5 Changes between 1.0.2e and 1.1.0 [xx XXX xxxx]
7 *) The distribution now has Makefile.in files, which are used to
8 create Makefile's when Configure is run. *Configure must be run
9 before trying to build now.*
12 *) The return value for SSL_CIPHER_description() for error conditions
16 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
18 Obtaining and performing DNSSEC validation of TLSA records is
19 the application's responsibility. The application provides
20 the TLSA records of its choice to OpenSSL, and these are then
21 used to authenticate the peer.
23 The TLSA records need not even come from DNS. They can, for
24 example, be used to implement local end-entity certificate or
25 trust-anchor "pinning", where the "pin" data takes the form
26 of TLSA records, which can augment or replace verification
27 based on the usual WebPKI public certification authorities.
30 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
31 continues to support deprecated interfaces in default builds.
32 However, applications are strongly advised to compile their
33 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
34 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
35 or the 1.1.0 releases.
37 In environments in which all applications have been ported to
38 not use any deprecated interfaces OpenSSL's Configure script
39 should be used with the --api=1.1.0 option to entirely remove
40 support for the deprecated features from the library and
41 unconditionally disable them in the installed headers.
42 Essentially the same effect can be achieved with the "no-deprecated"
43 argument to Configure, except that this will always restrict
44 the build to just the latest API, rather than a fixed API
47 As applications are ported to future revisions of the API,
48 they should update their compile-time OPENSSL_API_COMPAT define
49 accordingly, but in most cases should be able to continue to
50 compile with later releases.
52 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
53 0x10000000L and 0x00908000L, respectively. However those
54 versions did not support the OPENSSL_API_COMPAT feature, and
55 so applications are not typically tested for explicit support
56 of just the undeprecated features of either release.
59 *) Add support for setting the minimum and maximum supported protocol.
60 It can bet set via the SSL_set_min_proto_version() and
61 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
62 MaxProtcol. It's recommended to use the new APIs to disable
63 protocols instead of disabling individual protocols using
64 SSL_set_options() or SSL_CONF's Protocol. This change also
65 removes support for disabling TLS 1.2 in the OpenSSL TLS
66 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
69 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
72 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
73 and integrates ECDSA and ECDH functionality into EC. Implementations can
74 now redirect key generation and no longer need to convert to or from
77 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
78 include the ec.h header file instead.
81 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
82 ciphers who are no longer supported and drops support the ephemeral RSA key
83 exchange. The LOW ciphers currently doesn't have any ciphers in it.
86 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
87 opaque. For HMAC_CTX, the following constructors and destructors
90 HMAC_CTX *HMAC_CTX_new(void);
91 void HMAC_CTX_free(HMAC_CTX *ctx);
93 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
94 destroy such methods has been added. See EVP_MD_meth_new(3) and
95 EVP_CIPHER_meth_new(3) for documentation.
98 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
99 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
100 EVP_MD_CTX_reset() should be called instead to reinitialise
101 an already created structure.
102 2) For consistency with the majority of our object creators and
103 destructors, EVP_MD_CTX_(create|destroy) were renamed to
104 EVP_MD_CTX_(new|free). The old names are retained as macros
105 for deprecated builds.
108 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
109 cryptographic operations to be performed asynchronously as long as an
110 asynchronous capable engine is used. See the ASYNC_start_job() man page for
111 further details. Libssl has also had this capability integrated with the
112 introduction of the new mode SSL_MODE_ASYNC and associated error
113 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
114 pages. This work was developed in partnership with Intel Corp.
117 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
118 always enabled now. If you want to disable the support you should
119 exclude it using the list of supported ciphers. This also means that the
120 "-no_ecdhe" option has been removed from s_server.
123 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
124 SSL_{CTX_}set1_curves() which can set a list.
127 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
128 curve you want to support using SSL_{CTX_}set1_curves().
131 *) State machine rewrite. The state machine code has been significantly
132 refactored in order to remove much duplication of code and solve issues
133 with the old code (see ssl/statem/README for further details). This change
134 does have some associated API changes. Notably the SSL_state() function
135 has been removed and replaced by SSL_get_state which now returns an
136 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
137 altogether. The previous handshake states defined in ssl.h and ssl3.h have
141 *) All instances of the string "ssleay" in the public API were replaced
142 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
143 Some error codes related to internal RSA_eay API's were renamed.
146 *) The demo files in crypto/threads were moved to demo/threads.
149 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
151 [Matt Caswell, Rich Salz]
153 *) New ASN.1 embed macro.
155 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
156 structure is not allocated: it is part of the parent. That is instead of
164 This reduces memory fragmentation and make it impossible to accidentally
165 set a mandatory field to NULL.
167 This currently only works for some fields specifically a SEQUENCE, CHOICE,
168 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
169 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
173 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
176 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
177 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
178 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
179 DES and RC4 ciphersuites.
182 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
183 This changes the decoding behaviour for some invalid messages,
184 though the change is mostly in the more lenient direction, and
185 legacy behaviour is preserved as much as possible.
188 *) Fix no-stdio build.
189 [ David Woodhouse <David.Woodhouse@intel.com> and also
190 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
192 *) New testing framework
193 The testing framework has been largely rewritten and is now using
194 perl and the perl modules Test::Harness and an extended variant of
195 Test::More called OpenSSL::Test to do its work. All test scripts in
196 test/ have been rewritten into test recipes, and all direct calls to
197 executables in test/Makefile have become individual recipes using the
198 simplified testing OpenSSL::Test::Simple.
200 For documentation on our testing modules, do:
202 perldoc test/testlib/OpenSSL/Test/Simple.pm
203 perldoc test/testlib/OpenSSL/Test.pm
207 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
208 are used; the latter aborts on memory leaks (usually checked on exit).
209 Some undocumented "set malloc, etc., hooks" functions were removed
210 and others were changed. All are now documented.
213 *) In DSA_generate_parameters_ex, if the provided seed is too short,
215 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
217 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
218 from RFC4279, RFC4785, RFC5487, RFC5489.
220 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
221 original RSA_PSK patch.
224 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
225 era flag was never set throughout the codebase (only read). Also removed
226 SSL3_FLAGS_POP_BUFFER which was only used if
227 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
230 *) Changed the default name options in the "ca", "crl", "req" and "x509"
231 to be "oneline" instead of "compat".
234 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
235 not aware of clients that still exhibit this bug, and the workaround
236 hasn't been working properly for a while.
239 *) The return type of BIO_number_read() and BIO_number_written() as well as
240 the corresponding num_read and num_write members in the BIO structure has
241 changed from unsigned long to uint64_t. On platforms where an unsigned
242 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
246 *) Given the pervasive nature of TLS extensions it is inadvisable to run
247 OpenSSL without support for them. It also means that maintaining
248 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
249 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
252 *) Removed support for the two export grade static DH ciphersuites
253 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
254 were newly added (along with a number of other static DH ciphersuites) to
255 1.0.2. However the two export ones have *never* worked since they were
256 introduced. It seems strange in any case to be adding new export
257 ciphersuites, and given "logjam" it also does not seem correct to fix them.
260 *) Version negotiation has been rewritten. In particular SSLv23_method(),
261 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
262 and turned into macros which simply call the new preferred function names
263 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
264 should use the new names instead. Also as part of this change the ssl23.h
265 header file has been removed.
268 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
269 code and the associated standard is no longer considered fit-for-purpose.
272 *) RT2547 was closed. When generating a private key, try to make the
273 output file readable only by the owner. This behavior change might
274 be noticeable when interacting with other software.
276 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
280 *) Added HTTP GET support to the ocsp command.
283 *) Changed default digest for the dgst and enc commands from MD5 to
287 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
290 *) Added support for TLS extended master secret from
291 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
292 initial patch which was a great help during development.
295 *) All libssl internal structures have been removed from the public header
296 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
297 now redundant). Users should not attempt to access internal structures
298 directly. Instead they should use the provided API functions.
301 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
302 Access to deprecated functions can be re-enabled by running config with
303 "enable-deprecated". In addition applications wishing to use deprecated
304 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
305 will, by default, disable some transitive includes that previously existed
306 in the header files (e.g. ec.h will no longer, by default, include bn.h)
309 *) Added support for OCB mode. OpenSSL has been granted a patent license
310 compatible with the OpenSSL license for use of OCB. Details are available
311 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
312 for OCB can be removed by calling config with no-ocb.
315 *) SSLv2 support has been removed. It still supports receiving a SSLv2
316 compatible client hello.
319 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
320 done while fixing the error code for the key-too-small case.
321 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
323 *) CA.sh has been removmed; use CA.pl instead.
326 *) Removed old DES API.
329 *) Remove various unsupported platforms:
335 Sinix/ReliantUNIX RM400
340 16-bit platforms such as WIN16
343 *) Clean up OPENSSL_NO_xxx #define's
344 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
345 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
346 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
347 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
348 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
349 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
350 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
351 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
352 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
353 Remove MS_STATIC; it's a relic from platforms <32 bits.
356 *) Cleaned up dead code
357 Remove all but one '#ifdef undef' which is to be looked at.
360 *) Clean up calling of xxx_free routines.
361 Just like free(), fix most of the xxx_free routines to accept
362 NULL. Remove the non-null checks from callers. Save much code.
365 *) Add secure heap for storage of private keys (when possible).
366 Add BIO_s_secmem(), CBIGNUM, etc.
367 Contributed by Akamai Technologies under our Corporate CLA.
370 *) Experimental support for a new, fast, unbiased prime candidate generator,
371 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
372 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
374 *) New output format NSS in the sess_id command line tool. This allows
375 exporting the session id and the master key in NSS keylog format.
376 [Martin Kaiser <martin@kaiser.cx>]
378 *) Harmonize version and its documentation. -f flag is used to display
380 [mancha <mancha1@zoho.com>]
382 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
383 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
384 [mancha <mancha1@zoho.com>]
386 *) Fix some double frees. These are not thought to be exploitable.
387 [mancha <mancha1@zoho.com>]
389 *) A missing bounds check in the handling of the TLS heartbeat extension
390 can be used to reveal up to 64k of memory to a connected client or
393 Thanks for Neel Mehta of Google Security for discovering this bug and to
394 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
395 preparing the fix (CVE-2014-0160)
396 [Adam Langley, Bodo Moeller]
398 *) Fix for the attack described in the paper "Recovering OpenSSL
399 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
400 by Yuval Yarom and Naomi Benger. Details can be obtained from:
401 http://eprint.iacr.org/2014/140
403 Thanks to Yuval Yarom and Naomi Benger for discovering this
404 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
405 [Yuval Yarom and Naomi Benger]
407 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
408 this fixes a limitation in previous versions of OpenSSL.
411 *) Experimental encrypt-then-mac support.
413 Experimental support for encrypt then mac from
414 draft-gutmann-tls-encrypt-then-mac-02.txt
416 To enable it set the appropriate extension number (0x42 for the test
417 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
419 For non-compliant peers (i.e. just about everything) this should have no
422 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
426 *) Add EVP support for key wrapping algorithms, to avoid problems with
427 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
428 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
429 algorithms and include tests cases.
432 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
436 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
437 MGF1 digest and OAEP label.
440 *) Make openssl verify return errors.
441 [Chris Palmer <palmer@google.com> and Ben Laurie]
443 *) New function ASN1_TIME_diff to calculate the difference between two
444 ASN1_TIME structures or one structure and the current time.
447 *) Update fips_test_suite to support multiple command line options. New
448 test to induce all self test errors in sequence and check expected
452 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
453 sign or verify all in one operation.
456 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
457 test programs and fips_test_suite. Includes functionality to parse
458 the minimal script output of fipsalgest.pl directly.
461 *) Add authorisation parameter to FIPS_module_mode_set().
464 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
467 *) Use separate DRBG fields for internal and external flags. New function
468 FIPS_drbg_health_check() to perform on demand health checking. Add
469 generation tests to fips_test_suite with reduced health check interval to
470 demonstrate periodic health checking. Add "nodh" option to
471 fips_test_suite to skip very slow DH test.
474 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
478 *) More extensive health check for DRBG checking many more failure modes.
479 New function FIPS_selftest_drbg_all() to handle every possible DRBG
480 combination: call this in fips_test_suite.
483 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
484 and POST to handle Dual EC cases.
487 *) Add support for canonical generation of DSA parameter 'g'. See
490 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
491 POST to handle HMAC cases.
494 *) Add functions FIPS_module_version() and FIPS_module_version_text()
495 to return numerical and string versions of the FIPS module number.
498 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
499 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
500 outside the validated module in the FIPS capable OpenSSL.
503 *) Minor change to DRBG entropy callback semantics. In some cases
504 there is no multiple of the block length between min_len and
505 max_len. Allow the callback to return more than max_len bytes
506 of entropy but discard any extra: it is the callback's responsibility
507 to ensure that the extra data discarded does not impact the
508 requested amount of entropy.
511 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
512 information in FIPS186-3, SP800-57 and SP800-131A.
515 *) CCM support via EVP. Interface is very similar to GCM case except we
516 must supply all data in one chunk (i.e. no update, final) and the
517 message length must be supplied if AAD is used. Add algorithm test
521 *) Initial version of POST overhaul. Add POST callback to allow the status
522 of POST to be monitored and/or failures induced. Modify fips_test_suite
523 to use callback. Always run all selftests even if one fails.
526 *) XTS support including algorithm test driver in the fips_gcmtest program.
527 Note: this does increase the maximum key length from 32 to 64 bytes but
528 there should be no binary compatibility issues as existing applications
529 will never use XTS mode.
532 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
533 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
534 performs algorithm blocking for unapproved PRNG types. Also do not
535 set PRNG type in FIPS_mode_set(): leave this to the application.
536 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
537 the standard OpenSSL PRNG: set additional data to a date time vector.
540 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
541 This shouldn't present any incompatibility problems because applications
542 shouldn't be using these directly and any that are will need to rethink
543 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
546 *) Extensive self tests and health checking required by SP800-90 DRBG.
547 Remove strength parameter from FIPS_drbg_instantiate and always
548 instantiate at maximum supported strength.
551 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
554 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
557 *) New function DH_compute_key_padded() to compute a DH key and pad with
558 leading zeroes if needed: this complies with SP800-56A et al.
561 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
562 anything, incomplete, subject to change and largely untested at present.
565 *) Modify fipscanisteronly build option to only build the necessary object
566 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
569 *) Add experimental option FIPSSYMS to give all symbols in
570 fipscanister.o and FIPS or fips prefix. This will avoid
571 conflicts with future versions of OpenSSL. Add perl script
572 util/fipsas.pl to preprocess assembly language source files
573 and rename any affected symbols.
576 *) Add selftest checks and algorithm block of non-fips algorithms in
577 FIPS mode. Remove DES2 from selftests.
580 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
581 return internal method without any ENGINE dependencies. Add new
582 tiny fips sign and verify functions.
585 *) New build option no-ec2m to disable characteristic 2 code.
588 *) New build option "fipscanisteronly". This only builds fipscanister.o
589 and (currently) associated fips utilities. Uses the file Makefile.fips
590 instead of Makefile.org as the prototype.
593 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
594 Update fips_gcmtest to use IV generator.
597 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
598 setting output buffer to NULL. The *Final function must be
599 called although it will not retrieve any additional data. The tag
600 can be set or retrieved with a ctrl. The IV length is by default 12
601 bytes (96 bits) but can be set to an alternative value. If the IV
602 length exceeds the maximum IV length (currently 16 bytes) it cannot be
606 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
607 underlying do_cipher function handles all cipher semantics itself
608 including padding and finalisation. This is useful if (for example)
609 an ENGINE cipher handles block padding itself. The behaviour of
610 do_cipher is subtly changed if this flag is set: the return value
611 is the number of characters written to the output buffer (zero is
612 no longer an error code) or a negative error code. Also if the
613 input buffer is NULL and length 0 finalisation should be performed.
616 *) If a candidate issuer certificate is already part of the constructed
617 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
620 *) Improve forward-security support: add functions
622 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
623 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
625 for use by SSL/TLS servers; the callback function will be called whenever a
626 new session is created, and gets to decide whether the session may be
627 cached to make it resumable (return 0) or not (return 1). (As by the
628 SSL/TLS protocol specifications, the session_id sent by the server will be
629 empty to indicate that the session is not resumable; also, the server will
630 not generate RFC 4507 (RFC 5077) session tickets.)
632 A simple reasonable callback implementation is to return is_forward_secure.
633 This parameter will be set to 1 or 0 depending on the ciphersuite selected
634 by the SSL/TLS server library, indicating whether it can provide forward
636 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
638 *) New -verify_name option in command line utilities to set verification
642 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
643 Add CMAC pkey methods.
646 *) Experimental renegotiation in s_server -www mode. If the client
647 browses /reneg connection is renegotiated. If /renegcert it is
648 renegotiated requesting a certificate.
651 *) Add an "external" session cache for debugging purposes to s_server. This
652 should help trace issues which normally are only apparent in deployed
653 multi-process servers.
656 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
657 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
658 BIO_set_cipher() and some obscure PEM functions were changed so they
659 can now return an error. The RAND changes required a change to the
660 RAND_METHOD structure.
663 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
664 a gcc attribute to warn if the result of a function is ignored. This
665 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
666 whose return value is often ignored.
669 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
671 *) BN_mod_exp may produce incorrect results on x86_64
673 There is a carry propagating bug in the x86_64 Montgomery squaring
674 procedure. No EC algorithms are affected. Analysis suggests that attacks
675 against RSA and DSA as a result of this defect would be very difficult to
676 perform and are not believed likely. Attacks against DH are considered just
677 feasible (although very difficult) because most of the work necessary to
678 deduce information about a private key may be performed offline. The amount
679 of resources required for such an attack would be very significant and
680 likely only accessible to a limited number of attackers. An attacker would
681 additionally need online access to an unpatched system using the target
682 private key in a scenario with persistent DH parameters and a private
683 key that is shared between multiple clients. For example this can occur by
684 default in OpenSSL DHE based SSL/TLS ciphersuites.
686 This issue was reported to OpenSSL by Hanno Böck.
690 *) Certificate verify crash with missing PSS parameter
692 The signature verification routines will crash with a NULL pointer
693 dereference if presented with an ASN.1 signature using the RSA PSS
694 algorithm and absent mask generation function parameter. Since these
695 routines are used to verify certificate signature algorithms this can be
696 used to crash any certificate verification operation and exploited in a
697 DoS attack. Any application which performs certificate verification is
698 vulnerable including OpenSSL clients and servers which enable client
701 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
705 *) X509_ATTRIBUTE memory leak
707 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
708 memory. This structure is used by the PKCS#7 and CMS routines so any
709 application which reads PKCS#7 or CMS data from untrusted sources is
710 affected. SSL/TLS is not affected.
712 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
717 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
718 This changes the decoding behaviour for some invalid messages,
719 though the change is mostly in the more lenient direction, and
720 legacy behaviour is preserved as much as possible.
723 *) In DSA_generate_parameters_ex, if the provided seed is too short,
725 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
727 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
729 *) Alternate chains certificate forgery
731 During certificate verfification, OpenSSL will attempt to find an
732 alternative certificate chain if the first attempt to build such a chain
733 fails. An error in the implementation of this logic can mean that an
734 attacker could cause certain checks on untrusted certificates to be
735 bypassed, such as the CA flag, enabling them to use a valid leaf
736 certificate to act as a CA and "issue" an invalid certificate.
738 This issue was reported to OpenSSL by Adam Langley/David Benjamin
742 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
744 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
745 incompatibility in the handling of HMAC. The previous ABI has now been
749 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
751 *) Malformed ECParameters causes infinite loop
753 When processing an ECParameters structure OpenSSL enters an infinite loop
754 if the curve specified is over a specially malformed binary polynomial
757 This can be used to perform denial of service against any
758 system which processes public keys, certificate requests or
759 certificates. This includes TLS clients and TLS servers with
760 client authentication enabled.
762 This issue was reported to OpenSSL by Joseph Barr-Pixton.
766 *) Exploitable out-of-bounds read in X509_cmp_time
768 X509_cmp_time does not properly check the length of the ASN1_TIME
769 string and can read a few bytes out of bounds. In addition,
770 X509_cmp_time accepts an arbitrary number of fractional seconds in the
773 An attacker can use this to craft malformed certificates and CRLs of
774 various sizes and potentially cause a segmentation fault, resulting in
775 a DoS on applications that verify certificates or CRLs. TLS clients
776 that verify CRLs are affected. TLS clients and servers with client
777 authentication enabled may be affected if they use custom verification
780 This issue was reported to OpenSSL by Robert Swiecki (Google), and
781 independently by Hanno Böck.
785 *) PKCS7 crash with missing EnvelopedContent
787 The PKCS#7 parsing code does not handle missing inner EncryptedContent
788 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
789 with missing content and trigger a NULL pointer dereference on parsing.
791 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
792 structures from untrusted sources are affected. OpenSSL clients and
793 servers are not affected.
795 This issue was reported to OpenSSL by Michal Zalewski (Google).
799 *) CMS verify infinite loop with unknown hash function
801 When verifying a signedData message the CMS code can enter an infinite loop
802 if presented with an unknown hash function OID. This can be used to perform
803 denial of service against any system which verifies signedData messages using
805 This issue was reported to OpenSSL by Johannes Bauer.
809 *) Race condition handling NewSessionTicket
811 If a NewSessionTicket is received by a multi-threaded client when attempting to
812 reuse a previous ticket then a race condition can occur potentially leading to
813 a double free of the ticket data.
817 *) Only support 256-bit or stronger elliptic curves with the
818 'ecdh_auto' setting (server) or by default (client). Of supported
819 curves, prefer P-256 (both).
822 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
824 *) ClientHello sigalgs DoS fix
826 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
827 invalid signature algorithms extension a NULL pointer dereference will
828 occur. This can be exploited in a DoS attack against the server.
830 This issue was was reported to OpenSSL by David Ramos of Stanford
833 [Stephen Henson and Matt Caswell]
835 *) Multiblock corrupted pointer fix
837 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
838 feature only applies on 64 bit x86 architecture platforms that support AES
839 NI instructions. A defect in the implementation of "multiblock" can cause
840 OpenSSL's internal write buffer to become incorrectly set to NULL when
841 using non-blocking IO. Typically, when the user application is using a
842 socket BIO for writing, this will only result in a failed connection.
843 However if some other BIO is used then it is likely that a segmentation
844 fault will be triggered, thus enabling a potential DoS attack.
846 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
850 *) Segmentation fault in DTLSv1_listen fix
852 The DTLSv1_listen function is intended to be stateless and processes the
853 initial ClientHello from many peers. It is common for user code to loop
854 over the call to DTLSv1_listen until a valid ClientHello is received with
855 an associated cookie. A defect in the implementation of DTLSv1_listen means
856 that state is preserved in the SSL object from one invocation to the next
857 that can lead to a segmentation fault. Errors processing the initial
858 ClientHello can trigger this scenario. An example of such an error could be
859 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
862 This issue was reported to OpenSSL by Per Allansson.
866 *) Segmentation fault in ASN1_TYPE_cmp fix
868 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
869 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
870 certificate signature algorithm consistency this can be used to crash any
871 certificate verification operation and exploited in a DoS attack. Any
872 application which performs certificate verification is vulnerable including
873 OpenSSL clients and servers which enable client authentication.
877 *) Segmentation fault for invalid PSS parameters fix
879 The signature verification routines will crash with a NULL pointer
880 dereference if presented with an ASN.1 signature using the RSA PSS
881 algorithm and invalid parameters. Since these routines are used to verify
882 certificate signature algorithms this can be used to crash any
883 certificate verification operation and exploited in a DoS attack. Any
884 application which performs certificate verification is vulnerable including
885 OpenSSL clients and servers which enable client authentication.
887 This issue was was reported to OpenSSL by Brian Carpenter.
891 *) ASN.1 structure reuse memory corruption fix
893 Reusing a structure in ASN.1 parsing may allow an attacker to cause
894 memory corruption via an invalid write. Such reuse is and has been
895 strongly discouraged and is believed to be rare.
897 Applications that parse structures containing CHOICE or ANY DEFINED BY
898 components may be affected. Certificate parsing (d2i_X509 and related
899 functions) are however not affected. OpenSSL clients and servers are
904 *) PKCS7 NULL pointer dereferences fix
906 The PKCS#7 parsing code does not handle missing outer ContentInfo
907 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
908 missing content and trigger a NULL pointer dereference on parsing.
910 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
911 otherwise parse PKCS#7 structures from untrusted sources are
912 affected. OpenSSL clients and servers are not affected.
914 This issue was reported to OpenSSL by Michal Zalewski (Google).
918 *) DoS via reachable assert in SSLv2 servers fix
920 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
921 servers that both support SSLv2 and enable export cipher suites by sending
922 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
924 This issue was discovered by Sean Burford (Google) and Emilia Käsper
925 (OpenSSL development team).
929 *) Empty CKE with client auth and DHE fix
931 If client auth is used then a server can seg fault in the event of a DHE
932 ciphersuite being selected and a zero length ClientKeyExchange message
933 being sent by the client. This could be exploited in a DoS attack.
937 *) Handshake with unseeded PRNG fix
939 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
940 with an unseeded PRNG. The conditions are:
941 - The client is on a platform where the PRNG has not been seeded
942 automatically, and the user has not seeded manually
943 - A protocol specific client method version has been used (i.e. not
944 SSL_client_methodv23)
945 - A ciphersuite is used that does not require additional random data from
946 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
948 If the handshake succeeds then the client random that has been used will
949 have been generated from a PRNG with insufficient entropy and therefore the
950 output may be predictable.
952 For example using the following command with an unseeded openssl will
953 succeed on an unpatched platform:
955 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
959 *) Use After Free following d2i_ECPrivatekey error fix
961 A malformed EC private key file consumed via the d2i_ECPrivateKey function
962 could cause a use after free condition. This, in turn, could cause a double
963 free in several private key parsing functions (such as d2i_PrivateKey
964 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
965 for applications that receive EC private keys from untrusted
966 sources. This scenario is considered rare.
968 This issue was discovered by the BoringSSL project and fixed in their
973 *) X509_to_X509_REQ NULL pointer deref fix
975 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
976 the certificate key is invalid. This function is rarely used in practice.
978 This issue was discovered by Brian Carpenter.
982 *) Removed the export ciphers from the DEFAULT ciphers
985 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
987 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
988 ARMv5 through ARMv8, as opposite to "locking" it to single one.
989 So far those who have to target multiple plaforms would compromise
990 and argue that binary targeting say ARMv5 would still execute on
991 ARMv8. "Universal" build resolves this compromise by providing
992 near-optimal performance even on newer platforms.
995 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
996 (other platforms pending).
997 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
999 *) Add support for the SignedCertificateTimestampList certificate and
1000 OCSP response extensions from RFC6962.
1003 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1004 for corner cases. (Certain input points at infinity could lead to
1005 bogus results, with non-infinity inputs mapped to infinity too.)
1008 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1009 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1010 common cases are optimized and there still is room for further
1011 improvements. Vector Permutation AES for Altivec is also added.
1014 *) Add support for little-endian ppc64 Linux target.
1015 [Marcelo Cerri (IBM)]
1017 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1018 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1019 are optimized and there still is room for further improvements.
1020 Both 32- and 64-bit modes are supported.
1021 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1023 *) Improved ARMv7 NEON support.
1026 *) Support for SPARC Architecture 2011 crypto extensions, first
1027 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1028 SHA256/512, MD5, GHASH and modular exponentiation.
1029 [Andy Polyakov, David Miller]
1031 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1033 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1035 *) Support for new and upcoming Intel processors, including AVX2,
1036 BMI and SHA ISA extensions. This includes additional "stitched"
1037 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1040 This work was sponsored by Intel Corp.
1043 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1044 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1045 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1048 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1049 this fixes a limiation in previous versions of OpenSSL.
1052 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1053 MGF1 digest and OAEP label.
1056 *) Add EVP support for key wrapping algorithms, to avoid problems with
1057 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1058 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1059 algorithms and include tests cases.
1062 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1064 [Douglas E. Engert, Steve Henson]
1066 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1067 difference in days and seconds between two tm or ASN1_TIME structures.
1070 *) Add -rev test option to s_server to just reverse order of characters
1071 received by client and send back to server. Also prints an abbreviated
1072 summary of the connection parameters.
1075 *) New option -brief for s_client and s_server to print out a brief summary
1076 of connection parameters.
1079 *) Add callbacks for arbitrary TLS extensions.
1080 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1082 *) New option -crl_download in several openssl utilities to download CRLs
1083 from CRLDP extension in certificates.
1086 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1089 *) New function X509_CRL_diff to generate a delta CRL from the difference
1090 of two full CRLs. Add support to "crl" utility.
1093 *) New functions to set lookup_crls function and to retrieve
1094 X509_STORE from X509_STORE_CTX.
1097 *) Print out deprecated issuer and subject unique ID fields in
1101 *) Extend OCSP I/O functions so they can be used for simple general purpose
1102 HTTP as well as OCSP. New wrapper function which can be used to download
1103 CRLs using the OCSP API.
1106 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1109 *) SSL_CONF* functions. These provide a common framework for application
1110 configuration using configuration files or command lines.
1113 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1114 message callback and prints the results. Needs compile time option
1115 "enable-ssl-trace". New options to s_client and s_server to enable
1119 *) New ctrl and macro to retrieve supported points extensions.
1120 Print out extension in s_server and s_client.
1123 *) New functions to retrieve certificate signature and signature
1127 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1131 *) New Suite B modes for TLS code. These use and enforce the requirements
1132 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1133 only use Suite B curves. The Suite B modes can be set by using the
1134 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1137 *) New chain verification flags for Suite B levels of security. Check
1138 algorithms are acceptable when flags are set in X509_verify_cert.
1141 *) Make tls1_check_chain return a set of flags indicating checks passed
1142 by a certificate chain. Add additional tests to handle client
1143 certificates: checks for matching certificate type and issuer name
1147 *) If an attempt is made to use a signature algorithm not in the peer
1148 preference list abort the handshake. If client has no suitable
1149 signature algorithms in response to a certificate request do not
1150 use the certificate.
1153 *) If server EC tmp key is not in client preference list abort handshake.
1156 *) Add support for certificate stores in CERT structure. This makes it
1157 possible to have different stores per SSL structure or one store in
1158 the parent SSL_CTX. Include distint stores for certificate chain
1159 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1160 to build and store a certificate chain in CERT structure: returing
1161 an error if the chain cannot be built: this will allow applications
1162 to test if a chain is correctly configured.
1164 Note: if the CERT based stores are not set then the parent SSL_CTX
1165 store is used to retain compatibility with existing behaviour.
1169 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1170 mask based on the current session, check mask when sending client
1171 hello and checking the requested ciphersuite.
1174 *) New ctrls to retrieve and set certificate types in a certificate
1175 request message. Print out received values in s_client. If certificate
1176 types is not set with custom values set sensible values based on
1177 supported signature algorithms.
1180 *) Support for distinct client and server supported signature algorithms.
1183 *) Add certificate callback. If set this is called whenever a certificate
1184 is required by client or server. An application can decide which
1185 certificate chain to present based on arbitrary criteria: for example
1186 supported signature algorithms. Add very simple example to s_server.
1187 This fixes many of the problems and restrictions of the existing client
1188 certificate callback: for example you can now clear an existing
1189 certificate and specify the whole chain.
1192 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1193 the certificate can be used for (if anything). Set valid_flags field
1194 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1195 to have similar checks in it.
1197 Add new "cert_flags" field to CERT structure and include a "strict mode".
1198 This enforces some TLS certificate requirements (such as only permitting
1199 certificate signature algorithms contained in the supported algorithms
1200 extension) which some implementations ignore: this option should be used
1201 with caution as it could cause interoperability issues.
1204 *) Update and tidy signature algorithm extension processing. Work out
1205 shared signature algorithms based on preferences and peer algorithms
1206 and print them out in s_client and s_server. Abort handshake if no
1207 shared signature algorithms.
1210 *) Add new functions to allow customised supported signature algorithms
1211 for SSL and SSL_CTX structures. Add options to s_client and s_server
1215 *) New function SSL_certs_clear() to delete all references to certificates
1216 from an SSL structure. Before this once a certificate had been added
1217 it couldn't be removed.
1220 *) Integrate hostname, email address and IP address checking with certificate
1221 verification. New verify options supporting checking in opensl utility.
1224 *) Fixes and wildcard matching support to hostname and email checking
1225 functions. Add manual page.
1226 [Florian Weimer (Red Hat Product Security Team)]
1228 *) New functions to check a hostname email or IP address against a
1229 certificate. Add options x509 utility to print results of checks against
1233 *) Fix OCSP checking.
1234 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1236 *) Initial experimental support for explicitly trusted non-root CAs.
1237 OpenSSL still tries to build a complete chain to a root but if an
1238 intermediate CA has a trust setting included that is used. The first
1239 setting is used: whether to trust (e.g., -addtrust option to the x509
1243 *) Add -trusted_first option which attempts to find certificates in the
1244 trusted store even if an untrusted chain is also supplied.
1247 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1248 platform support for Linux and Android.
1251 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1254 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1255 When in FIPS mode the approved implementations are used as normal,
1256 when not in FIPS mode the internal unapproved versions are used instead.
1257 This means that the FIPS capable OpenSSL isn't forced to use the
1258 (often lower performance) FIPS implementations outside FIPS mode.
1261 *) Transparently support X9.42 DH parameters when calling
1262 PEM_read_bio_DHparameters. This means existing applications can handle
1263 the new parameter format automatically.
1266 *) Initial experimental support for X9.42 DH parameter format: mainly
1267 to support use of 'q' parameter for RFC5114 parameters.
1270 *) Add DH parameters from RFC5114 including test data to dhtest.
1273 *) Support for automatic EC temporary key parameter selection. If enabled
1274 the most preferred EC parameters are automatically used instead of
1275 hardcoded fixed parameters. Now a server just has to call:
1276 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1277 support ECDH and use the most appropriate parameters.
1280 *) Enhance and tidy EC curve and point format TLS extension code. Use
1281 static structures instead of allocation if default values are used.
1282 New ctrls to set curves we wish to support and to retrieve shared curves.
1283 Print out shared curves in s_server. New options to s_server and s_client
1284 to set list of supported curves.
1287 *) New ctrls to retrieve supported signature algorithms and
1288 supported curve values as an array of NIDs. Extend openssl utility
1289 to print out received values.
1292 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1293 between NIDs and the more common NIST names such as "P-256". Enhance
1294 ecparam utility and ECC method to recognise the NIST names for curves.
1297 *) Enhance SSL/TLS certificate chain handling to support different
1298 chains for each certificate instead of one chain in the parent SSL_CTX.
1301 *) Support for fixed DH ciphersuite client authentication: where both
1302 server and client use DH certificates with common parameters.
1305 *) Support for fixed DH ciphersuites: those requiring DH server
1309 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1311 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1312 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1313 X509_CINF_get_signature were reverted post internal team review.
1315 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1317 *) Build fixes for the Windows and OpenVMS platforms
1318 [Matt Caswell and Richard Levitte]
1320 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1322 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1323 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1324 dereference. This could lead to a Denial Of Service attack. Thanks to
1325 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1329 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1330 dtls1_buffer_record function under certain conditions. In particular this
1331 could occur if an attacker sent repeated DTLS records with the same
1332 sequence number but for the next epoch. The memory leak could be exploited
1333 by an attacker in a Denial of Service attack through memory exhaustion.
1334 Thanks to Chris Mueller for reporting this issue.
1338 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1339 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1340 method would be set to NULL which could later result in a NULL pointer
1341 dereference. Thanks to Frank Schmirler for reporting this issue.
1345 *) Abort handshake if server key exchange message is omitted for ephemeral
1348 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1349 reporting this issue.
1353 *) Remove non-export ephemeral RSA code on client and server. This code
1354 violated the TLS standard by allowing the use of temporary RSA keys in
1355 non-export ciphersuites and could be used by a server to effectively
1356 downgrade the RSA key length used to a value smaller than the server
1357 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1358 INRIA or reporting this issue.
1362 *) Fixed issue where DH client certificates are accepted without verification.
1363 An OpenSSL server will accept a DH certificate for client authentication
1364 without the certificate verify message. This effectively allows a client to
1365 authenticate without the use of a private key. This only affects servers
1366 which trust a client certificate authority which issues certificates
1367 containing DH keys: these are extremely rare and hardly ever encountered.
1368 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1373 *) Ensure that the session ID context of an SSL is updated when its
1374 SSL_CTX is updated via SSL_set_SSL_CTX.
1376 The session ID context is typically set from the parent SSL_CTX,
1377 and can vary with the CTX.
1380 *) Fix various certificate fingerprint issues.
1382 By using non-DER or invalid encodings outside the signed portion of a
1383 certificate the fingerprint can be changed without breaking the signature.
1384 Although no details of the signed portion of the certificate can be changed
1385 this can cause problems with some applications: e.g. those using the
1386 certificate fingerprint for blacklists.
1388 1. Reject signatures with non zero unused bits.
1390 If the BIT STRING containing the signature has non zero unused bits reject
1391 the signature. All current signature algorithms require zero unused bits.
1393 2. Check certificate algorithm consistency.
1395 Check the AlgorithmIdentifier inside TBS matches the one in the
1396 certificate signature. NB: this will result in signature failure
1397 errors for some broken certificates.
1399 Thanks to Konrad Kraszewski from Google for reporting this issue.
1401 3. Check DSA/ECDSA signatures use DER.
1403 Reencode DSA/ECDSA signatures and compare with the original received
1404 signature. Return an error if there is a mismatch.
1406 This will reject various cases including garbage after signature
1407 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1408 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1409 (negative or with leading zeroes).
1411 Further analysis was conducted and fixes were developed by Stephen Henson
1412 of the OpenSSL core team.
1417 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1418 results on some platforms, including x86_64. This bug occurs at random
1419 with a very low probability, and is not known to be exploitable in any
1420 way, though its exact impact is difficult to determine. Thanks to Pieter
1421 Wuille (Blockstream) who reported this issue and also suggested an initial
1422 fix. Further analysis was conducted by the OpenSSL development team and
1423 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1424 the OpenSSL core team.
1428 *) Do not resume sessions on the server if the negotiated protocol
1429 version does not match the session's version. Resuming with a different
1430 version, while not strictly forbidden by the RFC, is of questionable
1431 sanity and breaks all known clients.
1432 [David Benjamin, Emilia Käsper]
1434 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1435 early CCS messages during renegotiation. (Note that because
1436 renegotiation is encrypted, this early CCS was not exploitable.)
1439 *) Tighten client-side session ticket handling during renegotiation:
1440 ensure that the client only accepts a session ticket if the server sends
1441 the extension anew in the ServerHello. Previously, a TLS client would
1442 reuse the old extension state and thus accept a session ticket if one was
1443 announced in the initial ServerHello.
1445 Similarly, ensure that the client requires a session ticket if one
1446 was advertised in the ServerHello. Previously, a TLS client would
1447 ignore a missing NewSessionTicket message.
1450 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1452 *) SRTP Memory Leak.
1454 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1455 sends a carefully crafted handshake message, to cause OpenSSL to fail
1456 to free up to 64k of memory causing a memory leak. This could be
1457 exploited in a Denial Of Service attack. This issue affects OpenSSL
1458 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1459 whether SRTP is used or configured. Implementations of OpenSSL that
1460 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1462 The fix was developed by the OpenSSL team.
1466 *) Session Ticket Memory Leak.
1468 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1469 integrity of that ticket is first verified. In the event of a session
1470 ticket integrity check failing, OpenSSL will fail to free memory
1471 causing a memory leak. By sending a large number of invalid session
1472 tickets an attacker could exploit this issue in a Denial Of Service
1477 *) Build option no-ssl3 is incomplete.
1479 When OpenSSL is configured with "no-ssl3" as a build option, servers
1480 could accept and complete a SSL 3.0 handshake, and clients could be
1481 configured to send them.
1483 [Akamai and the OpenSSL team]
1485 *) Add support for TLS_FALLBACK_SCSV.
1486 Client applications doing fallback retries should call
1487 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1489 [Adam Langley, Bodo Moeller]
1491 *) Add additional DigestInfo checks.
1493 Reencode DigestInto in DER and check against the original when
1494 verifying RSA signature: this will reject any improperly encoded
1495 DigestInfo structures.
1497 Note: this is a precautionary measure and no attacks are currently known.
1501 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1503 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1504 SRP code can be overrun an internal buffer. Add sanity check that
1505 g, A, B < N to SRP code.
