5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
8 cryptographic operations to be performed asynchronously as long as an
9 asynchronous capable engine is used. See the ASYNC_start_job() man page for
10 further details. Libssl has also had this capability integrated with the
11 introduction of the new mode SSL_MODE_ASYNC and associated error
12 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
13 pages. This work was developed in partnership with Intel Corp.
16 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
17 SSL_{CTX_}set1_curves() which can set a list.
20 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
21 curve you want to support using SSL_{CTX_}set1_curves().
24 *) State machine rewrite. The state machine code has been significantly
25 refactored in order to remove much duplication of code and solve issues
26 with the old code (see ssl/statem/README for further details). This change
27 does have some associated API changes. Notably the SSL_state() function
28 has been removed and replaced by SSL_get_state which now returns an
29 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
30 altogether. The previous handshake states defined in ssl.h and ssl3.h have
34 *) All instances of the string "ssleay" in the public API were replaced
35 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
36 Some error codes related to internal RSA_eay API's were renamed.
39 *) The demo files in crypto/threads were moved to demo/threads.
42 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron and sureware.
45 *) New ASN.1 embed macro.
47 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
48 structure is not allocated: it is part of the parent. That is instead of
56 This reduces memory fragmentation and make it impossible to accidentally
57 set a mandatory field to NULL.
59 This currently only works for some fields specifically a SEQUENCE, CHOICE,
60 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
61 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
65 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
68 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
69 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
70 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
71 DES and RC4 ciphersuites.
74 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
75 This changes the decoding behaviour for some invalid messages,
76 though the change is mostly in the more lenient direction, and
77 legacy behaviour is preserved as much as possible.
80 *) Fix no-stdio build.
81 [ David Woodhouse <David.Woodhouse@intel.com> and also
82 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
84 *) New testing framework
85 The testing framework has been largely rewritten and is now using
86 perl and the perl modules Test::Harness and an extended variant of
87 Test::More called OpenSSL::Test to do its work. All test scripts in
88 test/ have been rewritten into test recipes, and all direct calls to
89 executables in test/Makefile have become individual recipes using the
90 simplified testing OpenSSL::Test::Simple.
92 For documentation on our testing modules, do:
94 perldoc test/testlib/OpenSSL/Test/Simple.pm
95 perldoc test/testlib/OpenSSL/Test.pm
99 *) In DSA_generate_parameters_ex, if the provided seed is too short,
101 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
103 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
104 from RFC4279, RFC4785, RFC5487, RFC5489.
106 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
107 original RSA_PSK patch.
110 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
111 era flag was never set throughout the codebase (only read). Also removed
112 SSL3_FLAGS_POP_BUFFER which was only used if
113 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
116 *) Changed the default name options in the "ca", "crl", "req" and "x509"
117 to be "oneline" instead of "compat".
120 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
121 not aware of clients that still exhibit this bug, and the workaround
122 hasn't been working properly for a while.
125 *) The return type of BIO_number_read() and BIO_number_written() as well as
126 the corresponding num_read and num_write members in the BIO structure has
127 changed from unsigned long to uint64_t. On platforms where an unsigned
128 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
132 *) Given the pervasive nature of TLS extensions it is inadvisable to run
133 OpenSSL without support for them. It also means that maintaining
134 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
135 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
138 *) Removed support for the two export grade static DH ciphersuites
139 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
140 were newly added (along with a number of other static DH ciphersuites) to
141 1.0.2. However the two export ones have *never* worked since they were
142 introduced. It seems strange in any case to be adding new export
143 ciphersuites, and given "logjam" it also does not seem correct to fix them.
146 *) Version negotiation has been rewritten. In particular SSLv23_method(),
147 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
148 and turned into macros which simply call the new preferred function names
149 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
150 should use the new names instead. Also as part of this change the ssl23.h
151 header file has been removed.
154 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
155 code and the associated standard is no longer considered fit-for-purpose.
158 *) RT2547 was closed. When generating a private key, try to make the
159 output file readable only by the owner. This behavior change might
160 be noticeable when interacting with other software.
162 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
166 *) Added HTTP GET support to the ocsp command.
169 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
172 *) Added support for TLS extended master secret from
173 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
174 initial patch which was a great help during development.
177 *) All libssl internal structures have been removed from the public header
178 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
179 now redundant). Users should not attempt to access internal structures
180 directly. Instead they should use the provided API functions.
183 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
184 Access to deprecated functions can be re-enabled by running config with
185 "enable-deprecated". In addition applications wishing to use deprecated
186 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
187 will, by default, disable some transitive includes that previously existed
188 in the header files (e.g. ec.h will no longer, by default, include bn.h)
191 *) Added support for OCB mode. OpenSSL has been granted a patent license
192 compatible with the OpenSSL license for use of OCB. Details are available
193 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
194 for OCB can be removed by calling config with no-ocb.
197 *) SSLv2 support has been removed. It still supports receiving a SSLv2
198 compatible client hello.
201 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
202 done while fixing the error code for the key-too-small case.
203 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
205 *) CA.sh has been removmed; use CA.pl instead.
208 *) Removed old DES API.
211 *) Remove various unsupported platforms:
217 Sinix/ReliantUNIX RM400
222 16-bit platforms such as WIN16
225 *) Clean up OPENSSL_NO_xxx #define's
226 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
227 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
228 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
229 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
230 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
231 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
232 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
233 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
234 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
235 Remove MS_STATIC; it's a relic from platforms <32 bits.
238 *) Cleaned up dead code
239 Remove all but one '#ifdef undef' which is to be looked at.
242 *) Clean up calling of xxx_free routines.
243 Just like free(), fix most of the xxx_free routines to accept
244 NULL. Remove the non-null checks from callers. Save much code.
247 *) Add secure heap for storage of private keys (when possible).
248 Add BIO_s_secmem(), CBIGNUM, etc.
249 Contributed by Akamai Technologies under our Corporate CLA.
252 *) Experimental support for a new, fast, unbiased prime candidate generator,
253 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
254 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
256 *) New output format NSS in the sess_id command line tool. This allows
257 exporting the session id and the master key in NSS keylog format.
258 [Martin Kaiser <martin@kaiser.cx>]
260 *) Harmonize version and its documentation. -f flag is used to display
262 [mancha <mancha1@zoho.com>]
264 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
265 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
266 [mancha <mancha1@zoho.com>]
268 *) Fix some double frees. These are not thought to be exploitable.
269 [mancha <mancha1@zoho.com>]
271 *) A missing bounds check in the handling of the TLS heartbeat extension
272 can be used to reveal up to 64k of memory to a connected client or
275 Thanks for Neel Mehta of Google Security for discovering this bug and to
276 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
277 preparing the fix (CVE-2014-0160)
278 [Adam Langley, Bodo Moeller]
280 *) Fix for the attack described in the paper "Recovering OpenSSL
281 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
282 by Yuval Yarom and Naomi Benger. Details can be obtained from:
283 http://eprint.iacr.org/2014/140
285 Thanks to Yuval Yarom and Naomi Benger for discovering this
286 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
287 [Yuval Yarom and Naomi Benger]
289 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
290 this fixes a limitation in previous versions of OpenSSL.
293 *) Experimental encrypt-then-mac support.
295 Experimental support for encrypt then mac from
296 draft-gutmann-tls-encrypt-then-mac-02.txt
298 To enable it set the appropriate extension number (0x42 for the test
299 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
301 For non-compliant peers (i.e. just about everything) this should have no
304 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
308 *) Add EVP support for key wrapping algorithms, to avoid problems with
309 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
310 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
311 algorithms and include tests cases.
314 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
318 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
319 MGF1 digest and OAEP label.
322 *) Make openssl verify return errors.
323 [Chris Palmer <palmer@google.com> and Ben Laurie]
325 *) New function ASN1_TIME_diff to calculate the difference between two
326 ASN1_TIME structures or one structure and the current time.
329 *) Update fips_test_suite to support multiple command line options. New
330 test to induce all self test errors in sequence and check expected
334 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
335 sign or verify all in one operation.
338 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
339 test programs and fips_test_suite. Includes functionality to parse
340 the minimal script output of fipsalgest.pl directly.
343 *) Add authorisation parameter to FIPS_module_mode_set().
346 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
349 *) Use separate DRBG fields for internal and external flags. New function
350 FIPS_drbg_health_check() to perform on demand health checking. Add
351 generation tests to fips_test_suite with reduced health check interval to
352 demonstrate periodic health checking. Add "nodh" option to
353 fips_test_suite to skip very slow DH test.
356 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
360 *) More extensive health check for DRBG checking many more failure modes.
361 New function FIPS_selftest_drbg_all() to handle every possible DRBG
362 combination: call this in fips_test_suite.
365 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
366 and POST to handle Dual EC cases.
369 *) Add support for canonical generation of DSA parameter 'g'. See
372 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
373 POST to handle HMAC cases.
376 *) Add functions FIPS_module_version() and FIPS_module_version_text()
377 to return numerical and string versions of the FIPS module number.
380 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
381 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
382 outside the validated module in the FIPS capable OpenSSL.
385 *) Minor change to DRBG entropy callback semantics. In some cases
386 there is no multiple of the block length between min_len and
387 max_len. Allow the callback to return more than max_len bytes
388 of entropy but discard any extra: it is the callback's responsibility
389 to ensure that the extra data discarded does not impact the
390 requested amount of entropy.
393 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
394 information in FIPS186-3, SP800-57 and SP800-131A.
397 *) CCM support via EVP. Interface is very similar to GCM case except we
398 must supply all data in one chunk (i.e. no update, final) and the
399 message length must be supplied if AAD is used. Add algorithm test
403 *) Initial version of POST overhaul. Add POST callback to allow the status
404 of POST to be monitored and/or failures induced. Modify fips_test_suite
405 to use callback. Always run all selftests even if one fails.
408 *) XTS support including algorithm test driver in the fips_gcmtest program.
409 Note: this does increase the maximum key length from 32 to 64 bytes but
410 there should be no binary compatibility issues as existing applications
411 will never use XTS mode.
414 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
415 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
416 performs algorithm blocking for unapproved PRNG types. Also do not
417 set PRNG type in FIPS_mode_set(): leave this to the application.
418 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
419 the standard OpenSSL PRNG: set additional data to a date time vector.
422 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
423 This shouldn't present any incompatibility problems because applications
424 shouldn't be using these directly and any that are will need to rethink
425 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
428 *) Extensive self tests and health checking required by SP800-90 DRBG.
429 Remove strength parameter from FIPS_drbg_instantiate and always
430 instantiate at maximum supported strength.
433 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
436 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
439 *) New function DH_compute_key_padded() to compute a DH key and pad with
440 leading zeroes if needed: this complies with SP800-56A et al.
443 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
444 anything, incomplete, subject to change and largely untested at present.
447 *) Modify fipscanisteronly build option to only build the necessary object
448 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
451 *) Add experimental option FIPSSYMS to give all symbols in
452 fipscanister.o and FIPS or fips prefix. This will avoid
453 conflicts with future versions of OpenSSL. Add perl script
454 util/fipsas.pl to preprocess assembly language source files
455 and rename any affected symbols.
458 *) Add selftest checks and algorithm block of non-fips algorithms in
459 FIPS mode. Remove DES2 from selftests.
462 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
463 return internal method without any ENGINE dependencies. Add new
464 tiny fips sign and verify functions.
467 *) New build option no-ec2m to disable characteristic 2 code.
470 *) New build option "fipscanisteronly". This only builds fipscanister.o
471 and (currently) associated fips utilities. Uses the file Makefile.fips
472 instead of Makefile.org as the prototype.
475 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
476 Update fips_gcmtest to use IV generator.
479 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
480 setting output buffer to NULL. The *Final function must be
481 called although it will not retrieve any additional data. The tag
482 can be set or retrieved with a ctrl. The IV length is by default 12
483 bytes (96 bits) but can be set to an alternative value. If the IV
484 length exceeds the maximum IV length (currently 16 bytes) it cannot be
488 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
489 underlying do_cipher function handles all cipher semantics itself
490 including padding and finalisation. This is useful if (for example)
491 an ENGINE cipher handles block padding itself. The behaviour of
492 do_cipher is subtly changed if this flag is set: the return value
493 is the number of characters written to the output buffer (zero is
494 no longer an error code) or a negative error code. Also if the
495 input buffer is NULL and length 0 finalisation should be performed.
498 *) If a candidate issuer certificate is already part of the constructed
499 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
502 *) Improve forward-security support: add functions
504 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
505 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
507 for use by SSL/TLS servers; the callback function will be called whenever a
508 new session is created, and gets to decide whether the session may be
509 cached to make it resumable (return 0) or not (return 1). (As by the
510 SSL/TLS protocol specifications, the session_id sent by the server will be
511 empty to indicate that the session is not resumable; also, the server will
512 not generate RFC 4507 (RFC 5077) session tickets.)
514 A simple reasonable callback implementation is to return is_forward_secure.
515 This parameter will be set to 1 or 0 depending on the ciphersuite selected
516 by the SSL/TLS server library, indicating whether it can provide forward
518 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
520 *) New -verify_name option in command line utilities to set verification
524 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
525 Add CMAC pkey methods.
528 *) Experimental renegotiation in s_server -www mode. If the client
529 browses /reneg connection is renegotiated. If /renegcert it is
530 renegotiated requesting a certificate.
533 *) Add an "external" session cache for debugging purposes to s_server. This
534 should help trace issues which normally are only apparent in deployed
535 multi-process servers.
538 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
539 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
540 BIO_set_cipher() and some obscure PEM functions were changed so they
541 can now return an error. The RAND changes required a change to the
542 RAND_METHOD structure.
545 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
546 a gcc attribute to warn if the result of a function is ignored. This
547 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
548 whose return value is often ignored.
551 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
553 *) Alternate chains certificate forgery
555 During certificate verfification, OpenSSL will attempt to find an
556 alternative certificate chain if the first attempt to build such a chain
557 fails. An error in the implementation of this logic can mean that an
558 attacker could cause certain checks on untrusted certificates to be
559 bypassed, such as the CA flag, enabling them to use a valid leaf
560 certificate to act as a CA and "issue" an invalid certificate.
562 This issue was reported to OpenSSL by Adam Langley/David Benjamin
566 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
568 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
569 incompatibility in the handling of HMAC. The previous ABI has now been
573 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
575 *) Malformed ECParameters causes infinite loop
577 When processing an ECParameters structure OpenSSL enters an infinite loop
578 if the curve specified is over a specially malformed binary polynomial
581 This can be used to perform denial of service against any
582 system which processes public keys, certificate requests or
583 certificates. This includes TLS clients and TLS servers with
584 client authentication enabled.
586 This issue was reported to OpenSSL by Joseph Barr-Pixton.
590 *) Exploitable out-of-bounds read in X509_cmp_time
592 X509_cmp_time does not properly check the length of the ASN1_TIME
593 string and can read a few bytes out of bounds. In addition,
594 X509_cmp_time accepts an arbitrary number of fractional seconds in the
597 An attacker can use this to craft malformed certificates and CRLs of
598 various sizes and potentially cause a segmentation fault, resulting in
599 a DoS on applications that verify certificates or CRLs. TLS clients
600 that verify CRLs are affected. TLS clients and servers with client
601 authentication enabled may be affected if they use custom verification
604 This issue was reported to OpenSSL by Robert Swiecki (Google), and
605 independently by Hanno Böck.
609 *) PKCS7 crash with missing EnvelopedContent
611 The PKCS#7 parsing code does not handle missing inner EncryptedContent
612 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
613 with missing content and trigger a NULL pointer dereference on parsing.
615 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
616 structures from untrusted sources are affected. OpenSSL clients and
617 servers are not affected.
619 This issue was reported to OpenSSL by Michal Zalewski (Google).
623 *) CMS verify infinite loop with unknown hash function
625 When verifying a signedData message the CMS code can enter an infinite loop
626 if presented with an unknown hash function OID. This can be used to perform
627 denial of service against any system which verifies signedData messages using
629 This issue was reported to OpenSSL by Johannes Bauer.
633 *) Race condition handling NewSessionTicket
635 If a NewSessionTicket is received by a multi-threaded client when attempting to
636 reuse a previous ticket then a race condition can occur potentially leading to
637 a double free of the ticket data.
641 *) Only support 256-bit or stronger elliptic curves with the
642 'ecdh_auto' setting (server) or by default (client). Of supported
643 curves, prefer P-256 (both).
646 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
648 *) ClientHello sigalgs DoS fix
650 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
651 invalid signature algorithms extension a NULL pointer dereference will
652 occur. This can be exploited in a DoS attack against the server.
654 This issue was was reported to OpenSSL by David Ramos of Stanford
657 [Stephen Henson and Matt Caswell]
659 *) Multiblock corrupted pointer fix
661 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
662 feature only applies on 64 bit x86 architecture platforms that support AES
663 NI instructions. A defect in the implementation of "multiblock" can cause
664 OpenSSL's internal write buffer to become incorrectly set to NULL when
665 using non-blocking IO. Typically, when the user application is using a
666 socket BIO for writing, this will only result in a failed connection.
667 However if some other BIO is used then it is likely that a segmentation
668 fault will be triggered, thus enabling a potential DoS attack.
670 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
674 *) Segmentation fault in DTLSv1_listen fix
676 The DTLSv1_listen function is intended to be stateless and processes the
677 initial ClientHello from many peers. It is common for user code to loop
678 over the call to DTLSv1_listen until a valid ClientHello is received with
679 an associated cookie. A defect in the implementation of DTLSv1_listen means
680 that state is preserved in the SSL object from one invocation to the next
681 that can lead to a segmentation fault. Errors processing the initial
682 ClientHello can trigger this scenario. An example of such an error could be
683 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
686 This issue was reported to OpenSSL by Per Allansson.
690 *) Segmentation fault in ASN1_TYPE_cmp fix
692 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
693 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
694 certificate signature algorithm consistency this can be used to crash any
695 certificate verification operation and exploited in a DoS attack. Any
696 application which performs certificate verification is vulnerable including
697 OpenSSL clients and servers which enable client authentication.
701 *) Segmentation fault for invalid PSS parameters fix
703 The signature verification routines will crash with a NULL pointer
704 dereference if presented with an ASN.1 signature using the RSA PSS
705 algorithm and invalid parameters. Since these routines are used to verify
706 certificate signature algorithms this can be used to crash any
707 certificate verification operation and exploited in a DoS attack. Any
708 application which performs certificate verification is vulnerable including
709 OpenSSL clients and servers which enable client authentication.
711 This issue was was reported to OpenSSL by Brian Carpenter.
715 *) ASN.1 structure reuse memory corruption fix
717 Reusing a structure in ASN.1 parsing may allow an attacker to cause
718 memory corruption via an invalid write. Such reuse is and has been
719 strongly discouraged and is believed to be rare.
721 Applications that parse structures containing CHOICE or ANY DEFINED BY
722 components may be affected. Certificate parsing (d2i_X509 and related
723 functions) are however not affected. OpenSSL clients and servers are
728 *) PKCS7 NULL pointer dereferences fix
730 The PKCS#7 parsing code does not handle missing outer ContentInfo
731 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
732 missing content and trigger a NULL pointer dereference on parsing.
734 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
735 otherwise parse PKCS#7 structures from untrusted sources are
736 affected. OpenSSL clients and servers are not affected.
738 This issue was reported to OpenSSL by Michal Zalewski (Google).
742 *) DoS via reachable assert in SSLv2 servers fix
744 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
745 servers that both support SSLv2 and enable export cipher suites by sending
746 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
748 This issue was discovered by Sean Burford (Google) and Emilia Käsper
749 (OpenSSL development team).
753 *) Empty CKE with client auth and DHE fix
755 If client auth is used then a server can seg fault in the event of a DHE
756 ciphersuite being selected and a zero length ClientKeyExchange message
757 being sent by the client. This could be exploited in a DoS attack.
761 *) Handshake with unseeded PRNG fix
763 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
764 with an unseeded PRNG. The conditions are:
765 - The client is on a platform where the PRNG has not been seeded
766 automatically, and the user has not seeded manually
767 - A protocol specific client method version has been used (i.e. not
768 SSL_client_methodv23)
769 - A ciphersuite is used that does not require additional random data from
770 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
772 If the handshake succeeds then the client random that has been used will
773 have been generated from a PRNG with insufficient entropy and therefore the
774 output may be predictable.
776 For example using the following command with an unseeded openssl will
777 succeed on an unpatched platform:
779 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
783 *) Use After Free following d2i_ECPrivatekey error fix
785 A malformed EC private key file consumed via the d2i_ECPrivateKey function
786 could cause a use after free condition. This, in turn, could cause a double
787 free in several private key parsing functions (such as d2i_PrivateKey
788 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
789 for applications that receive EC private keys from untrusted
790 sources. This scenario is considered rare.
792 This issue was discovered by the BoringSSL project and fixed in their
797 *) X509_to_X509_REQ NULL pointer deref fix
799 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
800 the certificate key is invalid. This function is rarely used in practice.
802 This issue was discovered by Brian Carpenter.
806 *) Removed the export ciphers from the DEFAULT ciphers
809 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
811 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
812 ARMv5 through ARMv8, as opposite to "locking" it to single one.
813 So far those who have to target multiple plaforms would compromise
814 and argue that binary targeting say ARMv5 would still execute on
815 ARMv8. "Universal" build resolves this compromise by providing
816 near-optimal performance even on newer platforms.
819 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
820 (other platforms pending).
821 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
823 *) Add support for the SignedCertificateTimestampList certificate and
824 OCSP response extensions from RFC6962.
827 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
828 for corner cases. (Certain input points at infinity could lead to
829 bogus results, with non-infinity inputs mapped to infinity too.)
832 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
833 This covers AES, SHA256/512 and GHASH. "Initial" means that most
834 common cases are optimized and there still is room for further
835 improvements. Vector Permutation AES for Altivec is also added.
838 *) Add support for little-endian ppc64 Linux target.
839 [Marcelo Cerri (IBM)]
841 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
842 SHA1, SHA256 and GHASH. "Initial" means that most common cases
843 are optimized and there still is room for further improvements.
844 Both 32- and 64-bit modes are supported.
845 [Andy Polyakov, Ard Biesheuvel (Linaro)]
847 *) Improved ARMv7 NEON support.
850 *) Support for SPARC Architecture 2011 crypto extensions, first
851 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
852 SHA256/512, MD5, GHASH and modular exponentiation.
853 [Andy Polyakov, David Miller]
855 *) Accelerated modular exponentiation for Intel processors, a.k.a.
857 [Shay Gueron & Vlad Krasnov (Intel Corp)]
859 *) Support for new and upcoming Intel processors, including AVX2,
860 BMI and SHA ISA extensions. This includes additional "stitched"
861 implementations, AESNI-SHA256 and GCM, and multi-buffer support
864 This work was sponsored by Intel Corp.
867 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
868 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
869 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
872 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
873 this fixes a limiation in previous versions of OpenSSL.
876 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
877 MGF1 digest and OAEP label.
880 *) Add EVP support for key wrapping algorithms, to avoid problems with
881 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
882 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
883 algorithms and include tests cases.
886 *) Add functions to allocate and set the fields of an ECDSA_METHOD
888 [Douglas E. Engert, Steve Henson]
890 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
891 difference in days and seconds between two tm or ASN1_TIME structures.
894 *) Add -rev test option to s_server to just reverse order of characters
895 received by client and send back to server. Also prints an abbreviated
896 summary of the connection parameters.
899 *) New option -brief for s_client and s_server to print out a brief summary
900 of connection parameters.
903 *) Add callbacks for arbitrary TLS extensions.
904 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
906 *) New option -crl_download in several openssl utilities to download CRLs
907 from CRLDP extension in certificates.
910 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
913 *) New function X509_CRL_diff to generate a delta CRL from the difference
914 of two full CRLs. Add support to "crl" utility.
917 *) New functions to set lookup_crls function and to retrieve
918 X509_STORE from X509_STORE_CTX.
921 *) Print out deprecated issuer and subject unique ID fields in
925 *) Extend OCSP I/O functions so they can be used for simple general purpose
926 HTTP as well as OCSP. New wrapper function which can be used to download
927 CRLs using the OCSP API.
930 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
933 *) SSL_CONF* functions. These provide a common framework for application
934 configuration using configuration files or command lines.
937 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
938 message callback and prints the results. Needs compile time option
939 "enable-ssl-trace". New options to s_client and s_server to enable
943 *) New ctrl and macro to retrieve supported points extensions.
944 Print out extension in s_server and s_client.
947 *) New functions to retrieve certificate signature and signature
951 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
955 *) New Suite B modes for TLS code. These use and enforce the requirements
956 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
957 only use Suite B curves. The Suite B modes can be set by using the
958 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
961 *) New chain verification flags for Suite B levels of security. Check
962 algorithms are acceptable when flags are set in X509_verify_cert.
965 *) Make tls1_check_chain return a set of flags indicating checks passed
966 by a certificate chain. Add additional tests to handle client
967 certificates: checks for matching certificate type and issuer name
971 *) If an attempt is made to use a signature algorithm not in the peer
972 preference list abort the handshake. If client has no suitable
973 signature algorithms in response to a certificate request do not
977 *) If server EC tmp key is not in client preference list abort handshake.
980 *) Add support for certificate stores in CERT structure. This makes it
981 possible to have different stores per SSL structure or one store in
982 the parent SSL_CTX. Include distint stores for certificate chain
983 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
984 to build and store a certificate chain in CERT structure: returing
985 an error if the chain cannot be built: this will allow applications
986 to test if a chain is correctly configured.
988 Note: if the CERT based stores are not set then the parent SSL_CTX
989 store is used to retain compatibility with existing behaviour.
993 *) New function ssl_set_client_disabled to set a ciphersuite disabled
994 mask based on the current session, check mask when sending client
995 hello and checking the requested ciphersuite.
998 *) New ctrls to retrieve and set certificate types in a certificate
999 request message. Print out received values in s_client. If certificate
1000 types is not set with custom values set sensible values based on
1001 supported signature algorithms.
1004 *) Support for distinct client and server supported signature algorithms.
1007 *) Add certificate callback. If set this is called whenever a certificate
1008 is required by client or server. An application can decide which
1009 certificate chain to present based on arbitrary criteria: for example
1010 supported signature algorithms. Add very simple example to s_server.
1011 This fixes many of the problems and restrictions of the existing client
1012 certificate callback: for example you can now clear an existing
1013 certificate and specify the whole chain.
1016 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1017 the certificate can be used for (if anything). Set valid_flags field
1018 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1019 to have similar checks in it.
1021 Add new "cert_flags" field to CERT structure and include a "strict mode".
1022 This enforces some TLS certificate requirements (such as only permitting
1023 certificate signature algorithms contained in the supported algorithms
1024 extension) which some implementations ignore: this option should be used
1025 with caution as it could cause interoperability issues.
1028 *) Update and tidy signature algorithm extension processing. Work out
1029 shared signature algorithms based on preferences and peer algorithms
1030 and print them out in s_client and s_server. Abort handshake if no
1031 shared signature algorithms.
1034 *) Add new functions to allow customised supported signature algorithms
1035 for SSL and SSL_CTX structures. Add options to s_client and s_server
1039 *) New function SSL_certs_clear() to delete all references to certificates
1040 from an SSL structure. Before this once a certificate had been added
1041 it couldn't be removed.
1044 *) Integrate hostname, email address and IP address checking with certificate
1045 verification. New verify options supporting checking in opensl utility.
1048 *) Fixes and wildcard matching support to hostname and email checking
1049 functions. Add manual page.
1050 [Florian Weimer (Red Hat Product Security Team)]
1052 *) New functions to check a hostname email or IP address against a
1053 certificate. Add options x509 utility to print results of checks against
1057 *) Fix OCSP checking.
1058 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1060 *) Initial experimental support for explicitly trusted non-root CAs.
1061 OpenSSL still tries to build a complete chain to a root but if an
1062 intermediate CA has a trust setting included that is used. The first
1063 setting is used: whether to trust (e.g., -addtrust option to the x509
1067 *) Add -trusted_first option which attempts to find certificates in the
1068 trusted store even if an untrusted chain is also supplied.
1071 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1072 platform support for Linux and Android.
1075 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1078 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1079 When in FIPS mode the approved implementations are used as normal,
1080 when not in FIPS mode the internal unapproved versions are used instead.
1081 This means that the FIPS capable OpenSSL isn't forced to use the
1082 (often lower performance) FIPS implementations outside FIPS mode.
1085 *) Transparently support X9.42 DH parameters when calling
1086 PEM_read_bio_DHparameters. This means existing applications can handle
1087 the new parameter format automatically.
1090 *) Initial experimental support for X9.42 DH parameter format: mainly
1091 to support use of 'q' parameter for RFC5114 parameters.
1094 *) Add DH parameters from RFC5114 including test data to dhtest.
1097 *) Support for automatic EC temporary key parameter selection. If enabled
1098 the most preferred EC parameters are automatically used instead of
1099 hardcoded fixed parameters. Now a server just has to call:
1100 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1101 support ECDH and use the most appropriate parameters.
1104 *) Enhance and tidy EC curve and point format TLS extension code. Use
1105 static structures instead of allocation if default values are used.
1106 New ctrls to set curves we wish to support and to retrieve shared curves.
1107 Print out shared curves in s_server. New options to s_server and s_client
1108 to set list of supported curves.
1111 *) New ctrls to retrieve supported signature algorithms and
1112 supported curve values as an array of NIDs. Extend openssl utility
1113 to print out received values.
1116 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1117 between NIDs and the more common NIST names such as "P-256". Enhance
1118 ecparam utility and ECC method to recognise the NIST names for curves.
1121 *) Enhance SSL/TLS certificate chain handling to support different
1122 chains for each certificate instead of one chain in the parent SSL_CTX.
1125 *) Support for fixed DH ciphersuite client authentication: where both
1126 server and client use DH certificates with common parameters.
1129 *) Support for fixed DH ciphersuites: those requiring DH server
1133 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1135 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1136 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1137 X509_CINF_get_signature were reverted post internal team review.
1139 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1141 *) Build fixes for the Windows and OpenVMS platforms
1142 [Matt Caswell and Richard Levitte]
1144 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1146 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1147 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1148 dereference. This could lead to a Denial Of Service attack. Thanks to
1149 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1153 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1154 dtls1_buffer_record function under certain conditions. In particular this
1155 could occur if an attacker sent repeated DTLS records with the same
1156 sequence number but for the next epoch. The memory leak could be exploited
1157 by an attacker in a Denial of Service attack through memory exhaustion.
1158 Thanks to Chris Mueller for reporting this issue.
1162 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1163 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1164 method would be set to NULL which could later result in a NULL pointer
1165 dereference. Thanks to Frank Schmirler for reporting this issue.
1169 *) Abort handshake if server key exchange message is omitted for ephemeral
1172 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1173 reporting this issue.
1177 *) Remove non-export ephemeral RSA code on client and server. This code
1178 violated the TLS standard by allowing the use of temporary RSA keys in
1179 non-export ciphersuites and could be used by a server to effectively
1180 downgrade the RSA key length used to a value smaller than the server
1181 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1182 INRIA or reporting this issue.
1186 *) Fixed issue where DH client certificates are accepted without verification.
1187 An OpenSSL server will accept a DH certificate for client authentication
1188 without the certificate verify message. This effectively allows a client to
1189 authenticate without the use of a private key. This only affects servers
1190 which trust a client certificate authority which issues certificates
1191 containing DH keys: these are extremely rare and hardly ever encountered.
1192 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1197 *) Ensure that the session ID context of an SSL is updated when its
1198 SSL_CTX is updated via SSL_set_SSL_CTX.
1200 The session ID context is typically set from the parent SSL_CTX,
1201 and can vary with the CTX.
1204 *) Fix various certificate fingerprint issues.
1206 By using non-DER or invalid encodings outside the signed portion of a
1207 certificate the fingerprint can be changed without breaking the signature.
1208 Although no details of the signed portion of the certificate can be changed
1209 this can cause problems with some applications: e.g. those using the
1210 certificate fingerprint for blacklists.
1212 1. Reject signatures with non zero unused bits.
1214 If the BIT STRING containing the signature has non zero unused bits reject
1215 the signature. All current signature algorithms require zero unused bits.
1217 2. Check certificate algorithm consistency.
1219 Check the AlgorithmIdentifier inside TBS matches the one in the
1220 certificate signature. NB: this will result in signature failure
1221 errors for some broken certificates.
1223 Thanks to Konrad Kraszewski from Google for reporting this issue.
1225 3. Check DSA/ECDSA signatures use DER.
1227 Reencode DSA/ECDSA signatures and compare with the original received
1228 signature. Return an error if there is a mismatch.
1230 This will reject various cases including garbage after signature
1231 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1232 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1233 (negative or with leading zeroes).
1235 Further analysis was conducted and fixes were developed by Stephen Henson
1236 of the OpenSSL core team.
1241 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1242 results on some platforms, including x86_64. This bug occurs at random
1243 with a very low probability, and is not known to be exploitable in any
1244 way, though its exact impact is difficult to determine. Thanks to Pieter
1245 Wuille (Blockstream) who reported this issue and also suggested an initial
1246 fix. Further analysis was conducted by the OpenSSL development team and
1247 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1248 the OpenSSL core team.
1252 *) Do not resume sessions on the server if the negotiated protocol
1253 version does not match the session's version. Resuming with a different
1254 version, while not strictly forbidden by the RFC, is of questionable
1255 sanity and breaks all known clients.
1256 [David Benjamin, Emilia Käsper]
1258 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1259 early CCS messages during renegotiation. (Note that because
1260 renegotiation is encrypted, this early CCS was not exploitable.)
1263 *) Tighten client-side session ticket handling during renegotiation:
1264 ensure that the client only accepts a session ticket if the server sends
1265 the extension anew in the ServerHello. Previously, a TLS client would
1266 reuse the old extension state and thus accept a session ticket if one was
1267 announced in the initial ServerHello.
1269 Similarly, ensure that the client requires a session ticket if one
1270 was advertised in the ServerHello. Previously, a TLS client would
1271 ignore a missing NewSessionTicket message.
1274 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1276 *) SRTP Memory Leak.
1278 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1279 sends a carefully crafted handshake message, to cause OpenSSL to fail
1280 to free up to 64k of memory causing a memory leak. This could be
1281 exploited in a Denial Of Service attack. This issue affects OpenSSL
1282 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1283 whether SRTP is used or configured. Implementations of OpenSSL that
1284 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1286 The fix was developed by the OpenSSL team.
1290 *) Session Ticket Memory Leak.
1292 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1293 integrity of that ticket is first verified. In the event of a session
1294 ticket integrity check failing, OpenSSL will fail to free memory
1295 causing a memory leak. By sending a large number of invalid session
1296 tickets an attacker could exploit this issue in a Denial Of Service
1301 *) Build option no-ssl3 is incomplete.
1303 When OpenSSL is configured with "no-ssl3" as a build option, servers
1304 could accept and complete a SSL 3.0 handshake, and clients could be
1305 configured to send them.
