5 Changes between 1.0.2e and 1.1.0 [xx XXX xxxx]
7 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
9 Obtaining and performing DNSSEC validation of TLSA records is
10 the application's responsibility. The application provides
11 the TLSA records of its choice to OpenSSL, and these are then
12 used to authenticate the peer.
14 The TLSA records need not even come from DNS. They can, for
15 example, be used to implement local end-entity certificate or
16 trust-anchor "pinning", where the "pin" data takes the form
17 of TLSA records, which can augment or replace verification
18 based on the usual WebPKI public certification authorities.
21 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
22 continues to support deprecated interfaces in default builds.
23 However, applications are strongly advised to compile their
24 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
25 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
26 or the 1.1.0 releases.
28 In environments in which all applications have been ported to
29 not use any deprecated interfaces OpenSSL's Configure script
30 should be used with the --api=1.1.0 option to entirely remove
31 support for the deprecated features from the library and
32 unconditionally disable them in the installed headers.
33 Essentially the same effect can be achieved with the "no-deprecated"
34 argument to Configure, except that this will always restrict
35 the build to just the latest API, rather than a fixed API
38 As applications are ported to future revisions of the API,
39 they should update their compile-time OPENSSL_API_COMPAT define
40 accordingly, but in most cases should be able to continue to
41 compile with later releases.
43 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
44 0x10000000L and 0x00908000L, respectively. However those
45 versions did not support the OPENSSL_API_COMPAT feature, and
46 so applications are not typically tested for explicit support
47 of just the undeprecated features of either release.
50 *) Add support for setting the minimum and maximum supported protocol.
51 It can bet set via the SSL_set_min_proto_version() and
52 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
53 MaxProtcol. It's recommended to use the new APIs to disable
54 protocols instead of disabling individual protocols using
55 SSL_set_options() or SSL_CONF's Protocol. This change also
56 removes support for disabling TLS 1.2 in the OpenSSL TLS
57 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
60 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
63 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
64 and integrates ECDSA and ECDH functionality into EC. Implementations can
65 now redirect key generation and no longer need to convert to or from
68 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
69 include the ec.h header file instead.
72 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
73 ciphers who are no longer supported and drops support the ephemeral RSA key
74 exchange. The LOW ciphers currently doesn't have any ciphers in it.
77 *) Make EVP_MD_CTX, EVP_MD and HMAC_CTX opaque. For HMAC_CTX, the
78 following constructors and destructors were added:
80 HMAC_CTX *HMAC_CTX_new(void);
81 void HMAC_CTX_free(HMAC_CTX *ctx);
83 For EVP_MD, a complete API to create, fill and destroy such
84 methods has been added. See EVP_MD_meth_new(3) for
88 1) HMAC_CTX_cleanup() and EVP_MD_CTX_cleanup() were removed,
89 HMAC_CTX_init() and EVP_MD_CTX_init() should be called instead
90 to reinitialise and already created structure. Also,
91 HMAC_CTX_init() and EVP_MD_CTX_init() now return 0 for failure
92 and 1 for success (they previously had the return type void).
93 2) For consistency with the majority of our object creators and
94 destructors, EVP_MD_CTX_(create|destroy) were renamed to
95 EVP_MD_CTX_(new|free). The old names are retained as macros
96 for deprecated builds.
99 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
100 cryptographic operations to be performed asynchronously as long as an
101 asynchronous capable engine is used. See the ASYNC_start_job() man page for
102 further details. Libssl has also had this capability integrated with the
103 introduction of the new mode SSL_MODE_ASYNC and associated error
104 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
105 pages. This work was developed in partnership with Intel Corp.
108 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
109 always enabled now. If you want to disable the support you should
110 exclude it using the list of supported ciphers. This also means that the
111 "-no_ecdhe" option has been removed from s_server.
114 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
115 SSL_{CTX_}set1_curves() which can set a list.
118 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
119 curve you want to support using SSL_{CTX_}set1_curves().
122 *) State machine rewrite. The state machine code has been significantly
123 refactored in order to remove much duplication of code and solve issues
124 with the old code (see ssl/statem/README for further details). This change
125 does have some associated API changes. Notably the SSL_state() function
126 has been removed and replaced by SSL_get_state which now returns an
127 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
128 altogether. The previous handshake states defined in ssl.h and ssl3.h have
132 *) All instances of the string "ssleay" in the public API were replaced
133 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
134 Some error codes related to internal RSA_eay API's were renamed.
137 *) The demo files in crypto/threads were moved to demo/threads.
140 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
142 [Matt Caswell, Rich Salz]
144 *) New ASN.1 embed macro.
146 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
147 structure is not allocated: it is part of the parent. That is instead of
155 This reduces memory fragmentation and make it impossible to accidentally
156 set a mandatory field to NULL.
158 This currently only works for some fields specifically a SEQUENCE, CHOICE,
159 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
160 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
164 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
167 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
168 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
169 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
170 DES and RC4 ciphersuites.
173 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
174 This changes the decoding behaviour for some invalid messages,
175 though the change is mostly in the more lenient direction, and
176 legacy behaviour is preserved as much as possible.
179 *) Fix no-stdio build.
180 [ David Woodhouse <David.Woodhouse@intel.com> and also
181 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
183 *) New testing framework
184 The testing framework has been largely rewritten and is now using
185 perl and the perl modules Test::Harness and an extended variant of
186 Test::More called OpenSSL::Test to do its work. All test scripts in
187 test/ have been rewritten into test recipes, and all direct calls to
188 executables in test/Makefile have become individual recipes using the
189 simplified testing OpenSSL::Test::Simple.
191 For documentation on our testing modules, do:
193 perldoc test/testlib/OpenSSL/Test/Simple.pm
194 perldoc test/testlib/OpenSSL/Test.pm
198 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
199 are used; the latter aborts on memory leaks (usually checked on exit).
200 Some undocumented "set malloc, etc., hooks" functions were removed
201 and others were changed. All are now documented.
204 *) In DSA_generate_parameters_ex, if the provided seed is too short,
206 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
208 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
209 from RFC4279, RFC4785, RFC5487, RFC5489.
211 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
212 original RSA_PSK patch.
215 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
216 era flag was never set throughout the codebase (only read). Also removed
217 SSL3_FLAGS_POP_BUFFER which was only used if
218 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
221 *) Changed the default name options in the "ca", "crl", "req" and "x509"
222 to be "oneline" instead of "compat".
225 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
226 not aware of clients that still exhibit this bug, and the workaround
227 hasn't been working properly for a while.
230 *) The return type of BIO_number_read() and BIO_number_written() as well as
231 the corresponding num_read and num_write members in the BIO structure has
232 changed from unsigned long to uint64_t. On platforms where an unsigned
233 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
237 *) Given the pervasive nature of TLS extensions it is inadvisable to run
238 OpenSSL without support for them. It also means that maintaining
239 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
240 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
243 *) Removed support for the two export grade static DH ciphersuites
244 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
245 were newly added (along with a number of other static DH ciphersuites) to
246 1.0.2. However the two export ones have *never* worked since they were
247 introduced. It seems strange in any case to be adding new export
248 ciphersuites, and given "logjam" it also does not seem correct to fix them.
251 *) Version negotiation has been rewritten. In particular SSLv23_method(),
252 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
253 and turned into macros which simply call the new preferred function names
254 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
255 should use the new names instead. Also as part of this change the ssl23.h
256 header file has been removed.
259 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
260 code and the associated standard is no longer considered fit-for-purpose.
263 *) RT2547 was closed. When generating a private key, try to make the
264 output file readable only by the owner. This behavior change might
265 be noticeable when interacting with other software.
267 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
271 *) Added HTTP GET support to the ocsp command.
274 *) Changed default digest for the dgst and enc commands from MD5 to
278 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
281 *) Added support for TLS extended master secret from
282 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
283 initial patch which was a great help during development.
286 *) All libssl internal structures have been removed from the public header
287 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
288 now redundant). Users should not attempt to access internal structures
289 directly. Instead they should use the provided API functions.
292 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
293 Access to deprecated functions can be re-enabled by running config with
294 "enable-deprecated". In addition applications wishing to use deprecated
295 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
296 will, by default, disable some transitive includes that previously existed
297 in the header files (e.g. ec.h will no longer, by default, include bn.h)
300 *) Added support for OCB mode. OpenSSL has been granted a patent license
301 compatible with the OpenSSL license for use of OCB. Details are available
302 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
303 for OCB can be removed by calling config with no-ocb.
306 *) SSLv2 support has been removed. It still supports receiving a SSLv2
307 compatible client hello.
310 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
311 done while fixing the error code for the key-too-small case.
312 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
314 *) CA.sh has been removmed; use CA.pl instead.
317 *) Removed old DES API.
320 *) Remove various unsupported platforms:
326 Sinix/ReliantUNIX RM400
331 16-bit platforms such as WIN16
334 *) Clean up OPENSSL_NO_xxx #define's
335 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
336 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
337 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
338 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
339 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
340 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
341 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
342 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
343 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
344 Remove MS_STATIC; it's a relic from platforms <32 bits.
347 *) Cleaned up dead code
348 Remove all but one '#ifdef undef' which is to be looked at.
351 *) Clean up calling of xxx_free routines.
352 Just like free(), fix most of the xxx_free routines to accept
353 NULL. Remove the non-null checks from callers. Save much code.
356 *) Add secure heap for storage of private keys (when possible).
357 Add BIO_s_secmem(), CBIGNUM, etc.
358 Contributed by Akamai Technologies under our Corporate CLA.
361 *) Experimental support for a new, fast, unbiased prime candidate generator,
362 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
363 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
365 *) New output format NSS in the sess_id command line tool. This allows
366 exporting the session id and the master key in NSS keylog format.
367 [Martin Kaiser <martin@kaiser.cx>]
369 *) Harmonize version and its documentation. -f flag is used to display
371 [mancha <mancha1@zoho.com>]
373 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
374 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
375 [mancha <mancha1@zoho.com>]
377 *) Fix some double frees. These are not thought to be exploitable.
378 [mancha <mancha1@zoho.com>]
380 *) A missing bounds check in the handling of the TLS heartbeat extension
381 can be used to reveal up to 64k of memory to a connected client or
384 Thanks for Neel Mehta of Google Security for discovering this bug and to
385 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
386 preparing the fix (CVE-2014-0160)
387 [Adam Langley, Bodo Moeller]
389 *) Fix for the attack described in the paper "Recovering OpenSSL
390 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
391 by Yuval Yarom and Naomi Benger. Details can be obtained from:
392 http://eprint.iacr.org/2014/140
394 Thanks to Yuval Yarom and Naomi Benger for discovering this
395 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
396 [Yuval Yarom and Naomi Benger]
398 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
399 this fixes a limitation in previous versions of OpenSSL.
402 *) Experimental encrypt-then-mac support.
404 Experimental support for encrypt then mac from
405 draft-gutmann-tls-encrypt-then-mac-02.txt
407 To enable it set the appropriate extension number (0x42 for the test
408 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
410 For non-compliant peers (i.e. just about everything) this should have no
413 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
417 *) Add EVP support for key wrapping algorithms, to avoid problems with
418 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
419 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
420 algorithms and include tests cases.
423 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
427 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
428 MGF1 digest and OAEP label.
431 *) Make openssl verify return errors.
432 [Chris Palmer <palmer@google.com> and Ben Laurie]
434 *) New function ASN1_TIME_diff to calculate the difference between two
435 ASN1_TIME structures or one structure and the current time.
438 *) Update fips_test_suite to support multiple command line options. New
439 test to induce all self test errors in sequence and check expected
443 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
444 sign or verify all in one operation.
447 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
448 test programs and fips_test_suite. Includes functionality to parse
449 the minimal script output of fipsalgest.pl directly.
452 *) Add authorisation parameter to FIPS_module_mode_set().
455 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
458 *) Use separate DRBG fields for internal and external flags. New function
459 FIPS_drbg_health_check() to perform on demand health checking. Add
460 generation tests to fips_test_suite with reduced health check interval to
461 demonstrate periodic health checking. Add "nodh" option to
462 fips_test_suite to skip very slow DH test.
465 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
469 *) More extensive health check for DRBG checking many more failure modes.
470 New function FIPS_selftest_drbg_all() to handle every possible DRBG
471 combination: call this in fips_test_suite.
474 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
475 and POST to handle Dual EC cases.
478 *) Add support for canonical generation of DSA parameter 'g'. See
481 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
482 POST to handle HMAC cases.
485 *) Add functions FIPS_module_version() and FIPS_module_version_text()
486 to return numerical and string versions of the FIPS module number.
489 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
490 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
491 outside the validated module in the FIPS capable OpenSSL.
494 *) Minor change to DRBG entropy callback semantics. In some cases
495 there is no multiple of the block length between min_len and
496 max_len. Allow the callback to return more than max_len bytes
497 of entropy but discard any extra: it is the callback's responsibility
498 to ensure that the extra data discarded does not impact the
499 requested amount of entropy.
502 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
503 information in FIPS186-3, SP800-57 and SP800-131A.
506 *) CCM support via EVP. Interface is very similar to GCM case except we
507 must supply all data in one chunk (i.e. no update, final) and the
508 message length must be supplied if AAD is used. Add algorithm test
512 *) Initial version of POST overhaul. Add POST callback to allow the status
513 of POST to be monitored and/or failures induced. Modify fips_test_suite
514 to use callback. Always run all selftests even if one fails.
517 *) XTS support including algorithm test driver in the fips_gcmtest program.
518 Note: this does increase the maximum key length from 32 to 64 bytes but
519 there should be no binary compatibility issues as existing applications
520 will never use XTS mode.
523 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
524 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
525 performs algorithm blocking for unapproved PRNG types. Also do not
526 set PRNG type in FIPS_mode_set(): leave this to the application.
527 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
528 the standard OpenSSL PRNG: set additional data to a date time vector.
531 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
532 This shouldn't present any incompatibility problems because applications
533 shouldn't be using these directly and any that are will need to rethink
534 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
537 *) Extensive self tests and health checking required by SP800-90 DRBG.
538 Remove strength parameter from FIPS_drbg_instantiate and always
539 instantiate at maximum supported strength.
542 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
545 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
548 *) New function DH_compute_key_padded() to compute a DH key and pad with
549 leading zeroes if needed: this complies with SP800-56A et al.
552 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
553 anything, incomplete, subject to change and largely untested at present.
556 *) Modify fipscanisteronly build option to only build the necessary object
557 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
560 *) Add experimental option FIPSSYMS to give all symbols in
561 fipscanister.o and FIPS or fips prefix. This will avoid
562 conflicts with future versions of OpenSSL. Add perl script
563 util/fipsas.pl to preprocess assembly language source files
564 and rename any affected symbols.
567 *) Add selftest checks and algorithm block of non-fips algorithms in
568 FIPS mode. Remove DES2 from selftests.
571 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
572 return internal method without any ENGINE dependencies. Add new
573 tiny fips sign and verify functions.
576 *) New build option no-ec2m to disable characteristic 2 code.
579 *) New build option "fipscanisteronly". This only builds fipscanister.o
580 and (currently) associated fips utilities. Uses the file Makefile.fips
581 instead of Makefile.org as the prototype.
584 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
585 Update fips_gcmtest to use IV generator.
588 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
589 setting output buffer to NULL. The *Final function must be
590 called although it will not retrieve any additional data. The tag
591 can be set or retrieved with a ctrl. The IV length is by default 12
592 bytes (96 bits) but can be set to an alternative value. If the IV
593 length exceeds the maximum IV length (currently 16 bytes) it cannot be
597 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
598 underlying do_cipher function handles all cipher semantics itself
599 including padding and finalisation. This is useful if (for example)
600 an ENGINE cipher handles block padding itself. The behaviour of
601 do_cipher is subtly changed if this flag is set: the return value
602 is the number of characters written to the output buffer (zero is
603 no longer an error code) or a negative error code. Also if the
604 input buffer is NULL and length 0 finalisation should be performed.
607 *) If a candidate issuer certificate is already part of the constructed
608 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
611 *) Improve forward-security support: add functions
613 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
614 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
616 for use by SSL/TLS servers; the callback function will be called whenever a
617 new session is created, and gets to decide whether the session may be
618 cached to make it resumable (return 0) or not (return 1). (As by the
619 SSL/TLS protocol specifications, the session_id sent by the server will be
620 empty to indicate that the session is not resumable; also, the server will
621 not generate RFC 4507 (RFC 5077) session tickets.)
623 A simple reasonable callback implementation is to return is_forward_secure.
624 This parameter will be set to 1 or 0 depending on the ciphersuite selected
625 by the SSL/TLS server library, indicating whether it can provide forward
627 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
629 *) New -verify_name option in command line utilities to set verification
633 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
634 Add CMAC pkey methods.
637 *) Experimental renegotiation in s_server -www mode. If the client
638 browses /reneg connection is renegotiated. If /renegcert it is
639 renegotiated requesting a certificate.
642 *) Add an "external" session cache for debugging purposes to s_server. This
643 should help trace issues which normally are only apparent in deployed
644 multi-process servers.
647 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
648 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
649 BIO_set_cipher() and some obscure PEM functions were changed so they
650 can now return an error. The RAND changes required a change to the
651 RAND_METHOD structure.
654 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
655 a gcc attribute to warn if the result of a function is ignored. This
656 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
657 whose return value is often ignored.
660 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
662 *) BN_mod_exp may produce incorrect results on x86_64
664 There is a carry propagating bug in the x86_64 Montgomery squaring
665 procedure. No EC algorithms are affected. Analysis suggests that attacks
666 against RSA and DSA as a result of this defect would be very difficult to
667 perform and are not believed likely. Attacks against DH are considered just
668 feasible (although very difficult) because most of the work necessary to
669 deduce information about a private key may be performed offline. The amount
670 of resources required for such an attack would be very significant and
671 likely only accessible to a limited number of attackers. An attacker would
672 additionally need online access to an unpatched system using the target
673 private key in a scenario with persistent DH parameters and a private
674 key that is shared between multiple clients. For example this can occur by
675 default in OpenSSL DHE based SSL/TLS ciphersuites.
677 This issue was reported to OpenSSL by Hanno Böck.
681 *) Certificate verify crash with missing PSS parameter
683 The signature verification routines will crash with a NULL pointer
684 dereference if presented with an ASN.1 signature using the RSA PSS
685 algorithm and absent mask generation function parameter. Since these
686 routines are used to verify certificate signature algorithms this can be
687 used to crash any certificate verification operation and exploited in a
688 DoS attack. Any application which performs certificate verification is
689 vulnerable including OpenSSL clients and servers which enable client
692 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
696 *) X509_ATTRIBUTE memory leak
698 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
699 memory. This structure is used by the PKCS#7 and CMS routines so any
700 application which reads PKCS#7 or CMS data from untrusted sources is
701 affected. SSL/TLS is not affected.
703 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
708 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
709 This changes the decoding behaviour for some invalid messages,
710 though the change is mostly in the more lenient direction, and
711 legacy behaviour is preserved as much as possible.
714 *) In DSA_generate_parameters_ex, if the provided seed is too short,
716 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
718 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
720 *) Alternate chains certificate forgery
722 During certificate verfification, OpenSSL will attempt to find an
723 alternative certificate chain if the first attempt to build such a chain
724 fails. An error in the implementation of this logic can mean that an
725 attacker could cause certain checks on untrusted certificates to be
726 bypassed, such as the CA flag, enabling them to use a valid leaf
727 certificate to act as a CA and "issue" an invalid certificate.
729 This issue was reported to OpenSSL by Adam Langley/David Benjamin
733 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
735 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
736 incompatibility in the handling of HMAC. The previous ABI has now been
740 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
742 *) Malformed ECParameters causes infinite loop
744 When processing an ECParameters structure OpenSSL enters an infinite loop
745 if the curve specified is over a specially malformed binary polynomial
748 This can be used to perform denial of service against any
749 system which processes public keys, certificate requests or
750 certificates. This includes TLS clients and TLS servers with
751 client authentication enabled.
753 This issue was reported to OpenSSL by Joseph Barr-Pixton.
757 *) Exploitable out-of-bounds read in X509_cmp_time
759 X509_cmp_time does not properly check the length of the ASN1_TIME
760 string and can read a few bytes out of bounds. In addition,
761 X509_cmp_time accepts an arbitrary number of fractional seconds in the
764 An attacker can use this to craft malformed certificates and CRLs of
765 various sizes and potentially cause a segmentation fault, resulting in
766 a DoS on applications that verify certificates or CRLs. TLS clients
767 that verify CRLs are affected. TLS clients and servers with client
768 authentication enabled may be affected if they use custom verification
771 This issue was reported to OpenSSL by Robert Swiecki (Google), and
772 independently by Hanno Böck.
776 *) PKCS7 crash with missing EnvelopedContent
778 The PKCS#7 parsing code does not handle missing inner EncryptedContent
779 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
780 with missing content and trigger a NULL pointer dereference on parsing.
782 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
783 structures from untrusted sources are affected. OpenSSL clients and
784 servers are not affected.
786 This issue was reported to OpenSSL by Michal Zalewski (Google).
790 *) CMS verify infinite loop with unknown hash function
792 When verifying a signedData message the CMS code can enter an infinite loop
793 if presented with an unknown hash function OID. This can be used to perform
794 denial of service against any system which verifies signedData messages using
796 This issue was reported to OpenSSL by Johannes Bauer.
800 *) Race condition handling NewSessionTicket
802 If a NewSessionTicket is received by a multi-threaded client when attempting to
803 reuse a previous ticket then a race condition can occur potentially leading to
804 a double free of the ticket data.
808 *) Only support 256-bit or stronger elliptic curves with the
809 'ecdh_auto' setting (server) or by default (client). Of supported
810 curves, prefer P-256 (both).
813 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
815 *) ClientHello sigalgs DoS fix
817 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
818 invalid signature algorithms extension a NULL pointer dereference will
819 occur. This can be exploited in a DoS attack against the server.
821 This issue was was reported to OpenSSL by David Ramos of Stanford
824 [Stephen Henson and Matt Caswell]
826 *) Multiblock corrupted pointer fix
828 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
829 feature only applies on 64 bit x86 architecture platforms that support AES
830 NI instructions. A defect in the implementation of "multiblock" can cause
831 OpenSSL's internal write buffer to become incorrectly set to NULL when
832 using non-blocking IO. Typically, when the user application is using a
833 socket BIO for writing, this will only result in a failed connection.
834 However if some other BIO is used then it is likely that a segmentation
835 fault will be triggered, thus enabling a potential DoS attack.
837 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
841 *) Segmentation fault in DTLSv1_listen fix
843 The DTLSv1_listen function is intended to be stateless and processes the
844 initial ClientHello from many peers. It is common for user code to loop
845 over the call to DTLSv1_listen until a valid ClientHello is received with
846 an associated cookie. A defect in the implementation of DTLSv1_listen means
847 that state is preserved in the SSL object from one invocation to the next
848 that can lead to a segmentation fault. Errors processing the initial
849 ClientHello can trigger this scenario. An example of such an error could be
850 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
853 This issue was reported to OpenSSL by Per Allansson.
857 *) Segmentation fault in ASN1_TYPE_cmp fix
859 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
860 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
861 certificate signature algorithm consistency this can be used to crash any
862 certificate verification operation and exploited in a DoS attack. Any
863 application which performs certificate verification is vulnerable including
864 OpenSSL clients and servers which enable client authentication.
868 *) Segmentation fault for invalid PSS parameters fix
870 The signature verification routines will crash with a NULL pointer
871 dereference if presented with an ASN.1 signature using the RSA PSS
872 algorithm and invalid parameters. Since these routines are used to verify
873 certificate signature algorithms this can be used to crash any
874 certificate verification operation and exploited in a DoS attack. Any
875 application which performs certificate verification is vulnerable including
876 OpenSSL clients and servers which enable client authentication.
878 This issue was was reported to OpenSSL by Brian Carpenter.
882 *) ASN.1 structure reuse memory corruption fix
884 Reusing a structure in ASN.1 parsing may allow an attacker to cause
885 memory corruption via an invalid write. Such reuse is and has been
886 strongly discouraged and is believed to be rare.
888 Applications that parse structures containing CHOICE or ANY DEFINED BY
889 components may be affected. Certificate parsing (d2i_X509 and related
890 functions) are however not affected. OpenSSL clients and servers are
895 *) PKCS7 NULL pointer dereferences fix
897 The PKCS#7 parsing code does not handle missing outer ContentInfo
898 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
899 missing content and trigger a NULL pointer dereference on parsing.
901 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
902 otherwise parse PKCS#7 structures from untrusted sources are
903 affected. OpenSSL clients and servers are not affected.
905 This issue was reported to OpenSSL by Michal Zalewski (Google).
909 *) DoS via reachable assert in SSLv2 servers fix
911 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
912 servers that both support SSLv2 and enable export cipher suites by sending
913 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
915 This issue was discovered by Sean Burford (Google) and Emilia Käsper
916 (OpenSSL development team).
920 *) Empty CKE with client auth and DHE fix
922 If client auth is used then a server can seg fault in the event of a DHE
923 ciphersuite being selected and a zero length ClientKeyExchange message
924 being sent by the client. This could be exploited in a DoS attack.
928 *) Handshake with unseeded PRNG fix
930 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
931 with an unseeded PRNG. The conditions are:
932 - The client is on a platform where the PRNG has not been seeded
933 automatically, and the user has not seeded manually
934 - A protocol specific client method version has been used (i.e. not
935 SSL_client_methodv23)
936 - A ciphersuite is used that does not require additional random data from
937 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
939 If the handshake succeeds then the client random that has been used will
940 have been generated from a PRNG with insufficient entropy and therefore the
941 output may be predictable.
943 For example using the following command with an unseeded openssl will
944 succeed on an unpatched platform:
946 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
950 *) Use After Free following d2i_ECPrivatekey error fix
952 A malformed EC private key file consumed via the d2i_ECPrivateKey function
953 could cause a use after free condition. This, in turn, could cause a double
954 free in several private key parsing functions (such as d2i_PrivateKey
955 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
956 for applications that receive EC private keys from untrusted
957 sources. This scenario is considered rare.
959 This issue was discovered by the BoringSSL project and fixed in their
964 *) X509_to_X509_REQ NULL pointer deref fix
966 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
967 the certificate key is invalid. This function is rarely used in practice.
969 This issue was discovered by Brian Carpenter.
973 *) Removed the export ciphers from the DEFAULT ciphers
976 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
978 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
979 ARMv5 through ARMv8, as opposite to "locking" it to single one.
980 So far those who have to target multiple plaforms would compromise
981 and argue that binary targeting say ARMv5 would still execute on
982 ARMv8. "Universal" build resolves this compromise by providing
983 near-optimal performance even on newer platforms.
986 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
987 (other platforms pending).
988 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
990 *) Add support for the SignedCertificateTimestampList certificate and
991 OCSP response extensions from RFC6962.
994 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
995 for corner cases. (Certain input points at infinity could lead to
996 bogus results, with non-infinity inputs mapped to infinity too.)
999 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1000 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1001 common cases are optimized and there still is room for further
1002 improvements. Vector Permutation AES for Altivec is also added.
1005 *) Add support for little-endian ppc64 Linux target.
1006 [Marcelo Cerri (IBM)]
1008 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1009 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1010 are optimized and there still is room for further improvements.
1011 Both 32- and 64-bit modes are supported.
1012 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1014 *) Improved ARMv7 NEON support.
1017 *) Support for SPARC Architecture 2011 crypto extensions, first
1018 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1019 SHA256/512, MD5, GHASH and modular exponentiation.
1020 [Andy Polyakov, David Miller]
1022 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1024 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1026 *) Support for new and upcoming Intel processors, including AVX2,
1027 BMI and SHA ISA extensions. This includes additional "stitched"
1028 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1031 This work was sponsored by Intel Corp.
1034 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1035 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1036 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1039 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1040 this fixes a limiation in previous versions of OpenSSL.
1043 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1044 MGF1 digest and OAEP label.
1047 *) Add EVP support for key wrapping algorithms, to avoid problems with
1048 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1049 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1050 algorithms and include tests cases.
1053 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1055 [Douglas E. Engert, Steve Henson]
1057 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1058 difference in days and seconds between two tm or ASN1_TIME structures.
1061 *) Add -rev test option to s_server to just reverse order of characters
1062 received by client and send back to server. Also prints an abbreviated
1063 summary of the connection parameters.
1066 *) New option -brief for s_client and s_server to print out a brief summary
1067 of connection parameters.
1070 *) Add callbacks for arbitrary TLS extensions.
1071 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1073 *) New option -crl_download in several openssl utilities to download CRLs
1074 from CRLDP extension in certificates.
1077 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1080 *) New function X509_CRL_diff to generate a delta CRL from the difference
1081 of two full CRLs. Add support to "crl" utility.
1084 *) New functions to set lookup_crls function and to retrieve
1085 X509_STORE from X509_STORE_CTX.
1088 *) Print out deprecated issuer and subject unique ID fields in
1092 *) Extend OCSP I/O functions so they can be used for simple general purpose
1093 HTTP as well as OCSP. New wrapper function which can be used to download
1094 CRLs using the OCSP API.
1097 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1100 *) SSL_CONF* functions. These provide a common framework for application
1101 configuration using configuration files or command lines.
1104 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1105 message callback and prints the results. Needs compile time option
1106 "enable-ssl-trace". New options to s_client and s_server to enable
1110 *) New ctrl and macro to retrieve supported points extensions.
1111 Print out extension in s_server and s_client.
1114 *) New functions to retrieve certificate signature and signature
1118 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1122 *) New Suite B modes for TLS code. These use and enforce the requirements
1123 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1124 only use Suite B curves. The Suite B modes can be set by using the
1125 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1128 *) New chain verification flags for Suite B levels of security. Check
1129 algorithms are acceptable when flags are set in X509_verify_cert.
1132 *) Make tls1_check_chain return a set of flags indicating checks passed
1133 by a certificate chain. Add additional tests to handle client
1134 certificates: checks for matching certificate type and issuer name
1138 *) If an attempt is made to use a signature algorithm not in the peer
1139 preference list abort the handshake. If client has no suitable
1140 signature algorithms in response to a certificate request do not
1141 use the certificate.
1144 *) If server EC tmp key is not in client preference list abort handshake.
1147 *) Add support for certificate stores in CERT structure. This makes it
1148 possible to have different stores per SSL structure or one store in
1149 the parent SSL_CTX. Include distint stores for certificate chain
1150 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1151 to build and store a certificate chain in CERT structure: returing
1152 an error if the chain cannot be built: this will allow applications
1153 to test if a chain is correctly configured.
1155 Note: if the CERT based stores are not set then the parent SSL_CTX
1156 store is used to retain compatibility with existing behaviour.
1160 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1161 mask based on the current session, check mask when sending client
1162 hello and checking the requested ciphersuite.
1165 *) New ctrls to retrieve and set certificate types in a certificate
1166 request message. Print out received values in s_client. If certificate
1167 types is not set with custom values set sensible values based on
1168 supported signature algorithms.
1171 *) Support for distinct client and server supported signature algorithms.
1174 *) Add certificate callback. If set this is called whenever a certificate
1175 is required by client or server. An application can decide which
1176 certificate chain to present based on arbitrary criteria: for example
1177 supported signature algorithms. Add very simple example to s_server.
1178 This fixes many of the problems and restrictions of the existing client
1179 certificate callback: for example you can now clear an existing
1180 certificate and specify the whole chain.
1183 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1184 the certificate can be used for (if anything). Set valid_flags field
1185 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1186 to have similar checks in it.
1188 Add new "cert_flags" field to CERT structure and include a "strict mode".
1189 This enforces some TLS certificate requirements (such as only permitting
1190 certificate signature algorithms contained in the supported algorithms
1191 extension) which some implementations ignore: this option should be used
1192 with caution as it could cause interoperability issues.
1195 *) Update and tidy signature algorithm extension processing. Work out
1196 shared signature algorithms based on preferences and peer algorithms
1197 and print them out in s_client and s_server. Abort handshake if no
1198 shared signature algorithms.
1201 *) Add new functions to allow customised supported signature algorithms
1202 for SSL and SSL_CTX structures. Add options to s_client and s_server
1206 *) New function SSL_certs_clear() to delete all references to certificates
1207 from an SSL structure. Before this once a certificate had been added
1208 it couldn't be removed.
1211 *) Integrate hostname, email address and IP address checking with certificate
1212 verification. New verify options supporting checking in opensl utility.
1215 *) Fixes and wildcard matching support to hostname and email checking
1216 functions. Add manual page.
1217 [Florian Weimer (Red Hat Product Security Team)]
1219 *) New functions to check a hostname email or IP address against a
1220 certificate. Add options x509 utility to print results of checks against
1224 *) Fix OCSP checking.
1225 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1227 *) Initial experimental support for explicitly trusted non-root CAs.
1228 OpenSSL still tries to build a complete chain to a root but if an
1229 intermediate CA has a trust setting included that is used. The first
1230 setting is used: whether to trust (e.g., -addtrust option to the x509
1234 *) Add -trusted_first option which attempts to find certificates in the
1235 trusted store even if an untrusted chain is also supplied.
1238 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1239 platform support for Linux and Android.
1242 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1245 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1246 When in FIPS mode the approved implementations are used as normal,
1247 when not in FIPS mode the internal unapproved versions are used instead.
1248 This means that the FIPS capable OpenSSL isn't forced to use the
1249 (often lower performance) FIPS implementations outside FIPS mode.
1252 *) Transparently support X9.42 DH parameters when calling
1253 PEM_read_bio_DHparameters. This means existing applications can handle
1254 the new parameter format automatically.
1257 *) Initial experimental support for X9.42 DH parameter format: mainly
1258 to support use of 'q' parameter for RFC5114 parameters.
1261 *) Add DH parameters from RFC5114 including test data to dhtest.
1264 *) Support for automatic EC temporary key parameter selection. If enabled
1265 the most preferred EC parameters are automatically used instead of
1266 hardcoded fixed parameters. Now a server just has to call:
1267 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1268 support ECDH and use the most appropriate parameters.
1271 *) Enhance and tidy EC curve and point format TLS extension code. Use
1272 static structures instead of allocation if default values are used.
1273 New ctrls to set curves we wish to support and to retrieve shared curves.
1274 Print out shared curves in s_server. New options to s_server and s_client
1275 to set list of supported curves.
1278 *) New ctrls to retrieve supported signature algorithms and
1279 supported curve values as an array of NIDs. Extend openssl utility
1280 to print out received values.
1283 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1284 between NIDs and the more common NIST names such as "P-256". Enhance
1285 ecparam utility and ECC method to recognise the NIST names for curves.
1288 *) Enhance SSL/TLS certificate chain handling to support different
1289 chains for each certificate instead of one chain in the parent SSL_CTX.
1292 *) Support for fixed DH ciphersuite client authentication: where both
1293 server and client use DH certificates with common parameters.
1296 *) Support for fixed DH ciphersuites: those requiring DH server
1300 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1302 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1303 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1304 X509_CINF_get_signature were reverted post internal team review.
1306 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1308 *) Build fixes for the Windows and OpenVMS platforms
1309 [Matt Caswell and Richard Levitte]
1311 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1313 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1314 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1315 dereference. This could lead to a Denial Of Service attack. Thanks to
1316 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1320 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1321 dtls1_buffer_record function under certain conditions. In particular this
1322 could occur if an attacker sent repeated DTLS records with the same
1323 sequence number but for the next epoch. The memory leak could be exploited
1324 by an attacker in a Denial of Service attack through memory exhaustion.
1325 Thanks to Chris Mueller for reporting this issue.
1329 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1330 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1331 method would be set to NULL which could later result in a NULL pointer
1332 dereference. Thanks to Frank Schmirler for reporting this issue.
1336 *) Abort handshake if server key exchange message is omitted for ephemeral
1339 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1340 reporting this issue.
1344 *) Remove non-export ephemeral RSA code on client and server. This code
1345 violated the TLS standard by allowing the use of temporary RSA keys in
1346 non-export ciphersuites and could be used by a server to effectively
1347 downgrade the RSA key length used to a value smaller than the server
1348 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1349 INRIA or reporting this issue.
1353 *) Fixed issue where DH client certificates are accepted without verification.
1354 An OpenSSL server will accept a DH certificate for client authentication
1355 without the certificate verify message. This effectively allows a client to
1356 authenticate without the use of a private key. This only affects servers
1357 which trust a client certificate authority which issues certificates
1358 containing DH keys: these are extremely rare and hardly ever encountered.
1359 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1364 *) Ensure that the session ID context of an SSL is updated when its
1365 SSL_CTX is updated via SSL_set_SSL_CTX.
1367 The session ID context is typically set from the parent SSL_CTX,
1368 and can vary with the CTX.
1371 *) Fix various certificate fingerprint issues.
1373 By using non-DER or invalid encodings outside the signed portion of a
1374 certificate the fingerprint can be changed without breaking the signature.
1375 Although no details of the signed portion of the certificate can be changed
1376 this can cause problems with some applications: e.g. those using the
1377 certificate fingerprint for blacklists.
1379 1. Reject signatures with non zero unused bits.
1381 If the BIT STRING containing the signature has non zero unused bits reject
1382 the signature. All current signature algorithms require zero unused bits.
1384 2. Check certificate algorithm consistency.
1386 Check the AlgorithmIdentifier inside TBS matches the one in the
1387 certificate signature. NB: this will result in signature failure
1388 errors for some broken certificates.
1390 Thanks to Konrad Kraszewski from Google for reporting this issue.
1392 3. Check DSA/ECDSA signatures use DER.
1394 Reencode DSA/ECDSA signatures and compare with the original received
1395 signature. Return an error if there is a mismatch.
1397 This will reject various cases including garbage after signature
1398 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1399 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1400 (negative or with leading zeroes).
1402 Further analysis was conducted and fixes were developed by Stephen Henson
1403 of the OpenSSL core team.
1408 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1409 results on some platforms, including x86_64. This bug occurs at random
1410 with a very low probability, and is not known to be exploitable in any
1411 way, though its exact impact is difficult to determine. Thanks to Pieter
1412 Wuille (Blockstream) who reported this issue and also suggested an initial
1413 fix. Further analysis was conducted by the OpenSSL development team and
1414 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1415 the OpenSSL core team.
1419 *) Do not resume sessions on the server if the negotiated protocol
1420 version does not match the session's version. Resuming with a different
1421 version, while not strictly forbidden by the RFC, is of questionable
1422 sanity and breaks all known clients.
1423 [David Benjamin, Emilia Käsper]
1425 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1426 early CCS messages during renegotiation. (Note that because
1427 renegotiation is encrypted, this early CCS was not exploitable.)
1430 *) Tighten client-side session ticket handling during renegotiation:
1431 ensure that the client only accepts a session ticket if the server sends
1432 the extension anew in the ServerHello. Previously, a TLS client would
1433 reuse the old extension state and thus accept a session ticket if one was
1434 announced in the initial ServerHello.
1436 Similarly, ensure that the client requires a session ticket if one
1437 was advertised in the ServerHello. Previously, a TLS client would
1438 ignore a missing NewSessionTicket message.
1441 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1443 *) SRTP Memory Leak.
1445 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1446 sends a carefully crafted handshake message, to cause OpenSSL to fail
1447 to free up to 64k of memory causing a memory leak. This could be
1448 exploited in a Denial Of Service attack. This issue affects OpenSSL
1449 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1450 whether SRTP is used or configured. Implementations of OpenSSL that
1451 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1453 The fix was developed by the OpenSSL team.
1457 *) Session Ticket Memory Leak.
1459 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1460 integrity of that ticket is first verified. In the event of a session
1461 ticket integrity check failing, OpenSSL will fail to free memory
1462 causing a memory leak. By sending a large number of invalid session
1463 tickets an attacker could exploit this issue in a Denial Of Service
1468 *) Build option no-ssl3 is incomplete.
1470 When OpenSSL is configured with "no-ssl3" as a build option, servers
1471 could accept and complete a SSL 3.0 handshake, and clients could be
1472 configured to send them.
1474 [Akamai and the OpenSSL team]
1476 *) Add support for TLS_FALLBACK_SCSV.
1477 Client applications doing fallback retries should call
1478 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1480 [Adam Langley, Bodo Moeller]
1482 *) Add additional DigestInfo checks.