1507 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1508 Group for discovering this issue.
1512 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1513 TLS 1.0 instead of higher protocol versions when the ClientHello message
1514 is badly fragmented. This allows a man-in-the-middle attacker to force a
1515 downgrade to TLS 1.0 even if both the server and the client support a
1516 higher protocol version, by modifying the client's TLS records.
1518 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1519 researching this issue.
1523 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1524 to a denial of service attack. A malicious server can crash the client
1525 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1526 ciphersuite and sending carefully crafted handshake messages.
1528 Thanks to Felix Gröbert (Google) for discovering and researching this
1533 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1534 to leak memory. This can be exploited through a Denial of Service attack.
1535 Thanks to Adam Langley for discovering and researching this issue.
1539 *) An attacker can force openssl to consume large amounts of memory whilst
1540 processing DTLS handshake messages. This can be exploited through a
1541 Denial of Service attack.
1542 Thanks to Adam Langley for discovering and researching this issue.
1546 *) An attacker can force an error condition which causes openssl to crash
1547 whilst processing DTLS packets due to memory being freed twice. This
1548 can be exploited through a Denial of Service attack.
1549 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1554 *) If a multithreaded client connects to a malicious server using a resumed
1555 session and the server sends an ec point format extension it could write
1556 up to 255 bytes to freed memory.
1558 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1563 *) A malicious server can crash an OpenSSL client with a null pointer
1564 dereference (read) by specifying an SRP ciphersuite even though it was not
1565 properly negotiated with the client. This can be exploited through a
1566 Denial of Service attack.
1568 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1569 discovering and researching this issue.
1573 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1574 X509_name_oneline, X509_name_print_ex et al. to leak some information
1575 from the stack. Applications may be affected if they echo pretty printing
1576 output to the attacker.
1578 Thanks to Ivan Fratric (Google) for discovering this issue.
1580 [Emilia Käsper, and Steve Henson]
1582 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1583 for corner cases. (Certain input points at infinity could lead to
1584 bogus results, with non-infinity inputs mapped to infinity too.)
1587 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1589 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1590 handshake can force the use of weak keying material in OpenSSL
1591 SSL/TLS clients and servers.
1593 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1594 researching this issue. (CVE-2014-0224)
1595 [KIKUCHI Masashi, Steve Henson]
1597 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1598 OpenSSL DTLS client the code can be made to recurse eventually crashing
1601 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1603 [Imre Rad, Steve Henson]
1605 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1606 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1607 client or server. This is potentially exploitable to run arbitrary
1608 code on a vulnerable client or server.
1610 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1611 [Jüri Aedla, Steve Henson]
1613 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1614 are subject to a denial of service attack.
1616 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1617 this issue. (CVE-2014-3470)
1618 [Felix Gröbert, Ivan Fratric, Steve Henson]
1620 *) Harmonize version and its documentation. -f flag is used to display
1622 [mancha <mancha1@zoho.com>]
1624 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1625 in i2d_ECPrivateKey.
1626 [mancha <mancha1@zoho.com>]
1628 *) Fix some double frees. These are not thought to be exploitable.
1629 [mancha <mancha1@zoho.com>]
1631 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1633 *) A missing bounds check in the handling of the TLS heartbeat extension
1634 can be used to reveal up to 64k of memory to a connected client or
1637 Thanks for Neel Mehta of Google Security for discovering this bug and to
1638 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1639 preparing the fix (CVE-2014-0160)
1640 [Adam Langley, Bodo Moeller]
1642 *) Fix for the attack described in the paper "Recovering OpenSSL
1643 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1644 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1645 http://eprint.iacr.org/2014/140
1647 Thanks to Yuval Yarom and Naomi Benger for discovering this
1648 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1649 [Yuval Yarom and Naomi Benger]
1651 *) TLS pad extension: draft-agl-tls-padding-03
1653 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1654 TLS client Hello record length value would otherwise be > 255 and
1655 less that 512 pad with a dummy extension containing zeroes so it
1656 is at least 512 bytes long.
1658 [Adam Langley, Steve Henson]
1660 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1662 *) Fix for TLS record tampering bug. A carefully crafted invalid
1663 handshake could crash OpenSSL with a NULL pointer exception.
1664 Thanks to Anton Johansson for reporting this issues.
1667 *) Keep original DTLS digest and encryption contexts in retransmission
1668 structures so we can use the previous session parameters if they need
1669 to be resent. (CVE-2013-6450)
1672 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1673 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1674 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1675 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1676 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1677 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1678 [Rob Stradling, Adam Langley]
1680 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1682 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1683 supporting platforms or when small records were transferred.
1684 [Andy Polyakov, Steve Henson]
1686 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1688 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1690 This addresses the flaw in CBC record processing discovered by
1691 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1692 at: http://www.isg.rhul.ac.uk/tls/
1694 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1695 Security Group at Royal Holloway, University of London
1696 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1697 Emilia Käsper for the initial patch.
1699 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1701 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1702 ciphersuites which can be exploited in a denial of service attack.
1703 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1704 and detecting this bug and to Wolfgang Ettlinger
1705 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1709 *) Return an error when checking OCSP signatures when key is NULL.
1710 This fixes a DoS attack. (CVE-2013-0166)
1713 *) Make openssl verify return errors.
1714 [Chris Palmer <palmer@google.com> and Ben Laurie]
1716 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1717 the right response is stapled. Also change SSL_get_certificate()
1718 so it returns the certificate actually sent.
1719 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1720 [Rob Stradling <rob.stradling@comodo.com>]
1722 *) Fix possible deadlock when decoding public keys.
1725 *) Don't use TLS 1.0 record version number in initial client hello
1729 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1731 *) Sanity check record length before skipping explicit IV in TLS
1732 1.2, 1.1 and DTLS to fix DoS attack.
1734 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1735 fuzzing as a service testing platform.
1739 *) Initialise tkeylen properly when encrypting CMS messages.
1740 Thanks to Solar Designer of Openwall for reporting this issue.
1743 *) In FIPS mode don't try to use composite ciphers as they are not
1747 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1749 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1750 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1751 mean any application compiled against OpenSSL 1.0.0 headers setting
1752 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1753 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1754 0x10000000L Any application which was previously compiled against
1755 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1756 will need to be recompiled as a result. Letting be results in
1757 inability to disable specifically TLS 1.1 and in client context,
1758 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1761 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1762 disable just protocol X, but all protocols above X *if* there are
1763 protocols *below* X still enabled. In more practical terms it means
1764 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1765 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1766 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1770 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1772 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1773 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1774 in CRYPTO_realloc_clean.
1776 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1777 issue and to Adam Langley <agl@chromium.org> for fixing it.
1779 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1781 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1784 *) Workarounds for some broken servers that "hang" if a client hello
1785 record length exceeds 255 bytes.
1787 1. Do not use record version number > TLS 1.0 in initial client
1788 hello: some (but not all) hanging servers will now work.
1789 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1790 the number of ciphers sent in the client hello. This should be
1791 set to an even number, such as 50, for example by passing:
1792 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1793 Most broken servers should now work.
1794 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1795 TLS 1.2 client support entirely.
1798 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1801 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1803 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1804 STRING form instead of a DigestInfo.
1807 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1808 and the RSA_sign/RSA_verify functions. This was made more apparent when
1809 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1810 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1811 the correct format in RSA_verify so both forms transparently work.
1814 *) Some servers which support TLS 1.0 can choke if we initially indicate
1815 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1816 encrypted premaster secret. As a workaround use the maximum permitted
1817 client version in client hello, this should keep such servers happy
1818 and still work with previous versions of OpenSSL.
1821 *) Add support for TLS/DTLS heartbeats.
1822 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1824 *) Add support for SCTP.
1825 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1827 *) Improved PRNG seeding for VOS.
1828 [Paul Green <Paul.Green@stratus.com>]
1830 *) Extensive assembler packs updates, most notably:
1832 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1833 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1834 - x86_64: bit-sliced AES implementation;
1835 - ARM: NEON support, contemporary platforms optimizations;
1836 - s390x: z196 support;
1837 - *: GHASH and GF(2^m) multiplication implementations;
1841 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1842 (removal of unnecessary code)
1843 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1845 *) Add TLS key material exporter from RFC 5705.
1848 *) Add DTLS-SRTP negotiation from RFC 5764.
1851 *) Add Next Protocol Negotiation,
1852 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1853 disabled with a no-npn flag to config or Configure. Code donated
1855 [Adam Langley <agl@google.com> and Ben Laurie]
1857 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1858 NIST-P256, NIST-P521, with constant-time single point multiplication on
1859 typical inputs. Compiler support for the nonstandard type __uint128_t is
1860 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1861 Code made available under Apache License version 2.0.
1863 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1864 line to include this in your build of OpenSSL, and run "make depend" (or
1865 "make update"). This enables the following EC_METHODs:
1867 EC_GFp_nistp224_method()
1868 EC_GFp_nistp256_method()
1869 EC_GFp_nistp521_method()
1871 EC_GROUP_new_by_curve_name() will automatically use these (while
1872 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1874 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1876 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1877 all platforms. Move ssize_t definition from e_os.h to the public
1878 header file e_os2.h as it now appears in public header file cms.h
1881 *) New -sigopt option to the ca, req and x509 utilities. Additional
1882 signature parameters can be passed using this option and in
1886 *) Add RSA PSS signing function. This will generate and set the
1887 appropriate AlgorithmIdentifiers for PSS based on those in the
1888 corresponding EVP_MD_CTX structure. No application support yet.
1891 *) Support for companion algorithm specific ASN1 signing routines.
1892 New function ASN1_item_sign_ctx() signs a pre-initialised
1893 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1894 the appropriate parameters.
1897 *) Add new algorithm specific ASN1 verification initialisation function
1898 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1899 handling will be the same no matter what EVP_PKEY_METHOD is used.
1900 Add a PSS handler to support verification of PSS signatures: checked
1901 against a number of sample certificates.
1904 *) Add signature printing for PSS. Add PSS OIDs.
1905 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1907 *) Add algorithm specific signature printing. An individual ASN1 method
1908 can now print out signatures instead of the standard hex dump.
1910 More complex signatures (e.g. PSS) can print out more meaningful
1911 information. Include DSA version that prints out the signature
1915 *) Password based recipient info support for CMS library: implementing
1919 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1920 neatly separates the code into cipher and PBE sections and is required
1921 for some algorithms that split PBES2 into separate pieces (such as
1922 password based CMS).
1925 *) Session-handling fixes:
1926 - Fix handling of connections that are resuming with a session ID,
1927 but also support Session Tickets.
1928 - Fix a bug that suppressed issuing of a new ticket if the client
1929 presented a ticket with an expired session.
1930 - Try to set the ticket lifetime hint to something reasonable.
1931 - Make tickets shorter by excluding irrelevant information.
1932 - On the client side, don't ignore renewed tickets.
1933 [Adam Langley, Bodo Moeller (Google)]
1935 *) Fix PSK session representation.
1938 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1940 This work was sponsored by Intel.
1943 *) Add GCM support to TLS library. Some custom code is needed to split
1944 the IV between the fixed (from PRF) and explicit (from TLS record)
1945 portions. This adds all GCM ciphersuites supported by RFC5288 and
1946 RFC5289. Generalise some AES* cipherstrings to include GCM and
1947 add a special AESGCM string for GCM only.
1950 *) Expand range of ctrls for AES GCM. Permit setting invocation
1951 field on decrypt and retrieval of invocation field only on encrypt.
1954 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1955 As required by RFC5289 these ciphersuites cannot be used if for
1956 versions of TLS earlier than 1.2.
1959 *) For FIPS capable OpenSSL interpret a NULL default public key method
1960 as unset and return the appropriate default but do *not* set the default.
1961 This means we can return the appropriate method in applications that
1962 switch between FIPS and non-FIPS modes.
1965 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1966 ENGINE is used then we cannot handle that in the FIPS module so we
1967 keep original code iff non-FIPS operations are allowed.
1970 *) Add -attime option to openssl utilities.
1971 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1973 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1976 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1977 FIPS EC methods unconditionally for now.
1980 *) New build option no-ec2m to disable characteristic 2 code.
1983 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1984 all cases can be covered as some introduce binary incompatibilities.
1987 *) Redirect RSA operations to FIPS module including keygen,
1988 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1991 *) Add similar low level API blocking to ciphers.
1994 *) Low level digest APIs are not approved in FIPS mode: any attempt
1995 to use these will cause a fatal error. Applications that *really* want
1996 to use them can use the private_* version instead.
1999 *) Redirect cipher operations to FIPS module for FIPS builds.
2002 *) Redirect digest operations to FIPS module for FIPS builds.
2005 *) Update build system to add "fips" flag which will link in fipscanister.o
2006 for static and shared library builds embedding a signature if needed.
2009 *) Output TLS supported curves in preference order instead of numerical
2010 order. This is currently hardcoded for the highest order curves first.
2011 This should be configurable so applications can judge speed vs strength.
2014 *) Add TLS v1.2 server support for client authentication.
2017 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2021 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2022 FIPS modules versions.
2025 *) Add TLS v1.2 client side support for client authentication. Keep cache
2026 of handshake records longer as we don't know the hash algorithm to use
2027 until after the certificate request message is received.
2030 *) Initial TLS v1.2 client support. Add a default signature algorithms
2031 extension including all the algorithms we support. Parse new signature
2032 format in client key exchange. Relax some ECC signing restrictions for
2033 TLS v1.2 as indicated in RFC5246.
2036 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2037 to new signature format when needed using client digest preference.
2038 All server ciphersuites should now work correctly in TLS v1.2. No client
2039 support yet and no support for client certificates.
2042 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2043 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2044 ciphersuites. At present only RSA key exchange ciphersuites work with
2045 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2046 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2047 and version checking.
2050 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2051 with this defined it will not be affected by any changes to ssl internal
2052 structures. Add several utility functions to allow openssl application
2053 to work with OPENSSL_NO_SSL_INTERN defined.
2057 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2059 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2062 *) Permit abbreviated handshakes when renegotiating using the function
2063 SSL_renegotiate_abbreviated().
2064 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2066 *) Add call to ENGINE_register_all_complete() to
2067 ENGINE_load_builtin_engines(), so some implementations get used
2068 automatically instead of needing explicit application support.
2071 *) Add support for TLS key exporter as described in RFC5705.
2072 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2074 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2075 a few changes are required:
2077 Add SSL_OP_NO_TLSv1_1 flag.
2078 Add TLSv1_1 methods.
2079 Update version checking logic to handle version 1.1.
2080 Add explicit IV handling (ported from DTLS code).
2081 Add command line options to s_client/s_server.
2084 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2086 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2087 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2088 content decryption and always return the same error. Note: this attack
2089 needs on average 2^20 messages so it only affects automated senders. The
2090 old behaviour can be reenabled in the CMS code by setting the
2091 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2092 an MMA defence is not necessary.
2093 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2094 this issue. (CVE-2012-0884)
2097 *) Fix CVE-2011-4619: make sure we really are receiving a
2098 client hello before rejecting multiple SGC restarts. Thanks to
2099 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2102 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2104 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2105 Thanks to Antonio Martin, Enterprise Secure Access Research and
2106 Development, Cisco Systems, Inc. for discovering this bug and
2107 preparing a fix. (CVE-2012-0050)
2110 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2112 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2113 of the Vaudenay padding oracle attack on CBC mode encryption
2114 which enables an efficient plaintext recovery attack against
2115 the OpenSSL implementation of DTLS. Their attack exploits timing
2116 differences arising during decryption processing. A research
2117 paper describing this attack can be found at:
2118 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2119 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2120 Security Group at Royal Holloway, University of London
2121 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2122 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2123 for preparing the fix. (CVE-2011-4108)
2124 [Robin Seggelmann, Michael Tuexen]
2126 *) Clear bytes used for block padding of SSL 3.0 records.
2128 [Adam Langley (Google)]
2130 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2131 Kadianakis <desnacked@gmail.com> for discovering this issue and
2132 Adam Langley for preparing the fix. (CVE-2011-4619)
2133 [Adam Langley (Google)]
2135 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2136 [Andrey Kulikov <amdeich@gmail.com>]
2138 *) Prevent malformed RFC3779 data triggering an assertion failure.
2139 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2140 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2141 [Rob Austein <sra@hactrn.net>]
2143 *) Improved PRNG seeding for VOS.
2144 [Paul Green <Paul.Green@stratus.com>]
2146 *) Fix ssl_ciph.c set-up race.
2147 [Adam Langley (Google)]
2149 *) Fix spurious failures in ecdsatest.c.
2150 [Emilia Käsper (Google)]
2152 *) Fix the BIO_f_buffer() implementation (which was mixing different
2153 interpretations of the '..._len' fields).
2154 [Adam Langley (Google)]
2156 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2157 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2158 threads won't reuse the same blinding coefficients.
2160 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2161 lock to call BN_BLINDING_invert_ex, and avoids one use of
2162 BN_BLINDING_update for each BN_BLINDING structure (previously,
2163 the last update always remained unused).
2164 [Emilia Käsper (Google)]
2166 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2167 [Bob Buckholz (Google)]
2169 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2171 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2172 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2173 [Kaspar Brand <ossl@velox.ch>]
2175 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2176 for multi-threaded use of ECDH. (CVE-2011-3210)
2177 [Adam Langley (Google)]
2179 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2182 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2183 signature public key algorithm by using OID xref utilities instead.
2184 Before this you could only use some ECC ciphersuites with SHA1 only.
2187 *) Add protection against ECDSA timing attacks as mentioned in the paper
2188 by Billy Bob Brumley and Nicola Tuveri, see:
2190 http://eprint.iacr.org/2011/232.pdf
2192 [Billy Bob Brumley and Nicola Tuveri]
2194 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2196 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2197 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2199 *) Fix bug in string printing code: if *any* escaping is enabled we must
2200 escape the escape character (backslash) or the resulting string is
2204 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2206 *) Disable code workaround for ancient and obsolete Netscape browsers
2207 and servers: an attacker can use it in a ciphersuite downgrade attack.
2208 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2211 *) Fixed J-PAKE implementation error, originally discovered by
2212 Sebastien Martini, further info and confirmation from Stefan
2213 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2216 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2218 *) Fix extension code to avoid race conditions which can result in a buffer
2219 overrun vulnerability: resumed sessions must not be modified as they can
2220 be shared by multiple threads. CVE-2010-3864
2223 *) Fix WIN32 build system to correctly link an ENGINE directory into
2227 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2229 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2231 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2233 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2235 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2236 context. The operation can be customised via the ctrl mechanism in
2237 case ENGINEs want to include additional functionality.
2240 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2243 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2244 output hashes compatible with older versions of OpenSSL.
2245 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2247 *) Fix compression algorithm handling: if resuming a session use the
2248 compression algorithm of the resumed session instead of determining
2249 it from client hello again. Don't allow server to change algorithm.
2252 *) Add load_crls() function to apps tidying load_certs() too. Add option
2253 to verify utility to allow additional CRLs to be included.
2256 *) Update OCSP request code to permit adding custom headers to the request:
2257 some responders need this.
2260 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2262 [Julia Lawall <julia@diku.dk>]
2264 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2265 needlessly dereferenced structures, used obsolete functions and
2266 didn't handle all updated verify codes correctly.
2269 *) Disable MD2 in the default configuration.
2272 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2273 indicate the initial BIO being pushed or popped. This makes it possible
2274 to determine whether the BIO is the one explicitly called or as a result
2275 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2276 it handles reference counts correctly and doesn't zero out the I/O bio
2277 when it is not being explicitly popped. WARNING: applications which
2278 included workarounds for the old buggy behaviour will need to be modified
2279 or they could free up already freed BIOs.
2282 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2283 renaming to all platforms (within the 0.9.8 branch, this was
2284 done conditionally on Netware platforms to avoid a name clash).
2285 [Guenter <lists@gknw.net>]
2287 *) Add ECDHE and PSK support to DTLS.
2288 [Michael Tuexen <tuexen@fh-muenster.de>]
2290 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2294 *) Add "missing" function EVP_MD_flags() (without this the only way to
2295 retrieve a digest flags is by accessing the structure directly. Update
2296 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2297 or cipher is registered as in the "from" argument. Print out all
2298 registered digests in the dgst usage message instead of manually
2299 attempting to work them out.
2302 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2303 this allows the use of compression and extensions. Change default cipher
2304 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2305 by default unless an application cipher string requests it.
2308 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2309 key ids to find matching certificates and keys but some PKCS#12 files
2310 don't follow the (somewhat unwritten) rules and this strategy fails.
2311 Now just gather all certificates together and the first private key
2312 then look for the first certificate that matches the key.
2315 *) Support use of registered digest and cipher names for dgst and cipher
2316 commands instead of having to add each one as a special case. So now
2323 openssl dgst -sha256 foo
2325 and this works for ENGINE based algorithms too.
2329 *) Update Gost ENGINE to support parameter files.
2330 [Victor B. Wagner <vitus@cryptocom.ru>]
2332 *) Support GeneralizedTime in ca utility.
2333 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2335 *) Enhance the hash format used for certificate directory links. The new
2336 form uses the canonical encoding (meaning equivalent names will work
2337 even if they aren't identical) and uses SHA1 instead of MD5. This form
2338 is incompatible with the older format and as a result c_rehash should
2339 be used to rebuild symbolic links.
2342 *) Make PKCS#8 the default write format for private keys, replacing the
2343 traditional format. This form is standardised, more secure and doesn't
2344 include an implicit MD5 dependency.
2347 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2348 committed to OpenSSL should pass this lot as a minimum.
2351 *) Add session ticket override functionality for use by EAP-FAST.
2352 [Jouni Malinen <j@w1.fi>]
2354 *) Modify HMAC functions to return a value. Since these can be implemented
2355 in an ENGINE errors can occur.
2358 *) Type-checked OBJ_bsearch_ex.
2361 *) Type-checked OBJ_bsearch. Also some constification necessitated
2362 by type-checking. Still to come: TXT_DB, bsearch(?),
2363 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2367 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2368 seconds to a tm structure directly, instead of going through OS
2369 specific date routines. This avoids any issues with OS routines such
2370 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2371 and X509_time_adj_ex() to cover the extended range. The existing
2372 X509_time_adj() is still usable and will no longer have any date issues.
2375 *) Delta CRL support. New use deltas option which will attempt to locate
2376 and search any appropriate delta CRLs available.
2378 This work was sponsored by Google.
2381 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2382 code and add additional score elements. Validate alternate CRL paths
2383 as part of the CRL checking and indicate a new error "CRL path validation
2384 error" in this case. Applications wanting additional details can use
2385 the verify callback and check the new "parent" field. If this is not
2386 NULL CRL path validation is taking place. Existing applications wont
2387 see this because it requires extended CRL support which is off by
2390 This work was sponsored by Google.
2393 *) Support for freshest CRL extension.
2395 This work was sponsored by Google.
2398 *) Initial indirect CRL support. Currently only supported in the CRLs
2399 passed directly and not via lookup. Process certificate issuer
2400 CRL entry extension and lookup CRL entries by bother issuer name
2401 and serial number. Check and process CRL issuer entry in IDP extension.
2403 This work was sponsored by Google.
2406 *) Add support for distinct certificate and CRL paths. The CRL issuer
2407 certificate is validated separately in this case. Only enabled if
2408 an extended CRL support flag is set: this flag will enable additional
2409 CRL functionality in future.
2411 This work was sponsored by Google.
2414 *) Add support for policy mappings extension.
2416 This work was sponsored by Google.
2419 *) Fixes to pathlength constraint, self issued certificate handling,
2420 policy processing to align with RFC3280 and PKITS tests.
2422 This work was sponsored by Google.
2425 *) Support for name constraints certificate extension. DN, email, DNS
2426 and URI types are currently supported.
2428 This work was sponsored by Google.
2431 *) To cater for systems that provide a pointer-based thread ID rather
2432 than numeric, deprecate the current numeric thread ID mechanism and
2433 replace it with a structure and associated callback type. This
2434 mechanism allows a numeric "hash" to be extracted from a thread ID in
2435 either case, and on platforms where pointers are larger than 'long',
2436 mixing is done to help ensure the numeric 'hash' is usable even if it
2437 can't be guaranteed unique. The default mechanism is to use "&errno"
2438 as a pointer-based thread ID to distinguish between threads.
2440 Applications that want to provide their own thread IDs should now use
2441 CRYPTO_THREADID_set_callback() to register a callback that will call
2442 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2444 Note that ERR_remove_state() is now deprecated, because it is tied
2445 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2446 to free the current thread's error state should be replaced by
2447 ERR_remove_thread_state(NULL).
2449 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2450 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2451 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2452 application was previously providing a numeric thread callback that
2453 was inappropriate for distinguishing threads, then uniqueness might
2454 have been obtained with &errno that happened immediately in the
2455 intermediate development versions of OpenSSL; this is no longer the
2456 case, the numeric thread callback will now override the automatic use
2458 [Geoff Thorpe, with help from Bodo Moeller]
2460 *) Initial support for different CRL issuing certificates. This covers a
2461 simple case where the self issued certificates in the chain exist and
2462 the real CRL issuer is higher in the existing chain.
2464 This work was sponsored by Google.
2467 *) Removed effectively defunct crypto/store from the build.
2470 *) Revamp of STACK to provide stronger type-checking. Still to come:
2471 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2472 ASN1_STRING, CONF_VALUE.
2475 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2476 RAM on SSL connections. This option can save about 34k per idle SSL.
2479 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2480 STACK, TXT_DB, bsearch, qsort.
2483 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2484 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2485 support for data, signedData, compressedData, digestedData and
2486 encryptedData, envelopedData types included. Scripts to check against
2487 RFC4134 examples draft and interop and consistency checks of many
2488 content types and variants.
2491 *) Add options to enc utility to support use of zlib compression BIO.
2494 *) Extend mk1mf to support importing of options and assembly language
2495 files from Configure script, currently only included in VC-WIN32.
2496 The assembly language rules can now optionally generate the source
2497 files from the associated perl scripts.
2500 *) Implement remaining functionality needed to support GOST ciphersuites.
2501 Interop testing has been performed using CryptoPro implementations.
2502 [Victor B. Wagner <vitus@cryptocom.ru>]
2504 *) s390x assembler pack.
2507 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2511 *) Implement Opaque PRF Input TLS extension as specified in
2512 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2513 official specification yet and no extension type assignment by
2514 IANA exists, this extension (for now) will have to be explicitly
2515 enabled when building OpenSSL by providing the extension number
2516 to use. For example, specify an option
2518 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2520 to the "config" or "Configure" script to enable the extension,
2521 assuming extension number 0x9527 (which is a completely arbitrary
2522 and unofficial assignment based on the MD5 hash of the Internet
2523 Draft). Note that by doing so, you potentially lose
2524 interoperability with other TLS implementations since these might
2525 be using the same extension number for other purposes.
2527 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2528 opaque PRF input value to use in the handshake. This will create
2529 an interal copy of the length-'len' string at 'src', and will
2530 return non-zero for success.
2532 To get more control and flexibility, provide a callback function
2535 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2536 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2540 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2543 Callback function 'cb' will be called in handshakes, and is
2544 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2545 Argument 'arg' is for application purposes (the value as given to
2546 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2547 be provided to the callback function). The callback function
2548 has to return non-zero to report success: usually 1 to use opaque
2549 PRF input just if possible, or 2 to enforce use of the opaque PRF
2550 input. In the latter case, the library will abort the handshake
2551 if opaque PRF input is not successfully negotiated.
2553 Arguments 'peerinput' and 'len' given to the callback function
2554 will always be NULL and 0 in the case of a client. A server will
2555 see the client's opaque PRF input through these variables if
2556 available (NULL and 0 otherwise). Note that if the server
2557 provides an opaque PRF input, the length must be the same as the
2558 length of the client's opaque PRF input.
2560 Note that the callback function will only be called when creating
2561 a new session (session resumption can resume whatever was
2562 previously negotiated), and will not be called in SSL 2.0
2563 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2564 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2565 for applications that need to enforce opaque PRF input.
2569 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2572 [Victor B. Wagner <vitus@cryptocom.ru>]
2574 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2575 RFC4507bis. The encrypted ticket format is an encrypted encoded
2576 SSL_SESSION structure, that way new session features are automatically
2579 If a client application caches session in an SSL_SESSION structure
2580 support is transparent because tickets are now stored in the encoded
2583 The SSL_CTX structure automatically generates keys for ticket
2584 protection in servers so again support should be possible
2585 with no application modification.
2587 If a client or server wishes to disable RFC4507 support then the option
2588 SSL_OP_NO_TICKET can be set.
2590 Add a TLS extension debugging callback to allow the contents of any client
2591 or server extensions to be examined.
2593 This work was sponsored by Google.
2596 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2597 OpenSSL should now compile cleanly on gcc 4.2
2598 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2600 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2601 support including streaming MAC support: this is required for GOST
2602 ciphersuite support.
2603 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2605 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2606 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2607 to output in BER and PEM format.
2610 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2611 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2612 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2613 ENGINE support for HMAC keys which are unextractable. New -mac and
2614 -macopt options to dgst utility.
2617 *) New option -sigopt to dgst utility. Update dgst to use
2618 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2619 alternative signing paramaters such as X9.31 or PSS in the dgst
2623 *) Change ssl_cipher_apply_rule(), the internal function that does
2624 the work each time a ciphersuite string requests enabling
2625 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2626 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2627 the order of disabled ciphersuites such that those ciphersuites
2628 that most recently went from enabled to disabled not only stay
2629 in order with respect to each other, but also have higher priority
2630 than other disabled ciphersuites the next time ciphersuites are
2633 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2634 the same ciphersuites as with "HIGH" alone, but in a specific
2635 order where the PSK ciphersuites come first (since they are the
2636 most recently disabled ciphersuites when "HIGH" is parsed).
2638 Also, change ssl_create_cipher_list() (using this new
2639 funcionality) such that between otherwise identical
2640 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2644 *) Change ssl_create_cipher_list() so that it automatically
2645 arranges the ciphersuites in reasonable order before starting
2646 to process the rule string. Thus, the definition for "DEFAULT"
2647 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2648 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2649 This makes it much easier to arrive at a reasonable default order
2650 in applications for which anonymous ciphers are OK (meaning
2651 that you can't actually use DEFAULT).
2652 [Bodo Moeller; suggested by Victor Duchovni]
2654 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2655 processing) into multiple integers instead of setting
2656 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2657 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2658 (These masks as well as the individual bit definitions are hidden
2659 away into the non-exported interface ssl/ssl_locl.h, so this
2660 change to the definition of the SSL_CIPHER structure shouldn't
2661 affect applications.) This give us more bits for each of these
2662 categories, so there is no longer a need to coagulate AES128 and
2663 AES256 into a single algorithm bit, and to coagulate Camellia128
2664 and Camellia256 into a single algorithm bit, which has led to all
2667 Thus, among other things, the kludge introduced in 0.9.7m and
2668 0.9.8e for masking out AES256 independently of AES128 or masking
2669 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2671 With the change, we also introduce new ciphersuite aliases that
2672 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2676 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2677 Use the leftmost N bytes of the signature input if the input is
2678 larger than the prime q (with N being the size in bytes of q).
2681 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2682 it yet and it is largely untested.
2685 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2688 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2689 some compilers (gcc 4.2 and later) reject their use. Safestack is
2690 reimplemented. Update ASN1 to avoid use of legacy functions.
2693 *) Win32/64 targets are linked with Winsock2.
2696 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2697 to external functions. This can be used to increase CRL handling
2698 efficiency especially when CRLs are very large by (for example) storing
2699 the CRL revoked certificates in a database.
2702 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2703 new CRLs added to a directory can be used. New command line option
2704 -verify_return_error to s_client and s_server. This causes real errors
2705 to be returned by the verify callback instead of carrying on no matter
2706 what. This reflects the way a "real world" verify callback would behave.
2709 *) GOST engine, supporting several GOST algorithms and public key formats.
2710 Kindly donated by Cryptocom.
2713 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2714 partitioned by DP are handled but no indirect CRL or reason partitioning
2715 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2716 selected via a scoring technique which handles IDP and AKID in CRLs.
2719 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2720 will ultimately be used for all verify operations: this will remove the
2721 X509_STORE dependency on certificate verification and allow alternative
2722 lookup methods. X509_STORE based implementations of these two callbacks.
2725 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2726 Modify get_crl() to find a valid (unexpired) CRL if possible.
2729 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2730 this would be called X509_CRL_cmp() but that name is already used by
2731 a function that just compares CRL issuer names. Cache several CRL
2732 extensions in X509_CRL structure and cache CRLDP in X509.
2735 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2736 this maps equivalent X509_NAME structures into a consistent structure.
2737 Name comparison can then be performed rapidly using memcmp().
2740 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2744 *) Allow digests to supply their own micalg string for S/MIME type using
2745 the ctrl EVP_MD_CTRL_MICALG.
2748 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2749 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2750 ctrl. It can then customise the structure before and/or after signing
2754 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2755 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2756 to free up any added signature OIDs.
2759 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2760 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2761 digest and cipher tables. New options added to openssl utility:
2762 list-message-digest-algorithms and list-cipher-algorithms.
2765 *) Change the array representation of binary polynomials: the list
2766 of degrees of non-zero coefficients is now terminated with -1.
2767 Previously it was terminated with 0, which was also part of the
2768 value; thus, the array representation was not applicable to
2769 polynomials where t^0 has coefficient zero. This change makes
2770 the array representation useful in a more general context.
2773 *) Various modifications and fixes to SSL/TLS cipher string
2774 handling. For ECC, the code now distinguishes between fixed ECDH
2775 with RSA certificates on the one hand and with ECDSA certificates
2776 on the other hand, since these are separate ciphersuites. The
2777 unused code for Fortezza ciphersuites has been removed.
2779 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2780 (not "ECDHE"). For consistency with the code for DH
2781 certificates, use of ECDH certificates is now considered ECDH
2782 authentication, not RSA or ECDSA authentication (the latter is
2783 merely the CA's signing algorithm and not actively used in the
2786 The temporary ciphersuite alias "ECCdraft" is no longer
2787 available, and ECC ciphersuites are no longer excluded from "ALL"
2788 and "DEFAULT". The following aliases now exist for RFC 4492
2789 ciphersuites, most of these by analogy with the DH case:
2791 kECDHr - ECDH cert, signed with RSA
2792 kECDHe - ECDH cert, signed with ECDSA
2793 kECDH - ECDH cert (signed with either RSA or ECDSA)
2794 kEECDH - ephemeral ECDH
2795 ECDH - ECDH cert or ephemeral ECDH
2801 AECDH - anonymous ECDH
2802 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2806 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2807 Use correct micalg parameters depending on digest(s) in signed message.
2810 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2811 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2814 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2815 an engine to register a method. Add ENGINE lookups for methods and
2816 functional reference processing.
2819 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2820 EVP_{Sign,Verify}* which allow an application to customise the signature
2824 *) New -resign option to smime utility. This adds one or more signers
2825 to an existing PKCS#7 signedData structure. Also -md option to use an
2826 alternative message digest algorithm for signing.
2829 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2830 create PKCS7 structures containing multiple signers. Update smime
2831 application to support multiple signers.
2834 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2838 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2839 Reorganize PBE internals to lookup from a static table using NIDs,
2840 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2841 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2842 PRF which will be automatically used with PBES2.
2845 *) Replace the algorithm specific calls to generate keys in "req" with the
2849 *) Update PKCS#7 enveloped data routines to use new API. This is now
2850 supported by any public key method supporting the encrypt operation. A
2851 ctrl is added to allow the public key algorithm to examine or modify
2852 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2856 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2857 a default digest type to use. In most cases this will be SHA1 but some
2858 algorithms (such as GOST) need to specify an alternative digest. The
2859 return value indicates how strong the preference is 1 means optional and
2860 2 is mandatory (that is it is the only supported type). Modify
2861 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2862 use the default md. Update openssl utilities to use the default digest
2863 type for signing if it is not explicitly indicated.
2866 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2867 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2868 signing method from the key type. This effectively removes the link
2869 between digests and public key types.
2872 *) Add an OID cross reference table and utility functions. Its purpose is to
2873 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2874 rsaEncryption. This will allow some of the algorithm specific hackery
2875 needed to use the correct OID to be removed.
2878 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2879 structures for PKCS7_sign(). They are now set up by the relevant public
2883 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2886 *) Add support for key derivation (agreement) in the API, DH method and
2890 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2891 public and private key formats. As a side effect these add additional
2892 command line functionality not previously available: DSA signatures can be
2893 generated and verified using pkeyutl and DH key support and generation in
2898 [Oliver Tappe <zooey@hirschkaefer.de>]
2900 *) New make target "install_html_docs" installs HTML renditions of the
2902 [Oliver Tappe <zooey@hirschkaefer.de>]
2904 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2905 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2906 support key and parameter generation and add initial key generation
2907 functionality for RSA.
2910 *) Add functions for main EVP_PKEY_method operations. The undocumented
2911 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2912 EVP_PKEY_{encrypt,decrypt}_old.
2915 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2916 key API, doesn't do much yet.
2919 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2920 public key algorithms. New option to openssl utility:
2921 "list-public-key-algorithms" to print out info.
2924 *) Implement the Supported Elliptic Curves Extension for
2925 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2928 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2929 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2932 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2933 utilities such as rsa, dsa, dsaparam etc except they process any key
2937 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2938 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2939 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2943 *) Initial support for pluggable public key ASN1.
2944 De-spaghettify the public key ASN1 handling. Move public and private
2945 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2946 algorithm specific handling to a single module within the relevant
2947 algorithm directory. Add functions to allow (near) opaque processing
2948 of public and private key structures.
2951 *) Implement the Supported Point Formats Extension for
2952 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2955 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2956 for the psk identity [hint] and the psk callback functions to the
2957 SSL_SESSION, SSL and SSL_CTX structure.
2960 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2964 SSL_CTX_use_psk_identity_hint
2965 SSL_get_psk_identity_hint
2966 SSL_get_psk_identity
2967 SSL_use_psk_identity_hint
2969 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2971 *) Add RFC 3161 compliant time stamp request creation, response generation
2972 and response verification functionality.
2973 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2975 *) Add initial support for TLS extensions, specifically for the server_name
2976 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2977 have new members for a host name. The SSL data structure has an
2978 additional member SSL_CTX *initial_ctx so that new sessions can be
2979 stored in that context to allow for session resumption, even after the
2980 SSL has been switched to a new SSL_CTX in reaction to a client's
2981 server_name extension.
2983 New functions (subject to change):
2985 SSL_get_servername()
2986 SSL_get_servername_type()
2989 New CTRL codes and macros (subject to change):
2991 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2992 - SSL_CTX_set_tlsext_servername_callback()
2993 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2994 - SSL_CTX_set_tlsext_servername_arg()
2995 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2997 openssl s_client has a new '-servername ...' option.
2999 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3000 '-key2 ...', '-servername_fatal' (subject to change). This allows
3001 testing the HostName extension for a specific single host name ('-cert'
3002 and '-key' remain fallbacks for handshakes without HostName
3003 negotiation). If the unrecognized_name alert has to be sent, this by
3004 default is a warning; it becomes fatal with the '-servername_fatal'
3007 [Peter Sylvester, Remy Allais, Christophe Renou]
3009 *) Whirlpool hash implementation is added.
3012 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3013 bn(64,32). Because of instruction set limitations it doesn't have
3014 any negative impact on performance. This was done mostly in order
3015 to make it possible to share assembler modules, such as bn_mul_mont
3016 implementations, between 32- and 64-bit builds without hassle.
3019 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3020 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3024 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3025 dedicated Montgomery multiplication procedure, is introduced.
3026 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3027 "64-bit" performance on certain 32-bit targets.
3030 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3031 in SSL structures. New SSL ctrl to set maximum send fragment size.
3032 Save memory by seeting the I/O buffer sizes dynamically instead of
3033 using the maximum available value.
3036 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3037 in addition to the text details.
3040 *) Very, very preliminary EXPERIMENTAL support for printing of general
3041 ASN1 structures. This currently produces rather ugly output and doesn't
3042 handle several customised structures at all.
3045 *) Integrated support for PVK file format and some related formats such
3046 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3047 these in the 'rsa' and 'dsa' utilities.