1307 [Akamai and the OpenSSL team]
1309 *) Add support for TLS_FALLBACK_SCSV.
1310 Client applications doing fallback retries should call
1311 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1313 [Adam Langley, Bodo Moeller]
1315 *) Add additional DigestInfo checks.
1317 Reencode DigestInto in DER and check against the original when
1318 verifying RSA signature: this will reject any improperly encoded
1319 DigestInfo structures.
1321 Note: this is a precautionary measure and no attacks are currently known.
1325 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1327 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1328 SRP code can be overrun an internal buffer. Add sanity check that
1329 g, A, B < N to SRP code.
1331 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1332 Group for discovering this issue.
1336 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1337 TLS 1.0 instead of higher protocol versions when the ClientHello message
1338 is badly fragmented. This allows a man-in-the-middle attacker to force a
1339 downgrade to TLS 1.0 even if both the server and the client support a
1340 higher protocol version, by modifying the client's TLS records.
1342 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1343 researching this issue.
1347 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1348 to a denial of service attack. A malicious server can crash the client
1349 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1350 ciphersuite and sending carefully crafted handshake messages.
1352 Thanks to Felix Gröbert (Google) for discovering and researching this
1357 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1358 to leak memory. This can be exploited through a Denial of Service attack.
1359 Thanks to Adam Langley for discovering and researching this issue.
1363 *) An attacker can force openssl to consume large amounts of memory whilst
1364 processing DTLS handshake messages. This can be exploited through a
1365 Denial of Service attack.
1366 Thanks to Adam Langley for discovering and researching this issue.
1370 *) An attacker can force an error condition which causes openssl to crash
1371 whilst processing DTLS packets due to memory being freed twice. This
1372 can be exploited through a Denial of Service attack.
1373 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1378 *) If a multithreaded client connects to a malicious server using a resumed
1379 session and the server sends an ec point format extension it could write
1380 up to 255 bytes to freed memory.
1382 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1387 *) A malicious server can crash an OpenSSL client with a null pointer
1388 dereference (read) by specifying an SRP ciphersuite even though it was not
1389 properly negotiated with the client. This can be exploited through a
1390 Denial of Service attack.
1392 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1393 discovering and researching this issue.
1397 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1398 X509_name_oneline, X509_name_print_ex et al. to leak some information
1399 from the stack. Applications may be affected if they echo pretty printing
1400 output to the attacker.
1402 Thanks to Ivan Fratric (Google) for discovering this issue.
1404 [Emilia Käsper, and Steve Henson]
1406 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1407 for corner cases. (Certain input points at infinity could lead to
1408 bogus results, with non-infinity inputs mapped to infinity too.)
1411 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1413 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1414 handshake can force the use of weak keying material in OpenSSL
1415 SSL/TLS clients and servers.
1417 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1418 researching this issue. (CVE-2014-0224)
1419 [KIKUCHI Masashi, Steve Henson]
1421 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1422 OpenSSL DTLS client the code can be made to recurse eventually crashing
1425 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1427 [Imre Rad, Steve Henson]
1429 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1430 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1431 client or server. This is potentially exploitable to run arbitrary
1432 code on a vulnerable client or server.
1434 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1435 [Jüri Aedla, Steve Henson]
1437 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1438 are subject to a denial of service attack.
1440 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1441 this issue. (CVE-2014-3470)
1442 [Felix Gröbert, Ivan Fratric, Steve Henson]
1444 *) Harmonize version and its documentation. -f flag is used to display
1446 [mancha <mancha1@zoho.com>]
1448 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1449 in i2d_ECPrivateKey.
1450 [mancha <mancha1@zoho.com>]
1452 *) Fix some double frees. These are not thought to be exploitable.
1453 [mancha <mancha1@zoho.com>]
1455 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1457 *) A missing bounds check in the handling of the TLS heartbeat extension
1458 can be used to reveal up to 64k of memory to a connected client or
1461 Thanks for Neel Mehta of Google Security for discovering this bug and to
1462 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1463 preparing the fix (CVE-2014-0160)
1464 [Adam Langley, Bodo Moeller]
1466 *) Fix for the attack described in the paper "Recovering OpenSSL
1467 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1468 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1469 http://eprint.iacr.org/2014/140
1471 Thanks to Yuval Yarom and Naomi Benger for discovering this
1472 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1473 [Yuval Yarom and Naomi Benger]
1475 *) TLS pad extension: draft-agl-tls-padding-03
1477 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1478 TLS client Hello record length value would otherwise be > 255 and
1479 less that 512 pad with a dummy extension containing zeroes so it
1480 is at least 512 bytes long.
1482 [Adam Langley, Steve Henson]
1484 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1486 *) Fix for TLS record tampering bug. A carefully crafted invalid
1487 handshake could crash OpenSSL with a NULL pointer exception.
1488 Thanks to Anton Johansson for reporting this issues.
1491 *) Keep original DTLS digest and encryption contexts in retransmission
1492 structures so we can use the previous session parameters if they need
1493 to be resent. (CVE-2013-6450)
1496 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1497 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1498 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1499 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1500 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1501 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1502 [Rob Stradling, Adam Langley]
1504 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1506 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1507 supporting platforms or when small records were transferred.
1508 [Andy Polyakov, Steve Henson]
1510 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1512 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1514 This addresses the flaw in CBC record processing discovered by
1515 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1516 at: http://www.isg.rhul.ac.uk/tls/
1518 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1519 Security Group at Royal Holloway, University of London
1520 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1521 Emilia Käsper for the initial patch.
1523 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1525 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1526 ciphersuites which can be exploited in a denial of service attack.
1527 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1528 and detecting this bug and to Wolfgang Ettlinger
1529 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1533 *) Return an error when checking OCSP signatures when key is NULL.
1534 This fixes a DoS attack. (CVE-2013-0166)
1537 *) Make openssl verify return errors.
1538 [Chris Palmer <palmer@google.com> and Ben Laurie]
1540 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1541 the right response is stapled. Also change SSL_get_certificate()
1542 so it returns the certificate actually sent.
1543 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1544 [Rob Stradling <rob.stradling@comodo.com>]
1546 *) Fix possible deadlock when decoding public keys.
1549 *) Don't use TLS 1.0 record version number in initial client hello
1553 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1555 *) Sanity check record length before skipping explicit IV in TLS
1556 1.2, 1.1 and DTLS to fix DoS attack.
1558 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1559 fuzzing as a service testing platform.
1563 *) Initialise tkeylen properly when encrypting CMS messages.
1564 Thanks to Solar Designer of Openwall for reporting this issue.
1567 *) In FIPS mode don't try to use composite ciphers as they are not
1571 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1573 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1574 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1575 mean any application compiled against OpenSSL 1.0.0 headers setting
1576 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1577 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1578 0x10000000L Any application which was previously compiled against
1579 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1580 will need to be recompiled as a result. Letting be results in
1581 inability to disable specifically TLS 1.1 and in client context,
1582 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1585 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1586 disable just protocol X, but all protocols above X *if* there are
1587 protocols *below* X still enabled. In more practical terms it means
1588 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1589 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1590 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1594 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1596 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1597 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1598 in CRYPTO_realloc_clean.
1600 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1601 issue and to Adam Langley <agl@chromium.org> for fixing it.
1603 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1605 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1608 *) Workarounds for some broken servers that "hang" if a client hello
1609 record length exceeds 255 bytes.
1611 1. Do not use record version number > TLS 1.0 in initial client
1612 hello: some (but not all) hanging servers will now work.
1613 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1614 the number of ciphers sent in the client hello. This should be
1615 set to an even number, such as 50, for example by passing:
1616 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1617 Most broken servers should now work.
1618 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1619 TLS 1.2 client support entirely.
1622 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1625 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1627 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1628 STRING form instead of a DigestInfo.
1631 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1632 and the RSA_sign/RSA_verify functions. This was made more apparent when
1633 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1634 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1635 the correct format in RSA_verify so both forms transparently work.
1638 *) Some servers which support TLS 1.0 can choke if we initially indicate
1639 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1640 encrypted premaster secret. As a workaround use the maximum permitted
1641 client version in client hello, this should keep such servers happy
1642 and still work with previous versions of OpenSSL.
1645 *) Add support for TLS/DTLS heartbeats.
1646 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1648 *) Add support for SCTP.
1649 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1651 *) Improved PRNG seeding for VOS.
1652 [Paul Green <Paul.Green@stratus.com>]
1654 *) Extensive assembler packs updates, most notably:
1656 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1657 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1658 - x86_64: bit-sliced AES implementation;
1659 - ARM: NEON support, contemporary platforms optimizations;
1660 - s390x: z196 support;
1661 - *: GHASH and GF(2^m) multiplication implementations;
1665 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1666 (removal of unnecessary code)
1667 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1669 *) Add TLS key material exporter from RFC 5705.
1672 *) Add DTLS-SRTP negotiation from RFC 5764.
1675 *) Add Next Protocol Negotiation,
1676 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1677 disabled with a no-npn flag to config or Configure. Code donated
1679 [Adam Langley <agl@google.com> and Ben Laurie]
1681 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1682 NIST-P256, NIST-P521, with constant-time single point multiplication on
1683 typical inputs. Compiler support for the nonstandard type __uint128_t is
1684 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1685 Code made available under Apache License version 2.0.
1687 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1688 line to include this in your build of OpenSSL, and run "make depend" (or
1689 "make update"). This enables the following EC_METHODs:
1691 EC_GFp_nistp224_method()
1692 EC_GFp_nistp256_method()
1693 EC_GFp_nistp521_method()
1695 EC_GROUP_new_by_curve_name() will automatically use these (while
1696 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1698 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1700 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1701 all platforms. Move ssize_t definition from e_os.h to the public
1702 header file e_os2.h as it now appears in public header file cms.h
1705 *) New -sigopt option to the ca, req and x509 utilities. Additional
1706 signature parameters can be passed using this option and in
1710 *) Add RSA PSS signing function. This will generate and set the
1711 appropriate AlgorithmIdentifiers for PSS based on those in the
1712 corresponding EVP_MD_CTX structure. No application support yet.
1715 *) Support for companion algorithm specific ASN1 signing routines.
1716 New function ASN1_item_sign_ctx() signs a pre-initialised
1717 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1718 the appropriate parameters.
1721 *) Add new algorithm specific ASN1 verification initialisation function
1722 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1723 handling will be the same no matter what EVP_PKEY_METHOD is used.
1724 Add a PSS handler to support verification of PSS signatures: checked
1725 against a number of sample certificates.
1728 *) Add signature printing for PSS. Add PSS OIDs.
1729 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1731 *) Add algorithm specific signature printing. An individual ASN1 method
1732 can now print out signatures instead of the standard hex dump.
1734 More complex signatures (e.g. PSS) can print out more meaningful
1735 information. Include DSA version that prints out the signature
1739 *) Password based recipient info support for CMS library: implementing
1743 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1744 neatly separates the code into cipher and PBE sections and is required
1745 for some algorithms that split PBES2 into separate pieces (such as
1746 password based CMS).
1749 *) Session-handling fixes:
1750 - Fix handling of connections that are resuming with a session ID,
1751 but also support Session Tickets.
1752 - Fix a bug that suppressed issuing of a new ticket if the client
1753 presented a ticket with an expired session.
1754 - Try to set the ticket lifetime hint to something reasonable.
1755 - Make tickets shorter by excluding irrelevant information.
1756 - On the client side, don't ignore renewed tickets.
1757 [Adam Langley, Bodo Moeller (Google)]
1759 *) Fix PSK session representation.
1762 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1764 This work was sponsored by Intel.
1767 *) Add GCM support to TLS library. Some custom code is needed to split
1768 the IV between the fixed (from PRF) and explicit (from TLS record)
1769 portions. This adds all GCM ciphersuites supported by RFC5288 and
1770 RFC5289. Generalise some AES* cipherstrings to include GCM and
1771 add a special AESGCM string for GCM only.
1774 *) Expand range of ctrls for AES GCM. Permit setting invocation
1775 field on decrypt and retrieval of invocation field only on encrypt.
1778 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1779 As required by RFC5289 these ciphersuites cannot be used if for
1780 versions of TLS earlier than 1.2.
1783 *) For FIPS capable OpenSSL interpret a NULL default public key method
1784 as unset and return the appropriate default but do *not* set the default.
1785 This means we can return the appropriate method in applications that
1786 switch between FIPS and non-FIPS modes.
1789 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1790 ENGINE is used then we cannot handle that in the FIPS module so we
1791 keep original code iff non-FIPS operations are allowed.
1794 *) Add -attime option to openssl utilities.
1795 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1797 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1800 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1801 FIPS EC methods unconditionally for now.
1804 *) New build option no-ec2m to disable characteristic 2 code.
1807 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1808 all cases can be covered as some introduce binary incompatibilities.
1811 *) Redirect RSA operations to FIPS module including keygen,
1812 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1815 *) Add similar low level API blocking to ciphers.
1818 *) Low level digest APIs are not approved in FIPS mode: any attempt
1819 to use these will cause a fatal error. Applications that *really* want
1820 to use them can use the private_* version instead.
1823 *) Redirect cipher operations to FIPS module for FIPS builds.
1826 *) Redirect digest operations to FIPS module for FIPS builds.
1829 *) Update build system to add "fips" flag which will link in fipscanister.o
1830 for static and shared library builds embedding a signature if needed.
1833 *) Output TLS supported curves in preference order instead of numerical
1834 order. This is currently hardcoded for the highest order curves first.
1835 This should be configurable so applications can judge speed vs strength.
1838 *) Add TLS v1.2 server support for client authentication.
1841 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1845 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1846 FIPS modules versions.
1849 *) Add TLS v1.2 client side support for client authentication. Keep cache
1850 of handshake records longer as we don't know the hash algorithm to use
1851 until after the certificate request message is received.
1854 *) Initial TLS v1.2 client support. Add a default signature algorithms
1855 extension including all the algorithms we support. Parse new signature
1856 format in client key exchange. Relax some ECC signing restrictions for
1857 TLS v1.2 as indicated in RFC5246.
1860 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1861 to new signature format when needed using client digest preference.
1862 All server ciphersuites should now work correctly in TLS v1.2. No client
1863 support yet and no support for client certificates.
1866 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1867 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1868 ciphersuites. At present only RSA key exchange ciphersuites work with
1869 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1870 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1871 and version checking.
1874 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1875 with this defined it will not be affected by any changes to ssl internal
1876 structures. Add several utility functions to allow openssl application
1877 to work with OPENSSL_NO_SSL_INTERN defined.
1881 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1883 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1886 *) Permit abbreviated handshakes when renegotiating using the function
1887 SSL_renegotiate_abbreviated().
1888 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1890 *) Add call to ENGINE_register_all_complete() to
1891 ENGINE_load_builtin_engines(), so some implementations get used
1892 automatically instead of needing explicit application support.
1895 *) Add support for TLS key exporter as described in RFC5705.
1896 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1898 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1899 a few changes are required:
1901 Add SSL_OP_NO_TLSv1_1 flag.
1902 Add TLSv1_1 methods.
1903 Update version checking logic to handle version 1.1.
1904 Add explicit IV handling (ported from DTLS code).
1905 Add command line options to s_client/s_server.
1908 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1910 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1911 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1912 content decryption and always return the same error. Note: this attack
1913 needs on average 2^20 messages so it only affects automated senders. The
1914 old behaviour can be reenabled in the CMS code by setting the
1915 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1916 an MMA defence is not necessary.
1917 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1918 this issue. (CVE-2012-0884)
1921 *) Fix CVE-2011-4619: make sure we really are receiving a
1922 client hello before rejecting multiple SGC restarts. Thanks to
1923 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1926 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1928 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1929 Thanks to Antonio Martin, Enterprise Secure Access Research and
1930 Development, Cisco Systems, Inc. for discovering this bug and
1931 preparing a fix. (CVE-2012-0050)
1934 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1936 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1937 of the Vaudenay padding oracle attack on CBC mode encryption
1938 which enables an efficient plaintext recovery attack against
1939 the OpenSSL implementation of DTLS. Their attack exploits timing
1940 differences arising during decryption processing. A research
1941 paper describing this attack can be found at:
1942 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1943 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1944 Security Group at Royal Holloway, University of London
1945 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1946 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1947 for preparing the fix. (CVE-2011-4108)
1948 [Robin Seggelmann, Michael Tuexen]
1950 *) Clear bytes used for block padding of SSL 3.0 records.
1952 [Adam Langley (Google)]
1954 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1955 Kadianakis <desnacked@gmail.com> for discovering this issue and
1956 Adam Langley for preparing the fix. (CVE-2011-4619)
1957 [Adam Langley (Google)]
1959 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1960 [Andrey Kulikov <amdeich@gmail.com>]
1962 *) Prevent malformed RFC3779 data triggering an assertion failure.
1963 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1964 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1965 [Rob Austein <sra@hactrn.net>]
1967 *) Improved PRNG seeding for VOS.
1968 [Paul Green <Paul.Green@stratus.com>]
1970 *) Fix ssl_ciph.c set-up race.
1971 [Adam Langley (Google)]
1973 *) Fix spurious failures in ecdsatest.c.
1974 [Emilia Käsper (Google)]
1976 *) Fix the BIO_f_buffer() implementation (which was mixing different
1977 interpretations of the '..._len' fields).
1978 [Adam Langley (Google)]
1980 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1981 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1982 threads won't reuse the same blinding coefficients.
1984 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1985 lock to call BN_BLINDING_invert_ex, and avoids one use of
1986 BN_BLINDING_update for each BN_BLINDING structure (previously,
1987 the last update always remained unused).
1988 [Emilia Käsper (Google)]
1990 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1991 [Bob Buckholz (Google)]
1993 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1995 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1996 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1997 [Kaspar Brand <ossl@velox.ch>]
1999 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2000 for multi-threaded use of ECDH. (CVE-2011-3210)
2001 [Adam Langley (Google)]
2003 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2006 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2007 signature public key algorithm by using OID xref utilities instead.
2008 Before this you could only use some ECC ciphersuites with SHA1 only.
2011 *) Add protection against ECDSA timing attacks as mentioned in the paper
2012 by Billy Bob Brumley and Nicola Tuveri, see:
2014 http://eprint.iacr.org/2011/232.pdf
2016 [Billy Bob Brumley and Nicola Tuveri]
2018 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2020 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2021 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2023 *) Fix bug in string printing code: if *any* escaping is enabled we must
2024 escape the escape character (backslash) or the resulting string is
2028 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2030 *) Disable code workaround for ancient and obsolete Netscape browsers
2031 and servers: an attacker can use it in a ciphersuite downgrade attack.
2032 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2035 *) Fixed J-PAKE implementation error, originally discovered by
2036 Sebastien Martini, further info and confirmation from Stefan
2037 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2040 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2042 *) Fix extension code to avoid race conditions which can result in a buffer
2043 overrun vulnerability: resumed sessions must not be modified as they can
2044 be shared by multiple threads. CVE-2010-3864
2047 *) Fix WIN32 build system to correctly link an ENGINE directory into
2051 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2053 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2055 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2057 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2059 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2060 context. The operation can be customised via the ctrl mechanism in
2061 case ENGINEs want to include additional functionality.
2064 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2067 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2068 output hashes compatible with older versions of OpenSSL.
2069 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2071 *) Fix compression algorithm handling: if resuming a session use the
2072 compression algorithm of the resumed session instead of determining
2073 it from client hello again. Don't allow server to change algorithm.
2076 *) Add load_crls() function to apps tidying load_certs() too. Add option
2077 to verify utility to allow additional CRLs to be included.
2080 *) Update OCSP request code to permit adding custom headers to the request:
2081 some responders need this.
2084 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2086 [Julia Lawall <julia@diku.dk>]
2088 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2089 needlessly dereferenced structures, used obsolete functions and
2090 didn't handle all updated verify codes correctly.
2093 *) Disable MD2 in the default configuration.
2096 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2097 indicate the initial BIO being pushed or popped. This makes it possible
2098 to determine whether the BIO is the one explicitly called or as a result
2099 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2100 it handles reference counts correctly and doesn't zero out the I/O bio
2101 when it is not being explicitly popped. WARNING: applications which
2102 included workarounds for the old buggy behaviour will need to be modified
2103 or they could free up already freed BIOs.
2106 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2107 renaming to all platforms (within the 0.9.8 branch, this was
2108 done conditionally on Netware platforms to avoid a name clash).
2109 [Guenter <lists@gknw.net>]
2111 *) Add ECDHE and PSK support to DTLS.
2112 [Michael Tuexen <tuexen@fh-muenster.de>]
2114 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2118 *) Add "missing" function EVP_MD_flags() (without this the only way to
2119 retrieve a digest flags is by accessing the structure directly. Update
2120 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2121 or cipher is registered as in the "from" argument. Print out all
2122 registered digests in the dgst usage message instead of manually
2123 attempting to work them out.
2126 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2127 this allows the use of compression and extensions. Change default cipher
2128 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2129 by default unless an application cipher string requests it.
2132 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2133 key ids to find matching certificates and keys but some PKCS#12 files
2134 don't follow the (somewhat unwritten) rules and this strategy fails.
2135 Now just gather all certificates together and the first private key
2136 then look for the first certificate that matches the key.
2139 *) Support use of registered digest and cipher names for dgst and cipher
2140 commands instead of having to add each one as a special case. So now
2147 openssl dgst -sha256 foo
2149 and this works for ENGINE based algorithms too.
2153 *) Update Gost ENGINE to support parameter files.
2154 [Victor B. Wagner <vitus@cryptocom.ru>]
2156 *) Support GeneralizedTime in ca utility.
2157 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2159 *) Enhance the hash format used for certificate directory links. The new
2160 form uses the canonical encoding (meaning equivalent names will work
2161 even if they aren't identical) and uses SHA1 instead of MD5. This form
2162 is incompatible with the older format and as a result c_rehash should
2163 be used to rebuild symbolic links.
2166 *) Make PKCS#8 the default write format for private keys, replacing the
2167 traditional format. This form is standardised, more secure and doesn't
2168 include an implicit MD5 dependency.
2171 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2172 committed to OpenSSL should pass this lot as a minimum.
2175 *) Add session ticket override functionality for use by EAP-FAST.
2176 [Jouni Malinen <j@w1.fi>]
2178 *) Modify HMAC functions to return a value. Since these can be implemented
2179 in an ENGINE errors can occur.
2182 *) Type-checked OBJ_bsearch_ex.
2185 *) Type-checked OBJ_bsearch. Also some constification necessitated
2186 by type-checking. Still to come: TXT_DB, bsearch(?),
2187 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2191 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2192 seconds to a tm structure directly, instead of going through OS
2193 specific date routines. This avoids any issues with OS routines such
2194 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2195 and X509_time_adj_ex() to cover the extended range. The existing
2196 X509_time_adj() is still usable and will no longer have any date issues.
2199 *) Delta CRL support. New use deltas option which will attempt to locate
2200 and search any appropriate delta CRLs available.
2202 This work was sponsored by Google.
2205 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2206 code and add additional score elements. Validate alternate CRL paths
2207 as part of the CRL checking and indicate a new error "CRL path validation
2208 error" in this case. Applications wanting additional details can use
2209 the verify callback and check the new "parent" field. If this is not
2210 NULL CRL path validation is taking place. Existing applications wont
2211 see this because it requires extended CRL support which is off by
2214 This work was sponsored by Google.
2217 *) Support for freshest CRL extension.
2219 This work was sponsored by Google.
2222 *) Initial indirect CRL support. Currently only supported in the CRLs
2223 passed directly and not via lookup. Process certificate issuer
2224 CRL entry extension and lookup CRL entries by bother issuer name
2225 and serial number. Check and process CRL issuer entry in IDP extension.
2227 This work was sponsored by Google.
2230 *) Add support for distinct certificate and CRL paths. The CRL issuer
2231 certificate is validated separately in this case. Only enabled if
2232 an extended CRL support flag is set: this flag will enable additional
2233 CRL functionality in future.
2235 This work was sponsored by Google.
2238 *) Add support for policy mappings extension.
2240 This work was sponsored by Google.
2243 *) Fixes to pathlength constraint, self issued certificate handling,
2244 policy processing to align with RFC3280 and PKITS tests.
2246 This work was sponsored by Google.
2249 *) Support for name constraints certificate extension. DN, email, DNS
2250 and URI types are currently supported.
2252 This work was sponsored by Google.
2255 *) To cater for systems that provide a pointer-based thread ID rather
2256 than numeric, deprecate the current numeric thread ID mechanism and
2257 replace it with a structure and associated callback type. This
2258 mechanism allows a numeric "hash" to be extracted from a thread ID in
2259 either case, and on platforms where pointers are larger than 'long',
2260 mixing is done to help ensure the numeric 'hash' is usable even if it
2261 can't be guaranteed unique. The default mechanism is to use "&errno"
2262 as a pointer-based thread ID to distinguish between threads.
2264 Applications that want to provide their own thread IDs should now use
2265 CRYPTO_THREADID_set_callback() to register a callback that will call
2266 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2268 Note that ERR_remove_state() is now deprecated, because it is tied
2269 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2270 to free the current thread's error state should be replaced by
2271 ERR_remove_thread_state(NULL).
2273 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2274 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2275 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2276 application was previously providing a numeric thread callback that
2277 was inappropriate for distinguishing threads, then uniqueness might
2278 have been obtained with &errno that happened immediately in the
2279 intermediate development versions of OpenSSL; this is no longer the
2280 case, the numeric thread callback will now override the automatic use
2282 [Geoff Thorpe, with help from Bodo Moeller]
2284 *) Initial support for different CRL issuing certificates. This covers a
2285 simple case where the self issued certificates in the chain exist and
2286 the real CRL issuer is higher in the existing chain.
2288 This work was sponsored by Google.
2291 *) Removed effectively defunct crypto/store from the build.
2294 *) Revamp of STACK to provide stronger type-checking. Still to come:
2295 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2296 ASN1_STRING, CONF_VALUE.
2299 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2300 RAM on SSL connections. This option can save about 34k per idle SSL.
2303 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2304 STACK, TXT_DB, bsearch, qsort.
2307 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2308 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2309 support for data, signedData, compressedData, digestedData and
2310 encryptedData, envelopedData types included. Scripts to check against
2311 RFC4134 examples draft and interop and consistency checks of many
2312 content types and variants.
2315 *) Add options to enc utility to support use of zlib compression BIO.
2318 *) Extend mk1mf to support importing of options and assembly language
2319 files from Configure script, currently only included in VC-WIN32.
2320 The assembly language rules can now optionally generate the source
2321 files from the associated perl scripts.
2324 *) Implement remaining functionality needed to support GOST ciphersuites.
2325 Interop testing has been performed using CryptoPro implementations.
2326 [Victor B. Wagner <vitus@cryptocom.ru>]
2328 *) s390x assembler pack.
2331 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2335 *) Implement Opaque PRF Input TLS extension as specified in
2336 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2337 official specification yet and no extension type assignment by
2338 IANA exists, this extension (for now) will have to be explicitly
2339 enabled when building OpenSSL by providing the extension number
2340 to use. For example, specify an option
2342 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2344 to the "config" or "Configure" script to enable the extension,
2345 assuming extension number 0x9527 (which is a completely arbitrary
2346 and unofficial assignment based on the MD5 hash of the Internet
2347 Draft). Note that by doing so, you potentially lose
2348 interoperability with other TLS implementations since these might
2349 be using the same extension number for other purposes.
2351 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2352 opaque PRF input value to use in the handshake. This will create
2353 an interal copy of the length-'len' string at 'src', and will
2354 return non-zero for success.
2356 To get more control and flexibility, provide a callback function
2359 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2360 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2364 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2367 Callback function 'cb' will be called in handshakes, and is
2368 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2369 Argument 'arg' is for application purposes (the value as given to
2370 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2371 be provided to the callback function). The callback function
2372 has to return non-zero to report success: usually 1 to use opaque
2373 PRF input just if possible, or 2 to enforce use of the opaque PRF
2374 input. In the latter case, the library will abort the handshake
2375 if opaque PRF input is not successfully negotiated.
2377 Arguments 'peerinput' and 'len' given to the callback function
2378 will always be NULL and 0 in the case of a client. A server will
2379 see the client's opaque PRF input through these variables if
2380 available (NULL and 0 otherwise). Note that if the server
2381 provides an opaque PRF input, the length must be the same as the
2382 length of the client's opaque PRF input.
2384 Note that the callback function will only be called when creating
2385 a new session (session resumption can resume whatever was
2386 previously negotiated), and will not be called in SSL 2.0
2387 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2388 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2389 for applications that need to enforce opaque PRF input.
2393 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2396 [Victor B. Wagner <vitus@cryptocom.ru>]
2398 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2399 RFC4507bis. The encrypted ticket format is an encrypted encoded
2400 SSL_SESSION structure, that way new session features are automatically
2403 If a client application caches session in an SSL_SESSION structure
2404 support is transparent because tickets are now stored in the encoded
2407 The SSL_CTX structure automatically generates keys for ticket
2408 protection in servers so again support should be possible
2409 with no application modification.
2411 If a client or server wishes to disable RFC4507 support then the option
2412 SSL_OP_NO_TICKET can be set.
2414 Add a TLS extension debugging callback to allow the contents of any client
2415 or server extensions to be examined.
2417 This work was sponsored by Google.
2420 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2421 OpenSSL should now compile cleanly on gcc 4.2
2422 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2424 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2425 support including streaming MAC support: this is required for GOST
2426 ciphersuite support.
2427 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2429 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2430 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2431 to output in BER and PEM format.
2434 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2435 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2436 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2437 ENGINE support for HMAC keys which are unextractable. New -mac and
2438 -macopt options to dgst utility.
2441 *) New option -sigopt to dgst utility. Update dgst to use
2442 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2443 alternative signing paramaters such as X9.31 or PSS in the dgst
2447 *) Change ssl_cipher_apply_rule(), the internal function that does
2448 the work each time a ciphersuite string requests enabling
2449 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2450 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2451 the order of disabled ciphersuites such that those ciphersuites
2452 that most recently went from enabled to disabled not only stay
2453 in order with respect to each other, but also have higher priority
2454 than other disabled ciphersuites the next time ciphersuites are
2457 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2458 the same ciphersuites as with "HIGH" alone, but in a specific
2459 order where the PSK ciphersuites come first (since they are the
2460 most recently disabled ciphersuites when "HIGH" is parsed).
2462 Also, change ssl_create_cipher_list() (using this new
2463 funcionality) such that between otherwise identical
2464 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2468 *) Change ssl_create_cipher_list() so that it automatically
2469 arranges the ciphersuites in reasonable order before starting
2470 to process the rule string. Thus, the definition for "DEFAULT"
2471 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2472 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2473 This makes it much easier to arrive at a reasonable default order
2474 in applications for which anonymous ciphers are OK (meaning
2475 that you can't actually use DEFAULT).
2476 [Bodo Moeller; suggested by Victor Duchovni]
2478 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2479 processing) into multiple integers instead of setting
2480 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2481 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2482 (These masks as well as the individual bit definitions are hidden
2483 away into the non-exported interface ssl/ssl_locl.h, so this
2484 change to the definition of the SSL_CIPHER structure shouldn't
2485 affect applications.) This give us more bits for each of these
2486 categories, so there is no longer a need to coagulate AES128 and
2487 AES256 into a single algorithm bit, and to coagulate Camellia128
2488 and Camellia256 into a single algorithm bit, which has led to all
2491 Thus, among other things, the kludge introduced in 0.9.7m and
2492 0.9.8e for masking out AES256 independently of AES128 or masking
2493 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2495 With the change, we also introduce new ciphersuite aliases that
2496 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2500 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2501 Use the leftmost N bytes of the signature input if the input is
2502 larger than the prime q (with N being the size in bytes of q).
2505 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2506 it yet and it is largely untested.
2509 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2512 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2513 some compilers (gcc 4.2 and later) reject their use. Safestack is
2514 reimplemented. Update ASN1 to avoid use of legacy functions.
2517 *) Win32/64 targets are linked with Winsock2.
2520 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2521 to external functions. This can be used to increase CRL handling
2522 efficiency especially when CRLs are very large by (for example) storing
2523 the CRL revoked certificates in a database.
2526 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2527 new CRLs added to a directory can be used. New command line option
2528 -verify_return_error to s_client and s_server. This causes real errors
2529 to be returned by the verify callback instead of carrying on no matter
2530 what. This reflects the way a "real world" verify callback would behave.
2533 *) GOST engine, supporting several GOST algorithms and public key formats.
2534 Kindly donated by Cryptocom.
2537 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2538 partitioned by DP are handled but no indirect CRL or reason partitioning
2539 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2540 selected via a scoring technique which handles IDP and AKID in CRLs.
2543 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2544 will ultimately be used for all verify operations: this will remove the
2545 X509_STORE dependency on certificate verification and allow alternative
2546 lookup methods. X509_STORE based implementations of these two callbacks.
2549 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2550 Modify get_crl() to find a valid (unexpired) CRL if possible.
2553 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2554 this would be called X509_CRL_cmp() but that name is already used by
2555 a function that just compares CRL issuer names. Cache several CRL
2556 extensions in X509_CRL structure and cache CRLDP in X509.
2559 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2560 this maps equivalent X509_NAME structures into a consistent structure.
2561 Name comparison can then be performed rapidly using memcmp().
2564 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2568 *) Allow digests to supply their own micalg string for S/MIME type using
2569 the ctrl EVP_MD_CTRL_MICALG.
2572 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2573 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2574 ctrl. It can then customise the structure before and/or after signing
2578 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2579 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2580 to free up any added signature OIDs.
2583 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2584 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2585 digest and cipher tables. New options added to openssl utility:
2586 list-message-digest-algorithms and list-cipher-algorithms.
2589 *) Change the array representation of binary polynomials: the list
2590 of degrees of non-zero coefficients is now terminated with -1.
2591 Previously it was terminated with 0, which was also part of the
2592 value; thus, the array representation was not applicable to
2593 polynomials where t^0 has coefficient zero. This change makes
2594 the array representation useful in a more general context.
2597 *) Various modifications and fixes to SSL/TLS cipher string
2598 handling. For ECC, the code now distinguishes between fixed ECDH
2599 with RSA certificates on the one hand and with ECDSA certificates
2600 on the other hand, since these are separate ciphersuites. The
2601 unused code for Fortezza ciphersuites has been removed.
2603 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2604 (not "ECDHE"). For consistency with the code for DH
2605 certificates, use of ECDH certificates is now considered ECDH
2606 authentication, not RSA or ECDSA authentication (the latter is
2607 merely the CA's signing algorithm and not actively used in the
2610 The temporary ciphersuite alias "ECCdraft" is no longer
2611 available, and ECC ciphersuites are no longer excluded from "ALL"
2612 and "DEFAULT". The following aliases now exist for RFC 4492
2613 ciphersuites, most of these by analogy with the DH case:
2615 kECDHr - ECDH cert, signed with RSA
2616 kECDHe - ECDH cert, signed with ECDSA
2617 kECDH - ECDH cert (signed with either RSA or ECDSA)
2618 kEECDH - ephemeral ECDH
2619 ECDH - ECDH cert or ephemeral ECDH
2625 AECDH - anonymous ECDH
2626 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2630 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2631 Use correct micalg parameters depending on digest(s) in signed message.