1484 Reencode DigestInto in DER and check against the original when
1485 verifying RSA signature: this will reject any improperly encoded
1486 DigestInfo structures.
1488 Note: this is a precautionary measure and no attacks are currently known.
1492 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1494 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1495 SRP code can be overrun an internal buffer. Add sanity check that
1496 g, A, B < N to SRP code.
1498 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1499 Group for discovering this issue.
1503 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1504 TLS 1.0 instead of higher protocol versions when the ClientHello message
1505 is badly fragmented. This allows a man-in-the-middle attacker to force a
1506 downgrade to TLS 1.0 even if both the server and the client support a
1507 higher protocol version, by modifying the client's TLS records.
1509 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1510 researching this issue.
1514 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1515 to a denial of service attack. A malicious server can crash the client
1516 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1517 ciphersuite and sending carefully crafted handshake messages.
1519 Thanks to Felix Gröbert (Google) for discovering and researching this
1524 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1525 to leak memory. This can be exploited through a Denial of Service attack.
1526 Thanks to Adam Langley for discovering and researching this issue.
1530 *) An attacker can force openssl to consume large amounts of memory whilst
1531 processing DTLS handshake messages. This can be exploited through a
1532 Denial of Service attack.
1533 Thanks to Adam Langley for discovering and researching this issue.
1537 *) An attacker can force an error condition which causes openssl to crash
1538 whilst processing DTLS packets due to memory being freed twice. This
1539 can be exploited through a Denial of Service attack.
1540 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1545 *) If a multithreaded client connects to a malicious server using a resumed
1546 session and the server sends an ec point format extension it could write
1547 up to 255 bytes to freed memory.
1549 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1554 *) A malicious server can crash an OpenSSL client with a null pointer
1555 dereference (read) by specifying an SRP ciphersuite even though it was not
1556 properly negotiated with the client. This can be exploited through a
1557 Denial of Service attack.
1559 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1560 discovering and researching this issue.
1564 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1565 X509_name_oneline, X509_name_print_ex et al. to leak some information
1566 from the stack. Applications may be affected if they echo pretty printing
1567 output to the attacker.
1569 Thanks to Ivan Fratric (Google) for discovering this issue.
1571 [Emilia Käsper, and Steve Henson]
1573 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1574 for corner cases. (Certain input points at infinity could lead to
1575 bogus results, with non-infinity inputs mapped to infinity too.)
1578 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1580 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1581 handshake can force the use of weak keying material in OpenSSL
1582 SSL/TLS clients and servers.
1584 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1585 researching this issue. (CVE-2014-0224)
1586 [KIKUCHI Masashi, Steve Henson]
1588 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1589 OpenSSL DTLS client the code can be made to recurse eventually crashing
1592 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1594 [Imre Rad, Steve Henson]
1596 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1597 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1598 client or server. This is potentially exploitable to run arbitrary
1599 code on a vulnerable client or server.
1601 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1602 [Jüri Aedla, Steve Henson]
1604 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1605 are subject to a denial of service attack.
1607 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1608 this issue. (CVE-2014-3470)
1609 [Felix Gröbert, Ivan Fratric, Steve Henson]
1611 *) Harmonize version and its documentation. -f flag is used to display
1613 [mancha <mancha1@zoho.com>]
1615 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1616 in i2d_ECPrivateKey.
1617 [mancha <mancha1@zoho.com>]
1619 *) Fix some double frees. These are not thought to be exploitable.
1620 [mancha <mancha1@zoho.com>]
1622 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1624 *) A missing bounds check in the handling of the TLS heartbeat extension
1625 can be used to reveal up to 64k of memory to a connected client or
1628 Thanks for Neel Mehta of Google Security for discovering this bug and to
1629 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1630 preparing the fix (CVE-2014-0160)
1631 [Adam Langley, Bodo Moeller]
1633 *) Fix for the attack described in the paper "Recovering OpenSSL
1634 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1635 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1636 http://eprint.iacr.org/2014/140
1638 Thanks to Yuval Yarom and Naomi Benger for discovering this
1639 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1640 [Yuval Yarom and Naomi Benger]
1642 *) TLS pad extension: draft-agl-tls-padding-03
1644 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1645 TLS client Hello record length value would otherwise be > 255 and
1646 less that 512 pad with a dummy extension containing zeroes so it
1647 is at least 512 bytes long.
1649 [Adam Langley, Steve Henson]
1651 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1653 *) Fix for TLS record tampering bug. A carefully crafted invalid
1654 handshake could crash OpenSSL with a NULL pointer exception.
1655 Thanks to Anton Johansson for reporting this issues.
1658 *) Keep original DTLS digest and encryption contexts in retransmission
1659 structures so we can use the previous session parameters if they need
1660 to be resent. (CVE-2013-6450)
1663 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1664 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1665 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1666 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1667 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1668 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1669 [Rob Stradling, Adam Langley]
1671 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1673 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1674 supporting platforms or when small records were transferred.
1675 [Andy Polyakov, Steve Henson]
1677 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1679 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1681 This addresses the flaw in CBC record processing discovered by
1682 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1683 at: http://www.isg.rhul.ac.uk/tls/
1685 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1686 Security Group at Royal Holloway, University of London
1687 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1688 Emilia Käsper for the initial patch.
1690 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1692 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1693 ciphersuites which can be exploited in a denial of service attack.
1694 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1695 and detecting this bug and to Wolfgang Ettlinger
1696 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1700 *) Return an error when checking OCSP signatures when key is NULL.
1701 This fixes a DoS attack. (CVE-2013-0166)
1704 *) Make openssl verify return errors.
1705 [Chris Palmer <palmer@google.com> and Ben Laurie]
1707 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1708 the right response is stapled. Also change SSL_get_certificate()
1709 so it returns the certificate actually sent.
1710 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1711 [Rob Stradling <rob.stradling@comodo.com>]
1713 *) Fix possible deadlock when decoding public keys.
1716 *) Don't use TLS 1.0 record version number in initial client hello
1720 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1722 *) Sanity check record length before skipping explicit IV in TLS
1723 1.2, 1.1 and DTLS to fix DoS attack.
1725 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1726 fuzzing as a service testing platform.
1730 *) Initialise tkeylen properly when encrypting CMS messages.
1731 Thanks to Solar Designer of Openwall for reporting this issue.
1734 *) In FIPS mode don't try to use composite ciphers as they are not
1738 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1740 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1741 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1742 mean any application compiled against OpenSSL 1.0.0 headers setting
1743 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1744 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1745 0x10000000L Any application which was previously compiled against
1746 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1747 will need to be recompiled as a result. Letting be results in
1748 inability to disable specifically TLS 1.1 and in client context,
1749 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1752 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1753 disable just protocol X, but all protocols above X *if* there are
1754 protocols *below* X still enabled. In more practical terms it means
1755 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1756 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1757 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1761 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1763 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1764 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1765 in CRYPTO_realloc_clean.
1767 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1768 issue and to Adam Langley <agl@chromium.org> for fixing it.
1770 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1772 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1775 *) Workarounds for some broken servers that "hang" if a client hello
1776 record length exceeds 255 bytes.
1778 1. Do not use record version number > TLS 1.0 in initial client
1779 hello: some (but not all) hanging servers will now work.
1780 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1781 the number of ciphers sent in the client hello. This should be
1782 set to an even number, such as 50, for example by passing:
1783 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1784 Most broken servers should now work.
1785 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1786 TLS 1.2 client support entirely.
1789 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1792 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1794 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1795 STRING form instead of a DigestInfo.
1798 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1799 and the RSA_sign/RSA_verify functions. This was made more apparent when
1800 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1801 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1802 the correct format in RSA_verify so both forms transparently work.
1805 *) Some servers which support TLS 1.0 can choke if we initially indicate
1806 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1807 encrypted premaster secret. As a workaround use the maximum permitted
1808 client version in client hello, this should keep such servers happy
1809 and still work with previous versions of OpenSSL.
1812 *) Add support for TLS/DTLS heartbeats.
1813 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1815 *) Add support for SCTP.
1816 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1818 *) Improved PRNG seeding for VOS.
1819 [Paul Green <Paul.Green@stratus.com>]
1821 *) Extensive assembler packs updates, most notably:
1823 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1824 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1825 - x86_64: bit-sliced AES implementation;
1826 - ARM: NEON support, contemporary platforms optimizations;
1827 - s390x: z196 support;
1828 - *: GHASH and GF(2^m) multiplication implementations;
1832 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1833 (removal of unnecessary code)
1834 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1836 *) Add TLS key material exporter from RFC 5705.
1839 *) Add DTLS-SRTP negotiation from RFC 5764.
1842 *) Add Next Protocol Negotiation,
1843 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1844 disabled with a no-npn flag to config or Configure. Code donated
1846 [Adam Langley <agl@google.com> and Ben Laurie]
1848 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1849 NIST-P256, NIST-P521, with constant-time single point multiplication on
1850 typical inputs. Compiler support for the nonstandard type __uint128_t is
1851 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1852 Code made available under Apache License version 2.0.
1854 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1855 line to include this in your build of OpenSSL, and run "make depend" (or
1856 "make update"). This enables the following EC_METHODs:
1858 EC_GFp_nistp224_method()
1859 EC_GFp_nistp256_method()
1860 EC_GFp_nistp521_method()
1862 EC_GROUP_new_by_curve_name() will automatically use these (while
1863 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1865 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1867 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1868 all platforms. Move ssize_t definition from e_os.h to the public
1869 header file e_os2.h as it now appears in public header file cms.h
1872 *) New -sigopt option to the ca, req and x509 utilities. Additional
1873 signature parameters can be passed using this option and in
1877 *) Add RSA PSS signing function. This will generate and set the
1878 appropriate AlgorithmIdentifiers for PSS based on those in the
1879 corresponding EVP_MD_CTX structure. No application support yet.
1882 *) Support for companion algorithm specific ASN1 signing routines.
1883 New function ASN1_item_sign_ctx() signs a pre-initialised
1884 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1885 the appropriate parameters.
1888 *) Add new algorithm specific ASN1 verification initialisation function
1889 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1890 handling will be the same no matter what EVP_PKEY_METHOD is used.
1891 Add a PSS handler to support verification of PSS signatures: checked
1892 against a number of sample certificates.
1895 *) Add signature printing for PSS. Add PSS OIDs.
1896 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1898 *) Add algorithm specific signature printing. An individual ASN1 method
1899 can now print out signatures instead of the standard hex dump.
1901 More complex signatures (e.g. PSS) can print out more meaningful
1902 information. Include DSA version that prints out the signature
1906 *) Password based recipient info support for CMS library: implementing
1910 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1911 neatly separates the code into cipher and PBE sections and is required
1912 for some algorithms that split PBES2 into separate pieces (such as
1913 password based CMS).
1916 *) Session-handling fixes:
1917 - Fix handling of connections that are resuming with a session ID,
1918 but also support Session Tickets.
1919 - Fix a bug that suppressed issuing of a new ticket if the client
1920 presented a ticket with an expired session.
1921 - Try to set the ticket lifetime hint to something reasonable.
1922 - Make tickets shorter by excluding irrelevant information.
1923 - On the client side, don't ignore renewed tickets.
1924 [Adam Langley, Bodo Moeller (Google)]
1926 *) Fix PSK session representation.
1929 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1931 This work was sponsored by Intel.
1934 *) Add GCM support to TLS library. Some custom code is needed to split
1935 the IV between the fixed (from PRF) and explicit (from TLS record)
1936 portions. This adds all GCM ciphersuites supported by RFC5288 and
1937 RFC5289. Generalise some AES* cipherstrings to include GCM and
1938 add a special AESGCM string for GCM only.
1941 *) Expand range of ctrls for AES GCM. Permit setting invocation
1942 field on decrypt and retrieval of invocation field only on encrypt.
1945 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1946 As required by RFC5289 these ciphersuites cannot be used if for
1947 versions of TLS earlier than 1.2.
1950 *) For FIPS capable OpenSSL interpret a NULL default public key method
1951 as unset and return the appropriate default but do *not* set the default.
1952 This means we can return the appropriate method in applications that
1953 switch between FIPS and non-FIPS modes.
1956 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1957 ENGINE is used then we cannot handle that in the FIPS module so we
1958 keep original code iff non-FIPS operations are allowed.
1961 *) Add -attime option to openssl utilities.
1962 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1964 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1967 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1968 FIPS EC methods unconditionally for now.
1971 *) New build option no-ec2m to disable characteristic 2 code.
1974 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1975 all cases can be covered as some introduce binary incompatibilities.
1978 *) Redirect RSA operations to FIPS module including keygen,
1979 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1982 *) Add similar low level API blocking to ciphers.
1985 *) Low level digest APIs are not approved in FIPS mode: any attempt
1986 to use these will cause a fatal error. Applications that *really* want
1987 to use them can use the private_* version instead.
1990 *) Redirect cipher operations to FIPS module for FIPS builds.
1993 *) Redirect digest operations to FIPS module for FIPS builds.
1996 *) Update build system to add "fips" flag which will link in fipscanister.o
1997 for static and shared library builds embedding a signature if needed.
2000 *) Output TLS supported curves in preference order instead of numerical
2001 order. This is currently hardcoded for the highest order curves first.
2002 This should be configurable so applications can judge speed vs strength.
2005 *) Add TLS v1.2 server support for client authentication.
2008 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2012 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2013 FIPS modules versions.
2016 *) Add TLS v1.2 client side support for client authentication. Keep cache
2017 of handshake records longer as we don't know the hash algorithm to use
2018 until after the certificate request message is received.
2021 *) Initial TLS v1.2 client support. Add a default signature algorithms
2022 extension including all the algorithms we support. Parse new signature
2023 format in client key exchange. Relax some ECC signing restrictions for
2024 TLS v1.2 as indicated in RFC5246.
2027 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2028 to new signature format when needed using client digest preference.
2029 All server ciphersuites should now work correctly in TLS v1.2. No client
2030 support yet and no support for client certificates.
2033 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2034 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2035 ciphersuites. At present only RSA key exchange ciphersuites work with
2036 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2037 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2038 and version checking.
2041 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2042 with this defined it will not be affected by any changes to ssl internal
2043 structures. Add several utility functions to allow openssl application
2044 to work with OPENSSL_NO_SSL_INTERN defined.
2048 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2050 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2053 *) Permit abbreviated handshakes when renegotiating using the function
2054 SSL_renegotiate_abbreviated().
2055 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2057 *) Add call to ENGINE_register_all_complete() to
2058 ENGINE_load_builtin_engines(), so some implementations get used
2059 automatically instead of needing explicit application support.
2062 *) Add support for TLS key exporter as described in RFC5705.
2063 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2065 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2066 a few changes are required:
2068 Add SSL_OP_NO_TLSv1_1 flag.
2069 Add TLSv1_1 methods.
2070 Update version checking logic to handle version 1.1.
2071 Add explicit IV handling (ported from DTLS code).
2072 Add command line options to s_client/s_server.
2075 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2077 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2078 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2079 content decryption and always return the same error. Note: this attack
2080 needs on average 2^20 messages so it only affects automated senders. The
2081 old behaviour can be reenabled in the CMS code by setting the
2082 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2083 an MMA defence is not necessary.
2084 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2085 this issue. (CVE-2012-0884)
2088 *) Fix CVE-2011-4619: make sure we really are receiving a
2089 client hello before rejecting multiple SGC restarts. Thanks to
2090 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2093 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2095 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2096 Thanks to Antonio Martin, Enterprise Secure Access Research and
2097 Development, Cisco Systems, Inc. for discovering this bug and
2098 preparing a fix. (CVE-2012-0050)
2101 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2103 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2104 of the Vaudenay padding oracle attack on CBC mode encryption
2105 which enables an efficient plaintext recovery attack against
2106 the OpenSSL implementation of DTLS. Their attack exploits timing
2107 differences arising during decryption processing. A research
2108 paper describing this attack can be found at:
2109 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2110 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2111 Security Group at Royal Holloway, University of London
2112 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2113 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2114 for preparing the fix. (CVE-2011-4108)
2115 [Robin Seggelmann, Michael Tuexen]
2117 *) Clear bytes used for block padding of SSL 3.0 records.
2119 [Adam Langley (Google)]
2121 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2122 Kadianakis <desnacked@gmail.com> for discovering this issue and
2123 Adam Langley for preparing the fix. (CVE-2011-4619)
2124 [Adam Langley (Google)]
2126 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2127 [Andrey Kulikov <amdeich@gmail.com>]
2129 *) Prevent malformed RFC3779 data triggering an assertion failure.
2130 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2131 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2132 [Rob Austein <sra@hactrn.net>]
2134 *) Improved PRNG seeding for VOS.
2135 [Paul Green <Paul.Green@stratus.com>]
2137 *) Fix ssl_ciph.c set-up race.
2138 [Adam Langley (Google)]
2140 *) Fix spurious failures in ecdsatest.c.
2141 [Emilia Käsper (Google)]
2143 *) Fix the BIO_f_buffer() implementation (which was mixing different
2144 interpretations of the '..._len' fields).
2145 [Adam Langley (Google)]
2147 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2148 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2149 threads won't reuse the same blinding coefficients.
2151 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2152 lock to call BN_BLINDING_invert_ex, and avoids one use of
2153 BN_BLINDING_update for each BN_BLINDING structure (previously,
2154 the last update always remained unused).
2155 [Emilia Käsper (Google)]
2157 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2158 [Bob Buckholz (Google)]
2160 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2162 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2163 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2164 [Kaspar Brand <ossl@velox.ch>]
2166 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2167 for multi-threaded use of ECDH. (CVE-2011-3210)
2168 [Adam Langley (Google)]
2170 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2173 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2174 signature public key algorithm by using OID xref utilities instead.
2175 Before this you could only use some ECC ciphersuites with SHA1 only.
2178 *) Add protection against ECDSA timing attacks as mentioned in the paper
2179 by Billy Bob Brumley and Nicola Tuveri, see:
2181 http://eprint.iacr.org/2011/232.pdf
2183 [Billy Bob Brumley and Nicola Tuveri]
2185 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2187 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2188 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2190 *) Fix bug in string printing code: if *any* escaping is enabled we must
2191 escape the escape character (backslash) or the resulting string is
2195 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2197 *) Disable code workaround for ancient and obsolete Netscape browsers
2198 and servers: an attacker can use it in a ciphersuite downgrade attack.
2199 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2202 *) Fixed J-PAKE implementation error, originally discovered by
2203 Sebastien Martini, further info and confirmation from Stefan
2204 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2207 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2209 *) Fix extension code to avoid race conditions which can result in a buffer
2210 overrun vulnerability: resumed sessions must not be modified as they can
2211 be shared by multiple threads. CVE-2010-3864
2214 *) Fix WIN32 build system to correctly link an ENGINE directory into
2218 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2220 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2222 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2224 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2226 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2227 context. The operation can be customised via the ctrl mechanism in
2228 case ENGINEs want to include additional functionality.
2231 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2234 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2235 output hashes compatible with older versions of OpenSSL.
2236 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2238 *) Fix compression algorithm handling: if resuming a session use the
2239 compression algorithm of the resumed session instead of determining
2240 it from client hello again. Don't allow server to change algorithm.
2243 *) Add load_crls() function to apps tidying load_certs() too. Add option
2244 to verify utility to allow additional CRLs to be included.
2247 *) Update OCSP request code to permit adding custom headers to the request:
2248 some responders need this.
2251 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2253 [Julia Lawall <julia@diku.dk>]
2255 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2256 needlessly dereferenced structures, used obsolete functions and
2257 didn't handle all updated verify codes correctly.
2260 *) Disable MD2 in the default configuration.
2263 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2264 indicate the initial BIO being pushed or popped. This makes it possible
2265 to determine whether the BIO is the one explicitly called or as a result
2266 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2267 it handles reference counts correctly and doesn't zero out the I/O bio
2268 when it is not being explicitly popped. WARNING: applications which
2269 included workarounds for the old buggy behaviour will need to be modified
2270 or they could free up already freed BIOs.
2273 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2274 renaming to all platforms (within the 0.9.8 branch, this was
2275 done conditionally on Netware platforms to avoid a name clash).
2276 [Guenter <lists@gknw.net>]
2278 *) Add ECDHE and PSK support to DTLS.
2279 [Michael Tuexen <tuexen@fh-muenster.de>]
2281 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2285 *) Add "missing" function EVP_MD_flags() (without this the only way to
2286 retrieve a digest flags is by accessing the structure directly. Update
2287 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2288 or cipher is registered as in the "from" argument. Print out all
2289 registered digests in the dgst usage message instead of manually
2290 attempting to work them out.
2293 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2294 this allows the use of compression and extensions. Change default cipher
2295 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2296 by default unless an application cipher string requests it.
2299 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2300 key ids to find matching certificates and keys but some PKCS#12 files
2301 don't follow the (somewhat unwritten) rules and this strategy fails.
2302 Now just gather all certificates together and the first private key
2303 then look for the first certificate that matches the key.
2306 *) Support use of registered digest and cipher names for dgst and cipher
2307 commands instead of having to add each one as a special case. So now
2314 openssl dgst -sha256 foo
2316 and this works for ENGINE based algorithms too.
2320 *) Update Gost ENGINE to support parameter files.
2321 [Victor B. Wagner <vitus@cryptocom.ru>]
2323 *) Support GeneralizedTime in ca utility.
2324 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2326 *) Enhance the hash format used for certificate directory links. The new
2327 form uses the canonical encoding (meaning equivalent names will work
2328 even if they aren't identical) and uses SHA1 instead of MD5. This form
2329 is incompatible with the older format and as a result c_rehash should
2330 be used to rebuild symbolic links.
2333 *) Make PKCS#8 the default write format for private keys, replacing the
2334 traditional format. This form is standardised, more secure and doesn't
2335 include an implicit MD5 dependency.
2338 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2339 committed to OpenSSL should pass this lot as a minimum.
2342 *) Add session ticket override functionality for use by EAP-FAST.
2343 [Jouni Malinen <j@w1.fi>]
2345 *) Modify HMAC functions to return a value. Since these can be implemented
2346 in an ENGINE errors can occur.
2349 *) Type-checked OBJ_bsearch_ex.
2352 *) Type-checked OBJ_bsearch. Also some constification necessitated
2353 by type-checking. Still to come: TXT_DB, bsearch(?),
2354 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2358 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2359 seconds to a tm structure directly, instead of going through OS
2360 specific date routines. This avoids any issues with OS routines such
2361 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2362 and X509_time_adj_ex() to cover the extended range. The existing
2363 X509_time_adj() is still usable and will no longer have any date issues.
2366 *) Delta CRL support. New use deltas option which will attempt to locate
2367 and search any appropriate delta CRLs available.
2369 This work was sponsored by Google.
2372 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2373 code and add additional score elements. Validate alternate CRL paths
2374 as part of the CRL checking and indicate a new error "CRL path validation
2375 error" in this case. Applications wanting additional details can use
2376 the verify callback and check the new "parent" field. If this is not
2377 NULL CRL path validation is taking place. Existing applications wont
2378 see this because it requires extended CRL support which is off by
2381 This work was sponsored by Google.
2384 *) Support for freshest CRL extension.
2386 This work was sponsored by Google.
2389 *) Initial indirect CRL support. Currently only supported in the CRLs
2390 passed directly and not via lookup. Process certificate issuer
2391 CRL entry extension and lookup CRL entries by bother issuer name
2392 and serial number. Check and process CRL issuer entry in IDP extension.
2394 This work was sponsored by Google.
2397 *) Add support for distinct certificate and CRL paths. The CRL issuer
2398 certificate is validated separately in this case. Only enabled if
2399 an extended CRL support flag is set: this flag will enable additional
2400 CRL functionality in future.
2402 This work was sponsored by Google.
2405 *) Add support for policy mappings extension.
2407 This work was sponsored by Google.
2410 *) Fixes to pathlength constraint, self issued certificate handling,
2411 policy processing to align with RFC3280 and PKITS tests.
2413 This work was sponsored by Google.
2416 *) Support for name constraints certificate extension. DN, email, DNS
2417 and URI types are currently supported.
2419 This work was sponsored by Google.
2422 *) To cater for systems that provide a pointer-based thread ID rather
2423 than numeric, deprecate the current numeric thread ID mechanism and
2424 replace it with a structure and associated callback type. This
2425 mechanism allows a numeric "hash" to be extracted from a thread ID in
2426 either case, and on platforms where pointers are larger than 'long',
2427 mixing is done to help ensure the numeric 'hash' is usable even if it
2428 can't be guaranteed unique. The default mechanism is to use "&errno"
2429 as a pointer-based thread ID to distinguish between threads.
2431 Applications that want to provide their own thread IDs should now use
2432 CRYPTO_THREADID_set_callback() to register a callback that will call
2433 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2435 Note that ERR_remove_state() is now deprecated, because it is tied
2436 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2437 to free the current thread's error state should be replaced by
2438 ERR_remove_thread_state(NULL).
2440 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2441 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2442 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2443 application was previously providing a numeric thread callback that
2444 was inappropriate for distinguishing threads, then uniqueness might
2445 have been obtained with &errno that happened immediately in the
2446 intermediate development versions of OpenSSL; this is no longer the
2447 case, the numeric thread callback will now override the automatic use
2449 [Geoff Thorpe, with help from Bodo Moeller]
2451 *) Initial support for different CRL issuing certificates. This covers a
2452 simple case where the self issued certificates in the chain exist and
2453 the real CRL issuer is higher in the existing chain.
2455 This work was sponsored by Google.
2458 *) Removed effectively defunct crypto/store from the build.
2461 *) Revamp of STACK to provide stronger type-checking. Still to come:
2462 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2463 ASN1_STRING, CONF_VALUE.
2466 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2467 RAM on SSL connections. This option can save about 34k per idle SSL.
2470 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2471 STACK, TXT_DB, bsearch, qsort.
2474 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2475 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2476 support for data, signedData, compressedData, digestedData and
2477 encryptedData, envelopedData types included. Scripts to check against
2478 RFC4134 examples draft and interop and consistency checks of many
2479 content types and variants.
2482 *) Add options to enc utility to support use of zlib compression BIO.
2485 *) Extend mk1mf to support importing of options and assembly language
2486 files from Configure script, currently only included in VC-WIN32.
2487 The assembly language rules can now optionally generate the source
2488 files from the associated perl scripts.
2491 *) Implement remaining functionality needed to support GOST ciphersuites.
2492 Interop testing has been performed using CryptoPro implementations.
2493 [Victor B. Wagner <vitus@cryptocom.ru>]
2495 *) s390x assembler pack.
2498 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2502 *) Implement Opaque PRF Input TLS extension as specified in
2503 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2504 official specification yet and no extension type assignment by
2505 IANA exists, this extension (for now) will have to be explicitly
2506 enabled when building OpenSSL by providing the extension number
2507 to use. For example, specify an option
2509 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2511 to the "config" or "Configure" script to enable the extension,
2512 assuming extension number 0x9527 (which is a completely arbitrary
2513 and unofficial assignment based on the MD5 hash of the Internet
2514 Draft). Note that by doing so, you potentially lose
2515 interoperability with other TLS implementations since these might
2516 be using the same extension number for other purposes.
2518 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2519 opaque PRF input value to use in the handshake. This will create
2520 an interal copy of the length-'len' string at 'src', and will
2521 return non-zero for success.
2523 To get more control and flexibility, provide a callback function
2526 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2527 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2531 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2534 Callback function 'cb' will be called in handshakes, and is
2535 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2536 Argument 'arg' is for application purposes (the value as given to
2537 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2538 be provided to the callback function). The callback function
2539 has to return non-zero to report success: usually 1 to use opaque
2540 PRF input just if possible, or 2 to enforce use of the opaque PRF
2541 input. In the latter case, the library will abort the handshake
2542 if opaque PRF input is not successfully negotiated.
2544 Arguments 'peerinput' and 'len' given to the callback function
2545 will always be NULL and 0 in the case of a client. A server will
2546 see the client's opaque PRF input through these variables if
2547 available (NULL and 0 otherwise). Note that if the server
2548 provides an opaque PRF input, the length must be the same as the
2549 length of the client's opaque PRF input.
2551 Note that the callback function will only be called when creating
2552 a new session (session resumption can resume whatever was
2553 previously negotiated), and will not be called in SSL 2.0
2554 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2555 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2556 for applications that need to enforce opaque PRF input.
2560 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2563 [Victor B. Wagner <vitus@cryptocom.ru>]
2565 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2566 RFC4507bis. The encrypted ticket format is an encrypted encoded
2567 SSL_SESSION structure, that way new session features are automatically
2570 If a client application caches session in an SSL_SESSION structure
2571 support is transparent because tickets are now stored in the encoded
2574 The SSL_CTX structure automatically generates keys for ticket
2575 protection in servers so again support should be possible
2576 with no application modification.
2578 If a client or server wishes to disable RFC4507 support then the option
2579 SSL_OP_NO_TICKET can be set.
2581 Add a TLS extension debugging callback to allow the contents of any client
2582 or server extensions to be examined.
2584 This work was sponsored by Google.
2587 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2588 OpenSSL should now compile cleanly on gcc 4.2
2589 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2591 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2592 support including streaming MAC support: this is required for GOST
2593 ciphersuite support.
2594 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2596 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2597 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2598 to output in BER and PEM format.
2601 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2602 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2603 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2604 ENGINE support for HMAC keys which are unextractable. New -mac and
2605 -macopt options to dgst utility.
2608 *) New option -sigopt to dgst utility. Update dgst to use
2609 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2610 alternative signing paramaters such as X9.31 or PSS in the dgst
2614 *) Change ssl_cipher_apply_rule(), the internal function that does
2615 the work each time a ciphersuite string requests enabling
2616 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2617 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2618 the order of disabled ciphersuites such that those ciphersuites
2619 that most recently went from enabled to disabled not only stay
2620 in order with respect to each other, but also have higher priority
2621 than other disabled ciphersuites the next time ciphersuites are
2624 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2625 the same ciphersuites as with "HIGH" alone, but in a specific
2626 order where the PSK ciphersuites come first (since they are the
2627 most recently disabled ciphersuites when "HIGH" is parsed).
2629 Also, change ssl_create_cipher_list() (using this new
2630 funcionality) such that between otherwise identical
2631 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2635 *) Change ssl_create_cipher_list() so that it automatically
2636 arranges the ciphersuites in reasonable order before starting
2637 to process the rule string. Thus, the definition for "DEFAULT"
2638 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2639 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2640 This makes it much easier to arrive at a reasonable default order
2641 in applications for which anonymous ciphers are OK (meaning
2642 that you can't actually use DEFAULT).
2643 [Bodo Moeller; suggested by Victor Duchovni]
2645 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2646 processing) into multiple integers instead of setting
2647 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2648 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2649 (These masks as well as the individual bit definitions are hidden
2650 away into the non-exported interface ssl/ssl_locl.h, so this
2651 change to the definition of the SSL_CIPHER structure shouldn't
2652 affect applications.) This give us more bits for each of these
2653 categories, so there is no longer a need to coagulate AES128 and
2654 AES256 into a single algorithm bit, and to coagulate Camellia128
2655 and Camellia256 into a single algorithm bit, which has led to all
2658 Thus, among other things, the kludge introduced in 0.9.7m and
2659 0.9.8e for masking out AES256 independently of AES128 or masking
2660 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2662 With the change, we also introduce new ciphersuite aliases that
2663 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2667 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2668 Use the leftmost N bytes of the signature input if the input is
2669 larger than the prime q (with N being the size in bytes of q).
2672 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2673 it yet and it is largely untested.
2676 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2679 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2680 some compilers (gcc 4.2 and later) reject their use. Safestack is
2681 reimplemented. Update ASN1 to avoid use of legacy functions.
2684 *) Win32/64 targets are linked with Winsock2.
2687 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2688 to external functions. This can be used to increase CRL handling
2689 efficiency especially when CRLs are very large by (for example) storing
2690 the CRL revoked certificates in a database.
2693 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2694 new CRLs added to a directory can be used. New command line option
2695 -verify_return_error to s_client and s_server. This causes real errors
2696 to be returned by the verify callback instead of carrying on no matter
2697 what. This reflects the way a "real world" verify callback would behave.
2700 *) GOST engine, supporting several GOST algorithms and public key formats.
2701 Kindly donated by Cryptocom.
2704 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2705 partitioned by DP are handled but no indirect CRL or reason partitioning
2706 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2707 selected via a scoring technique which handles IDP and AKID in CRLs.
2710 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2711 will ultimately be used for all verify operations: this will remove the
2712 X509_STORE dependency on certificate verification and allow alternative
2713 lookup methods. X509_STORE based implementations of these two callbacks.
2716 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2717 Modify get_crl() to find a valid (unexpired) CRL if possible.
2720 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2721 this would be called X509_CRL_cmp() but that name is already used by
2722 a function that just compares CRL issuer names. Cache several CRL
2723 extensions in X509_CRL structure and cache CRLDP in X509.
2726 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2727 this maps equivalent X509_NAME structures into a consistent structure.
2728 Name comparison can then be performed rapidly using memcmp().
2731 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2735 *) Allow digests to supply their own micalg string for S/MIME type using
2736 the ctrl EVP_MD_CTRL_MICALG.
2739 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2740 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2741 ctrl. It can then customise the structure before and/or after signing
2745 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2746 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2747 to free up any added signature OIDs.
2750 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2751 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2752 digest and cipher tables. New options added to openssl utility:
2753 list-message-digest-algorithms and list-cipher-algorithms.
2756 *) Change the array representation of binary polynomials: the list
2757 of degrees of non-zero coefficients is now terminated with -1.
2758 Previously it was terminated with 0, which was also part of the
2759 value; thus, the array representation was not applicable to
2760 polynomials where t^0 has coefficient zero. This change makes
2761 the array representation useful in a more general context.
2764 *) Various modifications and fixes to SSL/TLS cipher string
2765 handling. For ECC, the code now distinguishes between fixed ECDH
2766 with RSA certificates on the one hand and with ECDSA certificates
2767 on the other hand, since these are separate ciphersuites. The
2768 unused code for Fortezza ciphersuites has been removed.
2770 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2771 (not "ECDHE"). For consistency with the code for DH
2772 certificates, use of ECDH certificates is now considered ECDH
2773 authentication, not RSA or ECDSA authentication (the latter is
2774 merely the CA's signing algorithm and not actively used in the
2777 The temporary ciphersuite alias "ECCdraft" is no longer
2778 available, and ECC ciphersuites are no longer excluded from "ALL"
2779 and "DEFAULT". The following aliases now exist for RFC 4492
2780 ciphersuites, most of these by analogy with the DH case:
2782 kECDHr - ECDH cert, signed with RSA
2783 kECDHe - ECDH cert, signed with ECDSA
2784 kECDH - ECDH cert (signed with either RSA or ECDSA)
2785 kEECDH - ephemeral ECDH
2786 ECDH - ECDH cert or ephemeral ECDH
2792 AECDH - anonymous ECDH
2793 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2797 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2798 Use correct micalg parameters depending on digest(s) in signed message.
2801 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2802 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2805 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2806 an engine to register a method. Add ENGINE lookups for methods and
2807 functional reference processing.
2810 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2811 EVP_{Sign,Verify}* which allow an application to customise the signature
2815 *) New -resign option to smime utility. This adds one or more signers
2816 to an existing PKCS#7 signedData structure. Also -md option to use an
2817 alternative message digest algorithm for signing.
2820 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2821 create PKCS7 structures containing multiple signers. Update smime
2822 application to support multiple signers.
2825 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2829 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2830 Reorganize PBE internals to lookup from a static table using NIDs,
2831 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2832 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2833 PRF which will be automatically used with PBES2.
2836 *) Replace the algorithm specific calls to generate keys in "req" with the
2840 *) Update PKCS#7 enveloped data routines to use new API. This is now
2841 supported by any public key method supporting the encrypt operation. A
2842 ctrl is added to allow the public key algorithm to examine or modify
2843 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2847 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2848 a default digest type to use. In most cases this will be SHA1 but some
2849 algorithms (such as GOST) need to specify an alternative digest. The
2850 return value indicates how strong the preference is 1 means optional and
2851 2 is mandatory (that is it is the only supported type). Modify
2852 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2853 use the default md. Update openssl utilities to use the default digest
2854 type for signing if it is not explicitly indicated.
2857 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2858 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2859 signing method from the key type. This effectively removes the link
2860 between digests and public key types.
2863 *) Add an OID cross reference table and utility functions. Its purpose is to
2864 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2865 rsaEncryption. This will allow some of the algorithm specific hackery
2866 needed to use the correct OID to be removed.
2869 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2870 structures for PKCS7_sign(). They are now set up by the relevant public
2874 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2877 *) Add support for key derivation (agreement) in the API, DH method and
2881 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2882 public and private key formats. As a side effect these add additional
2883 command line functionality not previously available: DSA signatures can be
2884 generated and verified using pkeyutl and DH key support and generation in
2889 [Oliver Tappe <zooey@hirschkaefer.de>]
2891 *) New make target "install_html_docs" installs HTML renditions of the
2893 [Oliver Tappe <zooey@hirschkaefer.de>]
2895 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2896 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2897 support key and parameter generation and add initial key generation
2898 functionality for RSA.
2901 *) Add functions for main EVP_PKEY_method operations. The undocumented
2902 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2903 EVP_PKEY_{encrypt,decrypt}_old.
2906 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2907 key API, doesn't do much yet.
2910 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2911 public key algorithms. New option to openssl utility:
2912 "list-public-key-algorithms" to print out info.
2915 *) Implement the Supported Elliptic Curves Extension for
2916 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2919 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2920 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2923 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2924 utilities such as rsa, dsa, dsaparam etc except they process any key
2928 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2929 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2930 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2934 *) Initial support for pluggable public key ASN1.
2935 De-spaghettify the public key ASN1 handling. Move public and private
2936 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2937 algorithm specific handling to a single module within the relevant
2938 algorithm directory. Add functions to allow (near) opaque processing
2939 of public and private key structures.
2942 *) Implement the Supported Point Formats Extension for
2943 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2946 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2947 for the psk identity [hint] and the psk callback functions to the
2948 SSL_SESSION, SSL and SSL_CTX structure.
2951 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2955 SSL_CTX_use_psk_identity_hint
2956 SSL_get_psk_identity_hint
2957 SSL_get_psk_identity
2958 SSL_use_psk_identity_hint
2960 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2962 *) Add RFC 3161 compliant time stamp request creation, response generation
2963 and response verification functionality.
2964 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2966 *) Add initial support for TLS extensions, specifically for the server_name
2967 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2968 have new members for a host name. The SSL data structure has an
2969 additional member SSL_CTX *initial_ctx so that new sessions can be
2970 stored in that context to allow for session resumption, even after the
2971 SSL has been switched to a new SSL_CTX in reaction to a client's
2972 server_name extension.
2974 New functions (subject to change):
2976 SSL_get_servername()
2977 SSL_get_servername_type()
2980 New CTRL codes and macros (subject to change):
2982 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2983 - SSL_CTX_set_tlsext_servername_callback()
2984 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2985 - SSL_CTX_set_tlsext_servername_arg()
2986 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2988 openssl s_client has a new '-servername ...' option.
2990 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2991 '-key2 ...', '-servername_fatal' (subject to change). This allows
2992 testing the HostName extension for a specific single host name ('-cert'
2993 and '-key' remain fallbacks for handshakes without HostName
2994 negotiation). If the unrecognized_name alert has to be sent, this by
2995 default is a warning; it becomes fatal with the '-servername_fatal'
2998 [Peter Sylvester, Remy Allais, Christophe Renou]
3000 *) Whirlpool hash implementation is added.
3003 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3004 bn(64,32). Because of instruction set limitations it doesn't have
3005 any negative impact on performance. This was done mostly in order
3006 to make it possible to share assembler modules, such as bn_mul_mont
3007 implementations, between 32- and 64-bit builds without hassle.
3010 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3011 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3015 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3016 dedicated Montgomery multiplication procedure, is introduced.
3017 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3018 "64-bit" performance on certain 32-bit targets.
3021 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3022 in SSL structures. New SSL ctrl to set maximum send fragment size.
3023 Save memory by seeting the I/O buffer sizes dynamically instead of
3024 using the maximum available value.
3027 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3028 in addition to the text details.
3031 *) Very, very preliminary EXPERIMENTAL support for printing of general
3032 ASN1 structures. This currently produces rather ugly output and doesn't
3033 handle several customised structures at all.
3036 *) Integrated support for PVK file format and some related formats such
3037 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3038 these in the 'rsa' and 'dsa' utilities.
