5 Changes between 1.0.2e and 1.1.0 [xx XXX xxxx]
7 *) The GOST engine was out of date and therefore it has been removed. An up
8 to date GOST engine is now being maintained in an external repository.
9 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
10 support for GOST ciphersuites (these are only activated if a GOST engine
14 *) EGD is no longer supported by default; use enable-egd when
16 [Ben Kaduk and Rich Salz]
18 *) The distribution now has Makefile.in files, which are used to
19 create Makefile's when Configure is run. *Configure must be run
20 before trying to build now.*
23 *) The return value for SSL_CIPHER_description() for error conditions
27 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
29 Obtaining and performing DNSSEC validation of TLSA records is
30 the application's responsibility. The application provides
31 the TLSA records of its choice to OpenSSL, and these are then
32 used to authenticate the peer.
34 The TLSA records need not even come from DNS. They can, for
35 example, be used to implement local end-entity certificate or
36 trust-anchor "pinning", where the "pin" data takes the form
37 of TLSA records, which can augment or replace verification
38 based on the usual WebPKI public certification authorities.
41 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
42 continues to support deprecated interfaces in default builds.
43 However, applications are strongly advised to compile their
44 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
45 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
46 or the 1.1.0 releases.
48 In environments in which all applications have been ported to
49 not use any deprecated interfaces OpenSSL's Configure script
50 should be used with the --api=1.1.0 option to entirely remove
51 support for the deprecated features from the library and
52 unconditionally disable them in the installed headers.
53 Essentially the same effect can be achieved with the "no-deprecated"
54 argument to Configure, except that this will always restrict
55 the build to just the latest API, rather than a fixed API
58 As applications are ported to future revisions of the API,
59 they should update their compile-time OPENSSL_API_COMPAT define
60 accordingly, but in most cases should be able to continue to
61 compile with later releases.
63 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
64 0x10000000L and 0x00908000L, respectively. However those
65 versions did not support the OPENSSL_API_COMPAT feature, and
66 so applications are not typically tested for explicit support
67 of just the undeprecated features of either release.
70 *) Add support for setting the minimum and maximum supported protocol.
71 It can bet set via the SSL_set_min_proto_version() and
72 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
73 MaxProtcol. It's recommended to use the new APIs to disable
74 protocols instead of disabling individual protocols using
75 SSL_set_options() or SSL_CONF's Protocol. This change also
76 removes support for disabling TLS 1.2 in the OpenSSL TLS
77 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
80 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
83 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
84 and integrates ECDSA and ECDH functionality into EC. Implementations can
85 now redirect key generation and no longer need to convert to or from
88 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
89 include the ec.h header file instead.
92 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
93 ciphers who are no longer supported and drops support the ephemeral RSA key
94 exchange. The LOW ciphers currently doesn't have any ciphers in it.
97 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
98 opaque. For HMAC_CTX, the following constructors and destructors
101 HMAC_CTX *HMAC_CTX_new(void);
102 void HMAC_CTX_free(HMAC_CTX *ctx);
104 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
105 destroy such methods has been added. See EVP_MD_meth_new(3) and
106 EVP_CIPHER_meth_new(3) for documentation.
109 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
110 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
111 EVP_MD_CTX_reset() should be called instead to reinitialise
112 an already created structure.
113 2) For consistency with the majority of our object creators and
114 destructors, EVP_MD_CTX_(create|destroy) were renamed to
115 EVP_MD_CTX_(new|free). The old names are retained as macros
116 for deprecated builds.
119 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
120 cryptographic operations to be performed asynchronously as long as an
121 asynchronous capable engine is used. See the ASYNC_start_job() man page for
122 further details. Libssl has also had this capability integrated with the
123 introduction of the new mode SSL_MODE_ASYNC and associated error
124 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
125 pages. This work was developed in partnership with Intel Corp.
128 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
129 always enabled now. If you want to disable the support you should
130 exclude it using the list of supported ciphers. This also means that the
131 "-no_ecdhe" option has been removed from s_server.
134 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
135 SSL_{CTX_}set1_curves() which can set a list.
138 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
139 curve you want to support using SSL_{CTX_}set1_curves().
142 *) State machine rewrite. The state machine code has been significantly
143 refactored in order to remove much duplication of code and solve issues
144 with the old code (see ssl/statem/README for further details). This change
145 does have some associated API changes. Notably the SSL_state() function
146 has been removed and replaced by SSL_get_state which now returns an
147 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
148 altogether. The previous handshake states defined in ssl.h and ssl3.h have
152 *) All instances of the string "ssleay" in the public API were replaced
153 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
154 Some error codes related to internal RSA_eay API's were renamed.
157 *) The demo files in crypto/threads were moved to demo/threads.
160 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
162 [Matt Caswell, Rich Salz]
164 *) New ASN.1 embed macro.
166 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
167 structure is not allocated: it is part of the parent. That is instead of
175 This reduces memory fragmentation and make it impossible to accidentally
176 set a mandatory field to NULL.
178 This currently only works for some fields specifically a SEQUENCE, CHOICE,
179 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
180 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
184 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
187 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
188 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
189 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
190 DES and RC4 ciphersuites.
193 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
194 This changes the decoding behaviour for some invalid messages,
195 though the change is mostly in the more lenient direction, and
196 legacy behaviour is preserved as much as possible.
199 *) Fix no-stdio build.
200 [ David Woodhouse <David.Woodhouse@intel.com> and also
201 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
203 *) New testing framework
204 The testing framework has been largely rewritten and is now using
205 perl and the perl modules Test::Harness and an extended variant of
206 Test::More called OpenSSL::Test to do its work. All test scripts in
207 test/ have been rewritten into test recipes, and all direct calls to
208 executables in test/Makefile have become individual recipes using the
209 simplified testing OpenSSL::Test::Simple.
211 For documentation on our testing modules, do:
213 perldoc test/testlib/OpenSSL/Test/Simple.pm
214 perldoc test/testlib/OpenSSL/Test.pm
218 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
219 are used; the latter aborts on memory leaks (usually checked on exit).
220 Some undocumented "set malloc, etc., hooks" functions were removed
221 and others were changed. All are now documented.
224 *) In DSA_generate_parameters_ex, if the provided seed is too short,
226 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
228 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
229 from RFC4279, RFC4785, RFC5487, RFC5489.
231 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
232 original RSA_PSK patch.
235 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
236 era flag was never set throughout the codebase (only read). Also removed
237 SSL3_FLAGS_POP_BUFFER which was only used if
238 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
241 *) Changed the default name options in the "ca", "crl", "req" and "x509"
242 to be "oneline" instead of "compat".
245 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
246 not aware of clients that still exhibit this bug, and the workaround
247 hasn't been working properly for a while.
250 *) The return type of BIO_number_read() and BIO_number_written() as well as
251 the corresponding num_read and num_write members in the BIO structure has
252 changed from unsigned long to uint64_t. On platforms where an unsigned
253 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
257 *) Given the pervasive nature of TLS extensions it is inadvisable to run
258 OpenSSL without support for them. It also means that maintaining
259 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
260 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
263 *) Removed support for the two export grade static DH ciphersuites
264 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
265 were newly added (along with a number of other static DH ciphersuites) to
266 1.0.2. However the two export ones have *never* worked since they were
267 introduced. It seems strange in any case to be adding new export
268 ciphersuites, and given "logjam" it also does not seem correct to fix them.
271 *) Version negotiation has been rewritten. In particular SSLv23_method(),
272 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
273 and turned into macros which simply call the new preferred function names
274 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
275 should use the new names instead. Also as part of this change the ssl23.h
276 header file has been removed.
279 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
280 code and the associated standard is no longer considered fit-for-purpose.
283 *) RT2547 was closed. When generating a private key, try to make the
284 output file readable only by the owner. This behavior change might
285 be noticeable when interacting with other software.
287 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
291 *) Added HTTP GET support to the ocsp command.
294 *) Changed default digest for the dgst and enc commands from MD5 to
298 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
301 *) Added support for TLS extended master secret from
302 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
303 initial patch which was a great help during development.
306 *) All libssl internal structures have been removed from the public header
307 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
308 now redundant). Users should not attempt to access internal structures
309 directly. Instead they should use the provided API functions.
312 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
313 Access to deprecated functions can be re-enabled by running config with
314 "enable-deprecated". In addition applications wishing to use deprecated
315 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
316 will, by default, disable some transitive includes that previously existed
317 in the header files (e.g. ec.h will no longer, by default, include bn.h)
320 *) Added support for OCB mode. OpenSSL has been granted a patent license
321 compatible with the OpenSSL license for use of OCB. Details are available
322 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
323 for OCB can be removed by calling config with no-ocb.
326 *) SSLv2 support has been removed. It still supports receiving a SSLv2
327 compatible client hello.
330 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
331 done while fixing the error code for the key-too-small case.
332 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
334 *) CA.sh has been removmed; use CA.pl instead.
337 *) Removed old DES API.
340 *) Remove various unsupported platforms:
346 Sinix/ReliantUNIX RM400
351 16-bit platforms such as WIN16
354 *) Clean up OPENSSL_NO_xxx #define's
355 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
356 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
357 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
358 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
359 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
360 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
361 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
362 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
363 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
364 Remove MS_STATIC; it's a relic from platforms <32 bits.
367 *) Cleaned up dead code
368 Remove all but one '#ifdef undef' which is to be looked at.
371 *) Clean up calling of xxx_free routines.
372 Just like free(), fix most of the xxx_free routines to accept
373 NULL. Remove the non-null checks from callers. Save much code.
376 *) Add secure heap for storage of private keys (when possible).
377 Add BIO_s_secmem(), CBIGNUM, etc.
378 Contributed by Akamai Technologies under our Corporate CLA.
381 *) Experimental support for a new, fast, unbiased prime candidate generator,
382 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
383 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
385 *) New output format NSS in the sess_id command line tool. This allows
386 exporting the session id and the master key in NSS keylog format.
387 [Martin Kaiser <martin@kaiser.cx>]
389 *) Harmonize version and its documentation. -f flag is used to display
391 [mancha <mancha1@zoho.com>]
393 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
394 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
395 [mancha <mancha1@zoho.com>]
397 *) Fix some double frees. These are not thought to be exploitable.
398 [mancha <mancha1@zoho.com>]
400 *) A missing bounds check in the handling of the TLS heartbeat extension
401 can be used to reveal up to 64k of memory to a connected client or
404 Thanks for Neel Mehta of Google Security for discovering this bug and to
405 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
406 preparing the fix (CVE-2014-0160)
407 [Adam Langley, Bodo Moeller]
409 *) Fix for the attack described in the paper "Recovering OpenSSL
410 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
411 by Yuval Yarom and Naomi Benger. Details can be obtained from:
412 http://eprint.iacr.org/2014/140
414 Thanks to Yuval Yarom and Naomi Benger for discovering this
415 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
416 [Yuval Yarom and Naomi Benger]
418 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
419 this fixes a limitation in previous versions of OpenSSL.
422 *) Experimental encrypt-then-mac support.
424 Experimental support for encrypt then mac from
425 draft-gutmann-tls-encrypt-then-mac-02.txt
427 To enable it set the appropriate extension number (0x42 for the test
428 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
430 For non-compliant peers (i.e. just about everything) this should have no
433 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
437 *) Add EVP support for key wrapping algorithms, to avoid problems with
438 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
439 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
440 algorithms and include tests cases.
443 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
447 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
448 MGF1 digest and OAEP label.
451 *) Make openssl verify return errors.
452 [Chris Palmer <palmer@google.com> and Ben Laurie]
454 *) New function ASN1_TIME_diff to calculate the difference between two
455 ASN1_TIME structures or one structure and the current time.
458 *) Update fips_test_suite to support multiple command line options. New
459 test to induce all self test errors in sequence and check expected
463 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
464 sign or verify all in one operation.
467 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
468 test programs and fips_test_suite. Includes functionality to parse
469 the minimal script output of fipsalgest.pl directly.
472 *) Add authorisation parameter to FIPS_module_mode_set().
475 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
478 *) Use separate DRBG fields for internal and external flags. New function
479 FIPS_drbg_health_check() to perform on demand health checking. Add
480 generation tests to fips_test_suite with reduced health check interval to
481 demonstrate periodic health checking. Add "nodh" option to
482 fips_test_suite to skip very slow DH test.
485 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
489 *) More extensive health check for DRBG checking many more failure modes.
490 New function FIPS_selftest_drbg_all() to handle every possible DRBG
491 combination: call this in fips_test_suite.
494 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
495 and POST to handle Dual EC cases.
498 *) Add support for canonical generation of DSA parameter 'g'. See
501 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
502 POST to handle HMAC cases.
505 *) Add functions FIPS_module_version() and FIPS_module_version_text()
506 to return numerical and string versions of the FIPS module number.
509 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
510 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
511 outside the validated module in the FIPS capable OpenSSL.
514 *) Minor change to DRBG entropy callback semantics. In some cases
515 there is no multiple of the block length between min_len and
516 max_len. Allow the callback to return more than max_len bytes
517 of entropy but discard any extra: it is the callback's responsibility
518 to ensure that the extra data discarded does not impact the
519 requested amount of entropy.
522 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
523 information in FIPS186-3, SP800-57 and SP800-131A.
526 *) CCM support via EVP. Interface is very similar to GCM case except we
527 must supply all data in one chunk (i.e. no update, final) and the
528 message length must be supplied if AAD is used. Add algorithm test
532 *) Initial version of POST overhaul. Add POST callback to allow the status
533 of POST to be monitored and/or failures induced. Modify fips_test_suite
534 to use callback. Always run all selftests even if one fails.
537 *) XTS support including algorithm test driver in the fips_gcmtest program.
538 Note: this does increase the maximum key length from 32 to 64 bytes but
539 there should be no binary compatibility issues as existing applications
540 will never use XTS mode.
543 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
544 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
545 performs algorithm blocking for unapproved PRNG types. Also do not
546 set PRNG type in FIPS_mode_set(): leave this to the application.
547 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
548 the standard OpenSSL PRNG: set additional data to a date time vector.
551 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
552 This shouldn't present any incompatibility problems because applications
553 shouldn't be using these directly and any that are will need to rethink
554 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
557 *) Extensive self tests and health checking required by SP800-90 DRBG.
558 Remove strength parameter from FIPS_drbg_instantiate and always
559 instantiate at maximum supported strength.
562 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
565 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
568 *) New function DH_compute_key_padded() to compute a DH key and pad with
569 leading zeroes if needed: this complies with SP800-56A et al.
572 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
573 anything, incomplete, subject to change and largely untested at present.
576 *) Modify fipscanisteronly build option to only build the necessary object
577 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
580 *) Add experimental option FIPSSYMS to give all symbols in
581 fipscanister.o and FIPS or fips prefix. This will avoid
582 conflicts with future versions of OpenSSL. Add perl script
583 util/fipsas.pl to preprocess assembly language source files
584 and rename any affected symbols.
587 *) Add selftest checks and algorithm block of non-fips algorithms in
588 FIPS mode. Remove DES2 from selftests.
591 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
592 return internal method without any ENGINE dependencies. Add new
593 tiny fips sign and verify functions.
596 *) New build option no-ec2m to disable characteristic 2 code.
599 *) New build option "fipscanisteronly". This only builds fipscanister.o
600 and (currently) associated fips utilities. Uses the file Makefile.fips
601 instead of Makefile.org as the prototype.
604 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
605 Update fips_gcmtest to use IV generator.
608 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
609 setting output buffer to NULL. The *Final function must be
610 called although it will not retrieve any additional data. The tag
611 can be set or retrieved with a ctrl. The IV length is by default 12
612 bytes (96 bits) but can be set to an alternative value. If the IV
613 length exceeds the maximum IV length (currently 16 bytes) it cannot be
617 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
618 underlying do_cipher function handles all cipher semantics itself
619 including padding and finalisation. This is useful if (for example)
620 an ENGINE cipher handles block padding itself. The behaviour of
621 do_cipher is subtly changed if this flag is set: the return value
622 is the number of characters written to the output buffer (zero is
623 no longer an error code) or a negative error code. Also if the
624 input buffer is NULL and length 0 finalisation should be performed.
627 *) If a candidate issuer certificate is already part of the constructed
628 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
631 *) Improve forward-security support: add functions
633 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
634 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
636 for use by SSL/TLS servers; the callback function will be called whenever a
637 new session is created, and gets to decide whether the session may be
638 cached to make it resumable (return 0) or not (return 1). (As by the
639 SSL/TLS protocol specifications, the session_id sent by the server will be
640 empty to indicate that the session is not resumable; also, the server will
641 not generate RFC 4507 (RFC 5077) session tickets.)
643 A simple reasonable callback implementation is to return is_forward_secure.
644 This parameter will be set to 1 or 0 depending on the ciphersuite selected
645 by the SSL/TLS server library, indicating whether it can provide forward
647 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
649 *) New -verify_name option in command line utilities to set verification
653 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
654 Add CMAC pkey methods.
657 *) Experimental renegotiation in s_server -www mode. If the client
658 browses /reneg connection is renegotiated. If /renegcert it is
659 renegotiated requesting a certificate.
662 *) Add an "external" session cache for debugging purposes to s_server. This
663 should help trace issues which normally are only apparent in deployed
664 multi-process servers.
667 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
668 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
669 BIO_set_cipher() and some obscure PEM functions were changed so they
670 can now return an error. The RAND changes required a change to the
671 RAND_METHOD structure.
674 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
675 a gcc attribute to warn if the result of a function is ignored. This
676 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
677 whose return value is often ignored.
680 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
682 *) BN_mod_exp may produce incorrect results on x86_64
684 There is a carry propagating bug in the x86_64 Montgomery squaring
685 procedure. No EC algorithms are affected. Analysis suggests that attacks
686 against RSA and DSA as a result of this defect would be very difficult to
687 perform and are not believed likely. Attacks against DH are considered just
688 feasible (although very difficult) because most of the work necessary to
689 deduce information about a private key may be performed offline. The amount
690 of resources required for such an attack would be very significant and
691 likely only accessible to a limited number of attackers. An attacker would
692 additionally need online access to an unpatched system using the target
693 private key in a scenario with persistent DH parameters and a private
694 key that is shared between multiple clients. For example this can occur by
695 default in OpenSSL DHE based SSL/TLS ciphersuites.
697 This issue was reported to OpenSSL by Hanno Böck.
701 *) Certificate verify crash with missing PSS parameter
703 The signature verification routines will crash with a NULL pointer
704 dereference if presented with an ASN.1 signature using the RSA PSS
705 algorithm and absent mask generation function parameter. Since these
706 routines are used to verify certificate signature algorithms this can be
707 used to crash any certificate verification operation and exploited in a
708 DoS attack. Any application which performs certificate verification is
709 vulnerable including OpenSSL clients and servers which enable client
712 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
716 *) X509_ATTRIBUTE memory leak
718 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
719 memory. This structure is used by the PKCS#7 and CMS routines so any
720 application which reads PKCS#7 or CMS data from untrusted sources is
721 affected. SSL/TLS is not affected.
723 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
728 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
729 This changes the decoding behaviour for some invalid messages,
730 though the change is mostly in the more lenient direction, and
731 legacy behaviour is preserved as much as possible.
734 *) In DSA_generate_parameters_ex, if the provided seed is too short,
736 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
738 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
740 *) Alternate chains certificate forgery
742 During certificate verfification, OpenSSL will attempt to find an
743 alternative certificate chain if the first attempt to build such a chain
744 fails. An error in the implementation of this logic can mean that an
745 attacker could cause certain checks on untrusted certificates to be
746 bypassed, such as the CA flag, enabling them to use a valid leaf
747 certificate to act as a CA and "issue" an invalid certificate.
749 This issue was reported to OpenSSL by Adam Langley/David Benjamin
753 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
755 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
756 incompatibility in the handling of HMAC. The previous ABI has now been
760 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
762 *) Malformed ECParameters causes infinite loop
764 When processing an ECParameters structure OpenSSL enters an infinite loop
765 if the curve specified is over a specially malformed binary polynomial
768 This can be used to perform denial of service against any
769 system which processes public keys, certificate requests or
770 certificates. This includes TLS clients and TLS servers with
771 client authentication enabled.
773 This issue was reported to OpenSSL by Joseph Barr-Pixton.
777 *) Exploitable out-of-bounds read in X509_cmp_time
779 X509_cmp_time does not properly check the length of the ASN1_TIME
780 string and can read a few bytes out of bounds. In addition,
781 X509_cmp_time accepts an arbitrary number of fractional seconds in the
784 An attacker can use this to craft malformed certificates and CRLs of
785 various sizes and potentially cause a segmentation fault, resulting in
786 a DoS on applications that verify certificates or CRLs. TLS clients
787 that verify CRLs are affected. TLS clients and servers with client
788 authentication enabled may be affected if they use custom verification
791 This issue was reported to OpenSSL by Robert Swiecki (Google), and
792 independently by Hanno Böck.
796 *) PKCS7 crash with missing EnvelopedContent
798 The PKCS#7 parsing code does not handle missing inner EncryptedContent
799 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
800 with missing content and trigger a NULL pointer dereference on parsing.
802 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
803 structures from untrusted sources are affected. OpenSSL clients and
804 servers are not affected.
806 This issue was reported to OpenSSL by Michal Zalewski (Google).
810 *) CMS verify infinite loop with unknown hash function
812 When verifying a signedData message the CMS code can enter an infinite loop
813 if presented with an unknown hash function OID. This can be used to perform
814 denial of service against any system which verifies signedData messages using
816 This issue was reported to OpenSSL by Johannes Bauer.
820 *) Race condition handling NewSessionTicket
822 If a NewSessionTicket is received by a multi-threaded client when attempting to
823 reuse a previous ticket then a race condition can occur potentially leading to
824 a double free of the ticket data.
828 *) Only support 256-bit or stronger elliptic curves with the
829 'ecdh_auto' setting (server) or by default (client). Of supported
830 curves, prefer P-256 (both).
833 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
835 *) ClientHello sigalgs DoS fix
837 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
838 invalid signature algorithms extension a NULL pointer dereference will
839 occur. This can be exploited in a DoS attack against the server.
841 This issue was was reported to OpenSSL by David Ramos of Stanford
844 [Stephen Henson and Matt Caswell]
846 *) Multiblock corrupted pointer fix
848 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
849 feature only applies on 64 bit x86 architecture platforms that support AES
850 NI instructions. A defect in the implementation of "multiblock" can cause
851 OpenSSL's internal write buffer to become incorrectly set to NULL when
852 using non-blocking IO. Typically, when the user application is using a
853 socket BIO for writing, this will only result in a failed connection.
854 However if some other BIO is used then it is likely that a segmentation
855 fault will be triggered, thus enabling a potential DoS attack.
857 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
861 *) Segmentation fault in DTLSv1_listen fix
863 The DTLSv1_listen function is intended to be stateless and processes the
864 initial ClientHello from many peers. It is common for user code to loop
865 over the call to DTLSv1_listen until a valid ClientHello is received with
866 an associated cookie. A defect in the implementation of DTLSv1_listen means
867 that state is preserved in the SSL object from one invocation to the next
868 that can lead to a segmentation fault. Errors processing the initial
869 ClientHello can trigger this scenario. An example of such an error could be
870 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
873 This issue was reported to OpenSSL by Per Allansson.
877 *) Segmentation fault in ASN1_TYPE_cmp fix
879 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
880 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
881 certificate signature algorithm consistency this can be used to crash any
882 certificate verification operation and exploited in a DoS attack. Any
883 application which performs certificate verification is vulnerable including
884 OpenSSL clients and servers which enable client authentication.
888 *) Segmentation fault for invalid PSS parameters fix
890 The signature verification routines will crash with a NULL pointer
891 dereference if presented with an ASN.1 signature using the RSA PSS
892 algorithm and invalid parameters. Since these routines are used to verify
893 certificate signature algorithms this can be used to crash any
894 certificate verification operation and exploited in a DoS attack. Any
895 application which performs certificate verification is vulnerable including
896 OpenSSL clients and servers which enable client authentication.
898 This issue was was reported to OpenSSL by Brian Carpenter.
902 *) ASN.1 structure reuse memory corruption fix
904 Reusing a structure in ASN.1 parsing may allow an attacker to cause
905 memory corruption via an invalid write. Such reuse is and has been
906 strongly discouraged and is believed to be rare.
908 Applications that parse structures containing CHOICE or ANY DEFINED BY
909 components may be affected. Certificate parsing (d2i_X509 and related
910 functions) are however not affected. OpenSSL clients and servers are
915 *) PKCS7 NULL pointer dereferences fix
917 The PKCS#7 parsing code does not handle missing outer ContentInfo
918 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
919 missing content and trigger a NULL pointer dereference on parsing.
921 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
922 otherwise parse PKCS#7 structures from untrusted sources are
923 affected. OpenSSL clients and servers are not affected.
925 This issue was reported to OpenSSL by Michal Zalewski (Google).
929 *) DoS via reachable assert in SSLv2 servers fix
931 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
932 servers that both support SSLv2 and enable export cipher suites by sending
933 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
935 This issue was discovered by Sean Burford (Google) and Emilia Käsper
936 (OpenSSL development team).
940 *) Empty CKE with client auth and DHE fix
942 If client auth is used then a server can seg fault in the event of a DHE
943 ciphersuite being selected and a zero length ClientKeyExchange message
944 being sent by the client. This could be exploited in a DoS attack.
948 *) Handshake with unseeded PRNG fix
950 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
951 with an unseeded PRNG. The conditions are:
952 - The client is on a platform where the PRNG has not been seeded
953 automatically, and the user has not seeded manually
954 - A protocol specific client method version has been used (i.e. not
955 SSL_client_methodv23)
956 - A ciphersuite is used that does not require additional random data from
957 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
959 If the handshake succeeds then the client random that has been used will
960 have been generated from a PRNG with insufficient entropy and therefore the
961 output may be predictable.
963 For example using the following command with an unseeded openssl will
964 succeed on an unpatched platform:
966 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
970 *) Use After Free following d2i_ECPrivatekey error fix
972 A malformed EC private key file consumed via the d2i_ECPrivateKey function
973 could cause a use after free condition. This, in turn, could cause a double
974 free in several private key parsing functions (such as d2i_PrivateKey
975 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
976 for applications that receive EC private keys from untrusted
977 sources. This scenario is considered rare.
979 This issue was discovered by the BoringSSL project and fixed in their
984 *) X509_to_X509_REQ NULL pointer deref fix
986 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
987 the certificate key is invalid. This function is rarely used in practice.
989 This issue was discovered by Brian Carpenter.
993 *) Removed the export ciphers from the DEFAULT ciphers
996 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
998 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
999 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1000 So far those who have to target multiple plaforms would compromise
1001 and argue that binary targeting say ARMv5 would still execute on
1002 ARMv8. "Universal" build resolves this compromise by providing
1003 near-optimal performance even on newer platforms.
1006 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1007 (other platforms pending).
1008 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1010 *) Add support for the SignedCertificateTimestampList certificate and
1011 OCSP response extensions from RFC6962.
1014 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1015 for corner cases. (Certain input points at infinity could lead to
1016 bogus results, with non-infinity inputs mapped to infinity too.)
1019 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1020 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1021 common cases are optimized and there still is room for further
1022 improvements. Vector Permutation AES for Altivec is also added.
1025 *) Add support for little-endian ppc64 Linux target.
1026 [Marcelo Cerri (IBM)]
1028 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1029 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1030 are optimized and there still is room for further improvements.
1031 Both 32- and 64-bit modes are supported.
1032 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1034 *) Improved ARMv7 NEON support.
1037 *) Support for SPARC Architecture 2011 crypto extensions, first
1038 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1039 SHA256/512, MD5, GHASH and modular exponentiation.
1040 [Andy Polyakov, David Miller]
1042 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1044 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1046 *) Support for new and upcoming Intel processors, including AVX2,
1047 BMI and SHA ISA extensions. This includes additional "stitched"
1048 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1051 This work was sponsored by Intel Corp.
1054 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1055 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1056 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1059 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1060 this fixes a limiation in previous versions of OpenSSL.
1063 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1064 MGF1 digest and OAEP label.
1067 *) Add EVP support for key wrapping algorithms, to avoid problems with
1068 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1069 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1070 algorithms and include tests cases.
1073 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1075 [Douglas E. Engert, Steve Henson]
1077 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1078 difference in days and seconds between two tm or ASN1_TIME structures.
1081 *) Add -rev test option to s_server to just reverse order of characters
1082 received by client and send back to server. Also prints an abbreviated
1083 summary of the connection parameters.
1086 *) New option -brief for s_client and s_server to print out a brief summary
1087 of connection parameters.
1090 *) Add callbacks for arbitrary TLS extensions.
1091 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1093 *) New option -crl_download in several openssl utilities to download CRLs
1094 from CRLDP extension in certificates.
1097 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1100 *) New function X509_CRL_diff to generate a delta CRL from the difference
1101 of two full CRLs. Add support to "crl" utility.
1104 *) New functions to set lookup_crls function and to retrieve
1105 X509_STORE from X509_STORE_CTX.
1108 *) Print out deprecated issuer and subject unique ID fields in
1112 *) Extend OCSP I/O functions so they can be used for simple general purpose
1113 HTTP as well as OCSP. New wrapper function which can be used to download
1114 CRLs using the OCSP API.
1117 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1120 *) SSL_CONF* functions. These provide a common framework for application
1121 configuration using configuration files or command lines.
1124 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1125 message callback and prints the results. Needs compile time option
1126 "enable-ssl-trace". New options to s_client and s_server to enable
1130 *) New ctrl and macro to retrieve supported points extensions.
1131 Print out extension in s_server and s_client.
1134 *) New functions to retrieve certificate signature and signature
1138 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1142 *) New Suite B modes for TLS code. These use and enforce the requirements
1143 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1144 only use Suite B curves. The Suite B modes can be set by using the
1145 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1148 *) New chain verification flags for Suite B levels of security. Check
1149 algorithms are acceptable when flags are set in X509_verify_cert.
1152 *) Make tls1_check_chain return a set of flags indicating checks passed
1153 by a certificate chain. Add additional tests to handle client
1154 certificates: checks for matching certificate type and issuer name
1158 *) If an attempt is made to use a signature algorithm not in the peer
1159 preference list abort the handshake. If client has no suitable
1160 signature algorithms in response to a certificate request do not
1161 use the certificate.
1164 *) If server EC tmp key is not in client preference list abort handshake.
1167 *) Add support for certificate stores in CERT structure. This makes it
1168 possible to have different stores per SSL structure or one store in
1169 the parent SSL_CTX. Include distint stores for certificate chain
1170 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1171 to build and store a certificate chain in CERT structure: returing
1172 an error if the chain cannot be built: this will allow applications
1173 to test if a chain is correctly configured.
1175 Note: if the CERT based stores are not set then the parent SSL_CTX
1176 store is used to retain compatibility with existing behaviour.
1180 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1181 mask based on the current session, check mask when sending client
1182 hello and checking the requested ciphersuite.
1185 *) New ctrls to retrieve and set certificate types in a certificate
1186 request message. Print out received values in s_client. If certificate
1187 types is not set with custom values set sensible values based on
1188 supported signature algorithms.
1191 *) Support for distinct client and server supported signature algorithms.
1194 *) Add certificate callback. If set this is called whenever a certificate
1195 is required by client or server. An application can decide which
1196 certificate chain to present based on arbitrary criteria: for example
1197 supported signature algorithms. Add very simple example to s_server.
1198 This fixes many of the problems and restrictions of the existing client
1199 certificate callback: for example you can now clear an existing
1200 certificate and specify the whole chain.
1203 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1204 the certificate can be used for (if anything). Set valid_flags field
1205 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1206 to have similar checks in it.
1208 Add new "cert_flags" field to CERT structure and include a "strict mode".
1209 This enforces some TLS certificate requirements (such as only permitting
1210 certificate signature algorithms contained in the supported algorithms
1211 extension) which some implementations ignore: this option should be used
1212 with caution as it could cause interoperability issues.
1215 *) Update and tidy signature algorithm extension processing. Work out
1216 shared signature algorithms based on preferences and peer algorithms
1217 and print them out in s_client and s_server. Abort handshake if no
1218 shared signature algorithms.
1221 *) Add new functions to allow customised supported signature algorithms
1222 for SSL and SSL_CTX structures. Add options to s_client and s_server
1226 *) New function SSL_certs_clear() to delete all references to certificates
1227 from an SSL structure. Before this once a certificate had been added
1228 it couldn't be removed.
1231 *) Integrate hostname, email address and IP address checking with certificate
1232 verification. New verify options supporting checking in opensl utility.
1235 *) Fixes and wildcard matching support to hostname and email checking
1236 functions. Add manual page.
1237 [Florian Weimer (Red Hat Product Security Team)]
1239 *) New functions to check a hostname email or IP address against a
1240 certificate. Add options x509 utility to print results of checks against
1244 *) Fix OCSP checking.
1245 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1247 *) Initial experimental support for explicitly trusted non-root CAs.
1248 OpenSSL still tries to build a complete chain to a root but if an
1249 intermediate CA has a trust setting included that is used. The first
1250 setting is used: whether to trust (e.g., -addtrust option to the x509
1254 *) Add -trusted_first option which attempts to find certificates in the
1255 trusted store even if an untrusted chain is also supplied.
1258 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1259 platform support for Linux and Android.
1262 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1265 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1266 When in FIPS mode the approved implementations are used as normal,
1267 when not in FIPS mode the internal unapproved versions are used instead.
1268 This means that the FIPS capable OpenSSL isn't forced to use the
1269 (often lower performance) FIPS implementations outside FIPS mode.
1272 *) Transparently support X9.42 DH parameters when calling
1273 PEM_read_bio_DHparameters. This means existing applications can handle
1274 the new parameter format automatically.
1277 *) Initial experimental support for X9.42 DH parameter format: mainly
1278 to support use of 'q' parameter for RFC5114 parameters.
1281 *) Add DH parameters from RFC5114 including test data to dhtest.
1284 *) Support for automatic EC temporary key parameter selection. If enabled
1285 the most preferred EC parameters are automatically used instead of
1286 hardcoded fixed parameters. Now a server just has to call:
1287 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1288 support ECDH and use the most appropriate parameters.
1291 *) Enhance and tidy EC curve and point format TLS extension code. Use
1292 static structures instead of allocation if default values are used.
1293 New ctrls to set curves we wish to support and to retrieve shared curves.
1294 Print out shared curves in s_server. New options to s_server and s_client
1295 to set list of supported curves.
1298 *) New ctrls to retrieve supported signature algorithms and
1299 supported curve values as an array of NIDs. Extend openssl utility
1300 to print out received values.
1303 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1304 between NIDs and the more common NIST names such as "P-256". Enhance
1305 ecparam utility and ECC method to recognise the NIST names for curves.
1308 *) Enhance SSL/TLS certificate chain handling to support different
1309 chains for each certificate instead of one chain in the parent SSL_CTX.
1312 *) Support for fixed DH ciphersuite client authentication: where both
1313 server and client use DH certificates with common parameters.
1316 *) Support for fixed DH ciphersuites: those requiring DH server
1320 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1322 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1323 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1324 X509_CINF_get_signature were reverted post internal team review.
1326 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1328 *) Build fixes for the Windows and OpenVMS platforms
1329 [Matt Caswell and Richard Levitte]
1331 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1333 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1334 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1335 dereference. This could lead to a Denial Of Service attack. Thanks to
1336 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1340 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1341 dtls1_buffer_record function under certain conditions. In particular this
1342 could occur if an attacker sent repeated DTLS records with the same
1343 sequence number but for the next epoch. The memory leak could be exploited
1344 by an attacker in a Denial of Service attack through memory exhaustion.
1345 Thanks to Chris Mueller for reporting this issue.
1349 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1350 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1351 method would be set to NULL which could later result in a NULL pointer
1352 dereference. Thanks to Frank Schmirler for reporting this issue.
1356 *) Abort handshake if server key exchange message is omitted for ephemeral
1359 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1360 reporting this issue.
1364 *) Remove non-export ephemeral RSA code on client and server. This code
1365 violated the TLS standard by allowing the use of temporary RSA keys in
1366 non-export ciphersuites and could be used by a server to effectively
1367 downgrade the RSA key length used to a value smaller than the server
1368 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1369 INRIA or reporting this issue.
1373 *) Fixed issue where DH client certificates are accepted without verification.
1374 An OpenSSL server will accept a DH certificate for client authentication
1375 without the certificate verify message. This effectively allows a client to
1376 authenticate without the use of a private key. This only affects servers
1377 which trust a client certificate authority which issues certificates
1378 containing DH keys: these are extremely rare and hardly ever encountered.
1379 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1384 *) Ensure that the session ID context of an SSL is updated when its
1385 SSL_CTX is updated via SSL_set_SSL_CTX.
1387 The session ID context is typically set from the parent SSL_CTX,
1388 and can vary with the CTX.
1391 *) Fix various certificate fingerprint issues.
1393 By using non-DER or invalid encodings outside the signed portion of a
1394 certificate the fingerprint can be changed without breaking the signature.
1395 Although no details of the signed portion of the certificate can be changed
1396 this can cause problems with some applications: e.g. those using the
1397 certificate fingerprint for blacklists.
1399 1. Reject signatures with non zero unused bits.
1401 If the BIT STRING containing the signature has non zero unused bits reject
1402 the signature. All current signature algorithms require zero unused bits.
1404 2. Check certificate algorithm consistency.
1406 Check the AlgorithmIdentifier inside TBS matches the one in the
1407 certificate signature. NB: this will result in signature failure
1408 errors for some broken certificates.
1410 Thanks to Konrad Kraszewski from Google for reporting this issue.
1412 3. Check DSA/ECDSA signatures use DER.
1414 Reencode DSA/ECDSA signatures and compare with the original received
1415 signature. Return an error if there is a mismatch.
1417 This will reject various cases including garbage after signature
1418 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1419 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1420 (negative or with leading zeroes).
1422 Further analysis was conducted and fixes were developed by Stephen Henson
1423 of the OpenSSL core team.
1428 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1429 results on some platforms, including x86_64. This bug occurs at random
1430 with a very low probability, and is not known to be exploitable in any
1431 way, though its exact impact is difficult to determine. Thanks to Pieter
1432 Wuille (Blockstream) who reported this issue and also suggested an initial
1433 fix. Further analysis was conducted by the OpenSSL development team and
1434 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1435 the OpenSSL core team.
1439 *) Do not resume sessions on the server if the negotiated protocol
1440 version does not match the session's version. Resuming with a different
1441 version, while not strictly forbidden by the RFC, is of questionable
1442 sanity and breaks all known clients.
1443 [David Benjamin, Emilia Käsper]
1445 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1446 early CCS messages during renegotiation. (Note that because
1447 renegotiation is encrypted, this early CCS was not exploitable.)
1450 *) Tighten client-side session ticket handling during renegotiation:
1451 ensure that the client only accepts a session ticket if the server sends
1452 the extension anew in the ServerHello. Previously, a TLS client would
1453 reuse the old extension state and thus accept a session ticket if one was
1454 announced in the initial ServerHello.
1456 Similarly, ensure that the client requires a session ticket if one
1457 was advertised in the ServerHello. Previously, a TLS client would
1458 ignore a missing NewSessionTicket message.
1461 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1463 *) SRTP Memory Leak.
1465 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1466 sends a carefully crafted handshake message, to cause OpenSSL to fail
1467 to free up to 64k of memory causing a memory leak. This could be
1468 exploited in a Denial Of Service attack. This issue affects OpenSSL
1469 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1470 whether SRTP is used or configured. Implementations of OpenSSL that
1471 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1473 The fix was developed by the OpenSSL team.
1477 *) Session Ticket Memory Leak.
1479 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1480 integrity of that ticket is first verified. In the event of a session
1481 ticket integrity check failing, OpenSSL will fail to free memory
1482 causing a memory leak. By sending a large number of invalid session
1483 tickets an attacker could exploit this issue in a Denial Of Service
1488 *) Build option no-ssl3 is incomplete.
1490 When OpenSSL is configured with "no-ssl3" as a build option, servers
1491 could accept and complete a SSL 3.0 handshake, and clients could be
1492 configured to send them.
1494 [Akamai and the OpenSSL team]
1496 *) Add support for TLS_FALLBACK_SCSV.
1497 Client applications doing fallback retries should call
1498 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1500 [Adam Langley, Bodo Moeller]
1502 *) Add additional DigestInfo checks.