3050 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3053 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3054 place for the (very old) "NETSCAPE" format certificates which are now
3055 handled using new ASN1 code equivalents.
3058 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3059 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3060 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3063 *) Modify CRL distribution points extension code to print out previously
3064 unsupported fields. Enhance extension setting code to allow setting of
3068 *) Add print and set support for Issuing Distribution Point CRL extension.
3071 *) Change 'Configure' script to enable Camellia by default.
3074 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3076 *) When rejecting SSL/TLS records due to an incorrect version number, never
3077 update s->server with a new major version number. As of
3078 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3079 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3080 the previous behavior could result in a read attempt at NULL when
3081 receiving specific incorrect SSL/TLS records once record payload
3082 protection is active. (CVE-2010-0740)
3083 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3085 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3086 could be crashed if the relevant tables were not present (e.g. chrooted).
3087 [Tomas Hoger <thoger@redhat.com>]
3089 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3091 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3092 [Martin Olsson, Neel Mehta]
3094 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3095 accommodate for stack sorting, always a write lock!).
3098 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3099 excessive delays in the RAND_poll(): over a minute. As a workaround
3100 include a time check in the inner Heap32Next loop too.
3103 *) The code that handled flushing of data in SSL/TLS originally used the
3104 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3105 the problem outlined in PR#1949. The fix suggested there however can
3106 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3107 of Apache). So instead simplify the code to flush unconditionally.
3108 This should be fine since flushing with no data to flush is a no op.
3111 *) Handle TLS versions 2.0 and later properly and correctly use the
3112 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3113 off ancient servers have a habit of sticking around for a while...
3116 *) Modify compression code so it frees up structures without using the
3117 ex_data callbacks. This works around a problem where some applications
3118 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3119 restarting) then use compression (e.g. SSL with compression) later.
3120 This results in significant per-connection memory leaks and
3121 has caused some security issues including CVE-2008-1678 and
3125 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3126 change when encrypting or decrypting.
3129 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3130 connect and renegotiate with servers which do not support RI.
3131 Until RI is more widely deployed this option is enabled by default.
3134 *) Add "missing" ssl ctrls to clear options and mode.
3137 *) If client attempts to renegotiate and doesn't support RI respond with
3138 a no_renegotiation alert as required by RFC5746. Some renegotiating
3139 TLS clients will continue a connection gracefully when they receive
3140 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3141 waiting for a server hello which it will never receive. Now we treat a
3142 received no_renegotiation alert as a fatal error. This is because
3143 applications requesting a renegotiation might well expect it to succeed
3144 and would have no code in place to handle the server denying it so the
3145 only safe thing to do is to terminate the connection.
3148 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3149 peer supports secure renegotiation and 0 otherwise. Print out peer
3150 renegotiation support in s_client/s_server.
3153 *) Replace the highly broken and deprecated SPKAC certification method with
3154 the updated NID creation version. This should correctly handle UTF8.
3157 *) Implement RFC5746. Re-enable renegotiation but require the extension
3158 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3159 turns out to be a bad idea. It has been replaced by
3160 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3161 SSL_CTX_set_options(). This is really not recommended unless you
3162 know what you are doing.
3163 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3165 *) Fixes to stateless session resumption handling. Use initial_ctx when
3166 issuing and attempting to decrypt tickets in case it has changed during
3167 servername handling. Use a non-zero length session ID when attempting
3168 stateless session resumption: this makes it possible to determine if
3169 a resumption has occurred immediately after receiving server hello
3170 (several places in OpenSSL subtly assume this) instead of later in
3174 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3175 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3176 fixes for a few places where the return code is not checked
3178 [Julia Lawall <julia@diku.dk>]
3180 *) Add --strict-warnings option to Configure script to include devteam
3181 warnings in other configurations.
3184 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3185 makes it possible to install openssl libraries in locations which
3186 have names other than "lib", for example "/usr/lib64" which some
3188 [Steve Henson, based on patch from Jeremy Utley]
3190 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3191 X690 8.9.12 and can produce some misleading textual output of OIDs.
3192 [Steve Henson, reported by Dan Kaminsky]
3194 *) Delete MD2 from algorithm tables. This follows the recommendation in
3195 several standards that it is not used in new applications due to
3196 several cryptographic weaknesses. For binary compatibility reasons
3197 the MD2 API is still compiled in by default.
3200 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3204 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3205 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3207 [Guenter <lists@gknw.net>]
3209 *) Fix the server certificate chain building code to use X509_verify_cert(),
3210 it used to have an ad-hoc builder which was unable to cope with anything
3211 other than a simple chain.
3212 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3214 *) Don't check self signed certificate signatures in X509_verify_cert()
3215 by default (a flag can override this): it just wastes time without
3216 adding any security. As a useful side effect self signed root CAs
3217 with non-FIPS digests are now usable in FIPS mode.
3220 *) In dtls1_process_out_of_seq_message() the check if the current message
3221 is already buffered was missing. For every new message was memory
3222 allocated, allowing an attacker to perform an denial of service attack
3223 with sending out of seq handshake messages until there is no memory
3224 left. Additionally every future messege was buffered, even if the
3225 sequence number made no sense and would be part of another handshake.
3226 So only messages with sequence numbers less than 10 in advance will be
3227 buffered. (CVE-2009-1378)
3228 [Robin Seggelmann, discovered by Daniel Mentz]
3230 *) Records are buffered if they arrive with a future epoch to be
3231 processed after finishing the corresponding handshake. There is
3232 currently no limitation to this buffer allowing an attacker to perform
3233 a DOS attack with sending records with future epochs until there is no
3234 memory left. This patch adds the pqueue_size() function to determine
3235 the size of a buffer and limits the record buffer to 100 entries.
3237 [Robin Seggelmann, discovered by Daniel Mentz]
3239 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3240 parent structure is freed. (CVE-2009-1379)
3243 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3244 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3247 [Ilya O. <vrghost@gmail.com>]
3249 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3251 *) Disable renegotiation completely - this fixes a severe security
3252 problem (CVE-2009-3555) at the cost of breaking all
3253 renegotiation. Renegotiation can be re-enabled by setting
3254 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3255 run-time. This is really not recommended unless you know what
3259 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3261 *) Don't set val to NULL when freeing up structures, it is freed up by
3262 underlying code. If sizeof(void *) > sizeof(long) this can result in
3263 zeroing past the valid field. (CVE-2009-0789)
3264 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3266 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3267 checked correctly. This would allow some invalid signed attributes to
3268 appear to verify correctly. (CVE-2009-0591)
3269 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3271 *) Reject UniversalString and BMPString types with invalid lengths. This
3272 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3273 a legal length. (CVE-2009-0590)
3276 *) Set S/MIME signing as the default purpose rather than setting it
3277 unconditionally. This allows applications to override it at the store
3281 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3282 to handle some structures.
3285 *) Improve efficiency of mem_gets: don't search whole buffer each time
3287 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3289 *) New -hex option for openssl rand.
3292 *) Print out UTF8String and NumericString when parsing ASN1.
3295 *) Support NumericString type for name components.
3298 *) Allow CC in the environment to override the automatically chosen
3299 compiler. Note that nothing is done to ensure flags work with the
3303 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3305 *) Properly check EVP_VerifyFinal() and similar return values
3307 [Ben Laurie, Bodo Moeller, Google Security Team]
3309 *) Enable TLS extensions by default.
3312 *) Allow the CHIL engine to be loaded, whether the application is
3313 multithreaded or not. (This does not release the developer from the
3314 obligation to set up the dynamic locking callbacks.)
3315 [Sander Temme <sander@temme.net>]
3317 *) Use correct exit code if there is an error in dgst command.
3318 [Steve Henson; problem pointed out by Roland Dirlewanger]
3320 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3321 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3324 *) Add experimental JPAKE support, including demo authentication in
3325 s_client and s_server.
3328 *) Set the comparison function in v3_addr_canonize().
3329 [Rob Austein <sra@hactrn.net>]
3331 *) Add support for XMPP STARTTLS in s_client.
3332 [Philip Paeps <philip@freebsd.org>]
3334 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3335 to ensure that even with this option, only ciphersuites in the
3336 server's preference list will be accepted. (Note that the option
3337 applies only when resuming a session, so the earlier behavior was
3338 just about the algorithm choice for symmetric cryptography.)
3341 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3343 *) Fix NULL pointer dereference if a DTLS server received
3344 ChangeCipherSpec as first record (CVE-2009-1386).
3347 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3348 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3351 *) The fix in 0.9.8c that supposedly got rid of unsafe
3352 double-checked locking was incomplete for RSA blinding,
3353 addressing just one layer of what turns out to have been
3354 doubly unsafe triple-checked locking.
3356 So now fix this for real by retiring the MONT_HELPER macro
3357 in crypto/rsa/rsa_eay.c.
3359 [Bodo Moeller; problem pointed out by Marius Schilder]
3361 *) Various precautionary measures:
3363 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3365 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3366 (NB: This would require knowledge of the secret session ticket key
3367 to exploit, in which case you'd be SOL either way.)
3369 - Change bn_nist.c so that it will properly handle input BIGNUMs
3370 outside the expected range.
3372 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3375 [Neel Mehta, Bodo Moeller]
3377 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3378 the load fails. Useful for distros.
3379 [Ben Laurie and the FreeBSD team]
3381 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3384 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3387 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3389 This work was sponsored by Logica.
3392 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3393 keystores. Support for SSL/TLS client authentication too.
3394 Not compiled unless enable-capieng specified to Configure.
3396 This work was sponsored by Logica.
3399 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3400 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3401 attribute creation routines such as certifcate requests and PKCS#12
3405 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3407 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3408 handshake which could lead to a cilent crash as found using the
3409 Codenomicon TLS test suite (CVE-2008-1672)
3410 [Steve Henson, Mark Cox]
3412 *) Fix double free in TLS server name extensions which could lead to
3413 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3416 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3418 Clear the error queue to ensure that error entries left from
3419 older function calls do not interfere with the correct operation.
3420 [Lutz Jaenicke, Erik de Castro Lopo]
3422 *) Remove root CA certificates of commercial CAs:
3424 The OpenSSL project does not recommend any specific CA and does not
3425 have any policy with respect to including or excluding any CA.
3426 Therefore it does not make any sense to ship an arbitrary selection
3427 of root CA certificates with the OpenSSL software.
3430 *) RSA OAEP patches to fix two separate invalid memory reads.
3431 The first one involves inputs when 'lzero' is greater than
3432 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3433 before the beginning of from). The second one involves inputs where
3434 the 'db' section contains nothing but zeroes (there is a one-byte
3435 invalid read after the end of 'db').
3436 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3438 *) Partial backport from 0.9.9-dev:
3440 Introduce bn_mul_mont (dedicated Montgomery multiplication
3441 procedure) as a candidate for BIGNUM assembler implementation.
3442 While 0.9.9-dev uses assembler for various architectures, only
3443 x86_64 is available by default here in the 0.9.8 branch, and
3444 32-bit x86 is available through a compile-time setting.
3446 To try the 32-bit x86 assembler implementation, use Configure
3447 option "enable-montasm" (which exists only for this backport).
3449 As "enable-montasm" for 32-bit x86 disclaims code stability
3450 anyway, in this constellation we activate additional code
3451 backported from 0.9.9-dev for further performance improvements,
3452 namely BN_from_montgomery_word. (To enable this otherwise,
3453 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3455 [Andy Polyakov (backport partially by Bodo Moeller)]
3457 *) Add TLS session ticket callback. This allows an application to set
3458 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3459 values. This is useful for key rollover for example where several key
3460 sets may exist with different names.
3463 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3464 This was broken until now in 0.9.8 releases, such that the only way
3465 a registered ENGINE could be used (assuming it initialises
3466 successfully on the host) was to explicitly set it as the default
3467 for the relevant algorithms. This is in contradiction with 0.9.7
3468 behaviour and the documentation. With this fix, when an ENGINE is
3469 registered into a given algorithm's table of implementations, the
3470 'uptodate' flag is reset so that auto-discovery will be used next
3471 time a new context for that algorithm attempts to select an
3473 [Ian Lister (tweaked by Geoff Thorpe)]
3475 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3476 implemention in the following ways:
3478 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3481 Lack of BER streaming support means one pass streaming processing is
3482 only supported if data is detached: setting the streaming flag is
3483 ignored for embedded content.
3485 CMS support is disabled by default and must be explicitly enabled
3486 with the enable-cms configuration option.
3489 *) Update the GMP engine glue to do direct copies between BIGNUM and
3490 mpz_t when openssl and GMP use the same limb size. Otherwise the
3491 existing "conversion via a text string export" trick is still used.
3492 [Paul Sheer <paulsheer@gmail.com>]
3494 *) Zlib compression BIO. This is a filter BIO which compressed and
3495 uncompresses any data passed through it.
3498 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3499 RFC3394 compatible AES key wrapping.
3502 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3503 sets string data without copying. X509_ALGOR_set0() and
3504 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3505 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3506 from an X509_ATTRIBUTE structure optionally checking it occurs only
3507 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3511 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3512 to get the expected BN_FLG_CONSTTIME behavior.
3513 [Bodo Moeller (Google)]
3517 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3518 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3519 - added some more tests to do_tests.pl
3520 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3521 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3522 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3523 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3524 - various changes to netware.pl to enable gcc-cross builds on Win32
3526 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3527 - various changes to fix missing prototype warnings
3528 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3529 - added AES, WHIRLPOOL and CPUID assembler code to build files
3530 - added missing AES assembler make rules to mk1mf.pl
3531 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3532 [Guenter Knauf <eflash@gmx.net>]
3534 *) Implement certificate status request TLS extension defined in RFC3546.
3535 A client can set the appropriate parameters and receive the encoded
3536 OCSP response via a callback. A server can query the supplied parameters
3537 and set the encoded OCSP response in the callback. Add simplified examples
3538 to s_client and s_server.
3541 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3543 *) Fix various bugs:
3544 + Binary incompatibility of ssl_ctx_st structure
3545 + DTLS interoperation with non-compliant servers
3546 + Don't call get_session_cb() without proposed session
3547 + Fix ia64 assembler code
3548 [Andy Polyakov, Steve Henson]
3550 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3552 *) DTLS Handshake overhaul. There were longstanding issues with
3553 OpenSSL DTLS implementation, which were making it impossible for
3554 RFC 4347 compliant client to communicate with OpenSSL server.
3555 Unfortunately just fixing these incompatibilities would "cut off"
3556 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3557 server keeps tolerating non RFC compliant syntax. The opposite is
3558 not true, 0.9.8f client can not communicate with earlier server.
3559 This update even addresses CVE-2007-4995.
3562 *) Changes to avoid need for function casts in OpenSSL: some compilers
3563 (gcc 4.2 and later) reject their use.
3564 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3567 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3568 RFC4507bis. The encrypted ticket format is an encrypted encoded
3569 SSL_SESSION structure, that way new session features are automatically
3572 If a client application caches session in an SSL_SESSION structure
3573 support is transparent because tickets are now stored in the encoded
3576 The SSL_CTX structure automatically generates keys for ticket
3577 protection in servers so again support should be possible
3578 with no application modification.
3580 If a client or server wishes to disable RFC4507 support then the option
3581 SSL_OP_NO_TICKET can be set.
3583 Add a TLS extension debugging callback to allow the contents of any client
3584 or server extensions to be examined.
3586 This work was sponsored by Google.
3589 *) Add initial support for TLS extensions, specifically for the server_name
3590 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3591 have new members for a host name. The SSL data structure has an
3592 additional member SSL_CTX *initial_ctx so that new sessions can be
3593 stored in that context to allow for session resumption, even after the
3594 SSL has been switched to a new SSL_CTX in reaction to a client's
3595 server_name extension.
3597 New functions (subject to change):
3599 SSL_get_servername()
3600 SSL_get_servername_type()
3603 New CTRL codes and macros (subject to change):
3605 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3606 - SSL_CTX_set_tlsext_servername_callback()
3607 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3608 - SSL_CTX_set_tlsext_servername_arg()
3609 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3611 openssl s_client has a new '-servername ...' option.
3613 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3614 '-key2 ...', '-servername_fatal' (subject to change). This allows
3615 testing the HostName extension for a specific single host name ('-cert'
3616 and '-key' remain fallbacks for handshakes without HostName
3617 negotiation). If the unrecognized_name alert has to be sent, this by
3618 default is a warning; it becomes fatal with the '-servername_fatal'
3621 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3623 *) Add AES and SSE2 assembly language support to VC++ build.
3626 *) Mitigate attack on final subtraction in Montgomery reduction.
3629 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3630 (which previously caused an internal error).
3633 *) Squeeze another 10% out of IGE mode when in != out.
3636 *) AES IGE mode speedup.
3637 [Dean Gaudet (Google)]
3639 *) Add the Korean symmetric 128-bit cipher SEED (see
3640 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3641 add SEED ciphersuites from RFC 4162:
3643 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3644 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3645 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3646 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3648 To minimize changes between patchlevels in the OpenSSL 0.9.8
3649 series, SEED remains excluded from compilation unless OpenSSL
3650 is configured with 'enable-seed'.
3651 [KISA, Bodo Moeller]
3653 *) Mitigate branch prediction attacks, which can be practical if a
3654 single processor is shared, allowing a spy process to extract
3655 information. For detailed background information, see
3656 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3657 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3658 and Necessary Software Countermeasures"). The core of the change
3659 are new versions BN_div_no_branch() and
3660 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3661 respectively, which are slower, but avoid the security-relevant
3662 conditional branches. These are automatically called by BN_div()
3663 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3664 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3665 remove a conditional branch.
3667 BN_FLG_CONSTTIME is the new name for the previous
3668 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3669 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3670 in the exponent causes BN_mod_exp_mont() to use the alternative
3671 implementation in BN_mod_exp_mont_consttime().) The old name
3672 remains as a deprecated alias.
3674 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3675 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3676 constant-time implementations for more than just exponentiation.
3677 Here too the old name is kept as a deprecated alias.
3679 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3680 the BN_BLINDING structure gets an independent copy of the
3681 modulus. This means that the previous "BIGNUM *m" argument to
3682 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3683 essentially becomes "const BIGNUM *m", although we can't actually
3684 change this in the header file before 0.9.9. It allows
3685 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3686 enable BN_FLG_CONSTTIME.
3688 [Matthew D Wood (Intel Corp)]
3690 *) In the SSL/TLS server implementation, be strict about session ID
3691 context matching (which matters if an application uses a single
3692 external cache for different purposes). Previously,
3693 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3694 set. This did ensure strict client verification, but meant that,
3695 with applications using a single external cache for quite
3696 different requirements, clients could circumvent ciphersuite
3697 restrictions for a given session ID context by starting a session
3698 in a different context.
3701 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3702 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3703 authentication-only ciphersuites.
3706 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3707 not complete and could lead to a possible single byte overflow
3708 (CVE-2007-5135) [Ben Laurie]
3710 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3712 *) Since AES128 and AES256 (and similarly Camellia128 and
3713 Camellia256) share a single mask bit in the logic of
3714 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3715 kludge to work properly if AES128 is available and AES256 isn't
3716 (or if Camellia128 is available and Camellia256 isn't).
3719 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3720 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3721 When a point or a seed is encoded in a BIT STRING, we need to
3722 prevent the removal of trailing zero bits to get the proper DER
3723 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3724 of a NamedBitList, for which trailing 0 bits need to be removed.)
3727 *) Have SSL/TLS server implementation tolerate "mismatched" record
3728 protocol version while receiving ClientHello even if the
3729 ClientHello is fragmented. (The server can't insist on the
3730 particular protocol version it has chosen before the ServerHello
3731 message has informed the client about his choice.)
3734 *) Add RFC 3779 support.
3735 [Rob Austein for ARIN, Ben Laurie]
3737 *) Load error codes if they are not already present instead of using a
3738 static variable. This allows them to be cleanly unloaded and reloaded.
3739 Improve header file function name parsing.
3742 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3743 or CAPABILITY handshake as required by RFCs.
3746 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3748 *) Introduce limits to prevent malicious keys being able to
3749 cause a denial of service. (CVE-2006-2940)
3750 [Steve Henson, Bodo Moeller]
3752 *) Fix ASN.1 parsing of certain invalid structures that can result
3753 in a denial of service. (CVE-2006-2937) [Steve Henson]
3755 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3756 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3758 *) Fix SSL client code which could crash if connecting to a
3759 malicious SSLv2 server. (CVE-2006-4343)
3760 [Tavis Ormandy and Will Drewry, Google Security Team]
3762 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3763 match only those. Before that, "AES256-SHA" would be interpreted
3764 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3765 the same strength classification in 0.9.7h) as we currently only
3766 have a single AES bit in the ciphersuite description bitmap.
3767 That change, however, also applied to ciphersuite strings such as
3768 "RC4-MD5" that intentionally matched multiple ciphersuites --
3769 namely, SSL 2.0 ciphersuites in addition to the more common ones
3770 from SSL 3.0/TLS 1.0.
3772 So we change the selection algorithm again: Naming an explicit
3773 ciphersuite selects this one ciphersuite, and any other similar
3774 ciphersuite (same bitmap) from *other* protocol versions.
3775 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3776 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3778 Since SSL 2.0 does not have any ciphersuites for which the
3779 128/256 bit distinction would be relevant, this works for now.
3780 The proper fix will be to use different bits for AES128 and
3781 AES256, which would have avoided the problems from the beginning;
3782 however, bits are scarce, so we can only do this in a new release
3783 (not just a patchlevel) when we can change the SSL_CIPHER
3784 definition to split the single 'unsigned long mask' bitmap into
3785 multiple values to extend the available space.
3789 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3791 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3792 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3794 *) Add AES IGE and biIGE modes.
3797 *) Change the Unix randomness entropy gathering to use poll() when
3798 possible instead of select(), since the latter has some
3799 undesirable limitations.
3800 [Darryl Miles via Richard Levitte and Bodo Moeller]
3802 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3803 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3804 cannot be implicitly activated as part of, e.g., the "AES" alias.
3805 However, please upgrade to OpenSSL 0.9.9[-dev] for
3806 non-experimental use of the ECC ciphersuites to get TLS extension
3807 support, which is required for curve and point format negotiation
3808 to avoid potential handshake problems.
3811 *) Disable rogue ciphersuites:
3813 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3814 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3815 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3817 The latter two were purportedly from
3818 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3821 Also deactivate the remaining ciphersuites from
3822 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3823 unofficial, and the ID has long expired.
3826 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3827 dual-core machines) and other potential thread-safety issues.
3830 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3831 versions), which is now available for royalty-free use
3832 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3833 Also, add Camellia TLS ciphersuites from RFC 4132.
3835 To minimize changes between patchlevels in the OpenSSL 0.9.8
3836 series, Camellia remains excluded from compilation unless OpenSSL
3837 is configured with 'enable-camellia'.
3840 *) Disable the padding bug check when compression is in use. The padding
3841 bug check assumes the first packet is of even length, this is not
3842 necessarily true if compresssion is enabled and can result in false
3843 positives causing handshake failure. The actual bug test is ancient
3844 code so it is hoped that implementations will either have fixed it by
3845 now or any which still have the bug do not support compression.
3848 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3850 *) When applying a cipher rule check to see if string match is an explicit
3851 cipher suite and only match that one cipher suite if it is.
3854 *) Link in manifests for VC++ if needed.
3855 [Austin Ziegler <halostatue@gmail.com>]
3857 *) Update support for ECC-based TLS ciphersuites according to
3858 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3859 TLS extensions, which are supported starting with the 0.9.9
3860 branch, not in the OpenSSL 0.9.8 branch).
3863 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3864 opaque EVP_CIPHER_CTX handling.
3867 *) Fixes and enhancements to zlib compression code. We now only use
3868 "zlib1.dll" and use the default __cdecl calling convention on Win32
3869 to conform with the standards mentioned here:
3870 http://www.zlib.net/DLL_FAQ.txt
3871 Static zlib linking now works on Windows and the new --with-zlib-include
3872 --with-zlib-lib options to Configure can be used to supply the location
3873 of the headers and library. Gracefully handle case where zlib library
3877 *) Several fixes and enhancements to the OID generation code. The old code
3878 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3879 handle numbers larger than ULONG_MAX, truncated printing and had a
3880 non standard OBJ_obj2txt() behaviour.
3883 *) Add support for building of engines under engine/ as shared libraries
3884 under VC++ build system.
3887 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3888 Hopefully, we will not see any false combination of paths any more.
3891 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3893 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3894 (part of SSL_OP_ALL). This option used to disable the
3895 countermeasure against man-in-the-middle protocol-version
3896 rollback in the SSL 2.0 server implementation, which is a bad
3897 idea. (CVE-2005-2969)
3899 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3900 for Information Security, National Institute of Advanced Industrial
3901 Science and Technology [AIST], Japan)]
3903 *) Add two function to clear and return the verify parameter flags.
3906 *) Keep cipherlists sorted in the source instead of sorting them at
3907 runtime, thus removing the need for a lock.
3910 *) Avoid some small subgroup attacks in Diffie-Hellman.
3911 [Nick Mathewson and Ben Laurie]
3913 *) Add functions for well-known primes.
3916 *) Extended Windows CE support.
3917 [Satoshi Nakamura and Andy Polyakov]
3919 *) Initialize SSL_METHOD structures at compile time instead of during
3920 runtime, thus removing the need for a lock.
3923 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3924 attempting to decrypt each encrypted key in turn. Add support to
3928 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3930 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3933 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3936 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3937 key into the same file any more.
3940 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3943 *) Add -utf8 command line and config file option to 'ca'.
3944 [Stefan <stf@udoma.org]
3946 *) Removed the macro des_crypt(), as it seems to conflict with some
3947 libraries. Use DES_crypt().
3950 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3951 involves renaming the source and generated shared-libs for
3952 both. The engines will accept the corrected or legacy ids
3953 ('ncipher' and '4758_cca' respectively) when binding. NB,
3954 this only applies when building 'shared'.
3955 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3957 *) Add attribute functions to EVP_PKEY structure. Modify
3958 PKCS12_create() to recognize a CSP name attribute and
3959 use it. Make -CSP option work again in pkcs12 utility.
3962 *) Add new functionality to the bn blinding code:
3963 - automatic re-creation of the BN_BLINDING parameters after
3964 a fixed number of uses (currently 32)
3965 - add new function for parameter creation
3966 - introduce flags to control the update behaviour of the
3967 BN_BLINDING parameters
3968 - hide BN_BLINDING structure
3969 Add a second BN_BLINDING slot to the RSA structure to improve
3970 performance when a single RSA object is shared among several
3974 *) Add support for DTLS.
3975 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3977 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3978 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3981 *) Remove buggy and incomplete DH cert support from
3982 ssl/ssl_rsa.c and ssl/s3_both.c
3985 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3986 the apps/openssl applications.
3989 *) Compile clean with "-Wall -Wmissing-prototypes
3990 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3991 DEBUG_SAFESTACK must also be set.
3994 *) Change ./Configure so that certain algorithms can be disabled by default.
3995 The new counterpiece to "no-xxx" is "enable-xxx".
3997 The patented RC5 and MDC2 algorithms will now be disabled unless
3998 "enable-rc5" and "enable-mdc2", respectively, are specified.
4000 (IDEA remains enabled despite being patented. This is because IDEA
4001 is frequently required for interoperability, and there is no license
4002 fee for non-commercial use. As before, "no-idea" can be used to
4003 avoid this algorithm.)
4007 *) Add processing of proxy certificates (see RFC 3820). This work was
4008 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4009 EGEE (Enabling Grids for E-science in Europe).
4012 *) RC4 performance overhaul on modern architectures/implementations, such
4013 as Intel P4, IA-64 and AMD64.
4016 *) New utility extract-section.pl. This can be used specify an alternative
4017 section number in a pod file instead of having to treat each file as
4018 a separate case in Makefile. This can be done by adding two lines to the
4021 =for comment openssl_section:XXX
4023 The blank line is mandatory.
4027 *) New arguments -certform, -keyform and -pass for s_client and s_server
4028 to allow alternative format key and certificate files and passphrase
4032 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4033 update associated structures and add various utility functions.
4035 Add new policy related verify parameters, include policy checking in
4036 standard verify code. Enhance 'smime' application with extra parameters
4037 to support policy checking and print out.
4040 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4041 Nehemiah processors. These extensions support AES encryption in hardware
4042 as well as RNG (though RNG support is currently disabled).
4043 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4045 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4048 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4049 [Andy Polyakov and a number of other people]
4051 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4052 implementation contributed by IBM.
4053 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4055 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4056 exponent rather than 'unsigned long'. There is a corresponding change to
4057 the new 'rsa_keygen' element of the RSA_METHOD structure.
4058 [Jelte Jansen, Geoff Thorpe]
4060 *) Functionality for creating the initial serial number file is now
4061 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4063 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4064 number file to 1, which is bound to cause problems. To avoid
4065 the problems while respecting compatibility between different 0.9.7
4066 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4067 CA.pl for serial number initialization. With the new release 0.9.8,
4068 we can fix the problem directly in the 'ca' utility.)
4071 *) Reduced header interdepencies by declaring more opaque objects in
4072 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4073 give fewer recursive includes, which could break lazy source code - so
4074 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4075 developers should define this symbol when building and using openssl to
4076 ensure they track the recommended behaviour, interfaces, [etc], but
4077 backwards-compatible behaviour prevails when this isn't defined.
4080 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4083 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4084 This will generate a random key of the appropriate length based on the
4085 cipher context. The EVP_CIPHER can provide its own random key generation
4086 routine to support keys of a specific form. This is used in the des and
4087 3des routines to generate a key of the correct parity. Update S/MIME
4088 code to use new functions and hence generate correct parity DES keys.
4089 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4090 valid (weak or incorrect parity).
4093 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4094 as looking them up. This is useful when the verified structure may contain
4095 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4096 present unless the new PKCS7_NO_CRL flag is asserted.
4099 *) Extend ASN1 oid configuration module. It now additionally accepts the
4102 shortName = some long name, 1.2.3.4
4105 *) Reimplemented the BN_CTX implementation. There is now no more static
4106 limitation on the number of variables it can handle nor the depth of the
4107 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4108 information can now expand as required, and rather than having a single
4109 static array of bignums, BN_CTX now uses a linked-list of such arrays
4110 allowing it to expand on demand whilst maintaining the usefulness of
4111 BN_CTX's "bundling".
4114 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4115 to allow all RSA operations to function using a single BN_CTX.
4118 *) Preliminary support for certificate policy evaluation and checking. This
4119 is initially intended to pass the tests outlined in "Conformance Testing
4120 of Relying Party Client Certificate Path Processing Logic" v1.07.
4123 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4124 remained unused and not that useful. A variety of other little bignum
4125 tweaks and fixes have also been made continuing on from the audit (see
4129 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4130 associated ASN1, EVP and SSL functions and old ASN1 macros.
4133 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4134 and this should never fail. So the return value from the use of
4135 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4136 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4139 *) BN_CTX_get() should return zero-valued bignums, providing the same
4140 initialised value as BN_new().
4141 [Geoff Thorpe, suggested by Ulf Möller]
4143 *) Support for inhibitAnyPolicy certificate extension.
4146 *) An audit of the BIGNUM code is underway, for which debugging code is
4147 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4148 is considered valid when processing BIGNUMs, and causes execution to
4149 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4150 further steps are taken to deliberately pollute unused data in BIGNUM
4151 structures to try and expose faulty code further on. For now, openssl will
4152 (in its default mode of operation) continue to tolerate the inconsistent
4153 forms that it has tolerated in the past, but authors and packagers should
4154 consider trying openssl and their own applications when compiled with
4155 these debugging symbols defined. It will help highlight potential bugs in
4156 their own code, and will improve the test coverage for OpenSSL itself. At
4157 some point, these tighter rules will become openssl's default to improve
4158 maintainability, though the assert()s and other overheads will remain only
4159 in debugging configurations. See bn.h for more details.
4160 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4162 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4163 that can only be obtained through BN_CTX_new() (which implicitly
4164 initialises it). The presence of this function only made it possible
4165 to overwrite an existing structure (and cause memory leaks).
4168 *) Because of the callback-based approach for implementing LHASH as a
4169 template type, lh_insert() adds opaque objects to hash-tables and
4170 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4171 to clean up those corresponding objects before destroying the hash table
4172 (and losing the object pointers). So some over-zealous constifications in
4173 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4174 objects as "const" and the lh_doall[_arg] callback wrappers are not
4175 prototyped to have "const" restrictions on the object pointers they are
4176 given (and so aren't required to cast them away any more).
4179 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4180 (speed) prefers to use its own implementation. The two implementations
4181 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4182 its object type properly exposed (MS_TM) instead of casting to/from "char
4183 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4184 aren't necessarily the greatest nomenclatures - but this is what was used
4185 internally to the implementation so I've used that for now.
4188 *) Ensure that deprecated functions do not get compiled when
4189 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4190 the self-tests were still using deprecated key-generation functions so
4191 these have been updated also.
4194 *) Reorganise PKCS#7 code to separate the digest location functionality
4195 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4196 New function PKCS7_set_digest() to set the digest type for PKCS#7
4197 digestedData type. Add additional code to correctly generate the
4198 digestedData type and add support for this type in PKCS7 initialization
4202 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4203 structure of type "other".
4206 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4207 sure the loop does correctly stop and breaking ("division by zero")
4208 modulus operations are not performed. The (pre-generated) prime
4209 table crypto/bn/bn_prime.h was already correct, but it could not be
4210 re-generated on some platforms because of the "division by zero"
4211 situation in the script.
4212 [Ralf S. Engelschall]
4214 *) Update support for ECC-based TLS ciphersuites according to
4215 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4216 SHA-1 now is only used for "small" curves (where the
4217 representation of a field element takes up to 24 bytes); for
4218 larger curves, the field element resulting from ECDH is directly
4219 used as premaster secret.
4220 [Douglas Stebila (Sun Microsystems Laboratories)]
4222 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4223 curve secp160r1 to the tests.
4224 [Douglas Stebila (Sun Microsystems Laboratories)]
4226 *) Add the possibility to load symbols globally with DSO.
4227 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4229 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4230 control of the error stack.
4233 *) Add support for STORE in ENGINE.
4236 *) Add the STORE type. The intention is to provide a common interface
4237 to certificate and key stores, be they simple file-based stores, or
4238 HSM-type store, or LDAP stores, or...
4239 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4242 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4243 pass a list of arguments to any function as well as provide a way
4244 for a function to pass data back to the caller.
4247 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4248 works like BUF_strdup() but can be used to duplicate a portion of
4249 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4253 *) Add the function sk_find_ex() which works like sk_find(), but will
4254 return an index to an element even if an exact match couldn't be
4255 found. The index is guaranteed to point at the element where the
4256 searched-for key would be inserted to preserve sorting order.
4259 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4260 takes an extra flags argument for optional functionality. Currently,
4261 the following flags are defined:
4263 OBJ_BSEARCH_VALUE_ON_NOMATCH
4264 This one gets OBJ_bsearch_ex() to return a pointer to the first
4265 element where the comparing function returns a negative or zero
4268 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4269 This one gets OBJ_bsearch_ex() to return a pointer to the first
4270 element where the comparing function returns zero. This is useful
4271 if there are more than one element where the comparing function
4275 *) Make it possible to create self-signed certificates with 'openssl ca'
4276 in such a way that the self-signed certificate becomes part of the
4277 CA database and uses the same mechanisms for serial number generation
4278 as all other certificate signing. The new flag '-selfsign' enables
4279 this functionality. Adapt CA.sh and CA.pl.in.
4282 *) Add functionality to check the public key of a certificate request
4283 against a given private. This is useful to check that a certificate
4284 request can be signed by that key (self-signing).
4287 *) Make it possible to have multiple active certificates with the same
4288 subject in the CA index file. This is done only if the keyword
4289 'unique_subject' is set to 'no' in the main CA section (default
4290 if 'CA_default') of the configuration file. The value is saved
4291 with the database itself in a separate index attribute file,
4292 named like the index file with '.attr' appended to the name.
4295 *) Generate muti valued AVAs using '+' notation in config files for
4299 *) Support for nameConstraints certificate extension.
4302 *) Support for policyConstraints certificate extension.
4305 *) Support for policyMappings certificate extension.
4308 *) Make sure the default DSA_METHOD implementation only uses its
4309 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4310 and change its own handlers to be NULL so as to remove unnecessary
4311 indirection. This lets alternative implementations fallback to the
4312 default implementation more easily.
4315 *) Support for directoryName in GeneralName related extensions
4319 *) Make it possible to link applications using Makefile.shared.
4320 Make that possible even when linking against static libraries!
4323 *) Support for single pass processing for S/MIME signing. This now
4324 means that S/MIME signing can be done from a pipe, in addition
4325 cleartext signing (multipart/signed type) is effectively streaming
4326 and the signed data does not need to be all held in memory.
4328 This is done with a new flag PKCS7_STREAM. When this flag is set
4329 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4330 is done after the data is output (and digests calculated) in
4331 SMIME_write_PKCS7().
4334 *) Add full support for -rpath/-R, both in shared libraries and
4335 applications, at least on the platforms where it's known how
4339 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4340 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4341 will now compute a table of multiples of the generator that
4342 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4343 faster (notably in the case of a single point multiplication,
4344 scalar * generator).
4345 [Nils Larsch, Bodo Moeller]
4347 *) IPv6 support for certificate extensions. The various extensions
4348 which use the IP:a.b.c.d can now take IPv6 addresses using the
4349 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4353 *) Added an ENGINE that implements RSA by performing private key
4354 exponentiations with the GMP library. The conversions to and from
4355 GMP's mpz_t format aren't optimised nor are any montgomery forms
4356 cached, and on x86 it appears OpenSSL's own performance has caught up.
4357 However there are likely to be other architectures where GMP could
4358 provide a boost. This ENGINE is not built in by default, but it can be
4359 specified at Configure time and should be accompanied by the necessary
4360 linker additions, eg;
4361 ./config -DOPENSSL_USE_GMP -lgmp
4364 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4365 testing availability of engines with "-t" - the old behaviour is
4366 produced by increasing the feature's verbosity with "-tt".
4369 *) ECDSA routines: under certain error conditions uninitialized BN objects
4370 could be freed. Solution: make sure initialization is performed early
4371 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4375 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4376 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4377 software implementations. For DSA and DH, parameter generation can
4378 also be overriden by providing the appropriate method callbacks.
4381 *) Change the "progress" mechanism used in key-generation and
4382 primality testing to functions that take a new BN_GENCB pointer in
4383 place of callback/argument pairs. The new API functions have "_ex"
4384 postfixes and the older functions are reimplemented as wrappers for
4385 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4386 declarations of the old functions to help (graceful) attempts to
4387 migrate to the new functions. Also, the new key-generation API
4388 functions operate on a caller-supplied key-structure and return
4389 success/failure rather than returning a key or NULL - this is to
4390 help make "keygen" another member function of RSA_METHOD etc.
4392 Example for using the new callback interface:
4394 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4398 BN_GENCB_set(&my_cb, my_callback, my_arg);
4400 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4401 /* For the meaning of a, b in calls to my_callback(), see the
4402 * documentation of the function that calls the callback.
4403 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4404 * my_callback should return 1 if it wants BN_is_prime_ex()
4405 * to continue, or 0 to stop.
4410 *) Change the ZLIB compression method to be stateful, and make it
4411 available to TLS with the number defined in
4412 draft-ietf-tls-compression-04.txt.
4415 *) Add the ASN.1 structures and functions for CertificatePair, which
4416 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4418 CertificatePair ::= SEQUENCE {
4419 forward [0] Certificate OPTIONAL,
4420 reverse [1] Certificate OPTIONAL,
4421 -- at least one of the pair shall be present -- }
4423 Also implement the PEM functions to read and write certificate
4424 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4426 This needed to be defined, mostly for the sake of the LDAP
4427 attribute crossCertificatePair, but may prove useful elsewhere as
4431 *) Make it possible to inhibit symlinking of shared libraries in
4432 Makefile.shared, for Cygwin's sake.
4435 *) Extend the BIGNUM API by creating a function
4436 void BN_set_negative(BIGNUM *a, int neg);
4437 and a macro that behave like
4438 int BN_is_negative(const BIGNUM *a);
4440 to avoid the need to access 'a->neg' directly in applications.
4443 *) Implement fast modular reduction for pseudo-Mersenne primes
4444 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4445 EC_GROUP_new_curve_GFp() will now automatically use this
4447 [Nils Larsch <nla@trustcenter.de>]
4449 *) Add new lock type (CRYPTO_LOCK_BN).