2634 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2635 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2638 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2639 an engine to register a method. Add ENGINE lookups for methods and
2640 functional reference processing.
2643 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2644 EVP_{Sign,Verify}* which allow an application to customise the signature
2648 *) New -resign option to smime utility. This adds one or more signers
2649 to an existing PKCS#7 signedData structure. Also -md option to use an
2650 alternative message digest algorithm for signing.
2653 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2654 create PKCS7 structures containing multiple signers. Update smime
2655 application to support multiple signers.
2658 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2662 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2663 Reorganize PBE internals to lookup from a static table using NIDs,
2664 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2665 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2666 PRF which will be automatically used with PBES2.
2669 *) Replace the algorithm specific calls to generate keys in "req" with the
2673 *) Update PKCS#7 enveloped data routines to use new API. This is now
2674 supported by any public key method supporting the encrypt operation. A
2675 ctrl is added to allow the public key algorithm to examine or modify
2676 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2680 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2681 a default digest type to use. In most cases this will be SHA1 but some
2682 algorithms (such as GOST) need to specify an alternative digest. The
2683 return value indicates how strong the preference is 1 means optional and
2684 2 is mandatory (that is it is the only supported type). Modify
2685 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2686 use the default md. Update openssl utilities to use the default digest
2687 type for signing if it is not explicitly indicated.
2690 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2691 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2692 signing method from the key type. This effectively removes the link
2693 between digests and public key types.
2696 *) Add an OID cross reference table and utility functions. Its purpose is to
2697 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2698 rsaEncryption. This will allow some of the algorithm specific hackery
2699 needed to use the correct OID to be removed.
2702 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2703 structures for PKCS7_sign(). They are now set up by the relevant public
2707 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2710 *) Add support for key derivation (agreement) in the API, DH method and
2714 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2715 public and private key formats. As a side effect these add additional
2716 command line functionality not previously available: DSA signatures can be
2717 generated and verified using pkeyutl and DH key support and generation in
2722 [Oliver Tappe <zooey@hirschkaefer.de>]
2724 *) New make target "install_html_docs" installs HTML renditions of the
2726 [Oliver Tappe <zooey@hirschkaefer.de>]
2728 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2729 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2730 support key and parameter generation and add initial key generation
2731 functionality for RSA.
2734 *) Add functions for main EVP_PKEY_method operations. The undocumented
2735 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2736 EVP_PKEY_{encrypt,decrypt}_old.
2739 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2740 key API, doesn't do much yet.
2743 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2744 public key algorithms. New option to openssl utility:
2745 "list-public-key-algorithms" to print out info.
2748 *) Implement the Supported Elliptic Curves Extension for
2749 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2752 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2753 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2756 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2757 utilities such as rsa, dsa, dsaparam etc except they process any key
2761 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2762 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2763 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2767 *) Initial support for pluggable public key ASN1.
2768 De-spaghettify the public key ASN1 handling. Move public and private
2769 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2770 algorithm specific handling to a single module within the relevant
2771 algorithm directory. Add functions to allow (near) opaque processing
2772 of public and private key structures.
2775 *) Implement the Supported Point Formats Extension for
2776 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2779 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2780 for the psk identity [hint] and the psk callback functions to the
2781 SSL_SESSION, SSL and SSL_CTX structure.
2784 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2788 SSL_CTX_use_psk_identity_hint
2789 SSL_get_psk_identity_hint
2790 SSL_get_psk_identity
2791 SSL_use_psk_identity_hint
2793 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2795 *) Add RFC 3161 compliant time stamp request creation, response generation
2796 and response verification functionality.
2797 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2799 *) Add initial support for TLS extensions, specifically for the server_name
2800 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2801 have new members for a host name. The SSL data structure has an
2802 additional member SSL_CTX *initial_ctx so that new sessions can be
2803 stored in that context to allow for session resumption, even after the
2804 SSL has been switched to a new SSL_CTX in reaction to a client's
2805 server_name extension.
2807 New functions (subject to change):
2809 SSL_get_servername()
2810 SSL_get_servername_type()
2813 New CTRL codes and macros (subject to change):
2815 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2816 - SSL_CTX_set_tlsext_servername_callback()
2817 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2818 - SSL_CTX_set_tlsext_servername_arg()
2819 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2821 openssl s_client has a new '-servername ...' option.
2823 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2824 '-key2 ...', '-servername_fatal' (subject to change). This allows
2825 testing the HostName extension for a specific single host name ('-cert'
2826 and '-key' remain fallbacks for handshakes without HostName
2827 negotiation). If the unrecognized_name alert has to be sent, this by
2828 default is a warning; it becomes fatal with the '-servername_fatal'
2831 [Peter Sylvester, Remy Allais, Christophe Renou]
2833 *) Whirlpool hash implementation is added.
2836 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2837 bn(64,32). Because of instruction set limitations it doesn't have
2838 any negative impact on performance. This was done mostly in order
2839 to make it possible to share assembler modules, such as bn_mul_mont
2840 implementations, between 32- and 64-bit builds without hassle.
2843 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2844 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2848 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2849 dedicated Montgomery multiplication procedure, is introduced.
2850 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2851 "64-bit" performance on certain 32-bit targets.
2854 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2855 in SSL structures. New SSL ctrl to set maximum send fragment size.
2856 Save memory by seeting the I/O buffer sizes dynamically instead of
2857 using the maximum available value.
2860 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2861 in addition to the text details.
2864 *) Very, very preliminary EXPERIMENTAL support for printing of general
2865 ASN1 structures. This currently produces rather ugly output and doesn't
2866 handle several customised structures at all.
2869 *) Integrated support for PVK file format and some related formats such
2870 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2871 these in the 'rsa' and 'dsa' utilities.
2874 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2877 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2878 place for the (very old) "NETSCAPE" format certificates which are now
2879 handled using new ASN1 code equivalents.
2882 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2883 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2884 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2887 *) Modify CRL distribution points extension code to print out previously
2888 unsupported fields. Enhance extension setting code to allow setting of
2892 *) Add print and set support for Issuing Distribution Point CRL extension.
2895 *) Change 'Configure' script to enable Camellia by default.
2898 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2900 *) When rejecting SSL/TLS records due to an incorrect version number, never
2901 update s->server with a new major version number. As of
2902 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2903 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2904 the previous behavior could result in a read attempt at NULL when
2905 receiving specific incorrect SSL/TLS records once record payload
2906 protection is active. (CVE-2010-0740)
2907 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2909 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2910 could be crashed if the relevant tables were not present (e.g. chrooted).
2911 [Tomas Hoger <thoger@redhat.com>]
2913 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2915 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2916 [Martin Olsson, Neel Mehta]
2918 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2919 accommodate for stack sorting, always a write lock!).
2922 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2923 excessive delays in the RAND_poll(): over a minute. As a workaround
2924 include a time check in the inner Heap32Next loop too.
2927 *) The code that handled flushing of data in SSL/TLS originally used the
2928 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2929 the problem outlined in PR#1949. The fix suggested there however can
2930 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2931 of Apache). So instead simplify the code to flush unconditionally.
2932 This should be fine since flushing with no data to flush is a no op.
2935 *) Handle TLS versions 2.0 and later properly and correctly use the
2936 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2937 off ancient servers have a habit of sticking around for a while...
2940 *) Modify compression code so it frees up structures without using the
2941 ex_data callbacks. This works around a problem where some applications
2942 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2943 restarting) then use compression (e.g. SSL with compression) later.
2944 This results in significant per-connection memory leaks and
2945 has caused some security issues including CVE-2008-1678 and
2949 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2950 change when encrypting or decrypting.
2953 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2954 connect and renegotiate with servers which do not support RI.
2955 Until RI is more widely deployed this option is enabled by default.
2958 *) Add "missing" ssl ctrls to clear options and mode.
2961 *) If client attempts to renegotiate and doesn't support RI respond with
2962 a no_renegotiation alert as required by RFC5746. Some renegotiating
2963 TLS clients will continue a connection gracefully when they receive
2964 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2965 waiting for a server hello which it will never receive. Now we treat a
2966 received no_renegotiation alert as a fatal error. This is because
2967 applications requesting a renegotiation might well expect it to succeed
2968 and would have no code in place to handle the server denying it so the
2969 only safe thing to do is to terminate the connection.
2972 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2973 peer supports secure renegotiation and 0 otherwise. Print out peer
2974 renegotiation support in s_client/s_server.
2977 *) Replace the highly broken and deprecated SPKAC certification method with
2978 the updated NID creation version. This should correctly handle UTF8.
2981 *) Implement RFC5746. Re-enable renegotiation but require the extension
2982 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2983 turns out to be a bad idea. It has been replaced by
2984 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2985 SSL_CTX_set_options(). This is really not recommended unless you
2986 know what you are doing.
2987 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2989 *) Fixes to stateless session resumption handling. Use initial_ctx when
2990 issuing and attempting to decrypt tickets in case it has changed during
2991 servername handling. Use a non-zero length session ID when attempting
2992 stateless session resumption: this makes it possible to determine if
2993 a resumption has occurred immediately after receiving server hello
2994 (several places in OpenSSL subtly assume this) instead of later in
2998 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2999 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3000 fixes for a few places where the return code is not checked
3002 [Julia Lawall <julia@diku.dk>]
3004 *) Add --strict-warnings option to Configure script to include devteam
3005 warnings in other configurations.
3008 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3009 makes it possible to install openssl libraries in locations which
3010 have names other than "lib", for example "/usr/lib64" which some
3012 [Steve Henson, based on patch from Jeremy Utley]
3014 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3015 X690 8.9.12 and can produce some misleading textual output of OIDs.
3016 [Steve Henson, reported by Dan Kaminsky]
3018 *) Delete MD2 from algorithm tables. This follows the recommendation in
3019 several standards that it is not used in new applications due to
3020 several cryptographic weaknesses. For binary compatibility reasons
3021 the MD2 API is still compiled in by default.
3024 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3028 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3029 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3031 [Guenter <lists@gknw.net>]
3033 *) Fix the server certificate chain building code to use X509_verify_cert(),
3034 it used to have an ad-hoc builder which was unable to cope with anything
3035 other than a simple chain.
3036 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3038 *) Don't check self signed certificate signatures in X509_verify_cert()
3039 by default (a flag can override this): it just wastes time without
3040 adding any security. As a useful side effect self signed root CAs
3041 with non-FIPS digests are now usable in FIPS mode.
3044 *) In dtls1_process_out_of_seq_message() the check if the current message
3045 is already buffered was missing. For every new message was memory
3046 allocated, allowing an attacker to perform an denial of service attack
3047 with sending out of seq handshake messages until there is no memory
3048 left. Additionally every future messege was buffered, even if the
3049 sequence number made no sense and would be part of another handshake.
3050 So only messages with sequence numbers less than 10 in advance will be
3051 buffered. (CVE-2009-1378)
3052 [Robin Seggelmann, discovered by Daniel Mentz]
3054 *) Records are buffered if they arrive with a future epoch to be
3055 processed after finishing the corresponding handshake. There is
3056 currently no limitation to this buffer allowing an attacker to perform
3057 a DOS attack with sending records with future epochs until there is no
3058 memory left. This patch adds the pqueue_size() function to determine
3059 the size of a buffer and limits the record buffer to 100 entries.
3061 [Robin Seggelmann, discovered by Daniel Mentz]
3063 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3064 parent structure is freed. (CVE-2009-1379)
3067 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3068 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3071 [Ilya O. <vrghost@gmail.com>]
3073 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3075 *) Disable renegotiation completely - this fixes a severe security
3076 problem (CVE-2009-3555) at the cost of breaking all
3077 renegotiation. Renegotiation can be re-enabled by setting
3078 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3079 run-time. This is really not recommended unless you know what
3083 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3085 *) Don't set val to NULL when freeing up structures, it is freed up by
3086 underlying code. If sizeof(void *) > sizeof(long) this can result in
3087 zeroing past the valid field. (CVE-2009-0789)
3088 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3090 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3091 checked correctly. This would allow some invalid signed attributes to
3092 appear to verify correctly. (CVE-2009-0591)
3093 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3095 *) Reject UniversalString and BMPString types with invalid lengths. This
3096 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3097 a legal length. (CVE-2009-0590)
3100 *) Set S/MIME signing as the default purpose rather than setting it
3101 unconditionally. This allows applications to override it at the store
3105 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3106 to handle some structures.
3109 *) Improve efficiency of mem_gets: don't search whole buffer each time
3111 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3113 *) New -hex option for openssl rand.
3116 *) Print out UTF8String and NumericString when parsing ASN1.
3119 *) Support NumericString type for name components.
3122 *) Allow CC in the environment to override the automatically chosen
3123 compiler. Note that nothing is done to ensure flags work with the
3127 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3129 *) Properly check EVP_VerifyFinal() and similar return values
3131 [Ben Laurie, Bodo Moeller, Google Security Team]
3133 *) Enable TLS extensions by default.
3136 *) Allow the CHIL engine to be loaded, whether the application is
3137 multithreaded or not. (This does not release the developer from the
3138 obligation to set up the dynamic locking callbacks.)
3139 [Sander Temme <sander@temme.net>]
3141 *) Use correct exit code if there is an error in dgst command.
3142 [Steve Henson; problem pointed out by Roland Dirlewanger]
3144 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3145 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3148 *) Add experimental JPAKE support, including demo authentication in
3149 s_client and s_server.
3152 *) Set the comparison function in v3_addr_canonize().
3153 [Rob Austein <sra@hactrn.net>]
3155 *) Add support for XMPP STARTTLS in s_client.
3156 [Philip Paeps <philip@freebsd.org>]
3158 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3159 to ensure that even with this option, only ciphersuites in the
3160 server's preference list will be accepted. (Note that the option
3161 applies only when resuming a session, so the earlier behavior was
3162 just about the algorithm choice for symmetric cryptography.)
3165 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3167 *) Fix NULL pointer dereference if a DTLS server received
3168 ChangeCipherSpec as first record (CVE-2009-1386).
3171 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3172 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3175 *) The fix in 0.9.8c that supposedly got rid of unsafe
3176 double-checked locking was incomplete for RSA blinding,
3177 addressing just one layer of what turns out to have been
3178 doubly unsafe triple-checked locking.
3180 So now fix this for real by retiring the MONT_HELPER macro
3181 in crypto/rsa/rsa_eay.c.
3183 [Bodo Moeller; problem pointed out by Marius Schilder]
3185 *) Various precautionary measures:
3187 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3189 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3190 (NB: This would require knowledge of the secret session ticket key
3191 to exploit, in which case you'd be SOL either way.)
3193 - Change bn_nist.c so that it will properly handle input BIGNUMs
3194 outside the expected range.
3196 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3199 [Neel Mehta, Bodo Moeller]
3201 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3202 the load fails. Useful for distros.
3203 [Ben Laurie and the FreeBSD team]
3205 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3208 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3211 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3213 This work was sponsored by Logica.
3216 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3217 keystores. Support for SSL/TLS client authentication too.
3218 Not compiled unless enable-capieng specified to Configure.
3220 This work was sponsored by Logica.
3223 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3224 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3225 attribute creation routines such as certifcate requests and PKCS#12
3229 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3231 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3232 handshake which could lead to a cilent crash as found using the
3233 Codenomicon TLS test suite (CVE-2008-1672)
3234 [Steve Henson, Mark Cox]
3236 *) Fix double free in TLS server name extensions which could lead to
3237 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3240 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3242 Clear the error queue to ensure that error entries left from
3243 older function calls do not interfere with the correct operation.
3244 [Lutz Jaenicke, Erik de Castro Lopo]
3246 *) Remove root CA certificates of commercial CAs:
3248 The OpenSSL project does not recommend any specific CA and does not
3249 have any policy with respect to including or excluding any CA.
3250 Therefore it does not make any sense to ship an arbitrary selection
3251 of root CA certificates with the OpenSSL software.
3254 *) RSA OAEP patches to fix two separate invalid memory reads.
3255 The first one involves inputs when 'lzero' is greater than
3256 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3257 before the beginning of from). The second one involves inputs where
3258 the 'db' section contains nothing but zeroes (there is a one-byte
3259 invalid read after the end of 'db').
3260 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3262 *) Partial backport from 0.9.9-dev:
3264 Introduce bn_mul_mont (dedicated Montgomery multiplication
3265 procedure) as a candidate for BIGNUM assembler implementation.
3266 While 0.9.9-dev uses assembler for various architectures, only
3267 x86_64 is available by default here in the 0.9.8 branch, and
3268 32-bit x86 is available through a compile-time setting.
3270 To try the 32-bit x86 assembler implementation, use Configure
3271 option "enable-montasm" (which exists only for this backport).
3273 As "enable-montasm" for 32-bit x86 disclaims code stability
3274 anyway, in this constellation we activate additional code
3275 backported from 0.9.9-dev for further performance improvements,
3276 namely BN_from_montgomery_word. (To enable this otherwise,
3277 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3279 [Andy Polyakov (backport partially by Bodo Moeller)]
3281 *) Add TLS session ticket callback. This allows an application to set
3282 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3283 values. This is useful for key rollover for example where several key
3284 sets may exist with different names.
3287 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3288 This was broken until now in 0.9.8 releases, such that the only way
3289 a registered ENGINE could be used (assuming it initialises
3290 successfully on the host) was to explicitly set it as the default
3291 for the relevant algorithms. This is in contradiction with 0.9.7
3292 behaviour and the documentation. With this fix, when an ENGINE is
3293 registered into a given algorithm's table of implementations, the
3294 'uptodate' flag is reset so that auto-discovery will be used next
3295 time a new context for that algorithm attempts to select an
3297 [Ian Lister (tweaked by Geoff Thorpe)]
3299 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3300 implemention in the following ways:
3302 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3305 Lack of BER streaming support means one pass streaming processing is
3306 only supported if data is detached: setting the streaming flag is
3307 ignored for embedded content.
3309 CMS support is disabled by default and must be explicitly enabled
3310 with the enable-cms configuration option.
3313 *) Update the GMP engine glue to do direct copies between BIGNUM and
3314 mpz_t when openssl and GMP use the same limb size. Otherwise the
3315 existing "conversion via a text string export" trick is still used.
3316 [Paul Sheer <paulsheer@gmail.com>]
3318 *) Zlib compression BIO. This is a filter BIO which compressed and
3319 uncompresses any data passed through it.
3322 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3323 RFC3394 compatible AES key wrapping.
3326 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3327 sets string data without copying. X509_ALGOR_set0() and
3328 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3329 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3330 from an X509_ATTRIBUTE structure optionally checking it occurs only
3331 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3335 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3336 to get the expected BN_FLG_CONSTTIME behavior.
3337 [Bodo Moeller (Google)]
3341 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3342 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3343 - added some more tests to do_tests.pl
3344 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3345 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3346 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3347 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3348 - various changes to netware.pl to enable gcc-cross builds on Win32
3350 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3351 - various changes to fix missing prototype warnings
3352 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3353 - added AES, WHIRLPOOL and CPUID assembler code to build files
3354 - added missing AES assembler make rules to mk1mf.pl
3355 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3356 [Guenter Knauf <eflash@gmx.net>]
3358 *) Implement certificate status request TLS extension defined in RFC3546.
3359 A client can set the appropriate parameters and receive the encoded
3360 OCSP response via a callback. A server can query the supplied parameters
3361 and set the encoded OCSP response in the callback. Add simplified examples
3362 to s_client and s_server.
3365 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3367 *) Fix various bugs:
3368 + Binary incompatibility of ssl_ctx_st structure
3369 + DTLS interoperation with non-compliant servers
3370 + Don't call get_session_cb() without proposed session
3371 + Fix ia64 assembler code
3372 [Andy Polyakov, Steve Henson]
3374 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3376 *) DTLS Handshake overhaul. There were longstanding issues with
3377 OpenSSL DTLS implementation, which were making it impossible for
3378 RFC 4347 compliant client to communicate with OpenSSL server.
3379 Unfortunately just fixing these incompatibilities would "cut off"
3380 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3381 server keeps tolerating non RFC compliant syntax. The opposite is
3382 not true, 0.9.8f client can not communicate with earlier server.
3383 This update even addresses CVE-2007-4995.
3386 *) Changes to avoid need for function casts in OpenSSL: some compilers
3387 (gcc 4.2 and later) reject their use.
3388 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3391 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3392 RFC4507bis. The encrypted ticket format is an encrypted encoded
3393 SSL_SESSION structure, that way new session features are automatically
3396 If a client application caches session in an SSL_SESSION structure
3397 support is transparent because tickets are now stored in the encoded
3400 The SSL_CTX structure automatically generates keys for ticket
3401 protection in servers so again support should be possible
3402 with no application modification.
3404 If a client or server wishes to disable RFC4507 support then the option
3405 SSL_OP_NO_TICKET can be set.
3407 Add a TLS extension debugging callback to allow the contents of any client
3408 or server extensions to be examined.
3410 This work was sponsored by Google.
3413 *) Add initial support for TLS extensions, specifically for the server_name
3414 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3415 have new members for a host name. The SSL data structure has an
3416 additional member SSL_CTX *initial_ctx so that new sessions can be
3417 stored in that context to allow for session resumption, even after the
3418 SSL has been switched to a new SSL_CTX in reaction to a client's
3419 server_name extension.
3421 New functions (subject to change):
3423 SSL_get_servername()
3424 SSL_get_servername_type()
3427 New CTRL codes and macros (subject to change):
3429 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3430 - SSL_CTX_set_tlsext_servername_callback()
3431 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3432 - SSL_CTX_set_tlsext_servername_arg()
3433 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3435 openssl s_client has a new '-servername ...' option.
3437 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3438 '-key2 ...', '-servername_fatal' (subject to change). This allows
3439 testing the HostName extension for a specific single host name ('-cert'
3440 and '-key' remain fallbacks for handshakes without HostName
3441 negotiation). If the unrecognized_name alert has to be sent, this by
3442 default is a warning; it becomes fatal with the '-servername_fatal'
3445 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3447 *) Add AES and SSE2 assembly language support to VC++ build.
3450 *) Mitigate attack on final subtraction in Montgomery reduction.
3453 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3454 (which previously caused an internal error).
3457 *) Squeeze another 10% out of IGE mode when in != out.
3460 *) AES IGE mode speedup.
3461 [Dean Gaudet (Google)]
3463 *) Add the Korean symmetric 128-bit cipher SEED (see
3464 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3465 add SEED ciphersuites from RFC 4162:
3467 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3468 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3469 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3470 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3472 To minimize changes between patchlevels in the OpenSSL 0.9.8
3473 series, SEED remains excluded from compilation unless OpenSSL
3474 is configured with 'enable-seed'.
3475 [KISA, Bodo Moeller]
3477 *) Mitigate branch prediction attacks, which can be practical if a
3478 single processor is shared, allowing a spy process to extract
3479 information. For detailed background information, see
3480 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3481 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3482 and Necessary Software Countermeasures"). The core of the change
3483 are new versions BN_div_no_branch() and
3484 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3485 respectively, which are slower, but avoid the security-relevant
3486 conditional branches. These are automatically called by BN_div()
3487 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3488 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3489 remove a conditional branch.
3491 BN_FLG_CONSTTIME is the new name for the previous
3492 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3493 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3494 in the exponent causes BN_mod_exp_mont() to use the alternative
3495 implementation in BN_mod_exp_mont_consttime().) The old name
3496 remains as a deprecated alias.
3498 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3499 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3500 constant-time implementations for more than just exponentiation.
3501 Here too the old name is kept as a deprecated alias.
3503 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3504 the BN_BLINDING structure gets an independent copy of the
3505 modulus. This means that the previous "BIGNUM *m" argument to
3506 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3507 essentially becomes "const BIGNUM *m", although we can't actually
3508 change this in the header file before 0.9.9. It allows
3509 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3510 enable BN_FLG_CONSTTIME.
3512 [Matthew D Wood (Intel Corp)]
3514 *) In the SSL/TLS server implementation, be strict about session ID
3515 context matching (which matters if an application uses a single
3516 external cache for different purposes). Previously,
3517 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3518 set. This did ensure strict client verification, but meant that,
3519 with applications using a single external cache for quite
3520 different requirements, clients could circumvent ciphersuite
3521 restrictions for a given session ID context by starting a session
3522 in a different context.
3525 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3526 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3527 authentication-only ciphersuites.
3530 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3531 not complete and could lead to a possible single byte overflow
3532 (CVE-2007-5135) [Ben Laurie]
3534 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3536 *) Since AES128 and AES256 (and similarly Camellia128 and
3537 Camellia256) share a single mask bit in the logic of
3538 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3539 kludge to work properly if AES128 is available and AES256 isn't
3540 (or if Camellia128 is available and Camellia256 isn't).
3543 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3544 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3545 When a point or a seed is encoded in a BIT STRING, we need to
3546 prevent the removal of trailing zero bits to get the proper DER
3547 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3548 of a NamedBitList, for which trailing 0 bits need to be removed.)
3551 *) Have SSL/TLS server implementation tolerate "mismatched" record
3552 protocol version while receiving ClientHello even if the
3553 ClientHello is fragmented. (The server can't insist on the
3554 particular protocol version it has chosen before the ServerHello
3555 message has informed the client about his choice.)
3558 *) Add RFC 3779 support.
3559 [Rob Austein for ARIN, Ben Laurie]
3561 *) Load error codes if they are not already present instead of using a
3562 static variable. This allows them to be cleanly unloaded and reloaded.
3563 Improve header file function name parsing.
3566 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3567 or CAPABILITY handshake as required by RFCs.
3570 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3572 *) Introduce limits to prevent malicious keys being able to
3573 cause a denial of service. (CVE-2006-2940)
3574 [Steve Henson, Bodo Moeller]
3576 *) Fix ASN.1 parsing of certain invalid structures that can result
3577 in a denial of service. (CVE-2006-2937) [Steve Henson]
3579 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3580 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3582 *) Fix SSL client code which could crash if connecting to a
3583 malicious SSLv2 server. (CVE-2006-4343)
3584 [Tavis Ormandy and Will Drewry, Google Security Team]
3586 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3587 match only those. Before that, "AES256-SHA" would be interpreted
3588 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3589 the same strength classification in 0.9.7h) as we currently only
3590 have a single AES bit in the ciphersuite description bitmap.
3591 That change, however, also applied to ciphersuite strings such as
3592 "RC4-MD5" that intentionally matched multiple ciphersuites --
3593 namely, SSL 2.0 ciphersuites in addition to the more common ones
3594 from SSL 3.0/TLS 1.0.
3596 So we change the selection algorithm again: Naming an explicit
3597 ciphersuite selects this one ciphersuite, and any other similar
3598 ciphersuite (same bitmap) from *other* protocol versions.
3599 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3600 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3602 Since SSL 2.0 does not have any ciphersuites for which the
3603 128/256 bit distinction would be relevant, this works for now.
3604 The proper fix will be to use different bits for AES128 and
3605 AES256, which would have avoided the problems from the beginning;
3606 however, bits are scarce, so we can only do this in a new release
3607 (not just a patchlevel) when we can change the SSL_CIPHER
3608 definition to split the single 'unsigned long mask' bitmap into
3609 multiple values to extend the available space.
3613 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3615 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3616 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3618 *) Add AES IGE and biIGE modes.
3621 *) Change the Unix randomness entropy gathering to use poll() when
3622 possible instead of select(), since the latter has some
3623 undesirable limitations.
3624 [Darryl Miles via Richard Levitte and Bodo Moeller]
3626 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3627 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3628 cannot be implicitly activated as part of, e.g., the "AES" alias.
3629 However, please upgrade to OpenSSL 0.9.9[-dev] for
3630 non-experimental use of the ECC ciphersuites to get TLS extension
3631 support, which is required for curve and point format negotiation
3632 to avoid potential handshake problems.
3635 *) Disable rogue ciphersuites:
3637 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3638 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3639 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3641 The latter two were purportedly from
3642 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3645 Also deactivate the remaining ciphersuites from
3646 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3647 unofficial, and the ID has long expired.
3650 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3651 dual-core machines) and other potential thread-safety issues.
3654 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3655 versions), which is now available for royalty-free use
3656 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3657 Also, add Camellia TLS ciphersuites from RFC 4132.
3659 To minimize changes between patchlevels in the OpenSSL 0.9.8
3660 series, Camellia remains excluded from compilation unless OpenSSL
3661 is configured with 'enable-camellia'.
3664 *) Disable the padding bug check when compression is in use. The padding
3665 bug check assumes the first packet is of even length, this is not
3666 necessarily true if compresssion is enabled and can result in false
3667 positives causing handshake failure. The actual bug test is ancient
3668 code so it is hoped that implementations will either have fixed it by
3669 now or any which still have the bug do not support compression.
3672 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3674 *) When applying a cipher rule check to see if string match is an explicit
3675 cipher suite and only match that one cipher suite if it is.
3678 *) Link in manifests for VC++ if needed.
3679 [Austin Ziegler <halostatue@gmail.com>]
3681 *) Update support for ECC-based TLS ciphersuites according to
3682 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3683 TLS extensions, which are supported starting with the 0.9.9
3684 branch, not in the OpenSSL 0.9.8 branch).
3687 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3688 opaque EVP_CIPHER_CTX handling.
3691 *) Fixes and enhancements to zlib compression code. We now only use
3692 "zlib1.dll" and use the default __cdecl calling convention on Win32
3693 to conform with the standards mentioned here:
3694 http://www.zlib.net/DLL_FAQ.txt
3695 Static zlib linking now works on Windows and the new --with-zlib-include
3696 --with-zlib-lib options to Configure can be used to supply the location
3697 of the headers and library. Gracefully handle case where zlib library
3701 *) Several fixes and enhancements to the OID generation code. The old code
3702 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3703 handle numbers larger than ULONG_MAX, truncated printing and had a
3704 non standard OBJ_obj2txt() behaviour.
3707 *) Add support for building of engines under engine/ as shared libraries
3708 under VC++ build system.
3711 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3712 Hopefully, we will not see any false combination of paths any more.
3715 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3717 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3718 (part of SSL_OP_ALL). This option used to disable the
3719 countermeasure against man-in-the-middle protocol-version
3720 rollback in the SSL 2.0 server implementation, which is a bad
3721 idea. (CVE-2005-2969)
3723 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3724 for Information Security, National Institute of Advanced Industrial
3725 Science and Technology [AIST], Japan)]
3727 *) Add two function to clear and return the verify parameter flags.
3730 *) Keep cipherlists sorted in the source instead of sorting them at
3731 runtime, thus removing the need for a lock.
3734 *) Avoid some small subgroup attacks in Diffie-Hellman.
3735 [Nick Mathewson and Ben Laurie]
3737 *) Add functions for well-known primes.
3740 *) Extended Windows CE support.
3741 [Satoshi Nakamura and Andy Polyakov]
3743 *) Initialize SSL_METHOD structures at compile time instead of during
3744 runtime, thus removing the need for a lock.
3747 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3748 attempting to decrypt each encrypted key in turn. Add support to
3752 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3754 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3757 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3760 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3761 key into the same file any more.
3764 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3767 *) Add -utf8 command line and config file option to 'ca'.
3768 [Stefan <stf@udoma.org]
3770 *) Removed the macro des_crypt(), as it seems to conflict with some
3771 libraries. Use DES_crypt().
3774 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3775 involves renaming the source and generated shared-libs for
3776 both. The engines will accept the corrected or legacy ids
3777 ('ncipher' and '4758_cca' respectively) when binding. NB,
3778 this only applies when building 'shared'.
3779 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3781 *) Add attribute functions to EVP_PKEY structure. Modify
3782 PKCS12_create() to recognize a CSP name attribute and
3783 use it. Make -CSP option work again in pkcs12 utility.
3786 *) Add new functionality to the bn blinding code:
3787 - automatic re-creation of the BN_BLINDING parameters after
3788 a fixed number of uses (currently 32)
3789 - add new function for parameter creation
3790 - introduce flags to control the update behaviour of the
3791 BN_BLINDING parameters
3792 - hide BN_BLINDING structure
3793 Add a second BN_BLINDING slot to the RSA structure to improve
3794 performance when a single RSA object is shared among several
3798 *) Add support for DTLS.
3799 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3801 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3802 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3805 *) Remove buggy and incomplete DH cert support from
3806 ssl/ssl_rsa.c and ssl/s3_both.c
3809 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3810 the apps/openssl applications.
3813 *) Compile clean with "-Wall -Wmissing-prototypes
3814 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3815 DEBUG_SAFESTACK must also be set.
3818 *) Change ./Configure so that certain algorithms can be disabled by default.
3819 The new counterpiece to "no-xxx" is "enable-xxx".
3821 The patented RC5 and MDC2 algorithms will now be disabled unless
3822 "enable-rc5" and "enable-mdc2", respectively, are specified.
3824 (IDEA remains enabled despite being patented. This is because IDEA
3825 is frequently required for interoperability, and there is no license
3826 fee for non-commercial use. As before, "no-idea" can be used to
3827 avoid this algorithm.)
3831 *) Add processing of proxy certificates (see RFC 3820). This work was
3832 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3833 EGEE (Enabling Grids for E-science in Europe).
3836 *) RC4 performance overhaul on modern architectures/implementations, such
3837 as Intel P4, IA-64 and AMD64.
3840 *) New utility extract-section.pl. This can be used specify an alternative
3841 section number in a pod file instead of having to treat each file as
3842 a separate case in Makefile. This can be done by adding two lines to the
3845 =for comment openssl_section:XXX
3847 The blank line is mandatory.
3851 *) New arguments -certform, -keyform and -pass for s_client and s_server
3852 to allow alternative format key and certificate files and passphrase
3856 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3857 update associated structures and add various utility functions.
3859 Add new policy related verify parameters, include policy checking in
3860 standard verify code. Enhance 'smime' application with extra parameters
3861 to support policy checking and print out.
3864 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3865 Nehemiah processors. These extensions support AES encryption in hardware
3866 as well as RNG (though RNG support is currently disabled).
3867 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3869 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3872 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3873 [Andy Polyakov and a number of other people]
3875 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3876 implementation contributed by IBM.
3877 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3879 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3880 exponent rather than 'unsigned long'. There is a corresponding change to
3881 the new 'rsa_keygen' element of the RSA_METHOD structure.
3882 [Jelte Jansen, Geoff Thorpe]
3884 *) Functionality for creating the initial serial number file is now
3885 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3887 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3888 number file to 1, which is bound to cause problems. To avoid
3889 the problems while respecting compatibility between different 0.9.7
3890 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3891 CA.pl for serial number initialization. With the new release 0.9.8,
3892 we can fix the problem directly in the 'ca' utility.)
3895 *) Reduced header interdepencies by declaring more opaque objects in
3896 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3897 give fewer recursive includes, which could break lazy source code - so
3898 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3899 developers should define this symbol when building and using openssl to
3900 ensure they track the recommended behaviour, interfaces, [etc], but
3901 backwards-compatible behaviour prevails when this isn't defined.