1504 Reencode DigestInto in DER and check against the original when
1505 verifying RSA signature: this will reject any improperly encoded
1506 DigestInfo structures.
1508 Note: this is a precautionary measure and no attacks are currently known.
1512 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1514 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1515 SRP code can be overrun an internal buffer. Add sanity check that
1516 g, A, B < N to SRP code.
1518 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1519 Group for discovering this issue.
1523 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1524 TLS 1.0 instead of higher protocol versions when the ClientHello message
1525 is badly fragmented. This allows a man-in-the-middle attacker to force a
1526 downgrade to TLS 1.0 even if both the server and the client support a
1527 higher protocol version, by modifying the client's TLS records.
1529 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1530 researching this issue.
1534 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1535 to a denial of service attack. A malicious server can crash the client
1536 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1537 ciphersuite and sending carefully crafted handshake messages.
1539 Thanks to Felix Gröbert (Google) for discovering and researching this
1544 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1545 to leak memory. This can be exploited through a Denial of Service attack.
1546 Thanks to Adam Langley for discovering and researching this issue.
1550 *) An attacker can force openssl to consume large amounts of memory whilst
1551 processing DTLS handshake messages. This can be exploited through a
1552 Denial of Service attack.
1553 Thanks to Adam Langley for discovering and researching this issue.
1557 *) An attacker can force an error condition which causes openssl to crash
1558 whilst processing DTLS packets due to memory being freed twice. This
1559 can be exploited through a Denial of Service attack.
1560 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1565 *) If a multithreaded client connects to a malicious server using a resumed
1566 session and the server sends an ec point format extension it could write
1567 up to 255 bytes to freed memory.
1569 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1574 *) A malicious server can crash an OpenSSL client with a null pointer
1575 dereference (read) by specifying an SRP ciphersuite even though it was not
1576 properly negotiated with the client. This can be exploited through a
1577 Denial of Service attack.
1579 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1580 discovering and researching this issue.
1584 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1585 X509_name_oneline, X509_name_print_ex et al. to leak some information
1586 from the stack. Applications may be affected if they echo pretty printing
1587 output to the attacker.
1589 Thanks to Ivan Fratric (Google) for discovering this issue.
1591 [Emilia Käsper, and Steve Henson]
1593 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1594 for corner cases. (Certain input points at infinity could lead to
1595 bogus results, with non-infinity inputs mapped to infinity too.)
1598 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1600 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1601 handshake can force the use of weak keying material in OpenSSL
1602 SSL/TLS clients and servers.
1604 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1605 researching this issue. (CVE-2014-0224)
1606 [KIKUCHI Masashi, Steve Henson]
1608 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1609 OpenSSL DTLS client the code can be made to recurse eventually crashing
1612 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1614 [Imre Rad, Steve Henson]
1616 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1617 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1618 client or server. This is potentially exploitable to run arbitrary
1619 code on a vulnerable client or server.
1621 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1622 [Jüri Aedla, Steve Henson]
1624 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1625 are subject to a denial of service attack.
1627 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1628 this issue. (CVE-2014-3470)
1629 [Felix Gröbert, Ivan Fratric, Steve Henson]
1631 *) Harmonize version and its documentation. -f flag is used to display
1633 [mancha <mancha1@zoho.com>]
1635 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1636 in i2d_ECPrivateKey.
1637 [mancha <mancha1@zoho.com>]
1639 *) Fix some double frees. These are not thought to be exploitable.
1640 [mancha <mancha1@zoho.com>]
1642 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1644 *) A missing bounds check in the handling of the TLS heartbeat extension
1645 can be used to reveal up to 64k of memory to a connected client or
1648 Thanks for Neel Mehta of Google Security for discovering this bug and to
1649 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1650 preparing the fix (CVE-2014-0160)
1651 [Adam Langley, Bodo Moeller]
1653 *) Fix for the attack described in the paper "Recovering OpenSSL
1654 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1655 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1656 http://eprint.iacr.org/2014/140
1658 Thanks to Yuval Yarom and Naomi Benger for discovering this
1659 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1660 [Yuval Yarom and Naomi Benger]
1662 *) TLS pad extension: draft-agl-tls-padding-03
1664 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1665 TLS client Hello record length value would otherwise be > 255 and
1666 less that 512 pad with a dummy extension containing zeroes so it
1667 is at least 512 bytes long.
1669 [Adam Langley, Steve Henson]
1671 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1673 *) Fix for TLS record tampering bug. A carefully crafted invalid
1674 handshake could crash OpenSSL with a NULL pointer exception.
1675 Thanks to Anton Johansson for reporting this issues.
1678 *) Keep original DTLS digest and encryption contexts in retransmission
1679 structures so we can use the previous session parameters if they need
1680 to be resent. (CVE-2013-6450)
1683 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1684 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1685 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1686 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1687 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1688 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1689 [Rob Stradling, Adam Langley]
1691 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1693 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1694 supporting platforms or when small records were transferred.
1695 [Andy Polyakov, Steve Henson]
1697 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1699 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1701 This addresses the flaw in CBC record processing discovered by
1702 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1703 at: http://www.isg.rhul.ac.uk/tls/
1705 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1706 Security Group at Royal Holloway, University of London
1707 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1708 Emilia Käsper for the initial patch.
1710 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1712 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1713 ciphersuites which can be exploited in a denial of service attack.
1714 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1715 and detecting this bug and to Wolfgang Ettlinger
1716 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1720 *) Return an error when checking OCSP signatures when key is NULL.
1721 This fixes a DoS attack. (CVE-2013-0166)
1724 *) Make openssl verify return errors.
1725 [Chris Palmer <palmer@google.com> and Ben Laurie]
1727 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1728 the right response is stapled. Also change SSL_get_certificate()
1729 so it returns the certificate actually sent.
1730 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1731 [Rob Stradling <rob.stradling@comodo.com>]
1733 *) Fix possible deadlock when decoding public keys.
1736 *) Don't use TLS 1.0 record version number in initial client hello
1740 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1742 *) Sanity check record length before skipping explicit IV in TLS
1743 1.2, 1.1 and DTLS to fix DoS attack.
1745 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1746 fuzzing as a service testing platform.
1750 *) Initialise tkeylen properly when encrypting CMS messages.
1751 Thanks to Solar Designer of Openwall for reporting this issue.
1754 *) In FIPS mode don't try to use composite ciphers as they are not
1758 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1760 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1761 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1762 mean any application compiled against OpenSSL 1.0.0 headers setting
1763 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1764 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1765 0x10000000L Any application which was previously compiled against
1766 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1767 will need to be recompiled as a result. Letting be results in
1768 inability to disable specifically TLS 1.1 and in client context,
1769 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1772 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1773 disable just protocol X, but all protocols above X *if* there are
1774 protocols *below* X still enabled. In more practical terms it means
1775 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1776 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1777 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1781 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1783 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1784 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1785 in CRYPTO_realloc_clean.
1787 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1788 issue and to Adam Langley <agl@chromium.org> for fixing it.
1790 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1792 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1795 *) Workarounds for some broken servers that "hang" if a client hello
1796 record length exceeds 255 bytes.
1798 1. Do not use record version number > TLS 1.0 in initial client
1799 hello: some (but not all) hanging servers will now work.
1800 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1801 the number of ciphers sent in the client hello. This should be
1802 set to an even number, such as 50, for example by passing:
1803 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1804 Most broken servers should now work.
1805 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1806 TLS 1.2 client support entirely.
1809 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1812 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1814 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1815 STRING form instead of a DigestInfo.
1818 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1819 and the RSA_sign/RSA_verify functions. This was made more apparent when
1820 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1821 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1822 the correct format in RSA_verify so both forms transparently work.
1825 *) Some servers which support TLS 1.0 can choke if we initially indicate
1826 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1827 encrypted premaster secret. As a workaround use the maximum permitted
1828 client version in client hello, this should keep such servers happy
1829 and still work with previous versions of OpenSSL.
1832 *) Add support for TLS/DTLS heartbeats.
1833 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1835 *) Add support for SCTP.
1836 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1838 *) Improved PRNG seeding for VOS.
1839 [Paul Green <Paul.Green@stratus.com>]
1841 *) Extensive assembler packs updates, most notably:
1843 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1844 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1845 - x86_64: bit-sliced AES implementation;
1846 - ARM: NEON support, contemporary platforms optimizations;
1847 - s390x: z196 support;
1848 - *: GHASH and GF(2^m) multiplication implementations;
1852 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1853 (removal of unnecessary code)
1854 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1856 *) Add TLS key material exporter from RFC 5705.
1859 *) Add DTLS-SRTP negotiation from RFC 5764.
1862 *) Add Next Protocol Negotiation,
1863 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1864 disabled with a no-npn flag to config or Configure. Code donated
1866 [Adam Langley <agl@google.com> and Ben Laurie]
1868 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1869 NIST-P256, NIST-P521, with constant-time single point multiplication on
1870 typical inputs. Compiler support for the nonstandard type __uint128_t is
1871 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1872 Code made available under Apache License version 2.0.
1874 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1875 line to include this in your build of OpenSSL, and run "make depend" (or
1876 "make update"). This enables the following EC_METHODs:
1878 EC_GFp_nistp224_method()
1879 EC_GFp_nistp256_method()
1880 EC_GFp_nistp521_method()
1882 EC_GROUP_new_by_curve_name() will automatically use these (while
1883 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1885 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1887 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1888 all platforms. Move ssize_t definition from e_os.h to the public
1889 header file e_os2.h as it now appears in public header file cms.h
1892 *) New -sigopt option to the ca, req and x509 utilities. Additional
1893 signature parameters can be passed using this option and in
1897 *) Add RSA PSS signing function. This will generate and set the
1898 appropriate AlgorithmIdentifiers for PSS based on those in the
1899 corresponding EVP_MD_CTX structure. No application support yet.
1902 *) Support for companion algorithm specific ASN1 signing routines.
1903 New function ASN1_item_sign_ctx() signs a pre-initialised
1904 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1905 the appropriate parameters.
1908 *) Add new algorithm specific ASN1 verification initialisation function
1909 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1910 handling will be the same no matter what EVP_PKEY_METHOD is used.
1911 Add a PSS handler to support verification of PSS signatures: checked
1912 against a number of sample certificates.
1915 *) Add signature printing for PSS. Add PSS OIDs.
1916 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1918 *) Add algorithm specific signature printing. An individual ASN1 method
1919 can now print out signatures instead of the standard hex dump.
1921 More complex signatures (e.g. PSS) can print out more meaningful
1922 information. Include DSA version that prints out the signature
1926 *) Password based recipient info support for CMS library: implementing
1930 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1931 neatly separates the code into cipher and PBE sections and is required
1932 for some algorithms that split PBES2 into separate pieces (such as
1933 password based CMS).
1936 *) Session-handling fixes:
1937 - Fix handling of connections that are resuming with a session ID,
1938 but also support Session Tickets.
1939 - Fix a bug that suppressed issuing of a new ticket if the client
1940 presented a ticket with an expired session.
1941 - Try to set the ticket lifetime hint to something reasonable.
1942 - Make tickets shorter by excluding irrelevant information.
1943 - On the client side, don't ignore renewed tickets.
1944 [Adam Langley, Bodo Moeller (Google)]
1946 *) Fix PSK session representation.
1949 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1951 This work was sponsored by Intel.
1954 *) Add GCM support to TLS library. Some custom code is needed to split
1955 the IV between the fixed (from PRF) and explicit (from TLS record)
1956 portions. This adds all GCM ciphersuites supported by RFC5288 and
1957 RFC5289. Generalise some AES* cipherstrings to include GCM and
1958 add a special AESGCM string for GCM only.
1961 *) Expand range of ctrls for AES GCM. Permit setting invocation
1962 field on decrypt and retrieval of invocation field only on encrypt.
1965 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1966 As required by RFC5289 these ciphersuites cannot be used if for
1967 versions of TLS earlier than 1.2.
1970 *) For FIPS capable OpenSSL interpret a NULL default public key method
1971 as unset and return the appropriate default but do *not* set the default.
1972 This means we can return the appropriate method in applications that
1973 switch between FIPS and non-FIPS modes.
1976 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1977 ENGINE is used then we cannot handle that in the FIPS module so we
1978 keep original code iff non-FIPS operations are allowed.
1981 *) Add -attime option to openssl utilities.
1982 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1984 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1987 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1988 FIPS EC methods unconditionally for now.
1991 *) New build option no-ec2m to disable characteristic 2 code.
1994 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1995 all cases can be covered as some introduce binary incompatibilities.
1998 *) Redirect RSA operations to FIPS module including keygen,
1999 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2002 *) Add similar low level API blocking to ciphers.
2005 *) Low level digest APIs are not approved in FIPS mode: any attempt
2006 to use these will cause a fatal error. Applications that *really* want
2007 to use them can use the private_* version instead.
2010 *) Redirect cipher operations to FIPS module for FIPS builds.
2013 *) Redirect digest operations to FIPS module for FIPS builds.
2016 *) Update build system to add "fips" flag which will link in fipscanister.o
2017 for static and shared library builds embedding a signature if needed.
2020 *) Output TLS supported curves in preference order instead of numerical
2021 order. This is currently hardcoded for the highest order curves first.
2022 This should be configurable so applications can judge speed vs strength.
2025 *) Add TLS v1.2 server support for client authentication.
2028 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2032 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2033 FIPS modules versions.
2036 *) Add TLS v1.2 client side support for client authentication. Keep cache
2037 of handshake records longer as we don't know the hash algorithm to use
2038 until after the certificate request message is received.
2041 *) Initial TLS v1.2 client support. Add a default signature algorithms
2042 extension including all the algorithms we support. Parse new signature
2043 format in client key exchange. Relax some ECC signing restrictions for
2044 TLS v1.2 as indicated in RFC5246.
2047 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2048 to new signature format when needed using client digest preference.
2049 All server ciphersuites should now work correctly in TLS v1.2. No client
2050 support yet and no support for client certificates.
2053 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2054 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2055 ciphersuites. At present only RSA key exchange ciphersuites work with
2056 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2057 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2058 and version checking.
2061 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2062 with this defined it will not be affected by any changes to ssl internal
2063 structures. Add several utility functions to allow openssl application
2064 to work with OPENSSL_NO_SSL_INTERN defined.
2068 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2070 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2073 *) Permit abbreviated handshakes when renegotiating using the function
2074 SSL_renegotiate_abbreviated().
2075 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2077 *) Add call to ENGINE_register_all_complete() to
2078 ENGINE_load_builtin_engines(), so some implementations get used
2079 automatically instead of needing explicit application support.
2082 *) Add support for TLS key exporter as described in RFC5705.
2083 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2085 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2086 a few changes are required:
2088 Add SSL_OP_NO_TLSv1_1 flag.
2089 Add TLSv1_1 methods.
2090 Update version checking logic to handle version 1.1.
2091 Add explicit IV handling (ported from DTLS code).
2092 Add command line options to s_client/s_server.
2095 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2097 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2098 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2099 content decryption and always return the same error. Note: this attack
2100 needs on average 2^20 messages so it only affects automated senders. The
2101 old behaviour can be reenabled in the CMS code by setting the
2102 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2103 an MMA defence is not necessary.
2104 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2105 this issue. (CVE-2012-0884)
2108 *) Fix CVE-2011-4619: make sure we really are receiving a
2109 client hello before rejecting multiple SGC restarts. Thanks to
2110 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2113 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2115 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2116 Thanks to Antonio Martin, Enterprise Secure Access Research and
2117 Development, Cisco Systems, Inc. for discovering this bug and
2118 preparing a fix. (CVE-2012-0050)
2121 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2123 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2124 of the Vaudenay padding oracle attack on CBC mode encryption
2125 which enables an efficient plaintext recovery attack against
2126 the OpenSSL implementation of DTLS. Their attack exploits timing
2127 differences arising during decryption processing. A research
2128 paper describing this attack can be found at:
2129 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2130 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2131 Security Group at Royal Holloway, University of London
2132 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2133 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2134 for preparing the fix. (CVE-2011-4108)
2135 [Robin Seggelmann, Michael Tuexen]
2137 *) Clear bytes used for block padding of SSL 3.0 records.
2139 [Adam Langley (Google)]
2141 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2142 Kadianakis <desnacked@gmail.com> for discovering this issue and
2143 Adam Langley for preparing the fix. (CVE-2011-4619)
2144 [Adam Langley (Google)]
2146 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2147 [Andrey Kulikov <amdeich@gmail.com>]
2149 *) Prevent malformed RFC3779 data triggering an assertion failure.
2150 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2151 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2152 [Rob Austein <sra@hactrn.net>]
2154 *) Improved PRNG seeding for VOS.
2155 [Paul Green <Paul.Green@stratus.com>]
2157 *) Fix ssl_ciph.c set-up race.
2158 [Adam Langley (Google)]
2160 *) Fix spurious failures in ecdsatest.c.
2161 [Emilia Käsper (Google)]
2163 *) Fix the BIO_f_buffer() implementation (which was mixing different
2164 interpretations of the '..._len' fields).
2165 [Adam Langley (Google)]
2167 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2168 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2169 threads won't reuse the same blinding coefficients.
2171 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2172 lock to call BN_BLINDING_invert_ex, and avoids one use of
2173 BN_BLINDING_update for each BN_BLINDING structure (previously,
2174 the last update always remained unused).
2175 [Emilia Käsper (Google)]
2177 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2178 [Bob Buckholz (Google)]
2180 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2182 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2183 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2184 [Kaspar Brand <ossl@velox.ch>]
2186 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2187 for multi-threaded use of ECDH. (CVE-2011-3210)
2188 [Adam Langley (Google)]
2190 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2193 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2194 signature public key algorithm by using OID xref utilities instead.
2195 Before this you could only use some ECC ciphersuites with SHA1 only.
2198 *) Add protection against ECDSA timing attacks as mentioned in the paper
2199 by Billy Bob Brumley and Nicola Tuveri, see:
2201 http://eprint.iacr.org/2011/232.pdf
2203 [Billy Bob Brumley and Nicola Tuveri]
2205 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2207 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2208 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2210 *) Fix bug in string printing code: if *any* escaping is enabled we must
2211 escape the escape character (backslash) or the resulting string is
2215 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2217 *) Disable code workaround for ancient and obsolete Netscape browsers
2218 and servers: an attacker can use it in a ciphersuite downgrade attack.
2219 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2222 *) Fixed J-PAKE implementation error, originally discovered by
2223 Sebastien Martini, further info and confirmation from Stefan
2224 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2227 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2229 *) Fix extension code to avoid race conditions which can result in a buffer
2230 overrun vulnerability: resumed sessions must not be modified as they can
2231 be shared by multiple threads. CVE-2010-3864
2234 *) Fix WIN32 build system to correctly link an ENGINE directory into
2238 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2240 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2242 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2244 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2246 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2247 context. The operation can be customised via the ctrl mechanism in
2248 case ENGINEs want to include additional functionality.
2251 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2254 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2255 output hashes compatible with older versions of OpenSSL.
2256 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2258 *) Fix compression algorithm handling: if resuming a session use the
2259 compression algorithm of the resumed session instead of determining
2260 it from client hello again. Don't allow server to change algorithm.
2263 *) Add load_crls() function to apps tidying load_certs() too. Add option
2264 to verify utility to allow additional CRLs to be included.
2267 *) Update OCSP request code to permit adding custom headers to the request:
2268 some responders need this.
2271 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2273 [Julia Lawall <julia@diku.dk>]
2275 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2276 needlessly dereferenced structures, used obsolete functions and
2277 didn't handle all updated verify codes correctly.
2280 *) Disable MD2 in the default configuration.
2283 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2284 indicate the initial BIO being pushed or popped. This makes it possible
2285 to determine whether the BIO is the one explicitly called or as a result
2286 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2287 it handles reference counts correctly and doesn't zero out the I/O bio
2288 when it is not being explicitly popped. WARNING: applications which
2289 included workarounds for the old buggy behaviour will need to be modified
2290 or they could free up already freed BIOs.
2293 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2294 renaming to all platforms (within the 0.9.8 branch, this was
2295 done conditionally on Netware platforms to avoid a name clash).
2296 [Guenter <lists@gknw.net>]
2298 *) Add ECDHE and PSK support to DTLS.
2299 [Michael Tuexen <tuexen@fh-muenster.de>]
2301 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2305 *) Add "missing" function EVP_MD_flags() (without this the only way to
2306 retrieve a digest flags is by accessing the structure directly. Update
2307 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2308 or cipher is registered as in the "from" argument. Print out all
2309 registered digests in the dgst usage message instead of manually
2310 attempting to work them out.
2313 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2314 this allows the use of compression and extensions. Change default cipher
2315 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2316 by default unless an application cipher string requests it.
2319 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2320 key ids to find matching certificates and keys but some PKCS#12 files
2321 don't follow the (somewhat unwritten) rules and this strategy fails.
2322 Now just gather all certificates together and the first private key
2323 then look for the first certificate that matches the key.
2326 *) Support use of registered digest and cipher names for dgst and cipher
2327 commands instead of having to add each one as a special case. So now
2334 openssl dgst -sha256 foo
2336 and this works for ENGINE based algorithms too.
2340 *) Update Gost ENGINE to support parameter files.
2341 [Victor B. Wagner <vitus@cryptocom.ru>]
2343 *) Support GeneralizedTime in ca utility.
2344 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2346 *) Enhance the hash format used for certificate directory links. The new
2347 form uses the canonical encoding (meaning equivalent names will work
2348 even if they aren't identical) and uses SHA1 instead of MD5. This form
2349 is incompatible with the older format and as a result c_rehash should
2350 be used to rebuild symbolic links.
2353 *) Make PKCS#8 the default write format for private keys, replacing the
2354 traditional format. This form is standardised, more secure and doesn't
2355 include an implicit MD5 dependency.
2358 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2359 committed to OpenSSL should pass this lot as a minimum.
2362 *) Add session ticket override functionality for use by EAP-FAST.
2363 [Jouni Malinen <j@w1.fi>]
2365 *) Modify HMAC functions to return a value. Since these can be implemented
2366 in an ENGINE errors can occur.
2369 *) Type-checked OBJ_bsearch_ex.
2372 *) Type-checked OBJ_bsearch. Also some constification necessitated
2373 by type-checking. Still to come: TXT_DB, bsearch(?),
2374 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2378 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2379 seconds to a tm structure directly, instead of going through OS
2380 specific date routines. This avoids any issues with OS routines such
2381 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2382 and X509_time_adj_ex() to cover the extended range. The existing
2383 X509_time_adj() is still usable and will no longer have any date issues.
2386 *) Delta CRL support. New use deltas option which will attempt to locate
2387 and search any appropriate delta CRLs available.
2389 This work was sponsored by Google.
2392 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2393 code and add additional score elements. Validate alternate CRL paths
2394 as part of the CRL checking and indicate a new error "CRL path validation
2395 error" in this case. Applications wanting additional details can use
2396 the verify callback and check the new "parent" field. If this is not
2397 NULL CRL path validation is taking place. Existing applications wont
2398 see this because it requires extended CRL support which is off by
2401 This work was sponsored by Google.
2404 *) Support for freshest CRL extension.
2406 This work was sponsored by Google.
2409 *) Initial indirect CRL support. Currently only supported in the CRLs
2410 passed directly and not via lookup. Process certificate issuer
2411 CRL entry extension and lookup CRL entries by bother issuer name
2412 and serial number. Check and process CRL issuer entry in IDP extension.
2414 This work was sponsored by Google.
2417 *) Add support for distinct certificate and CRL paths. The CRL issuer
2418 certificate is validated separately in this case. Only enabled if
2419 an extended CRL support flag is set: this flag will enable additional
2420 CRL functionality in future.
2422 This work was sponsored by Google.
2425 *) Add support for policy mappings extension.
2427 This work was sponsored by Google.
2430 *) Fixes to pathlength constraint, self issued certificate handling,
2431 policy processing to align with RFC3280 and PKITS tests.
2433 This work was sponsored by Google.
2436 *) Support for name constraints certificate extension. DN, email, DNS
2437 and URI types are currently supported.
2439 This work was sponsored by Google.
2442 *) To cater for systems that provide a pointer-based thread ID rather
2443 than numeric, deprecate the current numeric thread ID mechanism and
2444 replace it with a structure and associated callback type. This
2445 mechanism allows a numeric "hash" to be extracted from a thread ID in
2446 either case, and on platforms where pointers are larger than 'long',
2447 mixing is done to help ensure the numeric 'hash' is usable even if it
2448 can't be guaranteed unique. The default mechanism is to use "&errno"
2449 as a pointer-based thread ID to distinguish between threads.
2451 Applications that want to provide their own thread IDs should now use
2452 CRYPTO_THREADID_set_callback() to register a callback that will call
2453 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2455 Note that ERR_remove_state() is now deprecated, because it is tied
2456 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2457 to free the current thread's error state should be replaced by
2458 ERR_remove_thread_state(NULL).
2460 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2461 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2462 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2463 application was previously providing a numeric thread callback that
2464 was inappropriate for distinguishing threads, then uniqueness might
2465 have been obtained with &errno that happened immediately in the
2466 intermediate development versions of OpenSSL; this is no longer the
2467 case, the numeric thread callback will now override the automatic use
2469 [Geoff Thorpe, with help from Bodo Moeller]
2471 *) Initial support for different CRL issuing certificates. This covers a
2472 simple case where the self issued certificates in the chain exist and
2473 the real CRL issuer is higher in the existing chain.
2475 This work was sponsored by Google.
2478 *) Removed effectively defunct crypto/store from the build.
2481 *) Revamp of STACK to provide stronger type-checking. Still to come:
2482 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2483 ASN1_STRING, CONF_VALUE.
2486 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2487 RAM on SSL connections. This option can save about 34k per idle SSL.
2490 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2491 STACK, TXT_DB, bsearch, qsort.
2494 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2495 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2496 support for data, signedData, compressedData, digestedData and
2497 encryptedData, envelopedData types included. Scripts to check against
2498 RFC4134 examples draft and interop and consistency checks of many
2499 content types and variants.
2502 *) Add options to enc utility to support use of zlib compression BIO.
2505 *) Extend mk1mf to support importing of options and assembly language
2506 files from Configure script, currently only included in VC-WIN32.
2507 The assembly language rules can now optionally generate the source
2508 files from the associated perl scripts.
2511 *) Implement remaining functionality needed to support GOST ciphersuites.
2512 Interop testing has been performed using CryptoPro implementations.
2513 [Victor B. Wagner <vitus@cryptocom.ru>]
2515 *) s390x assembler pack.
2518 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2522 *) Implement Opaque PRF Input TLS extension as specified in
2523 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2524 official specification yet and no extension type assignment by
2525 IANA exists, this extension (for now) will have to be explicitly
2526 enabled when building OpenSSL by providing the extension number
2527 to use. For example, specify an option
2529 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2531 to the "config" or "Configure" script to enable the extension,
2532 assuming extension number 0x9527 (which is a completely arbitrary
2533 and unofficial assignment based on the MD5 hash of the Internet
2534 Draft). Note that by doing so, you potentially lose
2535 interoperability with other TLS implementations since these might
2536 be using the same extension number for other purposes.
2538 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2539 opaque PRF input value to use in the handshake. This will create
2540 an interal copy of the length-'len' string at 'src', and will
2541 return non-zero for success.
2543 To get more control and flexibility, provide a callback function
2546 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2547 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2551 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2554 Callback function 'cb' will be called in handshakes, and is
2555 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2556 Argument 'arg' is for application purposes (the value as given to
2557 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2558 be provided to the callback function). The callback function
2559 has to return non-zero to report success: usually 1 to use opaque
2560 PRF input just if possible, or 2 to enforce use of the opaque PRF
2561 input. In the latter case, the library will abort the handshake
2562 if opaque PRF input is not successfully negotiated.
2564 Arguments 'peerinput' and 'len' given to the callback function
2565 will always be NULL and 0 in the case of a client. A server will
2566 see the client's opaque PRF input through these variables if
2567 available (NULL and 0 otherwise). Note that if the server
2568 provides an opaque PRF input, the length must be the same as the
2569 length of the client's opaque PRF input.
2571 Note that the callback function will only be called when creating
2572 a new session (session resumption can resume whatever was
2573 previously negotiated), and will not be called in SSL 2.0
2574 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2575 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2576 for applications that need to enforce opaque PRF input.
2580 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2583 [Victor B. Wagner <vitus@cryptocom.ru>]
2585 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2586 RFC4507bis. The encrypted ticket format is an encrypted encoded
2587 SSL_SESSION structure, that way new session features are automatically
2590 If a client application caches session in an SSL_SESSION structure
2591 support is transparent because tickets are now stored in the encoded
2594 The SSL_CTX structure automatically generates keys for ticket
2595 protection in servers so again support should be possible
2596 with no application modification.
2598 If a client or server wishes to disable RFC4507 support then the option
2599 SSL_OP_NO_TICKET can be set.
2601 Add a TLS extension debugging callback to allow the contents of any client
2602 or server extensions to be examined.
2604 This work was sponsored by Google.
2607 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2608 OpenSSL should now compile cleanly on gcc 4.2
2609 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2611 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2612 support including streaming MAC support: this is required for GOST
2613 ciphersuite support.
2614 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2616 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2617 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2618 to output in BER and PEM format.
2621 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2622 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2623 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2624 ENGINE support for HMAC keys which are unextractable. New -mac and
2625 -macopt options to dgst utility.
2628 *) New option -sigopt to dgst utility. Update dgst to use
2629 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2630 alternative signing paramaters such as X9.31 or PSS in the dgst
2634 *) Change ssl_cipher_apply_rule(), the internal function that does
2635 the work each time a ciphersuite string requests enabling
2636 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2637 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2638 the order of disabled ciphersuites such that those ciphersuites
2639 that most recently went from enabled to disabled not only stay
2640 in order with respect to each other, but also have higher priority
2641 than other disabled ciphersuites the next time ciphersuites are
2644 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2645 the same ciphersuites as with "HIGH" alone, but in a specific
2646 order where the PSK ciphersuites come first (since they are the
2647 most recently disabled ciphersuites when "HIGH" is parsed).
2649 Also, change ssl_create_cipher_list() (using this new
2650 funcionality) such that between otherwise identical
2651 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2655 *) Change ssl_create_cipher_list() so that it automatically
2656 arranges the ciphersuites in reasonable order before starting
2657 to process the rule string. Thus, the definition for "DEFAULT"
2658 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2659 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2660 This makes it much easier to arrive at a reasonable default order
2661 in applications for which anonymous ciphers are OK (meaning
2662 that you can't actually use DEFAULT).
2663 [Bodo Moeller; suggested by Victor Duchovni]
2665 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2666 processing) into multiple integers instead of setting
2667 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2668 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2669 (These masks as well as the individual bit definitions are hidden
2670 away into the non-exported interface ssl/ssl_locl.h, so this
2671 change to the definition of the SSL_CIPHER structure shouldn't
2672 affect applications.) This give us more bits for each of these
2673 categories, so there is no longer a need to coagulate AES128 and
2674 AES256 into a single algorithm bit, and to coagulate Camellia128
2675 and Camellia256 into a single algorithm bit, which has led to all
2678 Thus, among other things, the kludge introduced in 0.9.7m and
2679 0.9.8e for masking out AES256 independently of AES128 or masking
2680 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2682 With the change, we also introduce new ciphersuite aliases that
2683 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2687 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2688 Use the leftmost N bytes of the signature input if the input is
2689 larger than the prime q (with N being the size in bytes of q).
2692 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2693 it yet and it is largely untested.
2696 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2699 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2700 some compilers (gcc 4.2 and later) reject their use. Safestack is
2701 reimplemented. Update ASN1 to avoid use of legacy functions.
2704 *) Win32/64 targets are linked with Winsock2.
2707 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2708 to external functions. This can be used to increase CRL handling
2709 efficiency especially when CRLs are very large by (for example) storing
2710 the CRL revoked certificates in a database.
2713 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2714 new CRLs added to a directory can be used. New command line option
2715 -verify_return_error to s_client and s_server. This causes real errors
2716 to be returned by the verify callback instead of carrying on no matter
2717 what. This reflects the way a "real world" verify callback would behave.
2720 *) GOST engine, supporting several GOST algorithms and public key formats.
2721 Kindly donated by Cryptocom.
2724 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2725 partitioned by DP are handled but no indirect CRL or reason partitioning
2726 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2727 selected via a scoring technique which handles IDP and AKID in CRLs.
2730 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2731 will ultimately be used for all verify operations: this will remove the
2732 X509_STORE dependency on certificate verification and allow alternative
2733 lookup methods. X509_STORE based implementations of these two callbacks.
2736 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2737 Modify get_crl() to find a valid (unexpired) CRL if possible.
2740 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2741 this would be called X509_CRL_cmp() but that name is already used by
2742 a function that just compares CRL issuer names. Cache several CRL
2743 extensions in X509_CRL structure and cache CRLDP in X509.
2746 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2747 this maps equivalent X509_NAME structures into a consistent structure.
2748 Name comparison can then be performed rapidly using memcmp().
2751 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2755 *) Allow digests to supply their own micalg string for S/MIME type using
2756 the ctrl EVP_MD_CTRL_MICALG.
2759 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2760 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2761 ctrl. It can then customise the structure before and/or after signing
2765 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2766 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2767 to free up any added signature OIDs.
2770 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2771 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2772 digest and cipher tables. New options added to openssl utility:
2773 list-message-digest-algorithms and list-cipher-algorithms.
2776 *) Change the array representation of binary polynomials: the list
2777 of degrees of non-zero coefficients is now terminated with -1.
2778 Previously it was terminated with 0, which was also part of the
2779 value; thus, the array representation was not applicable to
2780 polynomials where t^0 has coefficient zero. This change makes
2781 the array representation useful in a more general context.
2784 *) Various modifications and fixes to SSL/TLS cipher string
2785 handling. For ECC, the code now distinguishes between fixed ECDH
2786 with RSA certificates on the one hand and with ECDSA certificates
2787 on the other hand, since these are separate ciphersuites. The
2788 unused code for Fortezza ciphersuites has been removed.
2790 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2791 (not "ECDHE"). For consistency with the code for DH
2792 certificates, use of ECDH certificates is now considered ECDH
2793 authentication, not RSA or ECDSA authentication (the latter is
2794 merely the CA's signing algorithm and not actively used in the
2797 The temporary ciphersuite alias "ECCdraft" is no longer
2798 available, and ECC ciphersuites are no longer excluded from "ALL"
2799 and "DEFAULT". The following aliases now exist for RFC 4492
2800 ciphersuites, most of these by analogy with the DH case:
2802 kECDHr - ECDH cert, signed with RSA
2803 kECDHe - ECDH cert, signed with ECDSA
2804 kECDH - ECDH cert (signed with either RSA or ECDSA)
2805 kEECDH - ephemeral ECDH
2806 ECDH - ECDH cert or ephemeral ECDH
2812 AECDH - anonymous ECDH
2813 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2817 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2818 Use correct micalg parameters depending on digest(s) in signed message.
2821 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2822 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2825 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2826 an engine to register a method. Add ENGINE lookups for methods and
2827 functional reference processing.
2830 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2831 EVP_{Sign,Verify}* which allow an application to customise the signature
2835 *) New -resign option to smime utility. This adds one or more signers
2836 to an existing PKCS#7 signedData structure. Also -md option to use an
2837 alternative message digest algorithm for signing.
2840 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2841 create PKCS7 structures containing multiple signers. Update smime
2842 application to support multiple signers.
2845 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2849 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2850 Reorganize PBE internals to lookup from a static table using NIDs,
2851 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2852 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2853 PRF which will be automatically used with PBES2.
2856 *) Replace the algorithm specific calls to generate keys in "req" with the
2860 *) Update PKCS#7 enveloped data routines to use new API. This is now
2861 supported by any public key method supporting the encrypt operation. A
2862 ctrl is added to allow the public key algorithm to examine or modify
2863 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2867 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2868 a default digest type to use. In most cases this will be SHA1 but some
2869 algorithms (such as GOST) need to specify an alternative digest. The
2870 return value indicates how strong the preference is 1 means optional and
2871 2 is mandatory (that is it is the only supported type). Modify
2872 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2873 use the default md. Update openssl utilities to use the default digest
2874 type for signing if it is not explicitly indicated.
2877 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2878 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2879 signing method from the key type. This effectively removes the link
2880 between digests and public key types.
2883 *) Add an OID cross reference table and utility functions. Its purpose is to
2884 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2885 rsaEncryption. This will allow some of the algorithm specific hackery
2886 needed to use the correct OID to be removed.
2889 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2890 structures for PKCS7_sign(). They are now set up by the relevant public
2894 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2897 *) Add support for key derivation (agreement) in the API, DH method and
2901 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2902 public and private key formats. As a side effect these add additional
2903 command line functionality not previously available: DSA signatures can be
2904 generated and verified using pkeyutl and DH key support and generation in
2909 [Oliver Tappe <zooey@hirschkaefer.de>]
2911 *) New make target "install_html_docs" installs HTML renditions of the
2913 [Oliver Tappe <zooey@hirschkaefer.de>]
2915 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2916 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2917 support key and parameter generation and add initial key generation
2918 functionality for RSA.
2921 *) Add functions for main EVP_PKEY_method operations. The undocumented
2922 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2923 EVP_PKEY_{encrypt,decrypt}_old.
2926 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2927 key API, doesn't do much yet.
2930 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2931 public key algorithms. New option to openssl utility:
2932 "list-public-key-algorithms" to print out info.
2935 *) Implement the Supported Elliptic Curves Extension for
2936 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2939 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2940 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2943 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2944 utilities such as rsa, dsa, dsaparam etc except they process any key
2948 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2949 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2950 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2954 *) Initial support for pluggable public key ASN1.
2955 De-spaghettify the public key ASN1 handling. Move public and private
2956 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2957 algorithm specific handling to a single module within the relevant
2958 algorithm directory. Add functions to allow (near) opaque processing
2959 of public and private key structures.
2962 *) Implement the Supported Point Formats Extension for
2963 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2966 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2967 for the psk identity [hint] and the psk callback functions to the
2968 SSL_SESSION, SSL and SSL_CTX structure.
2971 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2975 SSL_CTX_use_psk_identity_hint
2976 SSL_get_psk_identity_hint
2977 SSL_get_psk_identity
2978 SSL_use_psk_identity_hint
2980 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2982 *) Add RFC 3161 compliant time stamp request creation, response generation
2983 and response verification functionality.
2984 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2986 *) Add initial support for TLS extensions, specifically for the server_name
2987 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2988 have new members for a host name. The SSL data structure has an
2989 additional member SSL_CTX *initial_ctx so that new sessions can be
2990 stored in that context to allow for session resumption, even after the
2991 SSL has been switched to a new SSL_CTX in reaction to a client's
2992 server_name extension.
2994 New functions (subject to change):
2996 SSL_get_servername()
2997 SSL_get_servername_type()
3000 New CTRL codes and macros (subject to change):
3002 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3003 - SSL_CTX_set_tlsext_servername_callback()
3004 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3005 - SSL_CTX_set_tlsext_servername_arg()
3006 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3008 openssl s_client has a new '-servername ...' option.
3010 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3011 '-key2 ...', '-servername_fatal' (subject to change). This allows
3012 testing the HostName extension for a specific single host name ('-cert'
3013 and '-key' remain fallbacks for handshakes without HostName
3014 negotiation). If the unrecognized_name alert has to be sent, this by
3015 default is a warning; it becomes fatal with the '-servername_fatal'
3018 [Peter Sylvester, Remy Allais, Christophe Renou]
3020 *) Whirlpool hash implementation is added.
3023 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3024 bn(64,32). Because of instruction set limitations it doesn't have
3025 any negative impact on performance. This was done mostly in order
3026 to make it possible to share assembler modules, such as bn_mul_mont
3027 implementations, between 32- and 64-bit builds without hassle.
3030 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3031 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3035 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3036 dedicated Montgomery multiplication procedure, is introduced.
3037 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3038 "64-bit" performance on certain 32-bit targets.
3041 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3042 in SSL structures. New SSL ctrl to set maximum send fragment size.
3043 Save memory by seeting the I/O buffer sizes dynamically instead of
3044 using the maximum available value.
3047 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3048 in addition to the text details.
3051 *) Very, very preliminary EXPERIMENTAL support for printing of general
3052 ASN1 structures. This currently produces rather ugly output and doesn't
3053 handle several customised structures at all.