3041 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3044 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3045 place for the (very old) "NETSCAPE" format certificates which are now
3046 handled using new ASN1 code equivalents.
3049 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3050 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3051 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3054 *) Modify CRL distribution points extension code to print out previously
3055 unsupported fields. Enhance extension setting code to allow setting of
3059 *) Add print and set support for Issuing Distribution Point CRL extension.
3062 *) Change 'Configure' script to enable Camellia by default.
3065 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3067 *) When rejecting SSL/TLS records due to an incorrect version number, never
3068 update s->server with a new major version number. As of
3069 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3070 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3071 the previous behavior could result in a read attempt at NULL when
3072 receiving specific incorrect SSL/TLS records once record payload
3073 protection is active. (CVE-2010-0740)
3074 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3076 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3077 could be crashed if the relevant tables were not present (e.g. chrooted).
3078 [Tomas Hoger <thoger@redhat.com>]
3080 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3082 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3083 [Martin Olsson, Neel Mehta]
3085 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3086 accommodate for stack sorting, always a write lock!).
3089 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3090 excessive delays in the RAND_poll(): over a minute. As a workaround
3091 include a time check in the inner Heap32Next loop too.
3094 *) The code that handled flushing of data in SSL/TLS originally used the
3095 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3096 the problem outlined in PR#1949. The fix suggested there however can
3097 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3098 of Apache). So instead simplify the code to flush unconditionally.
3099 This should be fine since flushing with no data to flush is a no op.
3102 *) Handle TLS versions 2.0 and later properly and correctly use the
3103 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3104 off ancient servers have a habit of sticking around for a while...
3107 *) Modify compression code so it frees up structures without using the
3108 ex_data callbacks. This works around a problem where some applications
3109 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3110 restarting) then use compression (e.g. SSL with compression) later.
3111 This results in significant per-connection memory leaks and
3112 has caused some security issues including CVE-2008-1678 and
3116 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3117 change when encrypting or decrypting.
3120 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3121 connect and renegotiate with servers which do not support RI.
3122 Until RI is more widely deployed this option is enabled by default.
3125 *) Add "missing" ssl ctrls to clear options and mode.
3128 *) If client attempts to renegotiate and doesn't support RI respond with
3129 a no_renegotiation alert as required by RFC5746. Some renegotiating
3130 TLS clients will continue a connection gracefully when they receive
3131 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3132 waiting for a server hello which it will never receive. Now we treat a
3133 received no_renegotiation alert as a fatal error. This is because
3134 applications requesting a renegotiation might well expect it to succeed
3135 and would have no code in place to handle the server denying it so the
3136 only safe thing to do is to terminate the connection.
3139 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3140 peer supports secure renegotiation and 0 otherwise. Print out peer
3141 renegotiation support in s_client/s_server.
3144 *) Replace the highly broken and deprecated SPKAC certification method with
3145 the updated NID creation version. This should correctly handle UTF8.
3148 *) Implement RFC5746. Re-enable renegotiation but require the extension
3149 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3150 turns out to be a bad idea. It has been replaced by
3151 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3152 SSL_CTX_set_options(). This is really not recommended unless you
3153 know what you are doing.
3154 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3156 *) Fixes to stateless session resumption handling. Use initial_ctx when
3157 issuing and attempting to decrypt tickets in case it has changed during
3158 servername handling. Use a non-zero length session ID when attempting
3159 stateless session resumption: this makes it possible to determine if
3160 a resumption has occurred immediately after receiving server hello
3161 (several places in OpenSSL subtly assume this) instead of later in
3165 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3166 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3167 fixes for a few places where the return code is not checked
3169 [Julia Lawall <julia@diku.dk>]
3171 *) Add --strict-warnings option to Configure script to include devteam
3172 warnings in other configurations.
3175 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3176 makes it possible to install openssl libraries in locations which
3177 have names other than "lib", for example "/usr/lib64" which some
3179 [Steve Henson, based on patch from Jeremy Utley]
3181 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3182 X690 8.9.12 and can produce some misleading textual output of OIDs.
3183 [Steve Henson, reported by Dan Kaminsky]
3185 *) Delete MD2 from algorithm tables. This follows the recommendation in
3186 several standards that it is not used in new applications due to
3187 several cryptographic weaknesses. For binary compatibility reasons
3188 the MD2 API is still compiled in by default.
3191 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3195 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3196 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3198 [Guenter <lists@gknw.net>]
3200 *) Fix the server certificate chain building code to use X509_verify_cert(),
3201 it used to have an ad-hoc builder which was unable to cope with anything
3202 other than a simple chain.
3203 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3205 *) Don't check self signed certificate signatures in X509_verify_cert()
3206 by default (a flag can override this): it just wastes time without
3207 adding any security. As a useful side effect self signed root CAs
3208 with non-FIPS digests are now usable in FIPS mode.
3211 *) In dtls1_process_out_of_seq_message() the check if the current message
3212 is already buffered was missing. For every new message was memory
3213 allocated, allowing an attacker to perform an denial of service attack
3214 with sending out of seq handshake messages until there is no memory
3215 left. Additionally every future messege was buffered, even if the
3216 sequence number made no sense and would be part of another handshake.
3217 So only messages with sequence numbers less than 10 in advance will be
3218 buffered. (CVE-2009-1378)
3219 [Robin Seggelmann, discovered by Daniel Mentz]
3221 *) Records are buffered if they arrive with a future epoch to be
3222 processed after finishing the corresponding handshake. There is
3223 currently no limitation to this buffer allowing an attacker to perform
3224 a DOS attack with sending records with future epochs until there is no
3225 memory left. This patch adds the pqueue_size() function to determine
3226 the size of a buffer and limits the record buffer to 100 entries.
3228 [Robin Seggelmann, discovered by Daniel Mentz]
3230 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3231 parent structure is freed. (CVE-2009-1379)
3234 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3235 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3238 [Ilya O. <vrghost@gmail.com>]
3240 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3242 *) Disable renegotiation completely - this fixes a severe security
3243 problem (CVE-2009-3555) at the cost of breaking all
3244 renegotiation. Renegotiation can be re-enabled by setting
3245 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3246 run-time. This is really not recommended unless you know what
3250 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3252 *) Don't set val to NULL when freeing up structures, it is freed up by
3253 underlying code. If sizeof(void *) > sizeof(long) this can result in
3254 zeroing past the valid field. (CVE-2009-0789)
3255 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3257 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3258 checked correctly. This would allow some invalid signed attributes to
3259 appear to verify correctly. (CVE-2009-0591)
3260 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3262 *) Reject UniversalString and BMPString types with invalid lengths. This
3263 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3264 a legal length. (CVE-2009-0590)
3267 *) Set S/MIME signing as the default purpose rather than setting it
3268 unconditionally. This allows applications to override it at the store
3272 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3273 to handle some structures.
3276 *) Improve efficiency of mem_gets: don't search whole buffer each time
3278 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3280 *) New -hex option for openssl rand.
3283 *) Print out UTF8String and NumericString when parsing ASN1.
3286 *) Support NumericString type for name components.
3289 *) Allow CC in the environment to override the automatically chosen
3290 compiler. Note that nothing is done to ensure flags work with the
3294 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3296 *) Properly check EVP_VerifyFinal() and similar return values
3298 [Ben Laurie, Bodo Moeller, Google Security Team]
3300 *) Enable TLS extensions by default.
3303 *) Allow the CHIL engine to be loaded, whether the application is
3304 multithreaded or not. (This does not release the developer from the
3305 obligation to set up the dynamic locking callbacks.)
3306 [Sander Temme <sander@temme.net>]
3308 *) Use correct exit code if there is an error in dgst command.
3309 [Steve Henson; problem pointed out by Roland Dirlewanger]
3311 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3312 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3315 *) Add experimental JPAKE support, including demo authentication in
3316 s_client and s_server.
3319 *) Set the comparison function in v3_addr_canonize().
3320 [Rob Austein <sra@hactrn.net>]
3322 *) Add support for XMPP STARTTLS in s_client.
3323 [Philip Paeps <philip@freebsd.org>]
3325 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3326 to ensure that even with this option, only ciphersuites in the
3327 server's preference list will be accepted. (Note that the option
3328 applies only when resuming a session, so the earlier behavior was
3329 just about the algorithm choice for symmetric cryptography.)
3332 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3334 *) Fix NULL pointer dereference if a DTLS server received
3335 ChangeCipherSpec as first record (CVE-2009-1386).
3338 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3339 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3342 *) The fix in 0.9.8c that supposedly got rid of unsafe
3343 double-checked locking was incomplete for RSA blinding,
3344 addressing just one layer of what turns out to have been
3345 doubly unsafe triple-checked locking.
3347 So now fix this for real by retiring the MONT_HELPER macro
3348 in crypto/rsa/rsa_eay.c.
3350 [Bodo Moeller; problem pointed out by Marius Schilder]
3352 *) Various precautionary measures:
3354 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3356 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3357 (NB: This would require knowledge of the secret session ticket key
3358 to exploit, in which case you'd be SOL either way.)
3360 - Change bn_nist.c so that it will properly handle input BIGNUMs
3361 outside the expected range.
3363 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3366 [Neel Mehta, Bodo Moeller]
3368 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3369 the load fails. Useful for distros.
3370 [Ben Laurie and the FreeBSD team]
3372 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3375 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3378 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3380 This work was sponsored by Logica.
3383 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3384 keystores. Support for SSL/TLS client authentication too.
3385 Not compiled unless enable-capieng specified to Configure.
3387 This work was sponsored by Logica.
3390 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3391 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3392 attribute creation routines such as certifcate requests and PKCS#12
3396 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3398 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3399 handshake which could lead to a cilent crash as found using the
3400 Codenomicon TLS test suite (CVE-2008-1672)
3401 [Steve Henson, Mark Cox]
3403 *) Fix double free in TLS server name extensions which could lead to
3404 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3407 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3409 Clear the error queue to ensure that error entries left from
3410 older function calls do not interfere with the correct operation.
3411 [Lutz Jaenicke, Erik de Castro Lopo]
3413 *) Remove root CA certificates of commercial CAs:
3415 The OpenSSL project does not recommend any specific CA and does not
3416 have any policy with respect to including or excluding any CA.
3417 Therefore it does not make any sense to ship an arbitrary selection
3418 of root CA certificates with the OpenSSL software.
3421 *) RSA OAEP patches to fix two separate invalid memory reads.
3422 The first one involves inputs when 'lzero' is greater than
3423 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3424 before the beginning of from). The second one involves inputs where
3425 the 'db' section contains nothing but zeroes (there is a one-byte
3426 invalid read after the end of 'db').
3427 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3429 *) Partial backport from 0.9.9-dev:
3431 Introduce bn_mul_mont (dedicated Montgomery multiplication
3432 procedure) as a candidate for BIGNUM assembler implementation.
3433 While 0.9.9-dev uses assembler for various architectures, only
3434 x86_64 is available by default here in the 0.9.8 branch, and
3435 32-bit x86 is available through a compile-time setting.
3437 To try the 32-bit x86 assembler implementation, use Configure
3438 option "enable-montasm" (which exists only for this backport).
3440 As "enable-montasm" for 32-bit x86 disclaims code stability
3441 anyway, in this constellation we activate additional code
3442 backported from 0.9.9-dev for further performance improvements,
3443 namely BN_from_montgomery_word. (To enable this otherwise,
3444 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3446 [Andy Polyakov (backport partially by Bodo Moeller)]
3448 *) Add TLS session ticket callback. This allows an application to set
3449 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3450 values. This is useful for key rollover for example where several key
3451 sets may exist with different names.
3454 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3455 This was broken until now in 0.9.8 releases, such that the only way
3456 a registered ENGINE could be used (assuming it initialises
3457 successfully on the host) was to explicitly set it as the default
3458 for the relevant algorithms. This is in contradiction with 0.9.7
3459 behaviour and the documentation. With this fix, when an ENGINE is
3460 registered into a given algorithm's table of implementations, the
3461 'uptodate' flag is reset so that auto-discovery will be used next
3462 time a new context for that algorithm attempts to select an
3464 [Ian Lister (tweaked by Geoff Thorpe)]
3466 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3467 implemention in the following ways:
3469 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3472 Lack of BER streaming support means one pass streaming processing is
3473 only supported if data is detached: setting the streaming flag is
3474 ignored for embedded content.
3476 CMS support is disabled by default and must be explicitly enabled
3477 with the enable-cms configuration option.
3480 *) Update the GMP engine glue to do direct copies between BIGNUM and
3481 mpz_t when openssl and GMP use the same limb size. Otherwise the
3482 existing "conversion via a text string export" trick is still used.
3483 [Paul Sheer <paulsheer@gmail.com>]
3485 *) Zlib compression BIO. This is a filter BIO which compressed and
3486 uncompresses any data passed through it.
3489 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3490 RFC3394 compatible AES key wrapping.
3493 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3494 sets string data without copying. X509_ALGOR_set0() and
3495 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3496 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3497 from an X509_ATTRIBUTE structure optionally checking it occurs only
3498 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3502 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3503 to get the expected BN_FLG_CONSTTIME behavior.
3504 [Bodo Moeller (Google)]
3508 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3509 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3510 - added some more tests to do_tests.pl
3511 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3512 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3513 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3514 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3515 - various changes to netware.pl to enable gcc-cross builds on Win32
3517 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3518 - various changes to fix missing prototype warnings
3519 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3520 - added AES, WHIRLPOOL and CPUID assembler code to build files
3521 - added missing AES assembler make rules to mk1mf.pl
3522 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3523 [Guenter Knauf <eflash@gmx.net>]
3525 *) Implement certificate status request TLS extension defined in RFC3546.
3526 A client can set the appropriate parameters and receive the encoded
3527 OCSP response via a callback. A server can query the supplied parameters
3528 and set the encoded OCSP response in the callback. Add simplified examples
3529 to s_client and s_server.
3532 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3534 *) Fix various bugs:
3535 + Binary incompatibility of ssl_ctx_st structure
3536 + DTLS interoperation with non-compliant servers
3537 + Don't call get_session_cb() without proposed session
3538 + Fix ia64 assembler code
3539 [Andy Polyakov, Steve Henson]
3541 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3543 *) DTLS Handshake overhaul. There were longstanding issues with
3544 OpenSSL DTLS implementation, which were making it impossible for
3545 RFC 4347 compliant client to communicate with OpenSSL server.
3546 Unfortunately just fixing these incompatibilities would "cut off"
3547 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3548 server keeps tolerating non RFC compliant syntax. The opposite is
3549 not true, 0.9.8f client can not communicate with earlier server.
3550 This update even addresses CVE-2007-4995.
3553 *) Changes to avoid need for function casts in OpenSSL: some compilers
3554 (gcc 4.2 and later) reject their use.
3555 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3558 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3559 RFC4507bis. The encrypted ticket format is an encrypted encoded
3560 SSL_SESSION structure, that way new session features are automatically
3563 If a client application caches session in an SSL_SESSION structure
3564 support is transparent because tickets are now stored in the encoded
3567 The SSL_CTX structure automatically generates keys for ticket
3568 protection in servers so again support should be possible
3569 with no application modification.
3571 If a client or server wishes to disable RFC4507 support then the option
3572 SSL_OP_NO_TICKET can be set.
3574 Add a TLS extension debugging callback to allow the contents of any client
3575 or server extensions to be examined.
3577 This work was sponsored by Google.
3580 *) Add initial support for TLS extensions, specifically for the server_name
3581 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3582 have new members for a host name. The SSL data structure has an
3583 additional member SSL_CTX *initial_ctx so that new sessions can be
3584 stored in that context to allow for session resumption, even after the
3585 SSL has been switched to a new SSL_CTX in reaction to a client's
3586 server_name extension.
3588 New functions (subject to change):
3590 SSL_get_servername()
3591 SSL_get_servername_type()
3594 New CTRL codes and macros (subject to change):
3596 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3597 - SSL_CTX_set_tlsext_servername_callback()
3598 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3599 - SSL_CTX_set_tlsext_servername_arg()
3600 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3602 openssl s_client has a new '-servername ...' option.
3604 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3605 '-key2 ...', '-servername_fatal' (subject to change). This allows
3606 testing the HostName extension for a specific single host name ('-cert'
3607 and '-key' remain fallbacks for handshakes without HostName
3608 negotiation). If the unrecognized_name alert has to be sent, this by
3609 default is a warning; it becomes fatal with the '-servername_fatal'
3612 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3614 *) Add AES and SSE2 assembly language support to VC++ build.
3617 *) Mitigate attack on final subtraction in Montgomery reduction.
3620 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3621 (which previously caused an internal error).
3624 *) Squeeze another 10% out of IGE mode when in != out.
3627 *) AES IGE mode speedup.
3628 [Dean Gaudet (Google)]
3630 *) Add the Korean symmetric 128-bit cipher SEED (see
3631 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3632 add SEED ciphersuites from RFC 4162:
3634 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3635 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3636 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3637 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3639 To minimize changes between patchlevels in the OpenSSL 0.9.8
3640 series, SEED remains excluded from compilation unless OpenSSL
3641 is configured with 'enable-seed'.
3642 [KISA, Bodo Moeller]
3644 *) Mitigate branch prediction attacks, which can be practical if a
3645 single processor is shared, allowing a spy process to extract
3646 information. For detailed background information, see
3647 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3648 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3649 and Necessary Software Countermeasures"). The core of the change
3650 are new versions BN_div_no_branch() and
3651 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3652 respectively, which are slower, but avoid the security-relevant
3653 conditional branches. These are automatically called by BN_div()
3654 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3655 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3656 remove a conditional branch.
3658 BN_FLG_CONSTTIME is the new name for the previous
3659 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3660 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3661 in the exponent causes BN_mod_exp_mont() to use the alternative
3662 implementation in BN_mod_exp_mont_consttime().) The old name
3663 remains as a deprecated alias.
3665 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3666 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3667 constant-time implementations for more than just exponentiation.
3668 Here too the old name is kept as a deprecated alias.
3670 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3671 the BN_BLINDING structure gets an independent copy of the
3672 modulus. This means that the previous "BIGNUM *m" argument to
3673 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3674 essentially becomes "const BIGNUM *m", although we can't actually
3675 change this in the header file before 0.9.9. It allows
3676 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3677 enable BN_FLG_CONSTTIME.
3679 [Matthew D Wood (Intel Corp)]
3681 *) In the SSL/TLS server implementation, be strict about session ID
3682 context matching (which matters if an application uses a single
3683 external cache for different purposes). Previously,
3684 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3685 set. This did ensure strict client verification, but meant that,
3686 with applications using a single external cache for quite
3687 different requirements, clients could circumvent ciphersuite
3688 restrictions for a given session ID context by starting a session
3689 in a different context.
3692 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3693 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3694 authentication-only ciphersuites.
3697 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3698 not complete and could lead to a possible single byte overflow
3699 (CVE-2007-5135) [Ben Laurie]
3701 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3703 *) Since AES128 and AES256 (and similarly Camellia128 and
3704 Camellia256) share a single mask bit in the logic of
3705 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3706 kludge to work properly if AES128 is available and AES256 isn't
3707 (or if Camellia128 is available and Camellia256 isn't).
3710 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3711 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3712 When a point or a seed is encoded in a BIT STRING, we need to
3713 prevent the removal of trailing zero bits to get the proper DER
3714 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3715 of a NamedBitList, for which trailing 0 bits need to be removed.)
3718 *) Have SSL/TLS server implementation tolerate "mismatched" record
3719 protocol version while receiving ClientHello even if the
3720 ClientHello is fragmented. (The server can't insist on the
3721 particular protocol version it has chosen before the ServerHello
3722 message has informed the client about his choice.)
3725 *) Add RFC 3779 support.
3726 [Rob Austein for ARIN, Ben Laurie]
3728 *) Load error codes if they are not already present instead of using a
3729 static variable. This allows them to be cleanly unloaded and reloaded.
3730 Improve header file function name parsing.
3733 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3734 or CAPABILITY handshake as required by RFCs.
3737 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3739 *) Introduce limits to prevent malicious keys being able to
3740 cause a denial of service. (CVE-2006-2940)
3741 [Steve Henson, Bodo Moeller]
3743 *) Fix ASN.1 parsing of certain invalid structures that can result
3744 in a denial of service. (CVE-2006-2937) [Steve Henson]
3746 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3747 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3749 *) Fix SSL client code which could crash if connecting to a
3750 malicious SSLv2 server. (CVE-2006-4343)
3751 [Tavis Ormandy and Will Drewry, Google Security Team]
3753 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3754 match only those. Before that, "AES256-SHA" would be interpreted
3755 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3756 the same strength classification in 0.9.7h) as we currently only
3757 have a single AES bit in the ciphersuite description bitmap.
3758 That change, however, also applied to ciphersuite strings such as
3759 "RC4-MD5" that intentionally matched multiple ciphersuites --
3760 namely, SSL 2.0 ciphersuites in addition to the more common ones
3761 from SSL 3.0/TLS 1.0.
3763 So we change the selection algorithm again: Naming an explicit
3764 ciphersuite selects this one ciphersuite, and any other similar
3765 ciphersuite (same bitmap) from *other* protocol versions.
3766 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3767 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3769 Since SSL 2.0 does not have any ciphersuites for which the
3770 128/256 bit distinction would be relevant, this works for now.
3771 The proper fix will be to use different bits for AES128 and
3772 AES256, which would have avoided the problems from the beginning;
3773 however, bits are scarce, so we can only do this in a new release
3774 (not just a patchlevel) when we can change the SSL_CIPHER
3775 definition to split the single 'unsigned long mask' bitmap into
3776 multiple values to extend the available space.
3780 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3782 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3783 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3785 *) Add AES IGE and biIGE modes.
3788 *) Change the Unix randomness entropy gathering to use poll() when
3789 possible instead of select(), since the latter has some
3790 undesirable limitations.
3791 [Darryl Miles via Richard Levitte and Bodo Moeller]
3793 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3794 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3795 cannot be implicitly activated as part of, e.g., the "AES" alias.
3796 However, please upgrade to OpenSSL 0.9.9[-dev] for
3797 non-experimental use of the ECC ciphersuites to get TLS extension
3798 support, which is required for curve and point format negotiation
3799 to avoid potential handshake problems.
3802 *) Disable rogue ciphersuites:
3804 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3805 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3806 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3808 The latter two were purportedly from
3809 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3812 Also deactivate the remaining ciphersuites from
3813 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3814 unofficial, and the ID has long expired.
3817 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3818 dual-core machines) and other potential thread-safety issues.
3821 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3822 versions), which is now available for royalty-free use
3823 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3824 Also, add Camellia TLS ciphersuites from RFC 4132.
3826 To minimize changes between patchlevels in the OpenSSL 0.9.8
3827 series, Camellia remains excluded from compilation unless OpenSSL
3828 is configured with 'enable-camellia'.
3831 *) Disable the padding bug check when compression is in use. The padding
3832 bug check assumes the first packet is of even length, this is not
3833 necessarily true if compresssion is enabled and can result in false
3834 positives causing handshake failure. The actual bug test is ancient
3835 code so it is hoped that implementations will either have fixed it by
3836 now or any which still have the bug do not support compression.
3839 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3841 *) When applying a cipher rule check to see if string match is an explicit
3842 cipher suite and only match that one cipher suite if it is.
3845 *) Link in manifests for VC++ if needed.
3846 [Austin Ziegler <halostatue@gmail.com>]
3848 *) Update support for ECC-based TLS ciphersuites according to
3849 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3850 TLS extensions, which are supported starting with the 0.9.9
3851 branch, not in the OpenSSL 0.9.8 branch).
3854 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3855 opaque EVP_CIPHER_CTX handling.
3858 *) Fixes and enhancements to zlib compression code. We now only use
3859 "zlib1.dll" and use the default __cdecl calling convention on Win32
3860 to conform with the standards mentioned here:
3861 http://www.zlib.net/DLL_FAQ.txt
3862 Static zlib linking now works on Windows and the new --with-zlib-include
3863 --with-zlib-lib options to Configure can be used to supply the location
3864 of the headers and library. Gracefully handle case where zlib library
3868 *) Several fixes and enhancements to the OID generation code. The old code
3869 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3870 handle numbers larger than ULONG_MAX, truncated printing and had a
3871 non standard OBJ_obj2txt() behaviour.
3874 *) Add support for building of engines under engine/ as shared libraries
3875 under VC++ build system.
3878 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3879 Hopefully, we will not see any false combination of paths any more.
3882 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3884 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3885 (part of SSL_OP_ALL). This option used to disable the
3886 countermeasure against man-in-the-middle protocol-version
3887 rollback in the SSL 2.0 server implementation, which is a bad
3888 idea. (CVE-2005-2969)
3890 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3891 for Information Security, National Institute of Advanced Industrial
3892 Science and Technology [AIST], Japan)]
3894 *) Add two function to clear and return the verify parameter flags.
3897 *) Keep cipherlists sorted in the source instead of sorting them at
3898 runtime, thus removing the need for a lock.
3901 *) Avoid some small subgroup attacks in Diffie-Hellman.
3902 [Nick Mathewson and Ben Laurie]
3904 *) Add functions for well-known primes.
3907 *) Extended Windows CE support.
3908 [Satoshi Nakamura and Andy Polyakov]
3910 *) Initialize SSL_METHOD structures at compile time instead of during
3911 runtime, thus removing the need for a lock.
3914 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3915 attempting to decrypt each encrypted key in turn. Add support to
3919 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3921 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3924 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3927 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3928 key into the same file any more.
3931 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3934 *) Add -utf8 command line and config file option to 'ca'.
3935 [Stefan <stf@udoma.org]
3937 *) Removed the macro des_crypt(), as it seems to conflict with some
3938 libraries. Use DES_crypt().
3941 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3942 involves renaming the source and generated shared-libs for
3943 both. The engines will accept the corrected or legacy ids
3944 ('ncipher' and '4758_cca' respectively) when binding. NB,
3945 this only applies when building 'shared'.
3946 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3948 *) Add attribute functions to EVP_PKEY structure. Modify
3949 PKCS12_create() to recognize a CSP name attribute and
3950 use it. Make -CSP option work again in pkcs12 utility.
3953 *) Add new functionality to the bn blinding code:
3954 - automatic re-creation of the BN_BLINDING parameters after
3955 a fixed number of uses (currently 32)
3956 - add new function for parameter creation
3957 - introduce flags to control the update behaviour of the
3958 BN_BLINDING parameters
3959 - hide BN_BLINDING structure
3960 Add a second BN_BLINDING slot to the RSA structure to improve
3961 performance when a single RSA object is shared among several
3965 *) Add support for DTLS.
3966 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3968 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3969 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3972 *) Remove buggy and incomplete DH cert support from
3973 ssl/ssl_rsa.c and ssl/s3_both.c
3976 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3977 the apps/openssl applications.
3980 *) Compile clean with "-Wall -Wmissing-prototypes
3981 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3982 DEBUG_SAFESTACK must also be set.
3985 *) Change ./Configure so that certain algorithms can be disabled by default.
3986 The new counterpiece to "no-xxx" is "enable-xxx".
3988 The patented RC5 and MDC2 algorithms will now be disabled unless
3989 "enable-rc5" and "enable-mdc2", respectively, are specified.
3991 (IDEA remains enabled despite being patented. This is because IDEA
3992 is frequently required for interoperability, and there is no license
3993 fee for non-commercial use. As before, "no-idea" can be used to
3994 avoid this algorithm.)
3998 *) Add processing of proxy certificates (see RFC 3820). This work was
3999 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4000 EGEE (Enabling Grids for E-science in Europe).
4003 *) RC4 performance overhaul on modern architectures/implementations, such
4004 as Intel P4, IA-64 and AMD64.
4007 *) New utility extract-section.pl. This can be used specify an alternative
4008 section number in a pod file instead of having to treat each file as
4009 a separate case in Makefile. This can be done by adding two lines to the
4012 =for comment openssl_section:XXX
4014 The blank line is mandatory.
4018 *) New arguments -certform, -keyform and -pass for s_client and s_server
4019 to allow alternative format key and certificate files and passphrase
4023 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4024 update associated structures and add various utility functions.
4026 Add new policy related verify parameters, include policy checking in
4027 standard verify code. Enhance 'smime' application with extra parameters
4028 to support policy checking and print out.
4031 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4032 Nehemiah processors. These extensions support AES encryption in hardware
4033 as well as RNG (though RNG support is currently disabled).
4034 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4036 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4039 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4040 [Andy Polyakov and a number of other people]
4042 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4043 implementation contributed by IBM.
4044 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4046 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4047 exponent rather than 'unsigned long'. There is a corresponding change to
4048 the new 'rsa_keygen' element of the RSA_METHOD structure.
4049 [Jelte Jansen, Geoff Thorpe]
4051 *) Functionality for creating the initial serial number file is now
4052 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4054 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4055 number file to 1, which is bound to cause problems. To avoid
4056 the problems while respecting compatibility between different 0.9.7
4057 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4058 CA.pl for serial number initialization. With the new release 0.9.8,
4059 we can fix the problem directly in the 'ca' utility.)
4062 *) Reduced header interdepencies by declaring more opaque objects in
4063 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4064 give fewer recursive includes, which could break lazy source code - so
4065 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4066 developers should define this symbol when building and using openssl to
4067 ensure they track the recommended behaviour, interfaces, [etc], but
4068 backwards-compatible behaviour prevails when this isn't defined.
4071 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4074 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4075 This will generate a random key of the appropriate length based on the
4076 cipher context. The EVP_CIPHER can provide its own random key generation
4077 routine to support keys of a specific form. This is used in the des and
4078 3des routines to generate a key of the correct parity. Update S/MIME
4079 code to use new functions and hence generate correct parity DES keys.
4080 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4081 valid (weak or incorrect parity).
4084 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4085 as looking them up. This is useful when the verified structure may contain
4086 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4087 present unless the new PKCS7_NO_CRL flag is asserted.
4090 *) Extend ASN1 oid configuration module. It now additionally accepts the
4093 shortName = some long name, 1.2.3.4
4096 *) Reimplemented the BN_CTX implementation. There is now no more static
4097 limitation on the number of variables it can handle nor the depth of the
4098 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4099 information can now expand as required, and rather than having a single
4100 static array of bignums, BN_CTX now uses a linked-list of such arrays
4101 allowing it to expand on demand whilst maintaining the usefulness of
4102 BN_CTX's "bundling".
4105 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4106 to allow all RSA operations to function using a single BN_CTX.
4109 *) Preliminary support for certificate policy evaluation and checking. This
4110 is initially intended to pass the tests outlined in "Conformance Testing
4111 of Relying Party Client Certificate Path Processing Logic" v1.07.
4114 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4115 remained unused and not that useful. A variety of other little bignum
4116 tweaks and fixes have also been made continuing on from the audit (see
4120 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4121 associated ASN1, EVP and SSL functions and old ASN1 macros.
4124 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4125 and this should never fail. So the return value from the use of
4126 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4127 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4130 *) BN_CTX_get() should return zero-valued bignums, providing the same
4131 initialised value as BN_new().
4132 [Geoff Thorpe, suggested by Ulf Möller]
4134 *) Support for inhibitAnyPolicy certificate extension.
4137 *) An audit of the BIGNUM code is underway, for which debugging code is
4138 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4139 is considered valid when processing BIGNUMs, and causes execution to
4140 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4141 further steps are taken to deliberately pollute unused data in BIGNUM
4142 structures to try and expose faulty code further on. For now, openssl will
4143 (in its default mode of operation) continue to tolerate the inconsistent
4144 forms that it has tolerated in the past, but authors and packagers should
4145 consider trying openssl and their own applications when compiled with
4146 these debugging symbols defined. It will help highlight potential bugs in
4147 their own code, and will improve the test coverage for OpenSSL itself. At
4148 some point, these tighter rules will become openssl's default to improve
4149 maintainability, though the assert()s and other overheads will remain only
4150 in debugging configurations. See bn.h for more details.
4151 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4153 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4154 that can only be obtained through BN_CTX_new() (which implicitly
4155 initialises it). The presence of this function only made it possible
4156 to overwrite an existing structure (and cause memory leaks).
4159 *) Because of the callback-based approach for implementing LHASH as a
4160 template type, lh_insert() adds opaque objects to hash-tables and
4161 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4162 to clean up those corresponding objects before destroying the hash table
4163 (and losing the object pointers). So some over-zealous constifications in
4164 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4165 objects as "const" and the lh_doall[_arg] callback wrappers are not
4166 prototyped to have "const" restrictions on the object pointers they are
4167 given (and so aren't required to cast them away any more).
4170 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4171 (speed) prefers to use its own implementation. The two implementations
4172 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4173 its object type properly exposed (MS_TM) instead of casting to/from "char
4174 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4175 aren't necessarily the greatest nomenclatures - but this is what was used
4176 internally to the implementation so I've used that for now.
4179 *) Ensure that deprecated functions do not get compiled when
4180 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4181 the self-tests were still using deprecated key-generation functions so
4182 these have been updated also.
4185 *) Reorganise PKCS#7 code to separate the digest location functionality
4186 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4187 New function PKCS7_set_digest() to set the digest type for PKCS#7
4188 digestedData type. Add additional code to correctly generate the
4189 digestedData type and add support for this type in PKCS7 initialization
4193 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4194 structure of type "other".
4197 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4198 sure the loop does correctly stop and breaking ("division by zero")
4199 modulus operations are not performed. The (pre-generated) prime
4200 table crypto/bn/bn_prime.h was already correct, but it could not be
4201 re-generated on some platforms because of the "division by zero"
4202 situation in the script.
4203 [Ralf S. Engelschall]
4205 *) Update support for ECC-based TLS ciphersuites according to
4206 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4207 SHA-1 now is only used for "small" curves (where the
4208 representation of a field element takes up to 24 bytes); for
4209 larger curves, the field element resulting from ECDH is directly
4210 used as premaster secret.
4211 [Douglas Stebila (Sun Microsystems Laboratories)]
4213 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4214 curve secp160r1 to the tests.
4215 [Douglas Stebila (Sun Microsystems Laboratories)]
4217 *) Add the possibility to load symbols globally with DSO.
4218 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4220 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4221 control of the error stack.
4224 *) Add support for STORE in ENGINE.
4227 *) Add the STORE type. The intention is to provide a common interface
4228 to certificate and key stores, be they simple file-based stores, or
4229 HSM-type store, or LDAP stores, or...
4230 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4233 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4234 pass a list of arguments to any function as well as provide a way
4235 for a function to pass data back to the caller.
4238 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4239 works like BUF_strdup() but can be used to duplicate a portion of
4240 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4244 *) Add the function sk_find_ex() which works like sk_find(), but will
4245 return an index to an element even if an exact match couldn't be
4246 found. The index is guaranteed to point at the element where the
4247 searched-for key would be inserted to preserve sorting order.
4250 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4251 takes an extra flags argument for optional functionality. Currently,
4252 the following flags are defined:
4254 OBJ_BSEARCH_VALUE_ON_NOMATCH
4255 This one gets OBJ_bsearch_ex() to return a pointer to the first
4256 element where the comparing function returns a negative or zero
4259 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4260 This one gets OBJ_bsearch_ex() to return a pointer to the first
4261 element where the comparing function returns zero. This is useful
4262 if there are more than one element where the comparing function
4266 *) Make it possible to create self-signed certificates with 'openssl ca'
4267 in such a way that the self-signed certificate becomes part of the
4268 CA database and uses the same mechanisms for serial number generation
4269 as all other certificate signing. The new flag '-selfsign' enables
4270 this functionality. Adapt CA.sh and CA.pl.in.
4273 *) Add functionality to check the public key of a certificate request
4274 against a given private. This is useful to check that a certificate
4275 request can be signed by that key (self-signing).
4278 *) Make it possible to have multiple active certificates with the same
4279 subject in the CA index file. This is done only if the keyword
4280 'unique_subject' is set to 'no' in the main CA section (default
4281 if 'CA_default') of the configuration file. The value is saved
4282 with the database itself in a separate index attribute file,
4283 named like the index file with '.attr' appended to the name.
4286 *) Generate muti valued AVAs using '+' notation in config files for
4290 *) Support for nameConstraints certificate extension.
4293 *) Support for policyConstraints certificate extension.
4296 *) Support for policyMappings certificate extension.
4299 *) Make sure the default DSA_METHOD implementation only uses its
4300 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4301 and change its own handlers to be NULL so as to remove unnecessary
4302 indirection. This lets alternative implementations fallback to the
4303 default implementation more easily.
4306 *) Support for directoryName in GeneralName related extensions
4310 *) Make it possible to link applications using Makefile.shared.
4311 Make that possible even when linking against static libraries!
4314 *) Support for single pass processing for S/MIME signing. This now
4315 means that S/MIME signing can be done from a pipe, in addition
4316 cleartext signing (multipart/signed type) is effectively streaming
4317 and the signed data does not need to be all held in memory.
4319 This is done with a new flag PKCS7_STREAM. When this flag is set
4320 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4321 is done after the data is output (and digests calculated) in
4322 SMIME_write_PKCS7().
4325 *) Add full support for -rpath/-R, both in shared libraries and
4326 applications, at least on the platforms where it's known how
4330 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4331 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4332 will now compute a table of multiples of the generator that
4333 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4334 faster (notably in the case of a single point multiplication,
4335 scalar * generator).
4336 [Nils Larsch, Bodo Moeller]
4338 *) IPv6 support for certificate extensions. The various extensions
4339 which use the IP:a.b.c.d can now take IPv6 addresses using the
4340 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4344 *) Added an ENGINE that implements RSA by performing private key
4345 exponentiations with the GMP library. The conversions to and from
4346 GMP's mpz_t format aren't optimised nor are any montgomery forms
4347 cached, and on x86 it appears OpenSSL's own performance has caught up.
4348 However there are likely to be other architectures where GMP could
4349 provide a boost. This ENGINE is not built in by default, but it can be
4350 specified at Configure time and should be accompanied by the necessary
4351 linker additions, eg;
4352 ./config -DOPENSSL_USE_GMP -lgmp
4355 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4356 testing availability of engines with "-t" - the old behaviour is
4357 produced by increasing the feature's verbosity with "-tt".
4360 *) ECDSA routines: under certain error conditions uninitialized BN objects
4361 could be freed. Solution: make sure initialization is performed early
4362 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4366 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4367 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4368 software implementations. For DSA and DH, parameter generation can
4369 also be overriden by providing the appropriate method callbacks.
4372 *) Change the "progress" mechanism used in key-generation and
4373 primality testing to functions that take a new BN_GENCB pointer in
4374 place of callback/argument pairs. The new API functions have "_ex"
4375 postfixes and the older functions are reimplemented as wrappers for
4376 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4377 declarations of the old functions to help (graceful) attempts to
4378 migrate to the new functions. Also, the new key-generation API
4379 functions operate on a caller-supplied key-structure and return
4380 success/failure rather than returning a key or NULL - this is to
4381 help make "keygen" another member function of RSA_METHOD etc.
4383 Example for using the new callback interface:
4385 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4389 BN_GENCB_set(&my_cb, my_callback, my_arg);
4391 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4392 /* For the meaning of a, b in calls to my_callback(), see the
4393 * documentation of the function that calls the callback.
4394 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4395 * my_callback should return 1 if it wants BN_is_prime_ex()
4396 * to continue, or 0 to stop.
4401 *) Change the ZLIB compression method to be stateful, and make it
4402 available to TLS with the number defined in
4403 draft-ietf-tls-compression-04.txt.
4406 *) Add the ASN.1 structures and functions for CertificatePair, which
4407 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4409 CertificatePair ::= SEQUENCE {
4410 forward [0] Certificate OPTIONAL,
4411 reverse [1] Certificate OPTIONAL,
4412 -- at least one of the pair shall be present -- }
4414 Also implement the PEM functions to read and write certificate
4415 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4417 This needed to be defined, mostly for the sake of the LDAP
4418 attribute crossCertificatePair, but may prove useful elsewhere as
4422 *) Make it possible to inhibit symlinking of shared libraries in
4423 Makefile.shared, for Cygwin's sake.
4426 *) Extend the BIGNUM API by creating a function
4427 void BN_set_negative(BIGNUM *a, int neg);
4428 and a macro that behave like
4429 int BN_is_negative(const BIGNUM *a);
4431 to avoid the need to access 'a->neg' directly in applications.