4452 *) Change the ENGINE framework to automatically load engines
4453 dynamically from specific directories unless they could be
4454 found to already be built in or loaded. Move all the
4455 current engines except for the cryptodev one to a new
4457 The engines in engines/ are built as shared libraries if
4458 the "shared" options was given to ./Configure or ./config.
4459 Otherwise, they are inserted in libcrypto.a.
4460 /usr/local/ssl/engines is the default directory for dynamic
4461 engines, but that can be overriden at configure time through
4462 the usual use of --prefix and/or --openssldir, and at run
4463 time with the environment variable OPENSSL_ENGINES.
4464 [Geoff Thorpe and Richard Levitte]
4466 *) Add Makefile.shared, a helper makefile to build shared
4467 libraries. Adapt Makefile.org.
4470 *) Add version info to Win32 DLLs.
4471 [Peter 'Luna' Runestig" <peter@runestig.com>]
4473 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4474 can be added using this API to created arbitrary PKCS#12
4475 files while avoiding the low level API.
4477 New options to PKCS12_create(), key or cert can be NULL and
4478 will then be omitted from the output file. The encryption
4479 algorithm NIDs can be set to -1 for no encryption, the mac
4480 iteration count can be set to 0 to omit the mac.
4482 Enhance pkcs12 utility by making the -nokeys and -nocerts
4483 options work when creating a PKCS#12 file. New option -nomac
4484 to omit the mac, NONE can be set for an encryption algorithm.
4485 New code is modified to use the enhanced PKCS12_create()
4486 instead of the low level API.
4489 *) Extend ASN1 encoder to support indefinite length constructed
4490 encoding. This can output sequences tags and octet strings in
4491 this form. Modify pk7_asn1.c to support indefinite length
4492 encoding. This is experimental and needs additional code to
4493 be useful, such as an ASN1 bio and some enhanced streaming
4496 Extend template encode functionality so that tagging is passed
4497 down to the template encoder.
4500 *) Let 'openssl req' fail if an argument to '-newkey' is not
4501 recognized instead of using RSA as a default.
4504 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4505 As these are not official, they are not included in "ALL";
4506 the "ECCdraft" ciphersuite group alias can be used to select them.
4507 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4509 *) Add ECDH engine support.
4510 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4512 *) Add ECDH in new directory crypto/ecdh/.
4513 [Douglas Stebila (Sun Microsystems Laboratories)]
4515 *) Let BN_rand_range() abort with an error after 100 iterations
4516 without success (which indicates a broken PRNG).
4519 *) Change BN_mod_sqrt() so that it verifies that the input value
4520 is really the square of the return value. (Previously,
4521 BN_mod_sqrt would show GIGO behaviour.)
4524 *) Add named elliptic curves over binary fields from X9.62, SECG,
4525 and WAP/WTLS; add OIDs that were still missing.
4527 [Sheueling Chang Shantz and Douglas Stebila
4528 (Sun Microsystems Laboratories)]
4530 *) Extend the EC library for elliptic curves over binary fields
4531 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4534 EC_GF2m_simple_method
4538 EC_GROUP_new_curve_GF2m
4539 EC_GROUP_set_curve_GF2m
4540 EC_GROUP_get_curve_GF2m
4541 EC_POINT_set_affine_coordinates_GF2m
4542 EC_POINT_get_affine_coordinates_GF2m
4543 EC_POINT_set_compressed_coordinates_GF2m
4545 Point compression for binary fields is disabled by default for
4546 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4549 As binary polynomials are represented as BIGNUMs, various members
4550 of the EC_GROUP and EC_POINT data structures can be shared
4551 between the implementations for prime fields and binary fields;
4552 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4553 are essentially identical to their ..._GFp counterparts.
4554 (For simplicity, the '..._GFp' prefix has been dropped from
4555 various internal method names.)
4557 An internal 'field_div' method (similar to 'field_mul' and
4558 'field_sqr') has been added; this is used only for binary fields.
4560 [Sheueling Chang Shantz and Douglas Stebila
4561 (Sun Microsystems Laboratories)]
4563 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4564 through methods ('mul', 'precompute_mult').
4566 The generic implementations (now internally called 'ec_wNAF_mul'
4567 and 'ec_wNAF_precomputed_mult') remain the default if these
4568 methods are undefined.
4570 [Sheueling Chang Shantz and Douglas Stebila
4571 (Sun Microsystems Laboratories)]
4573 *) New function EC_GROUP_get_degree, which is defined through
4574 EC_METHOD. For curves over prime fields, this returns the bit
4575 length of the modulus.
4577 [Sheueling Chang Shantz and Douglas Stebila
4578 (Sun Microsystems Laboratories)]
4580 *) New functions EC_GROUP_dup, EC_POINT_dup.
4581 (These simply call ..._new and ..._copy).
4583 [Sheueling Chang Shantz and Douglas Stebila
4584 (Sun Microsystems Laboratories)]
4586 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4587 Polynomials are represented as BIGNUMs (where the sign bit is not
4588 used) in the following functions [macros]:
4591 BN_GF2m_sub [= BN_GF2m_add]
4592 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4593 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4594 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4596 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4597 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4598 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4599 BN_GF2m_cmp [= BN_ucmp]
4601 (Note that only the 'mod' functions are actually for fields GF(2^m).
4602 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4604 For some functions, an the irreducible polynomial defining a
4605 field can be given as an 'unsigned int[]' with strictly
4606 decreasing elements giving the indices of those bits that are set;
4607 i.e., p[] represents the polynomial
4608 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4610 p[0] > p[1] > ... > p[k] = 0.
4611 This applies to the following functions:
4616 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4617 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4619 BN_GF2m_mod_sqrt_arr
4620 BN_GF2m_mod_solve_quad_arr
4624 Conversion can be performed by the following functions:
4629 bntest.c has additional tests for binary polynomial arithmetic.
4631 Two implementations for BN_GF2m_mod_div() are available.
4632 The default algorithm simply uses BN_GF2m_mod_inv() and
4633 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4634 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4635 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4637 [Sheueling Chang Shantz and Douglas Stebila
4638 (Sun Microsystems Laboratories)]
4640 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4641 functionality is disabled at compile-time.
4642 [Douglas Stebila <douglas.stebila@sun.com>]
4644 *) Change default behaviour of 'openssl asn1parse' so that more
4645 information is visible when viewing, e.g., a certificate:
4647 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4648 mode the content of non-printable OCTET STRINGs is output in a
4649 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4650 avoid the appearance of a printable string.
4651 [Nils Larsch <nla@trustcenter.de>]
4653 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4655 EC_GROUP_set_asn1_flag()
4656 EC_GROUP_get_asn1_flag()
4657 EC_GROUP_set_point_conversion_form()
4658 EC_GROUP_get_point_conversion_form()
4659 These control ASN1 encoding details:
4660 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4661 has been set to OPENSSL_EC_NAMED_CURVE.
4662 - Points are encoded in uncompressed form by default; options for
4663 asn1_for are as for point2oct, namely
4664 POINT_CONVERSION_COMPRESSED
4665 POINT_CONVERSION_UNCOMPRESSED
4666 POINT_CONVERSION_HYBRID
4668 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4671 EC_GROUP_get0_seed()
4672 EC_GROUP_get_seed_len()
4673 This is used only for ASN1 purposes (so far).
4674 [Nils Larsch <nla@trustcenter.de>]
4676 *) Add 'field_type' member to EC_METHOD, which holds the NID
4677 of the appropriate field type OID. The new function
4678 EC_METHOD_get_field_type() returns this value.
4679 [Nils Larsch <nla@trustcenter.de>]
4684 EC_POINT_point2hex()
4685 EC_POINT_hex2point()
4686 providing useful interfaces to EC_POINT_point2oct() and
4687 EC_POINT_oct2point().
4688 [Nils Larsch <nla@trustcenter.de>]
4690 *) Change internals of the EC library so that the functions
4691 EC_GROUP_set_generator()
4692 EC_GROUP_get_generator()
4693 EC_GROUP_get_order()
4694 EC_GROUP_get_cofactor()
4695 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4696 to methods, which would lead to unnecessary code duplication when
4697 adding different types of curves.
4698 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4700 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4701 arithmetic, and such that modified wNAFs are generated
4702 (which avoid length expansion in many cases).
4705 *) Add a function EC_GROUP_check_discriminant() (defined via
4706 EC_METHOD) that verifies that the curve discriminant is non-zero.
4708 Add a function EC_GROUP_check() that makes some sanity tests
4709 on a EC_GROUP, its generator and order. This includes
4710 EC_GROUP_check_discriminant().
4711 [Nils Larsch <nla@trustcenter.de>]
4713 *) Add ECDSA in new directory crypto/ecdsa/.
4715 Add applications 'openssl ecparam' and 'openssl ecdsa'
4716 (these are based on 'openssl dsaparam' and 'openssl dsa').
4718 ECDSA support is also included in various other files across the
4719 library. Most notably,
4720 - 'openssl req' now has a '-newkey ecdsa:file' option;
4721 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4722 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4723 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4724 them suitable for ECDSA where domain parameters must be
4725 extracted before the specific public key;
4726 - ECDSA engine support has been added.
4727 [Nils Larsch <nla@trustcenter.de>]
4729 *) Include some named elliptic curves, and add OIDs from X9.62,
4730 SECG, and WAP/WTLS. Each curve can be obtained from the new
4732 EC_GROUP_new_by_curve_name(),
4733 and the list of available named curves can be obtained with
4734 EC_get_builtin_curves().
4735 Also add a 'curve_name' member to EC_GROUP objects, which can be
4737 EC_GROUP_set_curve_name()
4738 EC_GROUP_get_curve_name()
4739 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4741 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4742 was actually never needed) and in BN_mul(). The removal in BN_mul()
4743 required a small change in bn_mul_part_recursive() and the addition
4744 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4745 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4746 bn_sub_words() and bn_add_words() except they take arrays with
4750 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4752 *) Cleanse PEM buffers before freeing them since they may contain
4754 [Benjamin Bennett <ben@psc.edu>]
4756 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4757 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4758 authentication-only ciphersuites.
4761 *) Since AES128 and AES256 share a single mask bit in the logic of
4762 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4763 kludge to work properly if AES128 is available and AES256 isn't.
4766 *) Expand security boundary to match 1.1.1 module.
4769 *) Remove redundant features: hash file source, editing of test vectors
4770 modify fipsld to use external fips_premain.c signature.
4773 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4774 run algorithm test programs.
4777 *) Make algorithm test programs more tolerant of whitespace.
4780 *) Have SSL/TLS server implementation tolerate "mismatched" record
4781 protocol version while receiving ClientHello even if the
4782 ClientHello is fragmented. (The server can't insist on the
4783 particular protocol version it has chosen before the ServerHello
4784 message has informed the client about his choice.)
4787 *) Load error codes if they are not already present instead of using a
4788 static variable. This allows them to be cleanly unloaded and reloaded.
4791 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4793 *) Introduce limits to prevent malicious keys being able to
4794 cause a denial of service. (CVE-2006-2940)
4795 [Steve Henson, Bodo Moeller]
4797 *) Fix ASN.1 parsing of certain invalid structures that can result
4798 in a denial of service. (CVE-2006-2937) [Steve Henson]
4800 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4801 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4803 *) Fix SSL client code which could crash if connecting to a
4804 malicious SSLv2 server. (CVE-2006-4343)
4805 [Tavis Ormandy and Will Drewry, Google Security Team]
4807 *) Change ciphersuite string processing so that an explicit
4808 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4809 will no longer include "AES128-SHA"), and any other similar
4810 ciphersuite (same bitmap) from *other* protocol versions (so that
4811 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4812 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4813 changes from 0.9.8b and 0.9.8d.
4816 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4818 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4819 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4821 *) Change the Unix randomness entropy gathering to use poll() when
4822 possible instead of select(), since the latter has some
4823 undesirable limitations.
4824 [Darryl Miles via Richard Levitte and Bodo Moeller]
4826 *) Disable rogue ciphersuites:
4828 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4829 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4830 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4832 The latter two were purportedly from
4833 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4836 Also deactive the remaining ciphersuites from
4837 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4838 unofficial, and the ID has long expired.
4841 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4842 dual-core machines) and other potential thread-safety issues.
4845 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4847 *) Adapt fipsld and the build system to link against the validated FIPS
4848 module in FIPS mode.
4851 *) Fixes for VC++ 2005 build under Windows.
4854 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4855 from a Windows bash shell such as MSYS. It is autodetected from the
4856 "config" script when run from a VC++ environment. Modify standard VC++
4857 build to use fipscanister.o from the GNU make build.
4860 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4862 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4863 The value now differs depending on if you build for FIPS or not.
4864 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4865 safely run with a non-FIPSed libcrypto, as it may crash because of
4866 the difference induced by this change.
4869 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4871 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4872 (part of SSL_OP_ALL). This option used to disable the
4873 countermeasure against man-in-the-middle protocol-version
4874 rollback in the SSL 2.0 server implementation, which is a bad
4875 idea. (CVE-2005-2969)
4877 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4878 for Information Security, National Institute of Advanced Industrial
4879 Science and Technology [AIST], Japan)]
4881 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4882 mainly for FIPS compliance and not fully integrated at this stage.
4885 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4886 the exponentiation using a fixed-length exponent. (Otherwise,
4887 the information leaked through timing could expose the secret key
4888 after many signatures; cf. Bleichenbacher's attack on DSA with
4892 *) Make a new fixed-window mod_exp implementation the default for
4893 RSA, DSA, and DH private-key operations so that the sequence of
4894 squares and multiplies and the memory access pattern are
4895 independent of the particular secret key. This will mitigate
4896 cache-timing and potential related attacks.
4898 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4899 and this is automatically used by BN_mod_exp_mont() if the new flag
4900 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4901 will use this BN flag for private exponents unless the flag
4902 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4903 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4905 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4907 *) Change the client implementation for SSLv23_method() and
4908 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4909 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4910 (Previously, the SSL 2.0 backwards compatible Client Hello
4911 message format would be used even with SSL_OP_NO_SSLv2.)
4914 *) Add support for smime-type MIME parameter in S/MIME messages which some
4918 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4919 a threadsafe manner. Modify rsa code to use new function and add calls
4920 to dsa and dh code (which had race conditions before).
4923 *) Include the fixed error library code in the C error file definitions
4924 instead of fixing them up at runtime. This keeps the error code
4925 structures constant.
4928 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4930 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4933 *) Fixes for newer kerberos headers. NB: the casts are needed because
4934 the 'length' field is signed on one version and unsigned on another
4935 with no (?) obvious way to tell the difference, without these VC++
4936 complains. Also the "definition" of FAR (blank) is no longer included
4937 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4938 some needed definitions.
4941 *) Undo Cygwin change.
4944 *) Added support for proxy certificates according to RFC 3820.
4945 Because they may be a security thread to unaware applications,
4946 they must be explicitely allowed in run-time. See
4947 docs/HOWTO/proxy_certificates.txt for further information.
4950 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4952 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4953 server and client random values. Previously
4954 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4955 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4957 This change has negligible security impact because:
4959 1. Server and client random values still have 24 bytes of pseudo random
4962 2. Server and client random values are sent in the clear in the initial
4965 3. The master secret is derived using the premaster secret (48 bytes in
4966 size for static RSA ciphersuites) as well as client server and random
4969 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4972 [Stephen Henson, reported by UK NISCC]
4974 *) Use Windows randomness collection on Cygwin.
4977 *) Fix hang in EGD/PRNGD query when communication socket is closed
4978 prematurely by EGD/PRNGD.
4979 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4981 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4984 *) Back-port of selected performance improvements from development
4985 branch, as well as improved support for PowerPC platforms.
4988 *) Add lots of checks for memory allocation failure, error codes to indicate
4989 failure and freeing up memory if a failure occurs.
4990 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4992 *) Add new -passin argument to dgst.
4995 *) Perform some character comparisons of different types in X509_NAME_cmp:
4996 this is needed for some certificates that reencode DNs into UTF8Strings
4997 (in violation of RFC3280) and can't or wont issue name rollover
5001 *) Make an explicit check during certificate validation to see that
5002 the CA setting in each certificate on the chain is correct. As a
5003 side effect always do the following basic checks on extensions,
5004 not just when there's an associated purpose to the check:
5006 - if there is an unhandled critical extension (unless the user
5007 has chosen to ignore this fault)
5008 - if the path length has been exceeded (if one is set at all)
5009 - that certain extensions fit the associated purpose (if one has
5013 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5015 *) Avoid a race condition when CRLs are checked in a multi threaded
5016 environment. This would happen due to the reordering of the revoked
5017 entries during signature checking and serial number lookup. Now the
5018 encoding is cached and the serial number sort performed under a lock.
5019 Add new STACK function sk_is_sorted().
5022 *) Add Delta CRL to the extension code.
5025 *) Various fixes to s3_pkt.c so alerts are sent properly.
5026 [David Holmes <d.holmes@f5.com>]
5028 *) Reduce the chances of duplicate issuer name and serial numbers (in
5029 violation of RFC3280) using the OpenSSL certificate creation utilities.
5030 This is done by creating a random 64 bit value for the initial serial
5031 number when a serial number file is created or when a self signed
5032 certificate is created using 'openssl req -x509'. The initial serial
5033 number file is created using 'openssl x509 -next_serial' in CA.pl
5034 rather than being initialized to 1.
5037 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5039 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5040 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5041 [Joe Orton, Steve Henson]
5043 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5045 [Joe Orton, Steve Henson]
5047 *) Make it possible to have multiple active certificates with the same
5048 subject in the CA index file. This is done only if the keyword
5049 'unique_subject' is set to 'no' in the main CA section (default
5050 if 'CA_default') of the configuration file. The value is saved
5051 with the database itself in a separate index attribute file,
5052 named like the index file with '.attr' appended to the name.
5055 *) X509 verify fixes. Disable broken certificate workarounds when
5056 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5057 keyUsage extension present. Don't accept CRLs with unhandled critical
5058 extensions: since verify currently doesn't process CRL extensions this
5059 rejects a CRL with *any* critical extensions. Add new verify error codes
5063 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5064 A clarification of RFC2560 will require the use of OCTET STRINGs and
5065 some implementations cannot handle the current raw format. Since OpenSSL
5066 copies and compares OCSP nonces as opaque blobs without any attempt at
5067 parsing them this should not create any compatibility issues.
5070 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5071 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5072 this HMAC (and other) operations are several times slower than OpenSSL
5076 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5077 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5079 *) Use the correct content when signing type "other".
5082 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5084 *) Fix various bugs revealed by running the NISCC test suite:
5086 Stop out of bounds reads in the ASN1 code when presented with
5087 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5089 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5091 If verify callback ignores invalid public key errors don't try to check
5092 certificate signature with the NULL public key.
5096 *) New -ignore_err option in ocsp application to stop the server
5097 exiting on the first error in a request.
5100 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5101 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5105 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5106 extra data after the compression methods not only for TLS 1.0
5107 but also for SSL 3.0 (as required by the specification).
5108 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5110 *) Change X509_certificate_type() to mark the key as exported/exportable
5111 when it's 512 *bits* long, not 512 bytes.
5114 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5115 blocks during encryption.
5118 *) Various fixes to base64 BIO and non blocking I/O. On write
5119 flushes were not handled properly if the BIO retried. On read
5120 data was not being buffered properly and had various logic bugs.
5121 This also affects blocking I/O when the data being decoded is a
5125 *) Various S/MIME bugfixes and compatibility changes:
5126 output correct application/pkcs7 MIME type if
5127 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5128 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5129 of files as .eml work). Correctly handle very long lines in MIME
5133 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5135 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5136 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5137 a protocol version number mismatch like a decryption error
5138 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5141 *) Turn on RSA blinding by default in the default implementation
5142 to avoid a timing attack. Applications that don't want it can call
5143 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5144 They would be ill-advised to do so in most cases.
5145 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5147 *) Change RSA blinding code so that it works when the PRNG is not
5148 seeded (in this case, the secret RSA exponent is abused as
5149 an unpredictable seed -- if it is not unpredictable, there
5150 is no point in blinding anyway). Make RSA blinding thread-safe
5151 by remembering the creator's thread ID in rsa->blinding and
5152 having all other threads use local one-time blinding factors
5153 (this requires more computation than sharing rsa->blinding, but
5154 avoids excessive locking; and if an RSA object is not shared
5155 between threads, blinding will still be very fast).
5158 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5159 ENGINE as defaults for all supported algorithms irrespective of
5160 the 'flags' parameter. 'flags' is now honoured, so applications
5161 should make sure they are passing it correctly.
5164 *) Target "mingw" now allows native Windows code to be generated in
5165 the Cygwin environment as well as with the MinGW compiler.
5168 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5170 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5171 via timing by performing a MAC computation even if incorrect
5172 block cipher padding has been found. This is a countermeasure
5173 against active attacks where the attacker has to distinguish
5174 between bad padding and a MAC verification error. (CVE-2003-0078)
5176 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5177 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5178 Martin Vuagnoux (EPFL, Ilion)]
5180 *) Make the no-err option work as intended. The intention with no-err
5181 is not to have the whole error stack handling routines removed from
5182 libcrypto, it's only intended to remove all the function name and
5183 reason texts, thereby removing some of the footprint that may not
5184 be interesting if those errors aren't displayed anyway.
5186 NOTE: it's still possible for any application or module to have it's
5187 own set of error texts inserted. The routines are there, just not
5188 used by default when no-err is given.
5191 *) Add support for FreeBSD on IA64.
5192 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5194 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5195 Kerberos function mit_des_cbc_cksum(). Before this change,
5196 the value returned by DES_cbc_cksum() was like the one from
5197 mit_des_cbc_cksum(), except the bytes were swapped.
5198 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5200 *) Allow an application to disable the automatic SSL chain building.
5201 Before this a rather primitive chain build was always performed in
5202 ssl3_output_cert_chain(): an application had no way to send the
5203 correct chain if the automatic operation produced an incorrect result.
5205 Now the chain builder is disabled if either:
5207 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5209 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5211 The reasoning behind this is that an application would not want the
5212 auto chain building to take place if extra chain certificates are
5213 present and it might also want a means of sending no additional
5214 certificates (for example the chain has two certificates and the
5218 *) Add the possibility to build without the ENGINE framework.
5219 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5221 *) Under Win32 gmtime() can return NULL: check return value in
5222 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5225 *) DSA routines: under certain error conditions uninitialized BN objects
5226 could be freed. Solution: make sure initialization is performed early
5227 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5228 Nils Larsch <nla@trustcenter.de> via PR#459)
5231 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5232 checked on reconnect on the client side, therefore session resumption
5233 could still fail with a "ssl session id is different" error. This
5234 behaviour is masked when SSL_OP_ALL is used due to
5235 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5236 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5237 followup to PR #377.
5240 *) IA-32 assembler support enhancements: unified ELF targets, support
5241 for SCO/Caldera platforms, fix for Cygwin shared build.
5244 *) Add support for FreeBSD on sparc64. As a consequence, support for
5245 FreeBSD on non-x86 processors is separate from x86 processors on
5246 the config script, much like the NetBSD support.
5247 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5249 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5251 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5254 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5255 code (06) was taken as the first octet of the session ID and the last
5256 octet was ignored consequently. As a result SSLv2 client side session
5257 caching could not have worked due to the session ID mismatch between
5259 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5263 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5264 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5268 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5269 seems that in spite of existing for more than a year, many application
5270 author have done nothing to provide the necessary callbacks, which
5271 means that this particular engine will not work properly anywhere.
5272 This is a very unfortunate situation which forces us, in the name
5273 of usability, to give the hw_ncipher.c a static lock, which is part
5275 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5276 appear in 0.9.8 or later. We EXPECT application authors to have
5277 dealt properly with this when 0.9.8 is released (unless we actually
5278 make such changes in the libcrypto locking code that changes will
5279 have to be made anyway).
5282 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5283 octets have been read, EOF or an error occurs. Without this change
5284 some truncated ASN1 structures will not produce an error.
5287 *) Disable Heimdal support, since it hasn't been fully implemented.
5288 Still give the possibility to force the use of Heimdal, but with
5289 warnings and a request that patches get sent to openssl-dev.
5292 *) Add the VC-CE target, introduce the WINCE sysname, and add
5293 INSTALL.WCE and appropriate conditionals to make it build.
5294 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5296 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5297 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5298 edit numbers of the version.
5299 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5301 *) Introduce safe string copy and catenation functions
5302 (BUF_strlcpy() and BUF_strlcat()).
5303 [Ben Laurie (CHATS) and Richard Levitte]
5305 *) Avoid using fixed-size buffers for one-line DNs.
5306 [Ben Laurie (CHATS)]
5308 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5309 resizing buffers containing secrets, and use where appropriate.
5310 [Ben Laurie (CHATS)]
5312 *) Avoid using fixed size buffers for configuration file location.
5313 [Ben Laurie (CHATS)]
5315 *) Avoid filename truncation for various CA files.
5316 [Ben Laurie (CHATS)]
5318 *) Use sizeof in preference to magic numbers.
5319 [Ben Laurie (CHATS)]
5321 *) Avoid filename truncation in cert requests.
5322 [Ben Laurie (CHATS)]
5324 *) Add assertions to check for (supposedly impossible) buffer
5326 [Ben Laurie (CHATS)]
5328 *) Don't cache truncated DNS entries in the local cache (this could
5329 potentially lead to a spoofing attack).
5330 [Ben Laurie (CHATS)]
5332 *) Fix various buffers to be large enough for hex/decimal
5333 representations in a platform independent manner.
5334 [Ben Laurie (CHATS)]
5336 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5337 resizing buffers containing secrets, and use where appropriate.
5338 [Ben Laurie (CHATS)]
5340 *) Add BIO_indent() to avoid much slightly worrying code to do
5342 [Ben Laurie (CHATS)]
5344 *) Convert sprintf()/BIO_puts() to BIO_printf().
5345 [Ben Laurie (CHATS)]
5347 *) buffer_gets() could terminate with the buffer only half
5349 [Ben Laurie (CHATS)]
5351 *) Add assertions to prevent user-supplied crypto functions from
5352 overflowing internal buffers by having large block sizes, etc.
5353 [Ben Laurie (CHATS)]
5355 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5357 [Ben Laurie (CHATS)]
5359 *) Eliminate unused copy of key in RC4.
5360 [Ben Laurie (CHATS)]
5362 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5363 [Ben Laurie (CHATS)]
5365 *) Fix off-by-one error in EGD path.
5366 [Ben Laurie (CHATS)]
5368 *) If RANDFILE path is too long, ignore instead of truncating.
5369 [Ben Laurie (CHATS)]
5371 *) Eliminate unused and incorrectly sized X.509 structure
5373 [Ben Laurie (CHATS)]
5375 *) Eliminate unused and dangerous function knumber().
5376 [Ben Laurie (CHATS)]
5378 *) Eliminate unused and dangerous structure, KSSL_ERR.
5379 [Ben Laurie (CHATS)]
5381 *) Protect against overlong session ID context length in an encoded
5382 session object. Since these are local, this does not appear to be
5384 [Ben Laurie (CHATS)]
5386 *) Change from security patch (see 0.9.6e below) that did not affect
5387 the 0.9.6 release series:
5389 Remote buffer overflow in SSL3 protocol - an attacker could
5390 supply an oversized master key in Kerberos-enabled versions.
5392 [Ben Laurie (CHATS)]
5394 *) Change the SSL kerb5 codes to match RFC 2712.
5397 *) Make -nameopt work fully for req and add -reqopt switch.
5398 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5400 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5401 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5403 *) Make sure tests can be performed even if the corresponding algorithms
5404 have been removed entirely. This was also the last step to make
5405 OpenSSL compilable with DJGPP under all reasonable conditions.
5406 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5408 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5409 to allow version independent disabling of normally unselected ciphers,
5410 which may be activated as a side-effect of selecting a single cipher.
5412 (E.g., cipher list string "RSA" enables ciphersuites that are left
5413 out of "ALL" because they do not provide symmetric encryption.
5414 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5415 [Lutz Jaenicke, Bodo Moeller]
5417 *) Add appropriate support for separate platform-dependent build
5418 directories. The recommended way to make a platform-dependent
5419 build directory is the following (tested on Linux), maybe with
5422 # Place yourself outside of the OpenSSL source tree. In
5423 # this example, the environment variable OPENSSL_SOURCE
5424 # is assumed to contain the absolute OpenSSL source directory.
5425 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5426 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5427 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5428 mkdir -p `dirname $F`
5429 ln -s $OPENSSL_SOURCE/$F $F
5432 To be absolutely sure not to disturb the source tree, a "make clean"
5433 is a good thing. If it isn't successful, don't worry about it,
5434 it probably means the source directory is very clean.
5437 *) Make sure any ENGINE control commands make local copies of string
5438 pointers passed to them whenever necessary. Otherwise it is possible
5439 the caller may have overwritten (or deallocated) the original string
5440 data when a later ENGINE operation tries to use the stored values.
5441 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5443 *) Improve diagnostics in file reading and command-line digests.
5444 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5446 *) Add AES modes CFB and OFB to the object database. Correct an
5447 error in AES-CFB decryption.
5450 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5451 allows existing EVP_CIPHER_CTX structures to be reused after
5452 calling EVP_*Final(). This behaviour is used by encryption
5453 BIOs and some applications. This has the side effect that
5454 applications must explicitly clean up cipher contexts with
5455 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5458 *) Check the values of dna and dnb in bn_mul_recursive before calling
5459 bn_mul_comba (a non zero value means the a or b arrays do not contain
5460 n2 elements) and fallback to bn_mul_normal if either is not zero.
5463 *) Fix escaping of non-ASCII characters when using the -subj option
5464 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5467 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5468 form for "surname", serialNumber has no short form.
5469 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5470 therefore remove "mail" short name for "internet 7".
5471 The OID for unique identifiers in X509 certificates is
5472 x500UniqueIdentifier, not uniqueIdentifier.
5473 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5476 *) Add an "init" command to the ENGINE config module and auto initialize
5477 ENGINEs. Without any "init" command the ENGINE will be initialized
5478 after all ctrl commands have been executed on it. If init=1 the
5479 ENGINE is initailized at that point (ctrls before that point are run
5480 on the uninitialized ENGINE and after on the initialized one). If
5481 init=0 then the ENGINE will not be iniatialized at all.
5484 *) Fix the 'app_verify_callback' interface so that the user-defined
5485 argument is actually passed to the callback: In the
5486 SSL_CTX_set_cert_verify_callback() prototype, the callback
5487 declaration has been changed from
5490 int (*cb)(X509_STORE_CTX *,void *);
5491 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5492 i=s->ctx->app_verify_callback(&ctx)
5493 has been changed into
5494 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5496 To update applications using SSL_CTX_set_cert_verify_callback(),
5497 a dummy argument can be added to their callback functions.
5498 [D. K. Smetters <smetters@parc.xerox.com>]
5500 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5501 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5503 *) Add and OPENSSL_LOAD_CONF define which will cause
5504 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5505 This allows older applications to transparently support certain
5506 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5507 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5508 load the config file and OPENSSL_add_all_algorithms_conf() which will
5509 always load it have also been added.
5512 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5513 Adjust NIDs and EVP layer.
5514 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5516 *) Config modules support in openssl utility.
5518 Most commands now load modules from the config file,
5519 though in a few (such as version) this isn't done
5520 because it couldn't be used for anything.
5522 In the case of ca and req the config file used is
5523 the same as the utility itself: that is the -config
5524 command line option can be used to specify an
5528 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5529 use "openssl_conf" if filename is NULL use default openssl config file.
5532 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5533 config section name. Add a new flag to tolerate a missing config file
5534 and move code to CONF_modules_load_file().
5537 *) Support for crypto accelerator cards from Accelerated Encryption
5538 Processing, www.aep.ie. (Use engine 'aep')
5539 The support was copied from 0.9.6c [engine] and adapted/corrected
5540 to work with the new engine framework.
5541 [AEP Inc. and Richard Levitte]
5543 *) Support for SureWare crypto accelerator cards from Baltimore
5544 Technologies. (Use engine 'sureware')
5545 The support was copied from 0.9.6c [engine] and adapted
5546 to work with the new engine framework.
5549 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5550 make the newer ENGINE framework commands for the CHIL engine work.
5551 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5553 *) Make it possible to produce shared libraries on ReliantUNIX.
5554 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5556 *) Add the configuration target debug-linux-ppro.
5557 Make 'openssl rsa' use the general key loading routines
5558 implemented in apps.c, and make those routines able to
5559 handle the key format FORMAT_NETSCAPE and the variant
5561 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5563 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5564 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5566 *) Add -keyform to rsautl, and document -engine.
5567 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5569 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5570 BIO_R_NO_SUCH_FILE error code rather than the generic
5571 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5574 *) Add new functions
5576 ERR_peek_last_error_line
5577 ERR_peek_last_error_line_data.
5578 These are similar to
5581 ERR_peek_error_line_data,
5582 but report on the latest error recorded rather than the first one
5583 still in the error queue.
5584 [Ben Laurie, Bodo Moeller]
5586 *) default_algorithms option in ENGINE config module. This allows things
5588 default_algorithms = ALL
5589 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5592 *) Preliminary ENGINE config module.
5595 *) New experimental application configuration code.
5598 *) Change the AES code to follow the same name structure as all other
5599 symmetric ciphers, and behave the same way. Move everything to
5600 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5601 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5603 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5604 [Ben Laurie and Theo de Raadt]
5606 *) Add option to output public keys in req command.
5607 [Massimiliano Pala madwolf@openca.org]
5609 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5610 (up to about 10% better than before for P-192 and P-224).
5613 *) New functions/macros
5615 SSL_CTX_set_msg_callback(ctx, cb)
5616 SSL_CTX_set_msg_callback_arg(ctx, arg)
5617 SSL_set_msg_callback(ssl, cb)
5618 SSL_set_msg_callback_arg(ssl, arg)
5620 to request calling a callback function
5622 void cb(int write_p, int version, int content_type,
5623 const void *buf, size_t len, SSL *ssl, void *arg)
5625 whenever a protocol message has been completely received
5626 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5627 protocol version according to which the SSL library interprets
5628 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5629 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5630 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5631 specification (change_cipher_spec(20), alert(21), handshake(22)).
5632 'buf' and 'len' point to the actual message, 'ssl' to the
5633 SSL object, and 'arg' is the application-defined value set by
5634 SSL[_CTX]_set_msg_callback_arg().
5636 'openssl s_client' and 'openssl s_server' have new '-msg' options
5637 to enable a callback that displays all protocol messages.
5640 *) Change the shared library support so shared libraries are built as
5641 soon as the corresponding static library is finished, and thereby get
5642 openssl and the test programs linked against the shared library.
5643 This still only happens when the keyword "shard" has been given to
5644 the configuration scripts.
5646 NOTE: shared library support is still an experimental thing, and
5647 backward binary compatibility is still not guaranteed.
5648 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5650 *) Add support for Subject Information Access extension.
5651 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5653 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5654 additional bytes when new memory had to be allocated, not just
5655 when reusing an existing buffer.
5658 *) New command line and configuration option 'utf8' for the req command.
5659 This allows field values to be specified as UTF8 strings.
5662 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5663 runs for the former and machine-readable output for the latter.
5666 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5667 of the e-mail address in the DN (i.e., it will go into a certificate
5668 extension only). The new configuration file option 'email_in_dn = no'
5669 has the same effect.
5670 [Massimiliano Pala madwolf@openca.org]
5672 *) Change all functions with names starting with des_ to be starting
5673 with DES_ instead. Add wrappers that are compatible with libdes,
5674 but are named _ossl_old_des_*. Finally, add macros that map the
5675 des_* symbols to the corresponding _ossl_old_des_* if libdes
5676 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5677 desired, the des_* symbols will be mapped to DES_*, with one
5680 Since we provide two compatibility mappings, the user needs to
5681 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5682 compatibility is desired. The default (i.e., when that macro
5683 isn't defined) is OpenSSL 0.9.6c compatibility.
5685 There are also macros that enable and disable the support of old
5686 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5687 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5688 are defined, the default will apply: to support the old des routines.
5690 In either case, one must include openssl/des.h to get the correct
5691 definitions. Do not try to just include openssl/des_old.h, that
5694 NOTE: This is a major break of an old API into a new one. Software
5695 authors are encouraged to switch to the DES_ style functions. Some
5696 time in the future, des_old.h and the libdes compatibility functions
5697 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5698 default), and then completely removed.
5701 *) Test for certificates which contain unsupported critical extensions.
5702 If such a certificate is found during a verify operation it is
5703 rejected by default: this behaviour can be overridden by either
5704 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5705 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5706 X509_supported_extension() has also been added which returns 1 if a
5707 particular extension is supported.
5710 *) Modify the behaviour of EVP cipher functions in similar way to digests
5711 to retain compatibility with existing code.
5714 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5715 compatibility with existing code. In particular the 'ctx' parameter does
5716 not have to be to be initialized before the call to EVP_DigestInit() and
5717 it is tidied up after a call to EVP_DigestFinal(). New function
5718 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5719 EVP_MD_CTX_copy() changed to not require the destination to be
5720 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5721 requires the destination to be valid.
5723 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5724 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5727 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5728 so that complete 'Handshake' protocol structures are kept in memory
5729 instead of overwriting 'msg_type' and 'length' with 'body' data.
5732 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5733 [Massimo Santin via Richard Levitte]
5735 *) Major restructuring to the underlying ENGINE code. This includes
5736 reduction of linker bloat, separation of pure "ENGINE" manipulation
5737 (initialisation, etc) from functionality dealing with implementations
5738 of specific crypto iterfaces. This change also introduces integrated
5739 support for symmetric ciphers and digest implementations - so ENGINEs
5740 can now accelerate these by providing EVP_CIPHER and EVP_MD
5741 implementations of their own. This is detailed in crypto/engine/README
5742 as it couldn't be adequately described here. However, there are a few
5743 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5744 were changed in the original introduction of ENGINE code have now
5745 reverted back - the hooking from this code to ENGINE is now a good
5746 deal more passive and at run-time, operations deal directly with
5747 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5748 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5749 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5750 they were not being used by the framework as there is no concept of a
5751 BIGNUM_METHOD and they could not be generalised to the new
5752 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5753 ENGINE_cpy() has been removed as it cannot be consistently defined in
5757 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5760 *) Change mkdef.pl to sort symbols that get the same entry number,
5761 and make sure the automatically generated functions ERR_load_*
5762 become part of libeay.num as well.
5765 *) New function SSL_renegotiate_pending(). This returns true once
5766 renegotiation has been requested (either SSL_renegotiate() call
5767 or HelloRequest/ClientHello received from the peer) and becomes
5768 false once a handshake has been completed.
5769 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5770 sends a HelloRequest, but does not ensure that a handshake takes
5771 place. SSL_renegotiate_pending() is useful for checking if the
5772 client has followed the request.)
5775 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5776 By default, clients may request session resumption even during
5777 renegotiation (if session ID contexts permit); with this option,
5778 session resumption is possible only in the first handshake.
5780 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5781 more bits available for options that should not be part of
5782 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5785 *) Add some demos for certificate and certificate request creation.
5788 *) Make maximum certificate chain size accepted from the peer application
5789 settable (SSL*_get/set_max_cert_list()), as proposed by
5790 "Douglas E. Engert" <deengert@anl.gov>.
5793 *) Add support for shared libraries for Unixware-7
5794 (Boyd Lynn Gerber <gerberb@zenez.com>).
5797 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5798 be done prior to destruction. Use this to unload error strings from
5799 ENGINEs that load their own error strings. NB: This adds two new API
5800 functions to "get" and "set" this destroy handler in an ENGINE.
5803 *) Alter all existing ENGINE implementations (except "openssl" and
5804 "openbsd") to dynamically instantiate their own error strings. This
5805 makes them more flexible to be built both as statically-linked ENGINEs
5806 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5807 Also, add stub code to each that makes building them as self-contained
5808 shared-libraries easier (see README.ENGINE).
5811 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5812 implementations into applications that are completely implemented in
5813 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5814 commands that can be used to configure what shared-library to load and
5815 to control aspects of the way it is handled. Also, made an update to
5816 the README.ENGINE file that brings its information up-to-date and
5817 provides some information and instructions on the "dynamic" ENGINE
5818 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5821 *) Make it possible to unload ranges of ERR strings with a new
5822 "ERR_unload_strings" function.