3056 *) Integrated support for PVK file format and some related formats such
3057 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3058 these in the 'rsa' and 'dsa' utilities.
3061 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3064 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3065 place for the (very old) "NETSCAPE" format certificates which are now
3066 handled using new ASN1 code equivalents.
3069 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3070 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3071 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3074 *) Modify CRL distribution points extension code to print out previously
3075 unsupported fields. Enhance extension setting code to allow setting of
3079 *) Add print and set support for Issuing Distribution Point CRL extension.
3082 *) Change 'Configure' script to enable Camellia by default.
3085 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3087 *) When rejecting SSL/TLS records due to an incorrect version number, never
3088 update s->server with a new major version number. As of
3089 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3090 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3091 the previous behavior could result in a read attempt at NULL when
3092 receiving specific incorrect SSL/TLS records once record payload
3093 protection is active. (CVE-2010-0740)
3094 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3096 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3097 could be crashed if the relevant tables were not present (e.g. chrooted).
3098 [Tomas Hoger <thoger@redhat.com>]
3100 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3102 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3103 [Martin Olsson, Neel Mehta]
3105 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3106 accommodate for stack sorting, always a write lock!).
3109 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3110 excessive delays in the RAND_poll(): over a minute. As a workaround
3111 include a time check in the inner Heap32Next loop too.
3114 *) The code that handled flushing of data in SSL/TLS originally used the
3115 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3116 the problem outlined in PR#1949. The fix suggested there however can
3117 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3118 of Apache). So instead simplify the code to flush unconditionally.
3119 This should be fine since flushing with no data to flush is a no op.
3122 *) Handle TLS versions 2.0 and later properly and correctly use the
3123 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3124 off ancient servers have a habit of sticking around for a while...
3127 *) Modify compression code so it frees up structures without using the
3128 ex_data callbacks. This works around a problem where some applications
3129 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3130 restarting) then use compression (e.g. SSL with compression) later.
3131 This results in significant per-connection memory leaks and
3132 has caused some security issues including CVE-2008-1678 and
3136 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3137 change when encrypting or decrypting.
3140 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3141 connect and renegotiate with servers which do not support RI.
3142 Until RI is more widely deployed this option is enabled by default.
3145 *) Add "missing" ssl ctrls to clear options and mode.
3148 *) If client attempts to renegotiate and doesn't support RI respond with
3149 a no_renegotiation alert as required by RFC5746. Some renegotiating
3150 TLS clients will continue a connection gracefully when they receive
3151 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3152 waiting for a server hello which it will never receive. Now we treat a
3153 received no_renegotiation alert as a fatal error. This is because
3154 applications requesting a renegotiation might well expect it to succeed
3155 and would have no code in place to handle the server denying it so the
3156 only safe thing to do is to terminate the connection.
3159 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3160 peer supports secure renegotiation and 0 otherwise. Print out peer
3161 renegotiation support in s_client/s_server.
3164 *) Replace the highly broken and deprecated SPKAC certification method with
3165 the updated NID creation version. This should correctly handle UTF8.
3168 *) Implement RFC5746. Re-enable renegotiation but require the extension
3169 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3170 turns out to be a bad idea. It has been replaced by
3171 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3172 SSL_CTX_set_options(). This is really not recommended unless you
3173 know what you are doing.
3174 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3176 *) Fixes to stateless session resumption handling. Use initial_ctx when
3177 issuing and attempting to decrypt tickets in case it has changed during
3178 servername handling. Use a non-zero length session ID when attempting
3179 stateless session resumption: this makes it possible to determine if
3180 a resumption has occurred immediately after receiving server hello
3181 (several places in OpenSSL subtly assume this) instead of later in
3185 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3186 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3187 fixes for a few places where the return code is not checked
3189 [Julia Lawall <julia@diku.dk>]
3191 *) Add --strict-warnings option to Configure script to include devteam
3192 warnings in other configurations.
3195 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3196 makes it possible to install openssl libraries in locations which
3197 have names other than "lib", for example "/usr/lib64" which some
3199 [Steve Henson, based on patch from Jeremy Utley]
3201 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3202 X690 8.9.12 and can produce some misleading textual output of OIDs.
3203 [Steve Henson, reported by Dan Kaminsky]
3205 *) Delete MD2 from algorithm tables. This follows the recommendation in
3206 several standards that it is not used in new applications due to
3207 several cryptographic weaknesses. For binary compatibility reasons
3208 the MD2 API is still compiled in by default.
3211 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3215 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3216 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3218 [Guenter <lists@gknw.net>]
3220 *) Fix the server certificate chain building code to use X509_verify_cert(),
3221 it used to have an ad-hoc builder which was unable to cope with anything
3222 other than a simple chain.
3223 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3225 *) Don't check self signed certificate signatures in X509_verify_cert()
3226 by default (a flag can override this): it just wastes time without
3227 adding any security. As a useful side effect self signed root CAs
3228 with non-FIPS digests are now usable in FIPS mode.
3231 *) In dtls1_process_out_of_seq_message() the check if the current message
3232 is already buffered was missing. For every new message was memory
3233 allocated, allowing an attacker to perform an denial of service attack
3234 with sending out of seq handshake messages until there is no memory
3235 left. Additionally every future messege was buffered, even if the
3236 sequence number made no sense and would be part of another handshake.
3237 So only messages with sequence numbers less than 10 in advance will be
3238 buffered. (CVE-2009-1378)
3239 [Robin Seggelmann, discovered by Daniel Mentz]
3241 *) Records are buffered if they arrive with a future epoch to be
3242 processed after finishing the corresponding handshake. There is
3243 currently no limitation to this buffer allowing an attacker to perform
3244 a DOS attack with sending records with future epochs until there is no
3245 memory left. This patch adds the pqueue_size() function to determine
3246 the size of a buffer and limits the record buffer to 100 entries.
3248 [Robin Seggelmann, discovered by Daniel Mentz]
3250 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3251 parent structure is freed. (CVE-2009-1379)
3254 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3255 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3258 [Ilya O. <vrghost@gmail.com>]
3260 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3262 *) Disable renegotiation completely - this fixes a severe security
3263 problem (CVE-2009-3555) at the cost of breaking all
3264 renegotiation. Renegotiation can be re-enabled by setting
3265 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3266 run-time. This is really not recommended unless you know what
3270 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3272 *) Don't set val to NULL when freeing up structures, it is freed up by
3273 underlying code. If sizeof(void *) > sizeof(long) this can result in
3274 zeroing past the valid field. (CVE-2009-0789)
3275 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3277 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3278 checked correctly. This would allow some invalid signed attributes to
3279 appear to verify correctly. (CVE-2009-0591)
3280 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3282 *) Reject UniversalString and BMPString types with invalid lengths. This
3283 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3284 a legal length. (CVE-2009-0590)
3287 *) Set S/MIME signing as the default purpose rather than setting it
3288 unconditionally. This allows applications to override it at the store
3292 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3293 to handle some structures.
3296 *) Improve efficiency of mem_gets: don't search whole buffer each time
3298 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3300 *) New -hex option for openssl rand.
3303 *) Print out UTF8String and NumericString when parsing ASN1.
3306 *) Support NumericString type for name components.
3309 *) Allow CC in the environment to override the automatically chosen
3310 compiler. Note that nothing is done to ensure flags work with the
3314 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3316 *) Properly check EVP_VerifyFinal() and similar return values
3318 [Ben Laurie, Bodo Moeller, Google Security Team]
3320 *) Enable TLS extensions by default.
3323 *) Allow the CHIL engine to be loaded, whether the application is
3324 multithreaded or not. (This does not release the developer from the
3325 obligation to set up the dynamic locking callbacks.)
3326 [Sander Temme <sander@temme.net>]
3328 *) Use correct exit code if there is an error in dgst command.
3329 [Steve Henson; problem pointed out by Roland Dirlewanger]
3331 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3332 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3335 *) Add experimental JPAKE support, including demo authentication in
3336 s_client and s_server.
3339 *) Set the comparison function in v3_addr_canonize().
3340 [Rob Austein <sra@hactrn.net>]
3342 *) Add support for XMPP STARTTLS in s_client.
3343 [Philip Paeps <philip@freebsd.org>]
3345 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3346 to ensure that even with this option, only ciphersuites in the
3347 server's preference list will be accepted. (Note that the option
3348 applies only when resuming a session, so the earlier behavior was
3349 just about the algorithm choice for symmetric cryptography.)
3352 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3354 *) Fix NULL pointer dereference if a DTLS server received
3355 ChangeCipherSpec as first record (CVE-2009-1386).
3358 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3359 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3362 *) The fix in 0.9.8c that supposedly got rid of unsafe
3363 double-checked locking was incomplete for RSA blinding,
3364 addressing just one layer of what turns out to have been
3365 doubly unsafe triple-checked locking.
3367 So now fix this for real by retiring the MONT_HELPER macro
3368 in crypto/rsa/rsa_eay.c.
3370 [Bodo Moeller; problem pointed out by Marius Schilder]
3372 *) Various precautionary measures:
3374 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3376 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3377 (NB: This would require knowledge of the secret session ticket key
3378 to exploit, in which case you'd be SOL either way.)
3380 - Change bn_nist.c so that it will properly handle input BIGNUMs
3381 outside the expected range.
3383 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3386 [Neel Mehta, Bodo Moeller]
3388 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3389 the load fails. Useful for distros.
3390 [Ben Laurie and the FreeBSD team]
3392 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3395 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3398 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3400 This work was sponsored by Logica.
3403 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3404 keystores. Support for SSL/TLS client authentication too.
3405 Not compiled unless enable-capieng specified to Configure.
3407 This work was sponsored by Logica.
3410 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3411 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3412 attribute creation routines such as certifcate requests and PKCS#12
3416 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3418 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3419 handshake which could lead to a cilent crash as found using the
3420 Codenomicon TLS test suite (CVE-2008-1672)
3421 [Steve Henson, Mark Cox]
3423 *) Fix double free in TLS server name extensions which could lead to
3424 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3427 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3429 Clear the error queue to ensure that error entries left from
3430 older function calls do not interfere with the correct operation.
3431 [Lutz Jaenicke, Erik de Castro Lopo]
3433 *) Remove root CA certificates of commercial CAs:
3435 The OpenSSL project does not recommend any specific CA and does not
3436 have any policy with respect to including or excluding any CA.
3437 Therefore it does not make any sense to ship an arbitrary selection
3438 of root CA certificates with the OpenSSL software.
3441 *) RSA OAEP patches to fix two separate invalid memory reads.
3442 The first one involves inputs when 'lzero' is greater than
3443 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3444 before the beginning of from). The second one involves inputs where
3445 the 'db' section contains nothing but zeroes (there is a one-byte
3446 invalid read after the end of 'db').
3447 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3449 *) Partial backport from 0.9.9-dev:
3451 Introduce bn_mul_mont (dedicated Montgomery multiplication
3452 procedure) as a candidate for BIGNUM assembler implementation.
3453 While 0.9.9-dev uses assembler for various architectures, only
3454 x86_64 is available by default here in the 0.9.8 branch, and
3455 32-bit x86 is available through a compile-time setting.
3457 To try the 32-bit x86 assembler implementation, use Configure
3458 option "enable-montasm" (which exists only for this backport).
3460 As "enable-montasm" for 32-bit x86 disclaims code stability
3461 anyway, in this constellation we activate additional code
3462 backported from 0.9.9-dev for further performance improvements,
3463 namely BN_from_montgomery_word. (To enable this otherwise,
3464 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3466 [Andy Polyakov (backport partially by Bodo Moeller)]
3468 *) Add TLS session ticket callback. This allows an application to set
3469 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3470 values. This is useful for key rollover for example where several key
3471 sets may exist with different names.
3474 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3475 This was broken until now in 0.9.8 releases, such that the only way
3476 a registered ENGINE could be used (assuming it initialises
3477 successfully on the host) was to explicitly set it as the default
3478 for the relevant algorithms. This is in contradiction with 0.9.7
3479 behaviour and the documentation. With this fix, when an ENGINE is
3480 registered into a given algorithm's table of implementations, the
3481 'uptodate' flag is reset so that auto-discovery will be used next
3482 time a new context for that algorithm attempts to select an
3484 [Ian Lister (tweaked by Geoff Thorpe)]
3486 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3487 implemention in the following ways:
3489 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3492 Lack of BER streaming support means one pass streaming processing is
3493 only supported if data is detached: setting the streaming flag is
3494 ignored for embedded content.
3496 CMS support is disabled by default and must be explicitly enabled
3497 with the enable-cms configuration option.
3500 *) Update the GMP engine glue to do direct copies between BIGNUM and
3501 mpz_t when openssl and GMP use the same limb size. Otherwise the
3502 existing "conversion via a text string export" trick is still used.
3503 [Paul Sheer <paulsheer@gmail.com>]
3505 *) Zlib compression BIO. This is a filter BIO which compressed and
3506 uncompresses any data passed through it.
3509 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3510 RFC3394 compatible AES key wrapping.
3513 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3514 sets string data without copying. X509_ALGOR_set0() and
3515 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3516 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3517 from an X509_ATTRIBUTE structure optionally checking it occurs only
3518 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3522 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3523 to get the expected BN_FLG_CONSTTIME behavior.
3524 [Bodo Moeller (Google)]
3528 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3529 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3530 - added some more tests to do_tests.pl
3531 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3532 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3533 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3534 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3535 - various changes to netware.pl to enable gcc-cross builds on Win32
3537 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3538 - various changes to fix missing prototype warnings
3539 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3540 - added AES, WHIRLPOOL and CPUID assembler code to build files
3541 - added missing AES assembler make rules to mk1mf.pl
3542 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3543 [Guenter Knauf <eflash@gmx.net>]
3545 *) Implement certificate status request TLS extension defined in RFC3546.
3546 A client can set the appropriate parameters and receive the encoded
3547 OCSP response via a callback. A server can query the supplied parameters
3548 and set the encoded OCSP response in the callback. Add simplified examples
3549 to s_client and s_server.
3552 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3554 *) Fix various bugs:
3555 + Binary incompatibility of ssl_ctx_st structure
3556 + DTLS interoperation with non-compliant servers
3557 + Don't call get_session_cb() without proposed session
3558 + Fix ia64 assembler code
3559 [Andy Polyakov, Steve Henson]
3561 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3563 *) DTLS Handshake overhaul. There were longstanding issues with
3564 OpenSSL DTLS implementation, which were making it impossible for
3565 RFC 4347 compliant client to communicate with OpenSSL server.
3566 Unfortunately just fixing these incompatibilities would "cut off"
3567 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3568 server keeps tolerating non RFC compliant syntax. The opposite is
3569 not true, 0.9.8f client can not communicate with earlier server.
3570 This update even addresses CVE-2007-4995.
3573 *) Changes to avoid need for function casts in OpenSSL: some compilers
3574 (gcc 4.2 and later) reject their use.
3575 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3578 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3579 RFC4507bis. The encrypted ticket format is an encrypted encoded
3580 SSL_SESSION structure, that way new session features are automatically
3583 If a client application caches session in an SSL_SESSION structure
3584 support is transparent because tickets are now stored in the encoded
3587 The SSL_CTX structure automatically generates keys for ticket
3588 protection in servers so again support should be possible
3589 with no application modification.
3591 If a client or server wishes to disable RFC4507 support then the option
3592 SSL_OP_NO_TICKET can be set.
3594 Add a TLS extension debugging callback to allow the contents of any client
3595 or server extensions to be examined.
3597 This work was sponsored by Google.
3600 *) Add initial support for TLS extensions, specifically for the server_name
3601 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3602 have new members for a host name. The SSL data structure has an
3603 additional member SSL_CTX *initial_ctx so that new sessions can be
3604 stored in that context to allow for session resumption, even after the
3605 SSL has been switched to a new SSL_CTX in reaction to a client's
3606 server_name extension.
3608 New functions (subject to change):
3610 SSL_get_servername()
3611 SSL_get_servername_type()
3614 New CTRL codes and macros (subject to change):
3616 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3617 - SSL_CTX_set_tlsext_servername_callback()
3618 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3619 - SSL_CTX_set_tlsext_servername_arg()
3620 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3622 openssl s_client has a new '-servername ...' option.
3624 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3625 '-key2 ...', '-servername_fatal' (subject to change). This allows
3626 testing the HostName extension for a specific single host name ('-cert'
3627 and '-key' remain fallbacks for handshakes without HostName
3628 negotiation). If the unrecognized_name alert has to be sent, this by
3629 default is a warning; it becomes fatal with the '-servername_fatal'
3632 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3634 *) Add AES and SSE2 assembly language support to VC++ build.
3637 *) Mitigate attack on final subtraction in Montgomery reduction.
3640 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3641 (which previously caused an internal error).
3644 *) Squeeze another 10% out of IGE mode when in != out.
3647 *) AES IGE mode speedup.
3648 [Dean Gaudet (Google)]
3650 *) Add the Korean symmetric 128-bit cipher SEED (see
3651 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3652 add SEED ciphersuites from RFC 4162:
3654 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3655 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3656 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3657 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3659 To minimize changes between patchlevels in the OpenSSL 0.9.8
3660 series, SEED remains excluded from compilation unless OpenSSL
3661 is configured with 'enable-seed'.
3662 [KISA, Bodo Moeller]
3664 *) Mitigate branch prediction attacks, which can be practical if a
3665 single processor is shared, allowing a spy process to extract
3666 information. For detailed background information, see
3667 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3668 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3669 and Necessary Software Countermeasures"). The core of the change
3670 are new versions BN_div_no_branch() and
3671 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3672 respectively, which are slower, but avoid the security-relevant
3673 conditional branches. These are automatically called by BN_div()
3674 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3675 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3676 remove a conditional branch.
3678 BN_FLG_CONSTTIME is the new name for the previous
3679 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3680 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3681 in the exponent causes BN_mod_exp_mont() to use the alternative
3682 implementation in BN_mod_exp_mont_consttime().) The old name
3683 remains as a deprecated alias.
3685 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3686 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3687 constant-time implementations for more than just exponentiation.
3688 Here too the old name is kept as a deprecated alias.
3690 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3691 the BN_BLINDING structure gets an independent copy of the
3692 modulus. This means that the previous "BIGNUM *m" argument to
3693 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3694 essentially becomes "const BIGNUM *m", although we can't actually
3695 change this in the header file before 0.9.9. It allows
3696 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3697 enable BN_FLG_CONSTTIME.
3699 [Matthew D Wood (Intel Corp)]
3701 *) In the SSL/TLS server implementation, be strict about session ID
3702 context matching (which matters if an application uses a single
3703 external cache for different purposes). Previously,
3704 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3705 set. This did ensure strict client verification, but meant that,
3706 with applications using a single external cache for quite
3707 different requirements, clients could circumvent ciphersuite
3708 restrictions for a given session ID context by starting a session
3709 in a different context.
3712 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3713 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3714 authentication-only ciphersuites.
3717 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3718 not complete and could lead to a possible single byte overflow
3719 (CVE-2007-5135) [Ben Laurie]
3721 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3723 *) Since AES128 and AES256 (and similarly Camellia128 and
3724 Camellia256) share a single mask bit in the logic of
3725 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3726 kludge to work properly if AES128 is available and AES256 isn't
3727 (or if Camellia128 is available and Camellia256 isn't).
3730 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3731 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3732 When a point or a seed is encoded in a BIT STRING, we need to
3733 prevent the removal of trailing zero bits to get the proper DER
3734 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3735 of a NamedBitList, for which trailing 0 bits need to be removed.)
3738 *) Have SSL/TLS server implementation tolerate "mismatched" record
3739 protocol version while receiving ClientHello even if the
3740 ClientHello is fragmented. (The server can't insist on the
3741 particular protocol version it has chosen before the ServerHello
3742 message has informed the client about his choice.)
3745 *) Add RFC 3779 support.
3746 [Rob Austein for ARIN, Ben Laurie]
3748 *) Load error codes if they are not already present instead of using a
3749 static variable. This allows them to be cleanly unloaded and reloaded.
3750 Improve header file function name parsing.
3753 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3754 or CAPABILITY handshake as required by RFCs.
3757 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3759 *) Introduce limits to prevent malicious keys being able to
3760 cause a denial of service. (CVE-2006-2940)
3761 [Steve Henson, Bodo Moeller]
3763 *) Fix ASN.1 parsing of certain invalid structures that can result
3764 in a denial of service. (CVE-2006-2937) [Steve Henson]
3766 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3767 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3769 *) Fix SSL client code which could crash if connecting to a
3770 malicious SSLv2 server. (CVE-2006-4343)
3771 [Tavis Ormandy and Will Drewry, Google Security Team]
3773 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3774 match only those. Before that, "AES256-SHA" would be interpreted
3775 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3776 the same strength classification in 0.9.7h) as we currently only
3777 have a single AES bit in the ciphersuite description bitmap.
3778 That change, however, also applied to ciphersuite strings such as
3779 "RC4-MD5" that intentionally matched multiple ciphersuites --
3780 namely, SSL 2.0 ciphersuites in addition to the more common ones
3781 from SSL 3.0/TLS 1.0.
3783 So we change the selection algorithm again: Naming an explicit
3784 ciphersuite selects this one ciphersuite, and any other similar
3785 ciphersuite (same bitmap) from *other* protocol versions.
3786 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3787 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3789 Since SSL 2.0 does not have any ciphersuites for which the
3790 128/256 bit distinction would be relevant, this works for now.
3791 The proper fix will be to use different bits for AES128 and
3792 AES256, which would have avoided the problems from the beginning;
3793 however, bits are scarce, so we can only do this in a new release
3794 (not just a patchlevel) when we can change the SSL_CIPHER
3795 definition to split the single 'unsigned long mask' bitmap into
3796 multiple values to extend the available space.
3800 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3802 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3803 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3805 *) Add AES IGE and biIGE modes.
3808 *) Change the Unix randomness entropy gathering to use poll() when
3809 possible instead of select(), since the latter has some
3810 undesirable limitations.
3811 [Darryl Miles via Richard Levitte and Bodo Moeller]
3813 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3814 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3815 cannot be implicitly activated as part of, e.g., the "AES" alias.
3816 However, please upgrade to OpenSSL 0.9.9[-dev] for
3817 non-experimental use of the ECC ciphersuites to get TLS extension
3818 support, which is required for curve and point format negotiation
3819 to avoid potential handshake problems.
3822 *) Disable rogue ciphersuites:
3824 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3825 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3826 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3828 The latter two were purportedly from
3829 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3832 Also deactivate the remaining ciphersuites from
3833 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3834 unofficial, and the ID has long expired.
3837 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3838 dual-core machines) and other potential thread-safety issues.
3841 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3842 versions), which is now available for royalty-free use
3843 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3844 Also, add Camellia TLS ciphersuites from RFC 4132.
3846 To minimize changes between patchlevels in the OpenSSL 0.9.8
3847 series, Camellia remains excluded from compilation unless OpenSSL
3848 is configured with 'enable-camellia'.
3851 *) Disable the padding bug check when compression is in use. The padding
3852 bug check assumes the first packet is of even length, this is not
3853 necessarily true if compresssion is enabled and can result in false
3854 positives causing handshake failure. The actual bug test is ancient
3855 code so it is hoped that implementations will either have fixed it by
3856 now or any which still have the bug do not support compression.
3859 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3861 *) When applying a cipher rule check to see if string match is an explicit
3862 cipher suite and only match that one cipher suite if it is.
3865 *) Link in manifests for VC++ if needed.
3866 [Austin Ziegler <halostatue@gmail.com>]
3868 *) Update support for ECC-based TLS ciphersuites according to
3869 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3870 TLS extensions, which are supported starting with the 0.9.9
3871 branch, not in the OpenSSL 0.9.8 branch).
3874 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3875 opaque EVP_CIPHER_CTX handling.
3878 *) Fixes and enhancements to zlib compression code. We now only use
3879 "zlib1.dll" and use the default __cdecl calling convention on Win32
3880 to conform with the standards mentioned here:
3881 http://www.zlib.net/DLL_FAQ.txt
3882 Static zlib linking now works on Windows and the new --with-zlib-include
3883 --with-zlib-lib options to Configure can be used to supply the location
3884 of the headers and library. Gracefully handle case where zlib library
3888 *) Several fixes and enhancements to the OID generation code. The old code
3889 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3890 handle numbers larger than ULONG_MAX, truncated printing and had a
3891 non standard OBJ_obj2txt() behaviour.
3894 *) Add support for building of engines under engine/ as shared libraries
3895 under VC++ build system.
3898 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3899 Hopefully, we will not see any false combination of paths any more.
3902 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3904 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3905 (part of SSL_OP_ALL). This option used to disable the
3906 countermeasure against man-in-the-middle protocol-version
3907 rollback in the SSL 2.0 server implementation, which is a bad
3908 idea. (CVE-2005-2969)
3910 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3911 for Information Security, National Institute of Advanced Industrial
3912 Science and Technology [AIST], Japan)]
3914 *) Add two function to clear and return the verify parameter flags.
3917 *) Keep cipherlists sorted in the source instead of sorting them at
3918 runtime, thus removing the need for a lock.
3921 *) Avoid some small subgroup attacks in Diffie-Hellman.
3922 [Nick Mathewson and Ben Laurie]
3924 *) Add functions for well-known primes.
3927 *) Extended Windows CE support.
3928 [Satoshi Nakamura and Andy Polyakov]
3930 *) Initialize SSL_METHOD structures at compile time instead of during
3931 runtime, thus removing the need for a lock.
3934 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3935 attempting to decrypt each encrypted key in turn. Add support to
3939 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3941 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3944 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3947 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3948 key into the same file any more.
3951 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3954 *) Add -utf8 command line and config file option to 'ca'.
3955 [Stefan <stf@udoma.org]
3957 *) Removed the macro des_crypt(), as it seems to conflict with some
3958 libraries. Use DES_crypt().
3961 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3962 involves renaming the source and generated shared-libs for
3963 both. The engines will accept the corrected or legacy ids
3964 ('ncipher' and '4758_cca' respectively) when binding. NB,
3965 this only applies when building 'shared'.
3966 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3968 *) Add attribute functions to EVP_PKEY structure. Modify
3969 PKCS12_create() to recognize a CSP name attribute and
3970 use it. Make -CSP option work again in pkcs12 utility.
3973 *) Add new functionality to the bn blinding code:
3974 - automatic re-creation of the BN_BLINDING parameters after
3975 a fixed number of uses (currently 32)
3976 - add new function for parameter creation
3977 - introduce flags to control the update behaviour of the
3978 BN_BLINDING parameters
3979 - hide BN_BLINDING structure
3980 Add a second BN_BLINDING slot to the RSA structure to improve
3981 performance when a single RSA object is shared among several
3985 *) Add support for DTLS.
3986 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3988 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3989 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3992 *) Remove buggy and incomplete DH cert support from
3993 ssl/ssl_rsa.c and ssl/s3_both.c
3996 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3997 the apps/openssl applications.
4000 *) Compile clean with "-Wall -Wmissing-prototypes
4001 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4002 DEBUG_SAFESTACK must also be set.
4005 *) Change ./Configure so that certain algorithms can be disabled by default.
4006 The new counterpiece to "no-xxx" is "enable-xxx".
4008 The patented RC5 and MDC2 algorithms will now be disabled unless
4009 "enable-rc5" and "enable-mdc2", respectively, are specified.
4011 (IDEA remains enabled despite being patented. This is because IDEA
4012 is frequently required for interoperability, and there is no license
4013 fee for non-commercial use. As before, "no-idea" can be used to
4014 avoid this algorithm.)
4018 *) Add processing of proxy certificates (see RFC 3820). This work was
4019 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4020 EGEE (Enabling Grids for E-science in Europe).
4023 *) RC4 performance overhaul on modern architectures/implementations, such
4024 as Intel P4, IA-64 and AMD64.
4027 *) New utility extract-section.pl. This can be used specify an alternative
4028 section number in a pod file instead of having to treat each file as
4029 a separate case in Makefile. This can be done by adding two lines to the
4032 =for comment openssl_section:XXX
4034 The blank line is mandatory.
4038 *) New arguments -certform, -keyform and -pass for s_client and s_server
4039 to allow alternative format key and certificate files and passphrase
4043 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4044 update associated structures and add various utility functions.
4046 Add new policy related verify parameters, include policy checking in
4047 standard verify code. Enhance 'smime' application with extra parameters
4048 to support policy checking and print out.
4051 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4052 Nehemiah processors. These extensions support AES encryption in hardware
4053 as well as RNG (though RNG support is currently disabled).
4054 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4056 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4059 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4060 [Andy Polyakov and a number of other people]
4062 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4063 implementation contributed by IBM.
4064 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4066 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4067 exponent rather than 'unsigned long'. There is a corresponding change to
4068 the new 'rsa_keygen' element of the RSA_METHOD structure.
4069 [Jelte Jansen, Geoff Thorpe]
4071 *) Functionality for creating the initial serial number file is now
4072 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4074 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4075 number file to 1, which is bound to cause problems. To avoid
4076 the problems while respecting compatibility between different 0.9.7
4077 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4078 CA.pl for serial number initialization. With the new release 0.9.8,
4079 we can fix the problem directly in the 'ca' utility.)
4082 *) Reduced header interdepencies by declaring more opaque objects in
4083 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4084 give fewer recursive includes, which could break lazy source code - so
4085 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4086 developers should define this symbol when building and using openssl to
4087 ensure they track the recommended behaviour, interfaces, [etc], but
4088 backwards-compatible behaviour prevails when this isn't defined.
4091 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4094 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4095 This will generate a random key of the appropriate length based on the
4096 cipher context. The EVP_CIPHER can provide its own random key generation
4097 routine to support keys of a specific form. This is used in the des and
4098 3des routines to generate a key of the correct parity. Update S/MIME
4099 code to use new functions and hence generate correct parity DES keys.
4100 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4101 valid (weak or incorrect parity).
4104 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4105 as looking them up. This is useful when the verified structure may contain
4106 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4107 present unless the new PKCS7_NO_CRL flag is asserted.
4110 *) Extend ASN1 oid configuration module. It now additionally accepts the
4113 shortName = some long name, 1.2.3.4
4116 *) Reimplemented the BN_CTX implementation. There is now no more static
4117 limitation on the number of variables it can handle nor the depth of the
4118 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4119 information can now expand as required, and rather than having a single
4120 static array of bignums, BN_CTX now uses a linked-list of such arrays
4121 allowing it to expand on demand whilst maintaining the usefulness of
4122 BN_CTX's "bundling".
4125 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4126 to allow all RSA operations to function using a single BN_CTX.
4129 *) Preliminary support for certificate policy evaluation and checking. This
4130 is initially intended to pass the tests outlined in "Conformance Testing
4131 of Relying Party Client Certificate Path Processing Logic" v1.07.
4134 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4135 remained unused and not that useful. A variety of other little bignum
4136 tweaks and fixes have also been made continuing on from the audit (see
4140 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4141 associated ASN1, EVP and SSL functions and old ASN1 macros.
4144 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4145 and this should never fail. So the return value from the use of
4146 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4147 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4150 *) BN_CTX_get() should return zero-valued bignums, providing the same
4151 initialised value as BN_new().
4152 [Geoff Thorpe, suggested by Ulf Möller]
4154 *) Support for inhibitAnyPolicy certificate extension.
4157 *) An audit of the BIGNUM code is underway, for which debugging code is
4158 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4159 is considered valid when processing BIGNUMs, and causes execution to
4160 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4161 further steps are taken to deliberately pollute unused data in BIGNUM
4162 structures to try and expose faulty code further on. For now, openssl will
4163 (in its default mode of operation) continue to tolerate the inconsistent
4164 forms that it has tolerated in the past, but authors and packagers should
4165 consider trying openssl and their own applications when compiled with
4166 these debugging symbols defined. It will help highlight potential bugs in
4167 their own code, and will improve the test coverage for OpenSSL itself. At
4168 some point, these tighter rules will become openssl's default to improve
4169 maintainability, though the assert()s and other overheads will remain only
4170 in debugging configurations. See bn.h for more details.
4171 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4173 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4174 that can only be obtained through BN_CTX_new() (which implicitly
4175 initialises it). The presence of this function only made it possible
4176 to overwrite an existing structure (and cause memory leaks).
4179 *) Because of the callback-based approach for implementing LHASH as a
4180 template type, lh_insert() adds opaque objects to hash-tables and
4181 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4182 to clean up those corresponding objects before destroying the hash table
4183 (and losing the object pointers). So some over-zealous constifications in
4184 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4185 objects as "const" and the lh_doall[_arg] callback wrappers are not
4186 prototyped to have "const" restrictions on the object pointers they are
4187 given (and so aren't required to cast them away any more).
4190 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4191 (speed) prefers to use its own implementation. The two implementations
4192 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4193 its object type properly exposed (MS_TM) instead of casting to/from "char
4194 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4195 aren't necessarily the greatest nomenclatures - but this is what was used
4196 internally to the implementation so I've used that for now.
4199 *) Ensure that deprecated functions do not get compiled when
4200 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4201 the self-tests were still using deprecated key-generation functions so
4202 these have been updated also.
4205 *) Reorganise PKCS#7 code to separate the digest location functionality
4206 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4207 New function PKCS7_set_digest() to set the digest type for PKCS#7
4208 digestedData type. Add additional code to correctly generate the
4209 digestedData type and add support for this type in PKCS7 initialization
4213 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4214 structure of type "other".
4217 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4218 sure the loop does correctly stop and breaking ("division by zero")
4219 modulus operations are not performed. The (pre-generated) prime
4220 table crypto/bn/bn_prime.h was already correct, but it could not be
4221 re-generated on some platforms because of the "division by zero"
4222 situation in the script.
4223 [Ralf S. Engelschall]
4225 *) Update support for ECC-based TLS ciphersuites according to
4226 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4227 SHA-1 now is only used for "small" curves (where the
4228 representation of a field element takes up to 24 bytes); for
4229 larger curves, the field element resulting from ECDH is directly
4230 used as premaster secret.
4231 [Douglas Stebila (Sun Microsystems Laboratories)]
4233 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4234 curve secp160r1 to the tests.
4235 [Douglas Stebila (Sun Microsystems Laboratories)]
4237 *) Add the possibility to load symbols globally with DSO.
4238 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4240 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4241 control of the error stack.
4244 *) Add support for STORE in ENGINE.
4247 *) Add the STORE type. The intention is to provide a common interface
4248 to certificate and key stores, be they simple file-based stores, or
4249 HSM-type store, or LDAP stores, or...
4250 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4253 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4254 pass a list of arguments to any function as well as provide a way
4255 for a function to pass data back to the caller.
4258 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4259 works like BUF_strdup() but can be used to duplicate a portion of
4260 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4264 *) Add the function sk_find_ex() which works like sk_find(), but will
4265 return an index to an element even if an exact match couldn't be
4266 found. The index is guaranteed to point at the element where the
4267 searched-for key would be inserted to preserve sorting order.
4270 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4271 takes an extra flags argument for optional functionality. Currently,
4272 the following flags are defined:
4274 OBJ_BSEARCH_VALUE_ON_NOMATCH
4275 This one gets OBJ_bsearch_ex() to return a pointer to the first
4276 element where the comparing function returns a negative or zero
4279 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4280 This one gets OBJ_bsearch_ex() to return a pointer to the first
4281 element where the comparing function returns zero. This is useful
4282 if there are more than one element where the comparing function
4286 *) Make it possible to create self-signed certificates with 'openssl ca'
4287 in such a way that the self-signed certificate becomes part of the
4288 CA database and uses the same mechanisms for serial number generation
4289 as all other certificate signing. The new flag '-selfsign' enables
4290 this functionality. Adapt CA.sh and CA.pl.in.
4293 *) Add functionality to check the public key of a certificate request
4294 against a given private. This is useful to check that a certificate
4295 request can be signed by that key (self-signing).
4298 *) Make it possible to have multiple active certificates with the same
4299 subject in the CA index file. This is done only if the keyword
4300 'unique_subject' is set to 'no' in the main CA section (default
4301 if 'CA_default') of the configuration file. The value is saved
4302 with the database itself in a separate index attribute file,
4303 named like the index file with '.attr' appended to the name.
4306 *) Generate muti valued AVAs using '+' notation in config files for
4310 *) Support for nameConstraints certificate extension.
4313 *) Support for policyConstraints certificate extension.
4316 *) Support for policyMappings certificate extension.
4319 *) Make sure the default DSA_METHOD implementation only uses its
4320 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4321 and change its own handlers to be NULL so as to remove unnecessary
4322 indirection. This lets alternative implementations fallback to the
4323 default implementation more easily.
4326 *) Support for directoryName in GeneralName related extensions
4330 *) Make it possible to link applications using Makefile.shared.
4331 Make that possible even when linking against static libraries!
4334 *) Support for single pass processing for S/MIME signing. This now
4335 means that S/MIME signing can be done from a pipe, in addition
4336 cleartext signing (multipart/signed type) is effectively streaming
4337 and the signed data does not need to be all held in memory.
4339 This is done with a new flag PKCS7_STREAM. When this flag is set
4340 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4341 is done after the data is output (and digests calculated) in
4342 SMIME_write_PKCS7().
4345 *) Add full support for -rpath/-R, both in shared libraries and
4346 applications, at least on the platforms where it's known how
4350 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4351 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4352 will now compute a table of multiples of the generator that
4353 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4354 faster (notably in the case of a single point multiplication,
4355 scalar * generator).
4356 [Nils Larsch, Bodo Moeller]
4358 *) IPv6 support for certificate extensions. The various extensions
4359 which use the IP:a.b.c.d can now take IPv6 addresses using the
4360 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4364 *) Added an ENGINE that implements RSA by performing private key
4365 exponentiations with the GMP library. The conversions to and from
4366 GMP's mpz_t format aren't optimised nor are any montgomery forms
4367 cached, and on x86 it appears OpenSSL's own performance has caught up.
4368 However there are likely to be other architectures where GMP could
4369 provide a boost. This ENGINE is not built in by default, but it can be
4370 specified at Configure time and should be accompanied by the necessary
4371 linker additions, eg;
4372 ./config -DOPENSSL_USE_GMP -lgmp
4375 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4376 testing availability of engines with "-t" - the old behaviour is
4377 produced by increasing the feature's verbosity with "-tt".
4380 *) ECDSA routines: under certain error conditions uninitialized BN objects
4381 could be freed. Solution: make sure initialization is performed early
4382 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4386 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4387 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4388 software implementations. For DSA and DH, parameter generation can
4389 also be overriden by providing the appropriate method callbacks.
4392 *) Change the "progress" mechanism used in key-generation and
4393 primality testing to functions that take a new BN_GENCB pointer in
4394 place of callback/argument pairs. The new API functions have "_ex"
4395 postfixes and the older functions are reimplemented as wrappers for
4396 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4397 declarations of the old functions to help (graceful) attempts to
4398 migrate to the new functions. Also, the new key-generation API
4399 functions operate on a caller-supplied key-structure and return
4400 success/failure rather than returning a key or NULL - this is to
4401 help make "keygen" another member function of RSA_METHOD etc.
4403 Example for using the new callback interface:
4405 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4409 BN_GENCB_set(&my_cb, my_callback, my_arg);
4411 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4412 /* For the meaning of a, b in calls to my_callback(), see the
4413 * documentation of the function that calls the callback.
4414 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4415 * my_callback should return 1 if it wants BN_is_prime_ex()
4416 * to continue, or 0 to stop.
4421 *) Change the ZLIB compression method to be stateful, and make it
4422 available to TLS with the number defined in
4423 draft-ietf-tls-compression-04.txt.
4426 *) Add the ASN.1 structures and functions for CertificatePair, which
4427 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4429 CertificatePair ::= SEQUENCE {
4430 forward [0] Certificate OPTIONAL,
4431 reverse [1] Certificate OPTIONAL,
4432 -- at least one of the pair shall be present -- }
4434 Also implement the PEM functions to read and write certificate
4435 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4437 This needed to be defined, mostly for the sake of the LDAP
4438 attribute crossCertificatePair, but may prove useful elsewhere as
4442 *) Make it possible to inhibit symlinking of shared libraries in
4443 Makefile.shared, for Cygwin's sake.
4446 *) Extend the BIGNUM API by creating a function
4447 void BN_set_negative(BIGNUM *a, int neg);
4448 and a macro that behave like
4449 int BN_is_negative(const BIGNUM *a);
4451 to avoid the need to access 'a->neg' directly in applications.