4434 *) Implement fast modular reduction for pseudo-Mersenne primes
4435 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4436 EC_GROUP_new_curve_GFp() will now automatically use this
4438 [Nils Larsch <nla@trustcenter.de>]
4440 *) Add new lock type (CRYPTO_LOCK_BN).
4443 *) Change the ENGINE framework to automatically load engines
4444 dynamically from specific directories unless they could be
4445 found to already be built in or loaded. Move all the
4446 current engines except for the cryptodev one to a new
4448 The engines in engines/ are built as shared libraries if
4449 the "shared" options was given to ./Configure or ./config.
4450 Otherwise, they are inserted in libcrypto.a.
4451 /usr/local/ssl/engines is the default directory for dynamic
4452 engines, but that can be overriden at configure time through
4453 the usual use of --prefix and/or --openssldir, and at run
4454 time with the environment variable OPENSSL_ENGINES.
4455 [Geoff Thorpe and Richard Levitte]
4457 *) Add Makefile.shared, a helper makefile to build shared
4458 libraries. Adapt Makefile.org.
4461 *) Add version info to Win32 DLLs.
4462 [Peter 'Luna' Runestig" <peter@runestig.com>]
4464 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4465 can be added using this API to created arbitrary PKCS#12
4466 files while avoiding the low level API.
4468 New options to PKCS12_create(), key or cert can be NULL and
4469 will then be omitted from the output file. The encryption
4470 algorithm NIDs can be set to -1 for no encryption, the mac
4471 iteration count can be set to 0 to omit the mac.
4473 Enhance pkcs12 utility by making the -nokeys and -nocerts
4474 options work when creating a PKCS#12 file. New option -nomac
4475 to omit the mac, NONE can be set for an encryption algorithm.
4476 New code is modified to use the enhanced PKCS12_create()
4477 instead of the low level API.
4480 *) Extend ASN1 encoder to support indefinite length constructed
4481 encoding. This can output sequences tags and octet strings in
4482 this form. Modify pk7_asn1.c to support indefinite length
4483 encoding. This is experimental and needs additional code to
4484 be useful, such as an ASN1 bio and some enhanced streaming
4487 Extend template encode functionality so that tagging is passed
4488 down to the template encoder.
4491 *) Let 'openssl req' fail if an argument to '-newkey' is not
4492 recognized instead of using RSA as a default.
4495 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4496 As these are not official, they are not included in "ALL";
4497 the "ECCdraft" ciphersuite group alias can be used to select them.
4498 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4500 *) Add ECDH engine support.
4501 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4503 *) Add ECDH in new directory crypto/ecdh/.
4504 [Douglas Stebila (Sun Microsystems Laboratories)]
4506 *) Let BN_rand_range() abort with an error after 100 iterations
4507 without success (which indicates a broken PRNG).
4510 *) Change BN_mod_sqrt() so that it verifies that the input value
4511 is really the square of the return value. (Previously,
4512 BN_mod_sqrt would show GIGO behaviour.)
4515 *) Add named elliptic curves over binary fields from X9.62, SECG,
4516 and WAP/WTLS; add OIDs that were still missing.
4518 [Sheueling Chang Shantz and Douglas Stebila
4519 (Sun Microsystems Laboratories)]
4521 *) Extend the EC library for elliptic curves over binary fields
4522 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4525 EC_GF2m_simple_method
4529 EC_GROUP_new_curve_GF2m
4530 EC_GROUP_set_curve_GF2m
4531 EC_GROUP_get_curve_GF2m
4532 EC_POINT_set_affine_coordinates_GF2m
4533 EC_POINT_get_affine_coordinates_GF2m
4534 EC_POINT_set_compressed_coordinates_GF2m
4536 Point compression for binary fields is disabled by default for
4537 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4540 As binary polynomials are represented as BIGNUMs, various members
4541 of the EC_GROUP and EC_POINT data structures can be shared
4542 between the implementations for prime fields and binary fields;
4543 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4544 are essentially identical to their ..._GFp counterparts.
4545 (For simplicity, the '..._GFp' prefix has been dropped from
4546 various internal method names.)
4548 An internal 'field_div' method (similar to 'field_mul' and
4549 'field_sqr') has been added; this is used only for binary fields.
4551 [Sheueling Chang Shantz and Douglas Stebila
4552 (Sun Microsystems Laboratories)]
4554 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4555 through methods ('mul', 'precompute_mult').
4557 The generic implementations (now internally called 'ec_wNAF_mul'
4558 and 'ec_wNAF_precomputed_mult') remain the default if these
4559 methods are undefined.
4561 [Sheueling Chang Shantz and Douglas Stebila
4562 (Sun Microsystems Laboratories)]
4564 *) New function EC_GROUP_get_degree, which is defined through
4565 EC_METHOD. For curves over prime fields, this returns the bit
4566 length of the modulus.
4568 [Sheueling Chang Shantz and Douglas Stebila
4569 (Sun Microsystems Laboratories)]
4571 *) New functions EC_GROUP_dup, EC_POINT_dup.
4572 (These simply call ..._new and ..._copy).
4574 [Sheueling Chang Shantz and Douglas Stebila
4575 (Sun Microsystems Laboratories)]
4577 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4578 Polynomials are represented as BIGNUMs (where the sign bit is not
4579 used) in the following functions [macros]:
4582 BN_GF2m_sub [= BN_GF2m_add]
4583 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4584 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4585 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4587 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4588 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4589 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4590 BN_GF2m_cmp [= BN_ucmp]
4592 (Note that only the 'mod' functions are actually for fields GF(2^m).
4593 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4595 For some functions, an the irreducible polynomial defining a
4596 field can be given as an 'unsigned int[]' with strictly
4597 decreasing elements giving the indices of those bits that are set;
4598 i.e., p[] represents the polynomial
4599 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4601 p[0] > p[1] > ... > p[k] = 0.
4602 This applies to the following functions:
4607 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4608 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4610 BN_GF2m_mod_sqrt_arr
4611 BN_GF2m_mod_solve_quad_arr
4615 Conversion can be performed by the following functions:
4620 bntest.c has additional tests for binary polynomial arithmetic.
4622 Two implementations for BN_GF2m_mod_div() are available.
4623 The default algorithm simply uses BN_GF2m_mod_inv() and
4624 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4625 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4626 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4628 [Sheueling Chang Shantz and Douglas Stebila
4629 (Sun Microsystems Laboratories)]
4631 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4632 functionality is disabled at compile-time.
4633 [Douglas Stebila <douglas.stebila@sun.com>]
4635 *) Change default behaviour of 'openssl asn1parse' so that more
4636 information is visible when viewing, e.g., a certificate:
4638 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4639 mode the content of non-printable OCTET STRINGs is output in a
4640 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4641 avoid the appearance of a printable string.
4642 [Nils Larsch <nla@trustcenter.de>]
4644 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4646 EC_GROUP_set_asn1_flag()
4647 EC_GROUP_get_asn1_flag()
4648 EC_GROUP_set_point_conversion_form()
4649 EC_GROUP_get_point_conversion_form()
4650 These control ASN1 encoding details:
4651 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4652 has been set to OPENSSL_EC_NAMED_CURVE.
4653 - Points are encoded in uncompressed form by default; options for
4654 asn1_for are as for point2oct, namely
4655 POINT_CONVERSION_COMPRESSED
4656 POINT_CONVERSION_UNCOMPRESSED
4657 POINT_CONVERSION_HYBRID
4659 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4662 EC_GROUP_get0_seed()
4663 EC_GROUP_get_seed_len()
4664 This is used only for ASN1 purposes (so far).
4665 [Nils Larsch <nla@trustcenter.de>]
4667 *) Add 'field_type' member to EC_METHOD, which holds the NID
4668 of the appropriate field type OID. The new function
4669 EC_METHOD_get_field_type() returns this value.
4670 [Nils Larsch <nla@trustcenter.de>]
4675 EC_POINT_point2hex()
4676 EC_POINT_hex2point()
4677 providing useful interfaces to EC_POINT_point2oct() and
4678 EC_POINT_oct2point().
4679 [Nils Larsch <nla@trustcenter.de>]
4681 *) Change internals of the EC library so that the functions
4682 EC_GROUP_set_generator()
4683 EC_GROUP_get_generator()
4684 EC_GROUP_get_order()
4685 EC_GROUP_get_cofactor()
4686 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4687 to methods, which would lead to unnecessary code duplication when
4688 adding different types of curves.
4689 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4691 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4692 arithmetic, and such that modified wNAFs are generated
4693 (which avoid length expansion in many cases).
4696 *) Add a function EC_GROUP_check_discriminant() (defined via
4697 EC_METHOD) that verifies that the curve discriminant is non-zero.
4699 Add a function EC_GROUP_check() that makes some sanity tests
4700 on a EC_GROUP, its generator and order. This includes
4701 EC_GROUP_check_discriminant().
4702 [Nils Larsch <nla@trustcenter.de>]
4704 *) Add ECDSA in new directory crypto/ecdsa/.
4706 Add applications 'openssl ecparam' and 'openssl ecdsa'
4707 (these are based on 'openssl dsaparam' and 'openssl dsa').
4709 ECDSA support is also included in various other files across the
4710 library. Most notably,
4711 - 'openssl req' now has a '-newkey ecdsa:file' option;
4712 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4713 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4714 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4715 them suitable for ECDSA where domain parameters must be
4716 extracted before the specific public key;
4717 - ECDSA engine support has been added.
4718 [Nils Larsch <nla@trustcenter.de>]
4720 *) Include some named elliptic curves, and add OIDs from X9.62,
4721 SECG, and WAP/WTLS. Each curve can be obtained from the new
4723 EC_GROUP_new_by_curve_name(),
4724 and the list of available named curves can be obtained with
4725 EC_get_builtin_curves().
4726 Also add a 'curve_name' member to EC_GROUP objects, which can be
4728 EC_GROUP_set_curve_name()
4729 EC_GROUP_get_curve_name()
4730 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4732 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4733 was actually never needed) and in BN_mul(). The removal in BN_mul()
4734 required a small change in bn_mul_part_recursive() and the addition
4735 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4736 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4737 bn_sub_words() and bn_add_words() except they take arrays with
4741 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4743 *) Cleanse PEM buffers before freeing them since they may contain
4745 [Benjamin Bennett <ben@psc.edu>]
4747 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4748 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4749 authentication-only ciphersuites.
4752 *) Since AES128 and AES256 share a single mask bit in the logic of
4753 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4754 kludge to work properly if AES128 is available and AES256 isn't.
4757 *) Expand security boundary to match 1.1.1 module.
4760 *) Remove redundant features: hash file source, editing of test vectors
4761 modify fipsld to use external fips_premain.c signature.
4764 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4765 run algorithm test programs.
4768 *) Make algorithm test programs more tolerant of whitespace.
4771 *) Have SSL/TLS server implementation tolerate "mismatched" record
4772 protocol version while receiving ClientHello even if the
4773 ClientHello is fragmented. (The server can't insist on the
4774 particular protocol version it has chosen before the ServerHello
4775 message has informed the client about his choice.)
4778 *) Load error codes if they are not already present instead of using a
4779 static variable. This allows them to be cleanly unloaded and reloaded.
4782 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4784 *) Introduce limits to prevent malicious keys being able to
4785 cause a denial of service. (CVE-2006-2940)
4786 [Steve Henson, Bodo Moeller]
4788 *) Fix ASN.1 parsing of certain invalid structures that can result
4789 in a denial of service. (CVE-2006-2937) [Steve Henson]
4791 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4792 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4794 *) Fix SSL client code which could crash if connecting to a
4795 malicious SSLv2 server. (CVE-2006-4343)
4796 [Tavis Ormandy and Will Drewry, Google Security Team]
4798 *) Change ciphersuite string processing so that an explicit
4799 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4800 will no longer include "AES128-SHA"), and any other similar
4801 ciphersuite (same bitmap) from *other* protocol versions (so that
4802 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4803 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4804 changes from 0.9.8b and 0.9.8d.
4807 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4809 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4810 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4812 *) Change the Unix randomness entropy gathering to use poll() when
4813 possible instead of select(), since the latter has some
4814 undesirable limitations.
4815 [Darryl Miles via Richard Levitte and Bodo Moeller]
4817 *) Disable rogue ciphersuites:
4819 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4820 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4821 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4823 The latter two were purportedly from
4824 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4827 Also deactive the remaining ciphersuites from
4828 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4829 unofficial, and the ID has long expired.
4832 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4833 dual-core machines) and other potential thread-safety issues.
4836 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4838 *) Adapt fipsld and the build system to link against the validated FIPS
4839 module in FIPS mode.
4842 *) Fixes for VC++ 2005 build under Windows.
4845 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4846 from a Windows bash shell such as MSYS. It is autodetected from the
4847 "config" script when run from a VC++ environment. Modify standard VC++
4848 build to use fipscanister.o from the GNU make build.
4851 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4853 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4854 The value now differs depending on if you build for FIPS or not.
4855 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4856 safely run with a non-FIPSed libcrypto, as it may crash because of
4857 the difference induced by this change.
4860 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4862 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4863 (part of SSL_OP_ALL). This option used to disable the
4864 countermeasure against man-in-the-middle protocol-version
4865 rollback in the SSL 2.0 server implementation, which is a bad
4866 idea. (CVE-2005-2969)
4868 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4869 for Information Security, National Institute of Advanced Industrial
4870 Science and Technology [AIST], Japan)]
4872 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4873 mainly for FIPS compliance and not fully integrated at this stage.
4876 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4877 the exponentiation using a fixed-length exponent. (Otherwise,
4878 the information leaked through timing could expose the secret key
4879 after many signatures; cf. Bleichenbacher's attack on DSA with
4883 *) Make a new fixed-window mod_exp implementation the default for
4884 RSA, DSA, and DH private-key operations so that the sequence of
4885 squares and multiplies and the memory access pattern are
4886 independent of the particular secret key. This will mitigate
4887 cache-timing and potential related attacks.
4889 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4890 and this is automatically used by BN_mod_exp_mont() if the new flag
4891 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4892 will use this BN flag for private exponents unless the flag
4893 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4894 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4896 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4898 *) Change the client implementation for SSLv23_method() and
4899 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4900 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4901 (Previously, the SSL 2.0 backwards compatible Client Hello
4902 message format would be used even with SSL_OP_NO_SSLv2.)
4905 *) Add support for smime-type MIME parameter in S/MIME messages which some
4909 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4910 a threadsafe manner. Modify rsa code to use new function and add calls
4911 to dsa and dh code (which had race conditions before).
4914 *) Include the fixed error library code in the C error file definitions
4915 instead of fixing them up at runtime. This keeps the error code
4916 structures constant.
4919 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4921 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4924 *) Fixes for newer kerberos headers. NB: the casts are needed because
4925 the 'length' field is signed on one version and unsigned on another
4926 with no (?) obvious way to tell the difference, without these VC++
4927 complains. Also the "definition" of FAR (blank) is no longer included
4928 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4929 some needed definitions.
4932 *) Undo Cygwin change.
4935 *) Added support for proxy certificates according to RFC 3820.
4936 Because they may be a security thread to unaware applications,
4937 they must be explicitely allowed in run-time. See
4938 docs/HOWTO/proxy_certificates.txt for further information.
4941 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4943 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4944 server and client random values. Previously
4945 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4946 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4948 This change has negligible security impact because:
4950 1. Server and client random values still have 24 bytes of pseudo random
4953 2. Server and client random values are sent in the clear in the initial
4956 3. The master secret is derived using the premaster secret (48 bytes in
4957 size for static RSA ciphersuites) as well as client server and random
4960 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4963 [Stephen Henson, reported by UK NISCC]
4965 *) Use Windows randomness collection on Cygwin.
4968 *) Fix hang in EGD/PRNGD query when communication socket is closed
4969 prematurely by EGD/PRNGD.
4970 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4972 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4975 *) Back-port of selected performance improvements from development
4976 branch, as well as improved support for PowerPC platforms.
4979 *) Add lots of checks for memory allocation failure, error codes to indicate
4980 failure and freeing up memory if a failure occurs.
4981 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4983 *) Add new -passin argument to dgst.
4986 *) Perform some character comparisons of different types in X509_NAME_cmp:
4987 this is needed for some certificates that reencode DNs into UTF8Strings
4988 (in violation of RFC3280) and can't or wont issue name rollover
4992 *) Make an explicit check during certificate validation to see that
4993 the CA setting in each certificate on the chain is correct. As a
4994 side effect always do the following basic checks on extensions,
4995 not just when there's an associated purpose to the check:
4997 - if there is an unhandled critical extension (unless the user
4998 has chosen to ignore this fault)
4999 - if the path length has been exceeded (if one is set at all)
5000 - that certain extensions fit the associated purpose (if one has
5004 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5006 *) Avoid a race condition when CRLs are checked in a multi threaded
5007 environment. This would happen due to the reordering of the revoked
5008 entries during signature checking and serial number lookup. Now the
5009 encoding is cached and the serial number sort performed under a lock.
5010 Add new STACK function sk_is_sorted().
5013 *) Add Delta CRL to the extension code.
5016 *) Various fixes to s3_pkt.c so alerts are sent properly.
5017 [David Holmes <d.holmes@f5.com>]
5019 *) Reduce the chances of duplicate issuer name and serial numbers (in
5020 violation of RFC3280) using the OpenSSL certificate creation utilities.
5021 This is done by creating a random 64 bit value for the initial serial
5022 number when a serial number file is created or when a self signed
5023 certificate is created using 'openssl req -x509'. The initial serial
5024 number file is created using 'openssl x509 -next_serial' in CA.pl
5025 rather than being initialized to 1.
5028 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5030 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5031 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5032 [Joe Orton, Steve Henson]
5034 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5036 [Joe Orton, Steve Henson]
5038 *) Make it possible to have multiple active certificates with the same
5039 subject in the CA index file. This is done only if the keyword
5040 'unique_subject' is set to 'no' in the main CA section (default
5041 if 'CA_default') of the configuration file. The value is saved
5042 with the database itself in a separate index attribute file,
5043 named like the index file with '.attr' appended to the name.
5046 *) X509 verify fixes. Disable broken certificate workarounds when
5047 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5048 keyUsage extension present. Don't accept CRLs with unhandled critical
5049 extensions: since verify currently doesn't process CRL extensions this
5050 rejects a CRL with *any* critical extensions. Add new verify error codes
5054 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5055 A clarification of RFC2560 will require the use of OCTET STRINGs and
5056 some implementations cannot handle the current raw format. Since OpenSSL
5057 copies and compares OCSP nonces as opaque blobs without any attempt at
5058 parsing them this should not create any compatibility issues.
5061 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5062 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5063 this HMAC (and other) operations are several times slower than OpenSSL
5067 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5068 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5070 *) Use the correct content when signing type "other".
5073 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5075 *) Fix various bugs revealed by running the NISCC test suite:
5077 Stop out of bounds reads in the ASN1 code when presented with
5078 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5080 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5082 If verify callback ignores invalid public key errors don't try to check
5083 certificate signature with the NULL public key.
5087 *) New -ignore_err option in ocsp application to stop the server
5088 exiting on the first error in a request.
5091 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5092 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5096 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5097 extra data after the compression methods not only for TLS 1.0
5098 but also for SSL 3.0 (as required by the specification).
5099 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5101 *) Change X509_certificate_type() to mark the key as exported/exportable
5102 when it's 512 *bits* long, not 512 bytes.
5105 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5106 blocks during encryption.
5109 *) Various fixes to base64 BIO and non blocking I/O. On write
5110 flushes were not handled properly if the BIO retried. On read
5111 data was not being buffered properly and had various logic bugs.
5112 This also affects blocking I/O when the data being decoded is a
5116 *) Various S/MIME bugfixes and compatibility changes:
5117 output correct application/pkcs7 MIME type if
5118 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5119 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5120 of files as .eml work). Correctly handle very long lines in MIME
5124 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5126 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5127 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5128 a protocol version number mismatch like a decryption error
5129 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5132 *) Turn on RSA blinding by default in the default implementation
5133 to avoid a timing attack. Applications that don't want it can call
5134 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5135 They would be ill-advised to do so in most cases.
5136 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5138 *) Change RSA blinding code so that it works when the PRNG is not
5139 seeded (in this case, the secret RSA exponent is abused as
5140 an unpredictable seed -- if it is not unpredictable, there
5141 is no point in blinding anyway). Make RSA blinding thread-safe
5142 by remembering the creator's thread ID in rsa->blinding and
5143 having all other threads use local one-time blinding factors
5144 (this requires more computation than sharing rsa->blinding, but
5145 avoids excessive locking; and if an RSA object is not shared
5146 between threads, blinding will still be very fast).
5149 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5150 ENGINE as defaults for all supported algorithms irrespective of
5151 the 'flags' parameter. 'flags' is now honoured, so applications
5152 should make sure they are passing it correctly.
5155 *) Target "mingw" now allows native Windows code to be generated in
5156 the Cygwin environment as well as with the MinGW compiler.
5159 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5161 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5162 via timing by performing a MAC computation even if incorrect
5163 block cipher padding has been found. This is a countermeasure
5164 against active attacks where the attacker has to distinguish
5165 between bad padding and a MAC verification error. (CVE-2003-0078)
5167 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5168 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5169 Martin Vuagnoux (EPFL, Ilion)]
5171 *) Make the no-err option work as intended. The intention with no-err
5172 is not to have the whole error stack handling routines removed from
5173 libcrypto, it's only intended to remove all the function name and
5174 reason texts, thereby removing some of the footprint that may not
5175 be interesting if those errors aren't displayed anyway.
5177 NOTE: it's still possible for any application or module to have it's
5178 own set of error texts inserted. The routines are there, just not
5179 used by default when no-err is given.
5182 *) Add support for FreeBSD on IA64.
5183 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5185 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5186 Kerberos function mit_des_cbc_cksum(). Before this change,
5187 the value returned by DES_cbc_cksum() was like the one from
5188 mit_des_cbc_cksum(), except the bytes were swapped.
5189 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5191 *) Allow an application to disable the automatic SSL chain building.
5192 Before this a rather primitive chain build was always performed in
5193 ssl3_output_cert_chain(): an application had no way to send the
5194 correct chain if the automatic operation produced an incorrect result.
5196 Now the chain builder is disabled if either:
5198 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5200 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5202 The reasoning behind this is that an application would not want the
5203 auto chain building to take place if extra chain certificates are
5204 present and it might also want a means of sending no additional
5205 certificates (for example the chain has two certificates and the
5209 *) Add the possibility to build without the ENGINE framework.
5210 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5212 *) Under Win32 gmtime() can return NULL: check return value in
5213 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5216 *) DSA routines: under certain error conditions uninitialized BN objects
5217 could be freed. Solution: make sure initialization is performed early
5218 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5219 Nils Larsch <nla@trustcenter.de> via PR#459)
5222 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5223 checked on reconnect on the client side, therefore session resumption
5224 could still fail with a "ssl session id is different" error. This
5225 behaviour is masked when SSL_OP_ALL is used due to
5226 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5227 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5228 followup to PR #377.
5231 *) IA-32 assembler support enhancements: unified ELF targets, support
5232 for SCO/Caldera platforms, fix for Cygwin shared build.
5235 *) Add support for FreeBSD on sparc64. As a consequence, support for
5236 FreeBSD on non-x86 processors is separate from x86 processors on
5237 the config script, much like the NetBSD support.
5238 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5240 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5242 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5245 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5246 code (06) was taken as the first octet of the session ID and the last
5247 octet was ignored consequently. As a result SSLv2 client side session
5248 caching could not have worked due to the session ID mismatch between
5250 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5254 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5255 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5259 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5260 seems that in spite of existing for more than a year, many application
5261 author have done nothing to provide the necessary callbacks, which
5262 means that this particular engine will not work properly anywhere.
5263 This is a very unfortunate situation which forces us, in the name
5264 of usability, to give the hw_ncipher.c a static lock, which is part
5266 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5267 appear in 0.9.8 or later. We EXPECT application authors to have
5268 dealt properly with this when 0.9.8 is released (unless we actually
5269 make such changes in the libcrypto locking code that changes will
5270 have to be made anyway).
5273 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5274 octets have been read, EOF or an error occurs. Without this change
5275 some truncated ASN1 structures will not produce an error.
5278 *) Disable Heimdal support, since it hasn't been fully implemented.
5279 Still give the possibility to force the use of Heimdal, but with
5280 warnings and a request that patches get sent to openssl-dev.
5283 *) Add the VC-CE target, introduce the WINCE sysname, and add
5284 INSTALL.WCE and appropriate conditionals to make it build.
5285 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5287 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5288 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5289 edit numbers of the version.
5290 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5292 *) Introduce safe string copy and catenation functions
5293 (BUF_strlcpy() and BUF_strlcat()).
5294 [Ben Laurie (CHATS) and Richard Levitte]
5296 *) Avoid using fixed-size buffers for one-line DNs.
5297 [Ben Laurie (CHATS)]
5299 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5300 resizing buffers containing secrets, and use where appropriate.
5301 [Ben Laurie (CHATS)]
5303 *) Avoid using fixed size buffers for configuration file location.
5304 [Ben Laurie (CHATS)]
5306 *) Avoid filename truncation for various CA files.
5307 [Ben Laurie (CHATS)]
5309 *) Use sizeof in preference to magic numbers.
5310 [Ben Laurie (CHATS)]
5312 *) Avoid filename truncation in cert requests.
5313 [Ben Laurie (CHATS)]
5315 *) Add assertions to check for (supposedly impossible) buffer
5317 [Ben Laurie (CHATS)]
5319 *) Don't cache truncated DNS entries in the local cache (this could
5320 potentially lead to a spoofing attack).
5321 [Ben Laurie (CHATS)]
5323 *) Fix various buffers to be large enough for hex/decimal
5324 representations in a platform independent manner.
5325 [Ben Laurie (CHATS)]
5327 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5328 resizing buffers containing secrets, and use where appropriate.
5329 [Ben Laurie (CHATS)]
5331 *) Add BIO_indent() to avoid much slightly worrying code to do
5333 [Ben Laurie (CHATS)]
5335 *) Convert sprintf()/BIO_puts() to BIO_printf().
5336 [Ben Laurie (CHATS)]
5338 *) buffer_gets() could terminate with the buffer only half
5340 [Ben Laurie (CHATS)]
5342 *) Add assertions to prevent user-supplied crypto functions from
5343 overflowing internal buffers by having large block sizes, etc.
5344 [Ben Laurie (CHATS)]
5346 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5348 [Ben Laurie (CHATS)]
5350 *) Eliminate unused copy of key in RC4.
5351 [Ben Laurie (CHATS)]
5353 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5354 [Ben Laurie (CHATS)]
5356 *) Fix off-by-one error in EGD path.
5357 [Ben Laurie (CHATS)]
5359 *) If RANDFILE path is too long, ignore instead of truncating.
5360 [Ben Laurie (CHATS)]
5362 *) Eliminate unused and incorrectly sized X.509 structure
5364 [Ben Laurie (CHATS)]
5366 *) Eliminate unused and dangerous function knumber().
5367 [Ben Laurie (CHATS)]
5369 *) Eliminate unused and dangerous structure, KSSL_ERR.
5370 [Ben Laurie (CHATS)]
5372 *) Protect against overlong session ID context length in an encoded
5373 session object. Since these are local, this does not appear to be
5375 [Ben Laurie (CHATS)]
5377 *) Change from security patch (see 0.9.6e below) that did not affect
5378 the 0.9.6 release series:
5380 Remote buffer overflow in SSL3 protocol - an attacker could
5381 supply an oversized master key in Kerberos-enabled versions.
5383 [Ben Laurie (CHATS)]
5385 *) Change the SSL kerb5 codes to match RFC 2712.
5388 *) Make -nameopt work fully for req and add -reqopt switch.
5389 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5391 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5392 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5394 *) Make sure tests can be performed even if the corresponding algorithms
5395 have been removed entirely. This was also the last step to make
5396 OpenSSL compilable with DJGPP under all reasonable conditions.
5397 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5399 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5400 to allow version independent disabling of normally unselected ciphers,
5401 which may be activated as a side-effect of selecting a single cipher.
5403 (E.g., cipher list string "RSA" enables ciphersuites that are left
5404 out of "ALL" because they do not provide symmetric encryption.
5405 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5406 [Lutz Jaenicke, Bodo Moeller]
5408 *) Add appropriate support for separate platform-dependent build
5409 directories. The recommended way to make a platform-dependent
5410 build directory is the following (tested on Linux), maybe with
5413 # Place yourself outside of the OpenSSL source tree. In
5414 # this example, the environment variable OPENSSL_SOURCE
5415 # is assumed to contain the absolute OpenSSL source directory.
5416 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5417 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5418 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5419 mkdir -p `dirname $F`
5420 ln -s $OPENSSL_SOURCE/$F $F
5423 To be absolutely sure not to disturb the source tree, a "make clean"
5424 is a good thing. If it isn't successful, don't worry about it,
5425 it probably means the source directory is very clean.
5428 *) Make sure any ENGINE control commands make local copies of string
5429 pointers passed to them whenever necessary. Otherwise it is possible
5430 the caller may have overwritten (or deallocated) the original string
5431 data when a later ENGINE operation tries to use the stored values.
5432 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5434 *) Improve diagnostics in file reading and command-line digests.
5435 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5437 *) Add AES modes CFB and OFB to the object database. Correct an
5438 error in AES-CFB decryption.
5441 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5442 allows existing EVP_CIPHER_CTX structures to be reused after
5443 calling EVP_*Final(). This behaviour is used by encryption
5444 BIOs and some applications. This has the side effect that
5445 applications must explicitly clean up cipher contexts with
5446 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5449 *) Check the values of dna and dnb in bn_mul_recursive before calling
5450 bn_mul_comba (a non zero value means the a or b arrays do not contain
5451 n2 elements) and fallback to bn_mul_normal if either is not zero.
5454 *) Fix escaping of non-ASCII characters when using the -subj option
5455 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5458 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5459 form for "surname", serialNumber has no short form.
5460 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5461 therefore remove "mail" short name for "internet 7".
5462 The OID for unique identifiers in X509 certificates is
5463 x500UniqueIdentifier, not uniqueIdentifier.
5464 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5467 *) Add an "init" command to the ENGINE config module and auto initialize
5468 ENGINEs. Without any "init" command the ENGINE will be initialized
5469 after all ctrl commands have been executed on it. If init=1 the
5470 ENGINE is initailized at that point (ctrls before that point are run
5471 on the uninitialized ENGINE and after on the initialized one). If
5472 init=0 then the ENGINE will not be iniatialized at all.
5475 *) Fix the 'app_verify_callback' interface so that the user-defined
5476 argument is actually passed to the callback: In the
5477 SSL_CTX_set_cert_verify_callback() prototype, the callback
5478 declaration has been changed from
5481 int (*cb)(X509_STORE_CTX *,void *);
5482 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5483 i=s->ctx->app_verify_callback(&ctx)
5484 has been changed into
5485 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5487 To update applications using SSL_CTX_set_cert_verify_callback(),
5488 a dummy argument can be added to their callback functions.
5489 [D. K. Smetters <smetters@parc.xerox.com>]
5491 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5492 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5494 *) Add and OPENSSL_LOAD_CONF define which will cause
5495 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5496 This allows older applications to transparently support certain
5497 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5498 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5499 load the config file and OPENSSL_add_all_algorithms_conf() which will
5500 always load it have also been added.
5503 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5504 Adjust NIDs and EVP layer.
5505 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5507 *) Config modules support in openssl utility.
5509 Most commands now load modules from the config file,
5510 though in a few (such as version) this isn't done
5511 because it couldn't be used for anything.
5513 In the case of ca and req the config file used is
5514 the same as the utility itself: that is the -config
5515 command line option can be used to specify an
5519 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5520 use "openssl_conf" if filename is NULL use default openssl config file.
5523 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5524 config section name. Add a new flag to tolerate a missing config file
5525 and move code to CONF_modules_load_file().
5528 *) Support for crypto accelerator cards from Accelerated Encryption
5529 Processing, www.aep.ie. (Use engine 'aep')
5530 The support was copied from 0.9.6c [engine] and adapted/corrected
5531 to work with the new engine framework.
5532 [AEP Inc. and Richard Levitte]
5534 *) Support for SureWare crypto accelerator cards from Baltimore
5535 Technologies. (Use engine 'sureware')
5536 The support was copied from 0.9.6c [engine] and adapted
5537 to work with the new engine framework.
5540 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5541 make the newer ENGINE framework commands for the CHIL engine work.
5542 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5544 *) Make it possible to produce shared libraries on ReliantUNIX.
5545 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5547 *) Add the configuration target debug-linux-ppro.
5548 Make 'openssl rsa' use the general key loading routines
5549 implemented in apps.c, and make those routines able to
5550 handle the key format FORMAT_NETSCAPE and the variant
5552 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5554 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5555 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5557 *) Add -keyform to rsautl, and document -engine.
5558 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5560 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5561 BIO_R_NO_SUCH_FILE error code rather than the generic
5562 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5565 *) Add new functions
5567 ERR_peek_last_error_line
5568 ERR_peek_last_error_line_data.
5569 These are similar to
5572 ERR_peek_error_line_data,
5573 but report on the latest error recorded rather than the first one
5574 still in the error queue.
5575 [Ben Laurie, Bodo Moeller]
5577 *) default_algorithms option in ENGINE config module. This allows things
5579 default_algorithms = ALL
5580 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5583 *) Preliminary ENGINE config module.
5586 *) New experimental application configuration code.
5589 *) Change the AES code to follow the same name structure as all other
5590 symmetric ciphers, and behave the same way. Move everything to
5591 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5592 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5594 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5595 [Ben Laurie and Theo de Raadt]
5597 *) Add option to output public keys in req command.
5598 [Massimiliano Pala madwolf@openca.org]
5600 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5601 (up to about 10% better than before for P-192 and P-224).
5604 *) New functions/macros
5606 SSL_CTX_set_msg_callback(ctx, cb)
5607 SSL_CTX_set_msg_callback_arg(ctx, arg)
5608 SSL_set_msg_callback(ssl, cb)
5609 SSL_set_msg_callback_arg(ssl, arg)
5611 to request calling a callback function
5613 void cb(int write_p, int version, int content_type,
5614 const void *buf, size_t len, SSL *ssl, void *arg)
5616 whenever a protocol message has been completely received
5617 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5618 protocol version according to which the SSL library interprets
5619 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5620 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5621 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5622 specification (change_cipher_spec(20), alert(21), handshake(22)).
5623 'buf' and 'len' point to the actual message, 'ssl' to the
5624 SSL object, and 'arg' is the application-defined value set by
5625 SSL[_CTX]_set_msg_callback_arg().
5627 'openssl s_client' and 'openssl s_server' have new '-msg' options
5628 to enable a callback that displays all protocol messages.
5631 *) Change the shared library support so shared libraries are built as
5632 soon as the corresponding static library is finished, and thereby get
5633 openssl and the test programs linked against the shared library.
5634 This still only happens when the keyword "shard" has been given to
5635 the configuration scripts.
5637 NOTE: shared library support is still an experimental thing, and
5638 backward binary compatibility is still not guaranteed.
5639 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5641 *) Add support for Subject Information Access extension.
5642 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5644 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5645 additional bytes when new memory had to be allocated, not just
5646 when reusing an existing buffer.
5649 *) New command line and configuration option 'utf8' for the req command.
5650 This allows field values to be specified as UTF8 strings.
5653 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5654 runs for the former and machine-readable output for the latter.
5657 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5658 of the e-mail address in the DN (i.e., it will go into a certificate
5659 extension only). The new configuration file option 'email_in_dn = no'
5660 has the same effect.
5661 [Massimiliano Pala madwolf@openca.org]
5663 *) Change all functions with names starting with des_ to be starting
5664 with DES_ instead. Add wrappers that are compatible with libdes,
5665 but are named _ossl_old_des_*. Finally, add macros that map the
5666 des_* symbols to the corresponding _ossl_old_des_* if libdes
5667 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5668 desired, the des_* symbols will be mapped to DES_*, with one
5671 Since we provide two compatibility mappings, the user needs to
5672 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5673 compatibility is desired. The default (i.e., when that macro
5674 isn't defined) is OpenSSL 0.9.6c compatibility.
5676 There are also macros that enable and disable the support of old
5677 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5678 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5679 are defined, the default will apply: to support the old des routines.
5681 In either case, one must include openssl/des.h to get the correct
5682 definitions. Do not try to just include openssl/des_old.h, that
5685 NOTE: This is a major break of an old API into a new one. Software
5686 authors are encouraged to switch to the DES_ style functions. Some
5687 time in the future, des_old.h and the libdes compatibility functions
5688 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5689 default), and then completely removed.
5692 *) Test for certificates which contain unsupported critical extensions.
5693 If such a certificate is found during a verify operation it is
5694 rejected by default: this behaviour can be overridden by either
5695 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5696 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5697 X509_supported_extension() has also been added which returns 1 if a
5698 particular extension is supported.
5701 *) Modify the behaviour of EVP cipher functions in similar way to digests
5702 to retain compatibility with existing code.
5705 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5706 compatibility with existing code. In particular the 'ctx' parameter does
5707 not have to be to be initialized before the call to EVP_DigestInit() and
5708 it is tidied up after a call to EVP_DigestFinal(). New function
5709 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5710 EVP_MD_CTX_copy() changed to not require the destination to be
5711 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5712 requires the destination to be valid.
5714 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5715 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5718 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5719 so that complete 'Handshake' protocol structures are kept in memory
5720 instead of overwriting 'msg_type' and 'length' with 'body' data.
5723 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5724 [Massimo Santin via Richard Levitte]
5726 *) Major restructuring to the underlying ENGINE code. This includes
5727 reduction of linker bloat, separation of pure "ENGINE" manipulation
5728 (initialisation, etc) from functionality dealing with implementations
5729 of specific crypto iterfaces. This change also introduces integrated
5730 support for symmetric ciphers and digest implementations - so ENGINEs
5731 can now accelerate these by providing EVP_CIPHER and EVP_MD
5732 implementations of their own. This is detailed in crypto/engine/README
5733 as it couldn't be adequately described here. However, there are a few
5734 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5735 were changed in the original introduction of ENGINE code have now
5736 reverted back - the hooking from this code to ENGINE is now a good
5737 deal more passive and at run-time, operations deal directly with
5738 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5739 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5740 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5741 they were not being used by the framework as there is no concept of a
5742 BIGNUM_METHOD and they could not be generalised to the new
5743 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5744 ENGINE_cpy() has been removed as it cannot be consistently defined in
5748 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5751 *) Change mkdef.pl to sort symbols that get the same entry number,
5752 and make sure the automatically generated functions ERR_load_*
5753 become part of libeay.num as well.
5756 *) New function SSL_renegotiate_pending(). This returns true once
5757 renegotiation has been requested (either SSL_renegotiate() call
5758 or HelloRequest/ClientHello received from the peer) and becomes
5759 false once a handshake has been completed.
5760 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5761 sends a HelloRequest, but does not ensure that a handshake takes
5762 place. SSL_renegotiate_pending() is useful for checking if the
5763 client has followed the request.)
5766 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5767 By default, clients may request session resumption even during
5768 renegotiation (if session ID contexts permit); with this option,
5769 session resumption is possible only in the first handshake.
5771 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5772 more bits available for options that should not be part of
5773 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5776 *) Add some demos for certificate and certificate request creation.
5779 *) Make maximum certificate chain size accepted from the peer application
5780 settable (SSL*_get/set_max_cert_list()), as proposed by
5781 "Douglas E. Engert" <deengert@anl.gov>.
5784 *) Add support for shared libraries for Unixware-7
5785 (Boyd Lynn Gerber <gerberb@zenez.com>).
5788 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5789 be done prior to destruction. Use this to unload error strings from
5790 ENGINEs that load their own error strings. NB: This adds two new API
5791 functions to "get" and "set" this destroy handler in an ENGINE.
5794 *) Alter all existing ENGINE implementations (except "openssl" and
5795 "openbsd") to dynamically instantiate their own error strings. This
5796 makes them more flexible to be built both as statically-linked ENGINEs
5797 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5798 Also, add stub code to each that makes building them as self-contained
5799 shared-libraries easier (see README.ENGINE).
5802 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5803 implementations into applications that are completely implemented in
5804 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5805 commands that can be used to configure what shared-library to load and
5806 to control aspects of the way it is handled. Also, made an update to
5807 the README.ENGINE file that brings its information up-to-date and
5808 provides some information and instructions on the "dynamic" ENGINE
5809 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5812 *) Make it possible to unload ranges of ERR strings with a new
5813 "ERR_unload_strings" function.