5825 *) Add a copy() function to EVP_MD.
5828 *) Make EVP_MD routines take a context pointer instead of just the
5829 md_data void pointer.
5832 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5833 that the digest can only process a single chunk of data
5834 (typically because it is provided by a piece of
5835 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5836 is only going to provide a single chunk of data, and hence the
5837 framework needn't accumulate the data for oneshot drivers.
5840 *) As with "ERR", make it possible to replace the underlying "ex_data"
5841 functions. This change also alters the storage and management of global
5842 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5843 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5844 index counters. The API functions that use this state have been changed
5845 to take a "class_index" rather than pointers to the class's local STACK
5846 and counter, and there is now an API function to dynamically create new
5847 classes. This centralisation allows us to (a) plug a lot of the
5848 thread-safety problems that existed, and (b) makes it possible to clean
5849 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5850 such data would previously have always leaked in application code and
5851 workarounds were in place to make the memory debugging turn a blind eye
5852 to it. Application code that doesn't use this new function will still
5853 leak as before, but their memory debugging output will announce it now
5854 rather than letting it slide.
5856 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5857 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5858 has a return value to indicate success or failure.
5861 *) Make it possible to replace the underlying "ERR" functions such that the
5862 global state (2 LHASH tables and 2 locks) is only used by the "default"
5863 implementation. This change also adds two functions to "get" and "set"
5864 the implementation prior to it being automatically set the first time
5865 any other ERR function takes place. Ie. an application can call "get",
5866 pass the return value to a module it has just loaded, and that module
5867 can call its own "set" function using that value. This means the
5868 module's "ERR" operations will use (and modify) the error state in the
5869 application and not in its own statically linked copy of OpenSSL code.
5872 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5873 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5874 the operation, and provides a more encapsulated way for external code
5875 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5876 to use these functions rather than manually incrementing the counts.
5878 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5881 *) Add EVP test program.
5884 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5887 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5888 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5889 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5890 These allow a CRL to be built without having to access X509_CRL fields
5891 directly. Modify 'ca' application to use new functions.
5894 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5895 bug workarounds. Rollback attack detection is a security feature.
5896 The problem will only arise on OpenSSL servers when TLSv1 is not
5897 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5898 Software authors not wanting to support TLSv1 will have special reasons
5899 for their choice and can explicitly enable this option.
5900 [Bodo Moeller, Lutz Jaenicke]
5902 *) Rationalise EVP so it can be extended: don't include a union of
5903 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5904 (similar to those existing for EVP_CIPHER_CTX).
5909 EVP_MD_CTX_init(&md); /* new function call */
5910 EVP_DigestInit(&md, EVP_sha1());
5911 EVP_DigestUpdate(&md, in, len);
5912 EVP_DigestFinal(&md, out, NULL);
5913 EVP_MD_CTX_cleanup(&md); /* new function call */
5917 *) Make DES key schedule conform to the usual scheme, as well as
5918 correcting its structure. This means that calls to DES functions
5919 now have to pass a pointer to a des_key_schedule instead of a
5920 plain des_key_schedule (which was actually always a pointer
5923 des_key_schedule ks;
5925 des_set_key_checked(..., &ks);
5926 des_ncbc_encrypt(..., &ks, ...);
5928 (Note that a later change renames 'des_...' into 'DES_...'.)
5931 *) Initial reduction of linker bloat: the use of some functions, such as
5932 PEM causes large amounts of unused functions to be linked in due to
5933 poor organisation. For example pem_all.c contains every PEM function
5934 which has a knock on effect of linking in large amounts of (unused)
5935 ASN1 code. Grouping together similar functions and splitting unrelated
5936 functions prevents this.
5939 *) Cleanup of EVP macros.
5942 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5943 correct _ecb suffix.
5946 *) Add initial OCSP responder support to ocsp application. The
5947 revocation information is handled using the text based index
5948 use by the ca application. The responder can either handle
5949 requests generated internally, supplied in files (for example
5950 via a CGI script) or using an internal minimal server.
5953 *) Add configuration choices to get zlib compression for TLS.
5956 *) Changes to Kerberos SSL for RFC 2712 compliance:
5957 1. Implemented real KerberosWrapper, instead of just using
5958 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5959 2. Implemented optional authenticator field of KerberosWrapper.
5961 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5962 and authenticator structs; see crypto/krb5/.
5964 Generalized Kerberos calls to support multiple Kerberos libraries.
5965 [Vern Staats <staatsvr@asc.hpc.mil>,
5966 Jeffrey Altman <jaltman@columbia.edu>
5967 via Richard Levitte]
5969 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5970 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5971 values for each of the key sizes rather than having just
5972 parameters (and 'speed' generating keys each time).
5975 *) Speed up EVP routines.
5978 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5979 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5980 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5981 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5983 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5984 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5985 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5988 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5990 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5993 *) Added the OS2-EMX target.
5994 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5996 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5997 to support NCONF routines in extension code. New function CONF_set_nconf()
5998 to allow functions which take an NCONF to also handle the old LHASH
5999 structure: this means that the old CONF compatible routines can be
6000 retained (in particular wrt extensions) without having to duplicate the
6001 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6004 *) Enhance the general user interface with mechanisms for inner control
6005 and with possibilities to have yes/no kind of prompts.
6008 *) Change all calls to low level digest routines in the library and
6009 applications to use EVP. Add missing calls to HMAC_cleanup() and
6010 don't assume HMAC_CTX can be copied using memcpy().
6011 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6013 *) Add the possibility to control engines through control names but with
6014 arbitrary arguments instead of just a string.
6015 Change the key loaders to take a UI_METHOD instead of a callback
6016 function pointer. NOTE: this breaks binary compatibility with earlier
6017 versions of OpenSSL [engine].
6018 Adapt the nCipher code for these new conditions and add a card insertion
6022 *) Enhance the general user interface with mechanisms to better support
6023 dialog box interfaces, application-defined prompts, the possibility
6024 to use defaults (for example default passwords from somewhere else)
6025 and interrupts/cancellations.
6028 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6029 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6032 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6033 tidy up some unnecessarily weird code in 'sk_new()').
6034 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6036 *) Change the key loading routines for ENGINEs to use the same kind
6037 callback (pem_password_cb) as all other routines that need this
6041 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6042 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6043 than this minimum value is recommended.
6046 *) New random seeder for OpenVMS, using the system process statistics
6047 that are easily reachable.
6050 *) Windows apparently can't transparently handle global
6051 variables defined in DLLs. Initialisations such as:
6053 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6055 wont compile. This is used by the any applications that need to
6056 declare their own ASN1 modules. This was fixed by adding the option
6057 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6058 needed for static libraries under Win32.
6061 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6062 setting of purpose and trust fields. New X509_STORE trust and
6063 purpose functions and tidy up setting in other SSL functions.
6066 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6067 structure. These are inherited by X509_STORE_CTX when it is
6068 initialised. This allows various defaults to be set in the
6069 X509_STORE structure (such as flags for CRL checking and custom
6070 purpose or trust settings) for functions which only use X509_STORE_CTX
6071 internally such as S/MIME.
6073 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6074 trust settings if they are not set in X509_STORE. This allows X509_STORE
6075 purposes and trust (in S/MIME for example) to override any set by default.
6077 Add command line options for CRL checking to smime, s_client and s_server
6081 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6082 are set then the CRL is looked up in the X509_STORE structure and
6083 its validity and signature checked, then if the certificate is found
6084 in the CRL the verify fails with a revoked error.
6086 Various new CRL related callbacks added to X509_STORE_CTX structure.
6088 Command line options added to 'verify' application to support this.
6090 This needs some additional work, such as being able to handle multiple
6091 CRLs with different times, extension based lookup (rather than just
6092 by subject name) and ultimately more complete V2 CRL extension
6096 *) Add a general user interface API (crypto/ui/). This is designed
6097 to replace things like des_read_password and friends (backward
6098 compatibility functions using this new API are provided).
6099 The purpose is to remove prompting functions from the DES code
6100 section as well as provide for prompting through dialog boxes in
6101 a window system and the like.
6104 *) Add "ex_data" support to ENGINE so implementations can add state at a
6105 per-structure level rather than having to store it globally.
6108 *) Make it possible for ENGINE structures to be copied when retrieved by
6109 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6110 This causes the "original" ENGINE structure to act like a template,
6111 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6112 operational state can be localised to each ENGINE structure, despite the
6113 fact they all share the same "methods". New ENGINE structures returned in
6114 this case have no functional references and the return value is the single
6115 structural reference. This matches the single structural reference returned
6116 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6120 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6121 needs to match any other type at all we need to manually clear the
6125 *) Changes to the "openssl engine" utility to include;
6126 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6127 about an ENGINE's available control commands.
6128 - executing control commands from command line arguments using the
6129 '-pre' and '-post' switches. '-post' is only used if '-t' is
6130 specified and the ENGINE is successfully initialised. The syntax for
6131 the individual commands are colon-separated, for example;
6132 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6135 *) New dynamic control command support for ENGINEs. ENGINEs can now
6136 declare their own commands (numbers), names (strings), descriptions,
6137 and input types for run-time discovery by calling applications. A
6138 subset of these commands are implicitly classed as "executable"
6139 depending on their input type, and only these can be invoked through
6140 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6141 can be based on user input, config files, etc). The distinction is
6142 that "executable" commands cannot return anything other than a boolean
6143 result and can only support numeric or string input, whereas some
6144 discoverable commands may only be for direct use through
6145 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6146 pointers, or other custom uses. The "executable" commands are to
6147 support parameterisations of ENGINE behaviour that can be
6148 unambiguously defined by ENGINEs and used consistently across any
6149 OpenSSL-based application. Commands have been added to all the
6150 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6151 control over shared-library paths without source code alterations.
6154 *) Changed all ENGINE implementations to dynamically allocate their
6155 ENGINEs rather than declaring them statically. Apart from this being
6156 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6157 this also allows the implementations to compile without using the
6158 internal engine_int.h header.
6161 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6162 'const' value. Any code that should be able to modify a RAND_METHOD
6163 should already have non-const pointers to it (ie. they should only
6164 modify their own ones).
6167 *) Made a variety of little tweaks to the ENGINE code.
6168 - "atalla" and "ubsec" string definitions were moved from header files
6169 to C code. "nuron" string definitions were placed in variables
6170 rather than hard-coded - allowing parameterisation of these values
6171 later on via ctrl() commands.
6172 - Removed unused "#if 0"'d code.
6173 - Fixed engine list iteration code so it uses ENGINE_free() to release
6174 structural references.
6175 - Constified the RAND_METHOD element of ENGINE structures.
6176 - Constified various get/set functions as appropriate and added
6177 missing functions (including a catch-all ENGINE_cpy that duplicates
6178 all ENGINE values onto a new ENGINE except reference counts/state).
6179 - Removed NULL parameter checks in get/set functions. Setting a method
6180 or function to NULL is a way of cancelling out a previously set
6181 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6182 and doesn't justify the extra error symbols and code.
6183 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6184 flags from engine_int.h to engine.h.
6185 - Changed prototypes for ENGINE handler functions (init(), finish(),
6186 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6189 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6190 to the algorithm using long division. The binary algorithm can be
6191 used only if the modulus is odd. On 32-bit systems, it is faster
6192 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6193 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6194 up to 450 bits. In 64-bit environments, the binary algorithm
6195 appears to be advantageous for much longer moduli; here we use it
6196 for moduli up to 2048 bits.
6199 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6200 could not support the combine flag in choice fields.
6203 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6204 extensions from a certificate request to the certificate.
6207 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6208 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6209 file: this allows the display of the certificate about to be
6210 signed to be customised, to allow certain fields to be included
6211 or excluded and extension details. The old system didn't display
6212 multicharacter strings properly, omitted fields not in the policy
6213 and couldn't display additional details such as extensions.
6216 *) Function EC_POINTs_mul for multiple scalar multiplication
6217 of an arbitrary number of elliptic curve points
6218 \sum scalars[i]*points[i],
6219 optionally including the generator defined for the EC_GROUP:
6220 scalar*generator + \sum scalars[i]*points[i].
6222 EC_POINT_mul is a simple wrapper function for the typical case
6223 that the point list has just one item (besides the optional
6227 *) First EC_METHODs for curves over GF(p):
6229 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6230 operations and provides various method functions that can also
6231 operate with faster implementations of modular arithmetic.
6233 EC_GFp_mont_method() reuses most functions that are part of
6234 EC_GFp_simple_method, but uses Montgomery arithmetic.
6236 [Bodo Moeller; point addition and point doubling
6237 implementation directly derived from source code provided by
6238 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6240 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6241 crypto/ec/ec_lib.c):
6243 Curves are EC_GROUP objects (with an optional group generator)
6244 based on EC_METHODs that are built into the library.
6246 Points are EC_POINT objects based on EC_GROUP objects.
6248 Most of the framework would be able to handle curves over arbitrary
6249 finite fields, but as there are no obvious types for fields other
6250 than GF(p), some functions are limited to that for now.
6253 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6254 that the file contains a complete HTTP response.
6257 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6258 change the def and num file printf format specifier from "%-40sXXX"
6259 to "%-39s XXX". The latter will always guarantee a space after the
6260 field while the former will cause them to run together if the field
6261 is 40 of more characters long.
6264 *) Constify the cipher and digest 'method' functions and structures
6265 and modify related functions to take constant EVP_MD and EVP_CIPHER
6269 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6270 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6273 *) Modify EVP_Digest*() routines so they now return values. Although the
6274 internal software routines can never fail additional hardware versions
6278 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6280 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6281 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6284 ERR_R_NESTED_ASN1_ERROR
6286 ERR_R_MISSING_ASN1_EOS
6287 were 4 .. 9, conflicting with
6288 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6290 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6291 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6293 Add new error code 'ERR_R_INTERNAL_ERROR'.
6296 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6300 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6301 sets the subject name for a new request or supersedes the
6302 subject name in a given request. Formats that can be parsed are
6303 'CN=Some Name, OU=myOU, C=IT'
6305 'CN=Some Name/OU=myOU/C=IT'.
6307 Add options '-batch' and '-verbose' to 'openssl req'.
6308 [Massimiliano Pala <madwolf@hackmasters.net>]
6310 *) Introduce the possibility to access global variables through
6311 functions on platform were that's the best way to handle exporting
6312 global variables in shared libraries. To enable this functionality,
6313 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6314 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6315 is normally done by Configure or something similar).
6317 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6318 in the source file (foo.c) like this:
6320 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6321 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6323 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6324 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6326 OPENSSL_DECLARE_GLOBAL(int,foo);
6327 #define foo OPENSSL_GLOBAL_REF(foo)
6328 OPENSSL_DECLARE_GLOBAL(double,bar);
6329 #define bar OPENSSL_GLOBAL_REF(bar)
6331 The #defines are very important, and therefore so is including the
6332 header file everywhere where the defined globals are used.
6334 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6335 of ASN.1 items, but that structure is a bit different.
6337 The largest change is in util/mkdef.pl which has been enhanced with
6338 better and easier to understand logic to choose which symbols should
6339 go into the Windows .def files as well as a number of fixes and code
6340 cleanup (among others, algorithm keywords are now sorted
6341 lexicographically to avoid constant rewrites).
6344 *) In BN_div() keep a copy of the sign of 'num' before writing the
6345 result to 'rm' because if rm==num the value will be overwritten
6346 and produce the wrong result if 'num' is negative: this caused
6347 problems with BN_mod() and BN_nnmod().
6350 *) Function OCSP_request_verify(). This checks the signature on an
6351 OCSP request and verifies the signer certificate. The signer
6352 certificate is just checked for a generic purpose and OCSP request
6356 *) Add OCSP_check_validity() function to check the validity of OCSP
6357 responses. OCSP responses are prepared in real time and may only
6358 be a few seconds old. Simply checking that the current time lies
6359 between thisUpdate and nextUpdate max reject otherwise valid responses
6360 caused by either OCSP responder or client clock inaccuracy. Instead
6361 we allow thisUpdate and nextUpdate to fall within a certain period of
6362 the current time. The age of the response can also optionally be
6363 checked. Two new options -validity_period and -status_age added to
6367 *) If signature or public key algorithm is unrecognized print out its
6368 OID rather that just UNKNOWN.
6371 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6372 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6373 ID to be generated from the issuer certificate alone which can then be
6374 passed to OCSP_id_issuer_cmp().
6377 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6378 ASN1 modules to export functions returning ASN1_ITEM pointers
6379 instead of the ASN1_ITEM structures themselves. This adds several
6380 new macros which allow the underlying ASN1 function/structure to
6381 be accessed transparently. As a result code should not use ASN1_ITEM
6382 references directly (such as &X509_it) but instead use the relevant
6383 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6384 use of the new ASN1 code on platforms where exporting structures
6385 is problematical (for example in shared libraries) but exporting
6386 functions returning pointers to structures is not.
6389 *) Add support for overriding the generation of SSL/TLS session IDs.
6390 These callbacks can be registered either in an SSL_CTX or per SSL.
6391 The purpose of this is to allow applications to control, if they wish,
6392 the arbitrary values chosen for use as session IDs, particularly as it
6393 can be useful for session caching in multiple-server environments. A
6394 command-line switch for testing this (and any client code that wishes
6395 to use such a feature) has been added to "s_server".
6396 [Geoff Thorpe, Lutz Jaenicke]
6398 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6399 of the form '#if defined(...) || defined(...) || ...' and
6400 '#if !defined(...) && !defined(...) && ...'. This also avoids
6401 the growing number of special cases it was previously handling.
6404 *) Make all configuration macros available for application by making
6405 sure they are available in opensslconf.h, by giving them names starting
6406 with "OPENSSL_" to avoid conflicts with other packages and by making
6407 sure e_os2.h will cover all platform-specific cases together with
6409 Additionally, it is now possible to define configuration/platform-
6410 specific names (called "system identities"). In the C code, these
6411 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6412 macro with the name beginning with "OPENSSL_SYS_", which is determined
6413 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6417 *) New option -set_serial to 'req' and 'x509' this allows the serial
6418 number to use to be specified on the command line. Previously self
6419 signed certificates were hard coded with serial number 0 and the
6420 CA options of 'x509' had to use a serial number in a file which was
6424 *) New options to 'ca' utility to support V2 CRL entry extensions.
6425 Currently CRL reason, invalidity date and hold instruction are
6426 supported. Add new CRL extensions to V3 code and some new objects.
6429 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6430 disable standard block padding (aka PKCS#5 padding) in the EVP
6431 API, which was previously mandatory. This means that the data is
6432 not padded in any way and so the total length much be a multiple
6433 of the block size, otherwise an error occurs.
6436 *) Initial (incomplete) OCSP SSL support.
6439 *) New function OCSP_parse_url(). This splits up a URL into its host,
6440 port and path components: primarily to parse OCSP URLs. New -url
6441 option to ocsp utility.
6444 *) New nonce behavior. The return value of OCSP_check_nonce() now
6445 reflects the various checks performed. Applications can decide
6446 whether to tolerate certain situations such as an absent nonce
6447 in a response when one was present in a request: the ocsp application
6448 just prints out a warning. New function OCSP_add1_basic_nonce()
6449 this is to allow responders to include a nonce in a response even if
6450 the request is nonce-less.
6453 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6454 skipped when using openssl x509 multiple times on a single input file,
6455 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6458 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6459 set string type: to handle setting ASN1_TIME structures. Fix ca
6460 utility to correctly initialize revocation date of CRLs.
6463 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6464 the clients preferred ciphersuites and rather use its own preferences.
6465 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6466 Internet Explorer by ensuring unchanged hash method during stepup.
6467 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6470 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6471 to aes and add a new 'exist' option to print out symbols that don't
6475 *) Additional options to ocsp utility to allow flags to be set and
6476 additional certificates supplied.
6479 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6480 OCSP client a number of certificate to only verify the response
6484 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6485 handle the new API. Currently only ECB, CBC modes supported. Add new
6488 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6489 Encryption Standard (AES) Ciphersuites for Transport Layer
6490 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6491 not enabled by default and were not part of the "ALL" ciphersuite
6492 alias because they were not yet official; they could be
6493 explicitly requested by specifying the "AESdraft" ciphersuite
6494 group alias. In the final release of OpenSSL 0.9.7, the group
6495 alias is called "AES" and is part of "ALL".)
6496 [Ben Laurie, Steve Henson, Bodo Moeller]
6498 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6499 request to response.
6502 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6503 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6504 extract information from a certificate request. OCSP_response_create()
6505 creates a response and optionally adds a basic response structure.
6506 OCSP_basic_add1_status() adds a complete single response to a basic
6507 response and returns the OCSP_SINGLERESP structure just added (to allow
6508 extensions to be included for example). OCSP_basic_add1_cert() adds a
6509 certificate to a basic response and OCSP_basic_sign() signs a basic
6510 response with various flags. New helper functions ASN1_TIME_check()
6511 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6512 (converts ASN1_TIME to GeneralizedTime).
6515 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6516 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6517 structure from a certificate. X509_pubkey_digest() digests the public_key
6518 contents: this is used in various key identifiers.
6521 *) Make sk_sort() tolerate a NULL argument.
6522 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6524 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6525 passed by the function are trusted implicitly. If any of them signed the
6526 response then it is assumed to be valid and is not verified.
6529 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6530 to data. This was previously part of the PKCS7 ASN1 code. This
6531 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6532 [Steve Henson, reported by Kenneth R. Robinette
6533 <support@securenetterm.com>]
6535 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6536 routines: without these tracing memory leaks is very painful.
6537 Fix leaks in PKCS12 and PKCS7 routines.
6540 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6541 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6542 effectively meant GeneralizedTime would never be used. Now it
6543 is initialised to -1 but X509_time_adj() now has to check the value
6544 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6545 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6546 [Steve Henson, reported by Kenneth R. Robinette
6547 <support@securenetterm.com>]
6549 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6550 result in a zero length in the ASN1_INTEGER structure which was
6551 not consistent with the structure when d2i_ASN1_INTEGER() was used
6552 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6553 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6554 where it did not print out a minus for negative ASN1_INTEGER.
6557 *) Add summary printout to ocsp utility. The various functions which
6558 convert status values to strings have been renamed to:
6559 OCSP_response_status_str(), OCSP_cert_status_str() and
6560 OCSP_crl_reason_str() and are no longer static. New options
6561 to verify nonce values and to disable verification. OCSP response
6562 printout format cleaned up.
6565 *) Add additional OCSP certificate checks. These are those specified
6566 in RFC2560. This consists of two separate checks: the CA of the
6567 certificate being checked must either be the OCSP signer certificate
6568 or the issuer of the OCSP signer certificate. In the latter case the
6569 OCSP signer certificate must contain the OCSP signing extended key
6570 usage. This check is performed by attempting to match the OCSP
6571 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6572 in the OCSP_CERTID structures of the response.
6575 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6576 and related routines. This uses the standard OpenSSL certificate
6577 verify routines to perform initial checks (just CA validity) and
6578 to obtain the certificate chain. Then additional checks will be
6579 performed on the chain. Currently the root CA is checked to see
6580 if it is explicitly trusted for OCSP signing. This is used to set
6581 a root CA as a global signing root: that is any certificate that
6582 chains to that CA is an acceptable OCSP signing certificate.
6585 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6586 extensions from a separate configuration file.
6587 As when reading extensions from the main configuration file,
6588 the '-extensions ...' option may be used for specifying the
6590 [Massimiliano Pala <madwolf@comune.modena.it>]
6592 *) New OCSP utility. Allows OCSP requests to be generated or
6593 read. The request can be sent to a responder and the output
6594 parsed, outputed or printed in text form. Not complete yet:
6595 still needs to check the OCSP response validity.
6598 *) New subcommands for 'openssl ca':
6599 'openssl ca -status <serial>' prints the status of the cert with
6600 the given serial number (according to the index file).
6601 'openssl ca -updatedb' updates the expiry status of certificates
6603 [Massimiliano Pala <madwolf@comune.modena.it>]
6605 *) New '-newreq-nodes' command option to CA.pl. This is like
6606 '-newreq', but calls 'openssl req' with the '-nodes' option
6607 so that the resulting key is not encrypted.
6608 [Damien Miller <djm@mindrot.org>]
6610 *) New configuration for the GNU Hurd.
6611 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6613 *) Initial code to implement OCSP basic response verify. This
6614 is currently incomplete. Currently just finds the signer's
6615 certificate and verifies the signature on the response.
6618 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6619 value of OPENSSLDIR. This is available via the new '-d' option
6620 to 'openssl version', and is also included in 'openssl version -a'.
6623 *) Allowing defining memory allocation callbacks that will be given
6624 file name and line number information in additional arguments
6625 (a const char* and an int). The basic functionality remains, as
6626 well as the original possibility to just replace malloc(),
6627 realloc() and free() by functions that do not know about these
6628 additional arguments. To register and find out the current
6629 settings for extended allocation functions, the following
6630 functions are provided:
6632 CRYPTO_set_mem_ex_functions
6633 CRYPTO_set_locked_mem_ex_functions
6634 CRYPTO_get_mem_ex_functions
6635 CRYPTO_get_locked_mem_ex_functions
6637 These work the same way as CRYPTO_set_mem_functions and friends.
6638 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6639 extended allocation function is enabled.
6640 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6641 a conventional allocation function is enabled.
6642 [Richard Levitte, Bodo Moeller]
6644 *) Finish off removing the remaining LHASH function pointer casts.
6645 There should no longer be any prototype-casting required when using
6646 the LHASH abstraction, and any casts that remain are "bugs". See
6647 the callback types and macros at the head of lhash.h for details
6648 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6651 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6652 If /dev/[u]random devices are not available or do not return enough
6653 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6655 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6656 /etc/entropy will be queried once each in this sequence, quering stops
6657 when enough entropy was collected without querying more sockets.
6660 *) Change the Unix RAND_poll() variant to be able to poll several
6661 random devices, as specified by DEVRANDOM, until a sufficient amount
6662 of data has been collected. We spend at most 10 ms on each file
6663 (select timeout) and read in non-blocking mode. DEVRANDOM now
6664 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6665 (previously it was just the string "/dev/urandom"), so on typical
6666 platforms the 10 ms delay will never occur.
6667 Also separate out the Unix variant to its own file, rand_unix.c.
6668 For VMS, there's a currently-empty rand_vms.c.
6671 *) Move OCSP client related routines to ocsp_cl.c. These
6672 provide utility functions which an application needing
6673 to issue a request to an OCSP responder and analyse the
6674 response will typically need: as opposed to those which an
6675 OCSP responder itself would need which will be added later.
6677 OCSP_request_sign() signs an OCSP request with an API similar
6678 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6679 response. OCSP_response_get1_basic() extracts basic response
6680 from response. OCSP_resp_find_status(): finds and extracts status
6681 information from an OCSP_CERTID structure (which will be created
6682 when the request structure is built). These are built from lower
6683 level functions which work on OCSP_SINGLERESP structures but
6684 wont normally be used unless the application wishes to examine
6685 extensions in the OCSP response for example.
6687 Replace nonce routines with a pair of functions.
6688 OCSP_request_add1_nonce() adds a nonce value and optionally
6689 generates a random value. OCSP_check_nonce() checks the
6690 validity of the nonce in an OCSP response.
6693 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6694 This doesn't copy the supplied OCSP_CERTID and avoids the
6695 need to free up the newly created id. Change return type
6696 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6697 This can then be used to add extensions to the request.
6698 Deleted OCSP_request_new(), since most of its functionality
6699 is now in OCSP_REQUEST_new() (and the case insensitive name
6700 clash) apart from the ability to set the request name which
6701 will be added elsewhere.
6704 *) Update OCSP API. Remove obsolete extensions argument from
6705 various functions. Extensions are now handled using the new
6706 OCSP extension code. New simple OCSP HTTP function which
6707 can be used to send requests and parse the response.
6710 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6711 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6712 uses the special reorder version of SET OF to sort the attributes
6713 and reorder them to match the encoded order. This resolves a long
6714 standing problem: a verify on a PKCS7 structure just after signing
6715 it used to fail because the attribute order did not match the
6716 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6717 it uses the received order. This is necessary to tolerate some broken
6718 software that does not order SET OF. This is handled by encoding
6719 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6720 to produce the required SET OF.
6723 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6724 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6725 files to get correct declarations of the ASN.1 item variables.
6728 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6729 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6730 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6731 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6732 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6733 ASN1_ITEM and no wrapper functions.
6736 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6737 replace the old function pointer based I/O routines. Change most of
6738 the *_d2i_bio() and *_d2i_fp() functions to use these.
6741 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6742 lines, recognice more "algorithms" that can be deselected, and make
6743 it complain about algorithm deselection that isn't recognised.
6746 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6747 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6748 to use new functions. Add NO_ASN1_OLD which can be set to remove
6749 some old style ASN1 functions: this can be used to determine if old
6750 code will still work when these eventually go away.
6753 *) New extension functions for OCSP structures, these follow the
6754 same conventions as certificates and CRLs.
6757 *) New function X509V3_add1_i2d(). This automatically encodes and
6758 adds an extension. Its behaviour can be customised with various
6759 flags to append, replace or delete. Various wrappers added for
6760 certifcates and CRLs.
6763 *) Fix to avoid calling the underlying ASN1 print routine when
6764 an extension cannot be parsed. Correct a typo in the
6765 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6768 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6769 entries for variables.
6772 *) Add functionality to apps/openssl.c for detecting locking
6773 problems: As the program is single-threaded, all we have
6774 to do is register a locking callback using an array for
6775 storing which locks are currently held by the program.
6778 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6779 SSL_get_ex_data_X509_STORE_idx(), which is used in
6780 ssl_verify_cert_chain() and thus can be called at any time
6781 during TLS/SSL handshakes so that thread-safety is essential.
6782 Unfortunately, the ex_data design is not at all suited
6783 for multi-threaded use, so it probably should be abolished.
6786 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6787 [Broadcom, tweaked and integrated by Geoff Thorpe]
6789 *) Move common extension printing code to new function
6790 X509V3_print_extensions(). Reorganise OCSP print routines and
6791 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6794 *) New function X509_signature_print() to remove duplication in some
6798 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6799 set (this was treated exactly the same as SET OF previously). This
6800 is used to reorder the STACK representing the structure to match the
6801 encoding. This will be used to get round a problem where a PKCS7
6802 structure which was signed could not be verified because the STACK
6803 order did not reflect the encoded order.
6806 *) Reimplement the OCSP ASN1 module using the new code.
6809 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6810 for its ASN1 operations. The old style function pointers still exist
6811 for now but they will eventually go away.
6814 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6815 completely replaces the old ASN1 functionality with a table driven
6816 encoder and decoder which interprets an ASN1_ITEM structure describing
6817 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6818 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6819 has also been converted to the new form.
6822 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6823 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6824 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6825 for negative moduli.
6828 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6829 of not touching the result's sign bit.
6832 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6836 *) Changed the LHASH code to use prototypes for callbacks, and created
6837 macros to declare and implement thin (optionally static) functions
6838 that provide type-safety and avoid function pointer casting for the
6839 type-specific callbacks.
6842 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6844 [Veers Staats <staatsvr@asc.hpc.mil>,
6845 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6847 *) Reformat the FAQ so the different questions and answers can be divided
6848 in sections depending on the subject.
6851 *) Have the zlib compression code load ZLIB.DLL dynamically under
6855 *) New function BN_mod_sqrt for computing square roots modulo a prime
6856 (using the probabilistic Tonelli-Shanks algorithm unless
6857 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6858 be handled deterministically).
6859 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6861 *) Make BN_mod_inverse faster by explicitly handling small quotients
6862 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6863 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6866 *) New function BN_kronecker.
6869 *) Fix BN_gcd so that it works on negative inputs; the result is
6870 positive unless both parameters are zero.
6871 Previously something reasonably close to an infinite loop was
6872 possible because numbers could be growing instead of shrinking
6873 in the implementation of Euclid's algorithm.
6876 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6877 sign of the number in question.
6879 Fix BN_is_word(a,w) to work correctly for w == 0.
6881 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6882 because its test if the absolute value of 'a' equals 'w'.
6883 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6884 it exists mostly for use in the implementations of BN_is_zero(),
6885 BN_is_one(), and BN_is_word().
6888 *) New function BN_swap.
6891 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6892 the exponentiation functions are more likely to produce reasonable
6893 results on negative inputs.
6896 *) Change BN_mod_mul so that the result is always non-negative.
6897 Previously, it could be negative if one of the factors was negative;
6898 I don't think anyone really wanted that behaviour.
6901 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6902 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6903 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6904 and add new functions:
6913 BN_mod_lshift1_quick
6917 These functions always generate non-negative results.
6919 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6920 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6922 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6923 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6924 be reduced modulo m.
6925 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6928 The following entry accidentally appeared in the CHANGES file
6929 distributed with OpenSSL 0.9.7. The modifications described in
6930 it do *not* apply to OpenSSL 0.9.7.
6932 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6933 was actually never needed) and in BN_mul(). The removal in BN_mul()
6934 required a small change in bn_mul_part_recursive() and the addition
6935 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6936 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6937 bn_sub_words() and bn_add_words() except they take arrays with
6942 *) In 'openssl passwd', verify passwords read from the terminal
6943 unless the '-salt' option is used (which usually means that
6944 verification would just waste user's time since the resulting
6945 hash is going to be compared with some given password hash)
6946 or the new '-noverify' option is used.
6948 This is an incompatible change, but it does not affect
6949 non-interactive use of 'openssl passwd' (passwords on the command
6950 line, '-stdin' option, '-in ...' option) and thus should not
6954 *) Remove all references to RSAref, since there's no more need for it.
6957 *) Make DSO load along a path given through an environment variable
6958 (SHLIB_PATH) with shl_load().
6961 *) Constify the ENGINE code as a result of BIGNUM constification.
6962 Also constify the RSA code and most things related to it. In a
6963 few places, most notable in the depth of the ASN.1 code, ugly
6964 casts back to non-const were required (to be solved at a later
6968 *) Make it so the openssl application has all engines loaded by default.
6971 *) Constify the BIGNUM routines a little more.
6974 *) Add the following functions:
6976 ENGINE_load_cswift()
6978 ENGINE_load_atalla()
6980 ENGINE_load_builtin_engines()
6982 That way, an application can itself choose if external engines that
6983 are built-in in OpenSSL shall ever be used or not. The benefit is
6984 that applications won't have to be linked with libdl or other dso
6985 libraries unless it's really needed.
6987 Changed 'openssl engine' to load all engines on demand.
6988 Changed the engine header files to avoid the duplication of some
6989 declarations (they differed!).
6992 *) 'openssl engine' can now list capabilities.
6995 *) Better error reporting in 'openssl engine'.
6998 *) Never call load_dh_param(NULL) in s_server.
7001 *) Add engine application. It can currently list engines by name and
7002 identity, and test if they are actually available.
7005 *) Improve RPM specification file by forcing symbolic linking and making
7006 sure the installed documentation is also owned by root.root.
7007 [Damien Miller <djm@mindrot.org>]
7009 *) Give the OpenSSL applications more possibilities to make use of
7010 keys (public as well as private) handled by engines.
7013 *) Add OCSP code that comes from CertCo.
7016 *) Add VMS support for the Rijndael code.
7019 *) Added untested support for Nuron crypto accelerator.
7022 *) Add support for external cryptographic devices. This code was
7023 previously distributed separately as the "engine" branch.
7024 [Geoff Thorpe, Richard Levitte]
7026 *) Rework the filename-translation in the DSO code. It is now possible to
7027 have far greater control over how a "name" is turned into a filename
7028 depending on the operating environment and any oddities about the
7029 different shared library filenames on each system.
7032 *) Support threads on FreeBSD-elf in Configure.
7035 *) Fix for SHA1 assembly problem with MASM: it produces
7036 warnings about corrupt line number information when assembling
7037 with debugging information. This is caused by the overlapping
7039 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7042 NCONF_get_number() has no error checking at all. As a replacement,
7043 NCONF_get_number_e() is defined (_e for "error checking") and is
7044 promoted strongly. The old NCONF_get_number is kept around for
7045 binary backward compatibility.
7046 Make it possible for methods to load from something other than a BIO,
7047 by providing a function pointer that is given a name instead of a BIO.
7048 For example, this could be used to load configuration data from an
7052 *) Fix for non blocking accept BIOs. Added new I/O special reason
7053 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7054 with non blocking I/O was not possible because no retry code was
7055 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7059 *) Added the beginnings of Rijndael support.
7062 *) Fix for bug in DirectoryString mask setting. Add support for
7063 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7064 to allow certificate printing to more controllable, additional
7065 'certopt' option to 'x509' to allow new printing options to be
7069 *) Clean old EAY MD5 hack from e_os.h.
7072 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7074 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7075 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7076 [Joe Orton, Steve Henson]
7078 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7080 *) Fix additional bug revealed by the NISCC test suite:
7082 Stop bug triggering large recursion when presented with
7083 certain ASN.1 tags (CVE-2003-0851)
7086 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7088 *) Fix various bugs revealed by running the NISCC test suite:
7090 Stop out of bounds reads in the ASN1 code when presented with
7091 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7093 If verify callback ignores invalid public key errors don't try to check
7094 certificate signature with the NULL public key.
7098 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7099 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7103 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7104 extra data after the compression methods not only for TLS 1.0
7105 but also for SSL 3.0 (as required by the specification).
7106 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7108 *) Change X509_certificate_type() to mark the key as exported/exportable
7109 when it's 512 *bits* long, not 512 bytes.
7112 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7114 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7115 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7116 a protocol version number mismatch like a decryption error
7117 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7120 *) Turn on RSA blinding by default in the default implementation
7121 to avoid a timing attack. Applications that don't want it can call
7122 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7123 They would be ill-advised to do so in most cases.
7124 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7126 *) Change RSA blinding code so that it works when the PRNG is not
7127 seeded (in this case, the secret RSA exponent is abused as
7128 an unpredictable seed -- if it is not unpredictable, there
7129 is no point in blinding anyway). Make RSA blinding thread-safe
7130 by remembering the creator's thread ID in rsa->blinding and
7131 having all other threads use local one-time blinding factors
7132 (this requires more computation than sharing rsa->blinding, but
7133 avoids excessive locking; and if an RSA object is not shared
7134 between threads, blinding will still be very fast).
7137 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7139 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7140 via timing by performing a MAC computation even if incorrrect
7141 block cipher padding has been found. This is a countermeasure
7142 against active attacks where the attacker has to distinguish
7143 between bad padding and a MAC verification error. (CVE-2003-0078)
7145 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7146 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7147 Martin Vuagnoux (EPFL, Ilion)]
7149 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7151 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7152 memory from it's contents. This is done with a counter that will
7153 place alternating values in each byte. This can be used to solve
7154 two issues: 1) the removal of calls to memset() by highly optimizing
7155 compilers, and 2) cleansing with other values than 0, since those can
7156 be read through on certain media, for example a swap space on disk.
7159 *) Bugfix: client side session caching did not work with external caching,
7160 because the session->cipher setting was not restored when reloading
7161 from the external cache. This problem was masked, when
7162 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7163 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7166 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7167 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7168 [Zeev Lieber <zeev-l@yahoo.com>]
7170 *) Undo an undocumented change introduced in 0.9.6e which caused
7171 repeated calls to OpenSSL_add_all_ciphers() and
7172 OpenSSL_add_all_digests() to be ignored, even after calling
7176 *) Change the default configuration reader to deal with last line not
7177 being properly terminated.
7180 *) Change X509_NAME_cmp() so it applies the special rules on handling
7181 DN values that are of type PrintableString, as well as RDNs of type
7182 emailAddress where the value has the type ia5String.
7183 [stefank@valicert.com via Richard Levitte]
7185 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7186 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7187 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7188 the bitwise-OR of the two for use by the majority of applications
7189 wanting this behaviour, and update the docs. The documented
7190 behaviour and actual behaviour were inconsistent and had been
7191 changing anyway, so this is more a bug-fix than a behavioural
7193 [Geoff Thorpe, diagnosed by Nadav Har'El]
7195 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7196 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7199 *) Fix initialization code race conditions in
7200 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7201 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7202 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7203 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7204 ssl2_get_cipher_by_char(),
7205 ssl3_get_cipher_by_char().
7206 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7208 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7209 the cached sessions are flushed, as the remove_cb() might use ex_data
7210 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7211 (see [openssl.org #212]).