3904 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3907 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3908 This will generate a random key of the appropriate length based on the
3909 cipher context. The EVP_CIPHER can provide its own random key generation
3910 routine to support keys of a specific form. This is used in the des and
3911 3des routines to generate a key of the correct parity. Update S/MIME
3912 code to use new functions and hence generate correct parity DES keys.
3913 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3914 valid (weak or incorrect parity).
3917 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3918 as looking them up. This is useful when the verified structure may contain
3919 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3920 present unless the new PKCS7_NO_CRL flag is asserted.
3923 *) Extend ASN1 oid configuration module. It now additionally accepts the
3926 shortName = some long name, 1.2.3.4
3929 *) Reimplemented the BN_CTX implementation. There is now no more static
3930 limitation on the number of variables it can handle nor the depth of the
3931 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3932 information can now expand as required, and rather than having a single
3933 static array of bignums, BN_CTX now uses a linked-list of such arrays
3934 allowing it to expand on demand whilst maintaining the usefulness of
3935 BN_CTX's "bundling".
3938 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3939 to allow all RSA operations to function using a single BN_CTX.
3942 *) Preliminary support for certificate policy evaluation and checking. This
3943 is initially intended to pass the tests outlined in "Conformance Testing
3944 of Relying Party Client Certificate Path Processing Logic" v1.07.
3947 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3948 remained unused and not that useful. A variety of other little bignum
3949 tweaks and fixes have also been made continuing on from the audit (see
3953 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3954 associated ASN1, EVP and SSL functions and old ASN1 macros.
3957 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3958 and this should never fail. So the return value from the use of
3959 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3960 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3963 *) BN_CTX_get() should return zero-valued bignums, providing the same
3964 initialised value as BN_new().
3965 [Geoff Thorpe, suggested by Ulf Möller]
3967 *) Support for inhibitAnyPolicy certificate extension.
3970 *) An audit of the BIGNUM code is underway, for which debugging code is
3971 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3972 is considered valid when processing BIGNUMs, and causes execution to
3973 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3974 further steps are taken to deliberately pollute unused data in BIGNUM
3975 structures to try and expose faulty code further on. For now, openssl will
3976 (in its default mode of operation) continue to tolerate the inconsistent
3977 forms that it has tolerated in the past, but authors and packagers should
3978 consider trying openssl and their own applications when compiled with
3979 these debugging symbols defined. It will help highlight potential bugs in
3980 their own code, and will improve the test coverage for OpenSSL itself. At
3981 some point, these tighter rules will become openssl's default to improve
3982 maintainability, though the assert()s and other overheads will remain only
3983 in debugging configurations. See bn.h for more details.
3984 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3986 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3987 that can only be obtained through BN_CTX_new() (which implicitly
3988 initialises it). The presence of this function only made it possible
3989 to overwrite an existing structure (and cause memory leaks).
3992 *) Because of the callback-based approach for implementing LHASH as a
3993 template type, lh_insert() adds opaque objects to hash-tables and
3994 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3995 to clean up those corresponding objects before destroying the hash table
3996 (and losing the object pointers). So some over-zealous constifications in
3997 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3998 objects as "const" and the lh_doall[_arg] callback wrappers are not
3999 prototyped to have "const" restrictions on the object pointers they are
4000 given (and so aren't required to cast them away any more).
4003 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4004 (speed) prefers to use its own implementation. The two implementations
4005 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4006 its object type properly exposed (MS_TM) instead of casting to/from "char
4007 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4008 aren't necessarily the greatest nomenclatures - but this is what was used
4009 internally to the implementation so I've used that for now.
4012 *) Ensure that deprecated functions do not get compiled when
4013 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4014 the self-tests were still using deprecated key-generation functions so
4015 these have been updated also.
4018 *) Reorganise PKCS#7 code to separate the digest location functionality
4019 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4020 New function PKCS7_set_digest() to set the digest type for PKCS#7
4021 digestedData type. Add additional code to correctly generate the
4022 digestedData type and add support for this type in PKCS7 initialization
4026 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4027 structure of type "other".
4030 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4031 sure the loop does correctly stop and breaking ("division by zero")
4032 modulus operations are not performed. The (pre-generated) prime
4033 table crypto/bn/bn_prime.h was already correct, but it could not be
4034 re-generated on some platforms because of the "division by zero"
4035 situation in the script.
4036 [Ralf S. Engelschall]
4038 *) Update support for ECC-based TLS ciphersuites according to
4039 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4040 SHA-1 now is only used for "small" curves (where the
4041 representation of a field element takes up to 24 bytes); for
4042 larger curves, the field element resulting from ECDH is directly
4043 used as premaster secret.
4044 [Douglas Stebila (Sun Microsystems Laboratories)]
4046 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4047 curve secp160r1 to the tests.
4048 [Douglas Stebila (Sun Microsystems Laboratories)]
4050 *) Add the possibility to load symbols globally with DSO.
4051 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4053 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4054 control of the error stack.
4057 *) Add support for STORE in ENGINE.
4060 *) Add the STORE type. The intention is to provide a common interface
4061 to certificate and key stores, be they simple file-based stores, or
4062 HSM-type store, or LDAP stores, or...
4063 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4066 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4067 pass a list of arguments to any function as well as provide a way
4068 for a function to pass data back to the caller.
4071 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4072 works like BUF_strdup() but can be used to duplicate a portion of
4073 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4077 *) Add the function sk_find_ex() which works like sk_find(), but will
4078 return an index to an element even if an exact match couldn't be
4079 found. The index is guaranteed to point at the element where the
4080 searched-for key would be inserted to preserve sorting order.
4083 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4084 takes an extra flags argument for optional functionality. Currently,
4085 the following flags are defined:
4087 OBJ_BSEARCH_VALUE_ON_NOMATCH
4088 This one gets OBJ_bsearch_ex() to return a pointer to the first
4089 element where the comparing function returns a negative or zero
4092 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4093 This one gets OBJ_bsearch_ex() to return a pointer to the first
4094 element where the comparing function returns zero. This is useful
4095 if there are more than one element where the comparing function
4099 *) Make it possible to create self-signed certificates with 'openssl ca'
4100 in such a way that the self-signed certificate becomes part of the
4101 CA database and uses the same mechanisms for serial number generation
4102 as all other certificate signing. The new flag '-selfsign' enables
4103 this functionality. Adapt CA.sh and CA.pl.in.
4106 *) Add functionality to check the public key of a certificate request
4107 against a given private. This is useful to check that a certificate
4108 request can be signed by that key (self-signing).
4111 *) Make it possible to have multiple active certificates with the same
4112 subject in the CA index file. This is done only if the keyword
4113 'unique_subject' is set to 'no' in the main CA section (default
4114 if 'CA_default') of the configuration file. The value is saved
4115 with the database itself in a separate index attribute file,
4116 named like the index file with '.attr' appended to the name.
4119 *) Generate muti valued AVAs using '+' notation in config files for
4123 *) Support for nameConstraints certificate extension.
4126 *) Support for policyConstraints certificate extension.
4129 *) Support for policyMappings certificate extension.
4132 *) Make sure the default DSA_METHOD implementation only uses its
4133 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4134 and change its own handlers to be NULL so as to remove unnecessary
4135 indirection. This lets alternative implementations fallback to the
4136 default implementation more easily.
4139 *) Support for directoryName in GeneralName related extensions
4143 *) Make it possible to link applications using Makefile.shared.
4144 Make that possible even when linking against static libraries!
4147 *) Support for single pass processing for S/MIME signing. This now
4148 means that S/MIME signing can be done from a pipe, in addition
4149 cleartext signing (multipart/signed type) is effectively streaming
4150 and the signed data does not need to be all held in memory.
4152 This is done with a new flag PKCS7_STREAM. When this flag is set
4153 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4154 is done after the data is output (and digests calculated) in
4155 SMIME_write_PKCS7().
4158 *) Add full support for -rpath/-R, both in shared libraries and
4159 applications, at least on the platforms where it's known how
4163 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4164 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4165 will now compute a table of multiples of the generator that
4166 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4167 faster (notably in the case of a single point multiplication,
4168 scalar * generator).
4169 [Nils Larsch, Bodo Moeller]
4171 *) IPv6 support for certificate extensions. The various extensions
4172 which use the IP:a.b.c.d can now take IPv6 addresses using the
4173 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4177 *) Added an ENGINE that implements RSA by performing private key
4178 exponentiations with the GMP library. The conversions to and from
4179 GMP's mpz_t format aren't optimised nor are any montgomery forms
4180 cached, and on x86 it appears OpenSSL's own performance has caught up.
4181 However there are likely to be other architectures where GMP could
4182 provide a boost. This ENGINE is not built in by default, but it can be
4183 specified at Configure time and should be accompanied by the necessary
4184 linker additions, eg;
4185 ./config -DOPENSSL_USE_GMP -lgmp
4188 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4189 testing availability of engines with "-t" - the old behaviour is
4190 produced by increasing the feature's verbosity with "-tt".
4193 *) ECDSA routines: under certain error conditions uninitialized BN objects
4194 could be freed. Solution: make sure initialization is performed early
4195 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4199 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4200 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4201 software implementations. For DSA and DH, parameter generation can
4202 also be overriden by providing the appropriate method callbacks.
4205 *) Change the "progress" mechanism used in key-generation and
4206 primality testing to functions that take a new BN_GENCB pointer in
4207 place of callback/argument pairs. The new API functions have "_ex"
4208 postfixes and the older functions are reimplemented as wrappers for
4209 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4210 declarations of the old functions to help (graceful) attempts to
4211 migrate to the new functions. Also, the new key-generation API
4212 functions operate on a caller-supplied key-structure and return
4213 success/failure rather than returning a key or NULL - this is to
4214 help make "keygen" another member function of RSA_METHOD etc.
4216 Example for using the new callback interface:
4218 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4222 BN_GENCB_set(&my_cb, my_callback, my_arg);
4224 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4225 /* For the meaning of a, b in calls to my_callback(), see the
4226 * documentation of the function that calls the callback.
4227 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4228 * my_callback should return 1 if it wants BN_is_prime_ex()
4229 * to continue, or 0 to stop.
4234 *) Change the ZLIB compression method to be stateful, and make it
4235 available to TLS with the number defined in
4236 draft-ietf-tls-compression-04.txt.
4239 *) Add the ASN.1 structures and functions for CertificatePair, which
4240 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4242 CertificatePair ::= SEQUENCE {
4243 forward [0] Certificate OPTIONAL,
4244 reverse [1] Certificate OPTIONAL,
4245 -- at least one of the pair shall be present -- }
4247 Also implement the PEM functions to read and write certificate
4248 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4250 This needed to be defined, mostly for the sake of the LDAP
4251 attribute crossCertificatePair, but may prove useful elsewhere as
4255 *) Make it possible to inhibit symlinking of shared libraries in
4256 Makefile.shared, for Cygwin's sake.
4259 *) Extend the BIGNUM API by creating a function
4260 void BN_set_negative(BIGNUM *a, int neg);
4261 and a macro that behave like
4262 int BN_is_negative(const BIGNUM *a);
4264 to avoid the need to access 'a->neg' directly in applications.
4267 *) Implement fast modular reduction for pseudo-Mersenne primes
4268 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4269 EC_GROUP_new_curve_GFp() will now automatically use this
4271 [Nils Larsch <nla@trustcenter.de>]
4273 *) Add new lock type (CRYPTO_LOCK_BN).
4276 *) Change the ENGINE framework to automatically load engines
4277 dynamically from specific directories unless they could be
4278 found to already be built in or loaded. Move all the
4279 current engines except for the cryptodev one to a new
4281 The engines in engines/ are built as shared libraries if
4282 the "shared" options was given to ./Configure or ./config.
4283 Otherwise, they are inserted in libcrypto.a.
4284 /usr/local/ssl/engines is the default directory for dynamic
4285 engines, but that can be overriden at configure time through
4286 the usual use of --prefix and/or --openssldir, and at run
4287 time with the environment variable OPENSSL_ENGINES.
4288 [Geoff Thorpe and Richard Levitte]
4290 *) Add Makefile.shared, a helper makefile to build shared
4291 libraries. Adapt Makefile.org.
4294 *) Add version info to Win32 DLLs.
4295 [Peter 'Luna' Runestig" <peter@runestig.com>]
4297 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4298 can be added using this API to created arbitrary PKCS#12
4299 files while avoiding the low level API.
4301 New options to PKCS12_create(), key or cert can be NULL and
4302 will then be omitted from the output file. The encryption
4303 algorithm NIDs can be set to -1 for no encryption, the mac
4304 iteration count can be set to 0 to omit the mac.
4306 Enhance pkcs12 utility by making the -nokeys and -nocerts
4307 options work when creating a PKCS#12 file. New option -nomac
4308 to omit the mac, NONE can be set for an encryption algorithm.
4309 New code is modified to use the enhanced PKCS12_create()
4310 instead of the low level API.
4313 *) Extend ASN1 encoder to support indefinite length constructed
4314 encoding. This can output sequences tags and octet strings in
4315 this form. Modify pk7_asn1.c to support indefinite length
4316 encoding. This is experimental and needs additional code to
4317 be useful, such as an ASN1 bio and some enhanced streaming
4320 Extend template encode functionality so that tagging is passed
4321 down to the template encoder.
4324 *) Let 'openssl req' fail if an argument to '-newkey' is not
4325 recognized instead of using RSA as a default.
4328 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4329 As these are not official, they are not included in "ALL";
4330 the "ECCdraft" ciphersuite group alias can be used to select them.
4331 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4333 *) Add ECDH engine support.
4334 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4336 *) Add ECDH in new directory crypto/ecdh/.
4337 [Douglas Stebila (Sun Microsystems Laboratories)]
4339 *) Let BN_rand_range() abort with an error after 100 iterations
4340 without success (which indicates a broken PRNG).
4343 *) Change BN_mod_sqrt() so that it verifies that the input value
4344 is really the square of the return value. (Previously,
4345 BN_mod_sqrt would show GIGO behaviour.)
4348 *) Add named elliptic curves over binary fields from X9.62, SECG,
4349 and WAP/WTLS; add OIDs that were still missing.
4351 [Sheueling Chang Shantz and Douglas Stebila
4352 (Sun Microsystems Laboratories)]
4354 *) Extend the EC library for elliptic curves over binary fields
4355 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4358 EC_GF2m_simple_method
4362 EC_GROUP_new_curve_GF2m
4363 EC_GROUP_set_curve_GF2m
4364 EC_GROUP_get_curve_GF2m
4365 EC_POINT_set_affine_coordinates_GF2m
4366 EC_POINT_get_affine_coordinates_GF2m
4367 EC_POINT_set_compressed_coordinates_GF2m
4369 Point compression for binary fields is disabled by default for
4370 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4373 As binary polynomials are represented as BIGNUMs, various members
4374 of the EC_GROUP and EC_POINT data structures can be shared
4375 between the implementations for prime fields and binary fields;
4376 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4377 are essentially identical to their ..._GFp counterparts.
4378 (For simplicity, the '..._GFp' prefix has been dropped from
4379 various internal method names.)
4381 An internal 'field_div' method (similar to 'field_mul' and
4382 'field_sqr') has been added; this is used only for binary fields.
4384 [Sheueling Chang Shantz and Douglas Stebila
4385 (Sun Microsystems Laboratories)]
4387 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4388 through methods ('mul', 'precompute_mult').
4390 The generic implementations (now internally called 'ec_wNAF_mul'
4391 and 'ec_wNAF_precomputed_mult') remain the default if these
4392 methods are undefined.
4394 [Sheueling Chang Shantz and Douglas Stebila
4395 (Sun Microsystems Laboratories)]
4397 *) New function EC_GROUP_get_degree, which is defined through
4398 EC_METHOD. For curves over prime fields, this returns the bit
4399 length of the modulus.
4401 [Sheueling Chang Shantz and Douglas Stebila
4402 (Sun Microsystems Laboratories)]
4404 *) New functions EC_GROUP_dup, EC_POINT_dup.
4405 (These simply call ..._new and ..._copy).
4407 [Sheueling Chang Shantz and Douglas Stebila
4408 (Sun Microsystems Laboratories)]
4410 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4411 Polynomials are represented as BIGNUMs (where the sign bit is not
4412 used) in the following functions [macros]:
4415 BN_GF2m_sub [= BN_GF2m_add]
4416 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4417 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4418 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4420 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4421 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4422 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4423 BN_GF2m_cmp [= BN_ucmp]
4425 (Note that only the 'mod' functions are actually for fields GF(2^m).
4426 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4428 For some functions, an the irreducible polynomial defining a
4429 field can be given as an 'unsigned int[]' with strictly
4430 decreasing elements giving the indices of those bits that are set;
4431 i.e., p[] represents the polynomial
4432 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4434 p[0] > p[1] > ... > p[k] = 0.
4435 This applies to the following functions:
4440 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4441 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4443 BN_GF2m_mod_sqrt_arr
4444 BN_GF2m_mod_solve_quad_arr
4448 Conversion can be performed by the following functions:
4453 bntest.c has additional tests for binary polynomial arithmetic.
4455 Two implementations for BN_GF2m_mod_div() are available.
4456 The default algorithm simply uses BN_GF2m_mod_inv() and
4457 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4458 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4459 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4461 [Sheueling Chang Shantz and Douglas Stebila
4462 (Sun Microsystems Laboratories)]
4464 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4465 functionality is disabled at compile-time.
4466 [Douglas Stebila <douglas.stebila@sun.com>]
4468 *) Change default behaviour of 'openssl asn1parse' so that more
4469 information is visible when viewing, e.g., a certificate:
4471 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4472 mode the content of non-printable OCTET STRINGs is output in a
4473 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4474 avoid the appearance of a printable string.
4475 [Nils Larsch <nla@trustcenter.de>]
4477 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4479 EC_GROUP_set_asn1_flag()
4480 EC_GROUP_get_asn1_flag()
4481 EC_GROUP_set_point_conversion_form()
4482 EC_GROUP_get_point_conversion_form()
4483 These control ASN1 encoding details:
4484 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4485 has been set to OPENSSL_EC_NAMED_CURVE.
4486 - Points are encoded in uncompressed form by default; options for
4487 asn1_for are as for point2oct, namely
4488 POINT_CONVERSION_COMPRESSED
4489 POINT_CONVERSION_UNCOMPRESSED
4490 POINT_CONVERSION_HYBRID
4492 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4495 EC_GROUP_get0_seed()
4496 EC_GROUP_get_seed_len()
4497 This is used only for ASN1 purposes (so far).
4498 [Nils Larsch <nla@trustcenter.de>]
4500 *) Add 'field_type' member to EC_METHOD, which holds the NID
4501 of the appropriate field type OID. The new function
4502 EC_METHOD_get_field_type() returns this value.
4503 [Nils Larsch <nla@trustcenter.de>]
4508 EC_POINT_point2hex()
4509 EC_POINT_hex2point()
4510 providing useful interfaces to EC_POINT_point2oct() and
4511 EC_POINT_oct2point().
4512 [Nils Larsch <nla@trustcenter.de>]
4514 *) Change internals of the EC library so that the functions
4515 EC_GROUP_set_generator()
4516 EC_GROUP_get_generator()
4517 EC_GROUP_get_order()
4518 EC_GROUP_get_cofactor()
4519 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4520 to methods, which would lead to unnecessary code duplication when
4521 adding different types of curves.
4522 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4524 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4525 arithmetic, and such that modified wNAFs are generated
4526 (which avoid length expansion in many cases).
4529 *) Add a function EC_GROUP_check_discriminant() (defined via
4530 EC_METHOD) that verifies that the curve discriminant is non-zero.
4532 Add a function EC_GROUP_check() that makes some sanity tests
4533 on a EC_GROUP, its generator and order. This includes
4534 EC_GROUP_check_discriminant().
4535 [Nils Larsch <nla@trustcenter.de>]
4537 *) Add ECDSA in new directory crypto/ecdsa/.
4539 Add applications 'openssl ecparam' and 'openssl ecdsa'
4540 (these are based on 'openssl dsaparam' and 'openssl dsa').
4542 ECDSA support is also included in various other files across the
4543 library. Most notably,
4544 - 'openssl req' now has a '-newkey ecdsa:file' option;
4545 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4546 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4547 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4548 them suitable for ECDSA where domain parameters must be
4549 extracted before the specific public key;
4550 - ECDSA engine support has been added.
4551 [Nils Larsch <nla@trustcenter.de>]
4553 *) Include some named elliptic curves, and add OIDs from X9.62,
4554 SECG, and WAP/WTLS. Each curve can be obtained from the new
4556 EC_GROUP_new_by_curve_name(),
4557 and the list of available named curves can be obtained with
4558 EC_get_builtin_curves().
4559 Also add a 'curve_name' member to EC_GROUP objects, which can be
4561 EC_GROUP_set_curve_name()
4562 EC_GROUP_get_curve_name()
4563 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4565 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4566 was actually never needed) and in BN_mul(). The removal in BN_mul()
4567 required a small change in bn_mul_part_recursive() and the addition
4568 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4569 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4570 bn_sub_words() and bn_add_words() except they take arrays with
4574 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4576 *) Cleanse PEM buffers before freeing them since they may contain
4578 [Benjamin Bennett <ben@psc.edu>]
4580 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4581 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4582 authentication-only ciphersuites.
4585 *) Since AES128 and AES256 share a single mask bit in the logic of
4586 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4587 kludge to work properly if AES128 is available and AES256 isn't.
4590 *) Expand security boundary to match 1.1.1 module.
4593 *) Remove redundant features: hash file source, editing of test vectors
4594 modify fipsld to use external fips_premain.c signature.
4597 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4598 run algorithm test programs.
4601 *) Make algorithm test programs more tolerant of whitespace.
4604 *) Have SSL/TLS server implementation tolerate "mismatched" record
4605 protocol version while receiving ClientHello even if the
4606 ClientHello is fragmented. (The server can't insist on the
4607 particular protocol version it has chosen before the ServerHello
4608 message has informed the client about his choice.)
4611 *) Load error codes if they are not already present instead of using a
4612 static variable. This allows them to be cleanly unloaded and reloaded.
4615 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4617 *) Introduce limits to prevent malicious keys being able to
4618 cause a denial of service. (CVE-2006-2940)
4619 [Steve Henson, Bodo Moeller]
4621 *) Fix ASN.1 parsing of certain invalid structures that can result
4622 in a denial of service. (CVE-2006-2937) [Steve Henson]
4624 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4625 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4627 *) Fix SSL client code which could crash if connecting to a
4628 malicious SSLv2 server. (CVE-2006-4343)
4629 [Tavis Ormandy and Will Drewry, Google Security Team]
4631 *) Change ciphersuite string processing so that an explicit
4632 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4633 will no longer include "AES128-SHA"), and any other similar
4634 ciphersuite (same bitmap) from *other* protocol versions (so that
4635 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4636 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4637 changes from 0.9.8b and 0.9.8d.
4640 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4642 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4643 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4645 *) Change the Unix randomness entropy gathering to use poll() when
4646 possible instead of select(), since the latter has some
4647 undesirable limitations.
4648 [Darryl Miles via Richard Levitte and Bodo Moeller]
4650 *) Disable rogue ciphersuites:
4652 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4653 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4654 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4656 The latter two were purportedly from
4657 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4660 Also deactive the remaining ciphersuites from
4661 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4662 unofficial, and the ID has long expired.
4665 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4666 dual-core machines) and other potential thread-safety issues.
4669 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4671 *) Adapt fipsld and the build system to link against the validated FIPS
4672 module in FIPS mode.
4675 *) Fixes for VC++ 2005 build under Windows.
4678 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4679 from a Windows bash shell such as MSYS. It is autodetected from the
4680 "config" script when run from a VC++ environment. Modify standard VC++
4681 build to use fipscanister.o from the GNU make build.
4684 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4686 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4687 The value now differs depending on if you build for FIPS or not.
4688 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4689 safely run with a non-FIPSed libcrypto, as it may crash because of
4690 the difference induced by this change.
4693 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4695 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4696 (part of SSL_OP_ALL). This option used to disable the
4697 countermeasure against man-in-the-middle protocol-version
4698 rollback in the SSL 2.0 server implementation, which is a bad
4699 idea. (CVE-2005-2969)
4701 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4702 for Information Security, National Institute of Advanced Industrial
4703 Science and Technology [AIST], Japan)]
4705 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4706 mainly for FIPS compliance and not fully integrated at this stage.
4709 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4710 the exponentiation using a fixed-length exponent. (Otherwise,
4711 the information leaked through timing could expose the secret key
4712 after many signatures; cf. Bleichenbacher's attack on DSA with
4716 *) Make a new fixed-window mod_exp implementation the default for
4717 RSA, DSA, and DH private-key operations so that the sequence of
4718 squares and multiplies and the memory access pattern are
4719 independent of the particular secret key. This will mitigate
4720 cache-timing and potential related attacks.
4722 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4723 and this is automatically used by BN_mod_exp_mont() if the new flag
4724 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4725 will use this BN flag for private exponents unless the flag
4726 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4727 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4729 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4731 *) Change the client implementation for SSLv23_method() and
4732 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4733 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4734 (Previously, the SSL 2.0 backwards compatible Client Hello
4735 message format would be used even with SSL_OP_NO_SSLv2.)
4738 *) Add support for smime-type MIME parameter in S/MIME messages which some
4742 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4743 a threadsafe manner. Modify rsa code to use new function and add calls
4744 to dsa and dh code (which had race conditions before).
4747 *) Include the fixed error library code in the C error file definitions
4748 instead of fixing them up at runtime. This keeps the error code
4749 structures constant.
4752 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4754 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4757 *) Fixes for newer kerberos headers. NB: the casts are needed because
4758 the 'length' field is signed on one version and unsigned on another
4759 with no (?) obvious way to tell the difference, without these VC++
4760 complains. Also the "definition" of FAR (blank) is no longer included
4761 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4762 some needed definitions.
4765 *) Undo Cygwin change.
4768 *) Added support for proxy certificates according to RFC 3820.
4769 Because they may be a security thread to unaware applications,
4770 they must be explicitely allowed in run-time. See
4771 docs/HOWTO/proxy_certificates.txt for further information.
4774 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4776 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4777 server and client random values. Previously
4778 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4779 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4781 This change has negligible security impact because:
4783 1. Server and client random values still have 24 bytes of pseudo random
4786 2. Server and client random values are sent in the clear in the initial
4789 3. The master secret is derived using the premaster secret (48 bytes in
4790 size for static RSA ciphersuites) as well as client server and random
4793 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4796 [Stephen Henson, reported by UK NISCC]
4798 *) Use Windows randomness collection on Cygwin.
4801 *) Fix hang in EGD/PRNGD query when communication socket is closed
4802 prematurely by EGD/PRNGD.
4803 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4805 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4808 *) Back-port of selected performance improvements from development
4809 branch, as well as improved support for PowerPC platforms.
4812 *) Add lots of checks for memory allocation failure, error codes to indicate
4813 failure and freeing up memory if a failure occurs.
4814 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4816 *) Add new -passin argument to dgst.
4819 *) Perform some character comparisons of different types in X509_NAME_cmp:
4820 this is needed for some certificates that reencode DNs into UTF8Strings
4821 (in violation of RFC3280) and can't or wont issue name rollover
4825 *) Make an explicit check during certificate validation to see that
4826 the CA setting in each certificate on the chain is correct. As a
4827 side effect always do the following basic checks on extensions,
4828 not just when there's an associated purpose to the check:
4830 - if there is an unhandled critical extension (unless the user
4831 has chosen to ignore this fault)
4832 - if the path length has been exceeded (if one is set at all)
4833 - that certain extensions fit the associated purpose (if one has
4837 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4839 *) Avoid a race condition when CRLs are checked in a multi threaded
4840 environment. This would happen due to the reordering of the revoked
4841 entries during signature checking and serial number lookup. Now the
4842 encoding is cached and the serial number sort performed under a lock.
4843 Add new STACK function sk_is_sorted().
4846 *) Add Delta CRL to the extension code.
4849 *) Various fixes to s3_pkt.c so alerts are sent properly.
4850 [David Holmes <d.holmes@f5.com>]
4852 *) Reduce the chances of duplicate issuer name and serial numbers (in
4853 violation of RFC3280) using the OpenSSL certificate creation utilities.
4854 This is done by creating a random 64 bit value for the initial serial
4855 number when a serial number file is created or when a self signed
4856 certificate is created using 'openssl req -x509'. The initial serial
4857 number file is created using 'openssl x509 -next_serial' in CA.pl
4858 rather than being initialized to 1.
4861 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4863 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4864 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4865 [Joe Orton, Steve Henson]
4867 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4869 [Joe Orton, Steve Henson]
4871 *) Make it possible to have multiple active certificates with the same
4872 subject in the CA index file. This is done only if the keyword
4873 'unique_subject' is set to 'no' in the main CA section (default
4874 if 'CA_default') of the configuration file. The value is saved
4875 with the database itself in a separate index attribute file,
4876 named like the index file with '.attr' appended to the name.
4879 *) X509 verify fixes. Disable broken certificate workarounds when
4880 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4881 keyUsage extension present. Don't accept CRLs with unhandled critical
4882 extensions: since verify currently doesn't process CRL extensions this
4883 rejects a CRL with *any* critical extensions. Add new verify error codes
4887 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4888 A clarification of RFC2560 will require the use of OCTET STRINGs and
4889 some implementations cannot handle the current raw format. Since OpenSSL
4890 copies and compares OCSP nonces as opaque blobs without any attempt at
4891 parsing them this should not create any compatibility issues.
4894 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4895 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4896 this HMAC (and other) operations are several times slower than OpenSSL
4900 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4901 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4903 *) Use the correct content when signing type "other".
4906 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4908 *) Fix various bugs revealed by running the NISCC test suite:
4910 Stop out of bounds reads in the ASN1 code when presented with
4911 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4913 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4915 If verify callback ignores invalid public key errors don't try to check
4916 certificate signature with the NULL public key.
4920 *) New -ignore_err option in ocsp application to stop the server
4921 exiting on the first error in a request.
4924 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4925 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4929 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4930 extra data after the compression methods not only for TLS 1.0
4931 but also for SSL 3.0 (as required by the specification).
4932 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4934 *) Change X509_certificate_type() to mark the key as exported/exportable
4935 when it's 512 *bits* long, not 512 bytes.
4938 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4939 blocks during encryption.
4942 *) Various fixes to base64 BIO and non blocking I/O. On write
4943 flushes were not handled properly if the BIO retried. On read
4944 data was not being buffered properly and had various logic bugs.
4945 This also affects blocking I/O when the data being decoded is a
4949 *) Various S/MIME bugfixes and compatibility changes:
4950 output correct application/pkcs7 MIME type if
4951 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4952 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4953 of files as .eml work). Correctly handle very long lines in MIME
4957 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4959 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4960 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4961 a protocol version number mismatch like a decryption error
4962 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4965 *) Turn on RSA blinding by default in the default implementation
4966 to avoid a timing attack. Applications that don't want it can call
4967 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4968 They would be ill-advised to do so in most cases.
4969 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4971 *) Change RSA blinding code so that it works when the PRNG is not
4972 seeded (in this case, the secret RSA exponent is abused as
4973 an unpredictable seed -- if it is not unpredictable, there
4974 is no point in blinding anyway). Make RSA blinding thread-safe
4975 by remembering the creator's thread ID in rsa->blinding and
4976 having all other threads use local one-time blinding factors
4977 (this requires more computation than sharing rsa->blinding, but
4978 avoids excessive locking; and if an RSA object is not shared
4979 between threads, blinding will still be very fast).
4982 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4983 ENGINE as defaults for all supported algorithms irrespective of
4984 the 'flags' parameter. 'flags' is now honoured, so applications
4985 should make sure they are passing it correctly.
4988 *) Target "mingw" now allows native Windows code to be generated in
4989 the Cygwin environment as well as with the MinGW compiler.
4992 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4994 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4995 via timing by performing a MAC computation even if incorrect
4996 block cipher padding has been found. This is a countermeasure
4997 against active attacks where the attacker has to distinguish
4998 between bad padding and a MAC verification error. (CVE-2003-0078)
5000 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5001 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5002 Martin Vuagnoux (EPFL, Ilion)]
5004 *) Make the no-err option work as intended. The intention with no-err
5005 is not to have the whole error stack handling routines removed from
5006 libcrypto, it's only intended to remove all the function name and
5007 reason texts, thereby removing some of the footprint that may not
5008 be interesting if those errors aren't displayed anyway.
5010 NOTE: it's still possible for any application or module to have it's
5011 own set of error texts inserted. The routines are there, just not
5012 used by default when no-err is given.
5015 *) Add support for FreeBSD on IA64.
5016 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5018 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5019 Kerberos function mit_des_cbc_cksum(). Before this change,
5020 the value returned by DES_cbc_cksum() was like the one from
5021 mit_des_cbc_cksum(), except the bytes were swapped.
5022 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5024 *) Allow an application to disable the automatic SSL chain building.
5025 Before this a rather primitive chain build was always performed in
5026 ssl3_output_cert_chain(): an application had no way to send the
5027 correct chain if the automatic operation produced an incorrect result.
5029 Now the chain builder is disabled if either:
5031 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5033 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5035 The reasoning behind this is that an application would not want the
5036 auto chain building to take place if extra chain certificates are
5037 present and it might also want a means of sending no additional
5038 certificates (for example the chain has two certificates and the
5042 *) Add the possibility to build without the ENGINE framework.
5043 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5045 *) Under Win32 gmtime() can return NULL: check return value in
5046 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5049 *) DSA routines: under certain error conditions uninitialized BN objects
5050 could be freed. Solution: make sure initialization is performed early
5051 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5052 Nils Larsch <nla@trustcenter.de> via PR#459)
5055 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5056 checked on reconnect on the client side, therefore session resumption
5057 could still fail with a "ssl session id is different" error. This
5058 behaviour is masked when SSL_OP_ALL is used due to
5059 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5060 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5061 followup to PR #377.
5064 *) IA-32 assembler support enhancements: unified ELF targets, support
5065 for SCO/Caldera platforms, fix for Cygwin shared build.
5068 *) Add support for FreeBSD on sparc64. As a consequence, support for
5069 FreeBSD on non-x86 processors is separate from x86 processors on
5070 the config script, much like the NetBSD support.
5071 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5073 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5075 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5078 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5079 code (06) was taken as the first octet of the session ID and the last
5080 octet was ignored consequently. As a result SSLv2 client side session
5081 caching could not have worked due to the session ID mismatch between
5083 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5087 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5088 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5092 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5093 seems that in spite of existing for more than a year, many application
5094 author have done nothing to provide the necessary callbacks, which
5095 means that this particular engine will not work properly anywhere.