5816 *) Add a copy() function to EVP_MD.
5819 *) Make EVP_MD routines take a context pointer instead of just the
5820 md_data void pointer.
5823 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5824 that the digest can only process a single chunk of data
5825 (typically because it is provided by a piece of
5826 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5827 is only going to provide a single chunk of data, and hence the
5828 framework needn't accumulate the data for oneshot drivers.
5831 *) As with "ERR", make it possible to replace the underlying "ex_data"
5832 functions. This change also alters the storage and management of global
5833 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5834 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5835 index counters. The API functions that use this state have been changed
5836 to take a "class_index" rather than pointers to the class's local STACK
5837 and counter, and there is now an API function to dynamically create new
5838 classes. This centralisation allows us to (a) plug a lot of the
5839 thread-safety problems that existed, and (b) makes it possible to clean
5840 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5841 such data would previously have always leaked in application code and
5842 workarounds were in place to make the memory debugging turn a blind eye
5843 to it. Application code that doesn't use this new function will still
5844 leak as before, but their memory debugging output will announce it now
5845 rather than letting it slide.
5847 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5848 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5849 has a return value to indicate success or failure.
5852 *) Make it possible to replace the underlying "ERR" functions such that the
5853 global state (2 LHASH tables and 2 locks) is only used by the "default"
5854 implementation. This change also adds two functions to "get" and "set"
5855 the implementation prior to it being automatically set the first time
5856 any other ERR function takes place. Ie. an application can call "get",
5857 pass the return value to a module it has just loaded, and that module
5858 can call its own "set" function using that value. This means the
5859 module's "ERR" operations will use (and modify) the error state in the
5860 application and not in its own statically linked copy of OpenSSL code.
5863 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5864 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5865 the operation, and provides a more encapsulated way for external code
5866 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5867 to use these functions rather than manually incrementing the counts.
5869 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5872 *) Add EVP test program.
5875 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5878 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5879 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5880 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5881 These allow a CRL to be built without having to access X509_CRL fields
5882 directly. Modify 'ca' application to use new functions.
5885 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5886 bug workarounds. Rollback attack detection is a security feature.
5887 The problem will only arise on OpenSSL servers when TLSv1 is not
5888 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5889 Software authors not wanting to support TLSv1 will have special reasons
5890 for their choice and can explicitly enable this option.
5891 [Bodo Moeller, Lutz Jaenicke]
5893 *) Rationalise EVP so it can be extended: don't include a union of
5894 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5895 (similar to those existing for EVP_CIPHER_CTX).
5900 EVP_MD_CTX_init(&md); /* new function call */
5901 EVP_DigestInit(&md, EVP_sha1());
5902 EVP_DigestUpdate(&md, in, len);
5903 EVP_DigestFinal(&md, out, NULL);
5904 EVP_MD_CTX_cleanup(&md); /* new function call */
5908 *) Make DES key schedule conform to the usual scheme, as well as
5909 correcting its structure. This means that calls to DES functions
5910 now have to pass a pointer to a des_key_schedule instead of a
5911 plain des_key_schedule (which was actually always a pointer
5914 des_key_schedule ks;
5916 des_set_key_checked(..., &ks);
5917 des_ncbc_encrypt(..., &ks, ...);
5919 (Note that a later change renames 'des_...' into 'DES_...'.)
5922 *) Initial reduction of linker bloat: the use of some functions, such as
5923 PEM causes large amounts of unused functions to be linked in due to
5924 poor organisation. For example pem_all.c contains every PEM function
5925 which has a knock on effect of linking in large amounts of (unused)
5926 ASN1 code. Grouping together similar functions and splitting unrelated
5927 functions prevents this.
5930 *) Cleanup of EVP macros.
5933 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5934 correct _ecb suffix.
5937 *) Add initial OCSP responder support to ocsp application. The
5938 revocation information is handled using the text based index
5939 use by the ca application. The responder can either handle
5940 requests generated internally, supplied in files (for example
5941 via a CGI script) or using an internal minimal server.
5944 *) Add configuration choices to get zlib compression for TLS.
5947 *) Changes to Kerberos SSL for RFC 2712 compliance:
5948 1. Implemented real KerberosWrapper, instead of just using
5949 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5950 2. Implemented optional authenticator field of KerberosWrapper.
5952 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5953 and authenticator structs; see crypto/krb5/.
5955 Generalized Kerberos calls to support multiple Kerberos libraries.
5956 [Vern Staats <staatsvr@asc.hpc.mil>,
5957 Jeffrey Altman <jaltman@columbia.edu>
5958 via Richard Levitte]
5960 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5961 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5962 values for each of the key sizes rather than having just
5963 parameters (and 'speed' generating keys each time).
5966 *) Speed up EVP routines.
5969 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5970 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5971 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5972 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5974 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5975 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5976 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5979 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5981 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5984 *) Added the OS2-EMX target.
5985 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5987 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5988 to support NCONF routines in extension code. New function CONF_set_nconf()
5989 to allow functions which take an NCONF to also handle the old LHASH
5990 structure: this means that the old CONF compatible routines can be
5991 retained (in particular wrt extensions) without having to duplicate the
5992 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5995 *) Enhance the general user interface with mechanisms for inner control
5996 and with possibilities to have yes/no kind of prompts.
5999 *) Change all calls to low level digest routines in the library and
6000 applications to use EVP. Add missing calls to HMAC_cleanup() and
6001 don't assume HMAC_CTX can be copied using memcpy().
6002 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6004 *) Add the possibility to control engines through control names but with
6005 arbitrary arguments instead of just a string.
6006 Change the key loaders to take a UI_METHOD instead of a callback
6007 function pointer. NOTE: this breaks binary compatibility with earlier
6008 versions of OpenSSL [engine].
6009 Adapt the nCipher code for these new conditions and add a card insertion
6013 *) Enhance the general user interface with mechanisms to better support
6014 dialog box interfaces, application-defined prompts, the possibility
6015 to use defaults (for example default passwords from somewhere else)
6016 and interrupts/cancellations.
6019 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6020 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6023 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6024 tidy up some unnecessarily weird code in 'sk_new()').
6025 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6027 *) Change the key loading routines for ENGINEs to use the same kind
6028 callback (pem_password_cb) as all other routines that need this
6032 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6033 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6034 than this minimum value is recommended.
6037 *) New random seeder for OpenVMS, using the system process statistics
6038 that are easily reachable.
6041 *) Windows apparently can't transparently handle global
6042 variables defined in DLLs. Initialisations such as:
6044 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6046 wont compile. This is used by the any applications that need to
6047 declare their own ASN1 modules. This was fixed by adding the option
6048 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6049 needed for static libraries under Win32.
6052 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6053 setting of purpose and trust fields. New X509_STORE trust and
6054 purpose functions and tidy up setting in other SSL functions.
6057 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6058 structure. These are inherited by X509_STORE_CTX when it is
6059 initialised. This allows various defaults to be set in the
6060 X509_STORE structure (such as flags for CRL checking and custom
6061 purpose or trust settings) for functions which only use X509_STORE_CTX
6062 internally such as S/MIME.
6064 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6065 trust settings if they are not set in X509_STORE. This allows X509_STORE
6066 purposes and trust (in S/MIME for example) to override any set by default.
6068 Add command line options for CRL checking to smime, s_client and s_server
6072 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6073 are set then the CRL is looked up in the X509_STORE structure and
6074 its validity and signature checked, then if the certificate is found
6075 in the CRL the verify fails with a revoked error.
6077 Various new CRL related callbacks added to X509_STORE_CTX structure.
6079 Command line options added to 'verify' application to support this.
6081 This needs some additional work, such as being able to handle multiple
6082 CRLs with different times, extension based lookup (rather than just
6083 by subject name) and ultimately more complete V2 CRL extension
6087 *) Add a general user interface API (crypto/ui/). This is designed
6088 to replace things like des_read_password and friends (backward
6089 compatibility functions using this new API are provided).
6090 The purpose is to remove prompting functions from the DES code
6091 section as well as provide for prompting through dialog boxes in
6092 a window system and the like.
6095 *) Add "ex_data" support to ENGINE so implementations can add state at a
6096 per-structure level rather than having to store it globally.
6099 *) Make it possible for ENGINE structures to be copied when retrieved by
6100 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6101 This causes the "original" ENGINE structure to act like a template,
6102 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6103 operational state can be localised to each ENGINE structure, despite the
6104 fact they all share the same "methods". New ENGINE structures returned in
6105 this case have no functional references and the return value is the single
6106 structural reference. This matches the single structural reference returned
6107 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6111 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6112 needs to match any other type at all we need to manually clear the
6116 *) Changes to the "openssl engine" utility to include;
6117 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6118 about an ENGINE's available control commands.
6119 - executing control commands from command line arguments using the
6120 '-pre' and '-post' switches. '-post' is only used if '-t' is
6121 specified and the ENGINE is successfully initialised. The syntax for
6122 the individual commands are colon-separated, for example;
6123 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6126 *) New dynamic control command support for ENGINEs. ENGINEs can now
6127 declare their own commands (numbers), names (strings), descriptions,
6128 and input types for run-time discovery by calling applications. A
6129 subset of these commands are implicitly classed as "executable"
6130 depending on their input type, and only these can be invoked through
6131 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6132 can be based on user input, config files, etc). The distinction is
6133 that "executable" commands cannot return anything other than a boolean
6134 result and can only support numeric or string input, whereas some
6135 discoverable commands may only be for direct use through
6136 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6137 pointers, or other custom uses. The "executable" commands are to
6138 support parameterisations of ENGINE behaviour that can be
6139 unambiguously defined by ENGINEs and used consistently across any
6140 OpenSSL-based application. Commands have been added to all the
6141 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6142 control over shared-library paths without source code alterations.
6145 *) Changed all ENGINE implementations to dynamically allocate their
6146 ENGINEs rather than declaring them statically. Apart from this being
6147 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6148 this also allows the implementations to compile without using the
6149 internal engine_int.h header.
6152 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6153 'const' value. Any code that should be able to modify a RAND_METHOD
6154 should already have non-const pointers to it (ie. they should only
6155 modify their own ones).
6158 *) Made a variety of little tweaks to the ENGINE code.
6159 - "atalla" and "ubsec" string definitions were moved from header files
6160 to C code. "nuron" string definitions were placed in variables
6161 rather than hard-coded - allowing parameterisation of these values
6162 later on via ctrl() commands.
6163 - Removed unused "#if 0"'d code.
6164 - Fixed engine list iteration code so it uses ENGINE_free() to release
6165 structural references.
6166 - Constified the RAND_METHOD element of ENGINE structures.
6167 - Constified various get/set functions as appropriate and added
6168 missing functions (including a catch-all ENGINE_cpy that duplicates
6169 all ENGINE values onto a new ENGINE except reference counts/state).
6170 - Removed NULL parameter checks in get/set functions. Setting a method
6171 or function to NULL is a way of cancelling out a previously set
6172 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6173 and doesn't justify the extra error symbols and code.
6174 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6175 flags from engine_int.h to engine.h.
6176 - Changed prototypes for ENGINE handler functions (init(), finish(),
6177 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6180 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6181 to the algorithm using long division. The binary algorithm can be
6182 used only if the modulus is odd. On 32-bit systems, it is faster
6183 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6184 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6185 up to 450 bits. In 64-bit environments, the binary algorithm
6186 appears to be advantageous for much longer moduli; here we use it
6187 for moduli up to 2048 bits.
6190 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6191 could not support the combine flag in choice fields.
6194 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6195 extensions from a certificate request to the certificate.
6198 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6199 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6200 file: this allows the display of the certificate about to be
6201 signed to be customised, to allow certain fields to be included
6202 or excluded and extension details. The old system didn't display
6203 multicharacter strings properly, omitted fields not in the policy
6204 and couldn't display additional details such as extensions.
6207 *) Function EC_POINTs_mul for multiple scalar multiplication
6208 of an arbitrary number of elliptic curve points
6209 \sum scalars[i]*points[i],
6210 optionally including the generator defined for the EC_GROUP:
6211 scalar*generator + \sum scalars[i]*points[i].
6213 EC_POINT_mul is a simple wrapper function for the typical case
6214 that the point list has just one item (besides the optional
6218 *) First EC_METHODs for curves over GF(p):
6220 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6221 operations and provides various method functions that can also
6222 operate with faster implementations of modular arithmetic.
6224 EC_GFp_mont_method() reuses most functions that are part of
6225 EC_GFp_simple_method, but uses Montgomery arithmetic.
6227 [Bodo Moeller; point addition and point doubling
6228 implementation directly derived from source code provided by
6229 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6231 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6232 crypto/ec/ec_lib.c):
6234 Curves are EC_GROUP objects (with an optional group generator)
6235 based on EC_METHODs that are built into the library.
6237 Points are EC_POINT objects based on EC_GROUP objects.
6239 Most of the framework would be able to handle curves over arbitrary
6240 finite fields, but as there are no obvious types for fields other
6241 than GF(p), some functions are limited to that for now.
6244 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6245 that the file contains a complete HTTP response.
6248 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6249 change the def and num file printf format specifier from "%-40sXXX"
6250 to "%-39s XXX". The latter will always guarantee a space after the
6251 field while the former will cause them to run together if the field
6252 is 40 of more characters long.
6255 *) Constify the cipher and digest 'method' functions and structures
6256 and modify related functions to take constant EVP_MD and EVP_CIPHER
6260 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6261 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6264 *) Modify EVP_Digest*() routines so they now return values. Although the
6265 internal software routines can never fail additional hardware versions
6269 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6271 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6272 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6275 ERR_R_NESTED_ASN1_ERROR
6277 ERR_R_MISSING_ASN1_EOS
6278 were 4 .. 9, conflicting with
6279 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6281 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6282 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6284 Add new error code 'ERR_R_INTERNAL_ERROR'.
6287 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6291 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6292 sets the subject name for a new request or supersedes the
6293 subject name in a given request. Formats that can be parsed are
6294 'CN=Some Name, OU=myOU, C=IT'
6296 'CN=Some Name/OU=myOU/C=IT'.
6298 Add options '-batch' and '-verbose' to 'openssl req'.
6299 [Massimiliano Pala <madwolf@hackmasters.net>]
6301 *) Introduce the possibility to access global variables through
6302 functions on platform were that's the best way to handle exporting
6303 global variables in shared libraries. To enable this functionality,
6304 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6305 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6306 is normally done by Configure or something similar).
6308 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6309 in the source file (foo.c) like this:
6311 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6312 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6314 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6315 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6317 OPENSSL_DECLARE_GLOBAL(int,foo);
6318 #define foo OPENSSL_GLOBAL_REF(foo)
6319 OPENSSL_DECLARE_GLOBAL(double,bar);
6320 #define bar OPENSSL_GLOBAL_REF(bar)
6322 The #defines are very important, and therefore so is including the
6323 header file everywhere where the defined globals are used.
6325 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6326 of ASN.1 items, but that structure is a bit different.
6328 The largest change is in util/mkdef.pl which has been enhanced with
6329 better and easier to understand logic to choose which symbols should
6330 go into the Windows .def files as well as a number of fixes and code
6331 cleanup (among others, algorithm keywords are now sorted
6332 lexicographically to avoid constant rewrites).
6335 *) In BN_div() keep a copy of the sign of 'num' before writing the
6336 result to 'rm' because if rm==num the value will be overwritten
6337 and produce the wrong result if 'num' is negative: this caused
6338 problems with BN_mod() and BN_nnmod().
6341 *) Function OCSP_request_verify(). This checks the signature on an
6342 OCSP request and verifies the signer certificate. The signer
6343 certificate is just checked for a generic purpose and OCSP request
6347 *) Add OCSP_check_validity() function to check the validity of OCSP
6348 responses. OCSP responses are prepared in real time and may only
6349 be a few seconds old. Simply checking that the current time lies
6350 between thisUpdate and nextUpdate max reject otherwise valid responses
6351 caused by either OCSP responder or client clock inaccuracy. Instead
6352 we allow thisUpdate and nextUpdate to fall within a certain period of
6353 the current time. The age of the response can also optionally be
6354 checked. Two new options -validity_period and -status_age added to
6358 *) If signature or public key algorithm is unrecognized print out its
6359 OID rather that just UNKNOWN.
6362 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6363 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6364 ID to be generated from the issuer certificate alone which can then be
6365 passed to OCSP_id_issuer_cmp().
6368 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6369 ASN1 modules to export functions returning ASN1_ITEM pointers
6370 instead of the ASN1_ITEM structures themselves. This adds several
6371 new macros which allow the underlying ASN1 function/structure to
6372 be accessed transparently. As a result code should not use ASN1_ITEM
6373 references directly (such as &X509_it) but instead use the relevant
6374 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6375 use of the new ASN1 code on platforms where exporting structures
6376 is problematical (for example in shared libraries) but exporting
6377 functions returning pointers to structures is not.
6380 *) Add support for overriding the generation of SSL/TLS session IDs.
6381 These callbacks can be registered either in an SSL_CTX or per SSL.
6382 The purpose of this is to allow applications to control, if they wish,
6383 the arbitrary values chosen for use as session IDs, particularly as it
6384 can be useful for session caching in multiple-server environments. A
6385 command-line switch for testing this (and any client code that wishes
6386 to use such a feature) has been added to "s_server".
6387 [Geoff Thorpe, Lutz Jaenicke]
6389 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6390 of the form '#if defined(...) || defined(...) || ...' and
6391 '#if !defined(...) && !defined(...) && ...'. This also avoids
6392 the growing number of special cases it was previously handling.
6395 *) Make all configuration macros available for application by making
6396 sure they are available in opensslconf.h, by giving them names starting
6397 with "OPENSSL_" to avoid conflicts with other packages and by making
6398 sure e_os2.h will cover all platform-specific cases together with
6400 Additionally, it is now possible to define configuration/platform-
6401 specific names (called "system identities"). In the C code, these
6402 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6403 macro with the name beginning with "OPENSSL_SYS_", which is determined
6404 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6408 *) New option -set_serial to 'req' and 'x509' this allows the serial
6409 number to use to be specified on the command line. Previously self
6410 signed certificates were hard coded with serial number 0 and the
6411 CA options of 'x509' had to use a serial number in a file which was
6415 *) New options to 'ca' utility to support V2 CRL entry extensions.
6416 Currently CRL reason, invalidity date and hold instruction are
6417 supported. Add new CRL extensions to V3 code and some new objects.
6420 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6421 disable standard block padding (aka PKCS#5 padding) in the EVP
6422 API, which was previously mandatory. This means that the data is
6423 not padded in any way and so the total length much be a multiple
6424 of the block size, otherwise an error occurs.
6427 *) Initial (incomplete) OCSP SSL support.
6430 *) New function OCSP_parse_url(). This splits up a URL into its host,
6431 port and path components: primarily to parse OCSP URLs. New -url
6432 option to ocsp utility.
6435 *) New nonce behavior. The return value of OCSP_check_nonce() now
6436 reflects the various checks performed. Applications can decide
6437 whether to tolerate certain situations such as an absent nonce
6438 in a response when one was present in a request: the ocsp application
6439 just prints out a warning. New function OCSP_add1_basic_nonce()
6440 this is to allow responders to include a nonce in a response even if
6441 the request is nonce-less.
6444 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6445 skipped when using openssl x509 multiple times on a single input file,
6446 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6449 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6450 set string type: to handle setting ASN1_TIME structures. Fix ca
6451 utility to correctly initialize revocation date of CRLs.
6454 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6455 the clients preferred ciphersuites and rather use its own preferences.
6456 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6457 Internet Explorer by ensuring unchanged hash method during stepup.
6458 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6461 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6462 to aes and add a new 'exist' option to print out symbols that don't
6466 *) Additional options to ocsp utility to allow flags to be set and
6467 additional certificates supplied.
6470 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6471 OCSP client a number of certificate to only verify the response
6475 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6476 handle the new API. Currently only ECB, CBC modes supported. Add new
6479 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6480 Encryption Standard (AES) Ciphersuites for Transport Layer
6481 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6482 not enabled by default and were not part of the "ALL" ciphersuite
6483 alias because they were not yet official; they could be
6484 explicitly requested by specifying the "AESdraft" ciphersuite
6485 group alias. In the final release of OpenSSL 0.9.7, the group
6486 alias is called "AES" and is part of "ALL".)
6487 [Ben Laurie, Steve Henson, Bodo Moeller]
6489 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6490 request to response.
6493 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6494 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6495 extract information from a certificate request. OCSP_response_create()
6496 creates a response and optionally adds a basic response structure.
6497 OCSP_basic_add1_status() adds a complete single response to a basic
6498 response and returns the OCSP_SINGLERESP structure just added (to allow
6499 extensions to be included for example). OCSP_basic_add1_cert() adds a
6500 certificate to a basic response and OCSP_basic_sign() signs a basic
6501 response with various flags. New helper functions ASN1_TIME_check()
6502 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6503 (converts ASN1_TIME to GeneralizedTime).
6506 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6507 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6508 structure from a certificate. X509_pubkey_digest() digests the public_key
6509 contents: this is used in various key identifiers.
6512 *) Make sk_sort() tolerate a NULL argument.
6513 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6515 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6516 passed by the function are trusted implicitly. If any of them signed the
6517 response then it is assumed to be valid and is not verified.
6520 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6521 to data. This was previously part of the PKCS7 ASN1 code. This
6522 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6523 [Steve Henson, reported by Kenneth R. Robinette
6524 <support@securenetterm.com>]
6526 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6527 routines: without these tracing memory leaks is very painful.
6528 Fix leaks in PKCS12 and PKCS7 routines.
6531 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6532 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6533 effectively meant GeneralizedTime would never be used. Now it
6534 is initialised to -1 but X509_time_adj() now has to check the value
6535 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6536 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6537 [Steve Henson, reported by Kenneth R. Robinette
6538 <support@securenetterm.com>]
6540 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6541 result in a zero length in the ASN1_INTEGER structure which was
6542 not consistent with the structure when d2i_ASN1_INTEGER() was used
6543 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6544 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6545 where it did not print out a minus for negative ASN1_INTEGER.
6548 *) Add summary printout to ocsp utility. The various functions which
6549 convert status values to strings have been renamed to:
6550 OCSP_response_status_str(), OCSP_cert_status_str() and
6551 OCSP_crl_reason_str() and are no longer static. New options
6552 to verify nonce values and to disable verification. OCSP response
6553 printout format cleaned up.
6556 *) Add additional OCSP certificate checks. These are those specified
6557 in RFC2560. This consists of two separate checks: the CA of the
6558 certificate being checked must either be the OCSP signer certificate
6559 or the issuer of the OCSP signer certificate. In the latter case the
6560 OCSP signer certificate must contain the OCSP signing extended key
6561 usage. This check is performed by attempting to match the OCSP
6562 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6563 in the OCSP_CERTID structures of the response.
6566 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6567 and related routines. This uses the standard OpenSSL certificate
6568 verify routines to perform initial checks (just CA validity) and
6569 to obtain the certificate chain. Then additional checks will be
6570 performed on the chain. Currently the root CA is checked to see
6571 if it is explicitly trusted for OCSP signing. This is used to set
6572 a root CA as a global signing root: that is any certificate that
6573 chains to that CA is an acceptable OCSP signing certificate.
6576 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6577 extensions from a separate configuration file.
6578 As when reading extensions from the main configuration file,
6579 the '-extensions ...' option may be used for specifying the
6581 [Massimiliano Pala <madwolf@comune.modena.it>]
6583 *) New OCSP utility. Allows OCSP requests to be generated or
6584 read. The request can be sent to a responder and the output
6585 parsed, outputed or printed in text form. Not complete yet:
6586 still needs to check the OCSP response validity.
6589 *) New subcommands for 'openssl ca':
6590 'openssl ca -status <serial>' prints the status of the cert with
6591 the given serial number (according to the index file).
6592 'openssl ca -updatedb' updates the expiry status of certificates
6594 [Massimiliano Pala <madwolf@comune.modena.it>]
6596 *) New '-newreq-nodes' command option to CA.pl. This is like
6597 '-newreq', but calls 'openssl req' with the '-nodes' option
6598 so that the resulting key is not encrypted.
6599 [Damien Miller <djm@mindrot.org>]
6601 *) New configuration for the GNU Hurd.
6602 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6604 *) Initial code to implement OCSP basic response verify. This
6605 is currently incomplete. Currently just finds the signer's
6606 certificate and verifies the signature on the response.
6609 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6610 value of OPENSSLDIR. This is available via the new '-d' option
6611 to 'openssl version', and is also included in 'openssl version -a'.
6614 *) Allowing defining memory allocation callbacks that will be given
6615 file name and line number information in additional arguments
6616 (a const char* and an int). The basic functionality remains, as
6617 well as the original possibility to just replace malloc(),
6618 realloc() and free() by functions that do not know about these
6619 additional arguments. To register and find out the current
6620 settings for extended allocation functions, the following
6621 functions are provided:
6623 CRYPTO_set_mem_ex_functions
6624 CRYPTO_set_locked_mem_ex_functions
6625 CRYPTO_get_mem_ex_functions
6626 CRYPTO_get_locked_mem_ex_functions
6628 These work the same way as CRYPTO_set_mem_functions and friends.
6629 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6630 extended allocation function is enabled.
6631 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6632 a conventional allocation function is enabled.
6633 [Richard Levitte, Bodo Moeller]
6635 *) Finish off removing the remaining LHASH function pointer casts.
6636 There should no longer be any prototype-casting required when using
6637 the LHASH abstraction, and any casts that remain are "bugs". See
6638 the callback types and macros at the head of lhash.h for details
6639 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6642 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6643 If /dev/[u]random devices are not available or do not return enough
6644 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6646 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6647 /etc/entropy will be queried once each in this sequence, quering stops
6648 when enough entropy was collected without querying more sockets.
6651 *) Change the Unix RAND_poll() variant to be able to poll several
6652 random devices, as specified by DEVRANDOM, until a sufficient amount
6653 of data has been collected. We spend at most 10 ms on each file
6654 (select timeout) and read in non-blocking mode. DEVRANDOM now
6655 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6656 (previously it was just the string "/dev/urandom"), so on typical
6657 platforms the 10 ms delay will never occur.
6658 Also separate out the Unix variant to its own file, rand_unix.c.
6659 For VMS, there's a currently-empty rand_vms.c.
6662 *) Move OCSP client related routines to ocsp_cl.c. These
6663 provide utility functions which an application needing
6664 to issue a request to an OCSP responder and analyse the
6665 response will typically need: as opposed to those which an
6666 OCSP responder itself would need which will be added later.
6668 OCSP_request_sign() signs an OCSP request with an API similar
6669 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6670 response. OCSP_response_get1_basic() extracts basic response
6671 from response. OCSP_resp_find_status(): finds and extracts status
6672 information from an OCSP_CERTID structure (which will be created
6673 when the request structure is built). These are built from lower
6674 level functions which work on OCSP_SINGLERESP structures but
6675 wont normally be used unless the application wishes to examine
6676 extensions in the OCSP response for example.
6678 Replace nonce routines with a pair of functions.
6679 OCSP_request_add1_nonce() adds a nonce value and optionally
6680 generates a random value. OCSP_check_nonce() checks the
6681 validity of the nonce in an OCSP response.
6684 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6685 This doesn't copy the supplied OCSP_CERTID and avoids the
6686 need to free up the newly created id. Change return type
6687 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6688 This can then be used to add extensions to the request.
6689 Deleted OCSP_request_new(), since most of its functionality
6690 is now in OCSP_REQUEST_new() (and the case insensitive name
6691 clash) apart from the ability to set the request name which
6692 will be added elsewhere.
6695 *) Update OCSP API. Remove obsolete extensions argument from
6696 various functions. Extensions are now handled using the new
6697 OCSP extension code. New simple OCSP HTTP function which
6698 can be used to send requests and parse the response.
6701 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6702 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6703 uses the special reorder version of SET OF to sort the attributes
6704 and reorder them to match the encoded order. This resolves a long
6705 standing problem: a verify on a PKCS7 structure just after signing
6706 it used to fail because the attribute order did not match the
6707 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6708 it uses the received order. This is necessary to tolerate some broken
6709 software that does not order SET OF. This is handled by encoding
6710 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6711 to produce the required SET OF.
6714 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6715 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6716 files to get correct declarations of the ASN.1 item variables.
6719 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6720 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6721 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6722 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6723 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6724 ASN1_ITEM and no wrapper functions.
6727 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6728 replace the old function pointer based I/O routines. Change most of
6729 the *_d2i_bio() and *_d2i_fp() functions to use these.
6732 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6733 lines, recognice more "algorithms" that can be deselected, and make
6734 it complain about algorithm deselection that isn't recognised.
6737 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6738 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6739 to use new functions. Add NO_ASN1_OLD which can be set to remove
6740 some old style ASN1 functions: this can be used to determine if old
6741 code will still work when these eventually go away.
6744 *) New extension functions for OCSP structures, these follow the
6745 same conventions as certificates and CRLs.
6748 *) New function X509V3_add1_i2d(). This automatically encodes and
6749 adds an extension. Its behaviour can be customised with various
6750 flags to append, replace or delete. Various wrappers added for
6751 certifcates and CRLs.
6754 *) Fix to avoid calling the underlying ASN1 print routine when
6755 an extension cannot be parsed. Correct a typo in the
6756 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6759 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6760 entries for variables.
6763 *) Add functionality to apps/openssl.c for detecting locking
6764 problems: As the program is single-threaded, all we have
6765 to do is register a locking callback using an array for
6766 storing which locks are currently held by the program.
6769 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6770 SSL_get_ex_data_X509_STORE_idx(), which is used in
6771 ssl_verify_cert_chain() and thus can be called at any time
6772 during TLS/SSL handshakes so that thread-safety is essential.
6773 Unfortunately, the ex_data design is not at all suited
6774 for multi-threaded use, so it probably should be abolished.
6777 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6778 [Broadcom, tweaked and integrated by Geoff Thorpe]
6780 *) Move common extension printing code to new function
6781 X509V3_print_extensions(). Reorganise OCSP print routines and
6782 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6785 *) New function X509_signature_print() to remove duplication in some
6789 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6790 set (this was treated exactly the same as SET OF previously). This
6791 is used to reorder the STACK representing the structure to match the
6792 encoding. This will be used to get round a problem where a PKCS7
6793 structure which was signed could not be verified because the STACK
6794 order did not reflect the encoded order.
6797 *) Reimplement the OCSP ASN1 module using the new code.
6800 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6801 for its ASN1 operations. The old style function pointers still exist
6802 for now but they will eventually go away.
6805 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6806 completely replaces the old ASN1 functionality with a table driven
6807 encoder and decoder which interprets an ASN1_ITEM structure describing
6808 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6809 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6810 has also been converted to the new form.
6813 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6814 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6815 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6816 for negative moduli.
6819 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6820 of not touching the result's sign bit.
6823 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6827 *) Changed the LHASH code to use prototypes for callbacks, and created
6828 macros to declare and implement thin (optionally static) functions
6829 that provide type-safety and avoid function pointer casting for the
6830 type-specific callbacks.
6833 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6835 [Veers Staats <staatsvr@asc.hpc.mil>,
6836 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6838 *) Reformat the FAQ so the different questions and answers can be divided
6839 in sections depending on the subject.
6842 *) Have the zlib compression code load ZLIB.DLL dynamically under
6846 *) New function BN_mod_sqrt for computing square roots modulo a prime
6847 (using the probabilistic Tonelli-Shanks algorithm unless
6848 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6849 be handled deterministically).
6850 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6852 *) Make BN_mod_inverse faster by explicitly handling small quotients
6853 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6854 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6857 *) New function BN_kronecker.
6860 *) Fix BN_gcd so that it works on negative inputs; the result is
6861 positive unless both parameters are zero.
6862 Previously something reasonably close to an infinite loop was
6863 possible because numbers could be growing instead of shrinking
6864 in the implementation of Euclid's algorithm.
6867 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6868 sign of the number in question.
6870 Fix BN_is_word(a,w) to work correctly for w == 0.
6872 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6873 because its test if the absolute value of 'a' equals 'w'.
6874 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6875 it exists mostly for use in the implementations of BN_is_zero(),
6876 BN_is_one(), and BN_is_word().
6879 *) New function BN_swap.
6882 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6883 the exponentiation functions are more likely to produce reasonable
6884 results on negative inputs.
6887 *) Change BN_mod_mul so that the result is always non-negative.
6888 Previously, it could be negative if one of the factors was negative;
6889 I don't think anyone really wanted that behaviour.
6892 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6893 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6894 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6895 and add new functions:
6904 BN_mod_lshift1_quick
6908 These functions always generate non-negative results.
6910 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6911 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6913 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6914 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6915 be reduced modulo m.
6916 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6919 The following entry accidentally appeared in the CHANGES file
6920 distributed with OpenSSL 0.9.7. The modifications described in
6921 it do *not* apply to OpenSSL 0.9.7.
6923 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6924 was actually never needed) and in BN_mul(). The removal in BN_mul()
6925 required a small change in bn_mul_part_recursive() and the addition
6926 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6927 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6928 bn_sub_words() and bn_add_words() except they take arrays with
6933 *) In 'openssl passwd', verify passwords read from the terminal
6934 unless the '-salt' option is used (which usually means that
6935 verification would just waste user's time since the resulting
6936 hash is going to be compared with some given password hash)
6937 or the new '-noverify' option is used.
6939 This is an incompatible change, but it does not affect
6940 non-interactive use of 'openssl passwd' (passwords on the command
6941 line, '-stdin' option, '-in ...' option) and thus should not
6945 *) Remove all references to RSAref, since there's no more need for it.
6948 *) Make DSO load along a path given through an environment variable
6949 (SHLIB_PATH) with shl_load().
6952 *) Constify the ENGINE code as a result of BIGNUM constification.
6953 Also constify the RSA code and most things related to it. In a
6954 few places, most notable in the depth of the ASN.1 code, ugly
6955 casts back to non-const were required (to be solved at a later
6959 *) Make it so the openssl application has all engines loaded by default.
6962 *) Constify the BIGNUM routines a little more.
6965 *) Add the following functions:
6967 ENGINE_load_cswift()
6969 ENGINE_load_atalla()
6971 ENGINE_load_builtin_engines()
6973 That way, an application can itself choose if external engines that
6974 are built-in in OpenSSL shall ever be used or not. The benefit is
6975 that applications won't have to be linked with libdl or other dso
6976 libraries unless it's really needed.
6978 Changed 'openssl engine' to load all engines on demand.
6979 Changed the engine header files to avoid the duplication of some
6980 declarations (they differed!).
6983 *) 'openssl engine' can now list capabilities.
6986 *) Better error reporting in 'openssl engine'.
6989 *) Never call load_dh_param(NULL) in s_server.
6992 *) Add engine application. It can currently list engines by name and
6993 identity, and test if they are actually available.
6996 *) Improve RPM specification file by forcing symbolic linking and making
6997 sure the installed documentation is also owned by root.root.
6998 [Damien Miller <djm@mindrot.org>]
7000 *) Give the OpenSSL applications more possibilities to make use of
7001 keys (public as well as private) handled by engines.
7004 *) Add OCSP code that comes from CertCo.
7007 *) Add VMS support for the Rijndael code.
7010 *) Added untested support for Nuron crypto accelerator.
7013 *) Add support for external cryptographic devices. This code was
7014 previously distributed separately as the "engine" branch.
7015 [Geoff Thorpe, Richard Levitte]
7017 *) Rework the filename-translation in the DSO code. It is now possible to
7018 have far greater control over how a "name" is turned into a filename
7019 depending on the operating environment and any oddities about the
7020 different shared library filenames on each system.
7023 *) Support threads on FreeBSD-elf in Configure.
7026 *) Fix for SHA1 assembly problem with MASM: it produces
7027 warnings about corrupt line number information when assembling
7028 with debugging information. This is caused by the overlapping
7030 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7033 NCONF_get_number() has no error checking at all. As a replacement,
7034 NCONF_get_number_e() is defined (_e for "error checking") and is
7035 promoted strongly. The old NCONF_get_number is kept around for
7036 binary backward compatibility.
7037 Make it possible for methods to load from something other than a BIO,
7038 by providing a function pointer that is given a name instead of a BIO.
7039 For example, this could be used to load configuration data from an
7043 *) Fix for non blocking accept BIOs. Added new I/O special reason
7044 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7045 with non blocking I/O was not possible because no retry code was
7046 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7050 *) Added the beginnings of Rijndael support.
7053 *) Fix for bug in DirectoryString mask setting. Add support for
7054 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7055 to allow certificate printing to more controllable, additional
7056 'certopt' option to 'x509' to allow new printing options to be
7060 *) Clean old EAY MD5 hack from e_os.h.
7063 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7065 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7066 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7067 [Joe Orton, Steve Henson]
7069 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7071 *) Fix additional bug revealed by the NISCC test suite:
7073 Stop bug triggering large recursion when presented with
7074 certain ASN.1 tags (CVE-2003-0851)
7077 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7079 *) Fix various bugs revealed by running the NISCC test suite:
7081 Stop out of bounds reads in the ASN1 code when presented with
7082 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7084 If verify callback ignores invalid public key errors don't try to check
7085 certificate signature with the NULL public key.
7089 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7090 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7094 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7095 extra data after the compression methods not only for TLS 1.0
7096 but also for SSL 3.0 (as required by the specification).
7097 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7099 *) Change X509_certificate_type() to mark the key as exported/exportable
7100 when it's 512 *bits* long, not 512 bytes.
7103 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7105 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7106 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7107 a protocol version number mismatch like a decryption error
7108 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7111 *) Turn on RSA blinding by default in the default implementation
7112 to avoid a timing attack. Applications that don't want it can call
7113 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7114 They would be ill-advised to do so in most cases.
7115 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7117 *) Change RSA blinding code so that it works when the PRNG is not
7118 seeded (in this case, the secret RSA exponent is abused as
7119 an unpredictable seed -- if it is not unpredictable, there
7120 is no point in blinding anyway). Make RSA blinding thread-safe
7121 by remembering the creator's thread ID in rsa->blinding and
7122 having all other threads use local one-time blinding factors
7123 (this requires more computation than sharing rsa->blinding, but
7124 avoids excessive locking; and if an RSA object is not shared
7125 between threads, blinding will still be very fast).
7128 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7130 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7131 via timing by performing a MAC computation even if incorrrect
7132 block cipher padding has been found. This is a countermeasure
7133 against active attacks where the attacker has to distinguish
7134 between bad padding and a MAC verification error. (CVE-2003-0078)
7136 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7137 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7138 Martin Vuagnoux (EPFL, Ilion)]
7140 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7142 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7143 memory from it's contents. This is done with a counter that will
7144 place alternating values in each byte. This can be used to solve
7145 two issues: 1) the removal of calls to memset() by highly optimizing
7146 compilers, and 2) cleansing with other values than 0, since those can
7147 be read through on certain media, for example a swap space on disk.
7150 *) Bugfix: client side session caching did not work with external caching,
7151 because the session->cipher setting was not restored when reloading
7152 from the external cache. This problem was masked, when
7153 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7154 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7157 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7158 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7159 [Zeev Lieber <zeev-l@yahoo.com>]
7161 *) Undo an undocumented change introduced in 0.9.6e which caused
7162 repeated calls to OpenSSL_add_all_ciphers() and
7163 OpenSSL_add_all_digests() to be ignored, even after calling
7167 *) Change the default configuration reader to deal with last line not
7168 being properly terminated.
7171 *) Change X509_NAME_cmp() so it applies the special rules on handling
7172 DN values that are of type PrintableString, as well as RDNs of type
7173 emailAddress where the value has the type ia5String.
7174 [stefank@valicert.com via Richard Levitte]
7176 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7177 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7178 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7179 the bitwise-OR of the two for use by the majority of applications
7180 wanting this behaviour, and update the docs. The documented
7181 behaviour and actual behaviour were inconsistent and had been
7182 changing anyway, so this is more a bug-fix than a behavioural
7184 [Geoff Thorpe, diagnosed by Nadav Har'El]
7186 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7187 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7190 *) Fix initialization code race conditions in
7191 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7192 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7193 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7194 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7195 ssl2_get_cipher_by_char(),
7196 ssl3_get_cipher_by_char().