4454 *) Implement fast modular reduction for pseudo-Mersenne primes
4455 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4456 EC_GROUP_new_curve_GFp() will now automatically use this
4458 [Nils Larsch <nla@trustcenter.de>]
4460 *) Add new lock type (CRYPTO_LOCK_BN).
4463 *) Change the ENGINE framework to automatically load engines
4464 dynamically from specific directories unless they could be
4465 found to already be built in or loaded. Move all the
4466 current engines except for the cryptodev one to a new
4468 The engines in engines/ are built as shared libraries if
4469 the "shared" options was given to ./Configure or ./config.
4470 Otherwise, they are inserted in libcrypto.a.
4471 /usr/local/ssl/engines is the default directory for dynamic
4472 engines, but that can be overriden at configure time through
4473 the usual use of --prefix and/or --openssldir, and at run
4474 time with the environment variable OPENSSL_ENGINES.
4475 [Geoff Thorpe and Richard Levitte]
4477 *) Add Makefile.shared, a helper makefile to build shared
4478 libraries. Adapt Makefile.org.
4481 *) Add version info to Win32 DLLs.
4482 [Peter 'Luna' Runestig" <peter@runestig.com>]
4484 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4485 can be added using this API to created arbitrary PKCS#12
4486 files while avoiding the low level API.
4488 New options to PKCS12_create(), key or cert can be NULL and
4489 will then be omitted from the output file. The encryption
4490 algorithm NIDs can be set to -1 for no encryption, the mac
4491 iteration count can be set to 0 to omit the mac.
4493 Enhance pkcs12 utility by making the -nokeys and -nocerts
4494 options work when creating a PKCS#12 file. New option -nomac
4495 to omit the mac, NONE can be set for an encryption algorithm.
4496 New code is modified to use the enhanced PKCS12_create()
4497 instead of the low level API.
4500 *) Extend ASN1 encoder to support indefinite length constructed
4501 encoding. This can output sequences tags and octet strings in
4502 this form. Modify pk7_asn1.c to support indefinite length
4503 encoding. This is experimental and needs additional code to
4504 be useful, such as an ASN1 bio and some enhanced streaming
4507 Extend template encode functionality so that tagging is passed
4508 down to the template encoder.
4511 *) Let 'openssl req' fail if an argument to '-newkey' is not
4512 recognized instead of using RSA as a default.
4515 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4516 As these are not official, they are not included in "ALL";
4517 the "ECCdraft" ciphersuite group alias can be used to select them.
4518 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4520 *) Add ECDH engine support.
4521 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4523 *) Add ECDH in new directory crypto/ecdh/.
4524 [Douglas Stebila (Sun Microsystems Laboratories)]
4526 *) Let BN_rand_range() abort with an error after 100 iterations
4527 without success (which indicates a broken PRNG).
4530 *) Change BN_mod_sqrt() so that it verifies that the input value
4531 is really the square of the return value. (Previously,
4532 BN_mod_sqrt would show GIGO behaviour.)
4535 *) Add named elliptic curves over binary fields from X9.62, SECG,
4536 and WAP/WTLS; add OIDs that were still missing.
4538 [Sheueling Chang Shantz and Douglas Stebila
4539 (Sun Microsystems Laboratories)]
4541 *) Extend the EC library for elliptic curves over binary fields
4542 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4545 EC_GF2m_simple_method
4549 EC_GROUP_new_curve_GF2m
4550 EC_GROUP_set_curve_GF2m
4551 EC_GROUP_get_curve_GF2m
4552 EC_POINT_set_affine_coordinates_GF2m
4553 EC_POINT_get_affine_coordinates_GF2m
4554 EC_POINT_set_compressed_coordinates_GF2m
4556 Point compression for binary fields is disabled by default for
4557 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4560 As binary polynomials are represented as BIGNUMs, various members
4561 of the EC_GROUP and EC_POINT data structures can be shared
4562 between the implementations for prime fields and binary fields;
4563 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4564 are essentially identical to their ..._GFp counterparts.
4565 (For simplicity, the '..._GFp' prefix has been dropped from
4566 various internal method names.)
4568 An internal 'field_div' method (similar to 'field_mul' and
4569 'field_sqr') has been added; this is used only for binary fields.
4571 [Sheueling Chang Shantz and Douglas Stebila
4572 (Sun Microsystems Laboratories)]
4574 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4575 through methods ('mul', 'precompute_mult').
4577 The generic implementations (now internally called 'ec_wNAF_mul'
4578 and 'ec_wNAF_precomputed_mult') remain the default if these
4579 methods are undefined.
4581 [Sheueling Chang Shantz and Douglas Stebila
4582 (Sun Microsystems Laboratories)]
4584 *) New function EC_GROUP_get_degree, which is defined through
4585 EC_METHOD. For curves over prime fields, this returns the bit
4586 length of the modulus.
4588 [Sheueling Chang Shantz and Douglas Stebila
4589 (Sun Microsystems Laboratories)]
4591 *) New functions EC_GROUP_dup, EC_POINT_dup.
4592 (These simply call ..._new and ..._copy).
4594 [Sheueling Chang Shantz and Douglas Stebila
4595 (Sun Microsystems Laboratories)]
4597 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4598 Polynomials are represented as BIGNUMs (where the sign bit is not
4599 used) in the following functions [macros]:
4602 BN_GF2m_sub [= BN_GF2m_add]
4603 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4604 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4605 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4607 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4608 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4609 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4610 BN_GF2m_cmp [= BN_ucmp]
4612 (Note that only the 'mod' functions are actually for fields GF(2^m).
4613 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4615 For some functions, an the irreducible polynomial defining a
4616 field can be given as an 'unsigned int[]' with strictly
4617 decreasing elements giving the indices of those bits that are set;
4618 i.e., p[] represents the polynomial
4619 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4621 p[0] > p[1] > ... > p[k] = 0.
4622 This applies to the following functions:
4627 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4628 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4630 BN_GF2m_mod_sqrt_arr
4631 BN_GF2m_mod_solve_quad_arr
4635 Conversion can be performed by the following functions:
4640 bntest.c has additional tests for binary polynomial arithmetic.
4642 Two implementations for BN_GF2m_mod_div() are available.
4643 The default algorithm simply uses BN_GF2m_mod_inv() and
4644 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4645 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4646 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4648 [Sheueling Chang Shantz and Douglas Stebila
4649 (Sun Microsystems Laboratories)]
4651 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4652 functionality is disabled at compile-time.
4653 [Douglas Stebila <douglas.stebila@sun.com>]
4655 *) Change default behaviour of 'openssl asn1parse' so that more
4656 information is visible when viewing, e.g., a certificate:
4658 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4659 mode the content of non-printable OCTET STRINGs is output in a
4660 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4661 avoid the appearance of a printable string.
4662 [Nils Larsch <nla@trustcenter.de>]
4664 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4666 EC_GROUP_set_asn1_flag()
4667 EC_GROUP_get_asn1_flag()
4668 EC_GROUP_set_point_conversion_form()
4669 EC_GROUP_get_point_conversion_form()
4670 These control ASN1 encoding details:
4671 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4672 has been set to OPENSSL_EC_NAMED_CURVE.
4673 - Points are encoded in uncompressed form by default; options for
4674 asn1_for are as for point2oct, namely
4675 POINT_CONVERSION_COMPRESSED
4676 POINT_CONVERSION_UNCOMPRESSED
4677 POINT_CONVERSION_HYBRID
4679 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4682 EC_GROUP_get0_seed()
4683 EC_GROUP_get_seed_len()
4684 This is used only for ASN1 purposes (so far).
4685 [Nils Larsch <nla@trustcenter.de>]
4687 *) Add 'field_type' member to EC_METHOD, which holds the NID
4688 of the appropriate field type OID. The new function
4689 EC_METHOD_get_field_type() returns this value.
4690 [Nils Larsch <nla@trustcenter.de>]
4695 EC_POINT_point2hex()
4696 EC_POINT_hex2point()
4697 providing useful interfaces to EC_POINT_point2oct() and
4698 EC_POINT_oct2point().
4699 [Nils Larsch <nla@trustcenter.de>]
4701 *) Change internals of the EC library so that the functions
4702 EC_GROUP_set_generator()
4703 EC_GROUP_get_generator()
4704 EC_GROUP_get_order()
4705 EC_GROUP_get_cofactor()
4706 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4707 to methods, which would lead to unnecessary code duplication when
4708 adding different types of curves.
4709 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4711 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4712 arithmetic, and such that modified wNAFs are generated
4713 (which avoid length expansion in many cases).
4716 *) Add a function EC_GROUP_check_discriminant() (defined via
4717 EC_METHOD) that verifies that the curve discriminant is non-zero.
4719 Add a function EC_GROUP_check() that makes some sanity tests
4720 on a EC_GROUP, its generator and order. This includes
4721 EC_GROUP_check_discriminant().
4722 [Nils Larsch <nla@trustcenter.de>]
4724 *) Add ECDSA in new directory crypto/ecdsa/.
4726 Add applications 'openssl ecparam' and 'openssl ecdsa'
4727 (these are based on 'openssl dsaparam' and 'openssl dsa').
4729 ECDSA support is also included in various other files across the
4730 library. Most notably,
4731 - 'openssl req' now has a '-newkey ecdsa:file' option;
4732 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4733 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4734 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4735 them suitable for ECDSA where domain parameters must be
4736 extracted before the specific public key;
4737 - ECDSA engine support has been added.
4738 [Nils Larsch <nla@trustcenter.de>]
4740 *) Include some named elliptic curves, and add OIDs from X9.62,
4741 SECG, and WAP/WTLS. Each curve can be obtained from the new
4743 EC_GROUP_new_by_curve_name(),
4744 and the list of available named curves can be obtained with
4745 EC_get_builtin_curves().
4746 Also add a 'curve_name' member to EC_GROUP objects, which can be
4748 EC_GROUP_set_curve_name()
4749 EC_GROUP_get_curve_name()
4750 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4752 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4753 was actually never needed) and in BN_mul(). The removal in BN_mul()
4754 required a small change in bn_mul_part_recursive() and the addition
4755 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4756 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4757 bn_sub_words() and bn_add_words() except they take arrays with
4761 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4763 *) Cleanse PEM buffers before freeing them since they may contain
4765 [Benjamin Bennett <ben@psc.edu>]
4767 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4768 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4769 authentication-only ciphersuites.
4772 *) Since AES128 and AES256 share a single mask bit in the logic of
4773 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4774 kludge to work properly if AES128 is available and AES256 isn't.
4777 *) Expand security boundary to match 1.1.1 module.
4780 *) Remove redundant features: hash file source, editing of test vectors
4781 modify fipsld to use external fips_premain.c signature.
4784 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4785 run algorithm test programs.
4788 *) Make algorithm test programs more tolerant of whitespace.
4791 *) Have SSL/TLS server implementation tolerate "mismatched" record
4792 protocol version while receiving ClientHello even if the
4793 ClientHello is fragmented. (The server can't insist on the
4794 particular protocol version it has chosen before the ServerHello
4795 message has informed the client about his choice.)
4798 *) Load error codes if they are not already present instead of using a
4799 static variable. This allows them to be cleanly unloaded and reloaded.
4802 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4804 *) Introduce limits to prevent malicious keys being able to
4805 cause a denial of service. (CVE-2006-2940)
4806 [Steve Henson, Bodo Moeller]
4808 *) Fix ASN.1 parsing of certain invalid structures that can result
4809 in a denial of service. (CVE-2006-2937) [Steve Henson]
4811 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4812 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4814 *) Fix SSL client code which could crash if connecting to a
4815 malicious SSLv2 server. (CVE-2006-4343)
4816 [Tavis Ormandy and Will Drewry, Google Security Team]
4818 *) Change ciphersuite string processing so that an explicit
4819 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4820 will no longer include "AES128-SHA"), and any other similar
4821 ciphersuite (same bitmap) from *other* protocol versions (so that
4822 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4823 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4824 changes from 0.9.8b and 0.9.8d.
4827 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4829 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4830 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4832 *) Change the Unix randomness entropy gathering to use poll() when
4833 possible instead of select(), since the latter has some
4834 undesirable limitations.
4835 [Darryl Miles via Richard Levitte and Bodo Moeller]
4837 *) Disable rogue ciphersuites:
4839 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4840 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4841 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4843 The latter two were purportedly from
4844 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4847 Also deactive the remaining ciphersuites from
4848 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4849 unofficial, and the ID has long expired.
4852 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4853 dual-core machines) and other potential thread-safety issues.
4856 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4858 *) Adapt fipsld and the build system to link against the validated FIPS
4859 module in FIPS mode.
4862 *) Fixes for VC++ 2005 build under Windows.
4865 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4866 from a Windows bash shell such as MSYS. It is autodetected from the
4867 "config" script when run from a VC++ environment. Modify standard VC++
4868 build to use fipscanister.o from the GNU make build.
4871 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4873 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4874 The value now differs depending on if you build for FIPS or not.
4875 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4876 safely run with a non-FIPSed libcrypto, as it may crash because of
4877 the difference induced by this change.
4880 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4882 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4883 (part of SSL_OP_ALL). This option used to disable the
4884 countermeasure against man-in-the-middle protocol-version
4885 rollback in the SSL 2.0 server implementation, which is a bad
4886 idea. (CVE-2005-2969)
4888 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4889 for Information Security, National Institute of Advanced Industrial
4890 Science and Technology [AIST], Japan)]
4892 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4893 mainly for FIPS compliance and not fully integrated at this stage.
4896 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4897 the exponentiation using a fixed-length exponent. (Otherwise,
4898 the information leaked through timing could expose the secret key
4899 after many signatures; cf. Bleichenbacher's attack on DSA with
4903 *) Make a new fixed-window mod_exp implementation the default for
4904 RSA, DSA, and DH private-key operations so that the sequence of
4905 squares and multiplies and the memory access pattern are
4906 independent of the particular secret key. This will mitigate
4907 cache-timing and potential related attacks.
4909 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4910 and this is automatically used by BN_mod_exp_mont() if the new flag
4911 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4912 will use this BN flag for private exponents unless the flag
4913 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4914 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4916 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4918 *) Change the client implementation for SSLv23_method() and
4919 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4920 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4921 (Previously, the SSL 2.0 backwards compatible Client Hello
4922 message format would be used even with SSL_OP_NO_SSLv2.)
4925 *) Add support for smime-type MIME parameter in S/MIME messages which some
4929 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4930 a threadsafe manner. Modify rsa code to use new function and add calls
4931 to dsa and dh code (which had race conditions before).
4934 *) Include the fixed error library code in the C error file definitions
4935 instead of fixing them up at runtime. This keeps the error code
4936 structures constant.
4939 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4941 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4944 *) Fixes for newer kerberos headers. NB: the casts are needed because
4945 the 'length' field is signed on one version and unsigned on another
4946 with no (?) obvious way to tell the difference, without these VC++
4947 complains. Also the "definition" of FAR (blank) is no longer included
4948 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4949 some needed definitions.
4952 *) Undo Cygwin change.
4955 *) Added support for proxy certificates according to RFC 3820.
4956 Because they may be a security thread to unaware applications,
4957 they must be explicitely allowed in run-time. See
4958 docs/HOWTO/proxy_certificates.txt for further information.
4961 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4963 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4964 server and client random values. Previously
4965 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4966 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4968 This change has negligible security impact because:
4970 1. Server and client random values still have 24 bytes of pseudo random
4973 2. Server and client random values are sent in the clear in the initial
4976 3. The master secret is derived using the premaster secret (48 bytes in
4977 size for static RSA ciphersuites) as well as client server and random
4980 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4983 [Stephen Henson, reported by UK NISCC]
4985 *) Use Windows randomness collection on Cygwin.
4988 *) Fix hang in EGD/PRNGD query when communication socket is closed
4989 prematurely by EGD/PRNGD.
4990 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4992 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4995 *) Back-port of selected performance improvements from development
4996 branch, as well as improved support for PowerPC platforms.
4999 *) Add lots of checks for memory allocation failure, error codes to indicate
5000 failure and freeing up memory if a failure occurs.
5001 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5003 *) Add new -passin argument to dgst.
5006 *) Perform some character comparisons of different types in X509_NAME_cmp:
5007 this is needed for some certificates that reencode DNs into UTF8Strings
5008 (in violation of RFC3280) and can't or wont issue name rollover
5012 *) Make an explicit check during certificate validation to see that
5013 the CA setting in each certificate on the chain is correct. As a
5014 side effect always do the following basic checks on extensions,
5015 not just when there's an associated purpose to the check:
5017 - if there is an unhandled critical extension (unless the user
5018 has chosen to ignore this fault)
5019 - if the path length has been exceeded (if one is set at all)
5020 - that certain extensions fit the associated purpose (if one has
5024 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5026 *) Avoid a race condition when CRLs are checked in a multi threaded
5027 environment. This would happen due to the reordering of the revoked
5028 entries during signature checking and serial number lookup. Now the
5029 encoding is cached and the serial number sort performed under a lock.
5030 Add new STACK function sk_is_sorted().
5033 *) Add Delta CRL to the extension code.
5036 *) Various fixes to s3_pkt.c so alerts are sent properly.
5037 [David Holmes <d.holmes@f5.com>]
5039 *) Reduce the chances of duplicate issuer name and serial numbers (in
5040 violation of RFC3280) using the OpenSSL certificate creation utilities.
5041 This is done by creating a random 64 bit value for the initial serial
5042 number when a serial number file is created or when a self signed
5043 certificate is created using 'openssl req -x509'. The initial serial
5044 number file is created using 'openssl x509 -next_serial' in CA.pl
5045 rather than being initialized to 1.
5048 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5050 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5051 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5052 [Joe Orton, Steve Henson]
5054 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5056 [Joe Orton, Steve Henson]
5058 *) Make it possible to have multiple active certificates with the same
5059 subject in the CA index file. This is done only if the keyword
5060 'unique_subject' is set to 'no' in the main CA section (default
5061 if 'CA_default') of the configuration file. The value is saved
5062 with the database itself in a separate index attribute file,
5063 named like the index file with '.attr' appended to the name.
5066 *) X509 verify fixes. Disable broken certificate workarounds when
5067 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5068 keyUsage extension present. Don't accept CRLs with unhandled critical
5069 extensions: since verify currently doesn't process CRL extensions this
5070 rejects a CRL with *any* critical extensions. Add new verify error codes
5074 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5075 A clarification of RFC2560 will require the use of OCTET STRINGs and
5076 some implementations cannot handle the current raw format. Since OpenSSL
5077 copies and compares OCSP nonces as opaque blobs without any attempt at
5078 parsing them this should not create any compatibility issues.
5081 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5082 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5083 this HMAC (and other) operations are several times slower than OpenSSL
5087 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5088 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5090 *) Use the correct content when signing type "other".
5093 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5095 *) Fix various bugs revealed by running the NISCC test suite:
5097 Stop out of bounds reads in the ASN1 code when presented with
5098 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5100 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5102 If verify callback ignores invalid public key errors don't try to check
5103 certificate signature with the NULL public key.
5107 *) New -ignore_err option in ocsp application to stop the server
5108 exiting on the first error in a request.
5111 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5112 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5116 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5117 extra data after the compression methods not only for TLS 1.0
5118 but also for SSL 3.0 (as required by the specification).
5119 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5121 *) Change X509_certificate_type() to mark the key as exported/exportable
5122 when it's 512 *bits* long, not 512 bytes.
5125 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5126 blocks during encryption.
5129 *) Various fixes to base64 BIO and non blocking I/O. On write
5130 flushes were not handled properly if the BIO retried. On read
5131 data was not being buffered properly and had various logic bugs.
5132 This also affects blocking I/O when the data being decoded is a
5136 *) Various S/MIME bugfixes and compatibility changes:
5137 output correct application/pkcs7 MIME type if
5138 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5139 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5140 of files as .eml work). Correctly handle very long lines in MIME
5144 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5146 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5147 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5148 a protocol version number mismatch like a decryption error
5149 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5152 *) Turn on RSA blinding by default in the default implementation
5153 to avoid a timing attack. Applications that don't want it can call
5154 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5155 They would be ill-advised to do so in most cases.
5156 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5158 *) Change RSA blinding code so that it works when the PRNG is not
5159 seeded (in this case, the secret RSA exponent is abused as
5160 an unpredictable seed -- if it is not unpredictable, there
5161 is no point in blinding anyway). Make RSA blinding thread-safe
5162 by remembering the creator's thread ID in rsa->blinding and
5163 having all other threads use local one-time blinding factors
5164 (this requires more computation than sharing rsa->blinding, but
5165 avoids excessive locking; and if an RSA object is not shared
5166 between threads, blinding will still be very fast).
5169 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5170 ENGINE as defaults for all supported algorithms irrespective of
5171 the 'flags' parameter. 'flags' is now honoured, so applications
5172 should make sure they are passing it correctly.
5175 *) Target "mingw" now allows native Windows code to be generated in
5176 the Cygwin environment as well as with the MinGW compiler.
5179 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5181 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5182 via timing by performing a MAC computation even if incorrect
5183 block cipher padding has been found. This is a countermeasure
5184 against active attacks where the attacker has to distinguish
5185 between bad padding and a MAC verification error. (CVE-2003-0078)
5187 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5188 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5189 Martin Vuagnoux (EPFL, Ilion)]
5191 *) Make the no-err option work as intended. The intention with no-err
5192 is not to have the whole error stack handling routines removed from
5193 libcrypto, it's only intended to remove all the function name and
5194 reason texts, thereby removing some of the footprint that may not
5195 be interesting if those errors aren't displayed anyway.
5197 NOTE: it's still possible for any application or module to have it's
5198 own set of error texts inserted. The routines are there, just not
5199 used by default when no-err is given.
5202 *) Add support for FreeBSD on IA64.
5203 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5205 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5206 Kerberos function mit_des_cbc_cksum(). Before this change,
5207 the value returned by DES_cbc_cksum() was like the one from
5208 mit_des_cbc_cksum(), except the bytes were swapped.
5209 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5211 *) Allow an application to disable the automatic SSL chain building.
5212 Before this a rather primitive chain build was always performed in
5213 ssl3_output_cert_chain(): an application had no way to send the
5214 correct chain if the automatic operation produced an incorrect result.
5216 Now the chain builder is disabled if either:
5218 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5220 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5222 The reasoning behind this is that an application would not want the
5223 auto chain building to take place if extra chain certificates are
5224 present and it might also want a means of sending no additional
5225 certificates (for example the chain has two certificates and the
5229 *) Add the possibility to build without the ENGINE framework.
5230 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5232 *) Under Win32 gmtime() can return NULL: check return value in
5233 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5236 *) DSA routines: under certain error conditions uninitialized BN objects
5237 could be freed. Solution: make sure initialization is performed early
5238 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5239 Nils Larsch <nla@trustcenter.de> via PR#459)
5242 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5243 checked on reconnect on the client side, therefore session resumption
5244 could still fail with a "ssl session id is different" error. This
5245 behaviour is masked when SSL_OP_ALL is used due to
5246 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5247 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5248 followup to PR #377.
5251 *) IA-32 assembler support enhancements: unified ELF targets, support
5252 for SCO/Caldera platforms, fix for Cygwin shared build.
5255 *) Add support for FreeBSD on sparc64. As a consequence, support for
5256 FreeBSD on non-x86 processors is separate from x86 processors on
5257 the config script, much like the NetBSD support.
5258 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5260 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5262 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5265 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5266 code (06) was taken as the first octet of the session ID and the last
5267 octet was ignored consequently. As a result SSLv2 client side session
5268 caching could not have worked due to the session ID mismatch between
5270 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5274 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5275 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5279 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5280 seems that in spite of existing for more than a year, many application
5281 author have done nothing to provide the necessary callbacks, which
5282 means that this particular engine will not work properly anywhere.
5283 This is a very unfortunate situation which forces us, in the name
5284 of usability, to give the hw_ncipher.c a static lock, which is part
5286 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5287 appear in 0.9.8 or later. We EXPECT application authors to have
5288 dealt properly with this when 0.9.8 is released (unless we actually
5289 make such changes in the libcrypto locking code that changes will
5290 have to be made anyway).
5293 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5294 octets have been read, EOF or an error occurs. Without this change
5295 some truncated ASN1 structures will not produce an error.
5298 *) Disable Heimdal support, since it hasn't been fully implemented.
5299 Still give the possibility to force the use of Heimdal, but with
5300 warnings and a request that patches get sent to openssl-dev.
5303 *) Add the VC-CE target, introduce the WINCE sysname, and add
5304 INSTALL.WCE and appropriate conditionals to make it build.
5305 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5307 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5308 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5309 edit numbers of the version.
5310 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5312 *) Introduce safe string copy and catenation functions
5313 (BUF_strlcpy() and BUF_strlcat()).
5314 [Ben Laurie (CHATS) and Richard Levitte]
5316 *) Avoid using fixed-size buffers for one-line DNs.
5317 [Ben Laurie (CHATS)]
5319 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5320 resizing buffers containing secrets, and use where appropriate.
5321 [Ben Laurie (CHATS)]
5323 *) Avoid using fixed size buffers for configuration file location.
5324 [Ben Laurie (CHATS)]
5326 *) Avoid filename truncation for various CA files.
5327 [Ben Laurie (CHATS)]
5329 *) Use sizeof in preference to magic numbers.
5330 [Ben Laurie (CHATS)]
5332 *) Avoid filename truncation in cert requests.
5333 [Ben Laurie (CHATS)]
5335 *) Add assertions to check for (supposedly impossible) buffer
5337 [Ben Laurie (CHATS)]
5339 *) Don't cache truncated DNS entries in the local cache (this could
5340 potentially lead to a spoofing attack).
5341 [Ben Laurie (CHATS)]
5343 *) Fix various buffers to be large enough for hex/decimal
5344 representations in a platform independent manner.
5345 [Ben Laurie (CHATS)]
5347 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5348 resizing buffers containing secrets, and use where appropriate.
5349 [Ben Laurie (CHATS)]
5351 *) Add BIO_indent() to avoid much slightly worrying code to do
5353 [Ben Laurie (CHATS)]
5355 *) Convert sprintf()/BIO_puts() to BIO_printf().
5356 [Ben Laurie (CHATS)]
5358 *) buffer_gets() could terminate with the buffer only half
5360 [Ben Laurie (CHATS)]
5362 *) Add assertions to prevent user-supplied crypto functions from
5363 overflowing internal buffers by having large block sizes, etc.
5364 [Ben Laurie (CHATS)]
5366 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5368 [Ben Laurie (CHATS)]
5370 *) Eliminate unused copy of key in RC4.
5371 [Ben Laurie (CHATS)]
5373 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5374 [Ben Laurie (CHATS)]
5376 *) Fix off-by-one error in EGD path.
5377 [Ben Laurie (CHATS)]
5379 *) If RANDFILE path is too long, ignore instead of truncating.
5380 [Ben Laurie (CHATS)]
5382 *) Eliminate unused and incorrectly sized X.509 structure
5384 [Ben Laurie (CHATS)]
5386 *) Eliminate unused and dangerous function knumber().
5387 [Ben Laurie (CHATS)]
5389 *) Eliminate unused and dangerous structure, KSSL_ERR.
5390 [Ben Laurie (CHATS)]
5392 *) Protect against overlong session ID context length in an encoded
5393 session object. Since these are local, this does not appear to be
5395 [Ben Laurie (CHATS)]
5397 *) Change from security patch (see 0.9.6e below) that did not affect
5398 the 0.9.6 release series:
5400 Remote buffer overflow in SSL3 protocol - an attacker could
5401 supply an oversized master key in Kerberos-enabled versions.
5403 [Ben Laurie (CHATS)]
5405 *) Change the SSL kerb5 codes to match RFC 2712.
5408 *) Make -nameopt work fully for req and add -reqopt switch.
5409 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5411 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5412 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5414 *) Make sure tests can be performed even if the corresponding algorithms
5415 have been removed entirely. This was also the last step to make
5416 OpenSSL compilable with DJGPP under all reasonable conditions.
5417 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5419 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5420 to allow version independent disabling of normally unselected ciphers,
5421 which may be activated as a side-effect of selecting a single cipher.
5423 (E.g., cipher list string "RSA" enables ciphersuites that are left
5424 out of "ALL" because they do not provide symmetric encryption.
5425 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5426 [Lutz Jaenicke, Bodo Moeller]
5428 *) Add appropriate support for separate platform-dependent build
5429 directories. The recommended way to make a platform-dependent
5430 build directory is the following (tested on Linux), maybe with
5433 # Place yourself outside of the OpenSSL source tree. In
5434 # this example, the environment variable OPENSSL_SOURCE
5435 # is assumed to contain the absolute OpenSSL source directory.
5436 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5437 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5438 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5439 mkdir -p `dirname $F`
5440 ln -s $OPENSSL_SOURCE/$F $F
5443 To be absolutely sure not to disturb the source tree, a "make clean"
5444 is a good thing. If it isn't successful, don't worry about it,
5445 it probably means the source directory is very clean.
5448 *) Make sure any ENGINE control commands make local copies of string
5449 pointers passed to them whenever necessary. Otherwise it is possible
5450 the caller may have overwritten (or deallocated) the original string
5451 data when a later ENGINE operation tries to use the stored values.
5452 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5454 *) Improve diagnostics in file reading and command-line digests.
5455 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5457 *) Add AES modes CFB and OFB to the object database. Correct an
5458 error in AES-CFB decryption.
5461 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5462 allows existing EVP_CIPHER_CTX structures to be reused after
5463 calling EVP_*Final(). This behaviour is used by encryption
5464 BIOs and some applications. This has the side effect that
5465 applications must explicitly clean up cipher contexts with
5466 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5469 *) Check the values of dna and dnb in bn_mul_recursive before calling
5470 bn_mul_comba (a non zero value means the a or b arrays do not contain
5471 n2 elements) and fallback to bn_mul_normal if either is not zero.
5474 *) Fix escaping of non-ASCII characters when using the -subj option
5475 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5478 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5479 form for "surname", serialNumber has no short form.
5480 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5481 therefore remove "mail" short name for "internet 7".
5482 The OID for unique identifiers in X509 certificates is
5483 x500UniqueIdentifier, not uniqueIdentifier.
5484 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5487 *) Add an "init" command to the ENGINE config module and auto initialize
5488 ENGINEs. Without any "init" command the ENGINE will be initialized
5489 after all ctrl commands have been executed on it. If init=1 the
5490 ENGINE is initailized at that point (ctrls before that point are run
5491 on the uninitialized ENGINE and after on the initialized one). If
5492 init=0 then the ENGINE will not be iniatialized at all.
5495 *) Fix the 'app_verify_callback' interface so that the user-defined
5496 argument is actually passed to the callback: In the
5497 SSL_CTX_set_cert_verify_callback() prototype, the callback
5498 declaration has been changed from
5501 int (*cb)(X509_STORE_CTX *,void *);
5502 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5503 i=s->ctx->app_verify_callback(&ctx)
5504 has been changed into
5505 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5507 To update applications using SSL_CTX_set_cert_verify_callback(),
5508 a dummy argument can be added to their callback functions.
5509 [D. K. Smetters <smetters@parc.xerox.com>]
5511 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5512 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5514 *) Add and OPENSSL_LOAD_CONF define which will cause
5515 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5516 This allows older applications to transparently support certain
5517 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5518 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5519 load the config file and OPENSSL_add_all_algorithms_conf() which will
5520 always load it have also been added.
5523 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5524 Adjust NIDs and EVP layer.
5525 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5527 *) Config modules support in openssl utility.
5529 Most commands now load modules from the config file,
5530 though in a few (such as version) this isn't done
5531 because it couldn't be used for anything.
5533 In the case of ca and req the config file used is
5534 the same as the utility itself: that is the -config
5535 command line option can be used to specify an
5539 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5540 use "openssl_conf" if filename is NULL use default openssl config file.
5543 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5544 config section name. Add a new flag to tolerate a missing config file
5545 and move code to CONF_modules_load_file().
5548 *) Support for crypto accelerator cards from Accelerated Encryption
5549 Processing, www.aep.ie. (Use engine 'aep')
5550 The support was copied from 0.9.6c [engine] and adapted/corrected
5551 to work with the new engine framework.
5552 [AEP Inc. and Richard Levitte]
5554 *) Support for SureWare crypto accelerator cards from Baltimore
5555 Technologies. (Use engine 'sureware')
5556 The support was copied from 0.9.6c [engine] and adapted
5557 to work with the new engine framework.
5560 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5561 make the newer ENGINE framework commands for the CHIL engine work.
5562 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5564 *) Make it possible to produce shared libraries on ReliantUNIX.
5565 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5567 *) Add the configuration target debug-linux-ppro.
5568 Make 'openssl rsa' use the general key loading routines
5569 implemented in apps.c, and make those routines able to
5570 handle the key format FORMAT_NETSCAPE and the variant
5572 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5574 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5575 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5577 *) Add -keyform to rsautl, and document -engine.
5578 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5580 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5581 BIO_R_NO_SUCH_FILE error code rather than the generic
5582 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5585 *) Add new functions
5587 ERR_peek_last_error_line
5588 ERR_peek_last_error_line_data.
5589 These are similar to
5592 ERR_peek_error_line_data,
5593 but report on the latest error recorded rather than the first one
5594 still in the error queue.
5595 [Ben Laurie, Bodo Moeller]
5597 *) default_algorithms option in ENGINE config module. This allows things
5599 default_algorithms = ALL
5600 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5603 *) Preliminary ENGINE config module.
5606 *) New experimental application configuration code.
5609 *) Change the AES code to follow the same name structure as all other
5610 symmetric ciphers, and behave the same way. Move everything to
5611 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5612 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5614 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5615 [Ben Laurie and Theo de Raadt]
5617 *) Add option to output public keys in req command.
5618 [Massimiliano Pala madwolf@openca.org]
5620 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5621 (up to about 10% better than before for P-192 and P-224).
5624 *) New functions/macros
5626 SSL_CTX_set_msg_callback(ctx, cb)
5627 SSL_CTX_set_msg_callback_arg(ctx, arg)
5628 SSL_set_msg_callback(ssl, cb)
5629 SSL_set_msg_callback_arg(ssl, arg)
5631 to request calling a callback function
5633 void cb(int write_p, int version, int content_type,
5634 const void *buf, size_t len, SSL *ssl, void *arg)
5636 whenever a protocol message has been completely received
5637 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5638 protocol version according to which the SSL library interprets
5639 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5640 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5641 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5642 specification (change_cipher_spec(20), alert(21), handshake(22)).
5643 'buf' and 'len' point to the actual message, 'ssl' to the
5644 SSL object, and 'arg' is the application-defined value set by
5645 SSL[_CTX]_set_msg_callback_arg().
5647 'openssl s_client' and 'openssl s_server' have new '-msg' options
5648 to enable a callback that displays all protocol messages.
5651 *) Change the shared library support so shared libraries are built as
5652 soon as the corresponding static library is finished, and thereby get
5653 openssl and the test programs linked against the shared library.
5654 This still only happens when the keyword "shard" has been given to
5655 the configuration scripts.
5657 NOTE: shared library support is still an experimental thing, and
5658 backward binary compatibility is still not guaranteed.
5659 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5661 *) Add support for Subject Information Access extension.
5662 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5664 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5665 additional bytes when new memory had to be allocated, not just
5666 when reusing an existing buffer.
5669 *) New command line and configuration option 'utf8' for the req command.
5670 This allows field values to be specified as UTF8 strings.
5673 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5674 runs for the former and machine-readable output for the latter.
5677 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5678 of the e-mail address in the DN (i.e., it will go into a certificate
5679 extension only). The new configuration file option 'email_in_dn = no'
5680 has the same effect.
5681 [Massimiliano Pala madwolf@openca.org]
5683 *) Change all functions with names starting with des_ to be starting
5684 with DES_ instead. Add wrappers that are compatible with libdes,
5685 but are named _ossl_old_des_*. Finally, add macros that map the
5686 des_* symbols to the corresponding _ossl_old_des_* if libdes
5687 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5688 desired, the des_* symbols will be mapped to DES_*, with one
5691 Since we provide two compatibility mappings, the user needs to
5692 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5693 compatibility is desired. The default (i.e., when that macro
5694 isn't defined) is OpenSSL 0.9.6c compatibility.
5696 There are also macros that enable and disable the support of old
5697 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5698 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5699 are defined, the default will apply: to support the old des routines.
5701 In either case, one must include openssl/des.h to get the correct
5702 definitions. Do not try to just include openssl/des_old.h, that
5705 NOTE: This is a major break of an old API into a new one. Software
5706 authors are encouraged to switch to the DES_ style functions. Some
5707 time in the future, des_old.h and the libdes compatibility functions
5708 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5709 default), and then completely removed.
5712 *) Test for certificates which contain unsupported critical extensions.
5713 If such a certificate is found during a verify operation it is
5714 rejected by default: this behaviour can be overridden by either
5715 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5716 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5717 X509_supported_extension() has also been added which returns 1 if a
5718 particular extension is supported.
5721 *) Modify the behaviour of EVP cipher functions in similar way to digests
5722 to retain compatibility with existing code.
5725 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5726 compatibility with existing code. In particular the 'ctx' parameter does
5727 not have to be to be initialized before the call to EVP_DigestInit() and
5728 it is tidied up after a call to EVP_DigestFinal(). New function
5729 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5730 EVP_MD_CTX_copy() changed to not require the destination to be
5731 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5732 requires the destination to be valid.
5734 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5735 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5738 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5739 so that complete 'Handshake' protocol structures are kept in memory
5740 instead of overwriting 'msg_type' and 'length' with 'body' data.
5743 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5744 [Massimo Santin via Richard Levitte]
5746 *) Major restructuring to the underlying ENGINE code. This includes
5747 reduction of linker bloat, separation of pure "ENGINE" manipulation
5748 (initialisation, etc) from functionality dealing with implementations
5749 of specific crypto iterfaces. This change also introduces integrated
5750 support for symmetric ciphers and digest implementations - so ENGINEs
5751 can now accelerate these by providing EVP_CIPHER and EVP_MD
5752 implementations of their own. This is detailed in crypto/engine/README
5753 as it couldn't be adequately described here. However, there are a few
5754 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5755 were changed in the original introduction of ENGINE code have now
5756 reverted back - the hooking from this code to ENGINE is now a good
5757 deal more passive and at run-time, operations deal directly with
5758 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5759 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5760 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5761 they were not being used by the framework as there is no concept of a
5762 BIGNUM_METHOD and they could not be generalised to the new
5763 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5764 ENGINE_cpy() has been removed as it cannot be consistently defined in
5768 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5771 *) Change mkdef.pl to sort symbols that get the same entry number,
5772 and make sure the automatically generated functions ERR_load_*
5773 become part of libeay.num as well.
5776 *) New function SSL_renegotiate_pending(). This returns true once
5777 renegotiation has been requested (either SSL_renegotiate() call
5778 or HelloRequest/ClientHello received from the peer) and becomes
5779 false once a handshake has been completed.
5780 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5781 sends a HelloRequest, but does not ensure that a handshake takes
5782 place. SSL_renegotiate_pending() is useful for checking if the
5783 client has followed the request.)
5786 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5787 By default, clients may request session resumption even during
5788 renegotiation (if session ID contexts permit); with this option,
5789 session resumption is possible only in the first handshake.
5791 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5792 more bits available for options that should not be part of
5793 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5796 *) Add some demos for certificate and certificate request creation.
5799 *) Make maximum certificate chain size accepted from the peer application
5800 settable (SSL*_get/set_max_cert_list()), as proposed by
5801 "Douglas E. Engert" <deengert@anl.gov>.
5804 *) Add support for shared libraries for Unixware-7
5805 (Boyd Lynn Gerber <gerberb@zenez.com>).
5808 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5809 be done prior to destruction. Use this to unload error strings from
5810 ENGINEs that load their own error strings. NB: This adds two new API
5811 functions to "get" and "set" this destroy handler in an ENGINE.
5814 *) Alter all existing ENGINE implementations (except "openssl" and
5815 "openbsd") to dynamically instantiate their own error strings. This
5816 makes them more flexible to be built both as statically-linked ENGINEs
5817 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5818 Also, add stub code to each that makes building them as self-contained
5819 shared-libraries easier (see README.ENGINE).