5096 This is a very unfortunate situation which forces us, in the name
5097 of usability, to give the hw_ncipher.c a static lock, which is part
5099 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5100 appear in 0.9.8 or later. We EXPECT application authors to have
5101 dealt properly with this when 0.9.8 is released (unless we actually
5102 make such changes in the libcrypto locking code that changes will
5103 have to be made anyway).
5106 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5107 octets have been read, EOF or an error occurs. Without this change
5108 some truncated ASN1 structures will not produce an error.
5111 *) Disable Heimdal support, since it hasn't been fully implemented.
5112 Still give the possibility to force the use of Heimdal, but with
5113 warnings and a request that patches get sent to openssl-dev.
5116 *) Add the VC-CE target, introduce the WINCE sysname, and add
5117 INSTALL.WCE and appropriate conditionals to make it build.
5118 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5120 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5121 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5122 edit numbers of the version.
5123 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5125 *) Introduce safe string copy and catenation functions
5126 (BUF_strlcpy() and BUF_strlcat()).
5127 [Ben Laurie (CHATS) and Richard Levitte]
5129 *) Avoid using fixed-size buffers for one-line DNs.
5130 [Ben Laurie (CHATS)]
5132 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5133 resizing buffers containing secrets, and use where appropriate.
5134 [Ben Laurie (CHATS)]
5136 *) Avoid using fixed size buffers for configuration file location.
5137 [Ben Laurie (CHATS)]
5139 *) Avoid filename truncation for various CA files.
5140 [Ben Laurie (CHATS)]
5142 *) Use sizeof in preference to magic numbers.
5143 [Ben Laurie (CHATS)]
5145 *) Avoid filename truncation in cert requests.
5146 [Ben Laurie (CHATS)]
5148 *) Add assertions to check for (supposedly impossible) buffer
5150 [Ben Laurie (CHATS)]
5152 *) Don't cache truncated DNS entries in the local cache (this could
5153 potentially lead to a spoofing attack).
5154 [Ben Laurie (CHATS)]
5156 *) Fix various buffers to be large enough for hex/decimal
5157 representations in a platform independent manner.
5158 [Ben Laurie (CHATS)]
5160 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5161 resizing buffers containing secrets, and use where appropriate.
5162 [Ben Laurie (CHATS)]
5164 *) Add BIO_indent() to avoid much slightly worrying code to do
5166 [Ben Laurie (CHATS)]
5168 *) Convert sprintf()/BIO_puts() to BIO_printf().
5169 [Ben Laurie (CHATS)]
5171 *) buffer_gets() could terminate with the buffer only half
5173 [Ben Laurie (CHATS)]
5175 *) Add assertions to prevent user-supplied crypto functions from
5176 overflowing internal buffers by having large block sizes, etc.
5177 [Ben Laurie (CHATS)]
5179 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5181 [Ben Laurie (CHATS)]
5183 *) Eliminate unused copy of key in RC4.
5184 [Ben Laurie (CHATS)]
5186 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5187 [Ben Laurie (CHATS)]
5189 *) Fix off-by-one error in EGD path.
5190 [Ben Laurie (CHATS)]
5192 *) If RANDFILE path is too long, ignore instead of truncating.
5193 [Ben Laurie (CHATS)]
5195 *) Eliminate unused and incorrectly sized X.509 structure
5197 [Ben Laurie (CHATS)]
5199 *) Eliminate unused and dangerous function knumber().
5200 [Ben Laurie (CHATS)]
5202 *) Eliminate unused and dangerous structure, KSSL_ERR.
5203 [Ben Laurie (CHATS)]
5205 *) Protect against overlong session ID context length in an encoded
5206 session object. Since these are local, this does not appear to be
5208 [Ben Laurie (CHATS)]
5210 *) Change from security patch (see 0.9.6e below) that did not affect
5211 the 0.9.6 release series:
5213 Remote buffer overflow in SSL3 protocol - an attacker could
5214 supply an oversized master key in Kerberos-enabled versions.
5216 [Ben Laurie (CHATS)]
5218 *) Change the SSL kerb5 codes to match RFC 2712.
5221 *) Make -nameopt work fully for req and add -reqopt switch.
5222 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5224 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5225 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5227 *) Make sure tests can be performed even if the corresponding algorithms
5228 have been removed entirely. This was also the last step to make
5229 OpenSSL compilable with DJGPP under all reasonable conditions.
5230 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5232 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5233 to allow version independent disabling of normally unselected ciphers,
5234 which may be activated as a side-effect of selecting a single cipher.
5236 (E.g., cipher list string "RSA" enables ciphersuites that are left
5237 out of "ALL" because they do not provide symmetric encryption.
5238 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5239 [Lutz Jaenicke, Bodo Moeller]
5241 *) Add appropriate support for separate platform-dependent build
5242 directories. The recommended way to make a platform-dependent
5243 build directory is the following (tested on Linux), maybe with
5246 # Place yourself outside of the OpenSSL source tree. In
5247 # this example, the environment variable OPENSSL_SOURCE
5248 # is assumed to contain the absolute OpenSSL source directory.
5249 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5250 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5251 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5252 mkdir -p `dirname $F`
5253 ln -s $OPENSSL_SOURCE/$F $F
5256 To be absolutely sure not to disturb the source tree, a "make clean"
5257 is a good thing. If it isn't successful, don't worry about it,
5258 it probably means the source directory is very clean.
5261 *) Make sure any ENGINE control commands make local copies of string
5262 pointers passed to them whenever necessary. Otherwise it is possible
5263 the caller may have overwritten (or deallocated) the original string
5264 data when a later ENGINE operation tries to use the stored values.
5265 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5267 *) Improve diagnostics in file reading and command-line digests.
5268 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5270 *) Add AES modes CFB and OFB to the object database. Correct an
5271 error in AES-CFB decryption.
5274 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5275 allows existing EVP_CIPHER_CTX structures to be reused after
5276 calling EVP_*Final(). This behaviour is used by encryption
5277 BIOs and some applications. This has the side effect that
5278 applications must explicitly clean up cipher contexts with
5279 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5282 *) Check the values of dna and dnb in bn_mul_recursive before calling
5283 bn_mul_comba (a non zero value means the a or b arrays do not contain
5284 n2 elements) and fallback to bn_mul_normal if either is not zero.
5287 *) Fix escaping of non-ASCII characters when using the -subj option
5288 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5291 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5292 form for "surname", serialNumber has no short form.
5293 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5294 therefore remove "mail" short name for "internet 7".
5295 The OID for unique identifiers in X509 certificates is
5296 x500UniqueIdentifier, not uniqueIdentifier.
5297 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5300 *) Add an "init" command to the ENGINE config module and auto initialize
5301 ENGINEs. Without any "init" command the ENGINE will be initialized
5302 after all ctrl commands have been executed on it. If init=1 the
5303 ENGINE is initailized at that point (ctrls before that point are run
5304 on the uninitialized ENGINE and after on the initialized one). If
5305 init=0 then the ENGINE will not be iniatialized at all.
5308 *) Fix the 'app_verify_callback' interface so that the user-defined
5309 argument is actually passed to the callback: In the
5310 SSL_CTX_set_cert_verify_callback() prototype, the callback
5311 declaration has been changed from
5314 int (*cb)(X509_STORE_CTX *,void *);
5315 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5316 i=s->ctx->app_verify_callback(&ctx)
5317 has been changed into
5318 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5320 To update applications using SSL_CTX_set_cert_verify_callback(),
5321 a dummy argument can be added to their callback functions.
5322 [D. K. Smetters <smetters@parc.xerox.com>]
5324 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5325 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5327 *) Add and OPENSSL_LOAD_CONF define which will cause
5328 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5329 This allows older applications to transparently support certain
5330 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5331 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5332 load the config file and OPENSSL_add_all_algorithms_conf() which will
5333 always load it have also been added.
5336 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5337 Adjust NIDs and EVP layer.
5338 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5340 *) Config modules support in openssl utility.
5342 Most commands now load modules from the config file,
5343 though in a few (such as version) this isn't done
5344 because it couldn't be used for anything.
5346 In the case of ca and req the config file used is
5347 the same as the utility itself: that is the -config
5348 command line option can be used to specify an
5352 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5353 use "openssl_conf" if filename is NULL use default openssl config file.
5356 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5357 config section name. Add a new flag to tolerate a missing config file
5358 and move code to CONF_modules_load_file().
5361 *) Support for crypto accelerator cards from Accelerated Encryption
5362 Processing, www.aep.ie. (Use engine 'aep')
5363 The support was copied from 0.9.6c [engine] and adapted/corrected
5364 to work with the new engine framework.
5365 [AEP Inc. and Richard Levitte]
5367 *) Support for SureWare crypto accelerator cards from Baltimore
5368 Technologies. (Use engine 'sureware')
5369 The support was copied from 0.9.6c [engine] and adapted
5370 to work with the new engine framework.
5373 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5374 make the newer ENGINE framework commands for the CHIL engine work.
5375 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5377 *) Make it possible to produce shared libraries on ReliantUNIX.
5378 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5380 *) Add the configuration target debug-linux-ppro.
5381 Make 'openssl rsa' use the general key loading routines
5382 implemented in apps.c, and make those routines able to
5383 handle the key format FORMAT_NETSCAPE and the variant
5385 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5387 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5388 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5390 *) Add -keyform to rsautl, and document -engine.
5391 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5393 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5394 BIO_R_NO_SUCH_FILE error code rather than the generic
5395 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5398 *) Add new functions
5400 ERR_peek_last_error_line
5401 ERR_peek_last_error_line_data.
5402 These are similar to
5405 ERR_peek_error_line_data,
5406 but report on the latest error recorded rather than the first one
5407 still in the error queue.
5408 [Ben Laurie, Bodo Moeller]
5410 *) default_algorithms option in ENGINE config module. This allows things
5412 default_algorithms = ALL
5413 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5416 *) Preliminary ENGINE config module.
5419 *) New experimental application configuration code.
5422 *) Change the AES code to follow the same name structure as all other
5423 symmetric ciphers, and behave the same way. Move everything to
5424 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5425 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5427 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5428 [Ben Laurie and Theo de Raadt]
5430 *) Add option to output public keys in req command.
5431 [Massimiliano Pala madwolf@openca.org]
5433 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5434 (up to about 10% better than before for P-192 and P-224).
5437 *) New functions/macros
5439 SSL_CTX_set_msg_callback(ctx, cb)
5440 SSL_CTX_set_msg_callback_arg(ctx, arg)
5441 SSL_set_msg_callback(ssl, cb)
5442 SSL_set_msg_callback_arg(ssl, arg)
5444 to request calling a callback function
5446 void cb(int write_p, int version, int content_type,
5447 const void *buf, size_t len, SSL *ssl, void *arg)
5449 whenever a protocol message has been completely received
5450 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5451 protocol version according to which the SSL library interprets
5452 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5453 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5454 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5455 specification (change_cipher_spec(20), alert(21), handshake(22)).
5456 'buf' and 'len' point to the actual message, 'ssl' to the
5457 SSL object, and 'arg' is the application-defined value set by
5458 SSL[_CTX]_set_msg_callback_arg().
5460 'openssl s_client' and 'openssl s_server' have new '-msg' options
5461 to enable a callback that displays all protocol messages.
5464 *) Change the shared library support so shared libraries are built as
5465 soon as the corresponding static library is finished, and thereby get
5466 openssl and the test programs linked against the shared library.
5467 This still only happens when the keyword "shard" has been given to
5468 the configuration scripts.
5470 NOTE: shared library support is still an experimental thing, and
5471 backward binary compatibility is still not guaranteed.
5472 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5474 *) Add support for Subject Information Access extension.
5475 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5477 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5478 additional bytes when new memory had to be allocated, not just
5479 when reusing an existing buffer.
5482 *) New command line and configuration option 'utf8' for the req command.
5483 This allows field values to be specified as UTF8 strings.
5486 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5487 runs for the former and machine-readable output for the latter.
5490 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5491 of the e-mail address in the DN (i.e., it will go into a certificate
5492 extension only). The new configuration file option 'email_in_dn = no'
5493 has the same effect.
5494 [Massimiliano Pala madwolf@openca.org]
5496 *) Change all functions with names starting with des_ to be starting
5497 with DES_ instead. Add wrappers that are compatible with libdes,
5498 but are named _ossl_old_des_*. Finally, add macros that map the
5499 des_* symbols to the corresponding _ossl_old_des_* if libdes
5500 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5501 desired, the des_* symbols will be mapped to DES_*, with one
5504 Since we provide two compatibility mappings, the user needs to
5505 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5506 compatibility is desired. The default (i.e., when that macro
5507 isn't defined) is OpenSSL 0.9.6c compatibility.
5509 There are also macros that enable and disable the support of old
5510 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5511 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5512 are defined, the default will apply: to support the old des routines.
5514 In either case, one must include openssl/des.h to get the correct
5515 definitions. Do not try to just include openssl/des_old.h, that
5518 NOTE: This is a major break of an old API into a new one. Software
5519 authors are encouraged to switch to the DES_ style functions. Some
5520 time in the future, des_old.h and the libdes compatibility functions
5521 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5522 default), and then completely removed.
5525 *) Test for certificates which contain unsupported critical extensions.
5526 If such a certificate is found during a verify operation it is
5527 rejected by default: this behaviour can be overridden by either
5528 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5529 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5530 X509_supported_extension() has also been added which returns 1 if a
5531 particular extension is supported.
5534 *) Modify the behaviour of EVP cipher functions in similar way to digests
5535 to retain compatibility with existing code.
5538 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5539 compatibility with existing code. In particular the 'ctx' parameter does
5540 not have to be to be initialized before the call to EVP_DigestInit() and
5541 it is tidied up after a call to EVP_DigestFinal(). New function
5542 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5543 EVP_MD_CTX_copy() changed to not require the destination to be
5544 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5545 requires the destination to be valid.
5547 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5548 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5551 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5552 so that complete 'Handshake' protocol structures are kept in memory
5553 instead of overwriting 'msg_type' and 'length' with 'body' data.
5556 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5557 [Massimo Santin via Richard Levitte]
5559 *) Major restructuring to the underlying ENGINE code. This includes
5560 reduction of linker bloat, separation of pure "ENGINE" manipulation
5561 (initialisation, etc) from functionality dealing with implementations
5562 of specific crypto iterfaces. This change also introduces integrated
5563 support for symmetric ciphers and digest implementations - so ENGINEs
5564 can now accelerate these by providing EVP_CIPHER and EVP_MD
5565 implementations of their own. This is detailed in crypto/engine/README
5566 as it couldn't be adequately described here. However, there are a few
5567 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5568 were changed in the original introduction of ENGINE code have now
5569 reverted back - the hooking from this code to ENGINE is now a good
5570 deal more passive and at run-time, operations deal directly with
5571 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5572 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5573 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5574 they were not being used by the framework as there is no concept of a
5575 BIGNUM_METHOD and they could not be generalised to the new
5576 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5577 ENGINE_cpy() has been removed as it cannot be consistently defined in
5581 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5584 *) Change mkdef.pl to sort symbols that get the same entry number,
5585 and make sure the automatically generated functions ERR_load_*
5586 become part of libeay.num as well.
5589 *) New function SSL_renegotiate_pending(). This returns true once
5590 renegotiation has been requested (either SSL_renegotiate() call
5591 or HelloRequest/ClientHello received from the peer) and becomes
5592 false once a handshake has been completed.
5593 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5594 sends a HelloRequest, but does not ensure that a handshake takes
5595 place. SSL_renegotiate_pending() is useful for checking if the
5596 client has followed the request.)
5599 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5600 By default, clients may request session resumption even during
5601 renegotiation (if session ID contexts permit); with this option,
5602 session resumption is possible only in the first handshake.
5604 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5605 more bits available for options that should not be part of
5606 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5609 *) Add some demos for certificate and certificate request creation.
5612 *) Make maximum certificate chain size accepted from the peer application
5613 settable (SSL*_get/set_max_cert_list()), as proposed by
5614 "Douglas E. Engert" <deengert@anl.gov>.
5617 *) Add support for shared libraries for Unixware-7
5618 (Boyd Lynn Gerber <gerberb@zenez.com>).
5621 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5622 be done prior to destruction. Use this to unload error strings from
5623 ENGINEs that load their own error strings. NB: This adds two new API
5624 functions to "get" and "set" this destroy handler in an ENGINE.
5627 *) Alter all existing ENGINE implementations (except "openssl" and
5628 "openbsd") to dynamically instantiate their own error strings. This
5629 makes them more flexible to be built both as statically-linked ENGINEs
5630 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5631 Also, add stub code to each that makes building them as self-contained
5632 shared-libraries easier (see README.ENGINE).
5635 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5636 implementations into applications that are completely implemented in
5637 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5638 commands that can be used to configure what shared-library to load and
5639 to control aspects of the way it is handled. Also, made an update to
5640 the README.ENGINE file that brings its information up-to-date and
5641 provides some information and instructions on the "dynamic" ENGINE
5642 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5645 *) Make it possible to unload ranges of ERR strings with a new
5646 "ERR_unload_strings" function.
5649 *) Add a copy() function to EVP_MD.
5652 *) Make EVP_MD routines take a context pointer instead of just the
5653 md_data void pointer.
5656 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5657 that the digest can only process a single chunk of data
5658 (typically because it is provided by a piece of
5659 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5660 is only going to provide a single chunk of data, and hence the
5661 framework needn't accumulate the data for oneshot drivers.
5664 *) As with "ERR", make it possible to replace the underlying "ex_data"
5665 functions. This change also alters the storage and management of global
5666 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5667 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5668 index counters. The API functions that use this state have been changed
5669 to take a "class_index" rather than pointers to the class's local STACK
5670 and counter, and there is now an API function to dynamically create new
5671 classes. This centralisation allows us to (a) plug a lot of the
5672 thread-safety problems that existed, and (b) makes it possible to clean
5673 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5674 such data would previously have always leaked in application code and
5675 workarounds were in place to make the memory debugging turn a blind eye
5676 to it. Application code that doesn't use this new function will still
5677 leak as before, but their memory debugging output will announce it now
5678 rather than letting it slide.
5680 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5681 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5682 has a return value to indicate success or failure.
5685 *) Make it possible to replace the underlying "ERR" functions such that the
5686 global state (2 LHASH tables and 2 locks) is only used by the "default"
5687 implementation. This change also adds two functions to "get" and "set"
5688 the implementation prior to it being automatically set the first time
5689 any other ERR function takes place. Ie. an application can call "get",
5690 pass the return value to a module it has just loaded, and that module
5691 can call its own "set" function using that value. This means the
5692 module's "ERR" operations will use (and modify) the error state in the
5693 application and not in its own statically linked copy of OpenSSL code.
5696 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5697 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5698 the operation, and provides a more encapsulated way for external code
5699 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5700 to use these functions rather than manually incrementing the counts.
5702 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5705 *) Add EVP test program.
5708 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5711 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5712 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5713 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5714 These allow a CRL to be built without having to access X509_CRL fields
5715 directly. Modify 'ca' application to use new functions.
5718 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5719 bug workarounds. Rollback attack detection is a security feature.
5720 The problem will only arise on OpenSSL servers when TLSv1 is not
5721 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5722 Software authors not wanting to support TLSv1 will have special reasons
5723 for their choice and can explicitly enable this option.
5724 [Bodo Moeller, Lutz Jaenicke]
5726 *) Rationalise EVP so it can be extended: don't include a union of
5727 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5728 (similar to those existing for EVP_CIPHER_CTX).
5733 EVP_MD_CTX_init(&md); /* new function call */
5734 EVP_DigestInit(&md, EVP_sha1());
5735 EVP_DigestUpdate(&md, in, len);
5736 EVP_DigestFinal(&md, out, NULL);
5737 EVP_MD_CTX_cleanup(&md); /* new function call */
5741 *) Make DES key schedule conform to the usual scheme, as well as
5742 correcting its structure. This means that calls to DES functions
5743 now have to pass a pointer to a des_key_schedule instead of a
5744 plain des_key_schedule (which was actually always a pointer
5747 des_key_schedule ks;
5749 des_set_key_checked(..., &ks);
5750 des_ncbc_encrypt(..., &ks, ...);
5752 (Note that a later change renames 'des_...' into 'DES_...'.)
5755 *) Initial reduction of linker bloat: the use of some functions, such as
5756 PEM causes large amounts of unused functions to be linked in due to
5757 poor organisation. For example pem_all.c contains every PEM function
5758 which has a knock on effect of linking in large amounts of (unused)
5759 ASN1 code. Grouping together similar functions and splitting unrelated
5760 functions prevents this.
5763 *) Cleanup of EVP macros.
5766 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5767 correct _ecb suffix.
5770 *) Add initial OCSP responder support to ocsp application. The
5771 revocation information is handled using the text based index
5772 use by the ca application. The responder can either handle
5773 requests generated internally, supplied in files (for example
5774 via a CGI script) or using an internal minimal server.
5777 *) Add configuration choices to get zlib compression for TLS.
5780 *) Changes to Kerberos SSL for RFC 2712 compliance:
5781 1. Implemented real KerberosWrapper, instead of just using
5782 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5783 2. Implemented optional authenticator field of KerberosWrapper.
5785 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5786 and authenticator structs; see crypto/krb5/.
5788 Generalized Kerberos calls to support multiple Kerberos libraries.
5789 [Vern Staats <staatsvr@asc.hpc.mil>,
5790 Jeffrey Altman <jaltman@columbia.edu>
5791 via Richard Levitte]
5793 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5794 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5795 values for each of the key sizes rather than having just
5796 parameters (and 'speed' generating keys each time).
5799 *) Speed up EVP routines.
5802 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5803 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5804 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5805 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5807 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5808 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5809 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5812 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5814 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5817 *) Added the OS2-EMX target.
5818 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5820 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5821 to support NCONF routines in extension code. New function CONF_set_nconf()
5822 to allow functions which take an NCONF to also handle the old LHASH
5823 structure: this means that the old CONF compatible routines can be
5824 retained (in particular wrt extensions) without having to duplicate the
5825 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5828 *) Enhance the general user interface with mechanisms for inner control
5829 and with possibilities to have yes/no kind of prompts.
5832 *) Change all calls to low level digest routines in the library and
5833 applications to use EVP. Add missing calls to HMAC_cleanup() and
5834 don't assume HMAC_CTX can be copied using memcpy().
5835 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5837 *) Add the possibility to control engines through control names but with
5838 arbitrary arguments instead of just a string.
5839 Change the key loaders to take a UI_METHOD instead of a callback
5840 function pointer. NOTE: this breaks binary compatibility with earlier
5841 versions of OpenSSL [engine].
5842 Adapt the nCipher code for these new conditions and add a card insertion
5846 *) Enhance the general user interface with mechanisms to better support
5847 dialog box interfaces, application-defined prompts, the possibility
5848 to use defaults (for example default passwords from somewhere else)
5849 and interrupts/cancellations.
5852 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5853 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5856 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5857 tidy up some unnecessarily weird code in 'sk_new()').
5858 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5860 *) Change the key loading routines for ENGINEs to use the same kind
5861 callback (pem_password_cb) as all other routines that need this
5865 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5866 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5867 than this minimum value is recommended.
5870 *) New random seeder for OpenVMS, using the system process statistics
5871 that are easily reachable.
5874 *) Windows apparently can't transparently handle global
5875 variables defined in DLLs. Initialisations such as:
5877 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5879 wont compile. This is used by the any applications that need to
5880 declare their own ASN1 modules. This was fixed by adding the option
5881 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5882 needed for static libraries under Win32.
5885 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5886 setting of purpose and trust fields. New X509_STORE trust and
5887 purpose functions and tidy up setting in other SSL functions.
5890 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5891 structure. These are inherited by X509_STORE_CTX when it is
5892 initialised. This allows various defaults to be set in the
5893 X509_STORE structure (such as flags for CRL checking and custom
5894 purpose or trust settings) for functions which only use X509_STORE_CTX
5895 internally such as S/MIME.
5897 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5898 trust settings if they are not set in X509_STORE. This allows X509_STORE
5899 purposes and trust (in S/MIME for example) to override any set by default.
5901 Add command line options for CRL checking to smime, s_client and s_server
5905 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5906 are set then the CRL is looked up in the X509_STORE structure and
5907 its validity and signature checked, then if the certificate is found
5908 in the CRL the verify fails with a revoked error.
5910 Various new CRL related callbacks added to X509_STORE_CTX structure.
5912 Command line options added to 'verify' application to support this.
5914 This needs some additional work, such as being able to handle multiple
5915 CRLs with different times, extension based lookup (rather than just
5916 by subject name) and ultimately more complete V2 CRL extension
5920 *) Add a general user interface API (crypto/ui/). This is designed
5921 to replace things like des_read_password and friends (backward
5922 compatibility functions using this new API are provided).
5923 The purpose is to remove prompting functions from the DES code
5924 section as well as provide for prompting through dialog boxes in
5925 a window system and the like.
5928 *) Add "ex_data" support to ENGINE so implementations can add state at a
5929 per-structure level rather than having to store it globally.
5932 *) Make it possible for ENGINE structures to be copied when retrieved by
5933 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5934 This causes the "original" ENGINE structure to act like a template,
5935 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5936 operational state can be localised to each ENGINE structure, despite the
5937 fact they all share the same "methods". New ENGINE structures returned in
5938 this case have no functional references and the return value is the single
5939 structural reference. This matches the single structural reference returned
5940 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5944 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5945 needs to match any other type at all we need to manually clear the
5949 *) Changes to the "openssl engine" utility to include;
5950 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5951 about an ENGINE's available control commands.
5952 - executing control commands from command line arguments using the
5953 '-pre' and '-post' switches. '-post' is only used if '-t' is
5954 specified and the ENGINE is successfully initialised. The syntax for
5955 the individual commands are colon-separated, for example;
5956 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5959 *) New dynamic control command support for ENGINEs. ENGINEs can now
5960 declare their own commands (numbers), names (strings), descriptions,
5961 and input types for run-time discovery by calling applications. A
5962 subset of these commands are implicitly classed as "executable"
5963 depending on their input type, and only these can be invoked through
5964 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5965 can be based on user input, config files, etc). The distinction is
5966 that "executable" commands cannot return anything other than a boolean
5967 result and can only support numeric or string input, whereas some
5968 discoverable commands may only be for direct use through
5969 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5970 pointers, or other custom uses. The "executable" commands are to
5971 support parameterisations of ENGINE behaviour that can be
5972 unambiguously defined by ENGINEs and used consistently across any
5973 OpenSSL-based application. Commands have been added to all the
5974 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5975 control over shared-library paths without source code alterations.
5978 *) Changed all ENGINE implementations to dynamically allocate their
5979 ENGINEs rather than declaring them statically. Apart from this being
5980 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5981 this also allows the implementations to compile without using the
5982 internal engine_int.h header.
5985 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5986 'const' value. Any code that should be able to modify a RAND_METHOD
5987 should already have non-const pointers to it (ie. they should only
5988 modify their own ones).
5991 *) Made a variety of little tweaks to the ENGINE code.
5992 - "atalla" and "ubsec" string definitions were moved from header files
5993 to C code. "nuron" string definitions were placed in variables
5994 rather than hard-coded - allowing parameterisation of these values
5995 later on via ctrl() commands.
5996 - Removed unused "#if 0"'d code.
5997 - Fixed engine list iteration code so it uses ENGINE_free() to release
5998 structural references.
5999 - Constified the RAND_METHOD element of ENGINE structures.
6000 - Constified various get/set functions as appropriate and added
6001 missing functions (including a catch-all ENGINE_cpy that duplicates
6002 all ENGINE values onto a new ENGINE except reference counts/state).
6003 - Removed NULL parameter checks in get/set functions. Setting a method
6004 or function to NULL is a way of cancelling out a previously set
6005 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6006 and doesn't justify the extra error symbols and code.
6007 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6008 flags from engine_int.h to engine.h.
6009 - Changed prototypes for ENGINE handler functions (init(), finish(),
6010 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6013 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6014 to the algorithm using long division. The binary algorithm can be
6015 used only if the modulus is odd. On 32-bit systems, it is faster
6016 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6017 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6018 up to 450 bits. In 64-bit environments, the binary algorithm
6019 appears to be advantageous for much longer moduli; here we use it
6020 for moduli up to 2048 bits.
6023 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6024 could not support the combine flag in choice fields.
6027 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6028 extensions from a certificate request to the certificate.
6031 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6032 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6033 file: this allows the display of the certificate about to be
6034 signed to be customised, to allow certain fields to be included
6035 or excluded and extension details. The old system didn't display
6036 multicharacter strings properly, omitted fields not in the policy
6037 and couldn't display additional details such as extensions.
6040 *) Function EC_POINTs_mul for multiple scalar multiplication
6041 of an arbitrary number of elliptic curve points
6042 \sum scalars[i]*points[i],
6043 optionally including the generator defined for the EC_GROUP:
6044 scalar*generator + \sum scalars[i]*points[i].
6046 EC_POINT_mul is a simple wrapper function for the typical case
6047 that the point list has just one item (besides the optional
6051 *) First EC_METHODs for curves over GF(p):
6053 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6054 operations and provides various method functions that can also
6055 operate with faster implementations of modular arithmetic.
6057 EC_GFp_mont_method() reuses most functions that are part of
6058 EC_GFp_simple_method, but uses Montgomery arithmetic.
6060 [Bodo Moeller; point addition and point doubling
6061 implementation directly derived from source code provided by
6062 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6064 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6065 crypto/ec/ec_lib.c):
6067 Curves are EC_GROUP objects (with an optional group generator)
6068 based on EC_METHODs that are built into the library.
6070 Points are EC_POINT objects based on EC_GROUP objects.
6072 Most of the framework would be able to handle curves over arbitrary
6073 finite fields, but as there are no obvious types for fields other
6074 than GF(p), some functions are limited to that for now.
6077 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6078 that the file contains a complete HTTP response.
6081 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6082 change the def and num file printf format specifier from "%-40sXXX"
6083 to "%-39s XXX". The latter will always guarantee a space after the
6084 field while the former will cause them to run together if the field
6085 is 40 of more characters long.
6088 *) Constify the cipher and digest 'method' functions and structures
6089 and modify related functions to take constant EVP_MD and EVP_CIPHER
6093 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6094 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6097 *) Modify EVP_Digest*() routines so they now return values. Although the
6098 internal software routines can never fail additional hardware versions
6102 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6104 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6105 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6108 ERR_R_NESTED_ASN1_ERROR
6110 ERR_R_MISSING_ASN1_EOS
6111 were 4 .. 9, conflicting with
6112 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6114 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6115 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6117 Add new error code 'ERR_R_INTERNAL_ERROR'.
6120 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6124 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6125 sets the subject name for a new request or supersedes the
6126 subject name in a given request. Formats that can be parsed are
6127 'CN=Some Name, OU=myOU, C=IT'
6129 'CN=Some Name/OU=myOU/C=IT'.
6131 Add options '-batch' and '-verbose' to 'openssl req'.
6132 [Massimiliano Pala <madwolf@hackmasters.net>]
6134 *) Introduce the possibility to access global variables through
6135 functions on platform were that's the best way to handle exporting
6136 global variables in shared libraries. To enable this functionality,
6137 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6138 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6139 is normally done by Configure or something similar).
6141 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6142 in the source file (foo.c) like this:
6144 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6145 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6147 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6148 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6150 OPENSSL_DECLARE_GLOBAL(int,foo);
6151 #define foo OPENSSL_GLOBAL_REF(foo)
6152 OPENSSL_DECLARE_GLOBAL(double,bar);
6153 #define bar OPENSSL_GLOBAL_REF(bar)
6155 The #defines are very important, and therefore so is including the
6156 header file everywhere where the defined globals are used.
6158 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6159 of ASN.1 items, but that structure is a bit different.
6161 The largest change is in util/mkdef.pl which has been enhanced with
6162 better and easier to understand logic to choose which symbols should
6163 go into the Windows .def files as well as a number of fixes and code
6164 cleanup (among others, algorithm keywords are now sorted
6165 lexicographically to avoid constant rewrites).
6168 *) In BN_div() keep a copy of the sign of 'num' before writing the
6169 result to 'rm' because if rm==num the value will be overwritten
6170 and produce the wrong result if 'num' is negative: this caused
6171 problems with BN_mod() and BN_nnmod().
6174 *) Function OCSP_request_verify(). This checks the signature on an
6175 OCSP request and verifies the signer certificate. The signer
6176 certificate is just checked for a generic purpose and OCSP request
6180 *) Add OCSP_check_validity() function to check the validity of OCSP
6181 responses. OCSP responses are prepared in real time and may only
6182 be a few seconds old. Simply checking that the current time lies
6183 between thisUpdate and nextUpdate max reject otherwise valid responses
6184 caused by either OCSP responder or client clock inaccuracy. Instead
6185 we allow thisUpdate and nextUpdate to fall within a certain period of
6186 the current time. The age of the response can also optionally be
6187 checked. Two new options -validity_period and -status_age added to
6191 *) If signature or public key algorithm is unrecognized print out its
6192 OID rather that just UNKNOWN.
6195 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6196 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6197 ID to be generated from the issuer certificate alone which can then be
6198 passed to OCSP_id_issuer_cmp().
6201 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6202 ASN1 modules to export functions returning ASN1_ITEM pointers
6203 instead of the ASN1_ITEM structures themselves. This adds several
6204 new macros which allow the underlying ASN1 function/structure to
6205 be accessed transparently. As a result code should not use ASN1_ITEM
6206 references directly (such as &X509_it) but instead use the relevant
6207 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6208 use of the new ASN1 code on platforms where exporting structures
6209 is problematical (for example in shared libraries) but exporting
6210 functions returning pointers to structures is not.
6213 *) Add support for overriding the generation of SSL/TLS session IDs.
6214 These callbacks can be registered either in an SSL_CTX or per SSL.
6215 The purpose of this is to allow applications to control, if they wish,
6216 the arbitrary values chosen for use as session IDs, particularly as it
6217 can be useful for session caching in multiple-server environments. A
6218 command-line switch for testing this (and any client code that wishes
6219 to use such a feature) has been added to "s_server".
6220 [Geoff Thorpe, Lutz Jaenicke]
6222 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6223 of the form '#if defined(...) || defined(...) || ...' and
6224 '#if !defined(...) && !defined(...) && ...'. This also avoids
6225 the growing number of special cases it was previously handling.
6228 *) Make all configuration macros available for application by making
6229 sure they are available in opensslconf.h, by giving them names starting
6230 with "OPENSSL_" to avoid conflicts with other packages and by making
6231 sure e_os2.h will cover all platform-specific cases together with
6233 Additionally, it is now possible to define configuration/platform-
6234 specific names (called "system identities"). In the C code, these
6235 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6236 macro with the name beginning with "OPENSSL_SYS_", which is determined
6237 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6241 *) New option -set_serial to 'req' and 'x509' this allows the serial
6242 number to use to be specified on the command line. Previously self
6243 signed certificates were hard coded with serial number 0 and the
6244 CA options of 'x509' had to use a serial number in a file which was
6248 *) New options to 'ca' utility to support V2 CRL entry extensions.