7197 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7199 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7200 the cached sessions are flushed, as the remove_cb() might use ex_data
7201 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7202 (see [openssl.org #212]).
7203 [Geoff Thorpe, Lutz Jaenicke]
7205 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7206 length, instead of the encoding length to d2i_ASN1_OBJECT.
7209 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7211 *) [In 0.9.6g-engine release:]
7212 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7213 [Lynn Gazis <lgazis@rainbow.com>]
7215 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7217 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7218 and get fix the header length calculation.
7219 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7220 Alon Kantor <alonk@checkpoint.com> (and others),
7223 *) Use proper error handling instead of 'assertions' in buffer
7224 overflow checks added in 0.9.6e. This prevents DoS (the
7225 assertions could call abort()).
7226 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7228 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7230 *) Add various sanity checks to asn1_get_length() to reject
7231 the ASN1 length bytes if they exceed sizeof(long), will appear
7232 negative or the content length exceeds the length of the
7234 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7236 *) Fix cipher selection routines: ciphers without encryption had no flags
7237 for the cipher strength set and where therefore not handled correctly
7238 by the selection routines (PR #130).
7241 *) Fix EVP_dsa_sha macro.
7245 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7246 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7247 that was added in OpenSSL 0.9.6d.
7249 As the countermeasure turned out to be incompatible with some
7250 broken SSL implementations, the new option is part of SSL_OP_ALL.
7251 SSL_OP_ALL is usually employed when compatibility with weird SSL
7252 implementations is desired (e.g. '-bugs' option to 's_client' and
7253 's_server'), so the new option is automatically set in many
7257 *) Changes in security patch:
7259 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7260 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7261 Air Force Materiel Command, USAF, under agreement number
7264 *) Add various sanity checks to asn1_get_length() to reject
7265 the ASN1 length bytes if they exceed sizeof(long), will appear
7266 negative or the content length exceeds the length of the
7267 supplied buffer. (CVE-2002-0659)
7268 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7270 *) Assertions for various potential buffer overflows, not known to
7272 [Ben Laurie (CHATS)]
7274 *) Various temporary buffers to hold ASCII versions of integers were
7275 too small for 64 bit platforms. (CVE-2002-0655)
7276 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7278 *) Remote buffer overflow in SSL3 protocol - an attacker could
7279 supply an oversized session ID to a client. (CVE-2002-0656)
7280 [Ben Laurie (CHATS)]
7282 *) Remote buffer overflow in SSL2 protocol - an attacker could
7283 supply an oversized client master key. (CVE-2002-0656)
7284 [Ben Laurie (CHATS)]
7286 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7288 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7289 encoded as NULL) with id-dsa-with-sha1.
7290 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7292 *) Check various X509_...() return values in apps/req.c.
7293 [Nils Larsch <nla@trustcenter.de>]
7295 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7296 an end-of-file condition would erronously be flagged, when the CRLF
7297 was just at the end of a processed block. The bug was discovered when
7298 processing data through a buffering memory BIO handing the data to a
7299 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7300 <ptsekov@syntrex.com> and Nedelcho Stanev.
7303 *) Implement a countermeasure against a vulnerability recently found
7304 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7305 before application data chunks to avoid the use of known IVs
7306 with data potentially chosen by the attacker.
7309 *) Fix length checks in ssl3_get_client_hello().
7312 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7313 to prevent ssl3_read_internal() from incorrectly assuming that
7314 ssl3_read_bytes() found application data while handshake
7315 processing was enabled when in fact s->s3->in_read_app_data was
7316 merely automatically cleared during the initial handshake.
7317 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7319 *) Fix object definitions for Private and Enterprise: they were not
7320 recognized in their shortname (=lowercase) representation. Extend
7321 obj_dat.pl to issue an error when using undefined keywords instead
7322 of silently ignoring the problem (Svenning Sorensen
7323 <sss@sss.dnsalias.net>).
7326 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7327 generators, i.e. generators other than 2 and 5. (Previously, the
7328 code did not properly initialise the 'add' and 'rem' values to
7329 BN_generate_prime().)
7331 In the new general case, we do not insist that 'generator' is
7332 actually a primitive root: This requirement is rather pointless;
7333 a generator of the order-q subgroup is just as good, if not
7337 *) Map new X509 verification errors to alerts. Discovered and submitted by
7338 Tom Wu <tom@arcot.com>.
7341 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7342 returning non-zero before the data has been completely received
7343 when using non-blocking I/O.
7344 [Bodo Moeller; problem pointed out by John Hughes]
7346 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7347 [Ben Laurie, Lutz Jaenicke]
7349 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7350 Yoram Zahavi <YoramZ@gilian.com>).
7353 *) Add information about CygWin 1.3 and on, and preserve proper
7354 configuration for the versions before that.
7355 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7357 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7358 check whether we deal with a copy of a session and do not delete from
7359 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7360 <izhar@checkpoint.com>.
7363 *) Do not store session data into the internal session cache, if it
7364 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7365 flag is set). Proposed by Aslam <aslam@funk.com>.
7368 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7372 *) [In 0.9.6d-engine release:]
7373 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7374 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7376 *) Add the configuration target linux-s390x.
7377 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7379 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7380 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7381 variable as an indication that a ClientHello message has been
7382 received. As the flag value will be lost between multiple
7383 invocations of ssl3_accept when using non-blocking I/O, the
7384 function may not be aware that a handshake has actually taken
7385 place, thus preventing a new session from being added to the
7388 To avoid this problem, we now set s->new_session to 2 instead of
7389 using a local variable.
7390 [Lutz Jaenicke, Bodo Moeller]
7392 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7393 if the SSL_R_LENGTH_MISMATCH error is detected.
7394 [Geoff Thorpe, Bodo Moeller]
7396 *) New 'shared_ldflag' column in Configure platform table.
7399 *) Fix EVP_CIPHER_mode macro.
7400 ["Dan S. Camper" <dan@bti.net>]
7402 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7403 type, we must throw them away by setting rr->length to 0.
7404 [D P Chang <dpc@qualys.com>]
7406 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7408 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7409 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7410 worked incorrectly for those cases where range = 10..._2 and
7411 3*range is two bits longer than range.)
7414 *) Only add signing time to PKCS7 structures if it is not already
7418 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7419 OBJ_ld_ce should be OBJ_id_ce.
7420 Also some ip-pda OIDs in crypto/objects/objects.txt were
7421 incorrect (cf. RFC 3039).
7422 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7424 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7425 returns early because it has nothing to do.
7426 [Andy Schneider <andy.schneider@bjss.co.uk>]
7428 *) [In 0.9.6c-engine release:]
7429 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7430 [Andy Schneider <andy.schneider@bjss.co.uk>]
7432 *) [In 0.9.6c-engine release:]
7433 Add support for Cryptographic Appliance's keyserver technology.
7434 (Use engine 'keyclient')
7435 [Cryptographic Appliances and Geoff Thorpe]
7437 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7438 is called via tools/c89.sh because arguments have to be
7439 rearranged (all '-L' options must appear before the first object
7441 [Richard Shapiro <rshapiro@abinitio.com>]
7443 *) [In 0.9.6c-engine release:]
7444 Add support for Broadcom crypto accelerator cards, backported
7446 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7448 *) [In 0.9.6c-engine release:]
7449 Add support for SureWare crypto accelerator cards from
7450 Baltimore Technologies. (Use engine 'sureware')
7451 [Baltimore Technologies and Mark Cox]
7453 *) [In 0.9.6c-engine release:]
7454 Add support for crypto accelerator cards from Accelerated
7455 Encryption Processing, www.aep.ie. (Use engine 'aep')
7456 [AEP Inc. and Mark Cox]
7458 *) Add a configuration entry for gcc on UnixWare.
7459 [Gary Benson <gbenson@redhat.com>]
7461 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7462 messages are stored in a single piece (fixed-length part and
7463 variable-length part combined) and fix various bugs found on the way.
7466 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7467 instead. BIO_gethostbyname() does not know what timeouts are
7468 appropriate, so entries would stay in cache even when they have
7470 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7472 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7473 faced with a pathologically small ClientHello fragment that does
7474 not contain client_version: Instead of aborting with an error,
7475 simply choose the highest available protocol version (i.e.,
7476 TLS 1.0 unless it is disabled). In practice, ClientHello
7477 messages are never sent like this, but this change gives us
7478 strictly correct behaviour at least for TLS.
7481 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7482 never resets s->method to s->ctx->method when called from within
7483 one of the SSL handshake functions.
7484 [Bodo Moeller; problem pointed out by Niko Baric]
7486 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7487 (sent using the client's version number) if client_version is
7488 smaller than the protocol version in use. Also change
7489 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7490 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7491 the client will at least see that alert.
7494 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7498 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7499 client receives HelloRequest while in a handshake.
7500 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7502 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7503 should end in 'break', not 'goto end' which circumvents various
7504 cleanups done in state SSL_ST_OK. But session related stuff
7505 must be disabled for SSL_ST_OK in the case that we just sent a
7508 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7509 before just sending a HelloRequest.
7510 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7512 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7513 reveal whether illegal block cipher padding was found or a MAC
7514 verification error occured. (Neither SSLerr() codes nor alerts
7515 are directly visible to potential attackers, but the information
7516 may leak via logfiles.)
7518 Similar changes are not required for the SSL 2.0 implementation
7519 because the number of padding bytes is sent in clear for SSL 2.0,
7520 and the extra bytes are just ignored. However ssl/s2_pkt.c
7521 failed to verify that the purported number of padding bytes is in
7525 *) Add OpenUNIX-8 support including shared libraries
7526 (Boyd Lynn Gerber <gerberb@zenez.com>).
7529 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7530 'wristwatch attack' using huge encoding parameters (cf.
7531 James H. Manger's CRYPTO 2001 paper). Note that the
7532 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7533 encoding parameters and hence was not vulnerable.
7536 *) BN_sqr() bug fix.
7537 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7539 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7540 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7541 followed by modular reduction.
7542 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7544 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7545 equivalent based on BN_pseudo_rand() instead of BN_rand().
7548 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7549 This function was broken, as the check for a new client hello message
7550 to handle SGC did not allow these large messages.
7551 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7554 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7557 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7558 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7561 *) Rework the configuration and shared library support for Tru64 Unix.
7562 The configuration part makes use of modern compiler features and
7563 still retains old compiler behavior for those that run older versions
7564 of the OS. The shared library support part includes a variant that
7565 uses the RPATH feature, and is available through the special
7566 configuration target "alpha-cc-rpath", which will never be selected
7568 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7570 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7571 with the same message size as in ssl3_get_certificate_request().
7572 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7573 messages might inadvertently be reject as too long.
7574 [Petr Lampa <lampa@fee.vutbr.cz>]
7576 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7579 *) Modified SSL library such that the verify_callback that has been set
7580 specificly for an SSL object with SSL_set_verify() is actually being
7581 used. Before the change, a verify_callback set with this function was
7582 ignored and the verify_callback() set in the SSL_CTX at the time of
7583 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7584 to allow the necessary settings.
7587 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7588 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7589 done automatically (in contradiction to the requirements of the C
7590 standard). This made problems when used from OpenSSH.
7593 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7594 dh->length and always used
7596 BN_rand_range(priv_key, dh->p).
7598 BN_rand_range() is not necessary for Diffie-Hellman, and this
7599 specific range makes Diffie-Hellman unnecessarily inefficient if
7600 dh->length (recommended exponent length) is much smaller than the
7601 length of dh->p. We could use BN_rand_range() if the order of
7602 the subgroup was stored in the DH structure, but we only have
7607 BN_rand(priv_key, l, ...)
7609 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7615 RSA_eay_public_encrypt
7616 RSA_eay_private_decrypt
7617 RSA_eay_private_encrypt (signing)
7618 RSA_eay_public_decrypt (signature verification)
7620 (default implementations for RSA_public_encrypt,
7621 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7622 always reject numbers >= n.
7625 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7626 to synchronize access to 'locking_thread'. This is necessary on
7627 systems where access to 'locking_thread' (an 'unsigned long'
7628 variable) is not atomic.
7631 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7632 *before* setting the 'crypto_lock_rand' flag. The previous code had
7633 a race condition if 0 is a valid thread ID.
7634 [Travis Vitek <vitek@roguewave.com>]
7636 *) Add support for shared libraries under Irix.
7637 [Albert Chin-A-Young <china@thewrittenword.com>]
7639 *) Add configuration option to build on Linux on both big-endian and
7641 [Ralf Baechle <ralf@uni-koblenz.de>]
7643 *) Add the possibility to create shared libraries on HP-UX.
7646 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7648 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7649 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7650 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7651 PRNG state recovery was possible based on the output of
7652 one PRNG request appropriately sized to gain knowledge on
7653 'md' followed by enough consecutive 1-byte PRNG requests
7654 to traverse all of 'state'.
7656 1. When updating 'md_local' (the current thread's copy of 'md')
7657 during PRNG output generation, hash all of the previous
7658 'md_local' value, not just the half used for PRNG output.
7660 2. Make the number of bytes from 'state' included into the hash
7661 independent from the number of PRNG bytes requested.
7663 The first measure alone would be sufficient to avoid
7664 Markku-Juhani's attack. (Actually it had never occurred
7665 to me that the half of 'md_local' used for chaining was the
7666 half from which PRNG output bytes were taken -- I had always
7667 assumed that the secret half would be used.) The second
7668 measure makes sure that additional data from 'state' is never
7669 mixed into 'md_local' in small portions; this heuristically
7670 further strengthens the PRNG.
7673 *) Fix crypto/bn/asm/mips3.s.
7676 *) When only the key is given to "enc", the IV is undefined. Print out
7677 an error message in this case.
7680 *) Handle special case when X509_NAME is empty in X509 printing routines.
7683 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7684 positive and less than q.
7687 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7688 used: it isn't thread safe and the add_lock_callback should handle
7690 [Paul Rose <Paul.Rose@bridge.com>]
7692 *) Verify that incoming data obeys the block size in
7693 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7697 [Ulf Möller, Bodo Möller]
7699 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7700 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7701 when fixing the server behaviour for backwards-compatible 'client
7702 hello' messages. (Note that the attack is impractical against
7703 SSL 3.0 and TLS 1.0 anyway because length and version checking
7704 means that the probability of guessing a valid ciphertext is
7705 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7708 Before 0.9.5, the countermeasure (hide the error by generating a
7709 random 'decryption result') did not work properly because
7710 ERR_clear_error() was missing, meaning that SSL_get_error() would
7711 detect the supposedly ignored error.
7713 Both problems are now fixed.
7716 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7717 (previously it was 1024).
7720 *) Fix for compatibility mode trust settings: ignore trust settings
7721 unless some valid trust or reject settings are present.
7724 *) Fix for blowfish EVP: its a variable length cipher.
7727 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7728 parameters in DSA public key structures and return an error in the
7729 DSA routines if parameters are absent.
7732 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7733 in the current directory if neither $RANDFILE nor $HOME was set.
7734 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7735 caused some confusion to Windows users who haven't defined $HOME.
7736 Thus RAND_file_name() is changed again: e_os.h can define a
7737 DEFAULT_HOME, which will be used if $HOME is not set.
7738 For Windows, we use "C:"; on other platforms, we still require
7739 environment variables.
7741 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7742 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7743 having multiple threads call RAND_poll() concurrently.
7746 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7747 combination of a flag and a thread ID variable.
7748 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7749 flag), *other* threads can enter ssleay_add_bytes without obeying
7750 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7751 that they do not hold after the first thread unsets add_do_not_lock).
7754 *) Change bctest again: '-x' expressions are not available in all
7758 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7760 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7761 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7763 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7764 the default extension for executables, if any. Also, make the perl
7765 scripts that use symlink() to test if it really exists and use "cp"
7766 if it doesn't. All this made OpenSSL compilable and installable in
7770 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7771 If SEQUENCE is length is indefinite just set c->slen to the total
7772 amount of data available.
7773 [Steve Henson, reported by shige@FreeBSD.org]
7774 [This change does not apply to 0.9.7.]
7776 *) Change bctest to avoid here-documents inside command substitution
7777 (workaround for FreeBSD /bin/sh bug).
7778 For compatibility with Ultrix, avoid shell functions (introduced
7779 in the bctest version that searches along $PATH).
7782 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7783 with des_encrypt() defined on some operating systems, like Solaris
7787 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7788 On the Importance of Eliminating Errors in Cryptographic
7789 Computations, J. Cryptology 14 (2001) 2, 101-119,
7790 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7793 *) MIPS assembler BIGNUM division bug fix.
7796 *) Disabled incorrect Alpha assembler code.
7799 *) Fix PKCS#7 decode routines so they correctly update the length
7800 after reading an EOC for the EXPLICIT tag.
7802 [This change does not apply to 0.9.7.]
7804 *) Fix bug in PKCS#12 key generation routines. This was triggered
7805 if a 3DES key was generated with a 0 initial byte. Include
7806 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7807 (but broken) behaviour.
7810 *) Enhance bctest to search for a working bc along $PATH and print
7812 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7814 *) Fix memory leaks in err.c: free err_data string if necessary;
7815 don't write to the wrong index in ERR_set_error_data.
7818 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7822 *) Replace rdtsc with _emit statements for VC++ version 5.
7823 [Jeremy Cooper <jeremy@baymoo.org>]
7825 *) Make it possible to reuse SSLv2 sessions.
7828 *) In copy_email() check for >= 0 as a return value for
7829 X509_NAME_get_index_by_NID() since 0 is a valid index.
7830 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7832 *) Avoid coredump with unsupported or invalid public keys by checking if
7833 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7834 PKCS7_verify() fails with non detached data.
7837 *) Don't use getenv in library functions when run as setuid/setgid.
7838 New function OPENSSL_issetugid().
7841 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7842 due to incorrect handling of multi-threading:
7844 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7846 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7848 3. Count how many times MemCheck_off() has been called so that
7849 nested use can be treated correctly. This also avoids
7850 inband-signalling in the previous code (which relied on the
7851 assumption that thread ID 0 is impossible).
7854 *) Add "-rand" option also to s_client and s_server.
7857 *) Fix CPU detection on Irix 6.x.
7858 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7859 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7861 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7864 [This change does not apply to 0.9.7.]
7866 *) Use the cached encoding of an X509_NAME structure rather than
7867 copying it. This is apparently the reason for the libsafe "errors"
7868 but the code is actually correct.
7871 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7872 Bleichenbacher's DSA attack.
7873 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7874 to be set and top=0 forces the highest bit to be set; top=-1 is new
7875 and leaves the highest bit random.
7876 [Ulf Moeller, Bodo Moeller]
7878 *) In the NCONF_...-based implementations for CONF_... queries
7879 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7880 a temporary CONF structure with the data component set to NULL
7881 (which gives segmentation faults in lh_retrieve).
7882 Instead, use NULL for the CONF pointer in CONF_get_string and
7883 CONF_get_number (which may use environment variables) and directly
7884 return NULL from CONF_get_section.
7887 *) Fix potential buffer overrun for EBCDIC.
7890 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7891 keyUsage if basicConstraints absent for a CA.
7894 *) Make SMIME_write_PKCS7() write mail header values with a format that
7895 is more generally accepted (no spaces before the semicolon), since
7896 some programs can't parse those values properly otherwise. Also make
7897 sure BIO's that break lines after each write do not create invalid
7901 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7902 macros previously used would not encode an empty SEQUENCE OF
7903 and break the signature.
7905 [This change does not apply to 0.9.7.]
7907 *) Zero the premaster secret after deriving the master secret in
7911 *) Add some EVP_add_digest_alias registrations (as found in
7912 OpenSSL_add_all_digests()) to SSL_library_init()
7913 aka OpenSSL_add_ssl_algorithms(). This provides improved
7914 compatibility with peers using X.509 certificates
7915 with unconventional AlgorithmIdentifier OIDs.
7918 *) Fix for Irix with NO_ASM.
7919 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7921 *) ./config script fixes.
7922 [Ulf Moeller, Richard Levitte]
7924 *) Fix 'openssl passwd -1'.
7927 *) Change PKCS12_key_gen_asc() so it can cope with non null
7928 terminated strings whose length is passed in the passlen
7929 parameter, for example from PEM callbacks. This was done
7930 by adding an extra length parameter to asc2uni().
7931 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7933 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7934 call failed, free the DSA structure.
7937 *) Fix to uni2asc() to cope with zero length Unicode strings.
7938 These are present in some PKCS#12 files.
7941 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7942 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7943 when writing a 32767 byte record.
7944 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7946 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7947 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7949 (RSA objects have a reference count access to which is protected
7950 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7951 so they are meant to be shared between threads.)
7952 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7953 "Reddie, Steven" <Steven.Reddie@ca.com>]
7955 *) Fix a deadlock in CRYPTO_mem_leaks().
7958 *) Use better test patterns in bntest.
7961 *) rand_win.c fix for Borland C.
7964 *) BN_rshift bugfix for n == 0.
7967 *) Add a 'bctest' script that checks for some known 'bc' bugs
7968 so that 'make test' does not abort just because 'bc' is broken.
7971 *) Store verify_result within SSL_SESSION also for client side to
7972 avoid potential security hole. (Re-used sessions on the client side
7973 always resulted in verify_result==X509_V_OK, not using the original
7974 result of the server certificate verification.)
7977 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7978 SSL3_RT_APPLICATION_DATA, return 0.
7979 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7983 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7984 releases, have been re-implemented by renaming the previous
7985 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7986 and ssl3_read_internal, respectively, and adding 'peek' parameters
7987 to them. The new ssl[23]_{read,peek} functions are calls to
7988 ssl[23]_read_internal with the 'peek' flag set appropriately.
7989 A 'peek' parameter has also been added to ssl3_read_bytes, which
7990 does the actual work for ssl3_read_internal.
7993 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7994 the method-specific "init()" handler. Also clean up ex_data after
7995 calling the method-specific "finish()" handler. Previously, this was
7996 happening the other way round.
7999 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8000 The previous value, 12, was not always sufficient for BN_mod_exp().
8003 *) Make sure that shared libraries get the internal name engine with
8004 the full version number and not just 0. This should mark the
8005 shared libraries as not backward compatible. Of course, this should
8006 be changed again when we can guarantee backward binary compatibility.
8009 *) Fix typo in get_cert_by_subject() in by_dir.c
8010 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8012 *) Rework the system to generate shared libraries:
8014 - Make note of the expected extension for the shared libraries and
8015 if there is a need for symbolic links from for example libcrypto.so.0
8016 to libcrypto.so.0.9.7. There is extended info in Configure for
8019 - Make as few rebuilds of the shared libraries as possible.
8021 - Still avoid linking the OpenSSL programs with the shared libraries.
8023 - When installing, install the shared libraries separately from the
8027 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8029 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8030 and not in SSL_clear because the latter is also used by the
8031 accept/connect functions; previously, the settings made by
8032 SSL_set_read_ahead would be lost during the handshake.
8033 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8035 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8036 Previously, it would create entries for disabled algorithms no
8040 *) Added several new manual pages for SSL_* function.
8043 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8045 *) In ssl23_get_client_hello, generate an error message when faced
8046 with an initial SSL 3.0/TLS record that is too small to contain the
8047 first two bytes of the ClientHello message, i.e. client_version.
8048 (Note that this is a pathologic case that probably has never happened
8049 in real life.) The previous approach was to use the version number
8050 from the record header as a substitute; but our protocol choice
8051 should not depend on that one because it is not authenticated
8052 by the Finished messages.
8055 *) More robust randomness gathering functions for Windows.
8056 [Jeffrey Altman <jaltman@columbia.edu>]
8058 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8059 not set then we don't setup the error code for issuer check errors
8060 to avoid possibly overwriting other errors which the callback does
8061 handle. If an application does set the flag then we assume it knows
8062 what it is doing and can handle the new informational codes
8066 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8067 a general "ANY" type, as such it should be able to decode anything
8068 including tagged types. However it didn't check the class so it would
8069 wrongly interpret tagged types in the same way as their universal
8070 counterpart and unknown types were just rejected. Changed so that the
8071 tagged and unknown types are handled in the same way as a SEQUENCE:
8072 that is the encoding is stored intact. There is also a new type
8073 "V_ASN1_OTHER" which is used when the class is not universal, in this
8074 case we have no idea what the actual type is so we just lump them all
8078 *) On VMS, stdout may very well lead to a file that is written to
8079 in a record-oriented fashion. That means that every write() will
8080 write a separate record, which will be read separately by the
8081 programs trying to read from it. This can be very confusing.
8083 The solution is to put a BIO filter in the way that will buffer
8084 text until a linefeed is reached, and then write everything a
8085 line at a time, so every record written will be an actual line,
8086 not chunks of lines and not (usually doesn't happen, but I've
8087 seen it once) several lines in one record. BIO_f_linebuffer() is
8090 Currently, it's a VMS-only method, because that's where it has
8091 been tested well enough.
8094 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8095 it can return incorrect results.
8096 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8097 but it was in 0.9.6-beta[12].)
8100 *) Disable the check for content being present when verifying detached
8101 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8102 include zero length content when signing messages.
8105 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8106 BIO_ctrl (for BIO pairs).
8109 *) Add DSO method for VMS.
8112 *) Bug fix: Montgomery multiplication could produce results with the
8116 *) Add RPM specification openssl.spec and modify it to build three
8117 packages. The default package contains applications, application
8118 documentation and run-time libraries. The devel package contains
8119 include files, static libraries and function documentation. The
8120 doc package contains the contents of the doc directory. The original
8121 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8124 *) Add a large number of documentation files for many SSL routines.
8125 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8127 *) Add a configuration entry for Sony News 4.
8128 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8130 *) Don't set the two most significant bits to one when generating a
8131 random number < q in the DSA library.
8134 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8135 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8136 the underlying transport is blocking) if a handshake took place.
8137 (The default behaviour is needed by applications such as s_client
8138 and s_server that use select() to determine when to use SSL_read;
8139 but for applications that know in advance when to expect data, it
8140 just makes things more complicated.)
8143 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8147 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8148 work better on such systems.
8149 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8151 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8152 Update PKCS12_parse() so it copies the friendlyName and the
8153 keyid to the certificates aux info.
8156 *) Fix bug in PKCS7_verify() which caused an infinite loop
8157 if there was more than one signature.
8158 [Sven Uszpelkat <su@celocom.de>]
8160 *) Major change in util/mkdef.pl to include extra information
8161 about each symbol, as well as presenting variables as well
8162 as functions. This change means that there's n more need
8163 to rebuild the .num files when some algorithms are excluded.
8166 *) Allow the verify time to be set by an application,
8167 rather than always using the current time.
8170 *) Phase 2 verify code reorganisation. The certificate
8171 verify code now looks up an issuer certificate by a
8172 number of criteria: subject name, authority key id
8173 and key usage. It also verifies self signed certificates
8174 by the same criteria. The main comparison function is
8175 X509_check_issued() which performs these checks.
8177 Lot of changes were necessary in order to support this
8178 without completely rewriting the lookup code.
8180 Authority and subject key identifier are now cached.
8182 The LHASH 'certs' is X509_STORE has now been replaced
8183 by a STACK_OF(X509_OBJECT). This is mainly because an
8184 LHASH can't store or retrieve multiple objects with
8185 the same hash value.
8187 As a result various functions (which were all internal
8188 use only) have changed to handle the new X509_STORE
8189 structure. This will break anything that messed round
8190 with X509_STORE internally.
8192 The functions X509_STORE_add_cert() now checks for an
8193 exact match, rather than just subject name.
8195 The X509_STORE API doesn't directly support the retrieval
8196 of multiple certificates matching a given criteria, however
8197 this can be worked round by performing a lookup first
8198 (which will fill the cache with candidate certificates)
8199 and then examining the cache for matches. This is probably
8200 the best we can do without throwing out X509_LOOKUP
8201 entirely (maybe later...).
8203 The X509_VERIFY_CTX structure has been enhanced considerably.
8205 All certificate lookup operations now go via a get_issuer()
8206 callback. Although this currently uses an X509_STORE it
8207 can be replaced by custom lookups. This is a simple way
8208 to bypass the X509_STORE hackery necessary to make this
8209 work and makes it possible to use more efficient techniques
8210 in future. A very simple version which uses a simple
8211 STACK for its trusted certificate store is also provided
8212 using X509_STORE_CTX_trusted_stack().
8214 The verify_cb() and verify() callbacks now have equivalents
8215 in the X509_STORE_CTX structure.
8217 X509_STORE_CTX also has a 'flags' field which can be used
8218 to customise the verify behaviour.
8221 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8222 excludes S/MIME capabilities.
8225 *) When a certificate request is read in keep a copy of the
8226 original encoding of the signed data and use it when outputing
8227 again. Signatures then use the original encoding rather than
8228 a decoded, encoded version which may cause problems if the
8229 request is improperly encoded.
8232 *) For consistency with other BIO_puts implementations, call
8233 buffer_write(b, ...) directly in buffer_puts instead of calling
8236 In BIO_puts, increment b->num_write as in BIO_write.
8237 [Peter.Sylvester@EdelWeb.fr]
8239 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8240 BN_zero, we may not return a BIGNUM with an array consisting of
8244 *) Avoid calling abort() from within the library when problems are
8245 detected, except if preprocessor symbols have been defined
8246 (such as REF_CHECK, BN_DEBUG etc.).
8249 *) New openssl application 'rsautl'. This utility can be
8250 used for low level RSA operations. DER public key
8251 BIO/fp routines also added.
8254 *) New Configure entry and patches for compiling on QNX 4.
8255 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8257 *) A demo state-machine implementation was sponsored by
8258 Nuron (http://www.nuron.com/) and is now available in
8259 demos/state_machine.
8262 *) New options added to the 'dgst' utility for signature
8263 generation and verification.
8266 *) Unrecognized PKCS#7 content types are now handled via a
8267 catch all ASN1_TYPE structure. This allows unsupported
8268 types to be stored as a "blob" and an application can
8269 encode and decode it manually.
8272 *) Fix various signed/unsigned issues to make a_strex.c
8274 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8276 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8277 length if passed a buffer. ASN1_INTEGER_to_BN failed
8278 if passed a NULL BN and its argument was negative.
8279 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8281 *) Modification to PKCS#7 encoding routines to output definite
8282 length encoding. Since currently the whole structures are in
8283 memory there's not real point in using indefinite length
8284 constructed encoding. However if OpenSSL is compiled with
8285 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8288 *) Added BIO_vprintf() and BIO_vsnprintf().
8291 *) Added more prefixes to parse for in the the strings written
8292 through a logging bio, to cover all the levels that are available
8293 through syslog. The prefixes are now:
8295 PANIC, EMERG, EMR => LOG_EMERG
8296 ALERT, ALR => LOG_ALERT
8297 CRIT, CRI => LOG_CRIT
8298 ERROR, ERR => LOG_ERR
8299 WARNING, WARN, WAR => LOG_WARNING
8300 NOTICE, NOTE, NOT => LOG_NOTICE
8301 INFO, INF => LOG_INFO
8302 DEBUG, DBG => LOG_DEBUG
8304 and as before, if none of those prefixes are present at the
8305 beginning of the string, LOG_ERR is chosen.
8307 On Win32, the LOG_* levels are mapped according to this:
8309 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8310 LOG_WARNING => EVENTLOG_WARNING_TYPE
8311 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8315 *) Made it possible to reconfigure with just the configuration
8316 argument "reconf" or "reconfigure". The command line arguments
8317 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8318 and are retrieved from there when reconfiguring.
8322 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8324 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8327 *) The obj_dat.pl script was messing up the sorting of object
8328 names. The reason was that it compared the quoted version
8329 of strings as a result "OCSP" > "OCSP Signing" because
8330 " > SPACE. Changed script to store unquoted versions of
8331 names and add quotes on output. It was also omitting some
8332 names from the lookup table if they were given a default
8333 value (that is if SN is missing it is given the same
8334 value as LN and vice versa), these are now added on the
8335 grounds that if an object has a name we should be able to
8336 look it up. Finally added warning output when duplicate
8337 short or long names are found.
8340 *) Changes needed for Tandem NSK.
8341 [Scott Uroff <scott@xypro.com>]
8343 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8344 RSA_padding_check_SSLv23(), special padding was never detected
8345 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8346 version rollback attacks was not effective.
8348 In s23_clnt.c, don't use special rollback-attack detection padding
8349 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8350 client; similarly, in s23_srvr.c, don't do the rollback check if
8351 SSL 2.0 is the only protocol enabled in the server.
8354 *) Make it possible to get hexdumps of unprintable data with 'openssl
8355 asn1parse'. By implication, the functions ASN1_parse_dump() and
8356 BIO_dump_indent() are added.
8359 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8360 these print out strings and name structures based on various
8361 flags including RFC2253 support and proper handling of
8362 multibyte characters. Added options to the 'x509' utility
8363 to allow the various flags to be set.
8366 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8367 Also change the functions X509_cmp_current_time() and
8368 X509_gmtime_adj() work with an ASN1_TIME structure,
8369 this will enable certificates using GeneralizedTime in validity
8370 dates to be checked.
8373 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8374 negative public key encodings) on by default,
8375 NO_NEG_PUBKEY_BUG can be set to disable it.
8378 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8379 content octets. An i2c_ASN1_OBJECT is unnecessary because
8380 the encoding can be trivially obtained from the structure.
8383 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8384 not read locks (CRYPTO_r_[un]lock).
8387 *) A first attempt at creating official support for shared
8388 libraries through configuration. I've kept it so the
8389 default is static libraries only, and the OpenSSL programs
8390 are always statically linked for now, but there are
8391 preparations for dynamic linking in place.
8392 This has been tested on Linux and Tru64.
8395 *) Randomness polling function for Win9x, as described in:
8396 Peter Gutmann, Software Generation of Practically Strong
8400 *) Fix so PRNG is seeded in req if using an already existing
8404 *) New options to smime application. -inform and -outform
8405 allow alternative formats for the S/MIME message including
8406 PEM and DER. The -content option allows the content to be
8407 specified separately. This should allow things like Netscape
8408 form signing output easier to verify.
8411 *) Fix the ASN1 encoding of tags using the 'long form'.
8414 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8415 STRING types. These convert content octets to and from the
8416 underlying type. The actual tag and length octets are
8417 already assumed to have been read in and checked. These
8418 are needed because all other string types have virtually
8419 identical handling apart from the tag. By having versions
8420 of the ASN1 functions that just operate on content octets
8421 IMPLICIT tagging can be handled properly. It also allows
8422 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8423 and ASN1_INTEGER are identical apart from the tag.
8426 *) Change the handling of OID objects as follows:
8428 - New object identifiers are inserted in objects.txt, following
8429 the syntax given in objects.README.
8430 - objects.pl is used to process obj_mac.num and create a new
8432 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8435 This is currently kind of a hack, and the perl code in objects.pl
8436 isn't very elegant, but it works as I intended. The simplest way
8437 to check that it worked correctly is to look in obj_dat.h and
8438 check the array nid_objs and make sure the objects haven't moved
8439 around (this is important!). Additions are OK, as well as
8440 consistent name changes.
8443 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8446 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8447 The given file adds to whatever has already been seeded into the
8448 random pool through the RANDFILE configuration file option or
8449 environment variable, or the default random state file.
8452 *) mkstack.pl now sorts each macro group into lexical order.
8453 Previously the output order depended on the order the files
8454 appeared in the directory, resulting in needless rewriting
8458 *) Patches to make OpenSSL compile under Win32 again. Mostly
8459 work arounds for the VC++ problem that it treats func() as
8460 func(void). Also stripped out the parts of mkdef.pl that
8461 added extra typesafe functions: these no longer exist.
8464 *) Reorganisation of the stack code. The macros are now all
8465 collected in safestack.h . Each macro is defined in terms of
8466 a "stack macro" of the form SKM_<name>(type, a, b). The
8467 DEBUG_SAFESTACK is now handled in terms of function casts,
8468 this has the advantage of retaining type safety without the
8469 use of additional functions. If DEBUG_SAFESTACK is not defined
8470 then the non typesafe macros are used instead. Also modified the
8471 mkstack.pl script to handle the new form. Needs testing to see
8472 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8473 the default if no major problems. Similar behaviour for ASN1_SET_OF
8474 and PKCS12_STACK_OF.
8477 *) When some versions of IIS use the 'NET' form of private key the
8478 key derivation algorithm is different. Normally MD5(password) is
8479 used as a 128 bit RC4 key. In the modified case
8480 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8481 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8482 as the old Netscape_RSA functions except they have an additional
8483 'sgckey' parameter which uses the modified algorithm. Also added
8484 an -sgckey command line option to the rsa utility. Thanks to
8485 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8486 algorithm to openssl-dev.
8489 *) The evp_local.h macros were using 'c.##kname' which resulted in
8490 invalid expansion on some systems (SCO 5.0.5 for example).
8491 Corrected to 'c.kname'.
8492 [Phillip Porch <root@theporch.com>]
8494 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8495 a STACK of email addresses from a certificate or request, these look
8496 in the subject name and the subject alternative name extensions and
8497 omit any duplicate addresses.
8500 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8501 This makes DSA verification about 2 % faster.
8504 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8505 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8506 plus overhead for 1024 bit moduli).
8507 This makes exponentiations about 0.5 % faster for 1024 bit
8508 exponents (as measured by "openssl speed rsa2048").
8511 *) Rename memory handling macros to avoid conflicts with other
8513 Malloc => OPENSSL_malloc
8514 Malloc_locked => OPENSSL_malloc_locked
8515 Realloc => OPENSSL_realloc
8516 Free => OPENSSL_free
8519 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8520 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8523 *) CygWin32 support.
8524 [John Jarvie <jjarvie@newsguy.com>]
8526 *) The type-safe stack code has been rejigged. It is now only compiled
8527 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8528 by default all type-specific stack functions are "#define"d back to
8529 standard stack functions. This results in more streamlined output
8530 but retains the type-safety checking possibilities of the original
8534 *) The STACK code has been cleaned up, and certain type declarations
8535 that didn't make a lot of sense have been brought in line. This has
8536 also involved a cleanup of sorts in safestack.h to more correctly
8537 map type-safe stack functions onto their plain stack counterparts.
8538 This work has also resulted in a variety of "const"ifications of
8539 lots of the code, especially "_cmp" operations which should normally
8540 be prototyped with "const" parameters anyway.
8543 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8544 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8545 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8546 where all of 'md' is used each time the PRNG is used, but 'state'
8547 is used only indexed by a cyclic counter. As entropy may not be
8548 well distributed from the beginning, 'md' is important as a
8549 chaining variable. However, the output function chains only half
8550 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8551 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8552 in all of 'state' being rewritten, with the new values depending
8553 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8556 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8557 the handshake is continued after ssl_verify_cert_chain();
8558 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8559 can lead to 'unexplainable' connection aborts later.
8560 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8562 *) Major EVP API cipher revision.
8563 Add hooks for extra EVP features. This allows various cipher
8564 parameters to be set in the EVP interface. Support added for variable
8565 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8566 setting of RC2 and RC5 parameters.
8568 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8571 Remove lots of duplicated code from the EVP library. For example *every*
8572 cipher init() function handles the 'iv' in the same way according to the
8573 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8574 for CFB and OFB modes they zero ctx->num.
8576 New functionality allows removal of S/MIME code RC2 hack.
8578 Most of the routines have the same form and so can be declared in terms
8581 By shifting this to the top level EVP_CipherInit() it can be removed from
8582 all individual ciphers. If the cipher wants to handle IVs or keys
8583 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8586 Change lots of functions like EVP_EncryptUpdate() to now return a
8587 value: although software versions of the algorithms cannot fail
8588 any installed hardware versions can.
8591 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8592 this option is set, tolerate broken clients that send the negotiated
8593 protocol version number instead of the requested protocol version
8597 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8598 i.e. non-zero for export ciphersuites, zero otherwise.
8599 Previous versions had this flag inverted, inconsistent with
8600 rsa_tmp_cb (..._TMP_RSA_CB).
8601 [Bodo Moeller; problem reported by Amit Chopra]
8603 *) Add missing DSA library text string. Work around for some IIS
8604 key files with invalid SEQUENCE encoding.
8607 *) Add a document (doc/standards.txt) that list all kinds of standards
8608 and so on that are implemented in OpenSSL.
8611 *) Enhance c_rehash script. Old version would mishandle certificates
8612 with the same subject name hash and wouldn't handle CRLs at all.
8613 Added -fingerprint option to crl utility, to support new c_rehash
8617 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8620 *) Fix for SSL server purpose checking. Server checking was
8621 rejecting certificates which had extended key usage present
8622 but no ssl client purpose.