5822 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5823 implementations into applications that are completely implemented in
5824 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5825 commands that can be used to configure what shared-library to load and
5826 to control aspects of the way it is handled. Also, made an update to
5827 the README.ENGINE file that brings its information up-to-date and
5828 provides some information and instructions on the "dynamic" ENGINE
5829 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5832 *) Make it possible to unload ranges of ERR strings with a new
5833 "ERR_unload_strings" function.
5836 *) Add a copy() function to EVP_MD.
5839 *) Make EVP_MD routines take a context pointer instead of just the
5840 md_data void pointer.
5843 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5844 that the digest can only process a single chunk of data
5845 (typically because it is provided by a piece of
5846 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5847 is only going to provide a single chunk of data, and hence the
5848 framework needn't accumulate the data for oneshot drivers.
5851 *) As with "ERR", make it possible to replace the underlying "ex_data"
5852 functions. This change also alters the storage and management of global
5853 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5854 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5855 index counters. The API functions that use this state have been changed
5856 to take a "class_index" rather than pointers to the class's local STACK
5857 and counter, and there is now an API function to dynamically create new
5858 classes. This centralisation allows us to (a) plug a lot of the
5859 thread-safety problems that existed, and (b) makes it possible to clean
5860 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5861 such data would previously have always leaked in application code and
5862 workarounds were in place to make the memory debugging turn a blind eye
5863 to it. Application code that doesn't use this new function will still
5864 leak as before, but their memory debugging output will announce it now
5865 rather than letting it slide.
5867 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5868 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5869 has a return value to indicate success or failure.
5872 *) Make it possible to replace the underlying "ERR" functions such that the
5873 global state (2 LHASH tables and 2 locks) is only used by the "default"
5874 implementation. This change also adds two functions to "get" and "set"
5875 the implementation prior to it being automatically set the first time
5876 any other ERR function takes place. Ie. an application can call "get",
5877 pass the return value to a module it has just loaded, and that module
5878 can call its own "set" function using that value. This means the
5879 module's "ERR" operations will use (and modify) the error state in the
5880 application and not in its own statically linked copy of OpenSSL code.
5883 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5884 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5885 the operation, and provides a more encapsulated way for external code
5886 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5887 to use these functions rather than manually incrementing the counts.
5889 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5892 *) Add EVP test program.
5895 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5898 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5899 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5900 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5901 These allow a CRL to be built without having to access X509_CRL fields
5902 directly. Modify 'ca' application to use new functions.
5905 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5906 bug workarounds. Rollback attack detection is a security feature.
5907 The problem will only arise on OpenSSL servers when TLSv1 is not
5908 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5909 Software authors not wanting to support TLSv1 will have special reasons
5910 for their choice and can explicitly enable this option.
5911 [Bodo Moeller, Lutz Jaenicke]
5913 *) Rationalise EVP so it can be extended: don't include a union of
5914 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5915 (similar to those existing for EVP_CIPHER_CTX).
5920 EVP_MD_CTX_init(&md); /* new function call */
5921 EVP_DigestInit(&md, EVP_sha1());
5922 EVP_DigestUpdate(&md, in, len);
5923 EVP_DigestFinal(&md, out, NULL);
5924 EVP_MD_CTX_cleanup(&md); /* new function call */
5928 *) Make DES key schedule conform to the usual scheme, as well as
5929 correcting its structure. This means that calls to DES functions
5930 now have to pass a pointer to a des_key_schedule instead of a
5931 plain des_key_schedule (which was actually always a pointer
5934 des_key_schedule ks;
5936 des_set_key_checked(..., &ks);
5937 des_ncbc_encrypt(..., &ks, ...);
5939 (Note that a later change renames 'des_...' into 'DES_...'.)
5942 *) Initial reduction of linker bloat: the use of some functions, such as
5943 PEM causes large amounts of unused functions to be linked in due to
5944 poor organisation. For example pem_all.c contains every PEM function
5945 which has a knock on effect of linking in large amounts of (unused)
5946 ASN1 code. Grouping together similar functions and splitting unrelated
5947 functions prevents this.
5950 *) Cleanup of EVP macros.
5953 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5954 correct _ecb suffix.
5957 *) Add initial OCSP responder support to ocsp application. The
5958 revocation information is handled using the text based index
5959 use by the ca application. The responder can either handle
5960 requests generated internally, supplied in files (for example
5961 via a CGI script) or using an internal minimal server.
5964 *) Add configuration choices to get zlib compression for TLS.
5967 *) Changes to Kerberos SSL for RFC 2712 compliance:
5968 1. Implemented real KerberosWrapper, instead of just using
5969 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5970 2. Implemented optional authenticator field of KerberosWrapper.
5972 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5973 and authenticator structs; see crypto/krb5/.
5975 Generalized Kerberos calls to support multiple Kerberos libraries.
5976 [Vern Staats <staatsvr@asc.hpc.mil>,
5977 Jeffrey Altman <jaltman@columbia.edu>
5978 via Richard Levitte]
5980 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5981 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5982 values for each of the key sizes rather than having just
5983 parameters (and 'speed' generating keys each time).
5986 *) Speed up EVP routines.
5989 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5990 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5991 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5992 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5994 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5995 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5996 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5999 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6001 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6004 *) Added the OS2-EMX target.
6005 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6007 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6008 to support NCONF routines in extension code. New function CONF_set_nconf()
6009 to allow functions which take an NCONF to also handle the old LHASH
6010 structure: this means that the old CONF compatible routines can be
6011 retained (in particular wrt extensions) without having to duplicate the
6012 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6015 *) Enhance the general user interface with mechanisms for inner control
6016 and with possibilities to have yes/no kind of prompts.
6019 *) Change all calls to low level digest routines in the library and
6020 applications to use EVP. Add missing calls to HMAC_cleanup() and
6021 don't assume HMAC_CTX can be copied using memcpy().
6022 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6024 *) Add the possibility to control engines through control names but with
6025 arbitrary arguments instead of just a string.
6026 Change the key loaders to take a UI_METHOD instead of a callback
6027 function pointer. NOTE: this breaks binary compatibility with earlier
6028 versions of OpenSSL [engine].
6029 Adapt the nCipher code for these new conditions and add a card insertion
6033 *) Enhance the general user interface with mechanisms to better support
6034 dialog box interfaces, application-defined prompts, the possibility
6035 to use defaults (for example default passwords from somewhere else)
6036 and interrupts/cancellations.
6039 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6040 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6043 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6044 tidy up some unnecessarily weird code in 'sk_new()').
6045 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6047 *) Change the key loading routines for ENGINEs to use the same kind
6048 callback (pem_password_cb) as all other routines that need this
6052 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6053 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6054 than this minimum value is recommended.
6057 *) New random seeder for OpenVMS, using the system process statistics
6058 that are easily reachable.
6061 *) Windows apparently can't transparently handle global
6062 variables defined in DLLs. Initialisations such as:
6064 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6066 wont compile. This is used by the any applications that need to
6067 declare their own ASN1 modules. This was fixed by adding the option
6068 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6069 needed for static libraries under Win32.
6072 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6073 setting of purpose and trust fields. New X509_STORE trust and
6074 purpose functions and tidy up setting in other SSL functions.
6077 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6078 structure. These are inherited by X509_STORE_CTX when it is
6079 initialised. This allows various defaults to be set in the
6080 X509_STORE structure (such as flags for CRL checking and custom
6081 purpose or trust settings) for functions which only use X509_STORE_CTX
6082 internally such as S/MIME.
6084 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6085 trust settings if they are not set in X509_STORE. This allows X509_STORE
6086 purposes and trust (in S/MIME for example) to override any set by default.
6088 Add command line options for CRL checking to smime, s_client and s_server
6092 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6093 are set then the CRL is looked up in the X509_STORE structure and
6094 its validity and signature checked, then if the certificate is found
6095 in the CRL the verify fails with a revoked error.
6097 Various new CRL related callbacks added to X509_STORE_CTX structure.
6099 Command line options added to 'verify' application to support this.
6101 This needs some additional work, such as being able to handle multiple
6102 CRLs with different times, extension based lookup (rather than just
6103 by subject name) and ultimately more complete V2 CRL extension
6107 *) Add a general user interface API (crypto/ui/). This is designed
6108 to replace things like des_read_password and friends (backward
6109 compatibility functions using this new API are provided).
6110 The purpose is to remove prompting functions from the DES code
6111 section as well as provide for prompting through dialog boxes in
6112 a window system and the like.
6115 *) Add "ex_data" support to ENGINE so implementations can add state at a
6116 per-structure level rather than having to store it globally.
6119 *) Make it possible for ENGINE structures to be copied when retrieved by
6120 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6121 This causes the "original" ENGINE structure to act like a template,
6122 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6123 operational state can be localised to each ENGINE structure, despite the
6124 fact they all share the same "methods". New ENGINE structures returned in
6125 this case have no functional references and the return value is the single
6126 structural reference. This matches the single structural reference returned
6127 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6131 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6132 needs to match any other type at all we need to manually clear the
6136 *) Changes to the "openssl engine" utility to include;
6137 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6138 about an ENGINE's available control commands.
6139 - executing control commands from command line arguments using the
6140 '-pre' and '-post' switches. '-post' is only used if '-t' is
6141 specified and the ENGINE is successfully initialised. The syntax for
6142 the individual commands are colon-separated, for example;
6143 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6146 *) New dynamic control command support for ENGINEs. ENGINEs can now
6147 declare their own commands (numbers), names (strings), descriptions,
6148 and input types for run-time discovery by calling applications. A
6149 subset of these commands are implicitly classed as "executable"
6150 depending on their input type, and only these can be invoked through
6151 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6152 can be based on user input, config files, etc). The distinction is
6153 that "executable" commands cannot return anything other than a boolean
6154 result and can only support numeric or string input, whereas some
6155 discoverable commands may only be for direct use through
6156 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6157 pointers, or other custom uses. The "executable" commands are to
6158 support parameterisations of ENGINE behaviour that can be
6159 unambiguously defined by ENGINEs and used consistently across any
6160 OpenSSL-based application. Commands have been added to all the
6161 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6162 control over shared-library paths without source code alterations.
6165 *) Changed all ENGINE implementations to dynamically allocate their
6166 ENGINEs rather than declaring them statically. Apart from this being
6167 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6168 this also allows the implementations to compile without using the
6169 internal engine_int.h header.
6172 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6173 'const' value. Any code that should be able to modify a RAND_METHOD
6174 should already have non-const pointers to it (ie. they should only
6175 modify their own ones).
6178 *) Made a variety of little tweaks to the ENGINE code.
6179 - "atalla" and "ubsec" string definitions were moved from header files
6180 to C code. "nuron" string definitions were placed in variables
6181 rather than hard-coded - allowing parameterisation of these values
6182 later on via ctrl() commands.
6183 - Removed unused "#if 0"'d code.
6184 - Fixed engine list iteration code so it uses ENGINE_free() to release
6185 structural references.
6186 - Constified the RAND_METHOD element of ENGINE structures.
6187 - Constified various get/set functions as appropriate and added
6188 missing functions (including a catch-all ENGINE_cpy that duplicates
6189 all ENGINE values onto a new ENGINE except reference counts/state).
6190 - Removed NULL parameter checks in get/set functions. Setting a method
6191 or function to NULL is a way of cancelling out a previously set
6192 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6193 and doesn't justify the extra error symbols and code.
6194 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6195 flags from engine_int.h to engine.h.
6196 - Changed prototypes for ENGINE handler functions (init(), finish(),
6197 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6200 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6201 to the algorithm using long division. The binary algorithm can be
6202 used only if the modulus is odd. On 32-bit systems, it is faster
6203 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6204 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6205 up to 450 bits. In 64-bit environments, the binary algorithm
6206 appears to be advantageous for much longer moduli; here we use it
6207 for moduli up to 2048 bits.
6210 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6211 could not support the combine flag in choice fields.
6214 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6215 extensions from a certificate request to the certificate.
6218 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6219 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6220 file: this allows the display of the certificate about to be
6221 signed to be customised, to allow certain fields to be included
6222 or excluded and extension details. The old system didn't display
6223 multicharacter strings properly, omitted fields not in the policy
6224 and couldn't display additional details such as extensions.
6227 *) Function EC_POINTs_mul for multiple scalar multiplication
6228 of an arbitrary number of elliptic curve points
6229 \sum scalars[i]*points[i],
6230 optionally including the generator defined for the EC_GROUP:
6231 scalar*generator + \sum scalars[i]*points[i].
6233 EC_POINT_mul is a simple wrapper function for the typical case
6234 that the point list has just one item (besides the optional
6238 *) First EC_METHODs for curves over GF(p):
6240 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6241 operations and provides various method functions that can also
6242 operate with faster implementations of modular arithmetic.
6244 EC_GFp_mont_method() reuses most functions that are part of
6245 EC_GFp_simple_method, but uses Montgomery arithmetic.
6247 [Bodo Moeller; point addition and point doubling
6248 implementation directly derived from source code provided by
6249 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6251 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6252 crypto/ec/ec_lib.c):
6254 Curves are EC_GROUP objects (with an optional group generator)
6255 based on EC_METHODs that are built into the library.
6257 Points are EC_POINT objects based on EC_GROUP objects.
6259 Most of the framework would be able to handle curves over arbitrary
6260 finite fields, but as there are no obvious types for fields other
6261 than GF(p), some functions are limited to that for now.
6264 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6265 that the file contains a complete HTTP response.
6268 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6269 change the def and num file printf format specifier from "%-40sXXX"
6270 to "%-39s XXX". The latter will always guarantee a space after the
6271 field while the former will cause them to run together if the field
6272 is 40 of more characters long.
6275 *) Constify the cipher and digest 'method' functions and structures
6276 and modify related functions to take constant EVP_MD and EVP_CIPHER
6280 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6281 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6284 *) Modify EVP_Digest*() routines so they now return values. Although the
6285 internal software routines can never fail additional hardware versions
6289 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6291 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6292 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6295 ERR_R_NESTED_ASN1_ERROR
6297 ERR_R_MISSING_ASN1_EOS
6298 were 4 .. 9, conflicting with
6299 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6301 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6302 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6304 Add new error code 'ERR_R_INTERNAL_ERROR'.
6307 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6311 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6312 sets the subject name for a new request or supersedes the
6313 subject name in a given request. Formats that can be parsed are
6314 'CN=Some Name, OU=myOU, C=IT'
6316 'CN=Some Name/OU=myOU/C=IT'.
6318 Add options '-batch' and '-verbose' to 'openssl req'.
6319 [Massimiliano Pala <madwolf@hackmasters.net>]
6321 *) Introduce the possibility to access global variables through
6322 functions on platform were that's the best way to handle exporting
6323 global variables in shared libraries. To enable this functionality,
6324 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6325 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6326 is normally done by Configure or something similar).
6328 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6329 in the source file (foo.c) like this:
6331 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6332 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6334 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6335 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6337 OPENSSL_DECLARE_GLOBAL(int,foo);
6338 #define foo OPENSSL_GLOBAL_REF(foo)
6339 OPENSSL_DECLARE_GLOBAL(double,bar);
6340 #define bar OPENSSL_GLOBAL_REF(bar)
6342 The #defines are very important, and therefore so is including the
6343 header file everywhere where the defined globals are used.
6345 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6346 of ASN.1 items, but that structure is a bit different.
6348 The largest change is in util/mkdef.pl which has been enhanced with
6349 better and easier to understand logic to choose which symbols should
6350 go into the Windows .def files as well as a number of fixes and code
6351 cleanup (among others, algorithm keywords are now sorted
6352 lexicographically to avoid constant rewrites).
6355 *) In BN_div() keep a copy of the sign of 'num' before writing the
6356 result to 'rm' because if rm==num the value will be overwritten
6357 and produce the wrong result if 'num' is negative: this caused
6358 problems with BN_mod() and BN_nnmod().
6361 *) Function OCSP_request_verify(). This checks the signature on an
6362 OCSP request and verifies the signer certificate. The signer
6363 certificate is just checked for a generic purpose and OCSP request
6367 *) Add OCSP_check_validity() function to check the validity of OCSP
6368 responses. OCSP responses are prepared in real time and may only
6369 be a few seconds old. Simply checking that the current time lies
6370 between thisUpdate and nextUpdate max reject otherwise valid responses
6371 caused by either OCSP responder or client clock inaccuracy. Instead
6372 we allow thisUpdate and nextUpdate to fall within a certain period of
6373 the current time. The age of the response can also optionally be
6374 checked. Two new options -validity_period and -status_age added to
6378 *) If signature or public key algorithm is unrecognized print out its
6379 OID rather that just UNKNOWN.
6382 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6383 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6384 ID to be generated from the issuer certificate alone which can then be
6385 passed to OCSP_id_issuer_cmp().
6388 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6389 ASN1 modules to export functions returning ASN1_ITEM pointers
6390 instead of the ASN1_ITEM structures themselves. This adds several
6391 new macros which allow the underlying ASN1 function/structure to
6392 be accessed transparently. As a result code should not use ASN1_ITEM
6393 references directly (such as &X509_it) but instead use the relevant
6394 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6395 use of the new ASN1 code on platforms where exporting structures
6396 is problematical (for example in shared libraries) but exporting
6397 functions returning pointers to structures is not.
6400 *) Add support for overriding the generation of SSL/TLS session IDs.
6401 These callbacks can be registered either in an SSL_CTX or per SSL.
6402 The purpose of this is to allow applications to control, if they wish,
6403 the arbitrary values chosen for use as session IDs, particularly as it
6404 can be useful for session caching in multiple-server environments. A
6405 command-line switch for testing this (and any client code that wishes
6406 to use such a feature) has been added to "s_server".
6407 [Geoff Thorpe, Lutz Jaenicke]
6409 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6410 of the form '#if defined(...) || defined(...) || ...' and
6411 '#if !defined(...) && !defined(...) && ...'. This also avoids
6412 the growing number of special cases it was previously handling.
6415 *) Make all configuration macros available for application by making
6416 sure they are available in opensslconf.h, by giving them names starting
6417 with "OPENSSL_" to avoid conflicts with other packages and by making
6418 sure e_os2.h will cover all platform-specific cases together with
6420 Additionally, it is now possible to define configuration/platform-
6421 specific names (called "system identities"). In the C code, these
6422 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6423 macro with the name beginning with "OPENSSL_SYS_", which is determined
6424 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6428 *) New option -set_serial to 'req' and 'x509' this allows the serial
6429 number to use to be specified on the command line. Previously self
6430 signed certificates were hard coded with serial number 0 and the
6431 CA options of 'x509' had to use a serial number in a file which was
6435 *) New options to 'ca' utility to support V2 CRL entry extensions.
6436 Currently CRL reason, invalidity date and hold instruction are
6437 supported. Add new CRL extensions to V3 code and some new objects.
6440 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6441 disable standard block padding (aka PKCS#5 padding) in the EVP
6442 API, which was previously mandatory. This means that the data is
6443 not padded in any way and so the total length much be a multiple
6444 of the block size, otherwise an error occurs.
6447 *) Initial (incomplete) OCSP SSL support.
6450 *) New function OCSP_parse_url(). This splits up a URL into its host,
6451 port and path components: primarily to parse OCSP URLs. New -url
6452 option to ocsp utility.
6455 *) New nonce behavior. The return value of OCSP_check_nonce() now
6456 reflects the various checks performed. Applications can decide
6457 whether to tolerate certain situations such as an absent nonce
6458 in a response when one was present in a request: the ocsp application
6459 just prints out a warning. New function OCSP_add1_basic_nonce()
6460 this is to allow responders to include a nonce in a response even if
6461 the request is nonce-less.
6464 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6465 skipped when using openssl x509 multiple times on a single input file,
6466 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6469 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6470 set string type: to handle setting ASN1_TIME structures. Fix ca
6471 utility to correctly initialize revocation date of CRLs.
6474 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6475 the clients preferred ciphersuites and rather use its own preferences.
6476 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6477 Internet Explorer by ensuring unchanged hash method during stepup.
6478 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6481 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6482 to aes and add a new 'exist' option to print out symbols that don't
6486 *) Additional options to ocsp utility to allow flags to be set and
6487 additional certificates supplied.
6490 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6491 OCSP client a number of certificate to only verify the response
6495 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6496 handle the new API. Currently only ECB, CBC modes supported. Add new
6499 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6500 Encryption Standard (AES) Ciphersuites for Transport Layer
6501 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6502 not enabled by default and were not part of the "ALL" ciphersuite
6503 alias because they were not yet official; they could be
6504 explicitly requested by specifying the "AESdraft" ciphersuite
6505 group alias. In the final release of OpenSSL 0.9.7, the group
6506 alias is called "AES" and is part of "ALL".)
6507 [Ben Laurie, Steve Henson, Bodo Moeller]
6509 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6510 request to response.
6513 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6514 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6515 extract information from a certificate request. OCSP_response_create()
6516 creates a response and optionally adds a basic response structure.
6517 OCSP_basic_add1_status() adds a complete single response to a basic
6518 response and returns the OCSP_SINGLERESP structure just added (to allow
6519 extensions to be included for example). OCSP_basic_add1_cert() adds a
6520 certificate to a basic response and OCSP_basic_sign() signs a basic
6521 response with various flags. New helper functions ASN1_TIME_check()
6522 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6523 (converts ASN1_TIME to GeneralizedTime).
6526 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6527 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6528 structure from a certificate. X509_pubkey_digest() digests the public_key
6529 contents: this is used in various key identifiers.
6532 *) Make sk_sort() tolerate a NULL argument.
6533 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6535 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6536 passed by the function are trusted implicitly. If any of them signed the
6537 response then it is assumed to be valid and is not verified.
6540 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6541 to data. This was previously part of the PKCS7 ASN1 code. This
6542 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6543 [Steve Henson, reported by Kenneth R. Robinette
6544 <support@securenetterm.com>]
6546 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6547 routines: without these tracing memory leaks is very painful.
6548 Fix leaks in PKCS12 and PKCS7 routines.
6551 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6552 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6553 effectively meant GeneralizedTime would never be used. Now it
6554 is initialised to -1 but X509_time_adj() now has to check the value
6555 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6556 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6557 [Steve Henson, reported by Kenneth R. Robinette
6558 <support@securenetterm.com>]
6560 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6561 result in a zero length in the ASN1_INTEGER structure which was
6562 not consistent with the structure when d2i_ASN1_INTEGER() was used
6563 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6564 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6565 where it did not print out a minus for negative ASN1_INTEGER.
6568 *) Add summary printout to ocsp utility. The various functions which
6569 convert status values to strings have been renamed to:
6570 OCSP_response_status_str(), OCSP_cert_status_str() and
6571 OCSP_crl_reason_str() and are no longer static. New options
6572 to verify nonce values and to disable verification. OCSP response
6573 printout format cleaned up.
6576 *) Add additional OCSP certificate checks. These are those specified
6577 in RFC2560. This consists of two separate checks: the CA of the
6578 certificate being checked must either be the OCSP signer certificate
6579 or the issuer of the OCSP signer certificate. In the latter case the
6580 OCSP signer certificate must contain the OCSP signing extended key
6581 usage. This check is performed by attempting to match the OCSP
6582 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6583 in the OCSP_CERTID structures of the response.
6586 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6587 and related routines. This uses the standard OpenSSL certificate
6588 verify routines to perform initial checks (just CA validity) and
6589 to obtain the certificate chain. Then additional checks will be
6590 performed on the chain. Currently the root CA is checked to see
6591 if it is explicitly trusted for OCSP signing. This is used to set
6592 a root CA as a global signing root: that is any certificate that
6593 chains to that CA is an acceptable OCSP signing certificate.
6596 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6597 extensions from a separate configuration file.
6598 As when reading extensions from the main configuration file,
6599 the '-extensions ...' option may be used for specifying the
6601 [Massimiliano Pala <madwolf@comune.modena.it>]
6603 *) New OCSP utility. Allows OCSP requests to be generated or
6604 read. The request can be sent to a responder and the output
6605 parsed, outputed or printed in text form. Not complete yet:
6606 still needs to check the OCSP response validity.
6609 *) New subcommands for 'openssl ca':
6610 'openssl ca -status <serial>' prints the status of the cert with
6611 the given serial number (according to the index file).
6612 'openssl ca -updatedb' updates the expiry status of certificates
6614 [Massimiliano Pala <madwolf@comune.modena.it>]
6616 *) New '-newreq-nodes' command option to CA.pl. This is like
6617 '-newreq', but calls 'openssl req' with the '-nodes' option
6618 so that the resulting key is not encrypted.
6619 [Damien Miller <djm@mindrot.org>]
6621 *) New configuration for the GNU Hurd.
6622 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6624 *) Initial code to implement OCSP basic response verify. This
6625 is currently incomplete. Currently just finds the signer's
6626 certificate and verifies the signature on the response.
6629 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6630 value of OPENSSLDIR. This is available via the new '-d' option
6631 to 'openssl version', and is also included in 'openssl version -a'.
6634 *) Allowing defining memory allocation callbacks that will be given
6635 file name and line number information in additional arguments
6636 (a const char* and an int). The basic functionality remains, as
6637 well as the original possibility to just replace malloc(),
6638 realloc() and free() by functions that do not know about these
6639 additional arguments. To register and find out the current
6640 settings for extended allocation functions, the following
6641 functions are provided:
6643 CRYPTO_set_mem_ex_functions
6644 CRYPTO_set_locked_mem_ex_functions
6645 CRYPTO_get_mem_ex_functions
6646 CRYPTO_get_locked_mem_ex_functions
6648 These work the same way as CRYPTO_set_mem_functions and friends.
6649 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6650 extended allocation function is enabled.
6651 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6652 a conventional allocation function is enabled.
6653 [Richard Levitte, Bodo Moeller]
6655 *) Finish off removing the remaining LHASH function pointer casts.
6656 There should no longer be any prototype-casting required when using
6657 the LHASH abstraction, and any casts that remain are "bugs". See
6658 the callback types and macros at the head of lhash.h for details
6659 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6662 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6663 If /dev/[u]random devices are not available or do not return enough
6664 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6666 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6667 /etc/entropy will be queried once each in this sequence, quering stops
6668 when enough entropy was collected without querying more sockets.
6671 *) Change the Unix RAND_poll() variant to be able to poll several
6672 random devices, as specified by DEVRANDOM, until a sufficient amount
6673 of data has been collected. We spend at most 10 ms on each file
6674 (select timeout) and read in non-blocking mode. DEVRANDOM now
6675 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6676 (previously it was just the string "/dev/urandom"), so on typical
6677 platforms the 10 ms delay will never occur.
6678 Also separate out the Unix variant to its own file, rand_unix.c.
6679 For VMS, there's a currently-empty rand_vms.c.
6682 *) Move OCSP client related routines to ocsp_cl.c. These
6683 provide utility functions which an application needing
6684 to issue a request to an OCSP responder and analyse the
6685 response will typically need: as opposed to those which an
6686 OCSP responder itself would need which will be added later.
6688 OCSP_request_sign() signs an OCSP request with an API similar
6689 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6690 response. OCSP_response_get1_basic() extracts basic response
6691 from response. OCSP_resp_find_status(): finds and extracts status
6692 information from an OCSP_CERTID structure (which will be created
6693 when the request structure is built). These are built from lower
6694 level functions which work on OCSP_SINGLERESP structures but
6695 wont normally be used unless the application wishes to examine
6696 extensions in the OCSP response for example.
6698 Replace nonce routines with a pair of functions.
6699 OCSP_request_add1_nonce() adds a nonce value and optionally
6700 generates a random value. OCSP_check_nonce() checks the
6701 validity of the nonce in an OCSP response.
6704 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6705 This doesn't copy the supplied OCSP_CERTID and avoids the
6706 need to free up the newly created id. Change return type
6707 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6708 This can then be used to add extensions to the request.
6709 Deleted OCSP_request_new(), since most of its functionality
6710 is now in OCSP_REQUEST_new() (and the case insensitive name
6711 clash) apart from the ability to set the request name which
6712 will be added elsewhere.
6715 *) Update OCSP API. Remove obsolete extensions argument from
6716 various functions. Extensions are now handled using the new
6717 OCSP extension code. New simple OCSP HTTP function which
6718 can be used to send requests and parse the response.
6721 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6722 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6723 uses the special reorder version of SET OF to sort the attributes
6724 and reorder them to match the encoded order. This resolves a long
6725 standing problem: a verify on a PKCS7 structure just after signing
6726 it used to fail because the attribute order did not match the
6727 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6728 it uses the received order. This is necessary to tolerate some broken
6729 software that does not order SET OF. This is handled by encoding
6730 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6731 to produce the required SET OF.
6734 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6735 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6736 files to get correct declarations of the ASN.1 item variables.
6739 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6740 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6741 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6742 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6743 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6744 ASN1_ITEM and no wrapper functions.
6747 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6748 replace the old function pointer based I/O routines. Change most of
6749 the *_d2i_bio() and *_d2i_fp() functions to use these.
6752 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6753 lines, recognice more "algorithms" that can be deselected, and make
6754 it complain about algorithm deselection that isn't recognised.
6757 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6758 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6759 to use new functions. Add NO_ASN1_OLD which can be set to remove
6760 some old style ASN1 functions: this can be used to determine if old
6761 code will still work when these eventually go away.
6764 *) New extension functions for OCSP structures, these follow the
6765 same conventions as certificates and CRLs.
6768 *) New function X509V3_add1_i2d(). This automatically encodes and
6769 adds an extension. Its behaviour can be customised with various
6770 flags to append, replace or delete. Various wrappers added for
6771 certifcates and CRLs.
6774 *) Fix to avoid calling the underlying ASN1 print routine when
6775 an extension cannot be parsed. Correct a typo in the
6776 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6779 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6780 entries for variables.
6783 *) Add functionality to apps/openssl.c for detecting locking
6784 problems: As the program is single-threaded, all we have
6785 to do is register a locking callback using an array for
6786 storing which locks are currently held by the program.
6789 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6790 SSL_get_ex_data_X509_STORE_idx(), which is used in
6791 ssl_verify_cert_chain() and thus can be called at any time
6792 during TLS/SSL handshakes so that thread-safety is essential.
6793 Unfortunately, the ex_data design is not at all suited
6794 for multi-threaded use, so it probably should be abolished.
6797 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6798 [Broadcom, tweaked and integrated by Geoff Thorpe]
6800 *) Move common extension printing code to new function
6801 X509V3_print_extensions(). Reorganise OCSP print routines and
6802 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6805 *) New function X509_signature_print() to remove duplication in some
6809 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6810 set (this was treated exactly the same as SET OF previously). This
6811 is used to reorder the STACK representing the structure to match the
6812 encoding. This will be used to get round a problem where a PKCS7
6813 structure which was signed could not be verified because the STACK
6814 order did not reflect the encoded order.
6817 *) Reimplement the OCSP ASN1 module using the new code.
6820 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6821 for its ASN1 operations. The old style function pointers still exist
6822 for now but they will eventually go away.
6825 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6826 completely replaces the old ASN1 functionality with a table driven
6827 encoder and decoder which interprets an ASN1_ITEM structure describing
6828 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6829 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6830 has also been converted to the new form.
6833 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6834 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6835 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6836 for negative moduli.
6839 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6840 of not touching the result's sign bit.
6843 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6847 *) Changed the LHASH code to use prototypes for callbacks, and created
6848 macros to declare and implement thin (optionally static) functions
6849 that provide type-safety and avoid function pointer casting for the
6850 type-specific callbacks.
6853 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6855 [Veers Staats <staatsvr@asc.hpc.mil>,
6856 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6858 *) Reformat the FAQ so the different questions and answers can be divided
6859 in sections depending on the subject.
6862 *) Have the zlib compression code load ZLIB.DLL dynamically under
6866 *) New function BN_mod_sqrt for computing square roots modulo a prime
6867 (using the probabilistic Tonelli-Shanks algorithm unless
6868 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6869 be handled deterministically).
6870 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6872 *) Make BN_mod_inverse faster by explicitly handling small quotients
6873 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6874 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6877 *) New function BN_kronecker.
6880 *) Fix BN_gcd so that it works on negative inputs; the result is
6881 positive unless both parameters are zero.
6882 Previously something reasonably close to an infinite loop was
6883 possible because numbers could be growing instead of shrinking
6884 in the implementation of Euclid's algorithm.
6887 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6888 sign of the number in question.
6890 Fix BN_is_word(a,w) to work correctly for w == 0.
6892 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6893 because its test if the absolute value of 'a' equals 'w'.
6894 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6895 it exists mostly for use in the implementations of BN_is_zero(),
6896 BN_is_one(), and BN_is_word().
6899 *) New function BN_swap.
6902 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6903 the exponentiation functions are more likely to produce reasonable
6904 results on negative inputs.
6907 *) Change BN_mod_mul so that the result is always non-negative.
6908 Previously, it could be negative if one of the factors was negative;
6909 I don't think anyone really wanted that behaviour.
6912 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6913 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6914 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6915 and add new functions:
6924 BN_mod_lshift1_quick
6928 These functions always generate non-negative results.
6930 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6931 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6933 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6934 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6935 be reduced modulo m.
6936 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6939 The following entry accidentally appeared in the CHANGES file
6940 distributed with OpenSSL 0.9.7. The modifications described in
6941 it do *not* apply to OpenSSL 0.9.7.
6943 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6944 was actually never needed) and in BN_mul(). The removal in BN_mul()
6945 required a small change in bn_mul_part_recursive() and the addition
6946 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6947 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6948 bn_sub_words() and bn_add_words() except they take arrays with
6953 *) In 'openssl passwd', verify passwords read from the terminal
6954 unless the '-salt' option is used (which usually means that
6955 verification would just waste user's time since the resulting
6956 hash is going to be compared with some given password hash)
6957 or the new '-noverify' option is used.
6959 This is an incompatible change, but it does not affect
6960 non-interactive use of 'openssl passwd' (passwords on the command
6961 line, '-stdin' option, '-in ...' option) and thus should not
6965 *) Remove all references to RSAref, since there's no more need for it.
6968 *) Make DSO load along a path given through an environment variable
6969 (SHLIB_PATH) with shl_load().
6972 *) Constify the ENGINE code as a result of BIGNUM constification.
6973 Also constify the RSA code and most things related to it. In a
6974 few places, most notable in the depth of the ASN.1 code, ugly
6975 casts back to non-const were required (to be solved at a later
6979 *) Make it so the openssl application has all engines loaded by default.
6982 *) Constify the BIGNUM routines a little more.
6985 *) Add the following functions:
6987 ENGINE_load_cswift()
6989 ENGINE_load_atalla()
6991 ENGINE_load_builtin_engines()
6993 That way, an application can itself choose if external engines that
6994 are built-in in OpenSSL shall ever be used or not. The benefit is
6995 that applications won't have to be linked with libdl or other dso
6996 libraries unless it's really needed.
6998 Changed 'openssl engine' to load all engines on demand.
6999 Changed the engine header files to avoid the duplication of some
7000 declarations (they differed!).
7003 *) 'openssl engine' can now list capabilities.
7006 *) Better error reporting in 'openssl engine'.
7009 *) Never call load_dh_param(NULL) in s_server.
7012 *) Add engine application. It can currently list engines by name and
7013 identity, and test if they are actually available.
7016 *) Improve RPM specification file by forcing symbolic linking and making
7017 sure the installed documentation is also owned by root.root.
7018 [Damien Miller <djm@mindrot.org>]
7020 *) Give the OpenSSL applications more possibilities to make use of
7021 keys (public as well as private) handled by engines.
7024 *) Add OCSP code that comes from CertCo.
7027 *) Add VMS support for the Rijndael code.
7030 *) Added untested support for Nuron crypto accelerator.
7033 *) Add support for external cryptographic devices. This code was
7034 previously distributed separately as the "engine" branch.
7035 [Geoff Thorpe, Richard Levitte]
7037 *) Rework the filename-translation in the DSO code. It is now possible to
7038 have far greater control over how a "name" is turned into a filename
7039 depending on the operating environment and any oddities about the
7040 different shared library filenames on each system.
7043 *) Support threads on FreeBSD-elf in Configure.
7046 *) Fix for SHA1 assembly problem with MASM: it produces
7047 warnings about corrupt line number information when assembling
7048 with debugging information. This is caused by the overlapping
7050 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7053 NCONF_get_number() has no error checking at all. As a replacement,
7054 NCONF_get_number_e() is defined (_e for "error checking") and is
7055 promoted strongly. The old NCONF_get_number is kept around for
7056 binary backward compatibility.
7057 Make it possible for methods to load from something other than a BIO,
7058 by providing a function pointer that is given a name instead of a BIO.
7059 For example, this could be used to load configuration data from an
7063 *) Fix for non blocking accept BIOs. Added new I/O special reason
7064 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7065 with non blocking I/O was not possible because no retry code was
7066 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7070 *) Added the beginnings of Rijndael support.
7073 *) Fix for bug in DirectoryString mask setting. Add support for
7074 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7075 to allow certificate printing to more controllable, additional
7076 'certopt' option to 'x509' to allow new printing options to be
7080 *) Clean old EAY MD5 hack from e_os.h.
7083 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7085 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7086 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7087 [Joe Orton, Steve Henson]
7089 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7091 *) Fix additional bug revealed by the NISCC test suite:
7093 Stop bug triggering large recursion when presented with
7094 certain ASN.1 tags (CVE-2003-0851)
7097 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7099 *) Fix various bugs revealed by running the NISCC test suite:
7101 Stop out of bounds reads in the ASN1 code when presented with
7102 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7104 If verify callback ignores invalid public key errors don't try to check
7105 certificate signature with the NULL public key.
7109 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7110 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7114 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7115 extra data after the compression methods not only for TLS 1.0
7116 but also for SSL 3.0 (as required by the specification).
7117 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7119 *) Change X509_certificate_type() to mark the key as exported/exportable
7120 when it's 512 *bits* long, not 512 bytes.
7123 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7125 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7126 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7127 a protocol version number mismatch like a decryption error
7128 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7131 *) Turn on RSA blinding by default in the default implementation
7132 to avoid a timing attack. Applications that don't want it can call
7133 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7134 They would be ill-advised to do so in most cases.
7135 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7137 *) Change RSA blinding code so that it works when the PRNG is not
7138 seeded (in this case, the secret RSA exponent is abused as
7139 an unpredictable seed -- if it is not unpredictable, there
7140 is no point in blinding anyway). Make RSA blinding thread-safe
7141 by remembering the creator's thread ID in rsa->blinding and
7142 having all other threads use local one-time blinding factors
7143 (this requires more computation than sharing rsa->blinding, but
7144 avoids excessive locking; and if an RSA object is not shared
7145 between threads, blinding will still be very fast).
7148 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7150 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7151 via timing by performing a MAC computation even if incorrrect
7152 block cipher padding has been found. This is a countermeasure
7153 against active attacks where the attacker has to distinguish
7154 between bad padding and a MAC verification error. (CVE-2003-0078)
7156 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7157 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7158 Martin Vuagnoux (EPFL, Ilion)]
7160 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7162 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7163 memory from it's contents. This is done with a counter that will
7164 place alternating values in each byte. This can be used to solve
7165 two issues: 1) the removal of calls to memset() by highly optimizing
7166 compilers, and 2) cleansing with other values than 0, since those can
7167 be read through on certain media, for example a swap space on disk.
7170 *) Bugfix: client side session caching did not work with external caching,
7171 because the session->cipher setting was not restored when reloading
7172 from the external cache. This problem was masked, when
7173 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7174 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7177 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7178 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7179 [Zeev Lieber <zeev-l@yahoo.com>]
7181 *) Undo an undocumented change introduced in 0.9.6e which caused
7182 repeated calls to OpenSSL_add_all_ciphers() and
7183 OpenSSL_add_all_digests() to be ignored, even after calling
7187 *) Change the default configuration reader to deal with last line not
7188 being properly terminated.
7191 *) Change X509_NAME_cmp() so it applies the special rules on handling
7192 DN values that are of type PrintableString, as well as RDNs of type
7193 emailAddress where the value has the type ia5String.
7194 [stefank@valicert.com via Richard Levitte]
7196 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7197 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7198 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7199 the bitwise-OR of the two for use by the majority of applications
7200 wanting this behaviour, and update the docs. The documented
7201 behaviour and actual behaviour were inconsistent and had been
7202 changing anyway, so this is more a bug-fix than a behavioural
7204 [Geoff Thorpe, diagnosed by Nadav Har'El]
7206 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7207 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7210 *) Fix initialization code race conditions in
7211 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7212 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7213 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7214 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7215 ssl2_get_cipher_by_char(),
7216 ssl3_get_cipher_by_char().