6249 Currently CRL reason, invalidity date and hold instruction are
6250 supported. Add new CRL extensions to V3 code and some new objects.
6253 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6254 disable standard block padding (aka PKCS#5 padding) in the EVP
6255 API, which was previously mandatory. This means that the data is
6256 not padded in any way and so the total length much be a multiple
6257 of the block size, otherwise an error occurs.
6260 *) Initial (incomplete) OCSP SSL support.
6263 *) New function OCSP_parse_url(). This splits up a URL into its host,
6264 port and path components: primarily to parse OCSP URLs. New -url
6265 option to ocsp utility.
6268 *) New nonce behavior. The return value of OCSP_check_nonce() now
6269 reflects the various checks performed. Applications can decide
6270 whether to tolerate certain situations such as an absent nonce
6271 in a response when one was present in a request: the ocsp application
6272 just prints out a warning. New function OCSP_add1_basic_nonce()
6273 this is to allow responders to include a nonce in a response even if
6274 the request is nonce-less.
6277 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6278 skipped when using openssl x509 multiple times on a single input file,
6279 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6282 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6283 set string type: to handle setting ASN1_TIME structures. Fix ca
6284 utility to correctly initialize revocation date of CRLs.
6287 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6288 the clients preferred ciphersuites and rather use its own preferences.
6289 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6290 Internet Explorer by ensuring unchanged hash method during stepup.
6291 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6294 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6295 to aes and add a new 'exist' option to print out symbols that don't
6299 *) Additional options to ocsp utility to allow flags to be set and
6300 additional certificates supplied.
6303 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6304 OCSP client a number of certificate to only verify the response
6308 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6309 handle the new API. Currently only ECB, CBC modes supported. Add new
6312 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6313 Encryption Standard (AES) Ciphersuites for Transport Layer
6314 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6315 not enabled by default and were not part of the "ALL" ciphersuite
6316 alias because they were not yet official; they could be
6317 explicitly requested by specifying the "AESdraft" ciphersuite
6318 group alias. In the final release of OpenSSL 0.9.7, the group
6319 alias is called "AES" and is part of "ALL".)
6320 [Ben Laurie, Steve Henson, Bodo Moeller]
6322 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6323 request to response.
6326 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6327 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6328 extract information from a certificate request. OCSP_response_create()
6329 creates a response and optionally adds a basic response structure.
6330 OCSP_basic_add1_status() adds a complete single response to a basic
6331 response and returns the OCSP_SINGLERESP structure just added (to allow
6332 extensions to be included for example). OCSP_basic_add1_cert() adds a
6333 certificate to a basic response and OCSP_basic_sign() signs a basic
6334 response with various flags. New helper functions ASN1_TIME_check()
6335 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6336 (converts ASN1_TIME to GeneralizedTime).
6339 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6340 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6341 structure from a certificate. X509_pubkey_digest() digests the public_key
6342 contents: this is used in various key identifiers.
6345 *) Make sk_sort() tolerate a NULL argument.
6346 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6348 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6349 passed by the function are trusted implicitly. If any of them signed the
6350 response then it is assumed to be valid and is not verified.
6353 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6354 to data. This was previously part of the PKCS7 ASN1 code. This
6355 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6356 [Steve Henson, reported by Kenneth R. Robinette
6357 <support@securenetterm.com>]
6359 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6360 routines: without these tracing memory leaks is very painful.
6361 Fix leaks in PKCS12 and PKCS7 routines.
6364 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6365 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6366 effectively meant GeneralizedTime would never be used. Now it
6367 is initialised to -1 but X509_time_adj() now has to check the value
6368 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6369 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6370 [Steve Henson, reported by Kenneth R. Robinette
6371 <support@securenetterm.com>]
6373 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6374 result in a zero length in the ASN1_INTEGER structure which was
6375 not consistent with the structure when d2i_ASN1_INTEGER() was used
6376 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6377 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6378 where it did not print out a minus for negative ASN1_INTEGER.
6381 *) Add summary printout to ocsp utility. The various functions which
6382 convert status values to strings have been renamed to:
6383 OCSP_response_status_str(), OCSP_cert_status_str() and
6384 OCSP_crl_reason_str() and are no longer static. New options
6385 to verify nonce values and to disable verification. OCSP response
6386 printout format cleaned up.
6389 *) Add additional OCSP certificate checks. These are those specified
6390 in RFC2560. This consists of two separate checks: the CA of the
6391 certificate being checked must either be the OCSP signer certificate
6392 or the issuer of the OCSP signer certificate. In the latter case the
6393 OCSP signer certificate must contain the OCSP signing extended key
6394 usage. This check is performed by attempting to match the OCSP
6395 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6396 in the OCSP_CERTID structures of the response.
6399 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6400 and related routines. This uses the standard OpenSSL certificate
6401 verify routines to perform initial checks (just CA validity) and
6402 to obtain the certificate chain. Then additional checks will be
6403 performed on the chain. Currently the root CA is checked to see
6404 if it is explicitly trusted for OCSP signing. This is used to set
6405 a root CA as a global signing root: that is any certificate that
6406 chains to that CA is an acceptable OCSP signing certificate.
6409 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6410 extensions from a separate configuration file.
6411 As when reading extensions from the main configuration file,
6412 the '-extensions ...' option may be used for specifying the
6414 [Massimiliano Pala <madwolf@comune.modena.it>]
6416 *) New OCSP utility. Allows OCSP requests to be generated or
6417 read. The request can be sent to a responder and the output
6418 parsed, outputed or printed in text form. Not complete yet:
6419 still needs to check the OCSP response validity.
6422 *) New subcommands for 'openssl ca':
6423 'openssl ca -status <serial>' prints the status of the cert with
6424 the given serial number (according to the index file).
6425 'openssl ca -updatedb' updates the expiry status of certificates
6427 [Massimiliano Pala <madwolf@comune.modena.it>]
6429 *) New '-newreq-nodes' command option to CA.pl. This is like
6430 '-newreq', but calls 'openssl req' with the '-nodes' option
6431 so that the resulting key is not encrypted.
6432 [Damien Miller <djm@mindrot.org>]
6434 *) New configuration for the GNU Hurd.
6435 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6437 *) Initial code to implement OCSP basic response verify. This
6438 is currently incomplete. Currently just finds the signer's
6439 certificate and verifies the signature on the response.
6442 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6443 value of OPENSSLDIR. This is available via the new '-d' option
6444 to 'openssl version', and is also included in 'openssl version -a'.
6447 *) Allowing defining memory allocation callbacks that will be given
6448 file name and line number information in additional arguments
6449 (a const char* and an int). The basic functionality remains, as
6450 well as the original possibility to just replace malloc(),
6451 realloc() and free() by functions that do not know about these
6452 additional arguments. To register and find out the current
6453 settings for extended allocation functions, the following
6454 functions are provided:
6456 CRYPTO_set_mem_ex_functions
6457 CRYPTO_set_locked_mem_ex_functions
6458 CRYPTO_get_mem_ex_functions
6459 CRYPTO_get_locked_mem_ex_functions
6461 These work the same way as CRYPTO_set_mem_functions and friends.
6462 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6463 extended allocation function is enabled.
6464 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6465 a conventional allocation function is enabled.
6466 [Richard Levitte, Bodo Moeller]
6468 *) Finish off removing the remaining LHASH function pointer casts.
6469 There should no longer be any prototype-casting required when using
6470 the LHASH abstraction, and any casts that remain are "bugs". See
6471 the callback types and macros at the head of lhash.h for details
6472 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6475 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6476 If /dev/[u]random devices are not available or do not return enough
6477 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6479 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6480 /etc/entropy will be queried once each in this sequence, quering stops
6481 when enough entropy was collected without querying more sockets.
6484 *) Change the Unix RAND_poll() variant to be able to poll several
6485 random devices, as specified by DEVRANDOM, until a sufficient amount
6486 of data has been collected. We spend at most 10 ms on each file
6487 (select timeout) and read in non-blocking mode. DEVRANDOM now
6488 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6489 (previously it was just the string "/dev/urandom"), so on typical
6490 platforms the 10 ms delay will never occur.
6491 Also separate out the Unix variant to its own file, rand_unix.c.
6492 For VMS, there's a currently-empty rand_vms.c.
6495 *) Move OCSP client related routines to ocsp_cl.c. These
6496 provide utility functions which an application needing
6497 to issue a request to an OCSP responder and analyse the
6498 response will typically need: as opposed to those which an
6499 OCSP responder itself would need which will be added later.
6501 OCSP_request_sign() signs an OCSP request with an API similar
6502 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6503 response. OCSP_response_get1_basic() extracts basic response
6504 from response. OCSP_resp_find_status(): finds and extracts status
6505 information from an OCSP_CERTID structure (which will be created
6506 when the request structure is built). These are built from lower
6507 level functions which work on OCSP_SINGLERESP structures but
6508 wont normally be used unless the application wishes to examine
6509 extensions in the OCSP response for example.
6511 Replace nonce routines with a pair of functions.
6512 OCSP_request_add1_nonce() adds a nonce value and optionally
6513 generates a random value. OCSP_check_nonce() checks the
6514 validity of the nonce in an OCSP response.
6517 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6518 This doesn't copy the supplied OCSP_CERTID and avoids the
6519 need to free up the newly created id. Change return type
6520 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6521 This can then be used to add extensions to the request.
6522 Deleted OCSP_request_new(), since most of its functionality
6523 is now in OCSP_REQUEST_new() (and the case insensitive name
6524 clash) apart from the ability to set the request name which
6525 will be added elsewhere.
6528 *) Update OCSP API. Remove obsolete extensions argument from
6529 various functions. Extensions are now handled using the new
6530 OCSP extension code. New simple OCSP HTTP function which
6531 can be used to send requests and parse the response.
6534 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6535 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6536 uses the special reorder version of SET OF to sort the attributes
6537 and reorder them to match the encoded order. This resolves a long
6538 standing problem: a verify on a PKCS7 structure just after signing
6539 it used to fail because the attribute order did not match the
6540 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6541 it uses the received order. This is necessary to tolerate some broken
6542 software that does not order SET OF. This is handled by encoding
6543 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6544 to produce the required SET OF.
6547 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6548 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6549 files to get correct declarations of the ASN.1 item variables.
6552 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6553 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6554 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6555 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6556 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6557 ASN1_ITEM and no wrapper functions.
6560 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6561 replace the old function pointer based I/O routines. Change most of
6562 the *_d2i_bio() and *_d2i_fp() functions to use these.
6565 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6566 lines, recognice more "algorithms" that can be deselected, and make
6567 it complain about algorithm deselection that isn't recognised.
6570 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6571 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6572 to use new functions. Add NO_ASN1_OLD which can be set to remove
6573 some old style ASN1 functions: this can be used to determine if old
6574 code will still work when these eventually go away.
6577 *) New extension functions for OCSP structures, these follow the
6578 same conventions as certificates and CRLs.
6581 *) New function X509V3_add1_i2d(). This automatically encodes and
6582 adds an extension. Its behaviour can be customised with various
6583 flags to append, replace or delete. Various wrappers added for
6584 certifcates and CRLs.
6587 *) Fix to avoid calling the underlying ASN1 print routine when
6588 an extension cannot be parsed. Correct a typo in the
6589 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6592 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6593 entries for variables.
6596 *) Add functionality to apps/openssl.c for detecting locking
6597 problems: As the program is single-threaded, all we have
6598 to do is register a locking callback using an array for
6599 storing which locks are currently held by the program.
6602 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6603 SSL_get_ex_data_X509_STORE_idx(), which is used in
6604 ssl_verify_cert_chain() and thus can be called at any time
6605 during TLS/SSL handshakes so that thread-safety is essential.
6606 Unfortunately, the ex_data design is not at all suited
6607 for multi-threaded use, so it probably should be abolished.
6610 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6611 [Broadcom, tweaked and integrated by Geoff Thorpe]
6613 *) Move common extension printing code to new function
6614 X509V3_print_extensions(). Reorganise OCSP print routines and
6615 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6618 *) New function X509_signature_print() to remove duplication in some
6622 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6623 set (this was treated exactly the same as SET OF previously). This
6624 is used to reorder the STACK representing the structure to match the
6625 encoding. This will be used to get round a problem where a PKCS7
6626 structure which was signed could not be verified because the STACK
6627 order did not reflect the encoded order.
6630 *) Reimplement the OCSP ASN1 module using the new code.
6633 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6634 for its ASN1 operations. The old style function pointers still exist
6635 for now but they will eventually go away.
6638 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6639 completely replaces the old ASN1 functionality with a table driven
6640 encoder and decoder which interprets an ASN1_ITEM structure describing
6641 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6642 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6643 has also been converted to the new form.
6646 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6647 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6648 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6649 for negative moduli.
6652 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6653 of not touching the result's sign bit.
6656 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6660 *) Changed the LHASH code to use prototypes for callbacks, and created
6661 macros to declare and implement thin (optionally static) functions
6662 that provide type-safety and avoid function pointer casting for the
6663 type-specific callbacks.
6666 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6668 [Veers Staats <staatsvr@asc.hpc.mil>,
6669 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6671 *) Reformat the FAQ so the different questions and answers can be divided
6672 in sections depending on the subject.
6675 *) Have the zlib compression code load ZLIB.DLL dynamically under
6679 *) New function BN_mod_sqrt for computing square roots modulo a prime
6680 (using the probabilistic Tonelli-Shanks algorithm unless
6681 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6682 be handled deterministically).
6683 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6685 *) Make BN_mod_inverse faster by explicitly handling small quotients
6686 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6687 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6690 *) New function BN_kronecker.
6693 *) Fix BN_gcd so that it works on negative inputs; the result is
6694 positive unless both parameters are zero.
6695 Previously something reasonably close to an infinite loop was
6696 possible because numbers could be growing instead of shrinking
6697 in the implementation of Euclid's algorithm.
6700 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6701 sign of the number in question.
6703 Fix BN_is_word(a,w) to work correctly for w == 0.
6705 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6706 because its test if the absolute value of 'a' equals 'w'.
6707 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6708 it exists mostly for use in the implementations of BN_is_zero(),
6709 BN_is_one(), and BN_is_word().
6712 *) New function BN_swap.
6715 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6716 the exponentiation functions are more likely to produce reasonable
6717 results on negative inputs.
6720 *) Change BN_mod_mul so that the result is always non-negative.
6721 Previously, it could be negative if one of the factors was negative;
6722 I don't think anyone really wanted that behaviour.
6725 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6726 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6727 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6728 and add new functions:
6737 BN_mod_lshift1_quick
6741 These functions always generate non-negative results.
6743 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6744 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6746 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6747 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6748 be reduced modulo m.
6749 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6752 The following entry accidentally appeared in the CHANGES file
6753 distributed with OpenSSL 0.9.7. The modifications described in
6754 it do *not* apply to OpenSSL 0.9.7.
6756 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6757 was actually never needed) and in BN_mul(). The removal in BN_mul()
6758 required a small change in bn_mul_part_recursive() and the addition
6759 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6760 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6761 bn_sub_words() and bn_add_words() except they take arrays with
6766 *) In 'openssl passwd', verify passwords read from the terminal
6767 unless the '-salt' option is used (which usually means that
6768 verification would just waste user's time since the resulting
6769 hash is going to be compared with some given password hash)
6770 or the new '-noverify' option is used.
6772 This is an incompatible change, but it does not affect
6773 non-interactive use of 'openssl passwd' (passwords on the command
6774 line, '-stdin' option, '-in ...' option) and thus should not
6778 *) Remove all references to RSAref, since there's no more need for it.
6781 *) Make DSO load along a path given through an environment variable
6782 (SHLIB_PATH) with shl_load().
6785 *) Constify the ENGINE code as a result of BIGNUM constification.
6786 Also constify the RSA code and most things related to it. In a
6787 few places, most notable in the depth of the ASN.1 code, ugly
6788 casts back to non-const were required (to be solved at a later
6792 *) Make it so the openssl application has all engines loaded by default.
6795 *) Constify the BIGNUM routines a little more.
6798 *) Add the following functions:
6800 ENGINE_load_cswift()
6802 ENGINE_load_atalla()
6804 ENGINE_load_builtin_engines()
6806 That way, an application can itself choose if external engines that
6807 are built-in in OpenSSL shall ever be used or not. The benefit is
6808 that applications won't have to be linked with libdl or other dso
6809 libraries unless it's really needed.
6811 Changed 'openssl engine' to load all engines on demand.
6812 Changed the engine header files to avoid the duplication of some
6813 declarations (they differed!).
6816 *) 'openssl engine' can now list capabilities.
6819 *) Better error reporting in 'openssl engine'.
6822 *) Never call load_dh_param(NULL) in s_server.
6825 *) Add engine application. It can currently list engines by name and
6826 identity, and test if they are actually available.
6829 *) Improve RPM specification file by forcing symbolic linking and making
6830 sure the installed documentation is also owned by root.root.
6831 [Damien Miller <djm@mindrot.org>]
6833 *) Give the OpenSSL applications more possibilities to make use of
6834 keys (public as well as private) handled by engines.
6837 *) Add OCSP code that comes from CertCo.
6840 *) Add VMS support for the Rijndael code.
6843 *) Added untested support for Nuron crypto accelerator.
6846 *) Add support for external cryptographic devices. This code was
6847 previously distributed separately as the "engine" branch.
6848 [Geoff Thorpe, Richard Levitte]
6850 *) Rework the filename-translation in the DSO code. It is now possible to
6851 have far greater control over how a "name" is turned into a filename
6852 depending on the operating environment and any oddities about the
6853 different shared library filenames on each system.
6856 *) Support threads on FreeBSD-elf in Configure.
6859 *) Fix for SHA1 assembly problem with MASM: it produces
6860 warnings about corrupt line number information when assembling
6861 with debugging information. This is caused by the overlapping
6863 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6866 NCONF_get_number() has no error checking at all. As a replacement,
6867 NCONF_get_number_e() is defined (_e for "error checking") and is
6868 promoted strongly. The old NCONF_get_number is kept around for
6869 binary backward compatibility.
6870 Make it possible for methods to load from something other than a BIO,
6871 by providing a function pointer that is given a name instead of a BIO.
6872 For example, this could be used to load configuration data from an
6876 *) Fix for non blocking accept BIOs. Added new I/O special reason
6877 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6878 with non blocking I/O was not possible because no retry code was
6879 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6883 *) Added the beginnings of Rijndael support.
6886 *) Fix for bug in DirectoryString mask setting. Add support for
6887 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6888 to allow certificate printing to more controllable, additional
6889 'certopt' option to 'x509' to allow new printing options to be
6893 *) Clean old EAY MD5 hack from e_os.h.
6896 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6898 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6899 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6900 [Joe Orton, Steve Henson]
6902 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6904 *) Fix additional bug revealed by the NISCC test suite:
6906 Stop bug triggering large recursion when presented with
6907 certain ASN.1 tags (CVE-2003-0851)
6910 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6912 *) Fix various bugs revealed by running the NISCC test suite:
6914 Stop out of bounds reads in the ASN1 code when presented with
6915 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6917 If verify callback ignores invalid public key errors don't try to check
6918 certificate signature with the NULL public key.
6922 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6923 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6927 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6928 extra data after the compression methods not only for TLS 1.0
6929 but also for SSL 3.0 (as required by the specification).
6930 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6932 *) Change X509_certificate_type() to mark the key as exported/exportable
6933 when it's 512 *bits* long, not 512 bytes.
6936 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6938 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6939 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6940 a protocol version number mismatch like a decryption error
6941 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6944 *) Turn on RSA blinding by default in the default implementation
6945 to avoid a timing attack. Applications that don't want it can call
6946 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6947 They would be ill-advised to do so in most cases.
6948 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6950 *) Change RSA blinding code so that it works when the PRNG is not
6951 seeded (in this case, the secret RSA exponent is abused as
6952 an unpredictable seed -- if it is not unpredictable, there
6953 is no point in blinding anyway). Make RSA blinding thread-safe
6954 by remembering the creator's thread ID in rsa->blinding and
6955 having all other threads use local one-time blinding factors
6956 (this requires more computation than sharing rsa->blinding, but
6957 avoids excessive locking; and if an RSA object is not shared
6958 between threads, blinding will still be very fast).
6961 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6963 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6964 via timing by performing a MAC computation even if incorrrect
6965 block cipher padding has been found. This is a countermeasure
6966 against active attacks where the attacker has to distinguish
6967 between bad padding and a MAC verification error. (CVE-2003-0078)
6969 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6970 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6971 Martin Vuagnoux (EPFL, Ilion)]
6973 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6975 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6976 memory from it's contents. This is done with a counter that will
6977 place alternating values in each byte. This can be used to solve
6978 two issues: 1) the removal of calls to memset() by highly optimizing
6979 compilers, and 2) cleansing with other values than 0, since those can
6980 be read through on certain media, for example a swap space on disk.
6983 *) Bugfix: client side session caching did not work with external caching,
6984 because the session->cipher setting was not restored when reloading
6985 from the external cache. This problem was masked, when
6986 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6987 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6990 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6991 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6992 [Zeev Lieber <zeev-l@yahoo.com>]
6994 *) Undo an undocumented change introduced in 0.9.6e which caused
6995 repeated calls to OpenSSL_add_all_ciphers() and
6996 OpenSSL_add_all_digests() to be ignored, even after calling
7000 *) Change the default configuration reader to deal with last line not
7001 being properly terminated.
7004 *) Change X509_NAME_cmp() so it applies the special rules on handling
7005 DN values that are of type PrintableString, as well as RDNs of type
7006 emailAddress where the value has the type ia5String.
7007 [stefank@valicert.com via Richard Levitte]
7009 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7010 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7011 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7012 the bitwise-OR of the two for use by the majority of applications
7013 wanting this behaviour, and update the docs. The documented
7014 behaviour and actual behaviour were inconsistent and had been
7015 changing anyway, so this is more a bug-fix than a behavioural
7017 [Geoff Thorpe, diagnosed by Nadav Har'El]
7019 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7020 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7023 *) Fix initialization code race conditions in
7024 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7025 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7026 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7027 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7028 ssl2_get_cipher_by_char(),
7029 ssl3_get_cipher_by_char().
7030 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7032 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7033 the cached sessions are flushed, as the remove_cb() might use ex_data
7034 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7035 (see [openssl.org #212]).
7036 [Geoff Thorpe, Lutz Jaenicke]
7038 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7039 length, instead of the encoding length to d2i_ASN1_OBJECT.
7042 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7044 *) [In 0.9.6g-engine release:]
7045 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7046 [Lynn Gazis <lgazis@rainbow.com>]
7048 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7050 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7051 and get fix the header length calculation.
7052 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7053 Alon Kantor <alonk@checkpoint.com> (and others),
7056 *) Use proper error handling instead of 'assertions' in buffer
7057 overflow checks added in 0.9.6e. This prevents DoS (the
7058 assertions could call abort()).
7059 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7061 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7063 *) Add various sanity checks to asn1_get_length() to reject
7064 the ASN1 length bytes if they exceed sizeof(long), will appear
7065 negative or the content length exceeds the length of the
7067 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7069 *) Fix cipher selection routines: ciphers without encryption had no flags
7070 for the cipher strength set and where therefore not handled correctly
7071 by the selection routines (PR #130).
7074 *) Fix EVP_dsa_sha macro.
7078 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7079 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7080 that was added in OpenSSL 0.9.6d.
7082 As the countermeasure turned out to be incompatible with some
7083 broken SSL implementations, the new option is part of SSL_OP_ALL.
7084 SSL_OP_ALL is usually employed when compatibility with weird SSL
7085 implementations is desired (e.g. '-bugs' option to 's_client' and
7086 's_server'), so the new option is automatically set in many
7090 *) Changes in security patch:
7092 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7093 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7094 Air Force Materiel Command, USAF, under agreement number
7097 *) Add various sanity checks to asn1_get_length() to reject
7098 the ASN1 length bytes if they exceed sizeof(long), will appear
7099 negative or the content length exceeds the length of the
7100 supplied buffer. (CVE-2002-0659)
7101 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7103 *) Assertions for various potential buffer overflows, not known to
7105 [Ben Laurie (CHATS)]
7107 *) Various temporary buffers to hold ASCII versions of integers were
7108 too small for 64 bit platforms. (CVE-2002-0655)
7109 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7111 *) Remote buffer overflow in SSL3 protocol - an attacker could
7112 supply an oversized session ID to a client. (CVE-2002-0656)
7113 [Ben Laurie (CHATS)]
7115 *) Remote buffer overflow in SSL2 protocol - an attacker could
7116 supply an oversized client master key. (CVE-2002-0656)
7117 [Ben Laurie (CHATS)]
7119 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7121 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7122 encoded as NULL) with id-dsa-with-sha1.
7123 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7125 *) Check various X509_...() return values in apps/req.c.
7126 [Nils Larsch <nla@trustcenter.de>]
7128 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7129 an end-of-file condition would erronously be flagged, when the CRLF
7130 was just at the end of a processed block. The bug was discovered when
7131 processing data through a buffering memory BIO handing the data to a
7132 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7133 <ptsekov@syntrex.com> and Nedelcho Stanev.
7136 *) Implement a countermeasure against a vulnerability recently found
7137 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7138 before application data chunks to avoid the use of known IVs
7139 with data potentially chosen by the attacker.
7142 *) Fix length checks in ssl3_get_client_hello().
7145 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7146 to prevent ssl3_read_internal() from incorrectly assuming that
7147 ssl3_read_bytes() found application data while handshake
7148 processing was enabled when in fact s->s3->in_read_app_data was
7149 merely automatically cleared during the initial handshake.
7150 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7152 *) Fix object definitions for Private and Enterprise: they were not
7153 recognized in their shortname (=lowercase) representation. Extend
7154 obj_dat.pl to issue an error when using undefined keywords instead
7155 of silently ignoring the problem (Svenning Sorensen
7156 <sss@sss.dnsalias.net>).
7159 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7160 generators, i.e. generators other than 2 and 5. (Previously, the
7161 code did not properly initialise the 'add' and 'rem' values to
7162 BN_generate_prime().)
7164 In the new general case, we do not insist that 'generator' is
7165 actually a primitive root: This requirement is rather pointless;
7166 a generator of the order-q subgroup is just as good, if not
7170 *) Map new X509 verification errors to alerts. Discovered and submitted by
7171 Tom Wu <tom@arcot.com>.
7174 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7175 returning non-zero before the data has been completely received
7176 when using non-blocking I/O.
7177 [Bodo Moeller; problem pointed out by John Hughes]
7179 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7180 [Ben Laurie, Lutz Jaenicke]
7182 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7183 Yoram Zahavi <YoramZ@gilian.com>).
7186 *) Add information about CygWin 1.3 and on, and preserve proper
7187 configuration for the versions before that.
7188 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7190 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7191 check whether we deal with a copy of a session and do not delete from
7192 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7193 <izhar@checkpoint.com>.
7196 *) Do not store session data into the internal session cache, if it
7197 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7198 flag is set). Proposed by Aslam <aslam@funk.com>.
7201 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7205 *) [In 0.9.6d-engine release:]
7206 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7207 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7209 *) Add the configuration target linux-s390x.
7210 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7212 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7213 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7214 variable as an indication that a ClientHello message has been
7215 received. As the flag value will be lost between multiple
7216 invocations of ssl3_accept when using non-blocking I/O, the
7217 function may not be aware that a handshake has actually taken
7218 place, thus preventing a new session from being added to the
7221 To avoid this problem, we now set s->new_session to 2 instead of
7222 using a local variable.
7223 [Lutz Jaenicke, Bodo Moeller]
7225 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7226 if the SSL_R_LENGTH_MISMATCH error is detected.
7227 [Geoff Thorpe, Bodo Moeller]
7229 *) New 'shared_ldflag' column in Configure platform table.
7232 *) Fix EVP_CIPHER_mode macro.
7233 ["Dan S. Camper" <dan@bti.net>]
7235 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7236 type, we must throw them away by setting rr->length to 0.
7237 [D P Chang <dpc@qualys.com>]
7239 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7241 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7242 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7243 worked incorrectly for those cases where range = 10..._2 and
7244 3*range is two bits longer than range.)
7247 *) Only add signing time to PKCS7 structures if it is not already
7251 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7252 OBJ_ld_ce should be OBJ_id_ce.
7253 Also some ip-pda OIDs in crypto/objects/objects.txt were
7254 incorrect (cf. RFC 3039).
7255 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7257 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7258 returns early because it has nothing to do.
7259 [Andy Schneider <andy.schneider@bjss.co.uk>]
7261 *) [In 0.9.6c-engine release:]
7262 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7263 [Andy Schneider <andy.schneider@bjss.co.uk>]
7265 *) [In 0.9.6c-engine release:]
7266 Add support for Cryptographic Appliance's keyserver technology.
7267 (Use engine 'keyclient')
7268 [Cryptographic Appliances and Geoff Thorpe]
7270 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7271 is called via tools/c89.sh because arguments have to be
7272 rearranged (all '-L' options must appear before the first object
7274 [Richard Shapiro <rshapiro@abinitio.com>]
7276 *) [In 0.9.6c-engine release:]
7277 Add support for Broadcom crypto accelerator cards, backported
7279 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7281 *) [In 0.9.6c-engine release:]
7282 Add support for SureWare crypto accelerator cards from
7283 Baltimore Technologies. (Use engine 'sureware')
7284 [Baltimore Technologies and Mark Cox]
7286 *) [In 0.9.6c-engine release:]
7287 Add support for crypto accelerator cards from Accelerated
7288 Encryption Processing, www.aep.ie. (Use engine 'aep')
7289 [AEP Inc. and Mark Cox]
7291 *) Add a configuration entry for gcc on UnixWare.
7292 [Gary Benson <gbenson@redhat.com>]
7294 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7295 messages are stored in a single piece (fixed-length part and
7296 variable-length part combined) and fix various bugs found on the way.
7299 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7300 instead. BIO_gethostbyname() does not know what timeouts are
7301 appropriate, so entries would stay in cache even when they have
7303 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7305 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7306 faced with a pathologically small ClientHello fragment that does
7307 not contain client_version: Instead of aborting with an error,
7308 simply choose the highest available protocol version (i.e.,
7309 TLS 1.0 unless it is disabled). In practice, ClientHello
7310 messages are never sent like this, but this change gives us
7311 strictly correct behaviour at least for TLS.
7314 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7315 never resets s->method to s->ctx->method when called from within
7316 one of the SSL handshake functions.
7317 [Bodo Moeller; problem pointed out by Niko Baric]
7319 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7320 (sent using the client's version number) if client_version is
7321 smaller than the protocol version in use. Also change
7322 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7323 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7324 the client will at least see that alert.
7327 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7331 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7332 client receives HelloRequest while in a handshake.
7333 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7335 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7336 should end in 'break', not 'goto end' which circumvents various
7337 cleanups done in state SSL_ST_OK. But session related stuff
7338 must be disabled for SSL_ST_OK in the case that we just sent a
7341 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7342 before just sending a HelloRequest.
7343 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7345 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7346 reveal whether illegal block cipher padding was found or a MAC
7347 verification error occured. (Neither SSLerr() codes nor alerts
7348 are directly visible to potential attackers, but the information
7349 may leak via logfiles.)
7351 Similar changes are not required for the SSL 2.0 implementation
7352 because the number of padding bytes is sent in clear for SSL 2.0,
7353 and the extra bytes are just ignored. However ssl/s2_pkt.c
7354 failed to verify that the purported number of padding bytes is in
7358 *) Add OpenUNIX-8 support including shared libraries
7359 (Boyd Lynn Gerber <gerberb@zenez.com>).
7362 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7363 'wristwatch attack' using huge encoding parameters (cf.
7364 James H. Manger's CRYPTO 2001 paper). Note that the
7365 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7366 encoding parameters and hence was not vulnerable.
7369 *) BN_sqr() bug fix.
7370 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7372 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7373 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7374 followed by modular reduction.
7375 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7377 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7378 equivalent based on BN_pseudo_rand() instead of BN_rand().
7381 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7382 This function was broken, as the check for a new client hello message
7383 to handle SGC did not allow these large messages.
7384 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7387 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7390 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7391 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7394 *) Rework the configuration and shared library support for Tru64 Unix.
7395 The configuration part makes use of modern compiler features and
7396 still retains old compiler behavior for those that run older versions
7397 of the OS. The shared library support part includes a variant that
7398 uses the RPATH feature, and is available through the special
7399 configuration target "alpha-cc-rpath", which will never be selected
7401 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7403 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7404 with the same message size as in ssl3_get_certificate_request().
7405 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7406 messages might inadvertently be reject as too long.
7407 [Petr Lampa <lampa@fee.vutbr.cz>]
7409 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7412 *) Modified SSL library such that the verify_callback that has been set
7413 specificly for an SSL object with SSL_set_verify() is actually being
7414 used. Before the change, a verify_callback set with this function was
7415 ignored and the verify_callback() set in the SSL_CTX at the time of
7416 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7417 to allow the necessary settings.
7420 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7421 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7422 done automatically (in contradiction to the requirements of the C
7423 standard). This made problems when used from OpenSSH.
7426 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7427 dh->length and always used
7429 BN_rand_range(priv_key, dh->p).
7431 BN_rand_range() is not necessary for Diffie-Hellman, and this
7432 specific range makes Diffie-Hellman unnecessarily inefficient if
7433 dh->length (recommended exponent length) is much smaller than the
7434 length of dh->p. We could use BN_rand_range() if the order of
7435 the subgroup was stored in the DH structure, but we only have
7440 BN_rand(priv_key, l, ...)
7442 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7448 RSA_eay_public_encrypt
7449 RSA_eay_private_decrypt
7450 RSA_eay_private_encrypt (signing)
7451 RSA_eay_public_decrypt (signature verification)
7453 (default implementations for RSA_public_encrypt,
7454 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7455 always reject numbers >= n.
7458 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7459 to synchronize access to 'locking_thread'. This is necessary on
7460 systems where access to 'locking_thread' (an 'unsigned long'
7461 variable) is not atomic.
7464 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7465 *before* setting the 'crypto_lock_rand' flag. The previous code had
7466 a race condition if 0 is a valid thread ID.
7467 [Travis Vitek <vitek@roguewave.com>]
7469 *) Add support for shared libraries under Irix.
7470 [Albert Chin-A-Young <china@thewrittenword.com>]
7472 *) Add configuration option to build on Linux on both big-endian and
7474 [Ralf Baechle <ralf@uni-koblenz.de>]
7476 *) Add the possibility to create shared libraries on HP-UX.
7479 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7481 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7482 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7483 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7484 PRNG state recovery was possible based on the output of
7485 one PRNG request appropriately sized to gain knowledge on
7486 'md' followed by enough consecutive 1-byte PRNG requests
7487 to traverse all of 'state'.
7489 1. When updating 'md_local' (the current thread's copy of 'md')
7490 during PRNG output generation, hash all of the previous
7491 'md_local' value, not just the half used for PRNG output.
7493 2. Make the number of bytes from 'state' included into the hash
7494 independent from the number of PRNG bytes requested.