7212 [Geoff Thorpe, Lutz Jaenicke]
7214 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7215 length, instead of the encoding length to d2i_ASN1_OBJECT.
7218 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7220 *) [In 0.9.6g-engine release:]
7221 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7222 [Lynn Gazis <lgazis@rainbow.com>]
7224 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7226 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7227 and get fix the header length calculation.
7228 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7229 Alon Kantor <alonk@checkpoint.com> (and others),
7232 *) Use proper error handling instead of 'assertions' in buffer
7233 overflow checks added in 0.9.6e. This prevents DoS (the
7234 assertions could call abort()).
7235 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7237 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7239 *) Add various sanity checks to asn1_get_length() to reject
7240 the ASN1 length bytes if they exceed sizeof(long), will appear
7241 negative or the content length exceeds the length of the
7243 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7245 *) Fix cipher selection routines: ciphers without encryption had no flags
7246 for the cipher strength set and where therefore not handled correctly
7247 by the selection routines (PR #130).
7250 *) Fix EVP_dsa_sha macro.
7254 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7255 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7256 that was added in OpenSSL 0.9.6d.
7258 As the countermeasure turned out to be incompatible with some
7259 broken SSL implementations, the new option is part of SSL_OP_ALL.
7260 SSL_OP_ALL is usually employed when compatibility with weird SSL
7261 implementations is desired (e.g. '-bugs' option to 's_client' and
7262 's_server'), so the new option is automatically set in many
7266 *) Changes in security patch:
7268 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7269 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7270 Air Force Materiel Command, USAF, under agreement number
7273 *) Add various sanity checks to asn1_get_length() to reject
7274 the ASN1 length bytes if they exceed sizeof(long), will appear
7275 negative or the content length exceeds the length of the
7276 supplied buffer. (CVE-2002-0659)
7277 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7279 *) Assertions for various potential buffer overflows, not known to
7281 [Ben Laurie (CHATS)]
7283 *) Various temporary buffers to hold ASCII versions of integers were
7284 too small for 64 bit platforms. (CVE-2002-0655)
7285 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7287 *) Remote buffer overflow in SSL3 protocol - an attacker could
7288 supply an oversized session ID to a client. (CVE-2002-0656)
7289 [Ben Laurie (CHATS)]
7291 *) Remote buffer overflow in SSL2 protocol - an attacker could
7292 supply an oversized client master key. (CVE-2002-0656)
7293 [Ben Laurie (CHATS)]
7295 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7297 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7298 encoded as NULL) with id-dsa-with-sha1.
7299 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7301 *) Check various X509_...() return values in apps/req.c.
7302 [Nils Larsch <nla@trustcenter.de>]
7304 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7305 an end-of-file condition would erronously be flagged, when the CRLF
7306 was just at the end of a processed block. The bug was discovered when
7307 processing data through a buffering memory BIO handing the data to a
7308 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7309 <ptsekov@syntrex.com> and Nedelcho Stanev.
7312 *) Implement a countermeasure against a vulnerability recently found
7313 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7314 before application data chunks to avoid the use of known IVs
7315 with data potentially chosen by the attacker.
7318 *) Fix length checks in ssl3_get_client_hello().
7321 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7322 to prevent ssl3_read_internal() from incorrectly assuming that
7323 ssl3_read_bytes() found application data while handshake
7324 processing was enabled when in fact s->s3->in_read_app_data was
7325 merely automatically cleared during the initial handshake.
7326 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7328 *) Fix object definitions for Private and Enterprise: they were not
7329 recognized in their shortname (=lowercase) representation. Extend
7330 obj_dat.pl to issue an error when using undefined keywords instead
7331 of silently ignoring the problem (Svenning Sorensen
7332 <sss@sss.dnsalias.net>).
7335 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7336 generators, i.e. generators other than 2 and 5. (Previously, the
7337 code did not properly initialise the 'add' and 'rem' values to
7338 BN_generate_prime().)
7340 In the new general case, we do not insist that 'generator' is
7341 actually a primitive root: This requirement is rather pointless;
7342 a generator of the order-q subgroup is just as good, if not
7346 *) Map new X509 verification errors to alerts. Discovered and submitted by
7347 Tom Wu <tom@arcot.com>.
7350 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7351 returning non-zero before the data has been completely received
7352 when using non-blocking I/O.
7353 [Bodo Moeller; problem pointed out by John Hughes]
7355 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7356 [Ben Laurie, Lutz Jaenicke]
7358 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7359 Yoram Zahavi <YoramZ@gilian.com>).
7362 *) Add information about CygWin 1.3 and on, and preserve proper
7363 configuration for the versions before that.
7364 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7366 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7367 check whether we deal with a copy of a session and do not delete from
7368 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7369 <izhar@checkpoint.com>.
7372 *) Do not store session data into the internal session cache, if it
7373 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7374 flag is set). Proposed by Aslam <aslam@funk.com>.
7377 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7381 *) [In 0.9.6d-engine release:]
7382 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7383 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7385 *) Add the configuration target linux-s390x.
7386 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7388 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7389 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7390 variable as an indication that a ClientHello message has been
7391 received. As the flag value will be lost between multiple
7392 invocations of ssl3_accept when using non-blocking I/O, the
7393 function may not be aware that a handshake has actually taken
7394 place, thus preventing a new session from being added to the
7397 To avoid this problem, we now set s->new_session to 2 instead of
7398 using a local variable.
7399 [Lutz Jaenicke, Bodo Moeller]
7401 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7402 if the SSL_R_LENGTH_MISMATCH error is detected.
7403 [Geoff Thorpe, Bodo Moeller]
7405 *) New 'shared_ldflag' column in Configure platform table.
7408 *) Fix EVP_CIPHER_mode macro.
7409 ["Dan S. Camper" <dan@bti.net>]
7411 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7412 type, we must throw them away by setting rr->length to 0.
7413 [D P Chang <dpc@qualys.com>]
7415 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7417 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7418 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7419 worked incorrectly for those cases where range = 10..._2 and
7420 3*range is two bits longer than range.)
7423 *) Only add signing time to PKCS7 structures if it is not already
7427 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7428 OBJ_ld_ce should be OBJ_id_ce.
7429 Also some ip-pda OIDs in crypto/objects/objects.txt were
7430 incorrect (cf. RFC 3039).
7431 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7433 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7434 returns early because it has nothing to do.
7435 [Andy Schneider <andy.schneider@bjss.co.uk>]
7437 *) [In 0.9.6c-engine release:]
7438 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7439 [Andy Schneider <andy.schneider@bjss.co.uk>]
7441 *) [In 0.9.6c-engine release:]
7442 Add support for Cryptographic Appliance's keyserver technology.
7443 (Use engine 'keyclient')
7444 [Cryptographic Appliances and Geoff Thorpe]
7446 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7447 is called via tools/c89.sh because arguments have to be
7448 rearranged (all '-L' options must appear before the first object
7450 [Richard Shapiro <rshapiro@abinitio.com>]
7452 *) [In 0.9.6c-engine release:]
7453 Add support for Broadcom crypto accelerator cards, backported
7455 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7457 *) [In 0.9.6c-engine release:]
7458 Add support for SureWare crypto accelerator cards from
7459 Baltimore Technologies. (Use engine 'sureware')
7460 [Baltimore Technologies and Mark Cox]
7462 *) [In 0.9.6c-engine release:]
7463 Add support for crypto accelerator cards from Accelerated
7464 Encryption Processing, www.aep.ie. (Use engine 'aep')
7465 [AEP Inc. and Mark Cox]
7467 *) Add a configuration entry for gcc on UnixWare.
7468 [Gary Benson <gbenson@redhat.com>]
7470 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7471 messages are stored in a single piece (fixed-length part and
7472 variable-length part combined) and fix various bugs found on the way.
7475 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7476 instead. BIO_gethostbyname() does not know what timeouts are
7477 appropriate, so entries would stay in cache even when they have
7479 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7481 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7482 faced with a pathologically small ClientHello fragment that does
7483 not contain client_version: Instead of aborting with an error,
7484 simply choose the highest available protocol version (i.e.,
7485 TLS 1.0 unless it is disabled). In practice, ClientHello
7486 messages are never sent like this, but this change gives us
7487 strictly correct behaviour at least for TLS.
7490 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7491 never resets s->method to s->ctx->method when called from within
7492 one of the SSL handshake functions.
7493 [Bodo Moeller; problem pointed out by Niko Baric]
7495 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7496 (sent using the client's version number) if client_version is
7497 smaller than the protocol version in use. Also change
7498 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7499 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7500 the client will at least see that alert.
7503 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7507 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7508 client receives HelloRequest while in a handshake.
7509 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7511 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7512 should end in 'break', not 'goto end' which circumvents various
7513 cleanups done in state SSL_ST_OK. But session related stuff
7514 must be disabled for SSL_ST_OK in the case that we just sent a
7517 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7518 before just sending a HelloRequest.
7519 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7521 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7522 reveal whether illegal block cipher padding was found or a MAC
7523 verification error occured. (Neither SSLerr() codes nor alerts
7524 are directly visible to potential attackers, but the information
7525 may leak via logfiles.)
7527 Similar changes are not required for the SSL 2.0 implementation
7528 because the number of padding bytes is sent in clear for SSL 2.0,
7529 and the extra bytes are just ignored. However ssl/s2_pkt.c
7530 failed to verify that the purported number of padding bytes is in
7534 *) Add OpenUNIX-8 support including shared libraries
7535 (Boyd Lynn Gerber <gerberb@zenez.com>).
7538 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7539 'wristwatch attack' using huge encoding parameters (cf.
7540 James H. Manger's CRYPTO 2001 paper). Note that the
7541 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7542 encoding parameters and hence was not vulnerable.
7545 *) BN_sqr() bug fix.
7546 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7548 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7549 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7550 followed by modular reduction.
7551 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7553 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7554 equivalent based on BN_pseudo_rand() instead of BN_rand().
7557 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7558 This function was broken, as the check for a new client hello message
7559 to handle SGC did not allow these large messages.
7560 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7563 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7566 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7567 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7570 *) Rework the configuration and shared library support for Tru64 Unix.
7571 The configuration part makes use of modern compiler features and
7572 still retains old compiler behavior for those that run older versions
7573 of the OS. The shared library support part includes a variant that
7574 uses the RPATH feature, and is available through the special
7575 configuration target "alpha-cc-rpath", which will never be selected
7577 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7579 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7580 with the same message size as in ssl3_get_certificate_request().
7581 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7582 messages might inadvertently be reject as too long.
7583 [Petr Lampa <lampa@fee.vutbr.cz>]
7585 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7588 *) Modified SSL library such that the verify_callback that has been set
7589 specificly for an SSL object with SSL_set_verify() is actually being
7590 used. Before the change, a verify_callback set with this function was
7591 ignored and the verify_callback() set in the SSL_CTX at the time of
7592 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7593 to allow the necessary settings.
7596 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7597 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7598 done automatically (in contradiction to the requirements of the C
7599 standard). This made problems when used from OpenSSH.
7602 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7603 dh->length and always used
7605 BN_rand_range(priv_key, dh->p).
7607 BN_rand_range() is not necessary for Diffie-Hellman, and this
7608 specific range makes Diffie-Hellman unnecessarily inefficient if
7609 dh->length (recommended exponent length) is much smaller than the
7610 length of dh->p. We could use BN_rand_range() if the order of
7611 the subgroup was stored in the DH structure, but we only have
7616 BN_rand(priv_key, l, ...)
7618 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7624 RSA_eay_public_encrypt
7625 RSA_eay_private_decrypt
7626 RSA_eay_private_encrypt (signing)
7627 RSA_eay_public_decrypt (signature verification)
7629 (default implementations for RSA_public_encrypt,
7630 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7631 always reject numbers >= n.
7634 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7635 to synchronize access to 'locking_thread'. This is necessary on
7636 systems where access to 'locking_thread' (an 'unsigned long'
7637 variable) is not atomic.
7640 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7641 *before* setting the 'crypto_lock_rand' flag. The previous code had
7642 a race condition if 0 is a valid thread ID.
7643 [Travis Vitek <vitek@roguewave.com>]
7645 *) Add support for shared libraries under Irix.
7646 [Albert Chin-A-Young <china@thewrittenword.com>]
7648 *) Add configuration option to build on Linux on both big-endian and
7650 [Ralf Baechle <ralf@uni-koblenz.de>]
7652 *) Add the possibility to create shared libraries on HP-UX.
7655 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7657 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7658 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7659 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7660 PRNG state recovery was possible based on the output of
7661 one PRNG request appropriately sized to gain knowledge on
7662 'md' followed by enough consecutive 1-byte PRNG requests
7663 to traverse all of 'state'.
7665 1. When updating 'md_local' (the current thread's copy of 'md')
7666 during PRNG output generation, hash all of the previous
7667 'md_local' value, not just the half used for PRNG output.
7669 2. Make the number of bytes from 'state' included into the hash
7670 independent from the number of PRNG bytes requested.
7672 The first measure alone would be sufficient to avoid
7673 Markku-Juhani's attack. (Actually it had never occurred
7674 to me that the half of 'md_local' used for chaining was the
7675 half from which PRNG output bytes were taken -- I had always
7676 assumed that the secret half would be used.) The second
7677 measure makes sure that additional data from 'state' is never
7678 mixed into 'md_local' in small portions; this heuristically
7679 further strengthens the PRNG.
7682 *) Fix crypto/bn/asm/mips3.s.
7685 *) When only the key is given to "enc", the IV is undefined. Print out
7686 an error message in this case.
7689 *) Handle special case when X509_NAME is empty in X509 printing routines.
7692 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7693 positive and less than q.
7696 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7697 used: it isn't thread safe and the add_lock_callback should handle
7699 [Paul Rose <Paul.Rose@bridge.com>]
7701 *) Verify that incoming data obeys the block size in
7702 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7706 [Ulf Möller, Bodo Möller]
7708 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7709 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7710 when fixing the server behaviour for backwards-compatible 'client
7711 hello' messages. (Note that the attack is impractical against
7712 SSL 3.0 and TLS 1.0 anyway because length and version checking
7713 means that the probability of guessing a valid ciphertext is
7714 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7717 Before 0.9.5, the countermeasure (hide the error by generating a
7718 random 'decryption result') did not work properly because
7719 ERR_clear_error() was missing, meaning that SSL_get_error() would
7720 detect the supposedly ignored error.
7722 Both problems are now fixed.
7725 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7726 (previously it was 1024).
7729 *) Fix for compatibility mode trust settings: ignore trust settings
7730 unless some valid trust or reject settings are present.
7733 *) Fix for blowfish EVP: its a variable length cipher.
7736 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7737 parameters in DSA public key structures and return an error in the
7738 DSA routines if parameters are absent.
7741 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7742 in the current directory if neither $RANDFILE nor $HOME was set.
7743 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7744 caused some confusion to Windows users who haven't defined $HOME.
7745 Thus RAND_file_name() is changed again: e_os.h can define a
7746 DEFAULT_HOME, which will be used if $HOME is not set.
7747 For Windows, we use "C:"; on other platforms, we still require
7748 environment variables.
7750 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7751 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7752 having multiple threads call RAND_poll() concurrently.
7755 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7756 combination of a flag and a thread ID variable.
7757 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7758 flag), *other* threads can enter ssleay_add_bytes without obeying
7759 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7760 that they do not hold after the first thread unsets add_do_not_lock).
7763 *) Change bctest again: '-x' expressions are not available in all
7767 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7769 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7770 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7772 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7773 the default extension for executables, if any. Also, make the perl
7774 scripts that use symlink() to test if it really exists and use "cp"
7775 if it doesn't. All this made OpenSSL compilable and installable in
7779 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7780 If SEQUENCE is length is indefinite just set c->slen to the total
7781 amount of data available.
7782 [Steve Henson, reported by shige@FreeBSD.org]
7783 [This change does not apply to 0.9.7.]
7785 *) Change bctest to avoid here-documents inside command substitution
7786 (workaround for FreeBSD /bin/sh bug).
7787 For compatibility with Ultrix, avoid shell functions (introduced
7788 in the bctest version that searches along $PATH).
7791 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7792 with des_encrypt() defined on some operating systems, like Solaris
7796 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7797 On the Importance of Eliminating Errors in Cryptographic
7798 Computations, J. Cryptology 14 (2001) 2, 101-119,
7799 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7802 *) MIPS assembler BIGNUM division bug fix.
7805 *) Disabled incorrect Alpha assembler code.
7808 *) Fix PKCS#7 decode routines so they correctly update the length
7809 after reading an EOC for the EXPLICIT tag.
7811 [This change does not apply to 0.9.7.]
7813 *) Fix bug in PKCS#12 key generation routines. This was triggered
7814 if a 3DES key was generated with a 0 initial byte. Include
7815 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7816 (but broken) behaviour.
7819 *) Enhance bctest to search for a working bc along $PATH and print
7821 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7823 *) Fix memory leaks in err.c: free err_data string if necessary;
7824 don't write to the wrong index in ERR_set_error_data.
7827 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7831 *) Replace rdtsc with _emit statements for VC++ version 5.
7832 [Jeremy Cooper <jeremy@baymoo.org>]
7834 *) Make it possible to reuse SSLv2 sessions.
7837 *) In copy_email() check for >= 0 as a return value for
7838 X509_NAME_get_index_by_NID() since 0 is a valid index.
7839 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7841 *) Avoid coredump with unsupported or invalid public keys by checking if
7842 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7843 PKCS7_verify() fails with non detached data.
7846 *) Don't use getenv in library functions when run as setuid/setgid.
7847 New function OPENSSL_issetugid().
7850 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7851 due to incorrect handling of multi-threading:
7853 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7855 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7857 3. Count how many times MemCheck_off() has been called so that
7858 nested use can be treated correctly. This also avoids
7859 inband-signalling in the previous code (which relied on the
7860 assumption that thread ID 0 is impossible).
7863 *) Add "-rand" option also to s_client and s_server.
7866 *) Fix CPU detection on Irix 6.x.
7867 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7868 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7870 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7873 [This change does not apply to 0.9.7.]
7875 *) Use the cached encoding of an X509_NAME structure rather than
7876 copying it. This is apparently the reason for the libsafe "errors"
7877 but the code is actually correct.
7880 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7881 Bleichenbacher's DSA attack.
7882 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7883 to be set and top=0 forces the highest bit to be set; top=-1 is new
7884 and leaves the highest bit random.
7885 [Ulf Moeller, Bodo Moeller]
7887 *) In the NCONF_...-based implementations for CONF_... queries
7888 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7889 a temporary CONF structure with the data component set to NULL
7890 (which gives segmentation faults in lh_retrieve).
7891 Instead, use NULL for the CONF pointer in CONF_get_string and
7892 CONF_get_number (which may use environment variables) and directly
7893 return NULL from CONF_get_section.
7896 *) Fix potential buffer overrun for EBCDIC.
7899 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7900 keyUsage if basicConstraints absent for a CA.
7903 *) Make SMIME_write_PKCS7() write mail header values with a format that
7904 is more generally accepted (no spaces before the semicolon), since
7905 some programs can't parse those values properly otherwise. Also make
7906 sure BIO's that break lines after each write do not create invalid
7910 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7911 macros previously used would not encode an empty SEQUENCE OF
7912 and break the signature.
7914 [This change does not apply to 0.9.7.]
7916 *) Zero the premaster secret after deriving the master secret in
7920 *) Add some EVP_add_digest_alias registrations (as found in
7921 OpenSSL_add_all_digests()) to SSL_library_init()
7922 aka OpenSSL_add_ssl_algorithms(). This provides improved
7923 compatibility with peers using X.509 certificates
7924 with unconventional AlgorithmIdentifier OIDs.
7927 *) Fix for Irix with NO_ASM.
7928 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7930 *) ./config script fixes.
7931 [Ulf Moeller, Richard Levitte]
7933 *) Fix 'openssl passwd -1'.
7936 *) Change PKCS12_key_gen_asc() so it can cope with non null
7937 terminated strings whose length is passed in the passlen
7938 parameter, for example from PEM callbacks. This was done
7939 by adding an extra length parameter to asc2uni().
7940 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7942 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7943 call failed, free the DSA structure.
7946 *) Fix to uni2asc() to cope with zero length Unicode strings.
7947 These are present in some PKCS#12 files.
7950 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7951 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7952 when writing a 32767 byte record.
7953 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7955 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7956 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7958 (RSA objects have a reference count access to which is protected
7959 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7960 so they are meant to be shared between threads.)
7961 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7962 "Reddie, Steven" <Steven.Reddie@ca.com>]
7964 *) Fix a deadlock in CRYPTO_mem_leaks().
7967 *) Use better test patterns in bntest.
7970 *) rand_win.c fix for Borland C.
7973 *) BN_rshift bugfix for n == 0.
7976 *) Add a 'bctest' script that checks for some known 'bc' bugs
7977 so that 'make test' does not abort just because 'bc' is broken.
7980 *) Store verify_result within SSL_SESSION also for client side to
7981 avoid potential security hole. (Re-used sessions on the client side
7982 always resulted in verify_result==X509_V_OK, not using the original
7983 result of the server certificate verification.)
7986 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7987 SSL3_RT_APPLICATION_DATA, return 0.
7988 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7992 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7993 releases, have been re-implemented by renaming the previous
7994 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7995 and ssl3_read_internal, respectively, and adding 'peek' parameters
7996 to them. The new ssl[23]_{read,peek} functions are calls to
7997 ssl[23]_read_internal with the 'peek' flag set appropriately.
7998 A 'peek' parameter has also been added to ssl3_read_bytes, which
7999 does the actual work for ssl3_read_internal.
8002 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8003 the method-specific "init()" handler. Also clean up ex_data after
8004 calling the method-specific "finish()" handler. Previously, this was
8005 happening the other way round.
8008 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8009 The previous value, 12, was not always sufficient for BN_mod_exp().
8012 *) Make sure that shared libraries get the internal name engine with
8013 the full version number and not just 0. This should mark the
8014 shared libraries as not backward compatible. Of course, this should
8015 be changed again when we can guarantee backward binary compatibility.
8018 *) Fix typo in get_cert_by_subject() in by_dir.c
8019 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8021 *) Rework the system to generate shared libraries:
8023 - Make note of the expected extension for the shared libraries and
8024 if there is a need for symbolic links from for example libcrypto.so.0
8025 to libcrypto.so.0.9.7. There is extended info in Configure for
8028 - Make as few rebuilds of the shared libraries as possible.
8030 - Still avoid linking the OpenSSL programs with the shared libraries.
8032 - When installing, install the shared libraries separately from the
8036 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8038 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8039 and not in SSL_clear because the latter is also used by the
8040 accept/connect functions; previously, the settings made by
8041 SSL_set_read_ahead would be lost during the handshake.
8042 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8044 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8045 Previously, it would create entries for disabled algorithms no
8049 *) Added several new manual pages for SSL_* function.
8052 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8054 *) In ssl23_get_client_hello, generate an error message when faced
8055 with an initial SSL 3.0/TLS record that is too small to contain the
8056 first two bytes of the ClientHello message, i.e. client_version.
8057 (Note that this is a pathologic case that probably has never happened
8058 in real life.) The previous approach was to use the version number
8059 from the record header as a substitute; but our protocol choice
8060 should not depend on that one because it is not authenticated
8061 by the Finished messages.
8064 *) More robust randomness gathering functions for Windows.
8065 [Jeffrey Altman <jaltman@columbia.edu>]
8067 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8068 not set then we don't setup the error code for issuer check errors
8069 to avoid possibly overwriting other errors which the callback does
8070 handle. If an application does set the flag then we assume it knows
8071 what it is doing and can handle the new informational codes
8075 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8076 a general "ANY" type, as such it should be able to decode anything
8077 including tagged types. However it didn't check the class so it would
8078 wrongly interpret tagged types in the same way as their universal
8079 counterpart and unknown types were just rejected. Changed so that the
8080 tagged and unknown types are handled in the same way as a SEQUENCE:
8081 that is the encoding is stored intact. There is also a new type
8082 "V_ASN1_OTHER" which is used when the class is not universal, in this
8083 case we have no idea what the actual type is so we just lump them all
8087 *) On VMS, stdout may very well lead to a file that is written to
8088 in a record-oriented fashion. That means that every write() will
8089 write a separate record, which will be read separately by the
8090 programs trying to read from it. This can be very confusing.
8092 The solution is to put a BIO filter in the way that will buffer
8093 text until a linefeed is reached, and then write everything a
8094 line at a time, so every record written will be an actual line,
8095 not chunks of lines and not (usually doesn't happen, but I've
8096 seen it once) several lines in one record. BIO_f_linebuffer() is
8099 Currently, it's a VMS-only method, because that's where it has
8100 been tested well enough.
8103 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8104 it can return incorrect results.
8105 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8106 but it was in 0.9.6-beta[12].)
8109 *) Disable the check for content being present when verifying detached
8110 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8111 include zero length content when signing messages.
8114 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8115 BIO_ctrl (for BIO pairs).
8118 *) Add DSO method for VMS.
8121 *) Bug fix: Montgomery multiplication could produce results with the
8125 *) Add RPM specification openssl.spec and modify it to build three
8126 packages. The default package contains applications, application
8127 documentation and run-time libraries. The devel package contains
8128 include files, static libraries and function documentation. The
8129 doc package contains the contents of the doc directory. The original
8130 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8133 *) Add a large number of documentation files for many SSL routines.
8134 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8136 *) Add a configuration entry for Sony News 4.
8137 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8139 *) Don't set the two most significant bits to one when generating a
8140 random number < q in the DSA library.
8143 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8144 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8145 the underlying transport is blocking) if a handshake took place.
8146 (The default behaviour is needed by applications such as s_client
8147 and s_server that use select() to determine when to use SSL_read;
8148 but for applications that know in advance when to expect data, it
8149 just makes things more complicated.)
8152 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8156 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8157 work better on such systems.
8158 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8160 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8161 Update PKCS12_parse() so it copies the friendlyName and the
8162 keyid to the certificates aux info.
8165 *) Fix bug in PKCS7_verify() which caused an infinite loop
8166 if there was more than one signature.
8167 [Sven Uszpelkat <su@celocom.de>]
8169 *) Major change in util/mkdef.pl to include extra information
8170 about each symbol, as well as presenting variables as well
8171 as functions. This change means that there's n more need
8172 to rebuild the .num files when some algorithms are excluded.
8175 *) Allow the verify time to be set by an application,
8176 rather than always using the current time.
8179 *) Phase 2 verify code reorganisation. The certificate
8180 verify code now looks up an issuer certificate by a
8181 number of criteria: subject name, authority key id
8182 and key usage. It also verifies self signed certificates
8183 by the same criteria. The main comparison function is
8184 X509_check_issued() which performs these checks.
8186 Lot of changes were necessary in order to support this
8187 without completely rewriting the lookup code.
8189 Authority and subject key identifier are now cached.
8191 The LHASH 'certs' is X509_STORE has now been replaced
8192 by a STACK_OF(X509_OBJECT). This is mainly because an
8193 LHASH can't store or retrieve multiple objects with
8194 the same hash value.
8196 As a result various functions (which were all internal
8197 use only) have changed to handle the new X509_STORE
8198 structure. This will break anything that messed round
8199 with X509_STORE internally.
8201 The functions X509_STORE_add_cert() now checks for an
8202 exact match, rather than just subject name.
8204 The X509_STORE API doesn't directly support the retrieval
8205 of multiple certificates matching a given criteria, however
8206 this can be worked round by performing a lookup first
8207 (which will fill the cache with candidate certificates)
8208 and then examining the cache for matches. This is probably
8209 the best we can do without throwing out X509_LOOKUP
8210 entirely (maybe later...).
8212 The X509_VERIFY_CTX structure has been enhanced considerably.
8214 All certificate lookup operations now go via a get_issuer()
8215 callback. Although this currently uses an X509_STORE it
8216 can be replaced by custom lookups. This is a simple way
8217 to bypass the X509_STORE hackery necessary to make this
8218 work and makes it possible to use more efficient techniques
8219 in future. A very simple version which uses a simple
8220 STACK for its trusted certificate store is also provided
8221 using X509_STORE_CTX_trusted_stack().
8223 The verify_cb() and verify() callbacks now have equivalents
8224 in the X509_STORE_CTX structure.
8226 X509_STORE_CTX also has a 'flags' field which can be used
8227 to customise the verify behaviour.
8230 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8231 excludes S/MIME capabilities.
8234 *) When a certificate request is read in keep a copy of the
8235 original encoding of the signed data and use it when outputing
8236 again. Signatures then use the original encoding rather than
8237 a decoded, encoded version which may cause problems if the
8238 request is improperly encoded.
8241 *) For consistency with other BIO_puts implementations, call
8242 buffer_write(b, ...) directly in buffer_puts instead of calling
8245 In BIO_puts, increment b->num_write as in BIO_write.
8246 [Peter.Sylvester@EdelWeb.fr]
8248 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8249 BN_zero, we may not return a BIGNUM with an array consisting of
8253 *) Avoid calling abort() from within the library when problems are
8254 detected, except if preprocessor symbols have been defined
8255 (such as REF_CHECK, BN_DEBUG etc.).
8258 *) New openssl application 'rsautl'. This utility can be
8259 used for low level RSA operations. DER public key
8260 BIO/fp routines also added.
8263 *) New Configure entry and patches for compiling on QNX 4.
8264 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8266 *) A demo state-machine implementation was sponsored by
8267 Nuron (http://www.nuron.com/) and is now available in
8268 demos/state_machine.
8271 *) New options added to the 'dgst' utility for signature
8272 generation and verification.
8275 *) Unrecognized PKCS#7 content types are now handled via a
8276 catch all ASN1_TYPE structure. This allows unsupported
8277 types to be stored as a "blob" and an application can
8278 encode and decode it manually.
8281 *) Fix various signed/unsigned issues to make a_strex.c
8283 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8285 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8286 length if passed a buffer. ASN1_INTEGER_to_BN failed
8287 if passed a NULL BN and its argument was negative.
8288 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8290 *) Modification to PKCS#7 encoding routines to output definite
8291 length encoding. Since currently the whole structures are in
8292 memory there's not real point in using indefinite length
8293 constructed encoding. However if OpenSSL is compiled with
8294 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8297 *) Added BIO_vprintf() and BIO_vsnprintf().
8300 *) Added more prefixes to parse for in the the strings written
8301 through a logging bio, to cover all the levels that are available
8302 through syslog. The prefixes are now:
8304 PANIC, EMERG, EMR => LOG_EMERG
8305 ALERT, ALR => LOG_ALERT
8306 CRIT, CRI => LOG_CRIT
8307 ERROR, ERR => LOG_ERR
8308 WARNING, WARN, WAR => LOG_WARNING
8309 NOTICE, NOTE, NOT => LOG_NOTICE
8310 INFO, INF => LOG_INFO
8311 DEBUG, DBG => LOG_DEBUG
8313 and as before, if none of those prefixes are present at the
8314 beginning of the string, LOG_ERR is chosen.
8316 On Win32, the LOG_* levels are mapped according to this:
8318 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8319 LOG_WARNING => EVENTLOG_WARNING_TYPE
8320 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8324 *) Made it possible to reconfigure with just the configuration
8325 argument "reconf" or "reconfigure". The command line arguments
8326 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8327 and are retrieved from there when reconfiguring.
8331 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8333 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8336 *) The obj_dat.pl script was messing up the sorting of object
8337 names. The reason was that it compared the quoted version
8338 of strings as a result "OCSP" > "OCSP Signing" because
8339 " > SPACE. Changed script to store unquoted versions of
8340 names and add quotes on output. It was also omitting some
8341 names from the lookup table if they were given a default
8342 value (that is if SN is missing it is given the same
8343 value as LN and vice versa), these are now added on the
8344 grounds that if an object has a name we should be able to
8345 look it up. Finally added warning output when duplicate
8346 short or long names are found.
8349 *) Changes needed for Tandem NSK.
8350 [Scott Uroff <scott@xypro.com>]
8352 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8353 RSA_padding_check_SSLv23(), special padding was never detected
8354 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8355 version rollback attacks was not effective.
8357 In s23_clnt.c, don't use special rollback-attack detection padding
8358 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8359 client; similarly, in s23_srvr.c, don't do the rollback check if
8360 SSL 2.0 is the only protocol enabled in the server.
8363 *) Make it possible to get hexdumps of unprintable data with 'openssl
8364 asn1parse'. By implication, the functions ASN1_parse_dump() and
8365 BIO_dump_indent() are added.
8368 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8369 these print out strings and name structures based on various
8370 flags including RFC2253 support and proper handling of
8371 multibyte characters. Added options to the 'x509' utility
8372 to allow the various flags to be set.
8375 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8376 Also change the functions X509_cmp_current_time() and
8377 X509_gmtime_adj() work with an ASN1_TIME structure,
8378 this will enable certificates using GeneralizedTime in validity
8379 dates to be checked.
8382 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8383 negative public key encodings) on by default,
8384 NO_NEG_PUBKEY_BUG can be set to disable it.
8387 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8388 content octets. An i2c_ASN1_OBJECT is unnecessary because
8389 the encoding can be trivially obtained from the structure.
8392 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8393 not read locks (CRYPTO_r_[un]lock).
8396 *) A first attempt at creating official support for shared
8397 libraries through configuration. I've kept it so the
8398 default is static libraries only, and the OpenSSL programs
8399 are always statically linked for now, but there are
8400 preparations for dynamic linking in place.
8401 This has been tested on Linux and Tru64.
8404 *) Randomness polling function for Win9x, as described in:
8405 Peter Gutmann, Software Generation of Practically Strong
8409 *) Fix so PRNG is seeded in req if using an already existing
8413 *) New options to smime application. -inform and -outform
8414 allow alternative formats for the S/MIME message including
8415 PEM and DER. The -content option allows the content to be
8416 specified separately. This should allow things like Netscape
8417 form signing output easier to verify.
8420 *) Fix the ASN1 encoding of tags using the 'long form'.
8423 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8424 STRING types. These convert content octets to and from the
8425 underlying type. The actual tag and length octets are
8426 already assumed to have been read in and checked. These
8427 are needed because all other string types have virtually
8428 identical handling apart from the tag. By having versions
8429 of the ASN1 functions that just operate on content octets
8430 IMPLICIT tagging can be handled properly. It also allows
8431 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8432 and ASN1_INTEGER are identical apart from the tag.
8435 *) Change the handling of OID objects as follows:
8437 - New object identifiers are inserted in objects.txt, following
8438 the syntax given in objects.README.
8439 - objects.pl is used to process obj_mac.num and create a new
8441 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8444 This is currently kind of a hack, and the perl code in objects.pl
8445 isn't very elegant, but it works as I intended. The simplest way
8446 to check that it worked correctly is to look in obj_dat.h and
8447 check the array nid_objs and make sure the objects haven't moved
8448 around (this is important!). Additions are OK, as well as
8449 consistent name changes.
8452 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8455 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8456 The given file adds to whatever has already been seeded into the
8457 random pool through the RANDFILE configuration file option or
8458 environment variable, or the default random state file.
8461 *) mkstack.pl now sorts each macro group into lexical order.
8462 Previously the output order depended on the order the files
8463 appeared in the directory, resulting in needless rewriting
8467 *) Patches to make OpenSSL compile under Win32 again. Mostly
8468 work arounds for the VC++ problem that it treats func() as
8469 func(void). Also stripped out the parts of mkdef.pl that
8470 added extra typesafe functions: these no longer exist.
8473 *) Reorganisation of the stack code. The macros are now all
8474 collected in safestack.h . Each macro is defined in terms of
8475 a "stack macro" of the form SKM_<name>(type, a, b). The
8476 DEBUG_SAFESTACK is now handled in terms of function casts,
8477 this has the advantage of retaining type safety without the
8478 use of additional functions. If DEBUG_SAFESTACK is not defined
8479 then the non typesafe macros are used instead. Also modified the
8480 mkstack.pl script to handle the new form. Needs testing to see
8481 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8482 the default if no major problems. Similar behaviour for ASN1_SET_OF
8483 and PKCS12_STACK_OF.
8486 *) When some versions of IIS use the 'NET' form of private key the
8487 key derivation algorithm is different. Normally MD5(password) is
8488 used as a 128 bit RC4 key. In the modified case
8489 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8490 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8491 as the old Netscape_RSA functions except they have an additional
8492 'sgckey' parameter which uses the modified algorithm. Also added
8493 an -sgckey command line option to the rsa utility. Thanks to
8494 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8495 algorithm to openssl-dev.
8498 *) The evp_local.h macros were using 'c.##kname' which resulted in
8499 invalid expansion on some systems (SCO 5.0.5 for example).
8500 Corrected to 'c.kname'.
8501 [Phillip Porch <root@theporch.com>]
8503 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8504 a STACK of email addresses from a certificate or request, these look
8505 in the subject name and the subject alternative name extensions and
8506 omit any duplicate addresses.
8509 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8510 This makes DSA verification about 2 % faster.
8513 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8514 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8515 plus overhead for 1024 bit moduli).
8516 This makes exponentiations about 0.5 % faster for 1024 bit
8517 exponents (as measured by "openssl speed rsa2048").
8520 *) Rename memory handling macros to avoid conflicts with other
8522 Malloc => OPENSSL_malloc
8523 Malloc_locked => OPENSSL_malloc_locked
8524 Realloc => OPENSSL_realloc
8525 Free => OPENSSL_free
8528 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8529 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8532 *) CygWin32 support.
8533 [John Jarvie <jjarvie@newsguy.com>]
8535 *) The type-safe stack code has been rejigged. It is now only compiled
8536 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8537 by default all type-specific stack functions are "#define"d back to
8538 standard stack functions. This results in more streamlined output
8539 but retains the type-safety checking possibilities of the original
8543 *) The STACK code has been cleaned up, and certain type declarations
8544 that didn't make a lot of sense have been brought in line. This has
8545 also involved a cleanup of sorts in safestack.h to more correctly
8546 map type-safe stack functions onto their plain stack counterparts.
8547 This work has also resulted in a variety of "const"ifications of
8548 lots of the code, especially "_cmp" operations which should normally
8549 be prototyped with "const" parameters anyway.
8552 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8553 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8554 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8555 where all of 'md' is used each time the PRNG is used, but 'state'
8556 is used only indexed by a cyclic counter. As entropy may not be
8557 well distributed from the beginning, 'md' is important as a
8558 chaining variable. However, the output function chains only half
8559 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8560 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8561 in all of 'state' being rewritten, with the new values depending
8562 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8565 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8566 the handshake is continued after ssl_verify_cert_chain();
8567 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8568 can lead to 'unexplainable' connection aborts later.
8569 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8571 *) Major EVP API cipher revision.
8572 Add hooks for extra EVP features. This allows various cipher
8573 parameters to be set in the EVP interface. Support added for variable
8574 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8575 setting of RC2 and RC5 parameters.
8577 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8580 Remove lots of duplicated code from the EVP library. For example *every*
8581 cipher init() function handles the 'iv' in the same way according to the
8582 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8583 for CFB and OFB modes they zero ctx->num.
8585 New functionality allows removal of S/MIME code RC2 hack.
8587 Most of the routines have the same form and so can be declared in terms
8590 By shifting this to the top level EVP_CipherInit() it can be removed from
8591 all individual ciphers. If the cipher wants to handle IVs or keys
8592 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8595 Change lots of functions like EVP_EncryptUpdate() to now return a
8596 value: although software versions of the algorithms cannot fail
8597 any installed hardware versions can.
8600 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8601 this option is set, tolerate broken clients that send the negotiated
8602 protocol version number instead of the requested protocol version
8606 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8607 i.e. non-zero for export ciphersuites, zero otherwise.
8608 Previous versions had this flag inverted, inconsistent with
8609 rsa_tmp_cb (..._TMP_RSA_CB).
8610 [Bodo Moeller; problem reported by Amit Chopra]
8612 *) Add missing DSA library text string. Work around for some IIS
8613 key files with invalid SEQUENCE encoding.
8616 *) Add a document (doc/standards.txt) that list all kinds of standards
8617 and so on that are implemented in OpenSSL.
8620 *) Enhance c_rehash script. Old version would mishandle certificates
8621 with the same subject name hash and wouldn't handle CRLs at all.
8622 Added -fingerprint option to crl utility, to support new c_rehash
8626 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8629 *) Fix for SSL server purpose checking. Server checking was
8630 rejecting certificates which had extended key usage present
8631 but no ssl client purpose.