7217 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7219 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7220 the cached sessions are flushed, as the remove_cb() might use ex_data
7221 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7222 (see [openssl.org #212]).
7223 [Geoff Thorpe, Lutz Jaenicke]
7225 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7226 length, instead of the encoding length to d2i_ASN1_OBJECT.
7229 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7231 *) [In 0.9.6g-engine release:]
7232 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7233 [Lynn Gazis <lgazis@rainbow.com>]
7235 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7237 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7238 and get fix the header length calculation.
7239 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7240 Alon Kantor <alonk@checkpoint.com> (and others),
7243 *) Use proper error handling instead of 'assertions' in buffer
7244 overflow checks added in 0.9.6e. This prevents DoS (the
7245 assertions could call abort()).
7246 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7248 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7250 *) Add various sanity checks to asn1_get_length() to reject
7251 the ASN1 length bytes if they exceed sizeof(long), will appear
7252 negative or the content length exceeds the length of the
7254 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7256 *) Fix cipher selection routines: ciphers without encryption had no flags
7257 for the cipher strength set and where therefore not handled correctly
7258 by the selection routines (PR #130).
7261 *) Fix EVP_dsa_sha macro.
7265 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7266 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7267 that was added in OpenSSL 0.9.6d.
7269 As the countermeasure turned out to be incompatible with some
7270 broken SSL implementations, the new option is part of SSL_OP_ALL.
7271 SSL_OP_ALL is usually employed when compatibility with weird SSL
7272 implementations is desired (e.g. '-bugs' option to 's_client' and
7273 's_server'), so the new option is automatically set in many
7277 *) Changes in security patch:
7279 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7280 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7281 Air Force Materiel Command, USAF, under agreement number
7284 *) Add various sanity checks to asn1_get_length() to reject
7285 the ASN1 length bytes if they exceed sizeof(long), will appear
7286 negative or the content length exceeds the length of the
7287 supplied buffer. (CVE-2002-0659)
7288 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7290 *) Assertions for various potential buffer overflows, not known to
7292 [Ben Laurie (CHATS)]
7294 *) Various temporary buffers to hold ASCII versions of integers were
7295 too small for 64 bit platforms. (CVE-2002-0655)
7296 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7298 *) Remote buffer overflow in SSL3 protocol - an attacker could
7299 supply an oversized session ID to a client. (CVE-2002-0656)
7300 [Ben Laurie (CHATS)]
7302 *) Remote buffer overflow in SSL2 protocol - an attacker could
7303 supply an oversized client master key. (CVE-2002-0656)
7304 [Ben Laurie (CHATS)]
7306 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7308 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7309 encoded as NULL) with id-dsa-with-sha1.
7310 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7312 *) Check various X509_...() return values in apps/req.c.
7313 [Nils Larsch <nla@trustcenter.de>]
7315 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7316 an end-of-file condition would erronously be flagged, when the CRLF
7317 was just at the end of a processed block. The bug was discovered when
7318 processing data through a buffering memory BIO handing the data to a
7319 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7320 <ptsekov@syntrex.com> and Nedelcho Stanev.
7323 *) Implement a countermeasure against a vulnerability recently found
7324 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7325 before application data chunks to avoid the use of known IVs
7326 with data potentially chosen by the attacker.
7329 *) Fix length checks in ssl3_get_client_hello().
7332 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7333 to prevent ssl3_read_internal() from incorrectly assuming that
7334 ssl3_read_bytes() found application data while handshake
7335 processing was enabled when in fact s->s3->in_read_app_data was
7336 merely automatically cleared during the initial handshake.
7337 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7339 *) Fix object definitions for Private and Enterprise: they were not
7340 recognized in their shortname (=lowercase) representation. Extend
7341 obj_dat.pl to issue an error when using undefined keywords instead
7342 of silently ignoring the problem (Svenning Sorensen
7343 <sss@sss.dnsalias.net>).
7346 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7347 generators, i.e. generators other than 2 and 5. (Previously, the
7348 code did not properly initialise the 'add' and 'rem' values to
7349 BN_generate_prime().)
7351 In the new general case, we do not insist that 'generator' is
7352 actually a primitive root: This requirement is rather pointless;
7353 a generator of the order-q subgroup is just as good, if not
7357 *) Map new X509 verification errors to alerts. Discovered and submitted by
7358 Tom Wu <tom@arcot.com>.
7361 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7362 returning non-zero before the data has been completely received
7363 when using non-blocking I/O.
7364 [Bodo Moeller; problem pointed out by John Hughes]
7366 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7367 [Ben Laurie, Lutz Jaenicke]
7369 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7370 Yoram Zahavi <YoramZ@gilian.com>).
7373 *) Add information about CygWin 1.3 and on, and preserve proper
7374 configuration for the versions before that.
7375 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7377 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7378 check whether we deal with a copy of a session and do not delete from
7379 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7380 <izhar@checkpoint.com>.
7383 *) Do not store session data into the internal session cache, if it
7384 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7385 flag is set). Proposed by Aslam <aslam@funk.com>.
7388 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7392 *) [In 0.9.6d-engine release:]
7393 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7394 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7396 *) Add the configuration target linux-s390x.
7397 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7399 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7400 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7401 variable as an indication that a ClientHello message has been
7402 received. As the flag value will be lost between multiple
7403 invocations of ssl3_accept when using non-blocking I/O, the
7404 function may not be aware that a handshake has actually taken
7405 place, thus preventing a new session from being added to the
7408 To avoid this problem, we now set s->new_session to 2 instead of
7409 using a local variable.
7410 [Lutz Jaenicke, Bodo Moeller]
7412 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7413 if the SSL_R_LENGTH_MISMATCH error is detected.
7414 [Geoff Thorpe, Bodo Moeller]
7416 *) New 'shared_ldflag' column in Configure platform table.
7419 *) Fix EVP_CIPHER_mode macro.
7420 ["Dan S. Camper" <dan@bti.net>]
7422 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7423 type, we must throw them away by setting rr->length to 0.
7424 [D P Chang <dpc@qualys.com>]
7426 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7428 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7429 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7430 worked incorrectly for those cases where range = 10..._2 and
7431 3*range is two bits longer than range.)
7434 *) Only add signing time to PKCS7 structures if it is not already
7438 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7439 OBJ_ld_ce should be OBJ_id_ce.
7440 Also some ip-pda OIDs in crypto/objects/objects.txt were
7441 incorrect (cf. RFC 3039).
7442 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7444 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7445 returns early because it has nothing to do.
7446 [Andy Schneider <andy.schneider@bjss.co.uk>]
7448 *) [In 0.9.6c-engine release:]
7449 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7450 [Andy Schneider <andy.schneider@bjss.co.uk>]
7452 *) [In 0.9.6c-engine release:]
7453 Add support for Cryptographic Appliance's keyserver technology.
7454 (Use engine 'keyclient')
7455 [Cryptographic Appliances and Geoff Thorpe]
7457 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7458 is called via tools/c89.sh because arguments have to be
7459 rearranged (all '-L' options must appear before the first object
7461 [Richard Shapiro <rshapiro@abinitio.com>]
7463 *) [In 0.9.6c-engine release:]
7464 Add support for Broadcom crypto accelerator cards, backported
7466 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7468 *) [In 0.9.6c-engine release:]
7469 Add support for SureWare crypto accelerator cards from
7470 Baltimore Technologies. (Use engine 'sureware')
7471 [Baltimore Technologies and Mark Cox]
7473 *) [In 0.9.6c-engine release:]
7474 Add support for crypto accelerator cards from Accelerated
7475 Encryption Processing, www.aep.ie. (Use engine 'aep')
7476 [AEP Inc. and Mark Cox]
7478 *) Add a configuration entry for gcc on UnixWare.
7479 [Gary Benson <gbenson@redhat.com>]
7481 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7482 messages are stored in a single piece (fixed-length part and
7483 variable-length part combined) and fix various bugs found on the way.
7486 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7487 instead. BIO_gethostbyname() does not know what timeouts are
7488 appropriate, so entries would stay in cache even when they have
7490 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7492 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7493 faced with a pathologically small ClientHello fragment that does
7494 not contain client_version: Instead of aborting with an error,
7495 simply choose the highest available protocol version (i.e.,
7496 TLS 1.0 unless it is disabled). In practice, ClientHello
7497 messages are never sent like this, but this change gives us
7498 strictly correct behaviour at least for TLS.
7501 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7502 never resets s->method to s->ctx->method when called from within
7503 one of the SSL handshake functions.
7504 [Bodo Moeller; problem pointed out by Niko Baric]
7506 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7507 (sent using the client's version number) if client_version is
7508 smaller than the protocol version in use. Also change
7509 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7510 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7511 the client will at least see that alert.
7514 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7518 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7519 client receives HelloRequest while in a handshake.
7520 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7522 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7523 should end in 'break', not 'goto end' which circumvents various
7524 cleanups done in state SSL_ST_OK. But session related stuff
7525 must be disabled for SSL_ST_OK in the case that we just sent a
7528 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7529 before just sending a HelloRequest.
7530 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7532 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7533 reveal whether illegal block cipher padding was found or a MAC
7534 verification error occured. (Neither SSLerr() codes nor alerts
7535 are directly visible to potential attackers, but the information
7536 may leak via logfiles.)
7538 Similar changes are not required for the SSL 2.0 implementation
7539 because the number of padding bytes is sent in clear for SSL 2.0,
7540 and the extra bytes are just ignored. However ssl/s2_pkt.c
7541 failed to verify that the purported number of padding bytes is in
7545 *) Add OpenUNIX-8 support including shared libraries
7546 (Boyd Lynn Gerber <gerberb@zenez.com>).
7549 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7550 'wristwatch attack' using huge encoding parameters (cf.
7551 James H. Manger's CRYPTO 2001 paper). Note that the
7552 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7553 encoding parameters and hence was not vulnerable.
7556 *) BN_sqr() bug fix.
7557 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7559 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7560 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7561 followed by modular reduction.
7562 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7564 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7565 equivalent based on BN_pseudo_rand() instead of BN_rand().
7568 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7569 This function was broken, as the check for a new client hello message
7570 to handle SGC did not allow these large messages.
7571 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7574 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7577 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7578 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7581 *) Rework the configuration and shared library support for Tru64 Unix.
7582 The configuration part makes use of modern compiler features and
7583 still retains old compiler behavior for those that run older versions
7584 of the OS. The shared library support part includes a variant that
7585 uses the RPATH feature, and is available through the special
7586 configuration target "alpha-cc-rpath", which will never be selected
7588 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7590 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7591 with the same message size as in ssl3_get_certificate_request().
7592 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7593 messages might inadvertently be reject as too long.
7594 [Petr Lampa <lampa@fee.vutbr.cz>]
7596 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7599 *) Modified SSL library such that the verify_callback that has been set
7600 specificly for an SSL object with SSL_set_verify() is actually being
7601 used. Before the change, a verify_callback set with this function was
7602 ignored and the verify_callback() set in the SSL_CTX at the time of
7603 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7604 to allow the necessary settings.
7607 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7608 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7609 done automatically (in contradiction to the requirements of the C
7610 standard). This made problems when used from OpenSSH.
7613 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7614 dh->length and always used
7616 BN_rand_range(priv_key, dh->p).
7618 BN_rand_range() is not necessary for Diffie-Hellman, and this
7619 specific range makes Diffie-Hellman unnecessarily inefficient if
7620 dh->length (recommended exponent length) is much smaller than the
7621 length of dh->p. We could use BN_rand_range() if the order of
7622 the subgroup was stored in the DH structure, but we only have
7627 BN_rand(priv_key, l, ...)
7629 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7635 RSA_eay_public_encrypt
7636 RSA_eay_private_decrypt
7637 RSA_eay_private_encrypt (signing)
7638 RSA_eay_public_decrypt (signature verification)
7640 (default implementations for RSA_public_encrypt,
7641 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7642 always reject numbers >= n.
7645 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7646 to synchronize access to 'locking_thread'. This is necessary on
7647 systems where access to 'locking_thread' (an 'unsigned long'
7648 variable) is not atomic.
7651 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7652 *before* setting the 'crypto_lock_rand' flag. The previous code had
7653 a race condition if 0 is a valid thread ID.
7654 [Travis Vitek <vitek@roguewave.com>]
7656 *) Add support for shared libraries under Irix.
7657 [Albert Chin-A-Young <china@thewrittenword.com>]
7659 *) Add configuration option to build on Linux on both big-endian and
7661 [Ralf Baechle <ralf@uni-koblenz.de>]
7663 *) Add the possibility to create shared libraries on HP-UX.
7666 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7668 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7669 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7670 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7671 PRNG state recovery was possible based on the output of
7672 one PRNG request appropriately sized to gain knowledge on
7673 'md' followed by enough consecutive 1-byte PRNG requests
7674 to traverse all of 'state'.
7676 1. When updating 'md_local' (the current thread's copy of 'md')
7677 during PRNG output generation, hash all of the previous
7678 'md_local' value, not just the half used for PRNG output.
7680 2. Make the number of bytes from 'state' included into the hash
7681 independent from the number of PRNG bytes requested.
7683 The first measure alone would be sufficient to avoid
7684 Markku-Juhani's attack. (Actually it had never occurred
7685 to me that the half of 'md_local' used for chaining was the
7686 half from which PRNG output bytes were taken -- I had always
7687 assumed that the secret half would be used.) The second
7688 measure makes sure that additional data from 'state' is never
7689 mixed into 'md_local' in small portions; this heuristically
7690 further strengthens the PRNG.
7693 *) Fix crypto/bn/asm/mips3.s.
7696 *) When only the key is given to "enc", the IV is undefined. Print out
7697 an error message in this case.
7700 *) Handle special case when X509_NAME is empty in X509 printing routines.
7703 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7704 positive and less than q.
7707 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7708 used: it isn't thread safe and the add_lock_callback should handle
7710 [Paul Rose <Paul.Rose@bridge.com>]
7712 *) Verify that incoming data obeys the block size in
7713 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7717 [Ulf Möller, Bodo Möller]
7719 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7720 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7721 when fixing the server behaviour for backwards-compatible 'client
7722 hello' messages. (Note that the attack is impractical against
7723 SSL 3.0 and TLS 1.0 anyway because length and version checking
7724 means that the probability of guessing a valid ciphertext is
7725 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7728 Before 0.9.5, the countermeasure (hide the error by generating a
7729 random 'decryption result') did not work properly because
7730 ERR_clear_error() was missing, meaning that SSL_get_error() would
7731 detect the supposedly ignored error.
7733 Both problems are now fixed.
7736 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7737 (previously it was 1024).
7740 *) Fix for compatibility mode trust settings: ignore trust settings
7741 unless some valid trust or reject settings are present.
7744 *) Fix for blowfish EVP: its a variable length cipher.
7747 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7748 parameters in DSA public key structures and return an error in the
7749 DSA routines if parameters are absent.
7752 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7753 in the current directory if neither $RANDFILE nor $HOME was set.
7754 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7755 caused some confusion to Windows users who haven't defined $HOME.
7756 Thus RAND_file_name() is changed again: e_os.h can define a
7757 DEFAULT_HOME, which will be used if $HOME is not set.
7758 For Windows, we use "C:"; on other platforms, we still require
7759 environment variables.
7761 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7762 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7763 having multiple threads call RAND_poll() concurrently.
7766 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7767 combination of a flag and a thread ID variable.
7768 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7769 flag), *other* threads can enter ssleay_add_bytes without obeying
7770 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7771 that they do not hold after the first thread unsets add_do_not_lock).
7774 *) Change bctest again: '-x' expressions are not available in all
7778 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7780 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7781 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7783 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7784 the default extension for executables, if any. Also, make the perl
7785 scripts that use symlink() to test if it really exists and use "cp"
7786 if it doesn't. All this made OpenSSL compilable and installable in
7790 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7791 If SEQUENCE is length is indefinite just set c->slen to the total
7792 amount of data available.
7793 [Steve Henson, reported by shige@FreeBSD.org]
7794 [This change does not apply to 0.9.7.]
7796 *) Change bctest to avoid here-documents inside command substitution
7797 (workaround for FreeBSD /bin/sh bug).
7798 For compatibility with Ultrix, avoid shell functions (introduced
7799 in the bctest version that searches along $PATH).
7802 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7803 with des_encrypt() defined on some operating systems, like Solaris
7807 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7808 On the Importance of Eliminating Errors in Cryptographic
7809 Computations, J. Cryptology 14 (2001) 2, 101-119,
7810 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7813 *) MIPS assembler BIGNUM division bug fix.
7816 *) Disabled incorrect Alpha assembler code.
7819 *) Fix PKCS#7 decode routines so they correctly update the length
7820 after reading an EOC for the EXPLICIT tag.
7822 [This change does not apply to 0.9.7.]
7824 *) Fix bug in PKCS#12 key generation routines. This was triggered
7825 if a 3DES key was generated with a 0 initial byte. Include
7826 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7827 (but broken) behaviour.
7830 *) Enhance bctest to search for a working bc along $PATH and print
7832 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7834 *) Fix memory leaks in err.c: free err_data string if necessary;
7835 don't write to the wrong index in ERR_set_error_data.
7838 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7842 *) Replace rdtsc with _emit statements for VC++ version 5.
7843 [Jeremy Cooper <jeremy@baymoo.org>]
7845 *) Make it possible to reuse SSLv2 sessions.
7848 *) In copy_email() check for >= 0 as a return value for
7849 X509_NAME_get_index_by_NID() since 0 is a valid index.
7850 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7852 *) Avoid coredump with unsupported or invalid public keys by checking if
7853 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7854 PKCS7_verify() fails with non detached data.
7857 *) Don't use getenv in library functions when run as setuid/setgid.
7858 New function OPENSSL_issetugid().
7861 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7862 due to incorrect handling of multi-threading:
7864 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7866 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7868 3. Count how many times MemCheck_off() has been called so that
7869 nested use can be treated correctly. This also avoids
7870 inband-signalling in the previous code (which relied on the
7871 assumption that thread ID 0 is impossible).
7874 *) Add "-rand" option also to s_client and s_server.
7877 *) Fix CPU detection on Irix 6.x.
7878 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7879 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7881 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7884 [This change does not apply to 0.9.7.]
7886 *) Use the cached encoding of an X509_NAME structure rather than
7887 copying it. This is apparently the reason for the libsafe "errors"
7888 but the code is actually correct.
7891 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7892 Bleichenbacher's DSA attack.
7893 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7894 to be set and top=0 forces the highest bit to be set; top=-1 is new
7895 and leaves the highest bit random.
7896 [Ulf Moeller, Bodo Moeller]
7898 *) In the NCONF_...-based implementations for CONF_... queries
7899 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7900 a temporary CONF structure with the data component set to NULL
7901 (which gives segmentation faults in lh_retrieve).
7902 Instead, use NULL for the CONF pointer in CONF_get_string and
7903 CONF_get_number (which may use environment variables) and directly
7904 return NULL from CONF_get_section.
7907 *) Fix potential buffer overrun for EBCDIC.
7910 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7911 keyUsage if basicConstraints absent for a CA.
7914 *) Make SMIME_write_PKCS7() write mail header values with a format that
7915 is more generally accepted (no spaces before the semicolon), since
7916 some programs can't parse those values properly otherwise. Also make
7917 sure BIO's that break lines after each write do not create invalid
7921 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7922 macros previously used would not encode an empty SEQUENCE OF
7923 and break the signature.
7925 [This change does not apply to 0.9.7.]
7927 *) Zero the premaster secret after deriving the master secret in
7931 *) Add some EVP_add_digest_alias registrations (as found in
7932 OpenSSL_add_all_digests()) to SSL_library_init()
7933 aka OpenSSL_add_ssl_algorithms(). This provides improved
7934 compatibility with peers using X.509 certificates
7935 with unconventional AlgorithmIdentifier OIDs.
7938 *) Fix for Irix with NO_ASM.
7939 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7941 *) ./config script fixes.
7942 [Ulf Moeller, Richard Levitte]
7944 *) Fix 'openssl passwd -1'.
7947 *) Change PKCS12_key_gen_asc() so it can cope with non null
7948 terminated strings whose length is passed in the passlen
7949 parameter, for example from PEM callbacks. This was done
7950 by adding an extra length parameter to asc2uni().
7951 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7953 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7954 call failed, free the DSA structure.
7957 *) Fix to uni2asc() to cope with zero length Unicode strings.
7958 These are present in some PKCS#12 files.
7961 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7962 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7963 when writing a 32767 byte record.
7964 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7966 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7967 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7969 (RSA objects have a reference count access to which is protected
7970 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7971 so they are meant to be shared between threads.)
7972 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7973 "Reddie, Steven" <Steven.Reddie@ca.com>]
7975 *) Fix a deadlock in CRYPTO_mem_leaks().
7978 *) Use better test patterns in bntest.
7981 *) rand_win.c fix for Borland C.
7984 *) BN_rshift bugfix for n == 0.
7987 *) Add a 'bctest' script that checks for some known 'bc' bugs
7988 so that 'make test' does not abort just because 'bc' is broken.
7991 *) Store verify_result within SSL_SESSION also for client side to
7992 avoid potential security hole. (Re-used sessions on the client side
7993 always resulted in verify_result==X509_V_OK, not using the original
7994 result of the server certificate verification.)
7997 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7998 SSL3_RT_APPLICATION_DATA, return 0.
7999 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8003 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8004 releases, have been re-implemented by renaming the previous
8005 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8006 and ssl3_read_internal, respectively, and adding 'peek' parameters
8007 to them. The new ssl[23]_{read,peek} functions are calls to
8008 ssl[23]_read_internal with the 'peek' flag set appropriately.
8009 A 'peek' parameter has also been added to ssl3_read_bytes, which
8010 does the actual work for ssl3_read_internal.
8013 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8014 the method-specific "init()" handler. Also clean up ex_data after
8015 calling the method-specific "finish()" handler. Previously, this was
8016 happening the other way round.
8019 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8020 The previous value, 12, was not always sufficient for BN_mod_exp().
8023 *) Make sure that shared libraries get the internal name engine with
8024 the full version number and not just 0. This should mark the
8025 shared libraries as not backward compatible. Of course, this should
8026 be changed again when we can guarantee backward binary compatibility.
8029 *) Fix typo in get_cert_by_subject() in by_dir.c
8030 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8032 *) Rework the system to generate shared libraries:
8034 - Make note of the expected extension for the shared libraries and
8035 if there is a need for symbolic links from for example libcrypto.so.0
8036 to libcrypto.so.0.9.7. There is extended info in Configure for
8039 - Make as few rebuilds of the shared libraries as possible.
8041 - Still avoid linking the OpenSSL programs with the shared libraries.
8043 - When installing, install the shared libraries separately from the
8047 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8049 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8050 and not in SSL_clear because the latter is also used by the
8051 accept/connect functions; previously, the settings made by
8052 SSL_set_read_ahead would be lost during the handshake.
8053 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8055 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8056 Previously, it would create entries for disabled algorithms no
8060 *) Added several new manual pages for SSL_* function.
8063 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8065 *) In ssl23_get_client_hello, generate an error message when faced
8066 with an initial SSL 3.0/TLS record that is too small to contain the
8067 first two bytes of the ClientHello message, i.e. client_version.
8068 (Note that this is a pathologic case that probably has never happened
8069 in real life.) The previous approach was to use the version number
8070 from the record header as a substitute; but our protocol choice
8071 should not depend on that one because it is not authenticated
8072 by the Finished messages.
8075 *) More robust randomness gathering functions for Windows.
8076 [Jeffrey Altman <jaltman@columbia.edu>]
8078 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8079 not set then we don't setup the error code for issuer check errors
8080 to avoid possibly overwriting other errors which the callback does
8081 handle. If an application does set the flag then we assume it knows
8082 what it is doing and can handle the new informational codes
8086 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8087 a general "ANY" type, as such it should be able to decode anything
8088 including tagged types. However it didn't check the class so it would
8089 wrongly interpret tagged types in the same way as their universal
8090 counterpart and unknown types were just rejected. Changed so that the
8091 tagged and unknown types are handled in the same way as a SEQUENCE:
8092 that is the encoding is stored intact. There is also a new type
8093 "V_ASN1_OTHER" which is used when the class is not universal, in this
8094 case we have no idea what the actual type is so we just lump them all
8098 *) On VMS, stdout may very well lead to a file that is written to
8099 in a record-oriented fashion. That means that every write() will
8100 write a separate record, which will be read separately by the
8101 programs trying to read from it. This can be very confusing.
8103 The solution is to put a BIO filter in the way that will buffer
8104 text until a linefeed is reached, and then write everything a
8105 line at a time, so every record written will be an actual line,
8106 not chunks of lines and not (usually doesn't happen, but I've
8107 seen it once) several lines in one record. BIO_f_linebuffer() is
8110 Currently, it's a VMS-only method, because that's where it has
8111 been tested well enough.
8114 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8115 it can return incorrect results.
8116 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8117 but it was in 0.9.6-beta[12].)
8120 *) Disable the check for content being present when verifying detached
8121 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8122 include zero length content when signing messages.
8125 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8126 BIO_ctrl (for BIO pairs).
8129 *) Add DSO method for VMS.
8132 *) Bug fix: Montgomery multiplication could produce results with the
8136 *) Add RPM specification openssl.spec and modify it to build three
8137 packages. The default package contains applications, application
8138 documentation and run-time libraries. The devel package contains
8139 include files, static libraries and function documentation. The
8140 doc package contains the contents of the doc directory. The original
8141 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8144 *) Add a large number of documentation files for many SSL routines.
8145 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8147 *) Add a configuration entry for Sony News 4.
8148 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8150 *) Don't set the two most significant bits to one when generating a
8151 random number < q in the DSA library.
8154 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8155 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8156 the underlying transport is blocking) if a handshake took place.
8157 (The default behaviour is needed by applications such as s_client
8158 and s_server that use select() to determine when to use SSL_read;
8159 but for applications that know in advance when to expect data, it
8160 just makes things more complicated.)
8163 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8167 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8168 work better on such systems.
8169 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8171 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8172 Update PKCS12_parse() so it copies the friendlyName and the
8173 keyid to the certificates aux info.
8176 *) Fix bug in PKCS7_verify() which caused an infinite loop
8177 if there was more than one signature.
8178 [Sven Uszpelkat <su@celocom.de>]
8180 *) Major change in util/mkdef.pl to include extra information
8181 about each symbol, as well as presenting variables as well
8182 as functions. This change means that there's n more need
8183 to rebuild the .num files when some algorithms are excluded.
8186 *) Allow the verify time to be set by an application,
8187 rather than always using the current time.
8190 *) Phase 2 verify code reorganisation. The certificate
8191 verify code now looks up an issuer certificate by a
8192 number of criteria: subject name, authority key id
8193 and key usage. It also verifies self signed certificates
8194 by the same criteria. The main comparison function is
8195 X509_check_issued() which performs these checks.
8197 Lot of changes were necessary in order to support this
8198 without completely rewriting the lookup code.
8200 Authority and subject key identifier are now cached.
8202 The LHASH 'certs' is X509_STORE has now been replaced
8203 by a STACK_OF(X509_OBJECT). This is mainly because an
8204 LHASH can't store or retrieve multiple objects with
8205 the same hash value.
8207 As a result various functions (which were all internal
8208 use only) have changed to handle the new X509_STORE
8209 structure. This will break anything that messed round
8210 with X509_STORE internally.
8212 The functions X509_STORE_add_cert() now checks for an
8213 exact match, rather than just subject name.
8215 The X509_STORE API doesn't directly support the retrieval
8216 of multiple certificates matching a given criteria, however
8217 this can be worked round by performing a lookup first
8218 (which will fill the cache with candidate certificates)
8219 and then examining the cache for matches. This is probably
8220 the best we can do without throwing out X509_LOOKUP
8221 entirely (maybe later...).
8223 The X509_VERIFY_CTX structure has been enhanced considerably.
8225 All certificate lookup operations now go via a get_issuer()
8226 callback. Although this currently uses an X509_STORE it
8227 can be replaced by custom lookups. This is a simple way
8228 to bypass the X509_STORE hackery necessary to make this
8229 work and makes it possible to use more efficient techniques
8230 in future. A very simple version which uses a simple
8231 STACK for its trusted certificate store is also provided
8232 using X509_STORE_CTX_trusted_stack().
8234 The verify_cb() and verify() callbacks now have equivalents
8235 in the X509_STORE_CTX structure.
8237 X509_STORE_CTX also has a 'flags' field which can be used
8238 to customise the verify behaviour.
8241 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8242 excludes S/MIME capabilities.
8245 *) When a certificate request is read in keep a copy of the
8246 original encoding of the signed data and use it when outputing
8247 again. Signatures then use the original encoding rather than
8248 a decoded, encoded version which may cause problems if the
8249 request is improperly encoded.
8252 *) For consistency with other BIO_puts implementations, call
8253 buffer_write(b, ...) directly in buffer_puts instead of calling
8256 In BIO_puts, increment b->num_write as in BIO_write.
8257 [Peter.Sylvester@EdelWeb.fr]
8259 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8260 BN_zero, we may not return a BIGNUM with an array consisting of
8264 *) Avoid calling abort() from within the library when problems are
8265 detected, except if preprocessor symbols have been defined
8266 (such as REF_CHECK, BN_DEBUG etc.).
8269 *) New openssl application 'rsautl'. This utility can be
8270 used for low level RSA operations. DER public key
8271 BIO/fp routines also added.
8274 *) New Configure entry and patches for compiling on QNX 4.
8275 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8277 *) A demo state-machine implementation was sponsored by
8278 Nuron (http://www.nuron.com/) and is now available in
8279 demos/state_machine.
8282 *) New options added to the 'dgst' utility for signature
8283 generation and verification.
8286 *) Unrecognized PKCS#7 content types are now handled via a
8287 catch all ASN1_TYPE structure. This allows unsupported
8288 types to be stored as a "blob" and an application can
8289 encode and decode it manually.
8292 *) Fix various signed/unsigned issues to make a_strex.c
8294 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8296 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8297 length if passed a buffer. ASN1_INTEGER_to_BN failed
8298 if passed a NULL BN and its argument was negative.
8299 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8301 *) Modification to PKCS#7 encoding routines to output definite
8302 length encoding. Since currently the whole structures are in
8303 memory there's not real point in using indefinite length
8304 constructed encoding. However if OpenSSL is compiled with
8305 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8308 *) Added BIO_vprintf() and BIO_vsnprintf().
8311 *) Added more prefixes to parse for in the the strings written
8312 through a logging bio, to cover all the levels that are available
8313 through syslog. The prefixes are now:
8315 PANIC, EMERG, EMR => LOG_EMERG
8316 ALERT, ALR => LOG_ALERT
8317 CRIT, CRI => LOG_CRIT
8318 ERROR, ERR => LOG_ERR
8319 WARNING, WARN, WAR => LOG_WARNING
8320 NOTICE, NOTE, NOT => LOG_NOTICE
8321 INFO, INF => LOG_INFO
8322 DEBUG, DBG => LOG_DEBUG
8324 and as before, if none of those prefixes are present at the
8325 beginning of the string, LOG_ERR is chosen.
8327 On Win32, the LOG_* levels are mapped according to this:
8329 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8330 LOG_WARNING => EVENTLOG_WARNING_TYPE
8331 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8335 *) Made it possible to reconfigure with just the configuration
8336 argument "reconf" or "reconfigure". The command line arguments
8337 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8338 and are retrieved from there when reconfiguring.
8342 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8344 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8347 *) The obj_dat.pl script was messing up the sorting of object
8348 names. The reason was that it compared the quoted version
8349 of strings as a result "OCSP" > "OCSP Signing" because
8350 " > SPACE. Changed script to store unquoted versions of
8351 names and add quotes on output. It was also omitting some
8352 names from the lookup table if they were given a default
8353 value (that is if SN is missing it is given the same
8354 value as LN and vice versa), these are now added on the
8355 grounds that if an object has a name we should be able to
8356 look it up. Finally added warning output when duplicate
8357 short or long names are found.
8360 *) Changes needed for Tandem NSK.
8361 [Scott Uroff <scott@xypro.com>]
8363 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8364 RSA_padding_check_SSLv23(), special padding was never detected
8365 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8366 version rollback attacks was not effective.
8368 In s23_clnt.c, don't use special rollback-attack detection padding
8369 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8370 client; similarly, in s23_srvr.c, don't do the rollback check if
8371 SSL 2.0 is the only protocol enabled in the server.
8374 *) Make it possible to get hexdumps of unprintable data with 'openssl
8375 asn1parse'. By implication, the functions ASN1_parse_dump() and
8376 BIO_dump_indent() are added.
8379 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8380 these print out strings and name structures based on various
8381 flags including RFC2253 support and proper handling of
8382 multibyte characters. Added options to the 'x509' utility
8383 to allow the various flags to be set.
8386 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8387 Also change the functions X509_cmp_current_time() and
8388 X509_gmtime_adj() work with an ASN1_TIME structure,
8389 this will enable certificates using GeneralizedTime in validity
8390 dates to be checked.
8393 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8394 negative public key encodings) on by default,
8395 NO_NEG_PUBKEY_BUG can be set to disable it.
8398 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8399 content octets. An i2c_ASN1_OBJECT is unnecessary because
8400 the encoding can be trivially obtained from the structure.
8403 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8404 not read locks (CRYPTO_r_[un]lock).
8407 *) A first attempt at creating official support for shared
8408 libraries through configuration. I've kept it so the
8409 default is static libraries only, and the OpenSSL programs
8410 are always statically linked for now, but there are
8411 preparations for dynamic linking in place.
8412 This has been tested on Linux and Tru64.
8415 *) Randomness polling function for Win9x, as described in:
8416 Peter Gutmann, Software Generation of Practically Strong
8420 *) Fix so PRNG is seeded in req if using an already existing
8424 *) New options to smime application. -inform and -outform
8425 allow alternative formats for the S/MIME message including
8426 PEM and DER. The -content option allows the content to be
8427 specified separately. This should allow things like Netscape
8428 form signing output easier to verify.
8431 *) Fix the ASN1 encoding of tags using the 'long form'.
8434 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8435 STRING types. These convert content octets to and from the
8436 underlying type. The actual tag and length octets are
8437 already assumed to have been read in and checked. These
8438 are needed because all other string types have virtually
8439 identical handling apart from the tag. By having versions
8440 of the ASN1 functions that just operate on content octets
8441 IMPLICIT tagging can be handled properly. It also allows
8442 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8443 and ASN1_INTEGER are identical apart from the tag.
8446 *) Change the handling of OID objects as follows:
8448 - New object identifiers are inserted in objects.txt, following
8449 the syntax given in objects.README.
8450 - objects.pl is used to process obj_mac.num and create a new
8452 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8455 This is currently kind of a hack, and the perl code in objects.pl
8456 isn't very elegant, but it works as I intended. The simplest way
8457 to check that it worked correctly is to look in obj_dat.h and
8458 check the array nid_objs and make sure the objects haven't moved
8459 around (this is important!). Additions are OK, as well as
8460 consistent name changes.
8463 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8466 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8467 The given file adds to whatever has already been seeded into the
8468 random pool through the RANDFILE configuration file option or
8469 environment variable, or the default random state file.
8472 *) mkstack.pl now sorts each macro group into lexical order.
8473 Previously the output order depended on the order the files
8474 appeared in the directory, resulting in needless rewriting
8478 *) Patches to make OpenSSL compile under Win32 again. Mostly
8479 work arounds for the VC++ problem that it treats func() as
8480 func(void). Also stripped out the parts of mkdef.pl that
8481 added extra typesafe functions: these no longer exist.
8484 *) Reorganisation of the stack code. The macros are now all
8485 collected in safestack.h . Each macro is defined in terms of
8486 a "stack macro" of the form SKM_<name>(type, a, b). The
8487 DEBUG_SAFESTACK is now handled in terms of function casts,
8488 this has the advantage of retaining type safety without the
8489 use of additional functions. If DEBUG_SAFESTACK is not defined
8490 then the non typesafe macros are used instead. Also modified the
8491 mkstack.pl script to handle the new form. Needs testing to see
8492 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8493 the default if no major problems. Similar behaviour for ASN1_SET_OF
8494 and PKCS12_STACK_OF.
8497 *) When some versions of IIS use the 'NET' form of private key the
8498 key derivation algorithm is different. Normally MD5(password) is
8499 used as a 128 bit RC4 key. In the modified case
8500 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8501 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8502 as the old Netscape_RSA functions except they have an additional
8503 'sgckey' parameter which uses the modified algorithm. Also added
8504 an -sgckey command line option to the rsa utility. Thanks to
8505 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8506 algorithm to openssl-dev.
8509 *) The evp_local.h macros were using 'c.##kname' which resulted in
8510 invalid expansion on some systems (SCO 5.0.5 for example).
8511 Corrected to 'c.kname'.
8512 [Phillip Porch <root@theporch.com>]
8514 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8515 a STACK of email addresses from a certificate or request, these look
8516 in the subject name and the subject alternative name extensions and
8517 omit any duplicate addresses.
8520 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8521 This makes DSA verification about 2 % faster.
8524 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8525 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8526 plus overhead for 1024 bit moduli).
8527 This makes exponentiations about 0.5 % faster for 1024 bit
8528 exponents (as measured by "openssl speed rsa2048").
8531 *) Rename memory handling macros to avoid conflicts with other
8533 Malloc => OPENSSL_malloc
8534 Malloc_locked => OPENSSL_malloc_locked
8535 Realloc => OPENSSL_realloc
8536 Free => OPENSSL_free
8539 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8540 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8543 *) CygWin32 support.
8544 [John Jarvie <jjarvie@newsguy.com>]
8546 *) The type-safe stack code has been rejigged. It is now only compiled
8547 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8548 by default all type-specific stack functions are "#define"d back to
8549 standard stack functions. This results in more streamlined output
8550 but retains the type-safety checking possibilities of the original
8554 *) The STACK code has been cleaned up, and certain type declarations
8555 that didn't make a lot of sense have been brought in line. This has
8556 also involved a cleanup of sorts in safestack.h to more correctly
8557 map type-safe stack functions onto their plain stack counterparts.
8558 This work has also resulted in a variety of "const"ifications of
8559 lots of the code, especially "_cmp" operations which should normally
8560 be prototyped with "const" parameters anyway.
8563 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8564 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8565 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8566 where all of 'md' is used each time the PRNG is used, but 'state'
8567 is used only indexed by a cyclic counter. As entropy may not be
8568 well distributed from the beginning, 'md' is important as a
8569 chaining variable. However, the output function chains only half
8570 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8571 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8572 in all of 'state' being rewritten, with the new values depending
8573 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8576 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8577 the handshake is continued after ssl_verify_cert_chain();
8578 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8579 can lead to 'unexplainable' connection aborts later.
8580 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8582 *) Major EVP API cipher revision.
8583 Add hooks for extra EVP features. This allows various cipher
8584 parameters to be set in the EVP interface. Support added for variable
8585 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8586 setting of RC2 and RC5 parameters.
8588 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8591 Remove lots of duplicated code from the EVP library. For example *every*
8592 cipher init() function handles the 'iv' in the same way according to the
8593 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8594 for CFB and OFB modes they zero ctx->num.
8596 New functionality allows removal of S/MIME code RC2 hack.
8598 Most of the routines have the same form and so can be declared in terms
8601 By shifting this to the top level EVP_CipherInit() it can be removed from
8602 all individual ciphers. If the cipher wants to handle IVs or keys
8603 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8606 Change lots of functions like EVP_EncryptUpdate() to now return a
8607 value: although software versions of the algorithms cannot fail
8608 any installed hardware versions can.
8611 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8612 this option is set, tolerate broken clients that send the negotiated
8613 protocol version number instead of the requested protocol version
8617 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8618 i.e. non-zero for export ciphersuites, zero otherwise.
8619 Previous versions had this flag inverted, inconsistent with
8620 rsa_tmp_cb (..._TMP_RSA_CB).
8621 [Bodo Moeller; problem reported by Amit Chopra]
8623 *) Add missing DSA library text string. Work around for some IIS
8624 key files with invalid SEQUENCE encoding.
8627 *) Add a document (doc/standards.txt) that list all kinds of standards
8628 and so on that are implemented in OpenSSL.
8631 *) Enhance c_rehash script. Old version would mishandle certificates
8632 with the same subject name hash and wouldn't handle CRLs at all.