8623 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8625 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8626 is a little unclear about how a blank password is handled.
8627 Since the password in encoded as a BMPString with terminating
8628 double NULL a zero length password would end up as just the
8629 double NULL. However no password at all is different and is
8630 handled differently in the PKCS#12 key generation code. NS
8631 treats a blank password as zero length. MSIE treats it as no
8632 password on export: but it will try both on import. We now do
8633 the same: PKCS12_parse() tries zero length and no password if
8634 the password is set to "" or NULL (NULL is now a valid password:
8635 it wasn't before) as does the pkcs12 application.
8638 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8639 perror when PEM_read_bio_X509_REQ fails, the error message must
8640 be obtained from the error queue.
8643 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8644 it in ERR_remove_state if appropriate, and change ERR_get_state
8645 accordingly to avoid race conditions (this is necessary because
8646 thread_hash is no longer constant once set).
8649 *) Bugfix for linux-elf makefile.one.
8652 *) RSA_get_default_method() will now cause a default
8653 RSA_METHOD to be chosen if one doesn't exist already.
8654 Previously this was only set during a call to RSA_new()
8655 or RSA_new_method(NULL) meaning it was possible for
8656 RSA_get_default_method() to return NULL.
8659 *) Added native name translation to the existing DSO code
8660 that will convert (if the flag to do so is set) filenames
8661 that are sufficiently small and have no path information
8662 into a canonical native form. Eg. "blah" converted to
8663 "libblah.so" or "blah.dll" etc.
8666 *) New function ERR_error_string_n(e, buf, len) which is like
8667 ERR_error_string(e, buf), but writes at most 'len' bytes
8668 including the 0 terminator. For ERR_error_string_n, 'buf'
8670 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8672 *) CONF library reworked to become more general. A new CONF
8673 configuration file reader "class" is implemented as well as a
8674 new functions (NCONF_*, for "New CONF") to handle it. The now
8675 old CONF_* functions are still there, but are reimplemented to
8676 work in terms of the new functions. Also, a set of functions
8677 to handle the internal storage of the configuration data is
8678 provided to make it easier to write new configuration file
8679 reader "classes" (I can definitely see something reading a
8680 configuration file in XML format, for example), called _CONF_*,
8681 or "the configuration storage API"...
8683 The new configuration file reading functions are:
8685 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8686 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8688 NCONF_default, NCONF_WIN32
8690 NCONF_dump_fp, NCONF_dump_bio
8692 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8693 NCONF_new creates a new CONF object. This works in the same way
8694 as other interfaces in OpenSSL, like the BIO interface.
8695 NCONF_dump_* dump the internal storage of the configuration file,
8696 which is useful for debugging. All other functions take the same
8697 arguments as the old CONF_* functions wth the exception of the
8698 first that must be a `CONF *' instead of a `LHASH *'.
8700 To make it easer to use the new classes with the old CONF_* functions,
8701 the function CONF_set_default_method is provided.
8704 *) Add '-tls1' option to 'openssl ciphers', which was already
8705 mentioned in the documentation but had not been implemented.
8706 (This option is not yet really useful because even the additional
8707 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8710 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8711 OpenSSL-based applications) load shared libraries and bind to
8712 them in a portable way.
8713 [Geoff Thorpe, with contributions from Richard Levitte]
8715 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8717 *) Make sure _lrotl and _lrotr are only used with MSVC.
8719 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8720 (the default implementation of RAND_status).
8722 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8723 to '-clrext' (= clear extensions), as intended and documented.
8724 [Bodo Moeller; inconsistency pointed out by Michael Attili
8727 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8728 was larger than the MD block size.
8729 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8731 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8732 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8733 using the passed key: if the passed key was a private key the result
8734 of X509_print(), for example, would be to print out all the private key
8738 *) des_quad_cksum() byte order bug fix.
8739 [Ulf Möller, using the problem description in krb4-0.9.7, where
8740 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8742 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8744 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8746 *) For easily testing in shell scripts whether some command
8747 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8748 returns with exit code 0 iff no command of the given name is available.
8749 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8750 the output goes to stdout and nothing is printed to stderr.
8751 Additional arguments are always ignored.
8753 Since for each cipher there is a command of the same name,
8754 the 'no-cipher' compilation switches can be tested this way.
8756 ('openssl no-XXX' is not able to detect pseudo-commands such
8757 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8760 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8763 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8764 is set; it will be thrown away anyway because each handshake creates
8766 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8767 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8768 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8769 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8772 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8773 'Q' and 'R' lose their special meanings (quit/renegotiate).
8774 This is part of what -quiet does; unlike -quiet, -ign_eof
8775 does not suppress any output.
8778 *) Add compatibility options to the purpose and trust code. The
8779 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8780 accepts a certificate or CA, this was the previous behaviour,
8781 with all the associated security issues.
8783 X509_TRUST_COMPAT is the old trust behaviour: only and
8784 automatically trust self signed roots in certificate store. A
8785 new trust setting X509_TRUST_DEFAULT is used to specify that
8786 a purpose has no associated trust setting and it should instead
8787 use the value in the default purpose.
8790 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8791 and fix a memory leak.
8794 *) In util/mkerr.pl (which implements 'make errors'), preserve
8795 reason strings from the previous version of the .c file, as
8796 the default to have only downcase letters (and digits) in
8797 automatically generated reasons codes is not always appropriate.
8800 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8801 using strerror. Previously, ERR_reason_error_string() returned
8802 library names as reason strings for SYSerr; but SYSerr is a special
8803 case where small numbers are errno values, not library numbers.
8806 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8807 converts DSA parameters into DH parameters. (When creating parameters,
8808 DSA_generate_parameters is used.)
8811 *) Include 'length' (recommended exponent length) in C code generated
8812 by 'openssl dhparam -C'.
8815 *) The second argument to set_label in perlasm was already being used
8816 so couldn't be used as a "file scope" flag. Moved to third argument
8820 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8821 instead of RAND_bytes for encryption IVs and salts.
8824 *) Include RAND_status() into RAND_METHOD instead of implementing
8825 it only for md_rand.c Otherwise replacing the PRNG by calling
8826 RAND_set_rand_method would be impossible.
8829 *) Don't let DSA_generate_key() enter an infinite loop if the random
8830 number generation fails.
8833 *) New 'rand' application for creating pseudo-random output.
8836 *) Added configuration support for Linux/IA64
8837 [Rolf Haberrecker <rolf@suse.de>]
8839 *) Assembler module support for Mingw32.
8842 *) Shared library support for HPUX (in shlib/).
8843 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8845 *) Shared library support for Solaris gcc.
8846 [Lutz Behnke <behnke@trustcenter.de>]
8848 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8850 *) PKCS7_encrypt() was adding text MIME headers twice because they
8851 were added manually and by SMIME_crlf_copy().
8854 *) In bntest.c don't call BN_rand with zero bits argument.
8855 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8857 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8858 case was implemented. This caused BN_div_recp() to fail occasionally.
8861 *) Add an optional second argument to the set_label() in the perl
8862 assembly language builder. If this argument exists and is set
8863 to 1 it signals that the assembler should use a symbol whose
8864 scope is the entire file, not just the current function. This
8865 is needed with MASM which uses the format label:: for this scope.
8866 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8868 *) Change the ASN1 types so they are typedefs by default. Before
8869 almost all types were #define'd to ASN1_STRING which was causing
8870 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8874 *) Change names of new functions to the new get1/get0 naming
8875 convention: After 'get1', the caller owns a reference count
8876 and has to call ..._free; 'get0' returns a pointer to some
8877 data structure without incrementing reference counters.
8878 (Some of the existing 'get' functions increment a reference
8879 counter, some don't.)
8880 Similarly, 'set1' and 'add1' functions increase reference
8881 counters or duplicate objects.
8884 *) Allow for the possibility of temp RSA key generation failure:
8885 the code used to assume it always worked and crashed on failure.
8888 *) Fix potential buffer overrun problem in BIO_printf().
8889 [Ulf Möller, using public domain code by Patrick Powell; problem
8890 pointed out by David Sacerdote <das33@cornell.edu>]
8892 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8893 RAND_egd() and RAND_status(). In the command line application,
8894 the EGD socket can be specified like a seed file using RANDFILE
8898 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8899 Some CAs (e.g. Verisign) distribute certificates in this form.
8902 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8903 list to exclude them. This means that no special compilation option
8904 is needed to use anonymous DH: it just needs to be included in the
8908 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8909 EVP_MD_type. The old functionality is available in a new macro called
8910 EVP_MD_md(). Change code that uses it and update docs.
8913 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8914 where the 'void *' argument is replaced by a function pointer argument.
8915 Previously 'void *' was abused to point to functions, which works on
8916 many platforms, but is not correct. As these functions are usually
8917 called by macros defined in OpenSSL header files, most source code
8918 should work without changes.
8921 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8922 sections with information on -D... compiler switches used for
8923 compiling the library so that applications can see them. To enable
8924 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8925 must be defined. E.g.,
8926 #define OPENSSL_ALGORITHM_DEFINES
8927 #include <openssl/opensslconf.h>
8928 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8929 [Richard Levitte, Ulf and Bodo Möller]
8931 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8935 *) Change the 'other' type in certificate aux info to a STACK_OF
8936 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8937 the required ASN1 format: arbitrary types determined by an OID.
8940 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8941 argument to 'req'. This is not because the function is newer or
8942 better than others it just uses the work 'NEW' in the certificate
8943 request header lines. Some software needs this.
8946 *) Reorganise password command line arguments: now passwords can be
8947 obtained from various sources. Delete the PEM_cb function and make
8948 it the default behaviour: i.e. if the callback is NULL and the
8949 usrdata argument is not NULL interpret it as a null terminated pass
8950 phrase. If usrdata and the callback are NULL then the pass phrase
8951 is prompted for as usual.
8954 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8955 the support is automatically enabled. The resulting binaries will
8956 autodetect the card and use it if present.
8957 [Ben Laurie and Compaq Inc.]
8959 *) Work around for Netscape hang bug. This sends certificate request
8960 and server done in one record. Since this is perfectly legal in the
8961 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8962 the bugs/SSLv3 entry for more info.
8965 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8968 *) Add -rand argument to smime and pkcs12 applications and read/write
8972 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8975 *) Add command line password options to the remaining applications.
8978 *) Bug fix for BN_div_recp() for numerators with an even number of
8982 *) More tests in bntest.c, and changed test_bn output.
8985 *) ./config recognizes MacOS X now.
8988 *) Bug fix for BN_div() when the first words of num and divsor are
8989 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8992 *) Add support for various broken PKCS#8 formats, and command line
8993 options to produce them.
8996 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8997 get temporary BIGNUMs from a BN_CTX.
9000 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9004 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9005 include a #define from the old name to the new. The original intent
9006 was that statically linked binaries could for example just call
9007 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9008 link with digests. This never worked because SSLeay_add_all_digests()
9009 and SSLeay_add_all_ciphers() were in the same source file so calling
9010 one would link with the other. They are now in separate source files.
9013 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9016 *) Use a less unusual form of the Miller-Rabin primality test (it used
9017 a binary algorithm for exponentiation integrated into the Miller-Rabin
9018 loop, our standard modexp algorithms are faster).
9021 *) Support for the EBCDIC character set completed.
9022 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9024 *) Source code cleanups: use const where appropriate, eliminate casts,
9025 use void * instead of char * in lhash.
9028 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9029 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9030 this the server could overwrite ephemeral keys that the client
9034 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9035 using 50 iterations of the Rabin-Miller test.
9037 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9038 iterations of the Rabin-Miller test as required by the appendix
9039 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9040 As BN_is_prime_fasttest includes trial division, DSA parameter
9041 generation becomes much faster.
9043 This implies a change for the callback functions in DSA_is_prime
9044 and DSA_generate_parameters: The callback function is called once
9045 for each positive witness in the Rabin-Miller test, not just
9046 occasionally in the inner loop; and the parameters to the
9047 callback function now provide an iteration count for the outer
9048 loop rather than for the current invocation of the inner loop.
9049 DSA_generate_parameters additionally can call the callback
9050 function with an 'iteration count' of -1, meaning that a
9051 candidate has passed the trial division test (when q is generated
9052 from an application-provided seed, trial division is skipped).
9055 *) New function BN_is_prime_fasttest that optionally does trial
9056 division before starting the Rabin-Miller test and has
9057 an additional BN_CTX * argument (whereas BN_is_prime always
9058 has to allocate at least one BN_CTX).
9059 'callback(1, -1, cb_arg)' is called when a number has passed the
9060 trial division stage.
9063 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9067 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9070 *) New function BN_pseudo_rand().
9073 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9074 bignum version of BN_from_montgomery() with the working code from
9075 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9079 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9080 made it impossible to use the same SSL_SESSION data structure in
9081 SSL2 clients in multiple threads.
9084 *) The return value of RAND_load_file() no longer counts bytes obtained
9085 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9086 to seed the PRNG (previously an explicit byte count was required).
9087 [Ulf Möller, Bodo Möller]
9089 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9090 used (char *) instead of (void *) and had casts all over the place.
9093 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9096 *) Retain source code compatibility for BN_prime_checks macro:
9097 BN_is_prime(..., BN_prime_checks, ...) now uses
9098 BN_prime_checks_for_size to determine the appropriate number of
9099 Rabin-Miller iterations.
9102 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9103 DH_CHECK_P_NOT_SAFE_PRIME.
9104 (Check if this is true? OpenPGP calls them "strong".)
9107 *) Merge the functionality of "dh" and "gendh" programs into a new program
9108 "dhparam". The old programs are retained for now but will handle DH keys
9109 (instead of parameters) in future.
9112 *) Make the ciphers, s_server and s_client programs check the return values
9113 when a new cipher list is set.
9116 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9117 ciphers. Before when the 56bit ciphers were enabled the sorting was
9120 The syntax for the cipher sorting has been extended to support sorting by
9121 cipher-strength (using the strength_bits hard coded in the tables).
9122 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9124 Fix a bug in the cipher-command parser: when supplying a cipher command
9125 string with an "undefined" symbol (neither command nor alphanumeric
9126 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9127 an error is flagged.
9129 Due to the strength-sorting extension, the code of the
9130 ssl_create_cipher_list() function was completely rearranged. I hope that
9131 the readability was also increased :-)
9132 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9134 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9135 for the first serial number and places 2 in the serial number file. This
9136 avoids problems when the root CA is created with serial number zero and
9137 the first user certificate has the same issuer name and serial number
9141 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9142 the new code. Add documentation for this stuff.
9145 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9146 X509_*() to X509at_*() on the grounds that they don't handle X509
9147 structures and behave in an analogous way to the X509v3 functions:
9148 they shouldn't be called directly but wrapper functions should be used
9151 So we also now have some wrapper functions that call the X509at functions
9152 when passed certificate requests. (TO DO: similar things can be done with
9153 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9154 things. Some of these need some d2i or i2d and print functionality
9155 because they handle more complex structures.)
9158 *) Add missing #ifndefs that caused missing symbols when building libssl
9159 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9160 NO_RSA in ssl/s2*.c.
9161 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9163 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9164 has a return value which indicates the quality of the random data
9165 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9166 error queue. New function RAND_pseudo_bytes() generates output that is
9167 guaranteed to be unique but not unpredictable. RAND_add is like
9168 RAND_seed, but takes an extra argument for an entropy estimate
9169 (RAND_seed always assumes full entropy).
9172 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9173 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9174 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9175 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9176 false-positive rate of at most 2^-80 for random input.
9179 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9182 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9183 in the 0.9.5 release), this returns the chain
9184 from an X509_CTX structure with a dup of the stack and all
9185 the X509 reference counts upped: so the stack will exist
9186 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9189 Also make SSL_SESSION_print() print out the verify return
9193 *) Add manpage for the pkcs12 command. Also change the default
9194 behaviour so MAC iteration counts are used unless the new
9195 -nomaciter option is used. This improves file security and
9196 only older versions of MSIE (4.0 for example) need it.
9199 *) Honor the no-xxx Configure options when creating .DEF files.
9202 *) Add PKCS#10 attributes to field table: challengePassword,
9203 unstructuredName and unstructuredAddress. These are taken from
9204 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9205 international characters are used.
9207 More changes to X509_ATTRIBUTE code: allow the setting of types
9208 based on strings. Remove the 'loc' parameter when adding
9209 attributes because these will be a SET OF encoding which is sorted
9213 *) Initial changes to the 'req' utility to allow request generation
9214 automation. This will allow an application to just generate a template
9215 file containing all the field values and have req construct the
9218 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9219 used all over the place including certificate requests and PKCS#7
9220 structures. They are currently handled manually where necessary with
9221 some primitive wrappers for PKCS#7. The new functions behave in a
9222 manner analogous to the X509 extension functions: they allow
9223 attributes to be looked up by NID and added.
9225 Later something similar to the X509V3 code would be desirable to
9226 automatically handle the encoding, decoding and printing of the
9227 more complex types. The string types like challengePassword can
9228 be handled by the string table functions.
9230 Also modified the multi byte string table handling. Now there is
9231 a 'global mask' which masks out certain types. The table itself
9232 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9233 is useful when for example there is only one permissible type
9234 (as in countryName) and using the mask might result in no valid
9238 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9239 SSL_get_peer_finished to allow applications to obtain the latest
9240 Finished messages sent to the peer or expected from the peer,
9241 respectively. (SSL_get_peer_finished is usually the Finished message
9242 actually received from the peer, otherwise the protocol will be aborted.)
9244 As the Finished message are message digests of the complete handshake
9245 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9246 be used for external authentication procedures when the authentication
9247 provided by SSL/TLS is not desired or is not enough.
9250 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9251 the host supports BWX extension and if Compaq C is present on the
9252 $PATH. Just exploiting of the BWX extension results in 20-30%
9253 performance kick for some algorithms, e.g. DES and RC4 to mention
9254 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9258 *) Add support for MS "fast SGC". This is arguably a violation of the
9259 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9260 weak crypto and after checking the certificate is SGC a second one
9261 with strong crypto. MS SGC stops the first handshake after receiving
9262 the server certificate message and sends a second client hello. Since
9263 a server will typically do all the time consuming operations before
9264 expecting any further messages from the client (server key exchange
9265 is the most expensive) there is little difference between the two.
9267 To get OpenSSL to support MS SGC we have to permit a second client
9268 hello message after we have sent server done. In addition we have to
9269 reset the MAC if we do get this second client hello.
9272 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9273 if a DER encoded private key is RSA or DSA traditional format. Changed
9274 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9275 format DER encoded private key. Newer code should use PKCS#8 format which
9276 has the key type encoded in the ASN1 structure. Added DER private key
9277 support to pkcs8 application.
9280 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9281 ciphersuites has been selected (as required by the SSL 3/TLS 1
9282 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9283 is set, we interpret this as a request to violate the specification
9284 (the worst that can happen is a handshake failure, and 'correct'
9285 behaviour would result in a handshake failure anyway).
9288 *) In SSL_CTX_add_session, take into account that there might be multiple
9289 SSL_SESSION structures with the same session ID (e.g. when two threads
9290 concurrently obtain them from an external cache).
9291 The internal cache can handle only one SSL_SESSION with a given ID,
9292 so if there's a conflict, we now throw out the old one to achieve
9296 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9297 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9298 some routines that use cipher OIDs: some ciphers do not have OIDs
9299 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9303 *) Simplify the trust setting structure and code. Now we just have
9304 two sequences of OIDs for trusted and rejected settings. These will
9305 typically have values the same as the extended key usage extension
9306 and any application specific purposes.
9308 The trust checking code now has a default behaviour: it will just
9309 check for an object with the same NID as the passed id. Functions can
9310 be provided to override either the default behaviour or the behaviour
9311 for a given id. SSL client, server and email already have functions
9312 in place for compatibility: they check the NID and also return "trusted"
9313 if the certificate is self signed.
9316 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9317 traditional format into an EVP_PKEY structure.
9320 *) Add a password callback function PEM_cb() which either prompts for
9321 a password if usr_data is NULL or otherwise assumes it is a null
9322 terminated password. Allow passwords to be passed on command line
9323 environment or config files in a few more utilities.
9326 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9327 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9328 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9329 Update documentation.
9332 *) Support for ASN1 "NULL" type. This could be handled before by using
9333 ASN1_TYPE but there wasn't any function that would try to read a NULL
9334 and produce an error if it couldn't. For compatibility we also have
9335 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9336 don't allocate anything because they don't need to.
9339 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9341 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9343 *) Rebuild of the memory allocation routines used by OpenSSL code and
9344 possibly others as well. The purpose is to make an interface that
9345 provide hooks so anyone can build a separate set of allocation and
9346 deallocation routines to be used by OpenSSL, for example memory
9347 pool implementations, or something else, which was previously hard
9348 since Malloc(), Realloc() and Free() were defined as macros having
9349 the values malloc, realloc and free, respectively (except for Win32
9350 compilations). The same is provided for memory debugging code.
9351 OpenSSL already comes with functionality to find memory leaks, but
9352 this gives people a chance to debug other memory problems.
9354 With these changes, a new set of functions and macros have appeared:
9356 CRYPTO_set_mem_debug_functions() [F]
9357 CRYPTO_get_mem_debug_functions() [F]
9358 CRYPTO_dbg_set_options() [F]
9359 CRYPTO_dbg_get_options() [F]
9360 CRYPTO_malloc_debug_init() [M]
9362 The memory debug functions are NULL by default, unless the library
9363 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9364 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9365 gives the standard debugging functions that come with OpenSSL) or
9366 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9367 provided by the library user) must be used. When the standard
9368 debugging functions are used, CRYPTO_dbg_set_options can be used to
9369 request additional information:
9370 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9371 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9373 Also, things like CRYPTO_set_mem_functions will always give the
9374 expected result (the new set of functions is used for allocation
9375 and deallocation) at all times, regardless of platform and compiler
9378 To finish it up, some functions that were never use in any other
9379 way than through macros have a new API and new semantic:
9382 CRYPTO_dbg_realloc()
9385 All macros of value have retained their old syntax.
9386 [Richard Levitte and Bodo Moeller]
9388 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9389 ordering of SMIMECapabilities wasn't in "strength order" and there
9390 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9394 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9395 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9396 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9398 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9399 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9400 functionality to handle multipart/signed properly) and a utility
9401 called 'smime' to call all this stuff. This is based on code I
9402 originally wrote for Celo who have kindly allowed it to be
9403 included in OpenSSL.
9406 *) Add variants des_set_key_checked and des_set_key_unchecked of
9407 des_set_key (aka des_key_sched). Global variable des_check_key
9408 decides which of these is called by des_set_key; this way
9409 des_check_key behaves as it always did, but applications and
9410 the library itself, which was buggy for des_check_key == 1,
9411 have a cleaner way to pick the version they need.
9414 *) New function PKCS12_newpass() which changes the password of a
9418 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9419 dynamic mix. In both cases the ids can be used as an index into the
9420 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9421 functions so they accept a list of the field values and the
9422 application doesn't need to directly manipulate the X509_TRUST
9426 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9430 *) Modify the way the V3 extension code looks up extensions. This now
9431 works in a similar way to the object code: we have some "standard"
9432 extensions in a static table which is searched with OBJ_bsearch()
9433 and the application can add dynamic ones if needed. The file
9434 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9435 updated whenever a new extension is added to the core code and kept
9436 in ext_nid order. There is a simple program 'tabtest.c' which checks
9437 this. New extensions are not added too often so this file can readily
9438 be maintained manually.
9440 There are two big advantages in doing things this way. The extensions
9441 can be looked up immediately and no longer need to be "added" using
9442 X509V3_add_standard_extensions(): this function now does nothing.
9443 [Side note: I get *lots* of email saying the extension code doesn't
9444 work because people forget to call this function]
9445 Also no dynamic allocation is done unless new extensions are added:
9446 so if we don't add custom extensions there is no need to call
9447 X509V3_EXT_cleanup().
9450 *) Modify enc utility's salting as follows: make salting the default. Add a
9451 magic header, so unsalted files fail gracefully instead of just decrypting
9452 to garbage. This is because not salting is a big security hole, so people
9453 should be discouraged from doing it.
9456 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9457 digest to be passed on the command line but it only used this
9458 parameter when signing a certificate. Modified so all relevant
9459 operations are affected by the digest parameter including the
9460 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9461 DSA key was used because it didn't fix the digest.
9464 *) Initial certificate chain verify code. Currently tests the untrusted
9465 certificates for consistency with the verify purpose (which is set
9466 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9468 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9469 this is because it will reject chains with invalid extensions whereas
9470 every previous version of OpenSSL and SSLeay made no checks at all.
9472 Trust code: checks the root CA for the relevant trust settings. Trust
9473 settings have an initial value consistent with the verify purpose: e.g.
9474 if the verify purpose is for SSL client use it expects the CA to be
9475 trusted for SSL client use. However the default value can be changed to
9476 permit custom trust settings: one example of this would be to only trust
9477 certificates from a specific "secure" set of CAs.
9479 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9480 which should be used for version portability: especially since the
9481 verify structure is likely to change more often now.
9483 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9484 to set them. If not set then assume SSL clients will verify SSL servers
9487 Two new options to the verify program: -untrusted allows a set of
9488 untrusted certificates to be passed in and -purpose which sets the
9489 intended purpose of the certificate. If a purpose is set then the
9490 new chain verify code is used to check extension consistency.
9493 *) Support for the authority information access extension.
9496 *) Modify RSA and DSA PEM read routines to transparently handle
9497 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9498 public keys in a format compatible with certificate
9499 SubjectPublicKeyInfo structures. Unfortunately there were already
9500 functions called *_PublicKey_* which used various odd formats so
9501 these are retained for compatibility: however the DSA variants were
9502 never in a public release so they have been deleted. Changed dsa/rsa
9503 utilities to handle the new format: note no releases ever handled public
9504 keys so we should be OK.
9506 The primary motivation for this change is to avoid the same fiasco
9507 that dogs private keys: there are several incompatible private key
9508 formats some of which are standard and some OpenSSL specific and
9509 require various evil hacks to allow partial transparent handling and
9510 even then it doesn't work with DER formats. Given the option anything
9511 other than PKCS#8 should be dumped: but the other formats have to
9512 stay in the name of compatibility.
9514 With public keys and the benefit of hindsight one standard format
9515 is used which works with EVP_PKEY, RSA or DSA structures: though
9516 it clearly returns an error if you try to read the wrong kind of key.
9518 Added a -pubkey option to the 'x509' utility to output the public key.
9519 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9520 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9521 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9522 that do the same as the EVP_PKEY_assign_*() except they up the
9523 reference count of the added key (they don't "swallow" the
9527 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9528 CRLs would fail if the file contained no certificates or no CRLs:
9529 added a new function to read in both types and return the number
9530 read: this means that if none are read it will be an error. The
9531 DER versions of the certificate and CRL reader would always fail
9532 because it isn't possible to mix certificates and CRLs in DER format
9533 without choking one or the other routine. Changed this to just read
9534 a certificate: this is the best we can do. Also modified the code
9535 in apps/verify.c to take notice of return codes: it was previously
9536 attempting to read in certificates from NULL pointers and ignoring
9537 any errors: this is one reason why the cert and CRL reader seemed
9538 to work. It doesn't check return codes from the default certificate
9539 routines: these may well fail if the certificates aren't installed.
9542 *) Code to support otherName option in GeneralName.
9545 *) First update to verify code. Change the verify utility
9546 so it warns if it is passed a self signed certificate:
9547 for consistency with the normal behaviour. X509_verify
9548 has been modified to it will now verify a self signed
9549 certificate if *exactly* the same certificate appears
9550 in the store: it was previously impossible to trust a
9551 single self signed certificate. This means that:
9552 openssl verify ss.pem
9553 now gives a warning about a self signed certificate but
9554 openssl verify -CAfile ss.pem ss.pem
9558 *) For servers, store verify_result in SSL_SESSION data structure
9559 (and add it to external session representation).
9560 This is needed when client certificate verifications fails,
9561 but an application-provided verification callback (set by
9562 SSL_CTX_set_cert_verify_callback) allows accepting the session
9563 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9564 but returns 1): When the session is reused, we have to set
9565 ssl->verify_result to the appropriate error code to avoid
9567 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9569 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9570 case in PKCS7_dataInit() where the signed PKCS7 structure
9571 didn't contain any existing data because it was being created.
9572 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9574 *) Add a salt to the key derivation routines in enc.c. This
9575 forms the first 8 bytes of the encrypted file. Also add a
9576 -S option to allow a salt to be input on the command line.
9579 *) New function X509_cmp(). Oddly enough there wasn't a function
9580 to compare two certificates. We do this by working out the SHA1
9581 hash and comparing that. X509_cmp() will be needed by the trust
9585 *) SSL_get1_session() is like SSL_get_session(), but increments
9586 the reference count in the SSL_SESSION returned.
9587 [Geoff Thorpe <geoff@eu.c2.net>]
9589 *) Fix for 'req': it was adding a null to request attributes.
9590 Also change the X509_LOOKUP and X509_INFO code to handle
9591 certificate auxiliary information.
9594 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9598 *) Add the possibility to add extra information to the memory leak
9599 detecting output, to form tracebacks, showing from where each
9600 allocation was originated: CRYPTO_push_info("constant string") adds
9601 the string plus current file name and line number to a per-thread
9602 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9603 is like calling CYRPTO_pop_info() until the stack is empty.
9604 Also updated memory leak detection code to be multi-thread-safe.
9607 *) Add options -text and -noout to pkcs7 utility and delete the
9608 encryption options which never did anything. Update docs.
9611 *) Add options to some of the utilities to allow the pass phrase
9612 to be included on either the command line (not recommended on
9613 OSes like Unix) or read from the environment. Update the
9614 manpages and fix a few bugs.
9617 *) Add a few manpages for some of the openssl commands.
9620 *) Fix the -revoke option in ca. It was freeing up memory twice,
9621 leaking and not finding already revoked certificates.
9624 *) Extensive changes to support certificate auxiliary information.
9625 This involves the use of X509_CERT_AUX structure and X509_AUX
9626 functions. An X509_AUX function such as PEM_read_X509_AUX()
9627 can still read in a certificate file in the usual way but it
9628 will also read in any additional "auxiliary information". By
9629 doing things this way a fair degree of compatibility can be
9630 retained: existing certificates can have this information added
9631 using the new 'x509' options.
9633 Current auxiliary information includes an "alias" and some trust
9634 settings. The trust settings will ultimately be used in enhanced
9635 certificate chain verification routines: currently a certificate
9636 can only be trusted if it is self signed and then it is trusted
9640 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9641 The problem was that one of the replacement routines had not been working
9642 since SSLeay releases. For now the offending routine has been replaced
9643 with non-optimised assembler. Even so, this now gives around 95%
9644 performance improvement for 1024 bit RSA signs.
9647 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9648 handling. Most clients have the effective key size in bits equal to
9649 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9650 A few however don't do this and instead use the size of the decrypted key
9651 to determine the RC2 key length and the AlgorithmIdentifier to determine
9652 the effective key length. In this case the effective key length can still
9653 be 40 bits but the key length can be 168 bits for example. This is fixed
9654 by manually forcing an RC2 key into the EVP_PKEY structure because the
9655 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9656 the key length and effective key length are equal.
9659 *) Add a bunch of functions that should simplify the creation of
9660 X509_NAME structures. Now you should be able to do:
9661 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9662 and have it automatically work out the correct field type and fill in
9663 the structures. The more adventurous can try:
9664 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9665 and it will (hopefully) work out the correct multibyte encoding.
9668 *) Change the 'req' utility to use the new field handling and multibyte
9669 copy routines. Before the DN field creation was handled in an ad hoc
9670 way in req, ca, and x509 which was rather broken and didn't support
9671 BMPStrings or UTF8Strings. Since some software doesn't implement
9672 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9673 using the dirstring_type option. See the new comment in the default
9674 openssl.cnf for more info.
9677 *) Make crypto/rand/md_rand.c more robust:
9678 - Assure unique random numbers after fork().
9679 - Make sure that concurrent threads access the global counter and
9680 md serializably so that we never lose entropy in them
9681 or use exactly the same state in multiple threads.
9682 Access to the large state is not always serializable because
9683 the additional locking could be a performance killer, and
9684 md should be large enough anyway.
9687 *) New file apps/app_rand.c with commonly needed functionality
9688 for handling the random seed file.
9690 Use the random seed file in some applications that previously did not:
9692 dsaparam -genkey (which also ignored its '-rand' option),
9695 x509 (when signing).
9696 Except on systems with /dev/urandom, it is crucial to have a random
9697 seed file at least for key creation, DSA signing, and for DH exchanges;
9698 for RSA signatures we could do without one.
9700 gendh and gendsa (unlike genrsa) used to read only the first byte
9701 of each file listed in the '-rand' option. The function as previously
9702 found in genrsa is now in app_rand.c and is used by all programs
9703 that support '-rand'.
9706 *) In RAND_write_file, use mode 0600 for creating files;
9707 don't just chmod when it may be too late.
9710 *) Report an error from X509_STORE_load_locations
9711 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9714 *) New function ASN1_mbstring_copy() this copies a string in either
9715 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9716 into an ASN1_STRING type. A mask of permissible types is passed
9717 and it chooses the "minimal" type to use or an error if not type
9721 *) Add function equivalents to the various macros in asn1.h. The old
9722 macros are retained with an M_ prefix. Code inside the library can
9723 use the M_ macros. External code (including the openssl utility)
9724 should *NOT* in order to be "shared library friendly".
9727 *) Add various functions that can check a certificate's extensions
9728 to see if it usable for various purposes such as SSL client,
9729 server or S/MIME and CAs of these types. This is currently
9730 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9731 verification. Also added a -purpose flag to x509 utility to
9732 print out all the purposes.
9735 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9739 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9740 for, obtain and decode and extension and obtain its critical flag.
9741 This allows all the necessary extension code to be handled in a
9742 single function call.
9745 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9746 platforms. See crypto/rc4/rc4_enc.c for further details.
9749 *) New -noout option to asn1parse. This causes no output to be produced
9750 its main use is when combined with -strparse and -out to extract data
9751 from a file (which may not be in ASN.1 format).
9754 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9755 when producing the local key id.
9756 [Richard Levitte <levitte@stacken.kth.se>]
9758 *) New option -dhparam in s_server. This allows a DH parameter file to be
9759 stated explicitly. If it is not stated then it tries the first server
9760 certificate file. The previous behaviour hard coded the filename
9764 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9765 a public key to be input or output. For example:
9766 openssl rsa -in key.pem -pubout -out pubkey.pem
9767 Also added necessary DSA public key functions to handle this.
9770 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9771 in the message. This was handled by allowing
9772 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9773 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9775 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9776 to the end of the strings whereas this didn't. This would cause problems
9777 if strings read with d2i_ASN1_bytes() were later modified.
9778 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9780 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9781 data and it contains EOF it will end up returning an error. This is
9782 caused by input 46 bytes long. The cause is due to the way base64
9783 BIOs find the start of base64 encoded data. They do this by trying a
9784 trial decode on each line until they find one that works. When they
9785 do a flag is set and it starts again knowing it can pass all the
9786 data directly through the decoder. Unfortunately it doesn't reset
9787 the context it uses. This means that if EOF is reached an attempt
9788 is made to pass two EOFs through the context and this causes the
9789 resulting error. This can also cause other problems as well. As is
9790 usual with these problems it takes *ages* to find and the fix is
9791 trivial: move one line.
9792 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9794 *) Ugly workaround to get s_client and s_server working under Windows. The
9795 old code wouldn't work because it needed to select() on sockets and the
9796 tty (for keypresses and to see if data could be written). Win32 only
9797 supports select() on sockets so we select() with a 1s timeout on the
9798 sockets and then see if any characters are waiting to be read, if none
9799 are present then we retry, we also assume we can always write data to
9800 the tty. This isn't nice because the code then blocks until we've
9801 received a complete line of data and it is effectively polling the
9802 keyboard at 1s intervals: however it's quite a bit better than not
9803 working at all :-) A dedicated Windows application might handle this
9804 with an event loop for example.
9807 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9808 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9809 will be called when RSA_sign() and RSA_verify() are used. This is useful
9810 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9811 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9812 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9813 This necessitated the support of an extra signature type NID_md5_sha1
9814 for SSL signatures and modifications to the SSL library to use it instead
9815 of calling RSA_public_decrypt() and RSA_private_encrypt().
9818 *) Add new -verify -CAfile and -CApath options to the crl program, these
9819 will lookup a CRL issuers certificate and verify the signature in a
9820 similar way to the verify program. Tidy up the crl program so it
9821 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9822 less strict. It will now permit CRL extensions even if it is not
9823 a V2 CRL: this will allow it to tolerate some broken CRLs.
9826 *) Initialize all non-automatic variables each time one of the openssl
9827 sub-programs is started (this is necessary as they may be started
9828 multiple times from the "OpenSSL>" prompt).
9829 [Lennart Bang, Bodo Moeller]
9831 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9832 removing all other RSA functionality (this is what NO_RSA does). This
9833 is so (for example) those in the US can disable those operations covered
9834 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9838 *) Non-copying interface to BIO pairs.
9839 (still largely untested)
9842 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9843 ASCII string. This was handled independently in various places before.
9846 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9847 UTF8 strings a character at a time.
9850 *) Use client_version from client hello to select the protocol
9851 (s23_srvr.c) and for RSA client key exchange verification
9852 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9855 *) Add various utility functions to handle SPKACs, these were previously
9856 handled by poking round in the structure internals. Added new function
9857 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9858 print, verify and generate SPKACs. Based on an original idea from
9859 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9862 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9865 *) Allow the config file extension section to be overwritten on the
9866 command line. Based on an original idea from Massimiliano Pala
9867 <madwolf@comune.modena.it>. The new option is called -extensions
9868 and can be applied to ca, req and x509. Also -reqexts to override
9869 the request extensions in req and -crlexts to override the crl extensions
9873 *) Add new feature to the SPKAC handling in ca. Now you can include
9874 the same field multiple times by preceding it by "XXXX." for example:
9877 this is the same syntax as used in the req config file.
9880 *) Allow certificate extensions to be added to certificate requests. These
9881 are specified in a 'req_extensions' option of the req section of the
9882 config file. They can be printed out with the -text option to req but
9883 are otherwise ignored at present.
9886 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9887 data read consists of only the final block it would not decrypted because
9888 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9889 A misplaced 'break' also meant the decrypted final block might not be
9890 copied until the next read.
9893 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9894 a few extra parameters to the DH structure: these will be useful if
9895 for example we want the value of 'q' or implement X9.42 DH.
9898 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9899 provides hooks that allow the default DSA functions or functions on a
9900 "per key" basis to be replaced. This allows hardware acceleration and
9901 hardware key storage to be handled without major modification to the
9902 library. Also added low level modexp hooks and CRYPTO_EX structure and
9903 associated functions.
9906 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9907 as "read only": it can't be written to and the buffer it points to will
9908 not be freed. Reading from a read only BIO is much more efficient than
9909 a normal memory BIO. This was added because there are several times when
9910 an area of memory needs to be read from a BIO. The previous method was
9911 to create a memory BIO and write the data to it, this results in two
9912 copies of the data and an O(n^2) reading algorithm. There is a new
9913 function BIO_new_mem_buf() which creates a read only memory BIO from
9914 an area of memory. Also modified the PKCS#7 routines to use read only
9918 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9919 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9920 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9921 but a retry condition occured while trying to read the rest.
9924 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9925 NID_pkcs7_encrypted by default: this was wrong since this should almost
9926 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9927 the encrypted data type: this is a more sensible place to put it and it
9928 allows the PKCS#12 code to be tidied up that duplicated this
9932 *) Changed obj_dat.pl script so it takes its input and output files on
9933 the command line. This should avoid shell escape redirection problems
9937 *) Initial support for certificate extension requests, these are included
9938 in things like Xenroll certificate requests. Included functions to allow
9939 extensions to be obtained and added.
9942 *) -crlf option to s_client and s_server for sending newlines as
9943 CRLF (as required by many protocols).
9946 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9948 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9949 [Ralf S. Engelschall]
9951 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9952 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9954 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9958 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9959 DH parameters/keys (q is lost during that conversion, but the resulting
9960 DH parameters contain its length).
9962 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9963 much faster than DH_generate_parameters (which creates parameters
9964 where p = 2*q + 1), and also the smaller q makes DH computations
9965 much more efficient (160-bit exponentiation instead of 1024-bit
9966 exponentiation); so this provides a convenient way to support DHE
9967 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9968 utter importance to use
9969 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9971 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9972 when such DH parameters are used, because otherwise small subgroup
9973 attacks may become possible!