8632 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8634 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8635 is a little unclear about how a blank password is handled.
8636 Since the password in encoded as a BMPString with terminating
8637 double NULL a zero length password would end up as just the
8638 double NULL. However no password at all is different and is
8639 handled differently in the PKCS#12 key generation code. NS
8640 treats a blank password as zero length. MSIE treats it as no
8641 password on export: but it will try both on import. We now do
8642 the same: PKCS12_parse() tries zero length and no password if
8643 the password is set to "" or NULL (NULL is now a valid password:
8644 it wasn't before) as does the pkcs12 application.
8647 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8648 perror when PEM_read_bio_X509_REQ fails, the error message must
8649 be obtained from the error queue.
8652 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8653 it in ERR_remove_state if appropriate, and change ERR_get_state
8654 accordingly to avoid race conditions (this is necessary because
8655 thread_hash is no longer constant once set).
8658 *) Bugfix for linux-elf makefile.one.
8661 *) RSA_get_default_method() will now cause a default
8662 RSA_METHOD to be chosen if one doesn't exist already.
8663 Previously this was only set during a call to RSA_new()
8664 or RSA_new_method(NULL) meaning it was possible for
8665 RSA_get_default_method() to return NULL.
8668 *) Added native name translation to the existing DSO code
8669 that will convert (if the flag to do so is set) filenames
8670 that are sufficiently small and have no path information
8671 into a canonical native form. Eg. "blah" converted to
8672 "libblah.so" or "blah.dll" etc.
8675 *) New function ERR_error_string_n(e, buf, len) which is like
8676 ERR_error_string(e, buf), but writes at most 'len' bytes
8677 including the 0 terminator. For ERR_error_string_n, 'buf'
8679 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8681 *) CONF library reworked to become more general. A new CONF
8682 configuration file reader "class" is implemented as well as a
8683 new functions (NCONF_*, for "New CONF") to handle it. The now
8684 old CONF_* functions are still there, but are reimplemented to
8685 work in terms of the new functions. Also, a set of functions
8686 to handle the internal storage of the configuration data is
8687 provided to make it easier to write new configuration file
8688 reader "classes" (I can definitely see something reading a
8689 configuration file in XML format, for example), called _CONF_*,
8690 or "the configuration storage API"...
8692 The new configuration file reading functions are:
8694 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8695 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8697 NCONF_default, NCONF_WIN32
8699 NCONF_dump_fp, NCONF_dump_bio
8701 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8702 NCONF_new creates a new CONF object. This works in the same way
8703 as other interfaces in OpenSSL, like the BIO interface.
8704 NCONF_dump_* dump the internal storage of the configuration file,
8705 which is useful for debugging. All other functions take the same
8706 arguments as the old CONF_* functions wth the exception of the
8707 first that must be a `CONF *' instead of a `LHASH *'.
8709 To make it easer to use the new classes with the old CONF_* functions,
8710 the function CONF_set_default_method is provided.
8713 *) Add '-tls1' option to 'openssl ciphers', which was already
8714 mentioned in the documentation but had not been implemented.
8715 (This option is not yet really useful because even the additional
8716 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8719 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8720 OpenSSL-based applications) load shared libraries and bind to
8721 them in a portable way.
8722 [Geoff Thorpe, with contributions from Richard Levitte]
8724 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8726 *) Make sure _lrotl and _lrotr are only used with MSVC.
8728 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8729 (the default implementation of RAND_status).
8731 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8732 to '-clrext' (= clear extensions), as intended and documented.
8733 [Bodo Moeller; inconsistency pointed out by Michael Attili
8736 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8737 was larger than the MD block size.
8738 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8740 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8741 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8742 using the passed key: if the passed key was a private key the result
8743 of X509_print(), for example, would be to print out all the private key
8747 *) des_quad_cksum() byte order bug fix.
8748 [Ulf Möller, using the problem description in krb4-0.9.7, where
8749 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8751 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8753 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8755 *) For easily testing in shell scripts whether some command
8756 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8757 returns with exit code 0 iff no command of the given name is available.
8758 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8759 the output goes to stdout and nothing is printed to stderr.
8760 Additional arguments are always ignored.
8762 Since for each cipher there is a command of the same name,
8763 the 'no-cipher' compilation switches can be tested this way.
8765 ('openssl no-XXX' is not able to detect pseudo-commands such
8766 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8769 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8772 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8773 is set; it will be thrown away anyway because each handshake creates
8775 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8776 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8777 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8778 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8781 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8782 'Q' and 'R' lose their special meanings (quit/renegotiate).
8783 This is part of what -quiet does; unlike -quiet, -ign_eof
8784 does not suppress any output.
8787 *) Add compatibility options to the purpose and trust code. The
8788 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8789 accepts a certificate or CA, this was the previous behaviour,
8790 with all the associated security issues.
8792 X509_TRUST_COMPAT is the old trust behaviour: only and
8793 automatically trust self signed roots in certificate store. A
8794 new trust setting X509_TRUST_DEFAULT is used to specify that
8795 a purpose has no associated trust setting and it should instead
8796 use the value in the default purpose.
8799 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8800 and fix a memory leak.
8803 *) In util/mkerr.pl (which implements 'make errors'), preserve
8804 reason strings from the previous version of the .c file, as
8805 the default to have only downcase letters (and digits) in
8806 automatically generated reasons codes is not always appropriate.
8809 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8810 using strerror. Previously, ERR_reason_error_string() returned
8811 library names as reason strings for SYSerr; but SYSerr is a special
8812 case where small numbers are errno values, not library numbers.
8815 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8816 converts DSA parameters into DH parameters. (When creating parameters,
8817 DSA_generate_parameters is used.)
8820 *) Include 'length' (recommended exponent length) in C code generated
8821 by 'openssl dhparam -C'.
8824 *) The second argument to set_label in perlasm was already being used
8825 so couldn't be used as a "file scope" flag. Moved to third argument
8829 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8830 instead of RAND_bytes for encryption IVs and salts.
8833 *) Include RAND_status() into RAND_METHOD instead of implementing
8834 it only for md_rand.c Otherwise replacing the PRNG by calling
8835 RAND_set_rand_method would be impossible.
8838 *) Don't let DSA_generate_key() enter an infinite loop if the random
8839 number generation fails.
8842 *) New 'rand' application for creating pseudo-random output.
8845 *) Added configuration support for Linux/IA64
8846 [Rolf Haberrecker <rolf@suse.de>]
8848 *) Assembler module support for Mingw32.
8851 *) Shared library support for HPUX (in shlib/).
8852 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8854 *) Shared library support for Solaris gcc.
8855 [Lutz Behnke <behnke@trustcenter.de>]
8857 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8859 *) PKCS7_encrypt() was adding text MIME headers twice because they
8860 were added manually and by SMIME_crlf_copy().
8863 *) In bntest.c don't call BN_rand with zero bits argument.
8864 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8866 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8867 case was implemented. This caused BN_div_recp() to fail occasionally.
8870 *) Add an optional second argument to the set_label() in the perl
8871 assembly language builder. If this argument exists and is set
8872 to 1 it signals that the assembler should use a symbol whose
8873 scope is the entire file, not just the current function. This
8874 is needed with MASM which uses the format label:: for this scope.
8875 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8877 *) Change the ASN1 types so they are typedefs by default. Before
8878 almost all types were #define'd to ASN1_STRING which was causing
8879 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8883 *) Change names of new functions to the new get1/get0 naming
8884 convention: After 'get1', the caller owns a reference count
8885 and has to call ..._free; 'get0' returns a pointer to some
8886 data structure without incrementing reference counters.
8887 (Some of the existing 'get' functions increment a reference
8888 counter, some don't.)
8889 Similarly, 'set1' and 'add1' functions increase reference
8890 counters or duplicate objects.
8893 *) Allow for the possibility of temp RSA key generation failure:
8894 the code used to assume it always worked and crashed on failure.
8897 *) Fix potential buffer overrun problem in BIO_printf().
8898 [Ulf Möller, using public domain code by Patrick Powell; problem
8899 pointed out by David Sacerdote <das33@cornell.edu>]
8901 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8902 RAND_egd() and RAND_status(). In the command line application,
8903 the EGD socket can be specified like a seed file using RANDFILE
8907 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8908 Some CAs (e.g. Verisign) distribute certificates in this form.
8911 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8912 list to exclude them. This means that no special compilation option
8913 is needed to use anonymous DH: it just needs to be included in the
8917 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8918 EVP_MD_type. The old functionality is available in a new macro called
8919 EVP_MD_md(). Change code that uses it and update docs.
8922 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8923 where the 'void *' argument is replaced by a function pointer argument.
8924 Previously 'void *' was abused to point to functions, which works on
8925 many platforms, but is not correct. As these functions are usually
8926 called by macros defined in OpenSSL header files, most source code
8927 should work without changes.
8930 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8931 sections with information on -D... compiler switches used for
8932 compiling the library so that applications can see them. To enable
8933 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8934 must be defined. E.g.,
8935 #define OPENSSL_ALGORITHM_DEFINES
8936 #include <openssl/opensslconf.h>
8937 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8938 [Richard Levitte, Ulf and Bodo Möller]
8940 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8944 *) Change the 'other' type in certificate aux info to a STACK_OF
8945 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8946 the required ASN1 format: arbitrary types determined by an OID.
8949 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8950 argument to 'req'. This is not because the function is newer or
8951 better than others it just uses the work 'NEW' in the certificate
8952 request header lines. Some software needs this.
8955 *) Reorganise password command line arguments: now passwords can be
8956 obtained from various sources. Delete the PEM_cb function and make
8957 it the default behaviour: i.e. if the callback is NULL and the
8958 usrdata argument is not NULL interpret it as a null terminated pass
8959 phrase. If usrdata and the callback are NULL then the pass phrase
8960 is prompted for as usual.
8963 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8964 the support is automatically enabled. The resulting binaries will
8965 autodetect the card and use it if present.
8966 [Ben Laurie and Compaq Inc.]
8968 *) Work around for Netscape hang bug. This sends certificate request
8969 and server done in one record. Since this is perfectly legal in the
8970 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8971 the bugs/SSLv3 entry for more info.
8974 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8977 *) Add -rand argument to smime and pkcs12 applications and read/write
8981 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8984 *) Add command line password options to the remaining applications.
8987 *) Bug fix for BN_div_recp() for numerators with an even number of
8991 *) More tests in bntest.c, and changed test_bn output.
8994 *) ./config recognizes MacOS X now.
8997 *) Bug fix for BN_div() when the first words of num and divsor are
8998 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9001 *) Add support for various broken PKCS#8 formats, and command line
9002 options to produce them.
9005 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9006 get temporary BIGNUMs from a BN_CTX.
9009 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9013 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9014 include a #define from the old name to the new. The original intent
9015 was that statically linked binaries could for example just call
9016 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9017 link with digests. This never worked because SSLeay_add_all_digests()
9018 and SSLeay_add_all_ciphers() were in the same source file so calling
9019 one would link with the other. They are now in separate source files.
9022 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9025 *) Use a less unusual form of the Miller-Rabin primality test (it used
9026 a binary algorithm for exponentiation integrated into the Miller-Rabin
9027 loop, our standard modexp algorithms are faster).
9030 *) Support for the EBCDIC character set completed.
9031 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9033 *) Source code cleanups: use const where appropriate, eliminate casts,
9034 use void * instead of char * in lhash.
9037 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9038 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9039 this the server could overwrite ephemeral keys that the client
9043 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9044 using 50 iterations of the Rabin-Miller test.
9046 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9047 iterations of the Rabin-Miller test as required by the appendix
9048 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9049 As BN_is_prime_fasttest includes trial division, DSA parameter
9050 generation becomes much faster.
9052 This implies a change for the callback functions in DSA_is_prime
9053 and DSA_generate_parameters: The callback function is called once
9054 for each positive witness in the Rabin-Miller test, not just
9055 occasionally in the inner loop; and the parameters to the
9056 callback function now provide an iteration count for the outer
9057 loop rather than for the current invocation of the inner loop.
9058 DSA_generate_parameters additionally can call the callback
9059 function with an 'iteration count' of -1, meaning that a
9060 candidate has passed the trial division test (when q is generated
9061 from an application-provided seed, trial division is skipped).
9064 *) New function BN_is_prime_fasttest that optionally does trial
9065 division before starting the Rabin-Miller test and has
9066 an additional BN_CTX * argument (whereas BN_is_prime always
9067 has to allocate at least one BN_CTX).
9068 'callback(1, -1, cb_arg)' is called when a number has passed the
9069 trial division stage.
9072 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9076 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9079 *) New function BN_pseudo_rand().
9082 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9083 bignum version of BN_from_montgomery() with the working code from
9084 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9088 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9089 made it impossible to use the same SSL_SESSION data structure in
9090 SSL2 clients in multiple threads.
9093 *) The return value of RAND_load_file() no longer counts bytes obtained
9094 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9095 to seed the PRNG (previously an explicit byte count was required).
9096 [Ulf Möller, Bodo Möller]
9098 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9099 used (char *) instead of (void *) and had casts all over the place.
9102 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9105 *) Retain source code compatibility for BN_prime_checks macro:
9106 BN_is_prime(..., BN_prime_checks, ...) now uses
9107 BN_prime_checks_for_size to determine the appropriate number of
9108 Rabin-Miller iterations.
9111 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9112 DH_CHECK_P_NOT_SAFE_PRIME.
9113 (Check if this is true? OpenPGP calls them "strong".)
9116 *) Merge the functionality of "dh" and "gendh" programs into a new program
9117 "dhparam". The old programs are retained for now but will handle DH keys
9118 (instead of parameters) in future.
9121 *) Make the ciphers, s_server and s_client programs check the return values
9122 when a new cipher list is set.
9125 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9126 ciphers. Before when the 56bit ciphers were enabled the sorting was
9129 The syntax for the cipher sorting has been extended to support sorting by
9130 cipher-strength (using the strength_bits hard coded in the tables).
9131 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9133 Fix a bug in the cipher-command parser: when supplying a cipher command
9134 string with an "undefined" symbol (neither command nor alphanumeric
9135 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9136 an error is flagged.
9138 Due to the strength-sorting extension, the code of the
9139 ssl_create_cipher_list() function was completely rearranged. I hope that
9140 the readability was also increased :-)
9141 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9143 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9144 for the first serial number and places 2 in the serial number file. This
9145 avoids problems when the root CA is created with serial number zero and
9146 the first user certificate has the same issuer name and serial number
9150 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9151 the new code. Add documentation for this stuff.
9154 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9155 X509_*() to X509at_*() on the grounds that they don't handle X509
9156 structures and behave in an analogous way to the X509v3 functions:
9157 they shouldn't be called directly but wrapper functions should be used
9160 So we also now have some wrapper functions that call the X509at functions
9161 when passed certificate requests. (TO DO: similar things can be done with
9162 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9163 things. Some of these need some d2i or i2d and print functionality
9164 because they handle more complex structures.)
9167 *) Add missing #ifndefs that caused missing symbols when building libssl
9168 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9169 NO_RSA in ssl/s2*.c.
9170 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9172 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9173 has a return value which indicates the quality of the random data
9174 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9175 error queue. New function RAND_pseudo_bytes() generates output that is
9176 guaranteed to be unique but not unpredictable. RAND_add is like
9177 RAND_seed, but takes an extra argument for an entropy estimate
9178 (RAND_seed always assumes full entropy).
9181 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9182 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9183 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9184 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9185 false-positive rate of at most 2^-80 for random input.
9188 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9191 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9192 in the 0.9.5 release), this returns the chain
9193 from an X509_CTX structure with a dup of the stack and all
9194 the X509 reference counts upped: so the stack will exist
9195 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9198 Also make SSL_SESSION_print() print out the verify return
9202 *) Add manpage for the pkcs12 command. Also change the default
9203 behaviour so MAC iteration counts are used unless the new
9204 -nomaciter option is used. This improves file security and
9205 only older versions of MSIE (4.0 for example) need it.
9208 *) Honor the no-xxx Configure options when creating .DEF files.
9211 *) Add PKCS#10 attributes to field table: challengePassword,
9212 unstructuredName and unstructuredAddress. These are taken from
9213 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9214 international characters are used.
9216 More changes to X509_ATTRIBUTE code: allow the setting of types
9217 based on strings. Remove the 'loc' parameter when adding
9218 attributes because these will be a SET OF encoding which is sorted
9222 *) Initial changes to the 'req' utility to allow request generation
9223 automation. This will allow an application to just generate a template
9224 file containing all the field values and have req construct the
9227 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9228 used all over the place including certificate requests and PKCS#7
9229 structures. They are currently handled manually where necessary with
9230 some primitive wrappers for PKCS#7. The new functions behave in a
9231 manner analogous to the X509 extension functions: they allow
9232 attributes to be looked up by NID and added.
9234 Later something similar to the X509V3 code would be desirable to
9235 automatically handle the encoding, decoding and printing of the
9236 more complex types. The string types like challengePassword can
9237 be handled by the string table functions.
9239 Also modified the multi byte string table handling. Now there is
9240 a 'global mask' which masks out certain types. The table itself
9241 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9242 is useful when for example there is only one permissible type
9243 (as in countryName) and using the mask might result in no valid
9247 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9248 SSL_get_peer_finished to allow applications to obtain the latest
9249 Finished messages sent to the peer or expected from the peer,
9250 respectively. (SSL_get_peer_finished is usually the Finished message
9251 actually received from the peer, otherwise the protocol will be aborted.)
9253 As the Finished message are message digests of the complete handshake
9254 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9255 be used for external authentication procedures when the authentication
9256 provided by SSL/TLS is not desired or is not enough.
9259 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9260 the host supports BWX extension and if Compaq C is present on the
9261 $PATH. Just exploiting of the BWX extension results in 20-30%
9262 performance kick for some algorithms, e.g. DES and RC4 to mention
9263 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9267 *) Add support for MS "fast SGC". This is arguably a violation of the
9268 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9269 weak crypto and after checking the certificate is SGC a second one
9270 with strong crypto. MS SGC stops the first handshake after receiving
9271 the server certificate message and sends a second client hello. Since
9272 a server will typically do all the time consuming operations before
9273 expecting any further messages from the client (server key exchange
9274 is the most expensive) there is little difference between the two.
9276 To get OpenSSL to support MS SGC we have to permit a second client
9277 hello message after we have sent server done. In addition we have to
9278 reset the MAC if we do get this second client hello.
9281 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9282 if a DER encoded private key is RSA or DSA traditional format. Changed
9283 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9284 format DER encoded private key. Newer code should use PKCS#8 format which
9285 has the key type encoded in the ASN1 structure. Added DER private key
9286 support to pkcs8 application.
9289 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9290 ciphersuites has been selected (as required by the SSL 3/TLS 1
9291 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9292 is set, we interpret this as a request to violate the specification
9293 (the worst that can happen is a handshake failure, and 'correct'
9294 behaviour would result in a handshake failure anyway).
9297 *) In SSL_CTX_add_session, take into account that there might be multiple
9298 SSL_SESSION structures with the same session ID (e.g. when two threads
9299 concurrently obtain them from an external cache).
9300 The internal cache can handle only one SSL_SESSION with a given ID,
9301 so if there's a conflict, we now throw out the old one to achieve
9305 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9306 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9307 some routines that use cipher OIDs: some ciphers do not have OIDs
9308 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9312 *) Simplify the trust setting structure and code. Now we just have
9313 two sequences of OIDs for trusted and rejected settings. These will
9314 typically have values the same as the extended key usage extension
9315 and any application specific purposes.
9317 The trust checking code now has a default behaviour: it will just
9318 check for an object with the same NID as the passed id. Functions can
9319 be provided to override either the default behaviour or the behaviour
9320 for a given id. SSL client, server and email already have functions
9321 in place for compatibility: they check the NID and also return "trusted"
9322 if the certificate is self signed.
9325 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9326 traditional format into an EVP_PKEY structure.
9329 *) Add a password callback function PEM_cb() which either prompts for
9330 a password if usr_data is NULL or otherwise assumes it is a null
9331 terminated password. Allow passwords to be passed on command line
9332 environment or config files in a few more utilities.
9335 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9336 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9337 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9338 Update documentation.
9341 *) Support for ASN1 "NULL" type. This could be handled before by using
9342 ASN1_TYPE but there wasn't any function that would try to read a NULL
9343 and produce an error if it couldn't. For compatibility we also have
9344 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9345 don't allocate anything because they don't need to.
9348 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9350 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9352 *) Rebuild of the memory allocation routines used by OpenSSL code and
9353 possibly others as well. The purpose is to make an interface that
9354 provide hooks so anyone can build a separate set of allocation and
9355 deallocation routines to be used by OpenSSL, for example memory
9356 pool implementations, or something else, which was previously hard
9357 since Malloc(), Realloc() and Free() were defined as macros having
9358 the values malloc, realloc and free, respectively (except for Win32
9359 compilations). The same is provided for memory debugging code.
9360 OpenSSL already comes with functionality to find memory leaks, but
9361 this gives people a chance to debug other memory problems.
9363 With these changes, a new set of functions and macros have appeared:
9365 CRYPTO_set_mem_debug_functions() [F]
9366 CRYPTO_get_mem_debug_functions() [F]
9367 CRYPTO_dbg_set_options() [F]
9368 CRYPTO_dbg_get_options() [F]
9369 CRYPTO_malloc_debug_init() [M]
9371 The memory debug functions are NULL by default, unless the library
9372 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9373 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9374 gives the standard debugging functions that come with OpenSSL) or
9375 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9376 provided by the library user) must be used. When the standard
9377 debugging functions are used, CRYPTO_dbg_set_options can be used to
9378 request additional information:
9379 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9380 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9382 Also, things like CRYPTO_set_mem_functions will always give the
9383 expected result (the new set of functions is used for allocation
9384 and deallocation) at all times, regardless of platform and compiler
9387 To finish it up, some functions that were never use in any other
9388 way than through macros have a new API and new semantic:
9391 CRYPTO_dbg_realloc()
9394 All macros of value have retained their old syntax.
9395 [Richard Levitte and Bodo Moeller]
9397 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9398 ordering of SMIMECapabilities wasn't in "strength order" and there
9399 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9403 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9404 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9405 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9407 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9408 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9409 functionality to handle multipart/signed properly) and a utility
9410 called 'smime' to call all this stuff. This is based on code I
9411 originally wrote for Celo who have kindly allowed it to be
9412 included in OpenSSL.
9415 *) Add variants des_set_key_checked and des_set_key_unchecked of
9416 des_set_key (aka des_key_sched). Global variable des_check_key
9417 decides which of these is called by des_set_key; this way
9418 des_check_key behaves as it always did, but applications and
9419 the library itself, which was buggy for des_check_key == 1,
9420 have a cleaner way to pick the version they need.
9423 *) New function PKCS12_newpass() which changes the password of a
9427 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9428 dynamic mix. In both cases the ids can be used as an index into the
9429 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9430 functions so they accept a list of the field values and the
9431 application doesn't need to directly manipulate the X509_TRUST
9435 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9439 *) Modify the way the V3 extension code looks up extensions. This now
9440 works in a similar way to the object code: we have some "standard"
9441 extensions in a static table which is searched with OBJ_bsearch()
9442 and the application can add dynamic ones if needed. The file
9443 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9444 updated whenever a new extension is added to the core code and kept
9445 in ext_nid order. There is a simple program 'tabtest.c' which checks
9446 this. New extensions are not added too often so this file can readily
9447 be maintained manually.
9449 There are two big advantages in doing things this way. The extensions
9450 can be looked up immediately and no longer need to be "added" using
9451 X509V3_add_standard_extensions(): this function now does nothing.
9452 [Side note: I get *lots* of email saying the extension code doesn't
9453 work because people forget to call this function]
9454 Also no dynamic allocation is done unless new extensions are added:
9455 so if we don't add custom extensions there is no need to call
9456 X509V3_EXT_cleanup().
9459 *) Modify enc utility's salting as follows: make salting the default. Add a
9460 magic header, so unsalted files fail gracefully instead of just decrypting
9461 to garbage. This is because not salting is a big security hole, so people
9462 should be discouraged from doing it.
9465 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9466 digest to be passed on the command line but it only used this
9467 parameter when signing a certificate. Modified so all relevant
9468 operations are affected by the digest parameter including the
9469 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9470 DSA key was used because it didn't fix the digest.
9473 *) Initial certificate chain verify code. Currently tests the untrusted
9474 certificates for consistency with the verify purpose (which is set
9475 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9477 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9478 this is because it will reject chains with invalid extensions whereas
9479 every previous version of OpenSSL and SSLeay made no checks at all.
9481 Trust code: checks the root CA for the relevant trust settings. Trust
9482 settings have an initial value consistent with the verify purpose: e.g.
9483 if the verify purpose is for SSL client use it expects the CA to be
9484 trusted for SSL client use. However the default value can be changed to
9485 permit custom trust settings: one example of this would be to only trust
9486 certificates from a specific "secure" set of CAs.
9488 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9489 which should be used for version portability: especially since the
9490 verify structure is likely to change more often now.
9492 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9493 to set them. If not set then assume SSL clients will verify SSL servers
9496 Two new options to the verify program: -untrusted allows a set of
9497 untrusted certificates to be passed in and -purpose which sets the
9498 intended purpose of the certificate. If a purpose is set then the
9499 new chain verify code is used to check extension consistency.
9502 *) Support for the authority information access extension.
9505 *) Modify RSA and DSA PEM read routines to transparently handle
9506 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9507 public keys in a format compatible with certificate
9508 SubjectPublicKeyInfo structures. Unfortunately there were already
9509 functions called *_PublicKey_* which used various odd formats so
9510 these are retained for compatibility: however the DSA variants were
9511 never in a public release so they have been deleted. Changed dsa/rsa
9512 utilities to handle the new format: note no releases ever handled public
9513 keys so we should be OK.
9515 The primary motivation for this change is to avoid the same fiasco
9516 that dogs private keys: there are several incompatible private key
9517 formats some of which are standard and some OpenSSL specific and
9518 require various evil hacks to allow partial transparent handling and
9519 even then it doesn't work with DER formats. Given the option anything
9520 other than PKCS#8 should be dumped: but the other formats have to
9521 stay in the name of compatibility.
9523 With public keys and the benefit of hindsight one standard format
9524 is used which works with EVP_PKEY, RSA or DSA structures: though
9525 it clearly returns an error if you try to read the wrong kind of key.
9527 Added a -pubkey option to the 'x509' utility to output the public key.
9528 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9529 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9530 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9531 that do the same as the EVP_PKEY_assign_*() except they up the
9532 reference count of the added key (they don't "swallow" the
9536 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9537 CRLs would fail if the file contained no certificates or no CRLs:
9538 added a new function to read in both types and return the number
9539 read: this means that if none are read it will be an error. The
9540 DER versions of the certificate and CRL reader would always fail
9541 because it isn't possible to mix certificates and CRLs in DER format
9542 without choking one or the other routine. Changed this to just read
9543 a certificate: this is the best we can do. Also modified the code
9544 in apps/verify.c to take notice of return codes: it was previously
9545 attempting to read in certificates from NULL pointers and ignoring
9546 any errors: this is one reason why the cert and CRL reader seemed
9547 to work. It doesn't check return codes from the default certificate
9548 routines: these may well fail if the certificates aren't installed.
9551 *) Code to support otherName option in GeneralName.
9554 *) First update to verify code. Change the verify utility
9555 so it warns if it is passed a self signed certificate:
9556 for consistency with the normal behaviour. X509_verify
9557 has been modified to it will now verify a self signed
9558 certificate if *exactly* the same certificate appears
9559 in the store: it was previously impossible to trust a
9560 single self signed certificate. This means that:
9561 openssl verify ss.pem
9562 now gives a warning about a self signed certificate but
9563 openssl verify -CAfile ss.pem ss.pem
9567 *) For servers, store verify_result in SSL_SESSION data structure
9568 (and add it to external session representation).
9569 This is needed when client certificate verifications fails,
9570 but an application-provided verification callback (set by
9571 SSL_CTX_set_cert_verify_callback) allows accepting the session
9572 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9573 but returns 1): When the session is reused, we have to set
9574 ssl->verify_result to the appropriate error code to avoid
9576 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9578 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9579 case in PKCS7_dataInit() where the signed PKCS7 structure
9580 didn't contain any existing data because it was being created.
9581 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9583 *) Add a salt to the key derivation routines in enc.c. This
9584 forms the first 8 bytes of the encrypted file. Also add a
9585 -S option to allow a salt to be input on the command line.
9588 *) New function X509_cmp(). Oddly enough there wasn't a function
9589 to compare two certificates. We do this by working out the SHA1
9590 hash and comparing that. X509_cmp() will be needed by the trust
9594 *) SSL_get1_session() is like SSL_get_session(), but increments
9595 the reference count in the SSL_SESSION returned.
9596 [Geoff Thorpe <geoff@eu.c2.net>]
9598 *) Fix for 'req': it was adding a null to request attributes.
9599 Also change the X509_LOOKUP and X509_INFO code to handle
9600 certificate auxiliary information.
9603 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9607 *) Add the possibility to add extra information to the memory leak
9608 detecting output, to form tracebacks, showing from where each
9609 allocation was originated: CRYPTO_push_info("constant string") adds
9610 the string plus current file name and line number to a per-thread
9611 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9612 is like calling CYRPTO_pop_info() until the stack is empty.
9613 Also updated memory leak detection code to be multi-thread-safe.
9616 *) Add options -text and -noout to pkcs7 utility and delete the
9617 encryption options which never did anything. Update docs.
9620 *) Add options to some of the utilities to allow the pass phrase
9621 to be included on either the command line (not recommended on
9622 OSes like Unix) or read from the environment. Update the
9623 manpages and fix a few bugs.
9626 *) Add a few manpages for some of the openssl commands.
9629 *) Fix the -revoke option in ca. It was freeing up memory twice,
9630 leaking and not finding already revoked certificates.
9633 *) Extensive changes to support certificate auxiliary information.
9634 This involves the use of X509_CERT_AUX structure and X509_AUX
9635 functions. An X509_AUX function such as PEM_read_X509_AUX()
9636 can still read in a certificate file in the usual way but it
9637 will also read in any additional "auxiliary information". By
9638 doing things this way a fair degree of compatibility can be
9639 retained: existing certificates can have this information added
9640 using the new 'x509' options.
9642 Current auxiliary information includes an "alias" and some trust
9643 settings. The trust settings will ultimately be used in enhanced
9644 certificate chain verification routines: currently a certificate
9645 can only be trusted if it is self signed and then it is trusted
9649 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9650 The problem was that one of the replacement routines had not been working
9651 since SSLeay releases. For now the offending routine has been replaced
9652 with non-optimised assembler. Even so, this now gives around 95%
9653 performance improvement for 1024 bit RSA signs.
9656 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9657 handling. Most clients have the effective key size in bits equal to
9658 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9659 A few however don't do this and instead use the size of the decrypted key
9660 to determine the RC2 key length and the AlgorithmIdentifier to determine
9661 the effective key length. In this case the effective key length can still
9662 be 40 bits but the key length can be 168 bits for example. This is fixed
9663 by manually forcing an RC2 key into the EVP_PKEY structure because the
9664 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9665 the key length and effective key length are equal.
9668 *) Add a bunch of functions that should simplify the creation of
9669 X509_NAME structures. Now you should be able to do:
9670 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9671 and have it automatically work out the correct field type and fill in
9672 the structures. The more adventurous can try:
9673 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9674 and it will (hopefully) work out the correct multibyte encoding.
9677 *) Change the 'req' utility to use the new field handling and multibyte
9678 copy routines. Before the DN field creation was handled in an ad hoc
9679 way in req, ca, and x509 which was rather broken and didn't support
9680 BMPStrings or UTF8Strings. Since some software doesn't implement
9681 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9682 using the dirstring_type option. See the new comment in the default
9683 openssl.cnf for more info.
9686 *) Make crypto/rand/md_rand.c more robust:
9687 - Assure unique random numbers after fork().
9688 - Make sure that concurrent threads access the global counter and
9689 md serializably so that we never lose entropy in them
9690 or use exactly the same state in multiple threads.
9691 Access to the large state is not always serializable because
9692 the additional locking could be a performance killer, and
9693 md should be large enough anyway.
9696 *) New file apps/app_rand.c with commonly needed functionality
9697 for handling the random seed file.
9699 Use the random seed file in some applications that previously did not:
9701 dsaparam -genkey (which also ignored its '-rand' option),
9704 x509 (when signing).
9705 Except on systems with /dev/urandom, it is crucial to have a random
9706 seed file at least for key creation, DSA signing, and for DH exchanges;
9707 for RSA signatures we could do without one.
9709 gendh and gendsa (unlike genrsa) used to read only the first byte
9710 of each file listed in the '-rand' option. The function as previously
9711 found in genrsa is now in app_rand.c and is used by all programs
9712 that support '-rand'.
9715 *) In RAND_write_file, use mode 0600 for creating files;
9716 don't just chmod when it may be too late.
9719 *) Report an error from X509_STORE_load_locations
9720 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9723 *) New function ASN1_mbstring_copy() this copies a string in either
9724 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9725 into an ASN1_STRING type. A mask of permissible types is passed
9726 and it chooses the "minimal" type to use or an error if not type
9730 *) Add function equivalents to the various macros in asn1.h. The old
9731 macros are retained with an M_ prefix. Code inside the library can
9732 use the M_ macros. External code (including the openssl utility)
9733 should *NOT* in order to be "shared library friendly".
9736 *) Add various functions that can check a certificate's extensions
9737 to see if it usable for various purposes such as SSL client,
9738 server or S/MIME and CAs of these types. This is currently
9739 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9740 verification. Also added a -purpose flag to x509 utility to
9741 print out all the purposes.
9744 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9748 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9749 for, obtain and decode and extension and obtain its critical flag.
9750 This allows all the necessary extension code to be handled in a
9751 single function call.
9754 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9755 platforms. See crypto/rc4/rc4_enc.c for further details.
9758 *) New -noout option to asn1parse. This causes no output to be produced
9759 its main use is when combined with -strparse and -out to extract data
9760 from a file (which may not be in ASN.1 format).
9763 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9764 when producing the local key id.
9765 [Richard Levitte <levitte@stacken.kth.se>]
9767 *) New option -dhparam in s_server. This allows a DH parameter file to be
9768 stated explicitly. If it is not stated then it tries the first server
9769 certificate file. The previous behaviour hard coded the filename
9773 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9774 a public key to be input or output. For example:
9775 openssl rsa -in key.pem -pubout -out pubkey.pem
9776 Also added necessary DSA public key functions to handle this.
9779 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9780 in the message. This was handled by allowing
9781 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9782 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9784 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9785 to the end of the strings whereas this didn't. This would cause problems
9786 if strings read with d2i_ASN1_bytes() were later modified.
9787 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9789 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9790 data and it contains EOF it will end up returning an error. This is
9791 caused by input 46 bytes long. The cause is due to the way base64
9792 BIOs find the start of base64 encoded data. They do this by trying a
9793 trial decode on each line until they find one that works. When they
9794 do a flag is set and it starts again knowing it can pass all the
9795 data directly through the decoder. Unfortunately it doesn't reset
9796 the context it uses. This means that if EOF is reached an attempt
9797 is made to pass two EOFs through the context and this causes the
9798 resulting error. This can also cause other problems as well. As is
9799 usual with these problems it takes *ages* to find and the fix is
9800 trivial: move one line.
9801 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9803 *) Ugly workaround to get s_client and s_server working under Windows. The
9804 old code wouldn't work because it needed to select() on sockets and the
9805 tty (for keypresses and to see if data could be written). Win32 only
9806 supports select() on sockets so we select() with a 1s timeout on the
9807 sockets and then see if any characters are waiting to be read, if none
9808 are present then we retry, we also assume we can always write data to
9809 the tty. This isn't nice because the code then blocks until we've
9810 received a complete line of data and it is effectively polling the
9811 keyboard at 1s intervals: however it's quite a bit better than not
9812 working at all :-) A dedicated Windows application might handle this
9813 with an event loop for example.
9816 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9817 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9818 will be called when RSA_sign() and RSA_verify() are used. This is useful
9819 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9820 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9821 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9822 This necessitated the support of an extra signature type NID_md5_sha1
9823 for SSL signatures and modifications to the SSL library to use it instead
9824 of calling RSA_public_decrypt() and RSA_private_encrypt().
9827 *) Add new -verify -CAfile and -CApath options to the crl program, these
9828 will lookup a CRL issuers certificate and verify the signature in a
9829 similar way to the verify program. Tidy up the crl program so it
9830 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9831 less strict. It will now permit CRL extensions even if it is not
9832 a V2 CRL: this will allow it to tolerate some broken CRLs.
9835 *) Initialize all non-automatic variables each time one of the openssl
9836 sub-programs is started (this is necessary as they may be started
9837 multiple times from the "OpenSSL>" prompt).
9838 [Lennart Bang, Bodo Moeller]
9840 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9841 removing all other RSA functionality (this is what NO_RSA does). This
9842 is so (for example) those in the US can disable those operations covered
9843 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9847 *) Non-copying interface to BIO pairs.
9848 (still largely untested)
9851 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9852 ASCII string. This was handled independently in various places before.
9855 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9856 UTF8 strings a character at a time.
9859 *) Use client_version from client hello to select the protocol
9860 (s23_srvr.c) and for RSA client key exchange verification
9861 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9864 *) Add various utility functions to handle SPKACs, these were previously
9865 handled by poking round in the structure internals. Added new function
9866 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9867 print, verify and generate SPKACs. Based on an original idea from
9868 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9871 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9874 *) Allow the config file extension section to be overwritten on the
9875 command line. Based on an original idea from Massimiliano Pala
9876 <madwolf@comune.modena.it>. The new option is called -extensions
9877 and can be applied to ca, req and x509. Also -reqexts to override
9878 the request extensions in req and -crlexts to override the crl extensions
9882 *) Add new feature to the SPKAC handling in ca. Now you can include
9883 the same field multiple times by preceding it by "XXXX." for example:
9886 this is the same syntax as used in the req config file.
9889 *) Allow certificate extensions to be added to certificate requests. These
9890 are specified in a 'req_extensions' option of the req section of the
9891 config file. They can be printed out with the -text option to req but
9892 are otherwise ignored at present.
9895 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9896 data read consists of only the final block it would not decrypted because
9897 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9898 A misplaced 'break' also meant the decrypted final block might not be
9899 copied until the next read.
9902 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9903 a few extra parameters to the DH structure: these will be useful if
9904 for example we want the value of 'q' or implement X9.42 DH.
9907 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9908 provides hooks that allow the default DSA functions or functions on a
9909 "per key" basis to be replaced. This allows hardware acceleration and
9910 hardware key storage to be handled without major modification to the
9911 library. Also added low level modexp hooks and CRYPTO_EX structure and
9912 associated functions.
9915 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9916 as "read only": it can't be written to and the buffer it points to will
9917 not be freed. Reading from a read only BIO is much more efficient than
9918 a normal memory BIO. This was added because there are several times when
9919 an area of memory needs to be read from a BIO. The previous method was
9920 to create a memory BIO and write the data to it, this results in two
9921 copies of the data and an O(n^2) reading algorithm. There is a new
9922 function BIO_new_mem_buf() which creates a read only memory BIO from
9923 an area of memory. Also modified the PKCS#7 routines to use read only
9927 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9928 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9929 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9930 but a retry condition occured while trying to read the rest.
9933 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9934 NID_pkcs7_encrypted by default: this was wrong since this should almost
9935 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9936 the encrypted data type: this is a more sensible place to put it and it
9937 allows the PKCS#12 code to be tidied up that duplicated this
9941 *) Changed obj_dat.pl script so it takes its input and output files on
9942 the command line. This should avoid shell escape redirection problems
9946 *) Initial support for certificate extension requests, these are included
9947 in things like Xenroll certificate requests. Included functions to allow
9948 extensions to be obtained and added.
9951 *) -crlf option to s_client and s_server for sending newlines as
9952 CRLF (as required by many protocols).
9955 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9957 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9958 [Ralf S. Engelschall]
9960 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9961 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9963 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9967 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9968 DH parameters/keys (q is lost during that conversion, but the resulting
9969 DH parameters contain its length).
9971 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9972 much faster than DH_generate_parameters (which creates parameters
9973 where p = 2*q + 1), and also the smaller q makes DH computations
9974 much more efficient (160-bit exponentiation instead of 1024-bit
9975 exponentiation); so this provides a convenient way to support DHE
9976 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9977 utter importance to use
9978 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9980 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9981 when such DH parameters are used, because otherwise small subgroup
9982 attacks may become possible!