8633 Added -fingerprint option to crl utility, to support new c_rehash
8637 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8640 *) Fix for SSL server purpose checking. Server checking was
8641 rejecting certificates which had extended key usage present
8642 but no ssl client purpose.
8643 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8645 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8646 is a little unclear about how a blank password is handled.
8647 Since the password in encoded as a BMPString with terminating
8648 double NULL a zero length password would end up as just the
8649 double NULL. However no password at all is different and is
8650 handled differently in the PKCS#12 key generation code. NS
8651 treats a blank password as zero length. MSIE treats it as no
8652 password on export: but it will try both on import. We now do
8653 the same: PKCS12_parse() tries zero length and no password if
8654 the password is set to "" or NULL (NULL is now a valid password:
8655 it wasn't before) as does the pkcs12 application.
8658 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8659 perror when PEM_read_bio_X509_REQ fails, the error message must
8660 be obtained from the error queue.
8663 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8664 it in ERR_remove_state if appropriate, and change ERR_get_state
8665 accordingly to avoid race conditions (this is necessary because
8666 thread_hash is no longer constant once set).
8669 *) Bugfix for linux-elf makefile.one.
8672 *) RSA_get_default_method() will now cause a default
8673 RSA_METHOD to be chosen if one doesn't exist already.
8674 Previously this was only set during a call to RSA_new()
8675 or RSA_new_method(NULL) meaning it was possible for
8676 RSA_get_default_method() to return NULL.
8679 *) Added native name translation to the existing DSO code
8680 that will convert (if the flag to do so is set) filenames
8681 that are sufficiently small and have no path information
8682 into a canonical native form. Eg. "blah" converted to
8683 "libblah.so" or "blah.dll" etc.
8686 *) New function ERR_error_string_n(e, buf, len) which is like
8687 ERR_error_string(e, buf), but writes at most 'len' bytes
8688 including the 0 terminator. For ERR_error_string_n, 'buf'
8690 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8692 *) CONF library reworked to become more general. A new CONF
8693 configuration file reader "class" is implemented as well as a
8694 new functions (NCONF_*, for "New CONF") to handle it. The now
8695 old CONF_* functions are still there, but are reimplemented to
8696 work in terms of the new functions. Also, a set of functions
8697 to handle the internal storage of the configuration data is
8698 provided to make it easier to write new configuration file
8699 reader "classes" (I can definitely see something reading a
8700 configuration file in XML format, for example), called _CONF_*,
8701 or "the configuration storage API"...
8703 The new configuration file reading functions are:
8705 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8706 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8708 NCONF_default, NCONF_WIN32
8710 NCONF_dump_fp, NCONF_dump_bio
8712 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8713 NCONF_new creates a new CONF object. This works in the same way
8714 as other interfaces in OpenSSL, like the BIO interface.
8715 NCONF_dump_* dump the internal storage of the configuration file,
8716 which is useful for debugging. All other functions take the same
8717 arguments as the old CONF_* functions wth the exception of the
8718 first that must be a `CONF *' instead of a `LHASH *'.
8720 To make it easer to use the new classes with the old CONF_* functions,
8721 the function CONF_set_default_method is provided.
8724 *) Add '-tls1' option to 'openssl ciphers', which was already
8725 mentioned in the documentation but had not been implemented.
8726 (This option is not yet really useful because even the additional
8727 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8730 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8731 OpenSSL-based applications) load shared libraries and bind to
8732 them in a portable way.
8733 [Geoff Thorpe, with contributions from Richard Levitte]
8735 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8737 *) Make sure _lrotl and _lrotr are only used with MSVC.
8739 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8740 (the default implementation of RAND_status).
8742 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8743 to '-clrext' (= clear extensions), as intended and documented.
8744 [Bodo Moeller; inconsistency pointed out by Michael Attili
8747 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8748 was larger than the MD block size.
8749 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8751 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8752 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8753 using the passed key: if the passed key was a private key the result
8754 of X509_print(), for example, would be to print out all the private key
8758 *) des_quad_cksum() byte order bug fix.
8759 [Ulf Möller, using the problem description in krb4-0.9.7, where
8760 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8762 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8764 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8766 *) For easily testing in shell scripts whether some command
8767 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8768 returns with exit code 0 iff no command of the given name is available.
8769 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8770 the output goes to stdout and nothing is printed to stderr.
8771 Additional arguments are always ignored.
8773 Since for each cipher there is a command of the same name,
8774 the 'no-cipher' compilation switches can be tested this way.
8776 ('openssl no-XXX' is not able to detect pseudo-commands such
8777 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8780 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8783 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8784 is set; it will be thrown away anyway because each handshake creates
8786 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8787 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8788 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8789 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8792 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8793 'Q' and 'R' lose their special meanings (quit/renegotiate).
8794 This is part of what -quiet does; unlike -quiet, -ign_eof
8795 does not suppress any output.
8798 *) Add compatibility options to the purpose and trust code. The
8799 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8800 accepts a certificate or CA, this was the previous behaviour,
8801 with all the associated security issues.
8803 X509_TRUST_COMPAT is the old trust behaviour: only and
8804 automatically trust self signed roots in certificate store. A
8805 new trust setting X509_TRUST_DEFAULT is used to specify that
8806 a purpose has no associated trust setting and it should instead
8807 use the value in the default purpose.
8810 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8811 and fix a memory leak.
8814 *) In util/mkerr.pl (which implements 'make errors'), preserve
8815 reason strings from the previous version of the .c file, as
8816 the default to have only downcase letters (and digits) in
8817 automatically generated reasons codes is not always appropriate.
8820 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8821 using strerror. Previously, ERR_reason_error_string() returned
8822 library names as reason strings for SYSerr; but SYSerr is a special
8823 case where small numbers are errno values, not library numbers.
8826 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8827 converts DSA parameters into DH parameters. (When creating parameters,
8828 DSA_generate_parameters is used.)
8831 *) Include 'length' (recommended exponent length) in C code generated
8832 by 'openssl dhparam -C'.
8835 *) The second argument to set_label in perlasm was already being used
8836 so couldn't be used as a "file scope" flag. Moved to third argument
8840 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8841 instead of RAND_bytes for encryption IVs and salts.
8844 *) Include RAND_status() into RAND_METHOD instead of implementing
8845 it only for md_rand.c Otherwise replacing the PRNG by calling
8846 RAND_set_rand_method would be impossible.
8849 *) Don't let DSA_generate_key() enter an infinite loop if the random
8850 number generation fails.
8853 *) New 'rand' application for creating pseudo-random output.
8856 *) Added configuration support for Linux/IA64
8857 [Rolf Haberrecker <rolf@suse.de>]
8859 *) Assembler module support for Mingw32.
8862 *) Shared library support for HPUX (in shlib/).
8863 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8865 *) Shared library support for Solaris gcc.
8866 [Lutz Behnke <behnke@trustcenter.de>]
8868 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8870 *) PKCS7_encrypt() was adding text MIME headers twice because they
8871 were added manually and by SMIME_crlf_copy().
8874 *) In bntest.c don't call BN_rand with zero bits argument.
8875 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8877 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8878 case was implemented. This caused BN_div_recp() to fail occasionally.
8881 *) Add an optional second argument to the set_label() in the perl
8882 assembly language builder. If this argument exists and is set
8883 to 1 it signals that the assembler should use a symbol whose
8884 scope is the entire file, not just the current function. This
8885 is needed with MASM which uses the format label:: for this scope.
8886 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8888 *) Change the ASN1 types so they are typedefs by default. Before
8889 almost all types were #define'd to ASN1_STRING which was causing
8890 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8894 *) Change names of new functions to the new get1/get0 naming
8895 convention: After 'get1', the caller owns a reference count
8896 and has to call ..._free; 'get0' returns a pointer to some
8897 data structure without incrementing reference counters.
8898 (Some of the existing 'get' functions increment a reference
8899 counter, some don't.)
8900 Similarly, 'set1' and 'add1' functions increase reference
8901 counters or duplicate objects.
8904 *) Allow for the possibility of temp RSA key generation failure:
8905 the code used to assume it always worked and crashed on failure.
8908 *) Fix potential buffer overrun problem in BIO_printf().
8909 [Ulf Möller, using public domain code by Patrick Powell; problem
8910 pointed out by David Sacerdote <das33@cornell.edu>]
8912 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8913 RAND_egd() and RAND_status(). In the command line application,
8914 the EGD socket can be specified like a seed file using RANDFILE
8918 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8919 Some CAs (e.g. Verisign) distribute certificates in this form.
8922 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8923 list to exclude them. This means that no special compilation option
8924 is needed to use anonymous DH: it just needs to be included in the
8928 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8929 EVP_MD_type. The old functionality is available in a new macro called
8930 EVP_MD_md(). Change code that uses it and update docs.
8933 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8934 where the 'void *' argument is replaced by a function pointer argument.
8935 Previously 'void *' was abused to point to functions, which works on
8936 many platforms, but is not correct. As these functions are usually
8937 called by macros defined in OpenSSL header files, most source code
8938 should work without changes.
8941 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8942 sections with information on -D... compiler switches used for
8943 compiling the library so that applications can see them. To enable
8944 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8945 must be defined. E.g.,
8946 #define OPENSSL_ALGORITHM_DEFINES
8947 #include <openssl/opensslconf.h>
8948 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8949 [Richard Levitte, Ulf and Bodo Möller]
8951 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8955 *) Change the 'other' type in certificate aux info to a STACK_OF
8956 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8957 the required ASN1 format: arbitrary types determined by an OID.
8960 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8961 argument to 'req'. This is not because the function is newer or
8962 better than others it just uses the work 'NEW' in the certificate
8963 request header lines. Some software needs this.
8966 *) Reorganise password command line arguments: now passwords can be
8967 obtained from various sources. Delete the PEM_cb function and make
8968 it the default behaviour: i.e. if the callback is NULL and the
8969 usrdata argument is not NULL interpret it as a null terminated pass
8970 phrase. If usrdata and the callback are NULL then the pass phrase
8971 is prompted for as usual.
8974 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8975 the support is automatically enabled. The resulting binaries will
8976 autodetect the card and use it if present.
8977 [Ben Laurie and Compaq Inc.]
8979 *) Work around for Netscape hang bug. This sends certificate request
8980 and server done in one record. Since this is perfectly legal in the
8981 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8982 the bugs/SSLv3 entry for more info.
8985 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8988 *) Add -rand argument to smime and pkcs12 applications and read/write
8992 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8995 *) Add command line password options to the remaining applications.
8998 *) Bug fix for BN_div_recp() for numerators with an even number of
9002 *) More tests in bntest.c, and changed test_bn output.
9005 *) ./config recognizes MacOS X now.
9008 *) Bug fix for BN_div() when the first words of num and divsor are
9009 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9012 *) Add support for various broken PKCS#8 formats, and command line
9013 options to produce them.
9016 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9017 get temporary BIGNUMs from a BN_CTX.
9020 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9024 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9025 include a #define from the old name to the new. The original intent
9026 was that statically linked binaries could for example just call
9027 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9028 link with digests. This never worked because SSLeay_add_all_digests()
9029 and SSLeay_add_all_ciphers() were in the same source file so calling
9030 one would link with the other. They are now in separate source files.
9033 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9036 *) Use a less unusual form of the Miller-Rabin primality test (it used
9037 a binary algorithm for exponentiation integrated into the Miller-Rabin
9038 loop, our standard modexp algorithms are faster).
9041 *) Support for the EBCDIC character set completed.
9042 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9044 *) Source code cleanups: use const where appropriate, eliminate casts,
9045 use void * instead of char * in lhash.
9048 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9049 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9050 this the server could overwrite ephemeral keys that the client
9054 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9055 using 50 iterations of the Rabin-Miller test.
9057 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9058 iterations of the Rabin-Miller test as required by the appendix
9059 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9060 As BN_is_prime_fasttest includes trial division, DSA parameter
9061 generation becomes much faster.
9063 This implies a change for the callback functions in DSA_is_prime
9064 and DSA_generate_parameters: The callback function is called once
9065 for each positive witness in the Rabin-Miller test, not just
9066 occasionally in the inner loop; and the parameters to the
9067 callback function now provide an iteration count for the outer
9068 loop rather than for the current invocation of the inner loop.
9069 DSA_generate_parameters additionally can call the callback
9070 function with an 'iteration count' of -1, meaning that a
9071 candidate has passed the trial division test (when q is generated
9072 from an application-provided seed, trial division is skipped).
9075 *) New function BN_is_prime_fasttest that optionally does trial
9076 division before starting the Rabin-Miller test and has
9077 an additional BN_CTX * argument (whereas BN_is_prime always
9078 has to allocate at least one BN_CTX).
9079 'callback(1, -1, cb_arg)' is called when a number has passed the
9080 trial division stage.
9083 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9087 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9090 *) New function BN_pseudo_rand().
9093 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9094 bignum version of BN_from_montgomery() with the working code from
9095 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9099 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9100 made it impossible to use the same SSL_SESSION data structure in
9101 SSL2 clients in multiple threads.
9104 *) The return value of RAND_load_file() no longer counts bytes obtained
9105 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9106 to seed the PRNG (previously an explicit byte count was required).
9107 [Ulf Möller, Bodo Möller]
9109 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9110 used (char *) instead of (void *) and had casts all over the place.
9113 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9116 *) Retain source code compatibility for BN_prime_checks macro:
9117 BN_is_prime(..., BN_prime_checks, ...) now uses
9118 BN_prime_checks_for_size to determine the appropriate number of
9119 Rabin-Miller iterations.
9122 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9123 DH_CHECK_P_NOT_SAFE_PRIME.
9124 (Check if this is true? OpenPGP calls them "strong".)
9127 *) Merge the functionality of "dh" and "gendh" programs into a new program
9128 "dhparam". The old programs are retained for now but will handle DH keys
9129 (instead of parameters) in future.
9132 *) Make the ciphers, s_server and s_client programs check the return values
9133 when a new cipher list is set.
9136 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9137 ciphers. Before when the 56bit ciphers were enabled the sorting was
9140 The syntax for the cipher sorting has been extended to support sorting by
9141 cipher-strength (using the strength_bits hard coded in the tables).
9142 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9144 Fix a bug in the cipher-command parser: when supplying a cipher command
9145 string with an "undefined" symbol (neither command nor alphanumeric
9146 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9147 an error is flagged.
9149 Due to the strength-sorting extension, the code of the
9150 ssl_create_cipher_list() function was completely rearranged. I hope that
9151 the readability was also increased :-)
9152 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9154 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9155 for the first serial number and places 2 in the serial number file. This
9156 avoids problems when the root CA is created with serial number zero and
9157 the first user certificate has the same issuer name and serial number
9161 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9162 the new code. Add documentation for this stuff.
9165 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9166 X509_*() to X509at_*() on the grounds that they don't handle X509
9167 structures and behave in an analogous way to the X509v3 functions:
9168 they shouldn't be called directly but wrapper functions should be used
9171 So we also now have some wrapper functions that call the X509at functions
9172 when passed certificate requests. (TO DO: similar things can be done with
9173 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9174 things. Some of these need some d2i or i2d and print functionality
9175 because they handle more complex structures.)
9178 *) Add missing #ifndefs that caused missing symbols when building libssl
9179 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9180 NO_RSA in ssl/s2*.c.
9181 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9183 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9184 has a return value which indicates the quality of the random data
9185 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9186 error queue. New function RAND_pseudo_bytes() generates output that is
9187 guaranteed to be unique but not unpredictable. RAND_add is like
9188 RAND_seed, but takes an extra argument for an entropy estimate
9189 (RAND_seed always assumes full entropy).
9192 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9193 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9194 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9195 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9196 false-positive rate of at most 2^-80 for random input.
9199 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9202 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9203 in the 0.9.5 release), this returns the chain
9204 from an X509_CTX structure with a dup of the stack and all
9205 the X509 reference counts upped: so the stack will exist
9206 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9209 Also make SSL_SESSION_print() print out the verify return
9213 *) Add manpage for the pkcs12 command. Also change the default
9214 behaviour so MAC iteration counts are used unless the new
9215 -nomaciter option is used. This improves file security and
9216 only older versions of MSIE (4.0 for example) need it.
9219 *) Honor the no-xxx Configure options when creating .DEF files.
9222 *) Add PKCS#10 attributes to field table: challengePassword,
9223 unstructuredName and unstructuredAddress. These are taken from
9224 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9225 international characters are used.
9227 More changes to X509_ATTRIBUTE code: allow the setting of types
9228 based on strings. Remove the 'loc' parameter when adding
9229 attributes because these will be a SET OF encoding which is sorted
9233 *) Initial changes to the 'req' utility to allow request generation
9234 automation. This will allow an application to just generate a template
9235 file containing all the field values and have req construct the
9238 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9239 used all over the place including certificate requests and PKCS#7
9240 structures. They are currently handled manually where necessary with
9241 some primitive wrappers for PKCS#7. The new functions behave in a
9242 manner analogous to the X509 extension functions: they allow
9243 attributes to be looked up by NID and added.
9245 Later something similar to the X509V3 code would be desirable to
9246 automatically handle the encoding, decoding and printing of the
9247 more complex types. The string types like challengePassword can
9248 be handled by the string table functions.
9250 Also modified the multi byte string table handling. Now there is
9251 a 'global mask' which masks out certain types. The table itself
9252 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9253 is useful when for example there is only one permissible type
9254 (as in countryName) and using the mask might result in no valid
9258 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9259 SSL_get_peer_finished to allow applications to obtain the latest
9260 Finished messages sent to the peer or expected from the peer,
9261 respectively. (SSL_get_peer_finished is usually the Finished message
9262 actually received from the peer, otherwise the protocol will be aborted.)
9264 As the Finished message are message digests of the complete handshake
9265 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9266 be used for external authentication procedures when the authentication
9267 provided by SSL/TLS is not desired or is not enough.
9270 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9271 the host supports BWX extension and if Compaq C is present on the
9272 $PATH. Just exploiting of the BWX extension results in 20-30%
9273 performance kick for some algorithms, e.g. DES and RC4 to mention
9274 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9278 *) Add support for MS "fast SGC". This is arguably a violation of the
9279 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9280 weak crypto and after checking the certificate is SGC a second one
9281 with strong crypto. MS SGC stops the first handshake after receiving
9282 the server certificate message and sends a second client hello. Since
9283 a server will typically do all the time consuming operations before
9284 expecting any further messages from the client (server key exchange
9285 is the most expensive) there is little difference between the two.
9287 To get OpenSSL to support MS SGC we have to permit a second client
9288 hello message after we have sent server done. In addition we have to
9289 reset the MAC if we do get this second client hello.
9292 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9293 if a DER encoded private key is RSA or DSA traditional format. Changed
9294 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9295 format DER encoded private key. Newer code should use PKCS#8 format which
9296 has the key type encoded in the ASN1 structure. Added DER private key
9297 support to pkcs8 application.
9300 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9301 ciphersuites has been selected (as required by the SSL 3/TLS 1
9302 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9303 is set, we interpret this as a request to violate the specification
9304 (the worst that can happen is a handshake failure, and 'correct'
9305 behaviour would result in a handshake failure anyway).
9308 *) In SSL_CTX_add_session, take into account that there might be multiple
9309 SSL_SESSION structures with the same session ID (e.g. when two threads
9310 concurrently obtain them from an external cache).
9311 The internal cache can handle only one SSL_SESSION with a given ID,
9312 so if there's a conflict, we now throw out the old one to achieve
9316 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9317 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9318 some routines that use cipher OIDs: some ciphers do not have OIDs
9319 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9323 *) Simplify the trust setting structure and code. Now we just have
9324 two sequences of OIDs for trusted and rejected settings. These will
9325 typically have values the same as the extended key usage extension
9326 and any application specific purposes.
9328 The trust checking code now has a default behaviour: it will just
9329 check for an object with the same NID as the passed id. Functions can
9330 be provided to override either the default behaviour or the behaviour
9331 for a given id. SSL client, server and email already have functions
9332 in place for compatibility: they check the NID and also return "trusted"
9333 if the certificate is self signed.
9336 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9337 traditional format into an EVP_PKEY structure.
9340 *) Add a password callback function PEM_cb() which either prompts for
9341 a password if usr_data is NULL or otherwise assumes it is a null
9342 terminated password. Allow passwords to be passed on command line
9343 environment or config files in a few more utilities.
9346 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9347 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9348 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9349 Update documentation.
9352 *) Support for ASN1 "NULL" type. This could be handled before by using
9353 ASN1_TYPE but there wasn't any function that would try to read a NULL
9354 and produce an error if it couldn't. For compatibility we also have
9355 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9356 don't allocate anything because they don't need to.
9359 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9361 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9363 *) Rebuild of the memory allocation routines used by OpenSSL code and
9364 possibly others as well. The purpose is to make an interface that
9365 provide hooks so anyone can build a separate set of allocation and
9366 deallocation routines to be used by OpenSSL, for example memory
9367 pool implementations, or something else, which was previously hard
9368 since Malloc(), Realloc() and Free() were defined as macros having
9369 the values malloc, realloc and free, respectively (except for Win32
9370 compilations). The same is provided for memory debugging code.
9371 OpenSSL already comes with functionality to find memory leaks, but
9372 this gives people a chance to debug other memory problems.
9374 With these changes, a new set of functions and macros have appeared:
9376 CRYPTO_set_mem_debug_functions() [F]
9377 CRYPTO_get_mem_debug_functions() [F]
9378 CRYPTO_dbg_set_options() [F]
9379 CRYPTO_dbg_get_options() [F]
9380 CRYPTO_malloc_debug_init() [M]
9382 The memory debug functions are NULL by default, unless the library
9383 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9384 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9385 gives the standard debugging functions that come with OpenSSL) or
9386 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9387 provided by the library user) must be used. When the standard
9388 debugging functions are used, CRYPTO_dbg_set_options can be used to
9389 request additional information:
9390 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9391 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9393 Also, things like CRYPTO_set_mem_functions will always give the
9394 expected result (the new set of functions is used for allocation
9395 and deallocation) at all times, regardless of platform and compiler
9398 To finish it up, some functions that were never use in any other
9399 way than through macros have a new API and new semantic:
9402 CRYPTO_dbg_realloc()
9405 All macros of value have retained their old syntax.
9406 [Richard Levitte and Bodo Moeller]
9408 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9409 ordering of SMIMECapabilities wasn't in "strength order" and there
9410 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9414 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9415 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9416 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9418 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9419 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9420 functionality to handle multipart/signed properly) and a utility
9421 called 'smime' to call all this stuff. This is based on code I
9422 originally wrote for Celo who have kindly allowed it to be
9423 included in OpenSSL.
9426 *) Add variants des_set_key_checked and des_set_key_unchecked of
9427 des_set_key (aka des_key_sched). Global variable des_check_key
9428 decides which of these is called by des_set_key; this way
9429 des_check_key behaves as it always did, but applications and
9430 the library itself, which was buggy for des_check_key == 1,
9431 have a cleaner way to pick the version they need.
9434 *) New function PKCS12_newpass() which changes the password of a
9438 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9439 dynamic mix. In both cases the ids can be used as an index into the
9440 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9441 functions so they accept a list of the field values and the
9442 application doesn't need to directly manipulate the X509_TRUST
9446 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9450 *) Modify the way the V3 extension code looks up extensions. This now
9451 works in a similar way to the object code: we have some "standard"
9452 extensions in a static table which is searched with OBJ_bsearch()
9453 and the application can add dynamic ones if needed. The file
9454 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9455 updated whenever a new extension is added to the core code and kept
9456 in ext_nid order. There is a simple program 'tabtest.c' which checks
9457 this. New extensions are not added too often so this file can readily
9458 be maintained manually.
9460 There are two big advantages in doing things this way. The extensions
9461 can be looked up immediately and no longer need to be "added" using
9462 X509V3_add_standard_extensions(): this function now does nothing.
9463 [Side note: I get *lots* of email saying the extension code doesn't
9464 work because people forget to call this function]
9465 Also no dynamic allocation is done unless new extensions are added:
9466 so if we don't add custom extensions there is no need to call
9467 X509V3_EXT_cleanup().
9470 *) Modify enc utility's salting as follows: make salting the default. Add a
9471 magic header, so unsalted files fail gracefully instead of just decrypting
9472 to garbage. This is because not salting is a big security hole, so people
9473 should be discouraged from doing it.
9476 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9477 digest to be passed on the command line but it only used this
9478 parameter when signing a certificate. Modified so all relevant
9479 operations are affected by the digest parameter including the
9480 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9481 DSA key was used because it didn't fix the digest.
9484 *) Initial certificate chain verify code. Currently tests the untrusted
9485 certificates for consistency with the verify purpose (which is set
9486 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9488 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9489 this is because it will reject chains with invalid extensions whereas
9490 every previous version of OpenSSL and SSLeay made no checks at all.
9492 Trust code: checks the root CA for the relevant trust settings. Trust
9493 settings have an initial value consistent with the verify purpose: e.g.
9494 if the verify purpose is for SSL client use it expects the CA to be
9495 trusted for SSL client use. However the default value can be changed to
9496 permit custom trust settings: one example of this would be to only trust
9497 certificates from a specific "secure" set of CAs.
9499 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9500 which should be used for version portability: especially since the
9501 verify structure is likely to change more often now.
9503 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9504 to set them. If not set then assume SSL clients will verify SSL servers
9507 Two new options to the verify program: -untrusted allows a set of
9508 untrusted certificates to be passed in and -purpose which sets the
9509 intended purpose of the certificate. If a purpose is set then the
9510 new chain verify code is used to check extension consistency.
9513 *) Support for the authority information access extension.
9516 *) Modify RSA and DSA PEM read routines to transparently handle
9517 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9518 public keys in a format compatible with certificate
9519 SubjectPublicKeyInfo structures. Unfortunately there were already
9520 functions called *_PublicKey_* which used various odd formats so
9521 these are retained for compatibility: however the DSA variants were
9522 never in a public release so they have been deleted. Changed dsa/rsa
9523 utilities to handle the new format: note no releases ever handled public
9524 keys so we should be OK.
9526 The primary motivation for this change is to avoid the same fiasco
9527 that dogs private keys: there are several incompatible private key
9528 formats some of which are standard and some OpenSSL specific and
9529 require various evil hacks to allow partial transparent handling and
9530 even then it doesn't work with DER formats. Given the option anything
9531 other than PKCS#8 should be dumped: but the other formats have to
9532 stay in the name of compatibility.
9534 With public keys and the benefit of hindsight one standard format
9535 is used which works with EVP_PKEY, RSA or DSA structures: though
9536 it clearly returns an error if you try to read the wrong kind of key.
9538 Added a -pubkey option to the 'x509' utility to output the public key.
9539 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9540 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9541 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9542 that do the same as the EVP_PKEY_assign_*() except they up the
9543 reference count of the added key (they don't "swallow" the
9547 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9548 CRLs would fail if the file contained no certificates or no CRLs:
9549 added a new function to read in both types and return the number
9550 read: this means that if none are read it will be an error. The
9551 DER versions of the certificate and CRL reader would always fail
9552 because it isn't possible to mix certificates and CRLs in DER format
9553 without choking one or the other routine. Changed this to just read
9554 a certificate: this is the best we can do. Also modified the code
9555 in apps/verify.c to take notice of return codes: it was previously
9556 attempting to read in certificates from NULL pointers and ignoring
9557 any errors: this is one reason why the cert and CRL reader seemed
9558 to work. It doesn't check return codes from the default certificate
9559 routines: these may well fail if the certificates aren't installed.
9562 *) Code to support otherName option in GeneralName.
9565 *) First update to verify code. Change the verify utility
9566 so it warns if it is passed a self signed certificate:
9567 for consistency with the normal behaviour. X509_verify
9568 has been modified to it will now verify a self signed
9569 certificate if *exactly* the same certificate appears
9570 in the store: it was previously impossible to trust a
9571 single self signed certificate. This means that:
9572 openssl verify ss.pem
9573 now gives a warning about a self signed certificate but
9574 openssl verify -CAfile ss.pem ss.pem
9578 *) For servers, store verify_result in SSL_SESSION data structure
9579 (and add it to external session representation).
9580 This is needed when client certificate verifications fails,
9581 but an application-provided verification callback (set by
9582 SSL_CTX_set_cert_verify_callback) allows accepting the session
9583 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9584 but returns 1): When the session is reused, we have to set
9585 ssl->verify_result to the appropriate error code to avoid
9587 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9589 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9590 case in PKCS7_dataInit() where the signed PKCS7 structure
9591 didn't contain any existing data because it was being created.
9592 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9594 *) Add a salt to the key derivation routines in enc.c. This
9595 forms the first 8 bytes of the encrypted file. Also add a
9596 -S option to allow a salt to be input on the command line.
9599 *) New function X509_cmp(). Oddly enough there wasn't a function
9600 to compare two certificates. We do this by working out the SHA1
9601 hash and comparing that. X509_cmp() will be needed by the trust
9605 *) SSL_get1_session() is like SSL_get_session(), but increments
9606 the reference count in the SSL_SESSION returned.
9607 [Geoff Thorpe <geoff@eu.c2.net>]
9609 *) Fix for 'req': it was adding a null to request attributes.
9610 Also change the X509_LOOKUP and X509_INFO code to handle
9611 certificate auxiliary information.
9614 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9618 *) Add the possibility to add extra information to the memory leak
9619 detecting output, to form tracebacks, showing from where each
9620 allocation was originated: CRYPTO_push_info("constant string") adds
9621 the string plus current file name and line number to a per-thread
9622 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9623 is like calling CYRPTO_pop_info() until the stack is empty.
9624 Also updated memory leak detection code to be multi-thread-safe.
9627 *) Add options -text and -noout to pkcs7 utility and delete the
9628 encryption options which never did anything. Update docs.
9631 *) Add options to some of the utilities to allow the pass phrase
9632 to be included on either the command line (not recommended on
9633 OSes like Unix) or read from the environment. Update the
9634 manpages and fix a few bugs.
9637 *) Add a few manpages for some of the openssl commands.
9640 *) Fix the -revoke option in ca. It was freeing up memory twice,
9641 leaking and not finding already revoked certificates.
9644 *) Extensive changes to support certificate auxiliary information.
9645 This involves the use of X509_CERT_AUX structure and X509_AUX
9646 functions. An X509_AUX function such as PEM_read_X509_AUX()
9647 can still read in a certificate file in the usual way but it
9648 will also read in any additional "auxiliary information". By
9649 doing things this way a fair degree of compatibility can be
9650 retained: existing certificates can have this information added
9651 using the new 'x509' options.
9653 Current auxiliary information includes an "alias" and some trust
9654 settings. The trust settings will ultimately be used in enhanced
9655 certificate chain verification routines: currently a certificate
9656 can only be trusted if it is self signed and then it is trusted
9660 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9661 The problem was that one of the replacement routines had not been working
9662 since SSLeay releases. For now the offending routine has been replaced
9663 with non-optimised assembler. Even so, this now gives around 95%
9664 performance improvement for 1024 bit RSA signs.
9667 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9668 handling. Most clients have the effective key size in bits equal to
9669 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9670 A few however don't do this and instead use the size of the decrypted key
9671 to determine the RC2 key length and the AlgorithmIdentifier to determine
9672 the effective key length. In this case the effective key length can still
9673 be 40 bits but the key length can be 168 bits for example. This is fixed
9674 by manually forcing an RC2 key into the EVP_PKEY structure because the
9675 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9676 the key length and effective key length are equal.
9679 *) Add a bunch of functions that should simplify the creation of
9680 X509_NAME structures. Now you should be able to do:
9681 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9682 and have it automatically work out the correct field type and fill in
9683 the structures. The more adventurous can try:
9684 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9685 and it will (hopefully) work out the correct multibyte encoding.
9688 *) Change the 'req' utility to use the new field handling and multibyte
9689 copy routines. Before the DN field creation was handled in an ad hoc
9690 way in req, ca, and x509 which was rather broken and didn't support
9691 BMPStrings or UTF8Strings. Since some software doesn't implement
9692 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9693 using the dirstring_type option. See the new comment in the default
9694 openssl.cnf for more info.
9697 *) Make crypto/rand/md_rand.c more robust:
9698 - Assure unique random numbers after fork().
9699 - Make sure that concurrent threads access the global counter and
9700 md serializably so that we never lose entropy in them
9701 or use exactly the same state in multiple threads.
9702 Access to the large state is not always serializable because
9703 the additional locking could be a performance killer, and
9704 md should be large enough anyway.
9707 *) New file apps/app_rand.c with commonly needed functionality
9708 for handling the random seed file.
9710 Use the random seed file in some applications that previously did not:
9712 dsaparam -genkey (which also ignored its '-rand' option),
9715 x509 (when signing).
9716 Except on systems with /dev/urandom, it is crucial to have a random
9717 seed file at least for key creation, DSA signing, and for DH exchanges;
9718 for RSA signatures we could do without one.
9720 gendh and gendsa (unlike genrsa) used to read only the first byte
9721 of each file listed in the '-rand' option. The function as previously
9722 found in genrsa is now in app_rand.c and is used by all programs
9723 that support '-rand'.
9726 *) In RAND_write_file, use mode 0600 for creating files;
9727 don't just chmod when it may be too late.
9730 *) Report an error from X509_STORE_load_locations
9731 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9734 *) New function ASN1_mbstring_copy() this copies a string in either
9735 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9736 into an ASN1_STRING type. A mask of permissible types is passed
9737 and it chooses the "minimal" type to use or an error if not type
9741 *) Add function equivalents to the various macros in asn1.h. The old
9742 macros are retained with an M_ prefix. Code inside the library can
9743 use the M_ macros. External code (including the openssl utility)
9744 should *NOT* in order to be "shared library friendly".
9747 *) Add various functions that can check a certificate's extensions
9748 to see if it usable for various purposes such as SSL client,
9749 server or S/MIME and CAs of these types. This is currently
9750 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9751 verification. Also added a -purpose flag to x509 utility to
9752 print out all the purposes.
9755 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9759 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9760 for, obtain and decode and extension and obtain its critical flag.
9761 This allows all the necessary extension code to be handled in a
9762 single function call.
9765 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9766 platforms. See crypto/rc4/rc4_enc.c for further details.
9769 *) New -noout option to asn1parse. This causes no output to be produced
9770 its main use is when combined with -strparse and -out to extract data
9771 from a file (which may not be in ASN.1 format).
9774 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9775 when producing the local key id.
9776 [Richard Levitte <levitte@stacken.kth.se>]
9778 *) New option -dhparam in s_server. This allows a DH parameter file to be
9779 stated explicitly. If it is not stated then it tries the first server
9780 certificate file. The previous behaviour hard coded the filename
9784 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9785 a public key to be input or output. For example:
9786 openssl rsa -in key.pem -pubout -out pubkey.pem
9787 Also added necessary DSA public key functions to handle this.
9790 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9791 in the message. This was handled by allowing
9792 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9793 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9795 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9796 to the end of the strings whereas this didn't. This would cause problems
9797 if strings read with d2i_ASN1_bytes() were later modified.
9798 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9800 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9801 data and it contains EOF it will end up returning an error. This is
9802 caused by input 46 bytes long. The cause is due to the way base64
9803 BIOs find the start of base64 encoded data. They do this by trying a
9804 trial decode on each line until they find one that works. When they
9805 do a flag is set and it starts again knowing it can pass all the
9806 data directly through the decoder. Unfortunately it doesn't reset
9807 the context it uses. This means that if EOF is reached an attempt
9808 is made to pass two EOFs through the context and this causes the
9809 resulting error. This can also cause other problems as well. As is
9810 usual with these problems it takes *ages* to find and the fix is
9811 trivial: move one line.
9812 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9814 *) Ugly workaround to get s_client and s_server working under Windows. The
9815 old code wouldn't work because it needed to select() on sockets and the
9816 tty (for keypresses and to see if data could be written). Win32 only
9817 supports select() on sockets so we select() with a 1s timeout on the
9818 sockets and then see if any characters are waiting to be read, if none
9819 are present then we retry, we also assume we can always write data to
9820 the tty. This isn't nice because the code then blocks until we've
9821 received a complete line of data and it is effectively polling the
9822 keyboard at 1s intervals: however it's quite a bit better than not
9823 working at all :-) A dedicated Windows application might handle this
9824 with an event loop for example.
9827 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9828 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9829 will be called when RSA_sign() and RSA_verify() are used. This is useful
9830 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9831 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9832 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9833 This necessitated the support of an extra signature type NID_md5_sha1
9834 for SSL signatures and modifications to the SSL library to use it instead
9835 of calling RSA_public_decrypt() and RSA_private_encrypt().
9838 *) Add new -verify -CAfile and -CApath options to the crl program, these
9839 will lookup a CRL issuers certificate and verify the signature in a
9840 similar way to the verify program. Tidy up the crl program so it
9841 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9842 less strict. It will now permit CRL extensions even if it is not
9843 a V2 CRL: this will allow it to tolerate some broken CRLs.
9846 *) Initialize all non-automatic variables each time one of the openssl
9847 sub-programs is started (this is necessary as they may be started
9848 multiple times from the "OpenSSL>" prompt).
9849 [Lennart Bang, Bodo Moeller]
9851 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9852 removing all other RSA functionality (this is what NO_RSA does). This
9853 is so (for example) those in the US can disable those operations covered
9854 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9858 *) Non-copying interface to BIO pairs.
9859 (still largely untested)
9862 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9863 ASCII string. This was handled independently in various places before.
9866 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9867 UTF8 strings a character at a time.
9870 *) Use client_version from client hello to select the protocol
9871 (s23_srvr.c) and for RSA client key exchange verification
9872 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9875 *) Add various utility functions to handle SPKACs, these were previously
9876 handled by poking round in the structure internals. Added new function
9877 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9878 print, verify and generate SPKACs. Based on an original idea from
9879 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9882 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9885 *) Allow the config file extension section to be overwritten on the
9886 command line. Based on an original idea from Massimiliano Pala
9887 <madwolf@comune.modena.it>. The new option is called -extensions
9888 and can be applied to ca, req and x509. Also -reqexts to override
9889 the request extensions in req and -crlexts to override the crl extensions
9893 *) Add new feature to the SPKAC handling in ca. Now you can include
9894 the same field multiple times by preceding it by "XXXX." for example:
9897 this is the same syntax as used in the req config file.
9900 *) Allow certificate extensions to be added to certificate requests. These
9901 are specified in a 'req_extensions' option of the req section of the
9902 config file. They can be printed out with the -text option to req but
9903 are otherwise ignored at present.
9906 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9907 data read consists of only the final block it would not decrypted because
9908 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9909 A misplaced 'break' also meant the decrypted final block might not be
9910 copied until the next read.
9913 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9914 a few extra parameters to the DH structure: these will be useful if
9915 for example we want the value of 'q' or implement X9.42 DH.
9918 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9919 provides hooks that allow the default DSA functions or functions on a
9920 "per key" basis to be replaced. This allows hardware acceleration and
9921 hardware key storage to be handled without major modification to the
9922 library. Also added low level modexp hooks and CRYPTO_EX structure and
9923 associated functions.
9926 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9927 as "read only": it can't be written to and the buffer it points to will
9928 not be freed. Reading from a read only BIO is much more efficient than
9929 a normal memory BIO. This was added because there are several times when
9930 an area of memory needs to be read from a BIO. The previous method was
9931 to create a memory BIO and write the data to it, this results in two
9932 copies of the data and an O(n^2) reading algorithm. There is a new
9933 function BIO_new_mem_buf() which creates a read only memory BIO from
9934 an area of memory. Also modified the PKCS#7 routines to use read only
9938 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9939 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9940 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9941 but a retry condition occured while trying to read the rest.
9944 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9945 NID_pkcs7_encrypted by default: this was wrong since this should almost
9946 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9947 the encrypted data type: this is a more sensible place to put it and it
9948 allows the PKCS#12 code to be tidied up that duplicated this
9952 *) Changed obj_dat.pl script so it takes its input and output files on
9953 the command line. This should avoid shell escape redirection problems
9957 *) Initial support for certificate extension requests, these are included
9958 in things like Xenroll certificate requests. Included functions to allow
9959 extensions to be obtained and added.
9962 *) -crlf option to s_client and s_server for sending newlines as
9963 CRLF (as required by many protocols).
9966 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9968 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9969 [Ralf S. Engelschall]
9971 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9972 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9974 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9978 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9979 DH parameters/keys (q is lost during that conversion, but the resulting
9980 DH parameters contain its length).
9982 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9983 much faster than DH_generate_parameters (which creates parameters
9984 where p = 2*q + 1), and also the smaller q makes DH computations
9985 much more efficient (160-bit exponentiation instead of 1024-bit
9986 exponentiation); so this provides a convenient way to support DHE
9987 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9988 utter importance to use
9989 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9991 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9992 when such DH parameters are used, because otherwise small subgroup
9993 attacks may become possible!