9976 *) Avoid memory leak in i2d_DHparams.
9979 *) Allow the -k option to be used more than once in the enc program:
9980 this allows the same encrypted message to be read by multiple recipients.
9983 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9984 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9985 it will always use the numerical form of the OID, even if it has a short
9989 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9990 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9991 otherwise bn_mod_exp was called. In the case of hardware keys for example
9992 no private key components need be present and it might store extra data
9993 in the RSA structure, which cannot be accessed from bn_mod_exp.
9994 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9995 private key operations.
9998 *) Added support for SPARC Linux.
10001 *) pem_password_cb function type incompatibly changed from
10002 typedef int pem_password_cb(char *buf, int size, int rwflag);
10004 ....(char *buf, int size, int rwflag, void *userdata);
10005 so that applications can pass data to their callbacks:
10006 The PEM[_ASN1]_{read,write}... functions and macros now take an
10007 additional void * argument, which is just handed through whenever
10008 the password callback is called.
10009 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10011 New function SSL_CTX_set_default_passwd_cb_userdata.
10013 Compatibility note: As many C implementations push function arguments
10014 onto the stack in reverse order, the new library version is likely to
10015 interoperate with programs that have been compiled with the old
10016 pem_password_cb definition (PEM_whatever takes some data that
10017 happens to be on the stack as its last argument, and the callback
10018 just ignores this garbage); but there is no guarantee whatsoever that
10021 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10022 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10023 problems not only on Windows, but also on some Unix platforms.
10024 To avoid problematic command lines, these definitions are now in an
10025 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10026 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10029 *) MIPS III/IV assembler module is reimplemented.
10032 *) More DES library cleanups: remove references to srand/rand and
10033 delete an unused file.
10036 *) Add support for the the free Netwide assembler (NASM) under Win32,
10037 since not many people have MASM (ml) and it can be hard to obtain.
10038 This is currently experimental but it seems to work OK and pass all
10039 the tests. Check out INSTALL.W32 for info.
10042 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10043 without temporary keys kept an extra copy of the server key,
10044 and connections with temporary keys did not free everything in case
10048 *) New function RSA_check_key and new openssl rsa option -check
10049 for verifying the consistency of RSA keys.
10050 [Ulf Moeller, Bodo Moeller]
10052 *) Various changes to make Win32 compile work:
10053 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10054 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10055 comparison" warnings.
10056 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10059 *) Add a debugging option to PKCS#5 v2 key generation function: when
10060 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10061 derived keys are printed to stderr.
10064 *) Copy the flags in ASN1_STRING_dup().
10065 [Roman E. Pavlov <pre@mo.msk.ru>]
10067 *) The x509 application mishandled signing requests containing DSA
10068 keys when the signing key was also DSA and the parameters didn't match.
10070 It was supposed to omit the parameters when they matched the signing key:
10071 the verifying software was then supposed to automatically use the CA's
10072 parameters if they were absent from the end user certificate.
10074 Omitting parameters is no longer recommended. The test was also
10075 the wrong way round! This was probably due to unusual behaviour in
10076 EVP_cmp_parameters() which returns 1 if the parameters match.
10077 This meant that parameters were omitted when they *didn't* match and
10078 the certificate was useless. Certificates signed with 'ca' didn't have
10080 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10082 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10083 The interface is as follows:
10084 Applications can use
10085 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10086 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10087 "off" is now the default.
10088 The library internally uses
10089 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10090 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10091 to disable memory-checking temporarily.
10093 Some inconsistent states that previously were possible (and were
10094 even the default) are now avoided.
10096 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10097 with each memory chunk allocated; this is occasionally more helpful
10098 than just having a counter.
10100 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10102 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10106 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10107 which largely parallels "options", but is for changing API behaviour,
10108 whereas "options" are about protocol behaviour.
10109 Initial "mode" flags are:
10111 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10112 a single record has been written.
10113 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10114 retries use the same buffer location.
10115 (But all of the contents must be
10119 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10122 *) Fix problems with no-hmac etc.
10123 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10125 *) New functions RSA_get_default_method(), RSA_set_method() and
10126 RSA_get_method(). These allows replacement of RSA_METHODs without having
10127 to mess around with the internals of an RSA structure.
10130 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10131 Also really enable memory leak checks in openssl.c and in some
10133 [Chad C. Mulligan, Bodo Moeller]
10135 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10136 up the length of negative integers. This has now been simplified to just
10137 store the length when it is first determined and use it later, rather
10138 than trying to keep track of where data is copied and updating it to
10140 [Steve Henson, reported by Brien Wheeler
10141 <bwheeler@authentica-security.com>]
10143 *) Add a new function PKCS7_signatureVerify. This allows the verification
10144 of a PKCS#7 signature but with the signing certificate passed to the
10145 function itself. This contrasts with PKCS7_dataVerify which assumes the
10146 certificate is present in the PKCS#7 structure. This isn't always the
10147 case: certificates can be omitted from a PKCS#7 structure and be
10148 distributed by "out of band" means (such as a certificate database).
10151 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10152 function prototypes in pem.h, also change util/mkdef.pl to add the
10153 necessary function names.
10156 *) mk1mf.pl (used by Windows builds) did not properly read the
10157 options set by Configure in the top level Makefile, and Configure
10158 was not even able to write more than one option correctly.
10159 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10162 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10163 file to be loaded from a BIO or FILE pointer. The BIO version will
10164 for example allow memory BIOs to contain config info.
10167 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10168 Whoever hopes to achieve shared-library compatibility across versions
10169 must use this, not the compile-time macro.
10170 (Exercise 0.9.4: Which is the minimum library version required by
10172 Note: All this applies only to multi-threaded programs, others don't
10176 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10177 through a BIO pair triggered the default case, i.e.
10178 SSLerr(...,SSL_R_UNKNOWN_STATE).
10181 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10182 can use the SSL library even if none of the specific BIOs is
10186 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10187 for the encoded length.
10188 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10190 *) Add initial documentation of the X509V3 functions.
10193 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10194 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10195 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10196 secure PKCS#8 private key format with a high iteration count.
10199 *) Fix determination of Perl interpreter: A perl or perl5
10200 _directory_ in $PATH was also accepted as the interpreter.
10201 [Ralf S. Engelschall]
10203 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10204 wrong with it but it was very old and did things like calling
10205 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10206 unusual formatting.
10209 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10210 to use the new extension code.
10213 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10214 with macros. This should make it easier to change their form, add extra
10215 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10219 *) Add to configuration table a new entry that can specify an alternative
10220 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10221 according to Mark Crispin <MRC@Panda.COM>.
10225 *) DES CBC did not update the IV. Weird.
10228 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10229 Changing the behaviour of the former might break existing programs --
10230 where IV updating is needed, des_ncbc_encrypt can be used.
10233 *) When bntest is run from "make test" it drives bc to check its
10234 calculations, as well as internally checking them. If an internal check
10235 fails, it needs to cause bc to give a non-zero result or make test carries
10236 on without noticing the failure. Fixed.
10239 *) DES library cleanups.
10242 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10243 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10244 ciphers. NOTE: although the key derivation function has been verified
10245 against some published test vectors it has not been extensively tested
10246 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10250 *) Instead of "mkdir -p", which is not fully portable, use new
10251 Perl script "util/mkdir-p.pl".
10254 *) Rewrite the way password based encryption (PBE) is handled. It used to
10255 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10256 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10257 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10258 the 'parameter' field of the AlgorithmIdentifier is passed to the
10259 underlying key generation function so it must do its own ASN1 parsing.
10260 This has also changed the EVP_PBE_CipherInit() function which now has a
10261 'parameter' argument instead of literal salt and iteration count values
10262 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10265 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10266 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10267 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10268 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10269 value was just used as a "magic string" and not used directly its
10270 value doesn't matter.
10273 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10277 *) "linux-sparc64" configuration (ultrapenguin).
10278 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10279 "linux-sparc" configuration.
10280 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10282 *) config now generates no-xxx options for missing ciphers.
10285 *) Support the EBCDIC character set (work in progress).
10286 File ebcdic.c not yet included because it has a different license.
10287 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10289 *) Support BS2000/OSD-POSIX.
10290 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10292 *) Make callbacks for key generation use void * instead of char *.
10295 *) Make S/MIME samples compile (not yet tested).
10298 *) Additional typesafe stacks.
10301 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10305 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10307 *) New configuration variant "sco5-gcc".
10309 *) Updated some demos.
10310 [Sean O Riordain, Wade Scholine]
10312 *) Add missing BIO_free at exit of pkcs12 application.
10315 *) Fix memory leak in conf.c.
10318 *) Updates for Win32 to assembler version of MD5.
10321 *) Set #! path to perl in apps/der_chop to where we found it
10322 instead of using a fixed path.
10325 *) SHA library changes for irix64-mips4-cc.
10328 *) Improvements for VMS support.
10332 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10334 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10335 This also avoids the problems with SC4.2 and unpatched SC5.
10336 [Andy Polyakov <appro@fy.chalmers.se>]
10338 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10339 These are required because of the typesafe stack would otherwise break
10340 existing code. If old code used a structure member which used to be STACK
10341 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10342 sk_num or sk_value it would produce an error because the num, data members
10343 are not present in STACK_OF. Now it just produces a warning. sk_set
10344 replaces the old method of assigning a value to sk_value
10345 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10346 that does this will no longer work (and should use sk_set instead) but
10347 this could be regarded as a "questionable" behaviour anyway.
10350 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10351 correctly handle encrypted S/MIME data.
10354 *) Change type of various DES function arguments from des_cblock
10355 (which means, in function argument declarations, pointer to char)
10356 to des_cblock * (meaning pointer to array with 8 char elements),
10357 which allows the compiler to do more typechecking; it was like
10358 that back in SSLeay, but with lots of ugly casts.
10360 Introduce new type const_des_cblock.
10363 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10364 problems: find RecipientInfo structure that matches recipient certificate
10365 and initialise the ASN1 structures properly based on passed cipher.
10368 *) Belatedly make the BN tests actually check the results.
10371 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10372 to and from BNs: it was completely broken. New compilation option
10373 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10374 key elements as negative integers.
10377 *) Reorganize and speed up MD5.
10378 [Andy Polyakov <appro@fy.chalmers.se>]
10381 [Richard Levitte <richard@levitte.org>]
10383 *) New option -out to asn1parse to allow the parsed structure to be
10384 output to a file. This is most useful when combined with the -strparse
10385 option to examine the output of things like OCTET STRINGS.
10388 *) Make SSL library a little more fool-proof by not requiring any longer
10389 that SSL_set_{accept,connect}_state be called before
10390 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10391 in many applications because usually everything *appeared* to work as
10392 intended anyway -- now it really works as intended).
10395 *) Move openssl.cnf out of lib/.
10398 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10399 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10400 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10401 [Ralf S. Engelschall]
10403 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10404 handle PKCS#7 enveloped data properly.
10405 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10407 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10408 copying pointers. The cert_st handling is changed by this in
10409 various ways (and thus what used to be known as ctx->default_cert
10410 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10411 any longer when s->cert does not give us what we need).
10412 ssl_cert_instantiate becomes obsolete by this change.
10413 As soon as we've got the new code right (possibly it already is?),
10414 we have solved a couple of bugs of the earlier code where s->cert
10415 was used as if it could not have been shared with other SSL structures.
10417 Note that using the SSL API in certain dirty ways now will result
10418 in different behaviour than observed with earlier library versions:
10419 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10420 does not influence s as it used to.
10422 In order to clean up things more thoroughly, inside SSL_SESSION
10423 we don't use CERT any longer, but a new structure SESS_CERT
10424 that holds per-session data (if available); currently, this is
10425 the peer's certificate chain and, for clients, the server's certificate
10426 and temporary key. CERT holds only those values that can have
10427 meaningful defaults in an SSL_CTX.
10430 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10431 from the internal representation. Various PKCS#7 fixes: remove some
10432 evil casts and set the enc_dig_alg field properly based on the signing
10436 *) Allow PKCS#12 password to be set from the command line or the
10437 environment. Let 'ca' get its config file name from the environment
10438 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10442 *) Allow certificate policies extension to use an IA5STRING for the
10443 organization field. This is contrary to the PKIX definition but
10444 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10448 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10449 without disallowing inline assembler and the like for non-pedantic builds.
10452 *) Support Borland C++ builder.
10453 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10455 *) Support Mingw32.
10458 *) SHA-1 cleanups and performance enhancements.
10459 [Andy Polyakov <appro@fy.chalmers.se>]
10461 *) Sparc v8plus assembler for the bignum library.
10462 [Andy Polyakov <appro@fy.chalmers.se>]
10464 *) Accept any -xxx and +xxx compiler options in Configure.
10467 *) Update HPUX configuration.
10470 *) Add missing sk_<type>_unshift() function to safestack.h
10471 [Ralf S. Engelschall]
10473 *) New function SSL_CTX_use_certificate_chain_file that sets the
10474 "extra_cert"s in addition to the certificate. (This makes sense
10475 only for "PEM" format files, as chains as a whole are not
10479 *) Support verify_depth from the SSL API.
10480 x509_vfy.c had what can be considered an off-by-one-error:
10481 Its depth (which was not part of the external interface)
10482 was actually counting the number of certificates in a chain;
10483 now it really counts the depth.
10486 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10487 instead of X509err, which often resulted in confusing error
10488 messages since the error codes are not globally unique
10489 (e.g. an alleged error in ssl3_accept when a certificate
10490 didn't match the private key).
10492 *) New function SSL_CTX_set_session_id_context that allows to set a default
10493 value (so that you don't need SSL_set_session_id_context for each
10494 connection using the SSL_CTX).
10497 *) OAEP decoding bug fix.
10500 *) Support INSTALL_PREFIX for package builders, as proposed by
10504 *) New Configure options "threads" and "no-threads". For systems
10505 where the proper compiler options are known (currently Solaris
10506 and Linux), "threads" is the default.
10509 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10512 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10513 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10514 such as /usr/local/bin.
10517 *) "make linux-shared" to build shared libraries.
10518 [Niels Poppe <niels@netbox.org>]
10520 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10523 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10524 extension adding in x509 utility.
10527 *) Remove NOPROTO sections and error code comments.
10530 *) Partial rewrite of the DEF file generator to now parse the ANSI
10534 *) New Configure options --prefix=DIR and --openssldir=DIR.
10537 *) Complete rewrite of the error code script(s). It is all now handled
10538 by one script at the top level which handles error code gathering,
10539 header rewriting and C source file generation. It should be much better
10540 than the old method: it now uses a modified version of Ulf's parser to
10541 read the ANSI prototypes in all header files (thus the old K&R definitions
10542 aren't needed for error creation any more) and do a better job of
10543 translating function codes into names. The old 'ASN1 error code imbedded
10544 in a comment' is no longer necessary and it doesn't use .err files which
10545 have now been deleted. Also the error code call doesn't have to appear all
10546 on one line (which resulted in some large lines...).
10549 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10552 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10553 0 (which usually indicates a closed connection), but continue reading.
10556 *) Fix some race conditions.
10559 *) Add support for CRL distribution points extension. Add Certificate
10560 Policies and CRL distribution points documentation.
10563 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10566 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10567 8 of keying material. Merlin has also confirmed interop with this fix
10568 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10569 [Merlin Hughes <merlin@baltimore.ie>]
10571 *) Fix lots of warnings.
10572 [Richard Levitte <levitte@stacken.kth.se>]
10574 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10575 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10576 [Richard Levitte <levitte@stacken.kth.se>]
10578 *) Fix problems with sizeof(long) == 8.
10579 [Andy Polyakov <appro@fy.chalmers.se>]
10581 *) Change functions to ANSI C.
10584 *) Fix typos in error codes.
10585 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10587 *) Remove defunct assembler files from Configure.
10590 *) SPARC v8 assembler BIGNUM implementation.
10591 [Andy Polyakov <appro@fy.chalmers.se>]
10593 *) Support for Certificate Policies extension: both print and set.
10594 Various additions to support the r2i method this uses.
10597 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10598 return a const string when you are expecting an allocated buffer.
10601 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10602 types DirectoryString and DisplayText.
10605 *) Add code to allow r2i extensions to access the configuration database,
10606 add an LHASH database driver and add several ctx helper functions.
10609 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10610 fail when they extended the size of a BIGNUM.
10613 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10614 support typesafe stack.
10617 *) Fix typo in SSL_[gs]et_options().
10618 [Nils Frostberg <nils@medcom.se>]
10620 *) Delete various functions and files that belonged to the (now obsolete)
10621 old X509V3 handling code.
10624 *) New Configure option "rsaref".
10627 *) Don't auto-generate pem.h.
10630 *) Introduce type-safe ASN.1 SETs.
10633 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10634 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10636 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10637 that links with OpenSSL (well at least cause lots of warnings), but fear
10638 not: the conversion is trivial, and it eliminates loads of evil casts. A
10639 few STACKed things have been converted already. Feel free to convert more.
10640 In the fullness of time, I'll do away with the STACK type altogether.
10643 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10644 specified in <certfile> by updating the entry in the index.txt file.
10645 This way one no longer has to edit the index.txt file manually for
10646 revoking a certificate. The -revoke option does the gory details now.
10647 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10649 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10650 `-text' option at all and this way the `-noout -text' combination was
10651 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10652 [Ralf S. Engelschall]
10654 *) Make sure a corresponding plain text error message exists for the
10655 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10656 verify callback function determined that a certificate was revoked.
10657 [Ralf S. Engelschall]
10659 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10660 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10661 all available cipers including rc5, which was forgotten until now.
10662 In order to let the testing shell script know which algorithms
10663 are available, a new (up to now undocumented) command
10664 "openssl list-cipher-commands" is used.
10667 *) Bugfix: s_client occasionally would sleep in select() when
10668 it should have checked SSL_pending() first.
10671 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10672 the raw DSA values prior to ASN.1 encoding.
10675 *) Tweaks to Configure
10676 [Niels Poppe <niels@netbox.org>]
10678 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10682 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10685 *) New config option to avoid instructions that are illegal on the 80386.
10686 The default code is faster, but requires at least a 486.
10689 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10690 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10691 same as SSL2_VERSION anyway.
10694 *) New "-showcerts" option for s_client.
10697 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10698 application. Various cleanups and fixes.
10701 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10702 modify error routines to work internally. Add error codes and PBE init
10703 to library startup routines.
10706 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10707 packing functions to asn1 and evp. Changed function names and error
10708 codes along the way.
10711 *) PKCS12 integration: and so it begins... First of several patches to
10712 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10713 objects to objects.h
10716 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10717 and display support for Thawte strong extranet extension.
10720 *) Add LinuxPPC support.
10721 [Jeff Dubrule <igor@pobox.org>]
10723 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10724 bn_div_words in alpha.s.
10725 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10727 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10728 OAEP isn't supported when OpenSSL is built with RSAref.
10729 [Ulf Moeller <ulf@fitug.de>]
10731 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10732 so they no longer are missing under -DNOPROTO.
10733 [Soren S. Jorvang <soren@t.dk>]
10736 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10738 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10739 doesn't work when the session is reused. Coming soon!
10742 *) Fix a security hole, that allows sessions to be reused in the wrong
10743 context thus bypassing client cert protection! All software that uses
10744 client certs and session caches in multiple contexts NEEDS PATCHING to
10745 allow session reuse! A fuller solution is in the works.
10746 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10748 *) Some more source tree cleanups (removed obsolete files
10749 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10750 permission on "config" script to be executable) and a fix for the INSTALL
10752 [Ulf Moeller <ulf@fitug.de>]
10754 *) Remove some legacy and erroneous uses of malloc, free instead of
10756 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10758 *) Make rsa_oaep_test return non-zero on error.
10759 [Ulf Moeller <ulf@fitug.de>]
10761 *) Add support for native Solaris shared libraries. Configure
10762 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10763 if someone would make that last step automatic.
10764 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10766 *) ctx_size was not built with the right compiler during "make links". Fixed.
10769 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10770 except NULL ciphers". This means the default cipher list will no longer
10771 enable NULL ciphers. They need to be specifically enabled e.g. with
10772 the string "DEFAULT:eNULL".
10775 *) Fix to RSA private encryption routines: if p < q then it would
10776 occasionally produce an invalid result. This will only happen with
10777 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10780 *) Be less restrictive and allow also `perl util/perlpath.pl
10781 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10782 because this way one can also use an interpreter named `perl5' (which is
10783 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10784 installed as `perl').
10785 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10787 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10788 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10790 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10791 advapi32.lib to Win32 build and change the pem test comparison
10792 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10793 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10794 and crypto/des/ede_cbcm_enc.c.
10797 *) DES quad checksum was broken on big-endian architectures. Fixed.
10800 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10801 Win32 test batch file so it (might) work again. The Win32 test batch file
10802 is horrible: I feel ill....
10805 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10806 in e_os.h. Audit of header files to check ANSI and non ANSI
10807 sections: 10 functions were absent from non ANSI section and not exported
10808 from Windows DLLs. Fixed up libeay.num for new functions.
10811 *) Make `openssl version' output lines consistent.
10812 [Ralf S. Engelschall]
10814 *) Fix Win32 symbol export lists for BIO functions: Added
10815 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10816 to ms/libeay{16,32}.def.
10817 [Ralf S. Engelschall]
10819 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10820 fine under Unix and passes some trivial tests I've now added. But the
10821 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10822 added to make sure no one expects that this stuff really works in the
10823 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10824 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10826 [Ralf S. Engelschall]
10828 *) Fix the generation of two part addresses in perl.
10829 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10831 *) Add config entry for Linux on MIPS.
10832 [John Tobey <jtobey@channel1.com>]
10834 *) Make links whenever Configure is run, unless we are on Windoze.
10837 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10838 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10842 *) Add a useful kludge to allow package maintainers to specify compiler and
10843 other platforms details on the command line without having to patch the
10844 Configure script everytime: One now can use ``perl Configure
10845 <id>:<details>'', i.e. platform ids are allowed to have details appended
10846 to them (separated by colons). This is treated as there would be a static
10847 pre-configured entry in Configure's %table under key <id> with value
10848 <details> and ``perl Configure <id>'' is called. So, when you want to
10849 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10850 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10851 now, which overrides the FreeBSD-elf entry on-the-fly.
10852 [Ralf S. Engelschall]
10854 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10857 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10858 on the `perl Configure ...' command line. This way one can compile
10859 OpenSSL libraries with Position Independent Code (PIC) which is needed
10860 for linking it into DSOs.
10861 [Ralf S. Engelschall]
10863 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10867 *) Cleaned up the LICENSE document: The official contact for any license
10868 questions now is the OpenSSL core team under openssl-core@openssl.org.
10869 And add a paragraph about the dual-license situation to make sure people
10870 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10871 to the OpenSSL toolkit.
10872 [Ralf S. Engelschall]
10874 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10875 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10876 Additionally cleaned up the `make links' target: Remove unnecessary
10877 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10878 to speed processing and no longer clutter the display with confusing
10879 stuff. Instead only the actually done links are displayed.
10880 [Ralf S. Engelschall]
10882 *) Permit null encryption ciphersuites, used for authentication only. It used
10883 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10884 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10888 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10889 signed attributes when verifying signatures (this would break them),
10890 the detached data encoding was wrong and public keys obtained using
10891 X509_get_pubkey() weren't freed.
10894 *) Add text documentation for the BUFFER functions. Also added a work around
10895 to a Win95 console bug. This was triggered by the password read stuff: the
10896 last character typed gets carried over to the next fread(). If you were
10897 generating a new cert request using 'req' for example then the last
10898 character of the passphrase would be CR which would then enter the first
10902 *) Added the new `Includes OpenSSL Cryptography Software' button as
10903 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10904 button and can be used by applications based on OpenSSL to show the
10905 relationship to the OpenSSL project.
10906 [Ralf S. Engelschall]
10908 *) Remove confusing variables in function signatures in files
10909 ssl/ssl_lib.c and ssl/ssl.h.
10910 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10912 *) Don't install bss_file.c under PREFIX/include/
10913 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10915 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10916 functions that return function pointers and has support for NT specific
10917 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10918 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10919 unsigned to signed types: this was killing the Win32 compile.
10922 *) Add new certificate file to stack functions,
10923 SSL_add_dir_cert_subjects_to_stack() and
10924 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10925 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10926 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10927 This means that Apache-SSL and similar packages don't have to mess around
10928 to add as many CAs as they want to the preferred list.
10931 *) Experiment with doxygen documentation. Currently only partially applied to
10933 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10934 openssl.doxy as the configuration file.
10937 *) Get rid of remaining C++-style comments which strict C compilers hate.
10938 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10940 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10941 compiled in by default: it has problems with large keys.
10944 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10945 DH private keys and/or callback functions which directly correspond to
10946 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10947 is needed for applications which have to configure certificates on a
10948 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10950 For the RSA certificate situation is makes no difference, but
10951 for the DSA certificate situation this fixes the "no shared cipher"
10952 problem where the OpenSSL cipher selection procedure failed because the
10953 temporary keys were not overtaken from the context and the API provided
10954 no way to reconfigure them.
10955 The new functions now let applications reconfigure the stuff and they
10956 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10957 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10958 non-public-API function ssl_cert_instantiate() is used as a helper
10959 function and also to reduce code redundancy inside ssl_rsa.c.
10960 [Ralf S. Engelschall]
10962 *) Move s_server -dcert and -dkey options out of the undocumented feature
10963 area because they are useful for the DSA situation and should be
10964 recognized by the users.
10965 [Ralf S. Engelschall]
10967 *) Fix the cipher decision scheme for export ciphers: the export bits are
10968 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10969 SSL_EXP_MASK. So, the original variable has to be used instead of the
10970 already masked variable.
10971 [Richard Levitte <levitte@stacken.kth.se>]
10973 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10974 [Richard Levitte <levitte@stacken.kth.se>]
10976 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10977 from `int' to `unsigned int' because it's a length and initialized by
10978 EVP_DigestFinal() which expects an `unsigned int *'.
10979 [Richard Levitte <levitte@stacken.kth.se>]
10981 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10982 script. Instead use the usual Shell->Perl transition trick.
10983 [Ralf S. Engelschall]
10985 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10986 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10987 -noout -modulus' as it's already the case for `openssl rsa -noout
10988 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10989 currently the public key is printed (a decision which was already done by
10990 `openssl dsa -modulus' in the past) which serves a similar purpose.
10991 Additionally the NO_RSA no longer completely removes the whole -modulus
10992 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10994 [Ralf S. Engelschall]
10996 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10997 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10998 [Arne Ansper <arne@ats.cyber.ee>]
11000 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11001 to be added. Now both 'req' and 'ca' can use new objects defined in the
11005 *) Add cool BIO that does syslog (or event log on NT).
11006 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11008 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11009 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11010 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11011 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11014 *) Add preliminary config info for new extension code.
11017 *) Make RSA_NO_PADDING really use no padding.
11018 [Ulf Moeller <ulf@fitug.de>]
11020 *) Generate errors when private/public key check is done.
11023 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11024 for some CRL extensions and new objects added.
11027 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11028 key usage extension and fuller support for authority key id.
11031 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11032 padding method for RSA, which is recommended for new applications in PKCS
11033 #1 v2.0 (RFC 2437, October 1998).
11034 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11035 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11036 against Bleichbacher's attack on RSA.
11037 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11040 *) Updates to the new SSL compression code
11041 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11043 *) Fix so that the version number in the master secret, when passed
11044 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11045 (because the server will not accept higher), that the version number
11046 is 0x03,0x01, not 0x03,0x00
11047 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11049 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11050 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11051 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11054 *) Support for RAW extensions where an arbitrary extension can be
11055 created by including its DER encoding. See apps/openssl.cnf for
11059 *) Make sure latest Perl versions don't interpret some generated C array
11060 code as Perl array code in the crypto/err/err_genc.pl script.
11061 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11063 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11064 not many people have the assembler. Various Win32 compilation fixes and
11065 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11066 build instructions.
11069 *) Modify configure script 'Configure' to automatically create crypto/date.h
11070 file under Win32 and also build pem.h from pem.org. New script
11071 util/mkfiles.pl to create the MINFO file on environments that can't do a
11072 'make files': perl util/mkfiles.pl >MINFO should work.
11075 *) Major rework of DES function declarations, in the pursuit of correctness
11076 and purity. As a result, many evil casts evaporated, and some weirdness,
11077 too. You may find this causes warnings in your code. Zapping your evil
11078 casts will probably fix them. Mostly.
11081 *) Fix for a typo in asn1.h. Bug fix to object creation script
11082 obj_dat.pl. It considered a zero in an object definition to mean
11083 "end of object": none of the objects in objects.h have any zeros
11084 so it wasn't spotted.
11085 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11087 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11088 Masking (CBCM). In the absence of test vectors, the best I have been able
11089 to do is check that the decrypt undoes the encrypt, so far. Send me test
11090 vectors if you have them.
11093 *) Correct calculation of key length for export ciphers (too much space was
11094 allocated for null ciphers). This has not been tested!
11097 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11098 message is now correct (it understands "crypto" and "ssl" on its
11099 command line). There is also now an "update" option. This will update
11100 the util/ssleay.num and util/libeay.num files with any new functions.
11102 perl util/mkdef.pl crypto ssl update
11103 it will update them.
11106 *) Overhauled the Perl interface (perl/*):
11107 - ported BN stuff to OpenSSL's different BN library
11108 - made the perl/ source tree CVS-aware
11109 - renamed the package from SSLeay to OpenSSL (the files still contain
11110 their history because I've copied them in the repository)
11111 - removed obsolete files (the test scripts will be replaced
11112 by better Test::Harness variants in the future)
11113 [Ralf S. Engelschall]
11115 *) First cut for a very conservative source tree cleanup:
11116 1. merge various obsolete readme texts into doc/ssleay.txt
11117 where we collect the old documents and readme texts.
11118 2. remove the first part of files where I'm already sure that we no
11119 longer need them because of three reasons: either they are just temporary
11120 files which were left by Eric or they are preserved original files where
11121 I've verified that the diff is also available in the CVS via "cvs diff
11122 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11123 the crypto/md/ stuff).
11124 [Ralf S. Engelschall]
11126 *) More extension code. Incomplete support for subject and issuer alt
11127 name, issuer and authority key id. Change the i2v function parameters
11128 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11129 what that's for :-) Fix to ASN1 macro which messed up
11130 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11133 *) Preliminary support for ENUMERATED type. This is largely copied from the
11137 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11138 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11140 *) Make sure `make rehash' target really finds the `openssl' program.
11141 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11143 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11144 like to hear about it if this slows down other processors.
11147 *) Add CygWin32 platform information to Configure script.
11148 [Alan Batie <batie@aahz.jf.intel.com>]
11150 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11151 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11153 *) New program nseq to manipulate netscape certificate sequences
11156 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11160 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11161 but the BN code had some problems that would cause failures when
11162 doing certificate verification and some other functions.
11163 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11165 *) Add ASN1 and PEM code to support netscape certificate sequences.
11168 *) Add ASN1 and PEM code to support netscape certificate sequences.
11171 *) Add several PKIX and private extended key usage OIDs.
11174 *) Modify the 'ca' program to handle the new extension code. Modify
11175 openssl.cnf for new extension format, add comments.
11178 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11179 and add a sample to openssl.cnf so req -x509 now adds appropriate
11183 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11184 error code, add initial support to X509_print() and x509 application.
11187 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11188 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11189 stuff is currently isolated and isn't even compiled yet.
11192 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11193 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11194 Removed the versions check from X509 routines when loading extensions:
11195 this allows certain broken certificates that don't set the version
11196 properly to be processed.
11199 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11200 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11201 can still be regenerated with "make depend".
11204 *) Spelling mistake in C version of CAST-128.
11205 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11207 *) Changes to the error generation code. The perl script err-code.pl
11208 now reads in the old error codes and retains the old numbers, only
11209 adding new ones if necessary. It also only changes the .err files if new
11210 codes are added. The makefiles have been modified to only insert errors
11211 when needed (to avoid needlessly modifying header files). This is done
11212 by only inserting errors if the .err file is newer than the auto generated
11213 C file. To rebuild all the error codes from scratch (the old behaviour)
11214 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11215 or delete all the .err files.
11218 *) CAST-128 was incorrectly implemented for short keys. The C version has
11219 been fixed, but is untested. The assembler versions are also fixed, but
11220 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11221 to regenerate it if needed.
11222 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11223 Hagino <itojun@kame.net>]
11225 *) File was opened incorrectly in randfile.c.
11226 [Ulf Möller <ulf@fitug.de>]
11228 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11229 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11230 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11231 al: it's just almost always a UTCTime. Note this patch adds new error
11232 codes so do a "make errors" if there are problems.
11235 *) Correct Linux 1 recognition in config.
11236 [Ulf Möller <ulf@fitug.de>]
11238 *) Remove pointless MD5 hash when using DSA keys in ca.
11239 [Anonymous <nobody@replay.com>]
11241 *) Generate an error if given an empty string as a cert directory. Also
11242 generate an error if handed NULL (previously returned 0 to indicate an
11243 error, but didn't set one).
11244 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11246 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11249 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11250 parameters. This was causing a warning which killed off the Win32 compile.
11253 *) Remove C++ style comments from crypto/bn/bn_local.h.
11254 [Neil Costigan <neil.costigan@celocom.com>]
11256 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11257 based on a text string, looking up short and long names and finally
11258 "dot" format. The "dot" format stuff didn't work. Added new function
11259 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11260 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11261 OID is not part of the table.
11264 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11265 X509_LOOKUP_by_alias().
11268 *) Sort openssl functions by name.
11271 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11272 encryption from sample DSA keys (in case anyone is interested the password
11276 *) Make _all_ *_free functions accept a NULL pointer.
11277 [Frans Heymans <fheymans@isaserver.be>]
11279 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11281 [Anonymous <nobody@replay.com>]
11283 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11284 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11286 *) Don't blow it for numeric -newkey arguments to apps/req.
11287 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11289 *) Temp key "for export" tests were wrong in s3_srvr.c.
11290 [Anonymous <nobody@replay.com>]
11292 *) Add prototype for temp key callback functions
11293 SSL_CTX_set_tmp_{rsa,dh}_callback().
11296 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11297 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11300 *) X509_name_add_entry() freed the wrong thing after an error.
11301 [Arne Ansper <arne@ats.cyber.ee>]
11303 *) rsa_eay.c would attempt to free a NULL context.
11304 [Arne Ansper <arne@ats.cyber.ee>]
11306 *) BIO_s_socket() had a broken should_retry() on Windoze.
11307 [Arne Ansper <arne@ats.cyber.ee>]
11309 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11310 [Arne Ansper <arne@ats.cyber.ee>]
11312 *) Make sure the already existing X509_STORE->depth variable is initialized
11313 in X509_STORE_new(), but document the fact that this variable is still
11314 unused in the certificate verification process.
11315 [Ralf S. Engelschall]
11317 *) Fix the various library and apps files to free up pkeys obtained from
11318 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11321 *) Fix reference counting in X509_PUBKEY_get(). This makes
11322 demos/maurice/example2.c work, amongst others, probably.
11323 [Steve Henson and Ben Laurie]
11325 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11326 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11327 are no longer created. This way we have a single and consistent command
11328 line interface `openssl <command>', similar to `cvs <command>'.
11329 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11331 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11332 BIT STRING wrapper always have zero unused bits.
11335 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11338 *) Make the top-level INSTALL documentation easier to understand.
11341 *) Makefiles updated to exit if an error occurs in a sub-directory
11342 make (including if user presses ^C) [Paul Sutton]
11344 *) Make Montgomery context stuff explicit in RSA data structure.
11347 *) Fix build order of pem and err to allow for generated pem.h.
11350 *) Fix renumbering bug in X509_NAME_delete_entry().
11353 *) Enhanced the err-ins.pl script so it makes the error library number
11354 global and can add a library name. This is needed for external ASN1 and
11355 other error libraries.
11358 *) Fixed sk_insert which never worked properly.
11361 *) Fix ASN1 macros so they can handle indefinite length constructed
11362 EXPLICIT tags. Some non standard certificates use these: they can now
11366 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11367 into a single doc/ssleay.txt bundle. This way the information is still
11368 preserved but no longer messes up this directory. Now it's new room for
11369 the new set of documentation files.
11370 [Ralf S. Engelschall]
11372 *) SETs were incorrectly DER encoded. This was a major pain, because they
11373 shared code with SEQUENCEs, which aren't coded the same. This means that
11374 almost everything to do with SETs or SEQUENCEs has either changed name or
11375 number of arguments.
11376 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11378 *) Fix test data to work with the above.
11381 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11382 was already fixed by Eric for 0.9.1 it seems.
11383 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11385 *) Autodetect FreeBSD3.
11388 *) Fix various bugs in Configure. This affects the following platforms:
11392 unixware-2.0-pentium
11396 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11397 before they are needed.
11400 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11404 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11406 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11407 changed SSLeay to OpenSSL in version strings.
11408 [Ralf S. Engelschall]
11410 *) Some fixups to the top-level documents.
11413 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11414 because the symlink to include/ was missing.
11415 [Ralf S. Engelschall]
11417 *) Incorporated the popular no-RSA/DSA-only patches
11418 which allow to compile a RSA-free SSLeay.
11419 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11421 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11422 when "ssleay" is still not found.
11423 [Ralf S. Engelschall]
11425 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11426 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11428 *) Updated the README file.
11429 [Ralf S. Engelschall]
11431 *) Added various .cvsignore files in the CVS repository subdirs
11432 to make a "cvs update" really silent.
11433 [Ralf S. Engelschall]
11435 *) Recompiled the error-definition header files and added
11436 missing symbols to the Win32 linker tables.
11437 [Ralf S. Engelschall]
11439 *) Cleaned up the top-level documents;
11440 o new files: CHANGES and LICENSE
11441 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11442 o merged COPYRIGHT into LICENSE
11443 o removed obsolete TODO file
11444 o renamed MICROSOFT to INSTALL.W32
11445 [Ralf S. Engelschall]
11447 *) Removed dummy files from the 0.9.1b source tree:
11448 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11449 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11450 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11451 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11452 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11453 [Ralf S. Engelschall]
11455 *) Added various platform portability fixes.
11458 *) The Genesis of the OpenSSL rpject:
11459 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11460 Young and Tim J. Hudson created while they were working for C2Net until
11462 [The OpenSSL Project]
11465 Changes between 0.9.0b and 0.9.1b [not released]
11467 *) Updated a few CA certificates under certs/
11470 *) Changed some BIGNUM api stuff.
11473 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11474 DGUX x86, Linux Alpha, etc.
11477 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11478 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11482 *) Add -strparse option to asn1pars program which parses nested
11484 [Dr Stephen Henson <shenson@bigfoot.com>]
11486 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11489 *) DSA fix for "ca" program.
11492 *) Added "-genkey" option to "dsaparam" program.
11495 *) Added RIPE MD160 (rmd160) message digest.
11498 *) Added -a (all) option to "ssleay version" command.
11501 *) Added PLATFORM define which is the id given to Configure.
11504 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11507 *) Extended the ASN.1 parser routines.
11510 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11513 *) Added a BN_CTX to the BN library.
11516 *) Fixed the weak key values in DES library
11519 *) Changed API in EVP library for cipher aliases.
11522 *) Added support for RC2/64bit cipher.
11525 *) Converted the lhash library to the crypto/mem.c functions.
11528 *) Added more recognized ASN.1 object ids.
11531 *) Added more RSA padding checks for SSL/TLS.
11534 *) Added BIO proxy/filter functionality.
11537 *) Added extra_certs to SSL_CTX which can be used
11538 send extra CA certificates to the client in the CA cert chain sending
11539 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11542 *) Now Fortezza is denied in the authentication phase because
11543 this is key exchange mechanism is not supported by SSLeay at all.
11546 *) Additional PKCS1 checks.
11549 *) Support the string "TLSv1" for all TLS v1 ciphers.
11552 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11553 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11556 *) Fixed a few memory leaks.
11559 *) Fixed various code and comment typos.
11562 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11563 bytes sent in the client random.
11564 [Edward Bishop <ebishop@spyglass.com>]