9985 *) Avoid memory leak in i2d_DHparams.
9988 *) Allow the -k option to be used more than once in the enc program:
9989 this allows the same encrypted message to be read by multiple recipients.
9992 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9993 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9994 it will always use the numerical form of the OID, even if it has a short
9998 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9999 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10000 otherwise bn_mod_exp was called. In the case of hardware keys for example
10001 no private key components need be present and it might store extra data
10002 in the RSA structure, which cannot be accessed from bn_mod_exp.
10003 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10004 private key operations.
10007 *) Added support for SPARC Linux.
10010 *) pem_password_cb function type incompatibly changed from
10011 typedef int pem_password_cb(char *buf, int size, int rwflag);
10013 ....(char *buf, int size, int rwflag, void *userdata);
10014 so that applications can pass data to their callbacks:
10015 The PEM[_ASN1]_{read,write}... functions and macros now take an
10016 additional void * argument, which is just handed through whenever
10017 the password callback is called.
10018 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10020 New function SSL_CTX_set_default_passwd_cb_userdata.
10022 Compatibility note: As many C implementations push function arguments
10023 onto the stack in reverse order, the new library version is likely to
10024 interoperate with programs that have been compiled with the old
10025 pem_password_cb definition (PEM_whatever takes some data that
10026 happens to be on the stack as its last argument, and the callback
10027 just ignores this garbage); but there is no guarantee whatsoever that
10030 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10031 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10032 problems not only on Windows, but also on some Unix platforms.
10033 To avoid problematic command lines, these definitions are now in an
10034 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10035 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10038 *) MIPS III/IV assembler module is reimplemented.
10041 *) More DES library cleanups: remove references to srand/rand and
10042 delete an unused file.
10045 *) Add support for the the free Netwide assembler (NASM) under Win32,
10046 since not many people have MASM (ml) and it can be hard to obtain.
10047 This is currently experimental but it seems to work OK and pass all
10048 the tests. Check out INSTALL.W32 for info.
10051 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10052 without temporary keys kept an extra copy of the server key,
10053 and connections with temporary keys did not free everything in case
10057 *) New function RSA_check_key and new openssl rsa option -check
10058 for verifying the consistency of RSA keys.
10059 [Ulf Moeller, Bodo Moeller]
10061 *) Various changes to make Win32 compile work:
10062 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10063 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10064 comparison" warnings.
10065 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10068 *) Add a debugging option to PKCS#5 v2 key generation function: when
10069 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10070 derived keys are printed to stderr.
10073 *) Copy the flags in ASN1_STRING_dup().
10074 [Roman E. Pavlov <pre@mo.msk.ru>]
10076 *) The x509 application mishandled signing requests containing DSA
10077 keys when the signing key was also DSA and the parameters didn't match.
10079 It was supposed to omit the parameters when they matched the signing key:
10080 the verifying software was then supposed to automatically use the CA's
10081 parameters if they were absent from the end user certificate.
10083 Omitting parameters is no longer recommended. The test was also
10084 the wrong way round! This was probably due to unusual behaviour in
10085 EVP_cmp_parameters() which returns 1 if the parameters match.
10086 This meant that parameters were omitted when they *didn't* match and
10087 the certificate was useless. Certificates signed with 'ca' didn't have
10089 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10091 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10092 The interface is as follows:
10093 Applications can use
10094 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10095 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10096 "off" is now the default.
10097 The library internally uses
10098 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10099 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10100 to disable memory-checking temporarily.
10102 Some inconsistent states that previously were possible (and were
10103 even the default) are now avoided.
10105 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10106 with each memory chunk allocated; this is occasionally more helpful
10107 than just having a counter.
10109 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10111 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10115 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10116 which largely parallels "options", but is for changing API behaviour,
10117 whereas "options" are about protocol behaviour.
10118 Initial "mode" flags are:
10120 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10121 a single record has been written.
10122 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10123 retries use the same buffer location.
10124 (But all of the contents must be
10128 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10131 *) Fix problems with no-hmac etc.
10132 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10134 *) New functions RSA_get_default_method(), RSA_set_method() and
10135 RSA_get_method(). These allows replacement of RSA_METHODs without having
10136 to mess around with the internals of an RSA structure.
10139 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10140 Also really enable memory leak checks in openssl.c and in some
10142 [Chad C. Mulligan, Bodo Moeller]
10144 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10145 up the length of negative integers. This has now been simplified to just
10146 store the length when it is first determined and use it later, rather
10147 than trying to keep track of where data is copied and updating it to
10149 [Steve Henson, reported by Brien Wheeler
10150 <bwheeler@authentica-security.com>]
10152 *) Add a new function PKCS7_signatureVerify. This allows the verification
10153 of a PKCS#7 signature but with the signing certificate passed to the
10154 function itself. This contrasts with PKCS7_dataVerify which assumes the
10155 certificate is present in the PKCS#7 structure. This isn't always the
10156 case: certificates can be omitted from a PKCS#7 structure and be
10157 distributed by "out of band" means (such as a certificate database).
10160 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10161 function prototypes in pem.h, also change util/mkdef.pl to add the
10162 necessary function names.
10165 *) mk1mf.pl (used by Windows builds) did not properly read the
10166 options set by Configure in the top level Makefile, and Configure
10167 was not even able to write more than one option correctly.
10168 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10171 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10172 file to be loaded from a BIO or FILE pointer. The BIO version will
10173 for example allow memory BIOs to contain config info.
10176 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10177 Whoever hopes to achieve shared-library compatibility across versions
10178 must use this, not the compile-time macro.
10179 (Exercise 0.9.4: Which is the minimum library version required by
10181 Note: All this applies only to multi-threaded programs, others don't
10185 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10186 through a BIO pair triggered the default case, i.e.
10187 SSLerr(...,SSL_R_UNKNOWN_STATE).
10190 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10191 can use the SSL library even if none of the specific BIOs is
10195 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10196 for the encoded length.
10197 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10199 *) Add initial documentation of the X509V3 functions.
10202 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10203 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10204 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10205 secure PKCS#8 private key format with a high iteration count.
10208 *) Fix determination of Perl interpreter: A perl or perl5
10209 _directory_ in $PATH was also accepted as the interpreter.
10210 [Ralf S. Engelschall]
10212 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10213 wrong with it but it was very old and did things like calling
10214 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10215 unusual formatting.
10218 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10219 to use the new extension code.
10222 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10223 with macros. This should make it easier to change their form, add extra
10224 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10228 *) Add to configuration table a new entry that can specify an alternative
10229 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10230 according to Mark Crispin <MRC@Panda.COM>.
10234 *) DES CBC did not update the IV. Weird.
10237 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10238 Changing the behaviour of the former might break existing programs --
10239 where IV updating is needed, des_ncbc_encrypt can be used.
10242 *) When bntest is run from "make test" it drives bc to check its
10243 calculations, as well as internally checking them. If an internal check
10244 fails, it needs to cause bc to give a non-zero result or make test carries
10245 on without noticing the failure. Fixed.
10248 *) DES library cleanups.
10251 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10252 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10253 ciphers. NOTE: although the key derivation function has been verified
10254 against some published test vectors it has not been extensively tested
10255 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10259 *) Instead of "mkdir -p", which is not fully portable, use new
10260 Perl script "util/mkdir-p.pl".
10263 *) Rewrite the way password based encryption (PBE) is handled. It used to
10264 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10265 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10266 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10267 the 'parameter' field of the AlgorithmIdentifier is passed to the
10268 underlying key generation function so it must do its own ASN1 parsing.
10269 This has also changed the EVP_PBE_CipherInit() function which now has a
10270 'parameter' argument instead of literal salt and iteration count values
10271 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10274 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10275 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10276 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10277 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10278 value was just used as a "magic string" and not used directly its
10279 value doesn't matter.
10282 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10286 *) "linux-sparc64" configuration (ultrapenguin).
10287 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10288 "linux-sparc" configuration.
10289 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10291 *) config now generates no-xxx options for missing ciphers.
10294 *) Support the EBCDIC character set (work in progress).
10295 File ebcdic.c not yet included because it has a different license.
10296 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10298 *) Support BS2000/OSD-POSIX.
10299 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10301 *) Make callbacks for key generation use void * instead of char *.
10304 *) Make S/MIME samples compile (not yet tested).
10307 *) Additional typesafe stacks.
10310 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10314 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10316 *) New configuration variant "sco5-gcc".
10318 *) Updated some demos.
10319 [Sean O Riordain, Wade Scholine]
10321 *) Add missing BIO_free at exit of pkcs12 application.
10324 *) Fix memory leak in conf.c.
10327 *) Updates for Win32 to assembler version of MD5.
10330 *) Set #! path to perl in apps/der_chop to where we found it
10331 instead of using a fixed path.
10334 *) SHA library changes for irix64-mips4-cc.
10337 *) Improvements for VMS support.
10341 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10343 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10344 This also avoids the problems with SC4.2 and unpatched SC5.
10345 [Andy Polyakov <appro@fy.chalmers.se>]
10347 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10348 These are required because of the typesafe stack would otherwise break
10349 existing code. If old code used a structure member which used to be STACK
10350 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10351 sk_num or sk_value it would produce an error because the num, data members
10352 are not present in STACK_OF. Now it just produces a warning. sk_set
10353 replaces the old method of assigning a value to sk_value
10354 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10355 that does this will no longer work (and should use sk_set instead) but
10356 this could be regarded as a "questionable" behaviour anyway.
10359 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10360 correctly handle encrypted S/MIME data.
10363 *) Change type of various DES function arguments from des_cblock
10364 (which means, in function argument declarations, pointer to char)
10365 to des_cblock * (meaning pointer to array with 8 char elements),
10366 which allows the compiler to do more typechecking; it was like
10367 that back in SSLeay, but with lots of ugly casts.
10369 Introduce new type const_des_cblock.
10372 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10373 problems: find RecipientInfo structure that matches recipient certificate
10374 and initialise the ASN1 structures properly based on passed cipher.
10377 *) Belatedly make the BN tests actually check the results.
10380 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10381 to and from BNs: it was completely broken. New compilation option
10382 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10383 key elements as negative integers.
10386 *) Reorganize and speed up MD5.
10387 [Andy Polyakov <appro@fy.chalmers.se>]
10390 [Richard Levitte <richard@levitte.org>]
10392 *) New option -out to asn1parse to allow the parsed structure to be
10393 output to a file. This is most useful when combined with the -strparse
10394 option to examine the output of things like OCTET STRINGS.
10397 *) Make SSL library a little more fool-proof by not requiring any longer
10398 that SSL_set_{accept,connect}_state be called before
10399 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10400 in many applications because usually everything *appeared* to work as
10401 intended anyway -- now it really works as intended).
10404 *) Move openssl.cnf out of lib/.
10407 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10408 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10409 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10410 [Ralf S. Engelschall]
10412 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10413 handle PKCS#7 enveloped data properly.
10414 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10416 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10417 copying pointers. The cert_st handling is changed by this in
10418 various ways (and thus what used to be known as ctx->default_cert
10419 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10420 any longer when s->cert does not give us what we need).
10421 ssl_cert_instantiate becomes obsolete by this change.
10422 As soon as we've got the new code right (possibly it already is?),
10423 we have solved a couple of bugs of the earlier code where s->cert
10424 was used as if it could not have been shared with other SSL structures.
10426 Note that using the SSL API in certain dirty ways now will result
10427 in different behaviour than observed with earlier library versions:
10428 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10429 does not influence s as it used to.
10431 In order to clean up things more thoroughly, inside SSL_SESSION
10432 we don't use CERT any longer, but a new structure SESS_CERT
10433 that holds per-session data (if available); currently, this is
10434 the peer's certificate chain and, for clients, the server's certificate
10435 and temporary key. CERT holds only those values that can have
10436 meaningful defaults in an SSL_CTX.
10439 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10440 from the internal representation. Various PKCS#7 fixes: remove some
10441 evil casts and set the enc_dig_alg field properly based on the signing
10445 *) Allow PKCS#12 password to be set from the command line or the
10446 environment. Let 'ca' get its config file name from the environment
10447 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10451 *) Allow certificate policies extension to use an IA5STRING for the
10452 organization field. This is contrary to the PKIX definition but
10453 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10457 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10458 without disallowing inline assembler and the like for non-pedantic builds.
10461 *) Support Borland C++ builder.
10462 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10464 *) Support Mingw32.
10467 *) SHA-1 cleanups and performance enhancements.
10468 [Andy Polyakov <appro@fy.chalmers.se>]
10470 *) Sparc v8plus assembler for the bignum library.
10471 [Andy Polyakov <appro@fy.chalmers.se>]
10473 *) Accept any -xxx and +xxx compiler options in Configure.
10476 *) Update HPUX configuration.
10479 *) Add missing sk_<type>_unshift() function to safestack.h
10480 [Ralf S. Engelschall]
10482 *) New function SSL_CTX_use_certificate_chain_file that sets the
10483 "extra_cert"s in addition to the certificate. (This makes sense
10484 only for "PEM" format files, as chains as a whole are not
10488 *) Support verify_depth from the SSL API.
10489 x509_vfy.c had what can be considered an off-by-one-error:
10490 Its depth (which was not part of the external interface)
10491 was actually counting the number of certificates in a chain;
10492 now it really counts the depth.
10495 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10496 instead of X509err, which often resulted in confusing error
10497 messages since the error codes are not globally unique
10498 (e.g. an alleged error in ssl3_accept when a certificate
10499 didn't match the private key).
10501 *) New function SSL_CTX_set_session_id_context that allows to set a default
10502 value (so that you don't need SSL_set_session_id_context for each
10503 connection using the SSL_CTX).
10506 *) OAEP decoding bug fix.
10509 *) Support INSTALL_PREFIX for package builders, as proposed by
10513 *) New Configure options "threads" and "no-threads". For systems
10514 where the proper compiler options are known (currently Solaris
10515 and Linux), "threads" is the default.
10518 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10521 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10522 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10523 such as /usr/local/bin.
10526 *) "make linux-shared" to build shared libraries.
10527 [Niels Poppe <niels@netbox.org>]
10529 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10532 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10533 extension adding in x509 utility.
10536 *) Remove NOPROTO sections and error code comments.
10539 *) Partial rewrite of the DEF file generator to now parse the ANSI
10543 *) New Configure options --prefix=DIR and --openssldir=DIR.
10546 *) Complete rewrite of the error code script(s). It is all now handled
10547 by one script at the top level which handles error code gathering,
10548 header rewriting and C source file generation. It should be much better
10549 than the old method: it now uses a modified version of Ulf's parser to
10550 read the ANSI prototypes in all header files (thus the old K&R definitions
10551 aren't needed for error creation any more) and do a better job of
10552 translating function codes into names. The old 'ASN1 error code imbedded
10553 in a comment' is no longer necessary and it doesn't use .err files which
10554 have now been deleted. Also the error code call doesn't have to appear all
10555 on one line (which resulted in some large lines...).
10558 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10561 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10562 0 (which usually indicates a closed connection), but continue reading.
10565 *) Fix some race conditions.
10568 *) Add support for CRL distribution points extension. Add Certificate
10569 Policies and CRL distribution points documentation.
10572 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10575 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10576 8 of keying material. Merlin has also confirmed interop with this fix
10577 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10578 [Merlin Hughes <merlin@baltimore.ie>]
10580 *) Fix lots of warnings.
10581 [Richard Levitte <levitte@stacken.kth.se>]
10583 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10584 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10585 [Richard Levitte <levitte@stacken.kth.se>]
10587 *) Fix problems with sizeof(long) == 8.
10588 [Andy Polyakov <appro@fy.chalmers.se>]
10590 *) Change functions to ANSI C.
10593 *) Fix typos in error codes.
10594 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10596 *) Remove defunct assembler files from Configure.
10599 *) SPARC v8 assembler BIGNUM implementation.
10600 [Andy Polyakov <appro@fy.chalmers.se>]
10602 *) Support for Certificate Policies extension: both print and set.
10603 Various additions to support the r2i method this uses.
10606 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10607 return a const string when you are expecting an allocated buffer.
10610 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10611 types DirectoryString and DisplayText.
10614 *) Add code to allow r2i extensions to access the configuration database,
10615 add an LHASH database driver and add several ctx helper functions.
10618 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10619 fail when they extended the size of a BIGNUM.
10622 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10623 support typesafe stack.
10626 *) Fix typo in SSL_[gs]et_options().
10627 [Nils Frostberg <nils@medcom.se>]
10629 *) Delete various functions and files that belonged to the (now obsolete)
10630 old X509V3 handling code.
10633 *) New Configure option "rsaref".
10636 *) Don't auto-generate pem.h.
10639 *) Introduce type-safe ASN.1 SETs.
10642 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10643 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10645 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10646 that links with OpenSSL (well at least cause lots of warnings), but fear
10647 not: the conversion is trivial, and it eliminates loads of evil casts. A
10648 few STACKed things have been converted already. Feel free to convert more.
10649 In the fullness of time, I'll do away with the STACK type altogether.
10652 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10653 specified in <certfile> by updating the entry in the index.txt file.
10654 This way one no longer has to edit the index.txt file manually for
10655 revoking a certificate. The -revoke option does the gory details now.
10656 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10658 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10659 `-text' option at all and this way the `-noout -text' combination was
10660 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10661 [Ralf S. Engelschall]
10663 *) Make sure a corresponding plain text error message exists for the
10664 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10665 verify callback function determined that a certificate was revoked.
10666 [Ralf S. Engelschall]
10668 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10669 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10670 all available cipers including rc5, which was forgotten until now.
10671 In order to let the testing shell script know which algorithms
10672 are available, a new (up to now undocumented) command
10673 "openssl list-cipher-commands" is used.
10676 *) Bugfix: s_client occasionally would sleep in select() when
10677 it should have checked SSL_pending() first.
10680 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10681 the raw DSA values prior to ASN.1 encoding.
10684 *) Tweaks to Configure
10685 [Niels Poppe <niels@netbox.org>]
10687 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10691 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10694 *) New config option to avoid instructions that are illegal on the 80386.
10695 The default code is faster, but requires at least a 486.
10698 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10699 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10700 same as SSL2_VERSION anyway.
10703 *) New "-showcerts" option for s_client.
10706 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10707 application. Various cleanups and fixes.
10710 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10711 modify error routines to work internally. Add error codes and PBE init
10712 to library startup routines.
10715 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10716 packing functions to asn1 and evp. Changed function names and error
10717 codes along the way.
10720 *) PKCS12 integration: and so it begins... First of several patches to
10721 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10722 objects to objects.h
10725 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10726 and display support for Thawte strong extranet extension.
10729 *) Add LinuxPPC support.
10730 [Jeff Dubrule <igor@pobox.org>]
10732 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10733 bn_div_words in alpha.s.
10734 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10736 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10737 OAEP isn't supported when OpenSSL is built with RSAref.
10738 [Ulf Moeller <ulf@fitug.de>]
10740 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10741 so they no longer are missing under -DNOPROTO.
10742 [Soren S. Jorvang <soren@t.dk>]
10745 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10747 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10748 doesn't work when the session is reused. Coming soon!
10751 *) Fix a security hole, that allows sessions to be reused in the wrong
10752 context thus bypassing client cert protection! All software that uses
10753 client certs and session caches in multiple contexts NEEDS PATCHING to
10754 allow session reuse! A fuller solution is in the works.
10755 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10757 *) Some more source tree cleanups (removed obsolete files
10758 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10759 permission on "config" script to be executable) and a fix for the INSTALL
10761 [Ulf Moeller <ulf@fitug.de>]
10763 *) Remove some legacy and erroneous uses of malloc, free instead of
10765 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10767 *) Make rsa_oaep_test return non-zero on error.
10768 [Ulf Moeller <ulf@fitug.de>]
10770 *) Add support for native Solaris shared libraries. Configure
10771 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10772 if someone would make that last step automatic.
10773 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10775 *) ctx_size was not built with the right compiler during "make links". Fixed.
10778 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10779 except NULL ciphers". This means the default cipher list will no longer
10780 enable NULL ciphers. They need to be specifically enabled e.g. with
10781 the string "DEFAULT:eNULL".
10784 *) Fix to RSA private encryption routines: if p < q then it would
10785 occasionally produce an invalid result. This will only happen with
10786 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10789 *) Be less restrictive and allow also `perl util/perlpath.pl
10790 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10791 because this way one can also use an interpreter named `perl5' (which is
10792 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10793 installed as `perl').
10794 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10796 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10797 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10799 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10800 advapi32.lib to Win32 build and change the pem test comparison
10801 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10802 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10803 and crypto/des/ede_cbcm_enc.c.
10806 *) DES quad checksum was broken on big-endian architectures. Fixed.
10809 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10810 Win32 test batch file so it (might) work again. The Win32 test batch file
10811 is horrible: I feel ill....
10814 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10815 in e_os.h. Audit of header files to check ANSI and non ANSI
10816 sections: 10 functions were absent from non ANSI section and not exported
10817 from Windows DLLs. Fixed up libeay.num for new functions.
10820 *) Make `openssl version' output lines consistent.
10821 [Ralf S. Engelschall]
10823 *) Fix Win32 symbol export lists for BIO functions: Added
10824 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10825 to ms/libeay{16,32}.def.
10826 [Ralf S. Engelschall]
10828 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10829 fine under Unix and passes some trivial tests I've now added. But the
10830 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10831 added to make sure no one expects that this stuff really works in the
10832 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10833 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10835 [Ralf S. Engelschall]
10837 *) Fix the generation of two part addresses in perl.
10838 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10840 *) Add config entry for Linux on MIPS.
10841 [John Tobey <jtobey@channel1.com>]
10843 *) Make links whenever Configure is run, unless we are on Windoze.
10846 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10847 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10851 *) Add a useful kludge to allow package maintainers to specify compiler and
10852 other platforms details on the command line without having to patch the
10853 Configure script everytime: One now can use ``perl Configure
10854 <id>:<details>'', i.e. platform ids are allowed to have details appended
10855 to them (separated by colons). This is treated as there would be a static
10856 pre-configured entry in Configure's %table under key <id> with value
10857 <details> and ``perl Configure <id>'' is called. So, when you want to
10858 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10859 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10860 now, which overrides the FreeBSD-elf entry on-the-fly.
10861 [Ralf S. Engelschall]
10863 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10866 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10867 on the `perl Configure ...' command line. This way one can compile
10868 OpenSSL libraries with Position Independent Code (PIC) which is needed
10869 for linking it into DSOs.
10870 [Ralf S. Engelschall]
10872 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10876 *) Cleaned up the LICENSE document: The official contact for any license
10877 questions now is the OpenSSL core team under openssl-core@openssl.org.
10878 And add a paragraph about the dual-license situation to make sure people
10879 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10880 to the OpenSSL toolkit.
10881 [Ralf S. Engelschall]
10883 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10884 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10885 Additionally cleaned up the `make links' target: Remove unnecessary
10886 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10887 to speed processing and no longer clutter the display with confusing
10888 stuff. Instead only the actually done links are displayed.
10889 [Ralf S. Engelschall]
10891 *) Permit null encryption ciphersuites, used for authentication only. It used
10892 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10893 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10897 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10898 signed attributes when verifying signatures (this would break them),
10899 the detached data encoding was wrong and public keys obtained using
10900 X509_get_pubkey() weren't freed.
10903 *) Add text documentation for the BUFFER functions. Also added a work around
10904 to a Win95 console bug. This was triggered by the password read stuff: the
10905 last character typed gets carried over to the next fread(). If you were
10906 generating a new cert request using 'req' for example then the last
10907 character of the passphrase would be CR which would then enter the first
10911 *) Added the new `Includes OpenSSL Cryptography Software' button as
10912 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10913 button and can be used by applications based on OpenSSL to show the
10914 relationship to the OpenSSL project.
10915 [Ralf S. Engelschall]
10917 *) Remove confusing variables in function signatures in files
10918 ssl/ssl_lib.c and ssl/ssl.h.
10919 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10921 *) Don't install bss_file.c under PREFIX/include/
10922 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10924 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10925 functions that return function pointers and has support for NT specific
10926 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10927 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10928 unsigned to signed types: this was killing the Win32 compile.
10931 *) Add new certificate file to stack functions,
10932 SSL_add_dir_cert_subjects_to_stack() and
10933 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10934 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10935 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10936 This means that Apache-SSL and similar packages don't have to mess around
10937 to add as many CAs as they want to the preferred list.
10940 *) Experiment with doxygen documentation. Currently only partially applied to
10942 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10943 openssl.doxy as the configuration file.
10946 *) Get rid of remaining C++-style comments which strict C compilers hate.
10947 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10949 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10950 compiled in by default: it has problems with large keys.
10953 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10954 DH private keys and/or callback functions which directly correspond to
10955 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10956 is needed for applications which have to configure certificates on a
10957 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10959 For the RSA certificate situation is makes no difference, but
10960 for the DSA certificate situation this fixes the "no shared cipher"
10961 problem where the OpenSSL cipher selection procedure failed because the
10962 temporary keys were not overtaken from the context and the API provided
10963 no way to reconfigure them.
10964 The new functions now let applications reconfigure the stuff and they
10965 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10966 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10967 non-public-API function ssl_cert_instantiate() is used as a helper
10968 function and also to reduce code redundancy inside ssl_rsa.c.
10969 [Ralf S. Engelschall]
10971 *) Move s_server -dcert and -dkey options out of the undocumented feature
10972 area because they are useful for the DSA situation and should be
10973 recognized by the users.
10974 [Ralf S. Engelschall]
10976 *) Fix the cipher decision scheme for export ciphers: the export bits are
10977 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10978 SSL_EXP_MASK. So, the original variable has to be used instead of the
10979 already masked variable.
10980 [Richard Levitte <levitte@stacken.kth.se>]
10982 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10983 [Richard Levitte <levitte@stacken.kth.se>]
10985 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10986 from `int' to `unsigned int' because it's a length and initialized by
10987 EVP_DigestFinal() which expects an `unsigned int *'.
10988 [Richard Levitte <levitte@stacken.kth.se>]
10990 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10991 script. Instead use the usual Shell->Perl transition trick.
10992 [Ralf S. Engelschall]
10994 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10995 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10996 -noout -modulus' as it's already the case for `openssl rsa -noout
10997 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10998 currently the public key is printed (a decision which was already done by
10999 `openssl dsa -modulus' in the past) which serves a similar purpose.
11000 Additionally the NO_RSA no longer completely removes the whole -modulus
11001 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11003 [Ralf S. Engelschall]
11005 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11006 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11007 [Arne Ansper <arne@ats.cyber.ee>]
11009 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11010 to be added. Now both 'req' and 'ca' can use new objects defined in the
11014 *) Add cool BIO that does syslog (or event log on NT).
11015 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11017 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11018 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11019 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11020 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11023 *) Add preliminary config info for new extension code.
11026 *) Make RSA_NO_PADDING really use no padding.
11027 [Ulf Moeller <ulf@fitug.de>]
11029 *) Generate errors when private/public key check is done.
11032 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11033 for some CRL extensions and new objects added.
11036 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11037 key usage extension and fuller support for authority key id.
11040 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11041 padding method for RSA, which is recommended for new applications in PKCS
11042 #1 v2.0 (RFC 2437, October 1998).
11043 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11044 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11045 against Bleichbacher's attack on RSA.
11046 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11049 *) Updates to the new SSL compression code
11050 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11052 *) Fix so that the version number in the master secret, when passed
11053 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11054 (because the server will not accept higher), that the version number
11055 is 0x03,0x01, not 0x03,0x00
11056 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11058 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11059 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11060 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11063 *) Support for RAW extensions where an arbitrary extension can be
11064 created by including its DER encoding. See apps/openssl.cnf for
11068 *) Make sure latest Perl versions don't interpret some generated C array
11069 code as Perl array code in the crypto/err/err_genc.pl script.
11070 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11072 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11073 not many people have the assembler. Various Win32 compilation fixes and
11074 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11075 build instructions.
11078 *) Modify configure script 'Configure' to automatically create crypto/date.h
11079 file under Win32 and also build pem.h from pem.org. New script
11080 util/mkfiles.pl to create the MINFO file on environments that can't do a
11081 'make files': perl util/mkfiles.pl >MINFO should work.
11084 *) Major rework of DES function declarations, in the pursuit of correctness
11085 and purity. As a result, many evil casts evaporated, and some weirdness,
11086 too. You may find this causes warnings in your code. Zapping your evil
11087 casts will probably fix them. Mostly.
11090 *) Fix for a typo in asn1.h. Bug fix to object creation script
11091 obj_dat.pl. It considered a zero in an object definition to mean
11092 "end of object": none of the objects in objects.h have any zeros
11093 so it wasn't spotted.
11094 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11096 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11097 Masking (CBCM). In the absence of test vectors, the best I have been able
11098 to do is check that the decrypt undoes the encrypt, so far. Send me test
11099 vectors if you have them.
11102 *) Correct calculation of key length for export ciphers (too much space was
11103 allocated for null ciphers). This has not been tested!
11106 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11107 message is now correct (it understands "crypto" and "ssl" on its
11108 command line). There is also now an "update" option. This will update
11109 the util/ssleay.num and util/libeay.num files with any new functions.
11111 perl util/mkdef.pl crypto ssl update
11112 it will update them.
11115 *) Overhauled the Perl interface (perl/*):
11116 - ported BN stuff to OpenSSL's different BN library
11117 - made the perl/ source tree CVS-aware
11118 - renamed the package from SSLeay to OpenSSL (the files still contain
11119 their history because I've copied them in the repository)
11120 - removed obsolete files (the test scripts will be replaced
11121 by better Test::Harness variants in the future)
11122 [Ralf S. Engelschall]
11124 *) First cut for a very conservative source tree cleanup:
11125 1. merge various obsolete readme texts into doc/ssleay.txt
11126 where we collect the old documents and readme texts.
11127 2. remove the first part of files where I'm already sure that we no
11128 longer need them because of three reasons: either they are just temporary
11129 files which were left by Eric or they are preserved original files where
11130 I've verified that the diff is also available in the CVS via "cvs diff
11131 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11132 the crypto/md/ stuff).
11133 [Ralf S. Engelschall]
11135 *) More extension code. Incomplete support for subject and issuer alt
11136 name, issuer and authority key id. Change the i2v function parameters
11137 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11138 what that's for :-) Fix to ASN1 macro which messed up
11139 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11142 *) Preliminary support for ENUMERATED type. This is largely copied from the
11146 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11147 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11149 *) Make sure `make rehash' target really finds the `openssl' program.
11150 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11152 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11153 like to hear about it if this slows down other processors.
11156 *) Add CygWin32 platform information to Configure script.
11157 [Alan Batie <batie@aahz.jf.intel.com>]
11159 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11160 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11162 *) New program nseq to manipulate netscape certificate sequences
11165 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11169 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11170 but the BN code had some problems that would cause failures when
11171 doing certificate verification and some other functions.
11172 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11174 *) Add ASN1 and PEM code to support netscape certificate sequences.
11177 *) Add ASN1 and PEM code to support netscape certificate sequences.
11180 *) Add several PKIX and private extended key usage OIDs.
11183 *) Modify the 'ca' program to handle the new extension code. Modify
11184 openssl.cnf for new extension format, add comments.
11187 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11188 and add a sample to openssl.cnf so req -x509 now adds appropriate
11192 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11193 error code, add initial support to X509_print() and x509 application.
11196 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11197 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11198 stuff is currently isolated and isn't even compiled yet.
11201 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11202 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11203 Removed the versions check from X509 routines when loading extensions:
11204 this allows certain broken certificates that don't set the version
11205 properly to be processed.
11208 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11209 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11210 can still be regenerated with "make depend".
11213 *) Spelling mistake in C version of CAST-128.
11214 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11216 *) Changes to the error generation code. The perl script err-code.pl
11217 now reads in the old error codes and retains the old numbers, only
11218 adding new ones if necessary. It also only changes the .err files if new
11219 codes are added. The makefiles have been modified to only insert errors
11220 when needed (to avoid needlessly modifying header files). This is done
11221 by only inserting errors if the .err file is newer than the auto generated
11222 C file. To rebuild all the error codes from scratch (the old behaviour)
11223 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11224 or delete all the .err files.
11227 *) CAST-128 was incorrectly implemented for short keys. The C version has
11228 been fixed, but is untested. The assembler versions are also fixed, but
11229 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11230 to regenerate it if needed.
11231 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11232 Hagino <itojun@kame.net>]
11234 *) File was opened incorrectly in randfile.c.
11235 [Ulf Möller <ulf@fitug.de>]
11237 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11238 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11239 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11240 al: it's just almost always a UTCTime. Note this patch adds new error
11241 codes so do a "make errors" if there are problems.
11244 *) Correct Linux 1 recognition in config.
11245 [Ulf Möller <ulf@fitug.de>]
11247 *) Remove pointless MD5 hash when using DSA keys in ca.
11248 [Anonymous <nobody@replay.com>]
11250 *) Generate an error if given an empty string as a cert directory. Also
11251 generate an error if handed NULL (previously returned 0 to indicate an
11252 error, but didn't set one).
11253 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11255 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11258 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11259 parameters. This was causing a warning which killed off the Win32 compile.
11262 *) Remove C++ style comments from crypto/bn/bn_local.h.
11263 [Neil Costigan <neil.costigan@celocom.com>]
11265 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11266 based on a text string, looking up short and long names and finally
11267 "dot" format. The "dot" format stuff didn't work. Added new function
11268 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11269 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11270 OID is not part of the table.
11273 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11274 X509_LOOKUP_by_alias().
11277 *) Sort openssl functions by name.
11280 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11281 encryption from sample DSA keys (in case anyone is interested the password
11285 *) Make _all_ *_free functions accept a NULL pointer.
11286 [Frans Heymans <fheymans@isaserver.be>]
11288 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11290 [Anonymous <nobody@replay.com>]
11292 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11293 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11295 *) Don't blow it for numeric -newkey arguments to apps/req.
11296 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11298 *) Temp key "for export" tests were wrong in s3_srvr.c.
11299 [Anonymous <nobody@replay.com>]
11301 *) Add prototype for temp key callback functions
11302 SSL_CTX_set_tmp_{rsa,dh}_callback().
11305 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11306 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11309 *) X509_name_add_entry() freed the wrong thing after an error.
11310 [Arne Ansper <arne@ats.cyber.ee>]
11312 *) rsa_eay.c would attempt to free a NULL context.
11313 [Arne Ansper <arne@ats.cyber.ee>]
11315 *) BIO_s_socket() had a broken should_retry() on Windoze.
11316 [Arne Ansper <arne@ats.cyber.ee>]
11318 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11319 [Arne Ansper <arne@ats.cyber.ee>]
11321 *) Make sure the already existing X509_STORE->depth variable is initialized
11322 in X509_STORE_new(), but document the fact that this variable is still
11323 unused in the certificate verification process.
11324 [Ralf S. Engelschall]
11326 *) Fix the various library and apps files to free up pkeys obtained from
11327 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11330 *) Fix reference counting in X509_PUBKEY_get(). This makes
11331 demos/maurice/example2.c work, amongst others, probably.
11332 [Steve Henson and Ben Laurie]
11334 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11335 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11336 are no longer created. This way we have a single and consistent command
11337 line interface `openssl <command>', similar to `cvs <command>'.
11338 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11340 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11341 BIT STRING wrapper always have zero unused bits.
11344 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11347 *) Make the top-level INSTALL documentation easier to understand.
11350 *) Makefiles updated to exit if an error occurs in a sub-directory
11351 make (including if user presses ^C) [Paul Sutton]
11353 *) Make Montgomery context stuff explicit in RSA data structure.
11356 *) Fix build order of pem and err to allow for generated pem.h.
11359 *) Fix renumbering bug in X509_NAME_delete_entry().
11362 *) Enhanced the err-ins.pl script so it makes the error library number
11363 global and can add a library name. This is needed for external ASN1 and
11364 other error libraries.
11367 *) Fixed sk_insert which never worked properly.
11370 *) Fix ASN1 macros so they can handle indefinite length constructed
11371 EXPLICIT tags. Some non standard certificates use these: they can now
11375 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11376 into a single doc/ssleay.txt bundle. This way the information is still
11377 preserved but no longer messes up this directory. Now it's new room for
11378 the new set of documentation files.
11379 [Ralf S. Engelschall]
11381 *) SETs were incorrectly DER encoded. This was a major pain, because they
11382 shared code with SEQUENCEs, which aren't coded the same. This means that
11383 almost everything to do with SETs or SEQUENCEs has either changed name or
11384 number of arguments.
11385 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11387 *) Fix test data to work with the above.
11390 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11391 was already fixed by Eric for 0.9.1 it seems.
11392 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11394 *) Autodetect FreeBSD3.
11397 *) Fix various bugs in Configure. This affects the following platforms:
11401 unixware-2.0-pentium
11405 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11406 before they are needed.
11409 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11413 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11415 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11416 changed SSLeay to OpenSSL in version strings.
11417 [Ralf S. Engelschall]
11419 *) Some fixups to the top-level documents.
11422 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11423 because the symlink to include/ was missing.
11424 [Ralf S. Engelschall]
11426 *) Incorporated the popular no-RSA/DSA-only patches
11427 which allow to compile a RSA-free SSLeay.
11428 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11430 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11431 when "ssleay" is still not found.
11432 [Ralf S. Engelschall]
11434 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11435 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11437 *) Updated the README file.
11438 [Ralf S. Engelschall]
11440 *) Added various .cvsignore files in the CVS repository subdirs
11441 to make a "cvs update" really silent.
11442 [Ralf S. Engelschall]
11444 *) Recompiled the error-definition header files and added
11445 missing symbols to the Win32 linker tables.
11446 [Ralf S. Engelschall]
11448 *) Cleaned up the top-level documents;
11449 o new files: CHANGES and LICENSE
11450 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11451 o merged COPYRIGHT into LICENSE
11452 o removed obsolete TODO file
11453 o renamed MICROSOFT to INSTALL.W32
11454 [Ralf S. Engelschall]
11456 *) Removed dummy files from the 0.9.1b source tree:
11457 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11458 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11459 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11460 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11461 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11462 [Ralf S. Engelschall]
11464 *) Added various platform portability fixes.
11467 *) The Genesis of the OpenSSL rpject:
11468 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11469 Young and Tim J. Hudson created while they were working for C2Net until
11471 [The OpenSSL Project]
11474 Changes between 0.9.0b and 0.9.1b [not released]
11476 *) Updated a few CA certificates under certs/
11479 *) Changed some BIGNUM api stuff.
11482 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11483 DGUX x86, Linux Alpha, etc.
11486 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11487 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11491 *) Add -strparse option to asn1pars program which parses nested
11493 [Dr Stephen Henson <shenson@bigfoot.com>]
11495 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11498 *) DSA fix for "ca" program.
11501 *) Added "-genkey" option to "dsaparam" program.
11504 *) Added RIPE MD160 (rmd160) message digest.
11507 *) Added -a (all) option to "ssleay version" command.
11510 *) Added PLATFORM define which is the id given to Configure.
11513 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11516 *) Extended the ASN.1 parser routines.
11519 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11522 *) Added a BN_CTX to the BN library.
11525 *) Fixed the weak key values in DES library
11528 *) Changed API in EVP library for cipher aliases.
11531 *) Added support for RC2/64bit cipher.
11534 *) Converted the lhash library to the crypto/mem.c functions.
11537 *) Added more recognized ASN.1 object ids.
11540 *) Added more RSA padding checks for SSL/TLS.
11543 *) Added BIO proxy/filter functionality.
11546 *) Added extra_certs to SSL_CTX which can be used
11547 send extra CA certificates to the client in the CA cert chain sending
11548 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11551 *) Now Fortezza is denied in the authentication phase because
11552 this is key exchange mechanism is not supported by SSLeay at all.
11555 *) Additional PKCS1 checks.
11558 *) Support the string "TLSv1" for all TLS v1 ciphers.
11561 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11562 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11565 *) Fixed a few memory leaks.
11568 *) Fixed various code and comment typos.
11571 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11572 bytes sent in the client random.
11573 [Edward Bishop <ebishop@spyglass.com>]