9996 *) Avoid memory leak in i2d_DHparams.
9999 *) Allow the -k option to be used more than once in the enc program:
10000 this allows the same encrypted message to be read by multiple recipients.
10003 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10004 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10005 it will always use the numerical form of the OID, even if it has a short
10009 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10010 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10011 otherwise bn_mod_exp was called. In the case of hardware keys for example
10012 no private key components need be present and it might store extra data
10013 in the RSA structure, which cannot be accessed from bn_mod_exp.
10014 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10015 private key operations.
10018 *) Added support for SPARC Linux.
10021 *) pem_password_cb function type incompatibly changed from
10022 typedef int pem_password_cb(char *buf, int size, int rwflag);
10024 ....(char *buf, int size, int rwflag, void *userdata);
10025 so that applications can pass data to their callbacks:
10026 The PEM[_ASN1]_{read,write}... functions and macros now take an
10027 additional void * argument, which is just handed through whenever
10028 the password callback is called.
10029 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10031 New function SSL_CTX_set_default_passwd_cb_userdata.
10033 Compatibility note: As many C implementations push function arguments
10034 onto the stack in reverse order, the new library version is likely to
10035 interoperate with programs that have been compiled with the old
10036 pem_password_cb definition (PEM_whatever takes some data that
10037 happens to be on the stack as its last argument, and the callback
10038 just ignores this garbage); but there is no guarantee whatsoever that
10041 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10042 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10043 problems not only on Windows, but also on some Unix platforms.
10044 To avoid problematic command lines, these definitions are now in an
10045 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10046 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10049 *) MIPS III/IV assembler module is reimplemented.
10052 *) More DES library cleanups: remove references to srand/rand and
10053 delete an unused file.
10056 *) Add support for the the free Netwide assembler (NASM) under Win32,
10057 since not many people have MASM (ml) and it can be hard to obtain.
10058 This is currently experimental but it seems to work OK and pass all
10059 the tests. Check out INSTALL.W32 for info.
10062 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10063 without temporary keys kept an extra copy of the server key,
10064 and connections with temporary keys did not free everything in case
10068 *) New function RSA_check_key and new openssl rsa option -check
10069 for verifying the consistency of RSA keys.
10070 [Ulf Moeller, Bodo Moeller]
10072 *) Various changes to make Win32 compile work:
10073 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10074 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10075 comparison" warnings.
10076 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10079 *) Add a debugging option to PKCS#5 v2 key generation function: when
10080 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10081 derived keys are printed to stderr.
10084 *) Copy the flags in ASN1_STRING_dup().
10085 [Roman E. Pavlov <pre@mo.msk.ru>]
10087 *) The x509 application mishandled signing requests containing DSA
10088 keys when the signing key was also DSA and the parameters didn't match.
10090 It was supposed to omit the parameters when they matched the signing key:
10091 the verifying software was then supposed to automatically use the CA's
10092 parameters if they were absent from the end user certificate.
10094 Omitting parameters is no longer recommended. The test was also
10095 the wrong way round! This was probably due to unusual behaviour in
10096 EVP_cmp_parameters() which returns 1 if the parameters match.
10097 This meant that parameters were omitted when they *didn't* match and
10098 the certificate was useless. Certificates signed with 'ca' didn't have
10100 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10102 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10103 The interface is as follows:
10104 Applications can use
10105 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10106 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10107 "off" is now the default.
10108 The library internally uses
10109 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10110 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10111 to disable memory-checking temporarily.
10113 Some inconsistent states that previously were possible (and were
10114 even the default) are now avoided.
10116 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10117 with each memory chunk allocated; this is occasionally more helpful
10118 than just having a counter.
10120 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10122 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10126 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10127 which largely parallels "options", but is for changing API behaviour,
10128 whereas "options" are about protocol behaviour.
10129 Initial "mode" flags are:
10131 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10132 a single record has been written.
10133 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10134 retries use the same buffer location.
10135 (But all of the contents must be
10139 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10142 *) Fix problems with no-hmac etc.
10143 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10145 *) New functions RSA_get_default_method(), RSA_set_method() and
10146 RSA_get_method(). These allows replacement of RSA_METHODs without having
10147 to mess around with the internals of an RSA structure.
10150 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10151 Also really enable memory leak checks in openssl.c and in some
10153 [Chad C. Mulligan, Bodo Moeller]
10155 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10156 up the length of negative integers. This has now been simplified to just
10157 store the length when it is first determined and use it later, rather
10158 than trying to keep track of where data is copied and updating it to
10160 [Steve Henson, reported by Brien Wheeler
10161 <bwheeler@authentica-security.com>]
10163 *) Add a new function PKCS7_signatureVerify. This allows the verification
10164 of a PKCS#7 signature but with the signing certificate passed to the
10165 function itself. This contrasts with PKCS7_dataVerify which assumes the
10166 certificate is present in the PKCS#7 structure. This isn't always the
10167 case: certificates can be omitted from a PKCS#7 structure and be
10168 distributed by "out of band" means (such as a certificate database).
10171 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10172 function prototypes in pem.h, also change util/mkdef.pl to add the
10173 necessary function names.
10176 *) mk1mf.pl (used by Windows builds) did not properly read the
10177 options set by Configure in the top level Makefile, and Configure
10178 was not even able to write more than one option correctly.
10179 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10182 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10183 file to be loaded from a BIO or FILE pointer. The BIO version will
10184 for example allow memory BIOs to contain config info.
10187 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10188 Whoever hopes to achieve shared-library compatibility across versions
10189 must use this, not the compile-time macro.
10190 (Exercise 0.9.4: Which is the minimum library version required by
10192 Note: All this applies only to multi-threaded programs, others don't
10196 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10197 through a BIO pair triggered the default case, i.e.
10198 SSLerr(...,SSL_R_UNKNOWN_STATE).
10201 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10202 can use the SSL library even if none of the specific BIOs is
10206 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10207 for the encoded length.
10208 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10210 *) Add initial documentation of the X509V3 functions.
10213 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10214 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10215 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10216 secure PKCS#8 private key format with a high iteration count.
10219 *) Fix determination of Perl interpreter: A perl or perl5
10220 _directory_ in $PATH was also accepted as the interpreter.
10221 [Ralf S. Engelschall]
10223 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10224 wrong with it but it was very old and did things like calling
10225 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10226 unusual formatting.
10229 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10230 to use the new extension code.
10233 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10234 with macros. This should make it easier to change their form, add extra
10235 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10239 *) Add to configuration table a new entry that can specify an alternative
10240 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10241 according to Mark Crispin <MRC@Panda.COM>.
10245 *) DES CBC did not update the IV. Weird.
10248 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10249 Changing the behaviour of the former might break existing programs --
10250 where IV updating is needed, des_ncbc_encrypt can be used.
10253 *) When bntest is run from "make test" it drives bc to check its
10254 calculations, as well as internally checking them. If an internal check
10255 fails, it needs to cause bc to give a non-zero result or make test carries
10256 on without noticing the failure. Fixed.
10259 *) DES library cleanups.
10262 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10263 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10264 ciphers. NOTE: although the key derivation function has been verified
10265 against some published test vectors it has not been extensively tested
10266 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10270 *) Instead of "mkdir -p", which is not fully portable, use new
10271 Perl script "util/mkdir-p.pl".
10274 *) Rewrite the way password based encryption (PBE) is handled. It used to
10275 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10276 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10277 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10278 the 'parameter' field of the AlgorithmIdentifier is passed to the
10279 underlying key generation function so it must do its own ASN1 parsing.
10280 This has also changed the EVP_PBE_CipherInit() function which now has a
10281 'parameter' argument instead of literal salt and iteration count values
10282 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10285 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10286 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10287 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10288 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10289 value was just used as a "magic string" and not used directly its
10290 value doesn't matter.
10293 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10297 *) "linux-sparc64" configuration (ultrapenguin).
10298 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10299 "linux-sparc" configuration.
10300 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10302 *) config now generates no-xxx options for missing ciphers.
10305 *) Support the EBCDIC character set (work in progress).
10306 File ebcdic.c not yet included because it has a different license.
10307 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10309 *) Support BS2000/OSD-POSIX.
10310 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10312 *) Make callbacks for key generation use void * instead of char *.
10315 *) Make S/MIME samples compile (not yet tested).
10318 *) Additional typesafe stacks.
10321 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10325 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10327 *) New configuration variant "sco5-gcc".
10329 *) Updated some demos.
10330 [Sean O Riordain, Wade Scholine]
10332 *) Add missing BIO_free at exit of pkcs12 application.
10335 *) Fix memory leak in conf.c.
10338 *) Updates for Win32 to assembler version of MD5.
10341 *) Set #! path to perl in apps/der_chop to where we found it
10342 instead of using a fixed path.
10345 *) SHA library changes for irix64-mips4-cc.
10348 *) Improvements for VMS support.
10352 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10354 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10355 This also avoids the problems with SC4.2 and unpatched SC5.
10356 [Andy Polyakov <appro@fy.chalmers.se>]
10358 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10359 These are required because of the typesafe stack would otherwise break
10360 existing code. If old code used a structure member which used to be STACK
10361 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10362 sk_num or sk_value it would produce an error because the num, data members
10363 are not present in STACK_OF. Now it just produces a warning. sk_set
10364 replaces the old method of assigning a value to sk_value
10365 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10366 that does this will no longer work (and should use sk_set instead) but
10367 this could be regarded as a "questionable" behaviour anyway.
10370 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10371 correctly handle encrypted S/MIME data.
10374 *) Change type of various DES function arguments from des_cblock
10375 (which means, in function argument declarations, pointer to char)
10376 to des_cblock * (meaning pointer to array with 8 char elements),
10377 which allows the compiler to do more typechecking; it was like
10378 that back in SSLeay, but with lots of ugly casts.
10380 Introduce new type const_des_cblock.
10383 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10384 problems: find RecipientInfo structure that matches recipient certificate
10385 and initialise the ASN1 structures properly based on passed cipher.
10388 *) Belatedly make the BN tests actually check the results.
10391 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10392 to and from BNs: it was completely broken. New compilation option
10393 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10394 key elements as negative integers.
10397 *) Reorganize and speed up MD5.
10398 [Andy Polyakov <appro@fy.chalmers.se>]
10401 [Richard Levitte <richard@levitte.org>]
10403 *) New option -out to asn1parse to allow the parsed structure to be
10404 output to a file. This is most useful when combined with the -strparse
10405 option to examine the output of things like OCTET STRINGS.
10408 *) Make SSL library a little more fool-proof by not requiring any longer
10409 that SSL_set_{accept,connect}_state be called before
10410 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10411 in many applications because usually everything *appeared* to work as
10412 intended anyway -- now it really works as intended).
10415 *) Move openssl.cnf out of lib/.
10418 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10419 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10420 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10421 [Ralf S. Engelschall]
10423 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10424 handle PKCS#7 enveloped data properly.
10425 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10427 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10428 copying pointers. The cert_st handling is changed by this in
10429 various ways (and thus what used to be known as ctx->default_cert
10430 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10431 any longer when s->cert does not give us what we need).
10432 ssl_cert_instantiate becomes obsolete by this change.
10433 As soon as we've got the new code right (possibly it already is?),
10434 we have solved a couple of bugs of the earlier code where s->cert
10435 was used as if it could not have been shared with other SSL structures.
10437 Note that using the SSL API in certain dirty ways now will result
10438 in different behaviour than observed with earlier library versions:
10439 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10440 does not influence s as it used to.
10442 In order to clean up things more thoroughly, inside SSL_SESSION
10443 we don't use CERT any longer, but a new structure SESS_CERT
10444 that holds per-session data (if available); currently, this is
10445 the peer's certificate chain and, for clients, the server's certificate
10446 and temporary key. CERT holds only those values that can have
10447 meaningful defaults in an SSL_CTX.
10450 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10451 from the internal representation. Various PKCS#7 fixes: remove some
10452 evil casts and set the enc_dig_alg field properly based on the signing
10456 *) Allow PKCS#12 password to be set from the command line or the
10457 environment. Let 'ca' get its config file name from the environment
10458 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10462 *) Allow certificate policies extension to use an IA5STRING for the
10463 organization field. This is contrary to the PKIX definition but
10464 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10468 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10469 without disallowing inline assembler and the like for non-pedantic builds.
10472 *) Support Borland C++ builder.
10473 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10475 *) Support Mingw32.
10478 *) SHA-1 cleanups and performance enhancements.
10479 [Andy Polyakov <appro@fy.chalmers.se>]
10481 *) Sparc v8plus assembler for the bignum library.
10482 [Andy Polyakov <appro@fy.chalmers.se>]
10484 *) Accept any -xxx and +xxx compiler options in Configure.
10487 *) Update HPUX configuration.
10490 *) Add missing sk_<type>_unshift() function to safestack.h
10491 [Ralf S. Engelschall]
10493 *) New function SSL_CTX_use_certificate_chain_file that sets the
10494 "extra_cert"s in addition to the certificate. (This makes sense
10495 only for "PEM" format files, as chains as a whole are not
10499 *) Support verify_depth from the SSL API.
10500 x509_vfy.c had what can be considered an off-by-one-error:
10501 Its depth (which was not part of the external interface)
10502 was actually counting the number of certificates in a chain;
10503 now it really counts the depth.
10506 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10507 instead of X509err, which often resulted in confusing error
10508 messages since the error codes are not globally unique
10509 (e.g. an alleged error in ssl3_accept when a certificate
10510 didn't match the private key).
10512 *) New function SSL_CTX_set_session_id_context that allows to set a default
10513 value (so that you don't need SSL_set_session_id_context for each
10514 connection using the SSL_CTX).
10517 *) OAEP decoding bug fix.
10520 *) Support INSTALL_PREFIX for package builders, as proposed by
10524 *) New Configure options "threads" and "no-threads". For systems
10525 where the proper compiler options are known (currently Solaris
10526 and Linux), "threads" is the default.
10529 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10532 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10533 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10534 such as /usr/local/bin.
10537 *) "make linux-shared" to build shared libraries.
10538 [Niels Poppe <niels@netbox.org>]
10540 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10543 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10544 extension adding in x509 utility.
10547 *) Remove NOPROTO sections and error code comments.
10550 *) Partial rewrite of the DEF file generator to now parse the ANSI
10554 *) New Configure options --prefix=DIR and --openssldir=DIR.
10557 *) Complete rewrite of the error code script(s). It is all now handled
10558 by one script at the top level which handles error code gathering,
10559 header rewriting and C source file generation. It should be much better
10560 than the old method: it now uses a modified version of Ulf's parser to
10561 read the ANSI prototypes in all header files (thus the old K&R definitions
10562 aren't needed for error creation any more) and do a better job of
10563 translating function codes into names. The old 'ASN1 error code imbedded
10564 in a comment' is no longer necessary and it doesn't use .err files which
10565 have now been deleted. Also the error code call doesn't have to appear all
10566 on one line (which resulted in some large lines...).
10569 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10572 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10573 0 (which usually indicates a closed connection), but continue reading.
10576 *) Fix some race conditions.
10579 *) Add support for CRL distribution points extension. Add Certificate
10580 Policies and CRL distribution points documentation.
10583 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10586 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10587 8 of keying material. Merlin has also confirmed interop with this fix
10588 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10589 [Merlin Hughes <merlin@baltimore.ie>]
10591 *) Fix lots of warnings.
10592 [Richard Levitte <levitte@stacken.kth.se>]
10594 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10595 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10596 [Richard Levitte <levitte@stacken.kth.se>]
10598 *) Fix problems with sizeof(long) == 8.
10599 [Andy Polyakov <appro@fy.chalmers.se>]
10601 *) Change functions to ANSI C.
10604 *) Fix typos in error codes.
10605 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10607 *) Remove defunct assembler files from Configure.
10610 *) SPARC v8 assembler BIGNUM implementation.
10611 [Andy Polyakov <appro@fy.chalmers.se>]
10613 *) Support for Certificate Policies extension: both print and set.
10614 Various additions to support the r2i method this uses.
10617 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10618 return a const string when you are expecting an allocated buffer.
10621 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10622 types DirectoryString and DisplayText.
10625 *) Add code to allow r2i extensions to access the configuration database,
10626 add an LHASH database driver and add several ctx helper functions.
10629 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10630 fail when they extended the size of a BIGNUM.
10633 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10634 support typesafe stack.
10637 *) Fix typo in SSL_[gs]et_options().
10638 [Nils Frostberg <nils@medcom.se>]
10640 *) Delete various functions and files that belonged to the (now obsolete)
10641 old X509V3 handling code.
10644 *) New Configure option "rsaref".
10647 *) Don't auto-generate pem.h.
10650 *) Introduce type-safe ASN.1 SETs.
10653 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10654 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10656 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10657 that links with OpenSSL (well at least cause lots of warnings), but fear
10658 not: the conversion is trivial, and it eliminates loads of evil casts. A
10659 few STACKed things have been converted already. Feel free to convert more.
10660 In the fullness of time, I'll do away with the STACK type altogether.
10663 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10664 specified in <certfile> by updating the entry in the index.txt file.
10665 This way one no longer has to edit the index.txt file manually for
10666 revoking a certificate. The -revoke option does the gory details now.
10667 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10669 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10670 `-text' option at all and this way the `-noout -text' combination was
10671 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10672 [Ralf S. Engelschall]
10674 *) Make sure a corresponding plain text error message exists for the
10675 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10676 verify callback function determined that a certificate was revoked.
10677 [Ralf S. Engelschall]
10679 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10680 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10681 all available cipers including rc5, which was forgotten until now.
10682 In order to let the testing shell script know which algorithms
10683 are available, a new (up to now undocumented) command
10684 "openssl list-cipher-commands" is used.
10687 *) Bugfix: s_client occasionally would sleep in select() when
10688 it should have checked SSL_pending() first.
10691 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10692 the raw DSA values prior to ASN.1 encoding.
10695 *) Tweaks to Configure
10696 [Niels Poppe <niels@netbox.org>]
10698 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10702 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10705 *) New config option to avoid instructions that are illegal on the 80386.
10706 The default code is faster, but requires at least a 486.
10709 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10710 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10711 same as SSL2_VERSION anyway.
10714 *) New "-showcerts" option for s_client.
10717 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10718 application. Various cleanups and fixes.
10721 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10722 modify error routines to work internally. Add error codes and PBE init
10723 to library startup routines.
10726 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10727 packing functions to asn1 and evp. Changed function names and error
10728 codes along the way.
10731 *) PKCS12 integration: and so it begins... First of several patches to
10732 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10733 objects to objects.h
10736 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10737 and display support for Thawte strong extranet extension.
10740 *) Add LinuxPPC support.
10741 [Jeff Dubrule <igor@pobox.org>]
10743 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10744 bn_div_words in alpha.s.
10745 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10747 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10748 OAEP isn't supported when OpenSSL is built with RSAref.
10749 [Ulf Moeller <ulf@fitug.de>]
10751 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10752 so they no longer are missing under -DNOPROTO.
10753 [Soren S. Jorvang <soren@t.dk>]
10756 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10758 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10759 doesn't work when the session is reused. Coming soon!
10762 *) Fix a security hole, that allows sessions to be reused in the wrong
10763 context thus bypassing client cert protection! All software that uses
10764 client certs and session caches in multiple contexts NEEDS PATCHING to
10765 allow session reuse! A fuller solution is in the works.
10766 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10768 *) Some more source tree cleanups (removed obsolete files
10769 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10770 permission on "config" script to be executable) and a fix for the INSTALL
10772 [Ulf Moeller <ulf@fitug.de>]
10774 *) Remove some legacy and erroneous uses of malloc, free instead of
10776 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10778 *) Make rsa_oaep_test return non-zero on error.
10779 [Ulf Moeller <ulf@fitug.de>]
10781 *) Add support for native Solaris shared libraries. Configure
10782 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10783 if someone would make that last step automatic.
10784 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10786 *) ctx_size was not built with the right compiler during "make links". Fixed.
10789 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10790 except NULL ciphers". This means the default cipher list will no longer
10791 enable NULL ciphers. They need to be specifically enabled e.g. with
10792 the string "DEFAULT:eNULL".
10795 *) Fix to RSA private encryption routines: if p < q then it would
10796 occasionally produce an invalid result. This will only happen with
10797 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10800 *) Be less restrictive and allow also `perl util/perlpath.pl
10801 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10802 because this way one can also use an interpreter named `perl5' (which is
10803 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10804 installed as `perl').
10805 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10807 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10808 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10810 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10811 advapi32.lib to Win32 build and change the pem test comparison
10812 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10813 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10814 and crypto/des/ede_cbcm_enc.c.
10817 *) DES quad checksum was broken on big-endian architectures. Fixed.
10820 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10821 Win32 test batch file so it (might) work again. The Win32 test batch file
10822 is horrible: I feel ill....
10825 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10826 in e_os.h. Audit of header files to check ANSI and non ANSI
10827 sections: 10 functions were absent from non ANSI section and not exported
10828 from Windows DLLs. Fixed up libeay.num for new functions.
10831 *) Make `openssl version' output lines consistent.
10832 [Ralf S. Engelschall]
10834 *) Fix Win32 symbol export lists for BIO functions: Added
10835 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10836 to ms/libeay{16,32}.def.
10837 [Ralf S. Engelschall]
10839 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10840 fine under Unix and passes some trivial tests I've now added. But the
10841 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10842 added to make sure no one expects that this stuff really works in the
10843 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10844 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10846 [Ralf S. Engelschall]
10848 *) Fix the generation of two part addresses in perl.
10849 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10851 *) Add config entry for Linux on MIPS.
10852 [John Tobey <jtobey@channel1.com>]
10854 *) Make links whenever Configure is run, unless we are on Windoze.
10857 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10858 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10862 *) Add a useful kludge to allow package maintainers to specify compiler and
10863 other platforms details on the command line without having to patch the
10864 Configure script everytime: One now can use ``perl Configure
10865 <id>:<details>'', i.e. platform ids are allowed to have details appended
10866 to them (separated by colons). This is treated as there would be a static
10867 pre-configured entry in Configure's %table under key <id> with value
10868 <details> and ``perl Configure <id>'' is called. So, when you want to
10869 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10870 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10871 now, which overrides the FreeBSD-elf entry on-the-fly.
10872 [Ralf S. Engelschall]
10874 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10877 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10878 on the `perl Configure ...' command line. This way one can compile
10879 OpenSSL libraries with Position Independent Code (PIC) which is needed
10880 for linking it into DSOs.
10881 [Ralf S. Engelschall]
10883 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10887 *) Cleaned up the LICENSE document: The official contact for any license
10888 questions now is the OpenSSL core team under openssl-core@openssl.org.
10889 And add a paragraph about the dual-license situation to make sure people
10890 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10891 to the OpenSSL toolkit.
10892 [Ralf S. Engelschall]
10894 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10895 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10896 Additionally cleaned up the `make links' target: Remove unnecessary
10897 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10898 to speed processing and no longer clutter the display with confusing
10899 stuff. Instead only the actually done links are displayed.
10900 [Ralf S. Engelschall]
10902 *) Permit null encryption ciphersuites, used for authentication only. It used
10903 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10904 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10908 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10909 signed attributes when verifying signatures (this would break them),
10910 the detached data encoding was wrong and public keys obtained using
10911 X509_get_pubkey() weren't freed.
10914 *) Add text documentation for the BUFFER functions. Also added a work around
10915 to a Win95 console bug. This was triggered by the password read stuff: the
10916 last character typed gets carried over to the next fread(). If you were
10917 generating a new cert request using 'req' for example then the last
10918 character of the passphrase would be CR which would then enter the first
10922 *) Added the new `Includes OpenSSL Cryptography Software' button as
10923 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10924 button and can be used by applications based on OpenSSL to show the
10925 relationship to the OpenSSL project.
10926 [Ralf S. Engelschall]
10928 *) Remove confusing variables in function signatures in files
10929 ssl/ssl_lib.c and ssl/ssl.h.
10930 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10932 *) Don't install bss_file.c under PREFIX/include/
10933 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10935 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10936 functions that return function pointers and has support for NT specific
10937 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10938 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10939 unsigned to signed types: this was killing the Win32 compile.
10942 *) Add new certificate file to stack functions,
10943 SSL_add_dir_cert_subjects_to_stack() and
10944 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10945 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10946 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10947 This means that Apache-SSL and similar packages don't have to mess around
10948 to add as many CAs as they want to the preferred list.
10951 *) Experiment with doxygen documentation. Currently only partially applied to
10953 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10954 openssl.doxy as the configuration file.
10957 *) Get rid of remaining C++-style comments which strict C compilers hate.
10958 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10960 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10961 compiled in by default: it has problems with large keys.
10964 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10965 DH private keys and/or callback functions which directly correspond to
10966 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10967 is needed for applications which have to configure certificates on a
10968 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10970 For the RSA certificate situation is makes no difference, but
10971 for the DSA certificate situation this fixes the "no shared cipher"
10972 problem where the OpenSSL cipher selection procedure failed because the
10973 temporary keys were not overtaken from the context and the API provided
10974 no way to reconfigure them.
10975 The new functions now let applications reconfigure the stuff and they
10976 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10977 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10978 non-public-API function ssl_cert_instantiate() is used as a helper
10979 function and also to reduce code redundancy inside ssl_rsa.c.
10980 [Ralf S. Engelschall]
10982 *) Move s_server -dcert and -dkey options out of the undocumented feature
10983 area because they are useful for the DSA situation and should be
10984 recognized by the users.
10985 [Ralf S. Engelschall]
10987 *) Fix the cipher decision scheme for export ciphers: the export bits are
10988 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10989 SSL_EXP_MASK. So, the original variable has to be used instead of the
10990 already masked variable.
10991 [Richard Levitte <levitte@stacken.kth.se>]
10993 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10994 [Richard Levitte <levitte@stacken.kth.se>]
10996 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10997 from `int' to `unsigned int' because it's a length and initialized by
10998 EVP_DigestFinal() which expects an `unsigned int *'.
10999 [Richard Levitte <levitte@stacken.kth.se>]
11001 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11002 script. Instead use the usual Shell->Perl transition trick.
11003 [Ralf S. Engelschall]
11005 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11006 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11007 -noout -modulus' as it's already the case for `openssl rsa -noout
11008 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11009 currently the public key is printed (a decision which was already done by
11010 `openssl dsa -modulus' in the past) which serves a similar purpose.
11011 Additionally the NO_RSA no longer completely removes the whole -modulus
11012 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11014 [Ralf S. Engelschall]
11016 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11017 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11018 [Arne Ansper <arne@ats.cyber.ee>]
11020 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11021 to be added. Now both 'req' and 'ca' can use new objects defined in the
11025 *) Add cool BIO that does syslog (or event log on NT).
11026 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11028 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11029 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11030 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11031 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11034 *) Add preliminary config info for new extension code.
11037 *) Make RSA_NO_PADDING really use no padding.
11038 [Ulf Moeller <ulf@fitug.de>]
11040 *) Generate errors when private/public key check is done.
11043 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11044 for some CRL extensions and new objects added.
11047 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11048 key usage extension and fuller support for authority key id.
11051 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11052 padding method for RSA, which is recommended for new applications in PKCS
11053 #1 v2.0 (RFC 2437, October 1998).
11054 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11055 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11056 against Bleichbacher's attack on RSA.
11057 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11060 *) Updates to the new SSL compression code
11061 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11063 *) Fix so that the version number in the master secret, when passed
11064 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11065 (because the server will not accept higher), that the version number
11066 is 0x03,0x01, not 0x03,0x00
11067 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11069 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11070 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11071 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11074 *) Support for RAW extensions where an arbitrary extension can be
11075 created by including its DER encoding. See apps/openssl.cnf for
11079 *) Make sure latest Perl versions don't interpret some generated C array
11080 code as Perl array code in the crypto/err/err_genc.pl script.
11081 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11083 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11084 not many people have the assembler. Various Win32 compilation fixes and
11085 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11086 build instructions.
11089 *) Modify configure script 'Configure' to automatically create crypto/date.h
11090 file under Win32 and also build pem.h from pem.org. New script
11091 util/mkfiles.pl to create the MINFO file on environments that can't do a
11092 'make files': perl util/mkfiles.pl >MINFO should work.
11095 *) Major rework of DES function declarations, in the pursuit of correctness
11096 and purity. As a result, many evil casts evaporated, and some weirdness,
11097 too. You may find this causes warnings in your code. Zapping your evil
11098 casts will probably fix them. Mostly.
11101 *) Fix for a typo in asn1.h. Bug fix to object creation script
11102 obj_dat.pl. It considered a zero in an object definition to mean
11103 "end of object": none of the objects in objects.h have any zeros
11104 so it wasn't spotted.
11105 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11107 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11108 Masking (CBCM). In the absence of test vectors, the best I have been able
11109 to do is check that the decrypt undoes the encrypt, so far. Send me test
11110 vectors if you have them.
11113 *) Correct calculation of key length for export ciphers (too much space was
11114 allocated for null ciphers). This has not been tested!
11117 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11118 message is now correct (it understands "crypto" and "ssl" on its
11119 command line). There is also now an "update" option. This will update
11120 the util/ssleay.num and util/libeay.num files with any new functions.
11122 perl util/mkdef.pl crypto ssl update
11123 it will update them.
11126 *) Overhauled the Perl interface (perl/*):
11127 - ported BN stuff to OpenSSL's different BN library
11128 - made the perl/ source tree CVS-aware
11129 - renamed the package from SSLeay to OpenSSL (the files still contain
11130 their history because I've copied them in the repository)
11131 - removed obsolete files (the test scripts will be replaced
11132 by better Test::Harness variants in the future)
11133 [Ralf S. Engelschall]
11135 *) First cut for a very conservative source tree cleanup:
11136 1. merge various obsolete readme texts into doc/ssleay.txt
11137 where we collect the old documents and readme texts.
11138 2. remove the first part of files where I'm already sure that we no
11139 longer need them because of three reasons: either they are just temporary
11140 files which were left by Eric or they are preserved original files where
11141 I've verified that the diff is also available in the CVS via "cvs diff
11142 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11143 the crypto/md/ stuff).
11144 [Ralf S. Engelschall]
11146 *) More extension code. Incomplete support for subject and issuer alt
11147 name, issuer and authority key id. Change the i2v function parameters
11148 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11149 what that's for :-) Fix to ASN1 macro which messed up
11150 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11153 *) Preliminary support for ENUMERATED type. This is largely copied from the
11157 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11158 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11160 *) Make sure `make rehash' target really finds the `openssl' program.
11161 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11163 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11164 like to hear about it if this slows down other processors.
11167 *) Add CygWin32 platform information to Configure script.
11168 [Alan Batie <batie@aahz.jf.intel.com>]
11170 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11171 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11173 *) New program nseq to manipulate netscape certificate sequences
11176 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11180 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11181 but the BN code had some problems that would cause failures when
11182 doing certificate verification and some other functions.
11183 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11185 *) Add ASN1 and PEM code to support netscape certificate sequences.
11188 *) Add ASN1 and PEM code to support netscape certificate sequences.
11191 *) Add several PKIX and private extended key usage OIDs.
11194 *) Modify the 'ca' program to handle the new extension code. Modify
11195 openssl.cnf for new extension format, add comments.
11198 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11199 and add a sample to openssl.cnf so req -x509 now adds appropriate
11203 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11204 error code, add initial support to X509_print() and x509 application.
11207 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11208 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11209 stuff is currently isolated and isn't even compiled yet.
11212 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11213 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11214 Removed the versions check from X509 routines when loading extensions:
11215 this allows certain broken certificates that don't set the version
11216 properly to be processed.
11219 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11220 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11221 can still be regenerated with "make depend".
11224 *) Spelling mistake in C version of CAST-128.
11225 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11227 *) Changes to the error generation code. The perl script err-code.pl
11228 now reads in the old error codes and retains the old numbers, only
11229 adding new ones if necessary. It also only changes the .err files if new
11230 codes are added. The makefiles have been modified to only insert errors
11231 when needed (to avoid needlessly modifying header files). This is done
11232 by only inserting errors if the .err file is newer than the auto generated
11233 C file. To rebuild all the error codes from scratch (the old behaviour)
11234 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11235 or delete all the .err files.
11238 *) CAST-128 was incorrectly implemented for short keys. The C version has
11239 been fixed, but is untested. The assembler versions are also fixed, but
11240 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11241 to regenerate it if needed.
11242 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11243 Hagino <itojun@kame.net>]
11245 *) File was opened incorrectly in randfile.c.
11246 [Ulf Möller <ulf@fitug.de>]
11248 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11249 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11250 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11251 al: it's just almost always a UTCTime. Note this patch adds new error
11252 codes so do a "make errors" if there are problems.
11255 *) Correct Linux 1 recognition in config.
11256 [Ulf Möller <ulf@fitug.de>]
11258 *) Remove pointless MD5 hash when using DSA keys in ca.
11259 [Anonymous <nobody@replay.com>]
11261 *) Generate an error if given an empty string as a cert directory. Also
11262 generate an error if handed NULL (previously returned 0 to indicate an
11263 error, but didn't set one).
11264 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11266 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11269 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11270 parameters. This was causing a warning which killed off the Win32 compile.
11273 *) Remove C++ style comments from crypto/bn/bn_local.h.
11274 [Neil Costigan <neil.costigan@celocom.com>]
11276 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11277 based on a text string, looking up short and long names and finally
11278 "dot" format. The "dot" format stuff didn't work. Added new function
11279 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11280 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11281 OID is not part of the table.
11284 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11285 X509_LOOKUP_by_alias().
11288 *) Sort openssl functions by name.
11291 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11292 encryption from sample DSA keys (in case anyone is interested the password
11296 *) Make _all_ *_free functions accept a NULL pointer.
11297 [Frans Heymans <fheymans@isaserver.be>]
11299 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11301 [Anonymous <nobody@replay.com>]
11303 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11304 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11306 *) Don't blow it for numeric -newkey arguments to apps/req.
11307 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11309 *) Temp key "for export" tests were wrong in s3_srvr.c.
11310 [Anonymous <nobody@replay.com>]
11312 *) Add prototype for temp key callback functions
11313 SSL_CTX_set_tmp_{rsa,dh}_callback().
11316 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11317 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11320 *) X509_name_add_entry() freed the wrong thing after an error.
11321 [Arne Ansper <arne@ats.cyber.ee>]
11323 *) rsa_eay.c would attempt to free a NULL context.
11324 [Arne Ansper <arne@ats.cyber.ee>]
11326 *) BIO_s_socket() had a broken should_retry() on Windoze.
11327 [Arne Ansper <arne@ats.cyber.ee>]
11329 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11330 [Arne Ansper <arne@ats.cyber.ee>]
11332 *) Make sure the already existing X509_STORE->depth variable is initialized
11333 in X509_STORE_new(), but document the fact that this variable is still
11334 unused in the certificate verification process.
11335 [Ralf S. Engelschall]
11337 *) Fix the various library and apps files to free up pkeys obtained from
11338 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11341 *) Fix reference counting in X509_PUBKEY_get(). This makes
11342 demos/maurice/example2.c work, amongst others, probably.
11343 [Steve Henson and Ben Laurie]
11345 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11346 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11347 are no longer created. This way we have a single and consistent command
11348 line interface `openssl <command>', similar to `cvs <command>'.
11349 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11351 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11352 BIT STRING wrapper always have zero unused bits.
11355 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11358 *) Make the top-level INSTALL documentation easier to understand.
11361 *) Makefiles updated to exit if an error occurs in a sub-directory
11362 make (including if user presses ^C) [Paul Sutton]
11364 *) Make Montgomery context stuff explicit in RSA data structure.
11367 *) Fix build order of pem and err to allow for generated pem.h.
11370 *) Fix renumbering bug in X509_NAME_delete_entry().
11373 *) Enhanced the err-ins.pl script so it makes the error library number
11374 global and can add a library name. This is needed for external ASN1 and
11375 other error libraries.
11378 *) Fixed sk_insert which never worked properly.
11381 *) Fix ASN1 macros so they can handle indefinite length constructed
11382 EXPLICIT tags. Some non standard certificates use these: they can now
11386 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11387 into a single doc/ssleay.txt bundle. This way the information is still
11388 preserved but no longer messes up this directory. Now it's new room for
11389 the new set of documentation files.
11390 [Ralf S. Engelschall]
11392 *) SETs were incorrectly DER encoded. This was a major pain, because they
11393 shared code with SEQUENCEs, which aren't coded the same. This means that
11394 almost everything to do with SETs or SEQUENCEs has either changed name or
11395 number of arguments.
11396 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11398 *) Fix test data to work with the above.
11401 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11402 was already fixed by Eric for 0.9.1 it seems.
11403 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11405 *) Autodetect FreeBSD3.
11408 *) Fix various bugs in Configure. This affects the following platforms:
11412 unixware-2.0-pentium
11416 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11417 before they are needed.
11420 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11424 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11426 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11427 changed SSLeay to OpenSSL in version strings.
11428 [Ralf S. Engelschall]
11430 *) Some fixups to the top-level documents.
11433 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11434 because the symlink to include/ was missing.
11435 [Ralf S. Engelschall]
11437 *) Incorporated the popular no-RSA/DSA-only patches
11438 which allow to compile a RSA-free SSLeay.
11439 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11441 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11442 when "ssleay" is still not found.
11443 [Ralf S. Engelschall]
11445 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11446 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11448 *) Updated the README file.
11449 [Ralf S. Engelschall]
11451 *) Added various .cvsignore files in the CVS repository subdirs
11452 to make a "cvs update" really silent.
11453 [Ralf S. Engelschall]
11455 *) Recompiled the error-definition header files and added
11456 missing symbols to the Win32 linker tables.
11457 [Ralf S. Engelschall]
11459 *) Cleaned up the top-level documents;
11460 o new files: CHANGES and LICENSE
11461 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11462 o merged COPYRIGHT into LICENSE
11463 o removed obsolete TODO file
11464 o renamed MICROSOFT to INSTALL.W32
11465 [Ralf S. Engelschall]
11467 *) Removed dummy files from the 0.9.1b source tree:
11468 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11469 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11470 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11471 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11472 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11473 [Ralf S. Engelschall]
11475 *) Added various platform portability fixes.
11478 *) The Genesis of the OpenSSL rpject:
11479 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11480 Young and Tim J. Hudson created while they were working for C2Net until
11482 [The OpenSSL Project]
11485 Changes between 0.9.0b and 0.9.1b [not released]
11487 *) Updated a few CA certificates under certs/
11490 *) Changed some BIGNUM api stuff.
11493 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11494 DGUX x86, Linux Alpha, etc.
11497 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11498 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11502 *) Add -strparse option to asn1pars program which parses nested
11504 [Dr Stephen Henson <shenson@bigfoot.com>]
11506 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11509 *) DSA fix for "ca" program.
11512 *) Added "-genkey" option to "dsaparam" program.
11515 *) Added RIPE MD160 (rmd160) message digest.
11518 *) Added -a (all) option to "ssleay version" command.
11521 *) Added PLATFORM define which is the id given to Configure.
11524 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11527 *) Extended the ASN.1 parser routines.
11530 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11533 *) Added a BN_CTX to the BN library.
11536 *) Fixed the weak key values in DES library
11539 *) Changed API in EVP library for cipher aliases.
11542 *) Added support for RC2/64bit cipher.
11545 *) Converted the lhash library to the crypto/mem.c functions.
11548 *) Added more recognized ASN.1 object ids.
11551 *) Added more RSA padding checks for SSL/TLS.
11554 *) Added BIO proxy/filter functionality.
11557 *) Added extra_certs to SSL_CTX which can be used
11558 send extra CA certificates to the client in the CA cert chain sending
11559 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11562 *) Now Fortezza is denied in the authentication phase because
11563 this is key exchange mechanism is not supported by SSLeay at all.
11566 *) Additional PKCS1 checks.
11569 *) Support the string "TLSv1" for all TLS v1 ciphers.
11572 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11573 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11576 *) Fixed a few memory leaks.
11579 *) Fixed various code and comment typos.
11582 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11583 bytes sent in the client random.
11584 [Edward Bishop <ebishop@spyglass.com>]