This was a major pain, because they
11484 shared code with SEQUENCEs, which aren't coded the same. This means that
11485 almost everything to do with SETs or SEQUENCEs has either changed name or
11486 number of arguments.
11487 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11489 *) Fix test data to work with the above.
11492 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11493 was already fixed by Eric for 0.9.1 it seems.
11494 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11496 *) Autodetect FreeBSD3.
11499 *) Fix various bugs in Configure. This affects the following platforms:
11503 unixware-2.0-pentium
11507 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11508 before they are needed.
11511 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11515 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11517 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11518 changed SSLeay to OpenSSL in version strings.
11519 [Ralf S. Engelschall]
11521 *) Some fixups to the top-level documents.
11524 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11525 because the symlink to include/ was missing.
11526 [Ralf S. Engelschall]
11528 *) Incorporated the popular no-RSA/DSA-only patches
11529 which allow to compile a RSA-free SSLeay.
11530 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11532 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11533 when "ssleay" is still not found.
11534 [Ralf S. Engelschall]
11536 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11537 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11539 *) Updated the README file.
11540 [Ralf S. Engelschall]
11542 *) Added various .cvsignore files in the CVS repository subdirs
11543 to make a "cvs update" really silent.
11544 [Ralf S. Engelschall]
11546 *) Recompiled the error-definition header files and added
11547 missing symbols to the Win32 linker tables.
11548 [Ralf S. Engelschall]
11550 *) Cleaned up the top-level documents;
11551 o new files: CHANGES and LICENSE
11552 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11553 o merged COPYRIGHT into LICENSE
11554 o removed obsolete TODO file
11555 o renamed MICROSOFT to INSTALL.W32
11556 [Ralf S. Engelschall]
11558 *) Removed dummy files from the 0.9.1b source tree:
11559 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11560 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11561 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11562 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11563 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11564 [Ralf S. Engelschall]
11566 *) Added various platform portability fixes.
11569 *) The Genesis of the OpenSSL rpject:
11570 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11571 Young and Tim J. Hudson created while they were working for C2Net until
11573 [The OpenSSL Project]
11576 Changes between 0.9.0b and 0.9.1b [not released]
11578 *) Updated a few CA certificates under certs/
11581 *) Changed some BIGNUM api stuff.
11584 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11585 DGUX x86, Linux Alpha, etc.
11588 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11589 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11593 *) Add -strparse option to asn1pars program which parses nested
11595 [Dr Stephen Henson <shenson@bigfoot.com>]
11597 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11600 *) DSA fix for "ca" program.
11603 *) Added "-genkey" option to "dsaparam" program.
11606 *) Added RIPE MD160 (rmd160) message digest.
11609 *) Added -a (all) option to "ssleay version" command.
11612 *) Added PLATFORM define which is the id given to Configure.
11615 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11618 *) Extended the ASN.1 parser routines.
11621 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11624 *) Added a BN_CTX to the BN library.
11627 *) Fixed the weak key values in DES library
11630 *) Changed API in EVP library for cipher aliases.
11633 *) Added support for RC2/64bit cipher.
11636 *) Converted the lhash library to the crypto/mem.c functions.
11639 *) Added more recognized ASN.1 object ids.
11642 *) Added more RSA padding checks for SSL/TLS.
11645 *) Added BIO proxy/filter functionality.
11648 *) Added extra_certs to SSL_CTX which can be used
11649 send extra CA certificates to the client in the CA cert chain sending
11650 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11653 *) Now Fortezza is denied in the authentication phase because
11654 this is key exchange mechanism is not supported by SSLeay at all.
11657 *) Additional PKCS1 checks.
11660 *) Support the string "TLSv1" for all TLS v1 ciphers.
11663 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11664 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11667 *) Fixed a few memory leaks.
11670 *) Fixed various code and comment typos.
11673 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11674 bytes sent in the client random.
11675 [Edward Bishop <ebishop@spyglass.com>]