7496 The first measure alone would be sufficient to avoid
7497 Markku-Juhani's attack. (Actually it had never occurred
7498 to me that the half of 'md_local' used for chaining was the
7499 half from which PRNG output bytes were taken -- I had always
7500 assumed that the secret half would be used.) The second
7501 measure makes sure that additional data from 'state' is never
7502 mixed into 'md_local' in small portions; this heuristically
7503 further strengthens the PRNG.
7506 *) Fix crypto/bn/asm/mips3.s.
7509 *) When only the key is given to "enc", the IV is undefined. Print out
7510 an error message in this case.
7513 *) Handle special case when X509_NAME is empty in X509 printing routines.
7516 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7517 positive and less than q.
7520 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7521 used: it isn't thread safe and the add_lock_callback should handle
7523 [Paul Rose <Paul.Rose@bridge.com>]
7525 *) Verify that incoming data obeys the block size in
7526 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7530 [Ulf Möller, Bodo Möller]
7532 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7533 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7534 when fixing the server behaviour for backwards-compatible 'client
7535 hello' messages. (Note that the attack is impractical against
7536 SSL 3.0 and TLS 1.0 anyway because length and version checking
7537 means that the probability of guessing a valid ciphertext is
7538 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7541 Before 0.9.5, the countermeasure (hide the error by generating a
7542 random 'decryption result') did not work properly because
7543 ERR_clear_error() was missing, meaning that SSL_get_error() would
7544 detect the supposedly ignored error.
7546 Both problems are now fixed.
7549 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7550 (previously it was 1024).
7553 *) Fix for compatibility mode trust settings: ignore trust settings
7554 unless some valid trust or reject settings are present.
7557 *) Fix for blowfish EVP: its a variable length cipher.
7560 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7561 parameters in DSA public key structures and return an error in the
7562 DSA routines if parameters are absent.
7565 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7566 in the current directory if neither $RANDFILE nor $HOME was set.
7567 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7568 caused some confusion to Windows users who haven't defined $HOME.
7569 Thus RAND_file_name() is changed again: e_os.h can define a
7570 DEFAULT_HOME, which will be used if $HOME is not set.
7571 For Windows, we use "C:"; on other platforms, we still require
7572 environment variables.
7574 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7575 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7576 having multiple threads call RAND_poll() concurrently.
7579 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7580 combination of a flag and a thread ID variable.
7581 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7582 flag), *other* threads can enter ssleay_add_bytes without obeying
7583 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7584 that they do not hold after the first thread unsets add_do_not_lock).
7587 *) Change bctest again: '-x' expressions are not available in all
7591 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7593 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7594 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7596 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7597 the default extension for executables, if any. Also, make the perl
7598 scripts that use symlink() to test if it really exists and use "cp"
7599 if it doesn't. All this made OpenSSL compilable and installable in
7603 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7604 If SEQUENCE is length is indefinite just set c->slen to the total
7605 amount of data available.
7606 [Steve Henson, reported by shige@FreeBSD.org]
7607 [This change does not apply to 0.9.7.]
7609 *) Change bctest to avoid here-documents inside command substitution
7610 (workaround for FreeBSD /bin/sh bug).
7611 For compatibility with Ultrix, avoid shell functions (introduced
7612 in the bctest version that searches along $PATH).
7615 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7616 with des_encrypt() defined on some operating systems, like Solaris
7620 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7621 On the Importance of Eliminating Errors in Cryptographic
7622 Computations, J. Cryptology 14 (2001) 2, 101-119,
7623 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7626 *) MIPS assembler BIGNUM division bug fix.
7629 *) Disabled incorrect Alpha assembler code.
7632 *) Fix PKCS#7 decode routines so they correctly update the length
7633 after reading an EOC for the EXPLICIT tag.
7635 [This change does not apply to 0.9.7.]
7637 *) Fix bug in PKCS#12 key generation routines. This was triggered
7638 if a 3DES key was generated with a 0 initial byte. Include
7639 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7640 (but broken) behaviour.
7643 *) Enhance bctest to search for a working bc along $PATH and print
7645 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7647 *) Fix memory leaks in err.c: free err_data string if necessary;
7648 don't write to the wrong index in ERR_set_error_data.
7651 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7655 *) Replace rdtsc with _emit statements for VC++ version 5.
7656 [Jeremy Cooper <jeremy@baymoo.org>]
7658 *) Make it possible to reuse SSLv2 sessions.
7661 *) In copy_email() check for >= 0 as a return value for
7662 X509_NAME_get_index_by_NID() since 0 is a valid index.
7663 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7665 *) Avoid coredump with unsupported or invalid public keys by checking if
7666 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7667 PKCS7_verify() fails with non detached data.
7670 *) Don't use getenv in library functions when run as setuid/setgid.
7671 New function OPENSSL_issetugid().
7674 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7675 due to incorrect handling of multi-threading:
7677 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7679 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7681 3. Count how many times MemCheck_off() has been called so that
7682 nested use can be treated correctly. This also avoids
7683 inband-signalling in the previous code (which relied on the
7684 assumption that thread ID 0 is impossible).
7687 *) Add "-rand" option also to s_client and s_server.
7690 *) Fix CPU detection on Irix 6.x.
7691 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7692 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7694 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7697 [This change does not apply to 0.9.7.]
7699 *) Use the cached encoding of an X509_NAME structure rather than
7700 copying it. This is apparently the reason for the libsafe "errors"
7701 but the code is actually correct.
7704 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7705 Bleichenbacher's DSA attack.
7706 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7707 to be set and top=0 forces the highest bit to be set; top=-1 is new
7708 and leaves the highest bit random.
7709 [Ulf Moeller, Bodo Moeller]
7711 *) In the NCONF_...-based implementations for CONF_... queries
7712 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7713 a temporary CONF structure with the data component set to NULL
7714 (which gives segmentation faults in lh_retrieve).
7715 Instead, use NULL for the CONF pointer in CONF_get_string and
7716 CONF_get_number (which may use environment variables) and directly
7717 return NULL from CONF_get_section.
7720 *) Fix potential buffer overrun for EBCDIC.
7723 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7724 keyUsage if basicConstraints absent for a CA.
7727 *) Make SMIME_write_PKCS7() write mail header values with a format that
7728 is more generally accepted (no spaces before the semicolon), since
7729 some programs can't parse those values properly otherwise. Also make
7730 sure BIO's that break lines after each write do not create invalid
7734 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7735 macros previously used would not encode an empty SEQUENCE OF
7736 and break the signature.
7738 [This change does not apply to 0.9.7.]
7740 *) Zero the premaster secret after deriving the master secret in
7744 *) Add some EVP_add_digest_alias registrations (as found in
7745 OpenSSL_add_all_digests()) to SSL_library_init()
7746 aka OpenSSL_add_ssl_algorithms(). This provides improved
7747 compatibility with peers using X.509 certificates
7748 with unconventional AlgorithmIdentifier OIDs.
7751 *) Fix for Irix with NO_ASM.
7752 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7754 *) ./config script fixes.
7755 [Ulf Moeller, Richard Levitte]
7757 *) Fix 'openssl passwd -1'.
7760 *) Change PKCS12_key_gen_asc() so it can cope with non null
7761 terminated strings whose length is passed in the passlen
7762 parameter, for example from PEM callbacks. This was done
7763 by adding an extra length parameter to asc2uni().
7764 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7766 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7767 call failed, free the DSA structure.
7770 *) Fix to uni2asc() to cope with zero length Unicode strings.
7771 These are present in some PKCS#12 files.
7774 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7775 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7776 when writing a 32767 byte record.
7777 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7779 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7780 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7782 (RSA objects have a reference count access to which is protected
7783 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7784 so they are meant to be shared between threads.)
7785 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7786 "Reddie, Steven" <Steven.Reddie@ca.com>]
7788 *) Fix a deadlock in CRYPTO_mem_leaks().
7791 *) Use better test patterns in bntest.
7794 *) rand_win.c fix for Borland C.
7797 *) BN_rshift bugfix for n == 0.
7800 *) Add a 'bctest' script that checks for some known 'bc' bugs
7801 so that 'make test' does not abort just because 'bc' is broken.
7804 *) Store verify_result within SSL_SESSION also for client side to
7805 avoid potential security hole. (Re-used sessions on the client side
7806 always resulted in verify_result==X509_V_OK, not using the original
7807 result of the server certificate verification.)
7810 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7811 SSL3_RT_APPLICATION_DATA, return 0.
7812 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7816 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7817 releases, have been re-implemented by renaming the previous
7818 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7819 and ssl3_read_internal, respectively, and adding 'peek' parameters
7820 to them. The new ssl[23]_{read,peek} functions are calls to
7821 ssl[23]_read_internal with the 'peek' flag set appropriately.
7822 A 'peek' parameter has also been added to ssl3_read_bytes, which
7823 does the actual work for ssl3_read_internal.
7826 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7827 the method-specific "init()" handler. Also clean up ex_data after
7828 calling the method-specific "finish()" handler. Previously, this was
7829 happening the other way round.
7832 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7833 The previous value, 12, was not always sufficient for BN_mod_exp().
7836 *) Make sure that shared libraries get the internal name engine with
7837 the full version number and not just 0. This should mark the
7838 shared libraries as not backward compatible. Of course, this should
7839 be changed again when we can guarantee backward binary compatibility.
7842 *) Fix typo in get_cert_by_subject() in by_dir.c
7843 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7845 *) Rework the system to generate shared libraries:
7847 - Make note of the expected extension for the shared libraries and
7848 if there is a need for symbolic links from for example libcrypto.so.0
7849 to libcrypto.so.0.9.7. There is extended info in Configure for
7852 - Make as few rebuilds of the shared libraries as possible.
7854 - Still avoid linking the OpenSSL programs with the shared libraries.
7856 - When installing, install the shared libraries separately from the
7860 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7862 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7863 and not in SSL_clear because the latter is also used by the
7864 accept/connect functions; previously, the settings made by
7865 SSL_set_read_ahead would be lost during the handshake.
7866 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7868 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7869 Previously, it would create entries for disabled algorithms no
7873 *) Added several new manual pages for SSL_* function.
7876 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7878 *) In ssl23_get_client_hello, generate an error message when faced
7879 with an initial SSL 3.0/TLS record that is too small to contain the
7880 first two bytes of the ClientHello message, i.e. client_version.
7881 (Note that this is a pathologic case that probably has never happened
7882 in real life.) The previous approach was to use the version number
7883 from the record header as a substitute; but our protocol choice
7884 should not depend on that one because it is not authenticated
7885 by the Finished messages.
7888 *) More robust randomness gathering functions for Windows.
7889 [Jeffrey Altman <jaltman@columbia.edu>]
7891 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7892 not set then we don't setup the error code for issuer check errors
7893 to avoid possibly overwriting other errors which the callback does
7894 handle. If an application does set the flag then we assume it knows
7895 what it is doing and can handle the new informational codes
7899 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7900 a general "ANY" type, as such it should be able to decode anything
7901 including tagged types. However it didn't check the class so it would
7902 wrongly interpret tagged types in the same way as their universal
7903 counterpart and unknown types were just rejected. Changed so that the
7904 tagged and unknown types are handled in the same way as a SEQUENCE:
7905 that is the encoding is stored intact. There is also a new type
7906 "V_ASN1_OTHER" which is used when the class is not universal, in this
7907 case we have no idea what the actual type is so we just lump them all
7911 *) On VMS, stdout may very well lead to a file that is written to
7912 in a record-oriented fashion. That means that every write() will
7913 write a separate record, which will be read separately by the
7914 programs trying to read from it. This can be very confusing.
7916 The solution is to put a BIO filter in the way that will buffer
7917 text until a linefeed is reached, and then write everything a
7918 line at a time, so every record written will be an actual line,
7919 not chunks of lines and not (usually doesn't happen, but I've
7920 seen it once) several lines in one record. BIO_f_linebuffer() is
7923 Currently, it's a VMS-only method, because that's where it has
7924 been tested well enough.
7927 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7928 it can return incorrect results.
7929 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7930 but it was in 0.9.6-beta[12].)
7933 *) Disable the check for content being present when verifying detached
7934 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7935 include zero length content when signing messages.
7938 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7939 BIO_ctrl (for BIO pairs).
7942 *) Add DSO method for VMS.
7945 *) Bug fix: Montgomery multiplication could produce results with the
7949 *) Add RPM specification openssl.spec and modify it to build three
7950 packages. The default package contains applications, application
7951 documentation and run-time libraries. The devel package contains
7952 include files, static libraries and function documentation. The
7953 doc package contains the contents of the doc directory. The original
7954 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7957 *) Add a large number of documentation files for many SSL routines.
7958 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7960 *) Add a configuration entry for Sony News 4.
7961 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7963 *) Don't set the two most significant bits to one when generating a
7964 random number < q in the DSA library.
7967 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7968 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7969 the underlying transport is blocking) if a handshake took place.
7970 (The default behaviour is needed by applications such as s_client
7971 and s_server that use select() to determine when to use SSL_read;
7972 but for applications that know in advance when to expect data, it
7973 just makes things more complicated.)
7976 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7980 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7981 work better on such systems.
7982 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7984 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7985 Update PKCS12_parse() so it copies the friendlyName and the
7986 keyid to the certificates aux info.
7989 *) Fix bug in PKCS7_verify() which caused an infinite loop
7990 if there was more than one signature.
7991 [Sven Uszpelkat <su@celocom.de>]
7993 *) Major change in util/mkdef.pl to include extra information
7994 about each symbol, as well as presenting variables as well
7995 as functions. This change means that there's n more need
7996 to rebuild the .num files when some algorithms are excluded.
7999 *) Allow the verify time to be set by an application,
8000 rather than always using the current time.
8003 *) Phase 2 verify code reorganisation. The certificate
8004 verify code now looks up an issuer certificate by a
8005 number of criteria: subject name, authority key id
8006 and key usage. It also verifies self signed certificates
8007 by the same criteria. The main comparison function is
8008 X509_check_issued() which performs these checks.
8010 Lot of changes were necessary in order to support this
8011 without completely rewriting the lookup code.
8013 Authority and subject key identifier are now cached.
8015 The LHASH 'certs' is X509_STORE has now been replaced
8016 by a STACK_OF(X509_OBJECT). This is mainly because an
8017 LHASH can't store or retrieve multiple objects with
8018 the same hash value.
8020 As a result various functions (which were all internal
8021 use only) have changed to handle the new X509_STORE
8022 structure. This will break anything that messed round
8023 with X509_STORE internally.
8025 The functions X509_STORE_add_cert() now checks for an
8026 exact match, rather than just subject name.
8028 The X509_STORE API doesn't directly support the retrieval
8029 of multiple certificates matching a given criteria, however
8030 this can be worked round by performing a lookup first
8031 (which will fill the cache with candidate certificates)
8032 and then examining the cache for matches. This is probably
8033 the best we can do without throwing out X509_LOOKUP
8034 entirely (maybe later...).
8036 The X509_VERIFY_CTX structure has been enhanced considerably.
8038 All certificate lookup operations now go via a get_issuer()
8039 callback. Although this currently uses an X509_STORE it
8040 can be replaced by custom lookups. This is a simple way
8041 to bypass the X509_STORE hackery necessary to make this
8042 work and makes it possible to use more efficient techniques
8043 in future. A very simple version which uses a simple
8044 STACK for its trusted certificate store is also provided
8045 using X509_STORE_CTX_trusted_stack().
8047 The verify_cb() and verify() callbacks now have equivalents
8048 in the X509_STORE_CTX structure.
8050 X509_STORE_CTX also has a 'flags' field which can be used
8051 to customise the verify behaviour.
8054 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8055 excludes S/MIME capabilities.
8058 *) When a certificate request is read in keep a copy of the
8059 original encoding of the signed data and use it when outputing
8060 again. Signatures then use the original encoding rather than
8061 a decoded, encoded version which may cause problems if the
8062 request is improperly encoded.
8065 *) For consistency with other BIO_puts implementations, call
8066 buffer_write(b, ...) directly in buffer_puts instead of calling
8069 In BIO_puts, increment b->num_write as in BIO_write.
8070 [Peter.Sylvester@EdelWeb.fr]
8072 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8073 BN_zero, we may not return a BIGNUM with an array consisting of
8077 *) Avoid calling abort() from within the library when problems are
8078 detected, except if preprocessor symbols have been defined
8079 (such as REF_CHECK, BN_DEBUG etc.).
8082 *) New openssl application 'rsautl'. This utility can be
8083 used for low level RSA operations. DER public key
8084 BIO/fp routines also added.
8087 *) New Configure entry and patches for compiling on QNX 4.
8088 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8090 *) A demo state-machine implementation was sponsored by
8091 Nuron (http://www.nuron.com/) and is now available in
8092 demos/state_machine.
8095 *) New options added to the 'dgst' utility for signature
8096 generation and verification.
8099 *) Unrecognized PKCS#7 content types are now handled via a
8100 catch all ASN1_TYPE structure. This allows unsupported
8101 types to be stored as a "blob" and an application can
8102 encode and decode it manually.
8105 *) Fix various signed/unsigned issues to make a_strex.c
8107 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8109 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8110 length if passed a buffer. ASN1_INTEGER_to_BN failed
8111 if passed a NULL BN and its argument was negative.
8112 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8114 *) Modification to PKCS#7 encoding routines to output definite
8115 length encoding. Since currently the whole structures are in
8116 memory there's not real point in using indefinite length
8117 constructed encoding. However if OpenSSL is compiled with
8118 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8121 *) Added BIO_vprintf() and BIO_vsnprintf().
8124 *) Added more prefixes to parse for in the the strings written
8125 through a logging bio, to cover all the levels that are available
8126 through syslog. The prefixes are now:
8128 PANIC, EMERG, EMR => LOG_EMERG
8129 ALERT, ALR => LOG_ALERT
8130 CRIT, CRI => LOG_CRIT
8131 ERROR, ERR => LOG_ERR
8132 WARNING, WARN, WAR => LOG_WARNING
8133 NOTICE, NOTE, NOT => LOG_NOTICE
8134 INFO, INF => LOG_INFO
8135 DEBUG, DBG => LOG_DEBUG
8137 and as before, if none of those prefixes are present at the
8138 beginning of the string, LOG_ERR is chosen.
8140 On Win32, the LOG_* levels are mapped according to this:
8142 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8143 LOG_WARNING => EVENTLOG_WARNING_TYPE
8144 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8148 *) Made it possible to reconfigure with just the configuration
8149 argument "reconf" or "reconfigure". The command line arguments
8150 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8151 and are retrieved from there when reconfiguring.
8155 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8157 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8160 *) The obj_dat.pl script was messing up the sorting of object
8161 names. The reason was that it compared the quoted version
8162 of strings as a result "OCSP" > "OCSP Signing" because
8163 " > SPACE. Changed script to store unquoted versions of
8164 names and add quotes on output. It was also omitting some
8165 names from the lookup table if they were given a default
8166 value (that is if SN is missing it is given the same
8167 value as LN and vice versa), these are now added on the
8168 grounds that if an object has a name we should be able to
8169 look it up. Finally added warning output when duplicate
8170 short or long names are found.
8173 *) Changes needed for Tandem NSK.
8174 [Scott Uroff <scott@xypro.com>]
8176 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8177 RSA_padding_check_SSLv23(), special padding was never detected
8178 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8179 version rollback attacks was not effective.
8181 In s23_clnt.c, don't use special rollback-attack detection padding
8182 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8183 client; similarly, in s23_srvr.c, don't do the rollback check if
8184 SSL 2.0 is the only protocol enabled in the server.
8187 *) Make it possible to get hexdumps of unprintable data with 'openssl
8188 asn1parse'. By implication, the functions ASN1_parse_dump() and
8189 BIO_dump_indent() are added.
8192 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8193 these print out strings and name structures based on various
8194 flags including RFC2253 support and proper handling of
8195 multibyte characters. Added options to the 'x509' utility
8196 to allow the various flags to be set.
8199 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8200 Also change the functions X509_cmp_current_time() and
8201 X509_gmtime_adj() work with an ASN1_TIME structure,
8202 this will enable certificates using GeneralizedTime in validity
8203 dates to be checked.
8206 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8207 negative public key encodings) on by default,
8208 NO_NEG_PUBKEY_BUG can be set to disable it.
8211 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8212 content octets. An i2c_ASN1_OBJECT is unnecessary because
8213 the encoding can be trivially obtained from the structure.
8216 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8217 not read locks (CRYPTO_r_[un]lock).
8220 *) A first attempt at creating official support for shared
8221 libraries through configuration. I've kept it so the
8222 default is static libraries only, and the OpenSSL programs
8223 are always statically linked for now, but there are
8224 preparations for dynamic linking in place.
8225 This has been tested on Linux and Tru64.
8228 *) Randomness polling function for Win9x, as described in:
8229 Peter Gutmann, Software Generation of Practically Strong
8233 *) Fix so PRNG is seeded in req if using an already existing
8237 *) New options to smime application. -inform and -outform
8238 allow alternative formats for the S/MIME message including
8239 PEM and DER. The -content option allows the content to be
8240 specified separately. This should allow things like Netscape
8241 form signing output easier to verify.
8244 *) Fix the ASN1 encoding of tags using the 'long form'.
8247 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8248 STRING types. These convert content octets to and from the
8249 underlying type. The actual tag and length octets are
8250 already assumed to have been read in and checked. These
8251 are needed because all other string types have virtually
8252 identical handling apart from the tag. By having versions
8253 of the ASN1 functions that just operate on content octets
8254 IMPLICIT tagging can be handled properly. It also allows
8255 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8256 and ASN1_INTEGER are identical apart from the tag.
8259 *) Change the handling of OID objects as follows:
8261 - New object identifiers are inserted in objects.txt, following
8262 the syntax given in objects.README.
8263 - objects.pl is used to process obj_mac.num and create a new
8265 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8268 This is currently kind of a hack, and the perl code in objects.pl
8269 isn't very elegant, but it works as I intended. The simplest way
8270 to check that it worked correctly is to look in obj_dat.h and
8271 check the array nid_objs and make sure the objects haven't moved
8272 around (this is important!). Additions are OK, as well as
8273 consistent name changes.
8276 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8279 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8280 The given file adds to whatever has already been seeded into the
8281 random pool through the RANDFILE configuration file option or
8282 environment variable, or the default random state file.
8285 *) mkstack.pl now sorts each macro group into lexical order.
8286 Previously the output order depended on the order the files
8287 appeared in the directory, resulting in needless rewriting
8291 *) Patches to make OpenSSL compile under Win32 again. Mostly
8292 work arounds for the VC++ problem that it treats func() as
8293 func(void). Also stripped out the parts of mkdef.pl that
8294 added extra typesafe functions: these no longer exist.
8297 *) Reorganisation of the stack code. The macros are now all
8298 collected in safestack.h . Each macro is defined in terms of
8299 a "stack macro" of the form SKM_<name>(type, a, b). The
8300 DEBUG_SAFESTACK is now handled in terms of function casts,
8301 this has the advantage of retaining type safety without the
8302 use of additional functions. If DEBUG_SAFESTACK is not defined
8303 then the non typesafe macros are used instead. Also modified the
8304 mkstack.pl script to handle the new form. Needs testing to see
8305 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8306 the default if no major problems. Similar behaviour for ASN1_SET_OF
8307 and PKCS12_STACK_OF.
8310 *) When some versions of IIS use the 'NET' form of private key the
8311 key derivation algorithm is different. Normally MD5(password) is
8312 used as a 128 bit RC4 key. In the modified case
8313 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8314 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8315 as the old Netscape_RSA functions except they have an additional
8316 'sgckey' parameter which uses the modified algorithm. Also added
8317 an -sgckey command line option to the rsa utility. Thanks to
8318 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8319 algorithm to openssl-dev.
8322 *) The evp_local.h macros were using 'c.##kname' which resulted in
8323 invalid expansion on some systems (SCO 5.0.5 for example).
8324 Corrected to 'c.kname'.
8325 [Phillip Porch <root@theporch.com>]
8327 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8328 a STACK of email addresses from a certificate or request, these look
8329 in the subject name and the subject alternative name extensions and
8330 omit any duplicate addresses.
8333 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8334 This makes DSA verification about 2 % faster.
8337 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8338 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8339 plus overhead for 1024 bit moduli).
8340 This makes exponentiations about 0.5 % faster for 1024 bit
8341 exponents (as measured by "openssl speed rsa2048").
8344 *) Rename memory handling macros to avoid conflicts with other
8346 Malloc => OPENSSL_malloc
8347 Malloc_locked => OPENSSL_malloc_locked
8348 Realloc => OPENSSL_realloc
8349 Free => OPENSSL_free
8352 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8353 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8356 *) CygWin32 support.
8357 [John Jarvie <jjarvie@newsguy.com>]
8359 *) The type-safe stack code has been rejigged. It is now only compiled
8360 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8361 by default all type-specific stack functions are "#define"d back to
8362 standard stack functions. This results in more streamlined output
8363 but retains the type-safety checking possibilities of the original
8367 *) The STACK code has been cleaned up, and certain type declarations
8368 that didn't make a lot of sense have been brought in line. This has
8369 also involved a cleanup of sorts in safestack.h to more correctly
8370 map type-safe stack functions onto their plain stack counterparts.
8371 This work has also resulted in a variety of "const"ifications of
8372 lots of the code, especially "_cmp" operations which should normally
8373 be prototyped with "const" parameters anyway.
8376 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8377 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8378 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8379 where all of 'md' is used each time the PRNG is used, but 'state'
8380 is used only indexed by a cyclic counter. As entropy may not be
8381 well distributed from the beginning, 'md' is important as a
8382 chaining variable. However, the output function chains only half
8383 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8384 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8385 in all of 'state' being rewritten, with the new values depending
8386 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8389 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8390 the handshake is continued after ssl_verify_cert_chain();
8391 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8392 can lead to 'unexplainable' connection aborts later.
8393 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8395 *) Major EVP API cipher revision.
8396 Add hooks for extra EVP features. This allows various cipher
8397 parameters to be set in the EVP interface. Support added for variable
8398 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8399 setting of RC2 and RC5 parameters.
8401 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8404 Remove lots of duplicated code from the EVP library. For example *every*
8405 cipher init() function handles the 'iv' in the same way according to the
8406 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8407 for CFB and OFB modes they zero ctx->num.
8409 New functionality allows removal of S/MIME code RC2 hack.
8411 Most of the routines have the same form and so can be declared in terms
8414 By shifting this to the top level EVP_CipherInit() it can be removed from
8415 all individual ciphers. If the cipher wants to handle IVs or keys
8416 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8419 Change lots of functions like EVP_EncryptUpdate() to now return a
8420 value: although software versions of the algorithms cannot fail
8421 any installed hardware versions can.
8424 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8425 this option is set, tolerate broken clients that send the negotiated
8426 protocol version number instead of the requested protocol version
8430 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8431 i.e. non-zero for export ciphersuites, zero otherwise.
8432 Previous versions had this flag inverted, inconsistent with
8433 rsa_tmp_cb (..._TMP_RSA_CB).
8434 [Bodo Moeller; problem reported by Amit Chopra]
8436 *) Add missing DSA library text string. Work around for some IIS
8437 key files with invalid SEQUENCE encoding.
8440 *) Add a document (doc/standards.txt) that list all kinds of standards
8441 and so on that are implemented in OpenSSL.
8444 *) Enhance c_rehash script. Old version would mishandle certificates
8445 with the same subject name hash and wouldn't handle CRLs at all.
8446 Added -fingerprint option to crl utility, to support new c_rehash
8450 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8453 *) Fix for SSL server purpose checking. Server checking was
8454 rejecting certificates which had extended key usage present
8455 but no ssl client purpose.
8456 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8458 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8459 is a little unclear about how a blank password is handled.
8460 Since the password in encoded as a BMPString with terminating
8461 double NULL a zero length password would end up as just the
8462 double NULL. However no password at all is different and is
8463 handled differently in the PKCS#12 key generation code. NS
8464 treats a blank password as zero length. MSIE treats it as no
8465 password on export: but it will try both on import. We now do
8466 the same: PKCS12_parse() tries zero length and no password if
8467 the password is set to "" or NULL (NULL is now a valid password:
8468 it wasn't before) as does the pkcs12 application.
8471 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8472 perror when PEM_read_bio_X509_REQ fails, the error message must
8473 be obtained from the error queue.
8476 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8477 it in ERR_remove_state if appropriate, and change ERR_get_state
8478 accordingly to avoid race conditions (this is necessary because
8479 thread_hash is no longer constant once set).
8482 *) Bugfix for linux-elf makefile.one.
8485 *) RSA_get_default_method() will now cause a default
8486 RSA_METHOD to be chosen if one doesn't exist already.
8487 Previously this was only set during a call to RSA_new()
8488 or RSA_new_method(NULL) meaning it was possible for
8489 RSA_get_default_method() to return NULL.
8492 *) Added native name translation to the existing DSO code
8493 that will convert (if the flag to do so is set) filenames
8494 that are sufficiently small and have no path information
8495 into a canonical native form. Eg. "blah" converted to
8496 "libblah.so" or "blah.dll" etc.
8499 *) New function ERR_error_string_n(e, buf, len) which is like
8500 ERR_error_string(e, buf), but writes at most 'len' bytes
8501 including the 0 terminator. For ERR_error_string_n, 'buf'
8503 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8505 *) CONF library reworked to become more general. A new CONF
8506 configuration file reader "class" is implemented as well as a
8507 new functions (NCONF_*, for "New CONF") to handle it. The now
8508 old CONF_* functions are still there, but are reimplemented to
8509 work in terms of the new functions. Also, a set of functions
8510 to handle the internal storage of the configuration data is
8511 provided to make it easier to write new configuration file
8512 reader "classes" (I can definitely see something reading a
8513 configuration file in XML format, for example), called _CONF_*,
8514 or "the configuration storage API"...
8516 The new configuration file reading functions are:
8518 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8519 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8521 NCONF_default, NCONF_WIN32
8523 NCONF_dump_fp, NCONF_dump_bio
8525 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8526 NCONF_new creates a new CONF object. This works in the same way
8527 as other interfaces in OpenSSL, like the BIO interface.
8528 NCONF_dump_* dump the internal storage of the configuration file,
8529 which is useful for debugging. All other functions take the same
8530 arguments as the old CONF_* functions wth the exception of the
8531 first that must be a `CONF *' instead of a `LHASH *'.
8533 To make it easer to use the new classes with the old CONF_* functions,
8534 the function CONF_set_default_method is provided.
8537 *) Add '-tls1' option to 'openssl ciphers', which was already
8538 mentioned in the documentation but had not been implemented.
8539 (This option is not yet really useful because even the additional
8540 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8543 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8544 OpenSSL-based applications) load shared libraries and bind to
8545 them in a portable way.
8546 [Geoff Thorpe, with contributions from Richard Levitte]
8548 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8550 *) Make sure _lrotl and _lrotr are only used with MSVC.
8552 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8553 (the default implementation of RAND_status).
8555 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8556 to '-clrext' (= clear extensions), as intended and documented.
8557 [Bodo Moeller; inconsistency pointed out by Michael Attili
8560 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8561 was larger than the MD block size.
8562 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8564 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8565 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8566 using the passed key: if the passed key was a private key the result
8567 of X509_print(), for example, would be to print out all the private key
8571 *) des_quad_cksum() byte order bug fix.
8572 [Ulf Möller, using the problem description in krb4-0.9.7, where
8573 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8575 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8577 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8579 *) For easily testing in shell scripts whether some command
8580 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8581 returns with exit code 0 iff no command of the given name is available.
8582 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8583 the output goes to stdout and nothing is printed to stderr.
8584 Additional arguments are always ignored.
8586 Since for each cipher there is a command of the same name,
8587 the 'no-cipher' compilation switches can be tested this way.
8589 ('openssl no-XXX' is not able to detect pseudo-commands such
8590 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8593 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8596 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8597 is set; it will be thrown away anyway because each handshake creates
8599 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8600 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8601 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8602 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8605 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8606 'Q' and 'R' lose their special meanings (quit/renegotiate).
8607 This is part of what -quiet does; unlike -quiet, -ign_eof
8608 does not suppress any output.
8611 *) Add compatibility options to the purpose and trust code. The
8612 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8613 accepts a certificate or CA, this was the previous behaviour,
8614 with all the associated security issues.
8616 X509_TRUST_COMPAT is the old trust behaviour: only and
8617 automatically trust self signed roots in certificate store. A
8618 new trust setting X509_TRUST_DEFAULT is used to specify that
8619 a purpose has no associated trust setting and it should instead
8620 use the value in the default purpose.
8623 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8624 and fix a memory leak.
8627 *) In util/mkerr.pl (which implements 'make errors'), preserve
8628 reason strings from the previous version of the .c file, as
8629 the default to have only downcase letters (and digits) in
8630 automatically generated reasons codes is not always appropriate.
8633 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8634 using strerror. Previously, ERR_reason_error_string() returned
8635 library names as reason strings for SYSerr; but SYSerr is a special
8636 case where small numbers are errno values, not library numbers.
8639 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8640 converts DSA parameters into DH parameters. (When creating parameters,
8641 DSA_generate_parameters is used.)
8644 *) Include 'length' (recommended exponent length) in C code generated
8645 by 'openssl dhparam -C'.
8648 *) The second argument to set_label in perlasm was already being used
8649 so couldn't be used as a "file scope" flag. Moved to third argument
8653 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8654 instead of RAND_bytes for encryption IVs and salts.
8657 *) Include RAND_status() into RAND_METHOD instead of implementing
8658 it only for md_rand.c Otherwise replacing the PRNG by calling
8659 RAND_set_rand_method would be impossible.
8662 *) Don't let DSA_generate_key() enter an infinite loop if the random
8663 number generation fails.
8666 *) New 'rand' application for creating pseudo-random output.
8669 *) Added configuration support for Linux/IA64
8670 [Rolf Haberrecker <rolf@suse.de>]
8672 *) Assembler module support for Mingw32.
8675 *) Shared library support for HPUX (in shlib/).
8676 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8678 *) Shared library support for Solaris gcc.
8679 [Lutz Behnke <behnke@trustcenter.de>]
8681 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8683 *) PKCS7_encrypt() was adding text MIME headers twice because they
8684 were added manually and by SMIME_crlf_copy().
8687 *) In bntest.c don't call BN_rand with zero bits argument.
8688 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8690 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8691 case was implemented. This caused BN_div_recp() to fail occasionally.
8694 *) Add an optional second argument to the set_label() in the perl
8695 assembly language builder. If this argument exists and is set
8696 to 1 it signals that the assembler should use a symbol whose
8697 scope is the entire file, not just the current function. This
8698 is needed with MASM which uses the format label:: for this scope.
8699 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8701 *) Change the ASN1 types so they are typedefs by default. Before
8702 almost all types were #define'd to ASN1_STRING which was causing
8703 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8707 *) Change names of new functions to the new get1/get0 naming
8708 convention: After 'get1', the caller owns a reference count
8709 and has to call ..._free; 'get0' returns a pointer to some
8710 data structure without incrementing reference counters.
8711 (Some of the existing 'get' functions increment a reference
8712 counter, some don't.)
8713 Similarly, 'set1' and 'add1' functions increase reference
8714 counters or duplicate objects.
8717 *) Allow for the possibility of temp RSA key generation failure:
8718 the code used to assume it always worked and crashed on failure.
8721 *) Fix potential buffer overrun problem in BIO_printf().
8722 [Ulf Möller, using public domain code by Patrick Powell; problem
8723 pointed out by David Sacerdote <das33@cornell.edu>]
8725 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8726 RAND_egd() and RAND_status(). In the command line application,
8727 the EGD socket can be specified like a seed file using RANDFILE
8731 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8732 Some CAs (e.g. Verisign) distribute certificates in this form.
8735 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8736 list to exclude them. This means that no special compilation option
8737 is needed to use anonymous DH: it just needs to be included in the
8741 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8742 EVP_MD_type. The old functionality is available in a new macro called
8743 EVP_MD_md(). Change code that uses it and update docs.
8746 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8747 where the 'void *' argument is replaced by a function pointer argument.
8748 Previously 'void *' was abused to point to functions, which works on
8749 many platforms, but is not correct. As these functions are usually
8750 called by macros defined in OpenSSL header files, most source code
8751 should work without changes.
8754 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8755 sections with information on -D... compiler switches used for
8756 compiling the library so that applications can see them. To enable
8757 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8758 must be defined. E.g.,
8759 #define OPENSSL_ALGORITHM_DEFINES
8760 #include <openssl/opensslconf.h>
8761 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8762 [Richard Levitte, Ulf and Bodo Möller]
8764 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8768 *) Change the 'other' type in certificate aux info to a STACK_OF
8769 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8770 the required ASN1 format: arbitrary types determined by an OID.
8773 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8774 argument to 'req'. This is not because the function is newer or
8775 better than others it just uses the work 'NEW' in the certificate
8776 request header lines. Some software needs this.
8779 *) Reorganise password command line arguments: now passwords can be
8780 obtained from various sources. Delete the PEM_cb function and make
8781 it the default behaviour: i.e. if the callback is NULL and the
8782 usrdata argument is not NULL interpret it as a null terminated pass
8783 phrase. If usrdata and the callback are NULL then the pass phrase
8784 is prompted for as usual.
8787 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8788 the support is automatically enabled. The resulting binaries will
8789 autodetect the card and use it if present.
8790 [Ben Laurie and Compaq Inc.]
8792 *) Work around for Netscape hang bug. This sends certificate request
8793 and server done in one record. Since this is perfectly legal in the
8794 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8795 the bugs/SSLv3 entry for more info.
8798 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8801 *) Add -rand argument to smime and pkcs12 applications and read/write
8805 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8808 *) Add command line password options to the remaining applications.
8811 *) Bug fix for BN_div_recp() for numerators with an even number of
8815 *) More tests in bntest.c, and changed test_bn output.
8818 *) ./config recognizes MacOS X now.
8821 *) Bug fix for BN_div() when the first words of num and divsor are
8822 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8825 *) Add support for various broken PKCS#8 formats, and command line
8826 options to produce them.
8829 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8830 get temporary BIGNUMs from a BN_CTX.
8833 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8837 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8838 include a #define from the old name to the new. The original intent
8839 was that statically linked binaries could for example just call
8840 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8841 link with digests. This never worked because SSLeay_add_all_digests()
8842 and SSLeay_add_all_ciphers() were in the same source file so calling
8843 one would link with the other. They are now in separate source files.
8846 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8849 *) Use a less unusual form of the Miller-Rabin primality test (it used
8850 a binary algorithm for exponentiation integrated into the Miller-Rabin
8851 loop, our standard modexp algorithms are faster).
8854 *) Support for the EBCDIC character set completed.
8855 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8857 *) Source code cleanups: use const where appropriate, eliminate casts,
8858 use void * instead of char * in lhash.
8861 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8862 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8863 this the server could overwrite ephemeral keys that the client
8867 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8868 using 50 iterations of the Rabin-Miller test.
8870 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8871 iterations of the Rabin-Miller test as required by the appendix
8872 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8873 As BN_is_prime_fasttest includes trial division, DSA parameter
8874 generation becomes much faster.
8876 This implies a change for the callback functions in DSA_is_prime
8877 and DSA_generate_parameters: The callback function is called once
8878 for each positive witness in the Rabin-Miller test, not just
8879 occasionally in the inner loop; and the parameters to the
8880 callback function now provide an iteration count for the outer
8881 loop rather than for the current invocation of the inner loop.
8882 DSA_generate_parameters additionally can call the callback
8883 function with an 'iteration count' of -1, meaning that a
8884 candidate has passed the trial division test (when q is generated
8885 from an application-provided seed, trial division is skipped).
8888 *) New function BN_is_prime_fasttest that optionally does trial
8889 division before starting the Rabin-Miller test and has
8890 an additional BN_CTX * argument (whereas BN_is_prime always
8891 has to allocate at least one BN_CTX).
8892 'callback(1, -1, cb_arg)' is called when a number has passed the
8893 trial division stage.
8896 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8900 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8903 *) New function BN_pseudo_rand().
8906 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8907 bignum version of BN_from_montgomery() with the working code from
8908 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8912 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8913 made it impossible to use the same SSL_SESSION data structure in
8914 SSL2 clients in multiple threads.
8917 *) The return value of RAND_load_file() no longer counts bytes obtained
8918 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8919 to seed the PRNG (previously an explicit byte count was required).
8920 [Ulf Möller, Bodo Möller]
8922 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8923 used (char *) instead of (void *) and had casts all over the place.
8926 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8929 *) Retain source code compatibility for BN_prime_checks macro:
8930 BN_is_prime(..., BN_prime_checks, ...) now uses
8931 BN_prime_checks_for_size to determine the appropriate number of
8932 Rabin-Miller iterations.
8935 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8936 DH_CHECK_P_NOT_SAFE_PRIME.
8937 (Check if this is true? OpenPGP calls them "strong".)
8940 *) Merge the functionality of "dh" and "gendh" programs into a new program
8941 "dhparam". The old programs are retained for now but will handle DH keys
8942 (instead of parameters) in future.
8945 *) Make the ciphers, s_server and s_client programs check the return values
8946 when a new cipher list is set.
8949 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8950 ciphers. Before when the 56bit ciphers were enabled the sorting was
8953 The syntax for the cipher sorting has been extended to support sorting by
8954 cipher-strength (using the strength_bits hard coded in the tables).
8955 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8957 Fix a bug in the cipher-command parser: when supplying a cipher command
8958 string with an "undefined" symbol (neither command nor alphanumeric
8959 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8960 an error is flagged.
8962 Due to the strength-sorting extension, the code of the
8963 ssl_create_cipher_list() function was completely rearranged. I hope that
8964 the readability was also increased :-)
8965 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8967 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8968 for the first serial number and places 2 in the serial number file. This
8969 avoids problems when the root CA is created with serial number zero and
8970 the first user certificate has the same issuer name and serial number
8974 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8975 the new code. Add documentation for this stuff.
8978 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8979 X509_*() to X509at_*() on the grounds that they don't handle X509
8980 structures and behave in an analogous way to the X509v3 functions:
8981 they shouldn't be called directly but wrapper functions should be used
8984 So we also now have some wrapper functions that call the X509at functions
8985 when passed certificate requests. (TO DO: similar things can be done with
8986 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8987 things. Some of these need some d2i or i2d and print functionality
8988 because they handle more complex structures.)
8991 *) Add missing #ifndefs that caused missing symbols when building libssl
8992 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8993 NO_RSA in ssl/s2*.c.
8994 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8996 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8997 has a return value which indicates the quality of the random data
8998 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8999 error queue. New function RAND_pseudo_bytes() generates output that is
9000 guaranteed to be unique but not unpredictable. RAND_add is like
9001 RAND_seed, but takes an extra argument for an entropy estimate
9002 (RAND_seed always assumes full entropy).
9005 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9006 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9007 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9008 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9009 false-positive rate of at most 2^-80 for random input.
9012 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9015 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9016 in the 0.9.5 release), this returns the chain
9017 from an X509_CTX structure with a dup of the stack and all
9018 the X509 reference counts upped: so the stack will exist
9019 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9022 Also make SSL_SESSION_print() print out the verify return
9026 *) Add manpage for the pkcs12 command. Also change the default
9027 behaviour so MAC iteration counts are used unless the new
9028 -nomaciter option is used. This improves file security and
9029 only older versions of MSIE (4.0 for example) need it.
9032 *) Honor the no-xxx Configure options when creating .DEF files.
9035 *) Add PKCS#10 attributes to field table: challengePassword,
9036 unstructuredName and unstructuredAddress. These are taken from
9037 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9038 international characters are used.
9040 More changes to X509_ATTRIBUTE code: allow the setting of types
9041 based on strings. Remove the 'loc' parameter when adding
9042 attributes because these will be a SET OF encoding which is sorted
9046 *) Initial changes to the 'req' utility to allow request generation
9047 automation. This will allow an application to just generate a template
9048 file containing all the field values and have req construct the
9051 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9052 used all over the place including certificate requests and PKCS#7
9053 structures. They are currently handled manually where necessary with
9054 some primitive wrappers for PKCS#7. The new functions behave in a
9055 manner analogous to the X509 extension functions: they allow
9056 attributes to be looked up by NID and added.
9058 Later something similar to the X509V3 code would be desirable to
9059 automatically handle the encoding, decoding and printing of the
9060 more complex types. The string types like challengePassword can
9061 be handled by the string table functions.
9063 Also modified the multi byte string table handling. Now there is
9064 a 'global mask' which masks out certain types. The table itself
9065 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9066 is useful when for example there is only one permissible type
9067 (as in countryName) and using the mask might result in no valid
9071 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9072 SSL_get_peer_finished to allow applications to obtain the latest
9073 Finished messages sent to the peer or expected from the peer,
9074 respectively. (SSL_get_peer_finished is usually the Finished message
9075 actually received from the peer, otherwise the protocol will be aborted.)
9077 As the Finished message are message digests of the complete handshake
9078 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9079 be used for external authentication procedures when the authentication
9080 provided by SSL/TLS is not desired or is not enough.
9083 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9084 the host supports BWX extension and if Compaq C is present on the
9085 $PATH. Just exploiting of the BWX extension results in 20-30%
9086 performance kick for some algorithms, e.g. DES and RC4 to mention
9087 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9091 *) Add support for MS "fast SGC". This is arguably a violation of the
9092 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9093 weak crypto and after checking the certificate is SGC a second one
9094 with strong crypto. MS SGC stops the first handshake after receiving
9095 the server certificate message and sends a second client hello. Since
9096 a server will typically do all the time consuming operations before
9097 expecting any further messages from the client (server key exchange
9098 is the most expensive) there is little difference between the two.
9100 To get OpenSSL to support MS SGC we have to permit a second client
9101 hello message after we have sent server done. In addition we have to
9102 reset the MAC if we do get this second client hello.
9105 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9106 if a DER encoded private key is RSA or DSA traditional format. Changed
9107 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9108 format DER encoded private key. Newer code should use PKCS#8 format which
9109 has the key type encoded in the ASN1 structure. Added DER private key
9110 support to pkcs8 application.
9113 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9114 ciphersuites has been selected (as required by the SSL 3/TLS 1
9115 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9116 is set, we interpret this as a request to violate the specification
9117 (the worst that can happen is a handshake failure, and 'correct'
9118 behaviour would result in a handshake failure anyway).
9121 *) In SSL_CTX_add_session, take into account that there might be multiple
9122 SSL_SESSION structures with the same session ID (e.g. when two threads
9123 concurrently obtain them from an external cache).
9124 The internal cache can handle only one SSL_SESSION with a given ID,
9125 so if there's a conflict, we now throw out the old one to achieve
9129 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9130 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9131 some routines that use cipher OIDs: some ciphers do not have OIDs
9132 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9136 *) Simplify the trust setting structure and code. Now we just have
9137 two sequences of OIDs for trusted and rejected settings. These will
9138 typically have values the same as the extended key usage extension
9139 and any application specific purposes.
9141 The trust checking code now has a default behaviour: it will just
9142 check for an object with the same NID as the passed id. Functions can
9143 be provided to override either the default behaviour or the behaviour
9144 for a given id. SSL client, server and email already have functions
9145 in place for compatibility: they check the NID and also return "trusted"
9146 if the certificate is self signed.
9149 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9150 traditional format into an EVP_PKEY structure.
9153 *) Add a password callback function PEM_cb() which either prompts for
9154 a password if usr_data is NULL or otherwise assumes it is a null
9155 terminated password. Allow passwords to be passed on command line
9156 environment or config files in a few more utilities.
9159 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9160 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9161 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9162 Update documentation.
9165 *) Support for ASN1 "NULL" type. This could be handled before by using
9166 ASN1_TYPE but there wasn't any function that would try to read a NULL
9167 and produce an error if it couldn't. For compatibility we also have
9168 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9169 don't allocate anything because they don't need to.
9172 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9174 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9176 *) Rebuild of the memory allocation routines used by OpenSSL code and
9177 possibly others as well. The purpose is to make an interface that
9178 provide hooks so anyone can build a separate set of allocation and
9179 deallocation routines to be used by OpenSSL, for example memory
9180 pool implementations, or something else, which was previously hard
9181 since Malloc(), Realloc() and Free() were defined as macros having
9182 the values malloc, realloc and free, respectively (except for Win32
9183 compilations). The same is provided for memory debugging code.
9184 OpenSSL already comes with functionality to find memory leaks, but
9185 this gives people a chance to debug other memory problems.
9187 With these changes, a new set of functions and macros have appeared:
9189 CRYPTO_set_mem_debug_functions() [F]
9190 CRYPTO_get_mem_debug_functions() [F]
9191 CRYPTO_dbg_set_options() [F]
9192 CRYPTO_dbg_get_options() [F]
9193 CRYPTO_malloc_debug_init() [M]
9195 The memory debug functions are NULL by default, unless the library
9196 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9197 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9198 gives the standard debugging functions that come with OpenSSL) or
9199 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9200 provided by the library user) must be used. When the standard
9201 debugging functions are used, CRYPTO_dbg_set_options can be used to
9202 request additional information:
9203 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9204 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9206 Also, things like CRYPTO_set_mem_functions will always give the
9207 expected result (the new set of functions is used for allocation
9208 and deallocation) at all times, regardless of platform and compiler
9211 To finish it up, some functions that were never use in any other
9212 way than through macros have a new API and new semantic:
9215 CRYPTO_dbg_realloc()
9218 All macros of value have retained their old syntax.
9219 [Richard Levitte and Bodo Moeller]
9221 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9222 ordering of SMIMECapabilities wasn't in "strength order" and there
9223 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9227 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9228 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9229 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9231 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9232 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9233 functionality to handle multipart/signed properly) and a utility
9234 called 'smime' to call all this stuff. This is based on code I
9235 originally wrote for Celo who have kindly allowed it to be
9236 included in OpenSSL.
9239 *) Add variants des_set_key_checked and des_set_key_unchecked of
9240 des_set_key (aka des_key_sched). Global variable des_check_key
9241 decides which of these is called by des_set_key; this way
9242 des_check_key behaves as it always did, but applications and
9243 the library itself, which was buggy for des_check_key == 1,
9244 have a cleaner way to pick the version they need.
9247 *) New function PKCS12_newpass() which changes the password of a
9251 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9252 dynamic mix. In both cases the ids can be used as an index into the
9253 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9254 functions so they accept a list of the field values and the
9255 application doesn't need to directly manipulate the X509_TRUST
9259 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9263 *) Modify the way the V3 extension code looks up extensions. This now
9264 works in a similar way to the object code: we have some "standard"
9265 extensions in a static table which is searched with OBJ_bsearch()
9266 and the application can add dynamic ones if needed. The file
9267 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9268 updated whenever a new extension is added to the core code and kept
9269 in ext_nid order. There is a simple program 'tabtest.c' which checks
9270 this. New extensions are not added too often so this file can readily
9271 be maintained manually.
9273 There are two big advantages in doing things this way. The extensions
9274 can be looked up immediately and no longer need to be "added" using
9275 X509V3_add_standard_extensions(): this function now does nothing.
9276 [Side note: I get *lots* of email saying the extension code doesn't
9277 work because people forget to call this function]
9278 Also no dynamic allocation is done unless new extensions are added:
9279 so if we don't add custom extensions there is no need to call
9280 X509V3_EXT_cleanup().
9283 *) Modify enc utility's salting as follows: make salting the default. Add a
9284 magic header, so unsalted files fail gracefully instead of just decrypting
9285 to garbage. This is because not salting is a big security hole, so people
9286 should be discouraged from doing it.
9289 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9290 digest to be passed on the command line but it only used this
9291 parameter when signing a certificate. Modified so all relevant
9292 operations are affected by the digest parameter including the
9293 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9294 DSA key was used because it didn't fix the digest.
9297 *) Initial certificate chain verify code. Currently tests the untrusted
9298 certificates for consistency with the verify purpose (which is set
9299 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9301 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9302 this is because it will reject chains with invalid extensions whereas
9303 every previous version of OpenSSL and SSLeay made no checks at all.
9305 Trust code: checks the root CA for the relevant trust settings. Trust
9306 settings have an initial value consistent with the verify purpose: e.g.
9307 if the verify purpose is for SSL client use it expects the CA to be
9308 trusted for SSL client use. However the default value can be changed to
9309 permit custom trust settings: one example of this would be to only trust
9310 certificates from a specific "secure" set of CAs.
9312 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9313 which should be used for version portability: especially since the
9314 verify structure is likely to change more often now.
9316 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9317 to set them. If not set then assume SSL clients will verify SSL servers
9320 Two new options to the verify program: -untrusted allows a set of
9321 untrusted certificates to be passed in and -purpose which sets the
9322 intended purpose of the certificate. If a purpose is set then the
9323 new chain verify code is used to check extension consistency.
9326 *) Support for the authority information access extension.
9329 *) Modify RSA and DSA PEM read routines to transparently handle
9330 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9331 public keys in a format compatible with certificate
9332 SubjectPublicKeyInfo structures. Unfortunately there were already
9333 functions called *_PublicKey_* which used various odd formats so
9334 these are retained for compatibility: however the DSA variants were
9335 never in a public release so they have been deleted. Changed dsa/rsa
9336 utilities to handle the new format: note no releases ever handled public
9337 keys so we should be OK.
9339 The primary motivation for this change is to avoid the same fiasco
9340 that dogs private keys: there are several incompatible private key
9341 formats some of which are standard and some OpenSSL specific and
9342 require various evil hacks to allow partial transparent handling and
9343 even then it doesn't work with DER formats. Given the option anything
9344 other than PKCS#8 should be dumped: but the other formats have to
9345 stay in the name of compatibility.
9347 With public keys and the benefit of hindsight one standard format
9348 is used which works with EVP_PKEY, RSA or DSA structures: though
9349 it clearly returns an error if you try to read the wrong kind of key.
9351 Added a -pubkey option to the 'x509' utility to output the public key.
9352 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9353 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9354 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9355 that do the same as the EVP_PKEY_assign_*() except they up the
9356 reference count of the added key (they don't "swallow" the
9360 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9361 CRLs would fail if the file contained no certificates or no CRLs:
9362 added a new function to read in both types and return the number
9363 read: this means that if none are read it will be an error. The
9364 DER versions of the certificate and CRL reader would always fail
9365 because it isn't possible to mix certificates and CRLs in DER format
9366 without choking one or the other routine. Changed this to just read
9367 a certificate: this is the best we can do. Also modified the code
9368 in apps/verify.c to take notice of return codes: it was previously
9369 attempting to read in certificates from NULL pointers and ignoring
9370 any errors: this is one reason why the cert and CRL reader seemed
9371 to work. It doesn't check return codes from the default certificate
9372 routines: these may well fail if the certificates aren't installed.
9375 *) Code to support otherName option in GeneralName.
9378 *) First update to verify code. Change the verify utility
9379 so it warns if it is passed a self signed certificate:
9380 for consistency with the normal behaviour. X509_verify
9381 has been modified to it will now verify a self signed
9382 certificate if *exactly* the same certificate appears
9383 in the store: it was previously impossible to trust a
9384 single self signed certificate. This means that:
9385 openssl verify ss.pem
9386 now gives a warning about a self signed certificate but
9387 openssl verify -CAfile ss.pem ss.pem
9391 *) For servers, store verify_result in SSL_SESSION data structure
9392 (and add it to external session representation).
9393 This is needed when client certificate verifications fails,
9394 but an application-provided verification callback (set by
9395 SSL_CTX_set_cert_verify_callback) allows accepting the session
9396 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9397 but returns 1): When the session is reused, we have to set
9398 ssl->verify_result to the appropriate error code to avoid
9400 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9402 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9403 case in PKCS7_dataInit() where the signed PKCS7 structure
9404 didn't contain any existing data because it was being created.
9405 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9407 *) Add a salt to the key derivation routines in enc.c. This
9408 forms the first 8 bytes of the encrypted file. Also add a
9409 -S option to allow a salt to be input on the command line.
9412 *) New function X509_cmp(). Oddly enough there wasn't a function
9413 to compare two certificates. We do this by working out the SHA1
9414 hash and comparing that. X509_cmp() will be needed by the trust
9418 *) SSL_get1_session() is like SSL_get_session(), but increments
9419 the reference count in the SSL_SESSION returned.
9420 [Geoff Thorpe <geoff@eu.c2.net>]
9422 *) Fix for 'req': it was adding a null to request attributes.
9423 Also change the X509_LOOKUP and X509_INFO code to handle
9424 certificate auxiliary information.
9427 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9431 *) Add the possibility to add extra information to the memory leak
9432 detecting output, to form tracebacks, showing from where each
9433 allocation was originated: CRYPTO_push_info("constant string") adds
9434 the string plus current file name and line number to a per-thread
9435 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9436 is like calling CYRPTO_pop_info() until the stack is empty.
9437 Also updated memory leak detection code to be multi-thread-safe.
9440 *) Add options -text and -noout to pkcs7 utility and delete the
9441 encryption options which never did anything. Update docs.
9444 *) Add options to some of the utilities to allow the pass phrase
9445 to be included on either the command line (not recommended on
9446 OSes like Unix) or read from the environment. Update the
9447 manpages and fix a few bugs.
9450 *) Add a few manpages for some of the openssl commands.
9453 *) Fix the -revoke option in ca. It was freeing up memory twice,
9454 leaking and not finding already revoked certificates.
9457 *) Extensive changes to support certificate auxiliary information.
9458 This involves the use of X509_CERT_AUX structure and X509_AUX
9459 functions. An X509_AUX function such as PEM_read_X509_AUX()
9460 can still read in a certificate file in the usual way but it
9461 will also read in any additional "auxiliary information". By
9462 doing things this way a fair degree of compatibility can be
9463 retained: existing certificates can have this information added
9464 using the new 'x509' options.
9466 Current auxiliary information includes an "alias" and some trust
9467 settings. The trust settings will ultimately be used in enhanced
9468 certificate chain verification routines: currently a certificate
9469 can only be trusted if it is self signed and then it is trusted
9473 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9474 The problem was that one of the replacement routines had not been working
9475 since SSLeay releases. For now the offending routine has been replaced
9476 with non-optimised assembler. Even so, this now gives around 95%
9477 performance improvement for 1024 bit RSA signs.
9480 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9481 handling. Most clients have the effective key size in bits equal to
9482 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9483 A few however don't do this and instead use the size of the decrypted key
9484 to determine the RC2 key length and the AlgorithmIdentifier to determine
9485 the effective key length. In this case the effective key length can still
9486 be 40 bits but the key length can be 168 bits for example. This is fixed
9487 by manually forcing an RC2 key into the EVP_PKEY structure because the
9488 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9489 the key length and effective key length are equal.
9492 *) Add a bunch of functions that should simplify the creation of
9493 X509_NAME structures. Now you should be able to do:
9494 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9495 and have it automatically work out the correct field type and fill in
9496 the structures. The more adventurous can try:
9497 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9498 and it will (hopefully) work out the correct multibyte encoding.
9501 *) Change the 'req' utility to use the new field handling and multibyte
9502 copy routines. Before the DN field creation was handled in an ad hoc
9503 way in req, ca, and x509 which was rather broken and didn't support
9504 BMPStrings or UTF8Strings. Since some software doesn't implement
9505 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9506 using the dirstring_type option. See the new comment in the default
9507 openssl.cnf for more info.
9510 *) Make crypto/rand/md_rand.c more robust:
9511 - Assure unique random numbers after fork().
9512 - Make sure that concurrent threads access the global counter and
9513 md serializably so that we never lose entropy in them
9514 or use exactly the same state in multiple threads.
9515 Access to the large state is not always serializable because
9516 the additional locking could be a performance killer, and
9517 md should be large enough anyway.
9520 *) New file apps/app_rand.c with commonly needed functionality
9521 for handling the random seed file.
9523 Use the random seed file in some applications that previously did not:
9525 dsaparam -genkey (which also ignored its '-rand' option),
9528 x509 (when signing).
9529 Except on systems with /dev/urandom, it is crucial to have a random
9530 seed file at least for key creation, DSA signing, and for DH exchanges;
9531 for RSA signatures we could do without one.
9533 gendh and gendsa (unlike genrsa) used to read only the first byte
9534 of each file listed in the '-rand' option. The function as previously
9535 found in genrsa is now in app_rand.c and is used by all programs
9536 that support '-rand'.
9539 *) In RAND_write_file, use mode 0600 for creating files;
9540 don't just chmod when it may be too late.
9543 *) Report an error from X509_STORE_load_locations
9544 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9547 *) New function ASN1_mbstring_copy() this copies a string in either
9548 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9549 into an ASN1_STRING type. A mask of permissible types is passed
9550 and it chooses the "minimal" type to use or an error if not type
9554 *) Add function equivalents to the various macros in asn1.h. The old
9555 macros are retained with an M_ prefix. Code inside the library can
9556 use the M_ macros. External code (including the openssl utility)
9557 should *NOT* in order to be "shared library friendly".
9560 *) Add various functions that can check a certificate's extensions
9561 to see if it usable for various purposes such as SSL client,
9562 server or S/MIME and CAs of these types. This is currently
9563 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9564 verification. Also added a -purpose flag to x509 utility to
9565 print out all the purposes.
9568 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9572 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9573 for, obtain and decode and extension and obtain its critical flag.
9574 This allows all the necessary extension code to be handled in a
9575 single function call.
9578 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9579 platforms. See crypto/rc4/rc4_enc.c for further details.
9582 *) New -noout option to asn1parse. This causes no output to be produced
9583 its main use is when combined with -strparse and -out to extract data
9584 from a file (which may not be in ASN.1 format).
9587 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9588 when producing the local key id.
9589 [Richard Levitte <levitte@stacken.kth.se>]
9591 *) New option -dhparam in s_server. This allows a DH parameter file to be
9592 stated explicitly. If it is not stated then it tries the first server
9593 certificate file. The previous behaviour hard coded the filename
9597 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9598 a public key to be input or output. For example:
9599 openssl rsa -in key.pem -pubout -out pubkey.pem
9600 Also added necessary DSA public key functions to handle this.
9603 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9604 in the message. This was handled by allowing
9605 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9606 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9608 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9609 to the end of the strings whereas this didn't. This would cause problems
9610 if strings read with d2i_ASN1_bytes() were later modified.
9611 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9613 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9614 data and it contains EOF it will end up returning an error. This is
9615 caused by input 46 bytes long. The cause is due to the way base64
9616 BIOs find the start of base64 encoded data. They do this by trying a
9617 trial decode on each line until they find one that works. When they
9618 do a flag is set and it starts again knowing it can pass all the
9619 data directly through the decoder. Unfortunately it doesn't reset
9620 the context it uses. This means that if EOF is reached an attempt
9621 is made to pass two EOFs through the context and this causes the
9622 resulting error. This can also cause other problems as well. As is
9623 usual with these problems it takes *ages* to find and the fix is
9624 trivial: move one line.
9625 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9627 *) Ugly workaround to get s_client and s_server working under Windows. The
9628 old code wouldn't work because it needed to select() on sockets and the
9629 tty (for keypresses and to see if data could be written). Win32 only
9630 supports select() on sockets so we select() with a 1s timeout on the
9631 sockets and then see if any characters are waiting to be read, if none
9632 are present then we retry, we also assume we can always write data to
9633 the tty. This isn't nice because the code then blocks until we've
9634 received a complete line of data and it is effectively polling the
9635 keyboard at 1s intervals: however it's quite a bit better than not
9636 working at all :-) A dedicated Windows application might handle this
9637 with an event loop for example.
9640 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9641 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9642 will be called when RSA_sign() and RSA_verify() are used. This is useful
9643 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9644 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9645 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9646 This necessitated the support of an extra signature type NID_md5_sha1
9647 for SSL signatures and modifications to the SSL library to use it instead
9648 of calling RSA_public_decrypt() and RSA_private_encrypt().
9651 *) Add new -verify -CAfile and -CApath options to the crl program, these
9652 will lookup a CRL issuers certificate and verify the signature in a
9653 similar way to the verify program. Tidy up the crl program so it
9654 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9655 less strict. It will now permit CRL extensions even if it is not
9656 a V2 CRL: this will allow it to tolerate some broken CRLs.
9659 *) Initialize all non-automatic variables each time one of the openssl
9660 sub-programs is started (this is necessary as they may be started
9661 multiple times from the "OpenSSL>" prompt).
9662 [Lennart Bang, Bodo Moeller]
9664 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9665 removing all other RSA functionality (this is what NO_RSA does). This
9666 is so (for example) those in the US can disable those operations covered
9667 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9671 *) Non-copying interface to BIO pairs.
9672 (still largely untested)
9675 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9676 ASCII string. This was handled independently in various places before.
9679 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9680 UTF8 strings a character at a time.
9683 *) Use client_version from client hello to select the protocol
9684 (s23_srvr.c) and for RSA client key exchange verification
9685 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9688 *) Add various utility functions to handle SPKACs, these were previously
9689 handled by poking round in the structure internals. Added new function
9690 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9691 print, verify and generate SPKACs. Based on an original idea from
9692 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9695 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9698 *) Allow the config file extension section to be overwritten on the
9699 command line. Based on an original idea from Massimiliano Pala
9700 <madwolf@comune.modena.it>. The new option is called -extensions
9701 and can be applied to ca, req and x509. Also -reqexts to override
9702 the request extensions in req and -crlexts to override the crl extensions
9706 *) Add new feature to the SPKAC handling in ca. Now you can include
9707 the same field multiple times by preceding it by "XXXX." for example:
9710 this is the same syntax as used in the req config file.
9713 *) Allow certificate extensions to be added to certificate requests. These
9714 are specified in a 'req_extensions' option of the req section of the
9715 config file. They can be printed out with the -text option to req but
9716 are otherwise ignored at present.
9719 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9720 data read consists of only the final block it would not decrypted because
9721 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9722 A misplaced 'break' also meant the decrypted final block might not be
9723 copied until the next read.
9726 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9727 a few extra parameters to the DH structure: these will be useful if
9728 for example we want the value of 'q' or implement X9.42 DH.
9731 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9732 provides hooks that allow the default DSA functions or functions on a
9733 "per key" basis to be replaced. This allows hardware acceleration and
9734 hardware key storage to be handled without major modification to the
9735 library. Also added low level modexp hooks and CRYPTO_EX structure and
9736 associated functions.
9739 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9740 as "read only": it can't be written to and the buffer it points to will
9741 not be freed. Reading from a read only BIO is much more efficient than
9742 a normal memory BIO. This was added because there are several times when
9743 an area of memory needs to be read from a BIO. The previous method was
9744 to create a memory BIO and write the data to it, this results in two
9745 copies of the data and an O(n^2) reading algorithm. There is a new
9746 function BIO_new_mem_buf() which creates a read only memory BIO from
9747 an area of memory. Also modified the PKCS#7 routines to use read only
9751 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9752 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9753 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9754 but a retry condition occured while trying to read the rest.
9757 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9758 NID_pkcs7_encrypted by default: this was wrong since this should almost
9759 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9760 the encrypted data type: this is a more sensible place to put it and it
9761 allows the PKCS#12 code to be tidied up that duplicated this
9765 *) Changed obj_dat.pl script so it takes its input and output files on
9766 the command line. This should avoid shell escape redirection problems
9770 *) Initial support for certificate extension requests, these are included
9771 in things like Xenroll certificate requests. Included functions to allow
9772 extensions to be obtained and added.
9775 *) -crlf option to s_client and s_server for sending newlines as
9776 CRLF (as required by many protocols).
9779 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9781 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9782 [Ralf S. Engelschall]
9784 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9785 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9787 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9791 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9792 DH parameters/keys (q is lost during that conversion, but the resulting
9793 DH parameters contain its length).
9795 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9796 much faster than DH_generate_parameters (which creates parameters
9797 where p = 2*q + 1), and also the smaller q makes DH computations
9798 much more efficient (160-bit exponentiation instead of 1024-bit
9799 exponentiation); so this provides a convenient way to support DHE
9800 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9801 utter importance to use
9802 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9804 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9805 when such DH parameters are used, because otherwise small subgroup
9806 attacks may become possible!
9809 *) Avoid memory leak in i2d_DHparams.
9812 *) Allow the -k option to be used more than once in the enc program:
9813 this allows the same encrypted message to be read by multiple recipients.
9816 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9817 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9818 it will always use the numerical form of the OID, even if it has a short
9822 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9823 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9824 otherwise bn_mod_exp was called. In the case of hardware keys for example
9825 no private key components need be present and it might store extra data
9826 in the RSA structure, which cannot be accessed from bn_mod_exp.
9827 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9828 private key operations.
9831 *) Added support for SPARC Linux.
9834 *) pem_password_cb function type incompatibly changed from
9835 typedef int pem_password_cb(char *buf, int size, int rwflag);
9837 ....(char *buf, int size, int rwflag, void *userdata);
9838 so that applications can pass data to their callbacks:
9839 The PEM[_ASN1]_{read,write}... functions and macros now take an
9840 additional void * argument, which is just handed through whenever
9841 the password callback is called.
9842 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9844 New function SSL_CTX_set_default_passwd_cb_userdata.
9846 Compatibility note: As many C implementations push function arguments
9847 onto the stack in reverse order, the new library version is likely to
9848 interoperate with programs that have been compiled with the old
9849 pem_password_cb definition (PEM_whatever takes some data that
9850 happens to be on the stack as its last argument, and the callback
9851 just ignores this garbage); but there is no guarantee whatsoever that
9854 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9855 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9856 problems not only on Windows, but also on some Unix platforms.
9857 To avoid problematic command lines, these definitions are now in an
9858 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9859 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9862 *) MIPS III/IV assembler module is reimplemented.
9865 *) More DES library cleanups: remove references to srand/rand and
9866 delete an unused file.
9869 *) Add support for the the free Netwide assembler (NASM) under Win32,
9870 since not many people have MASM (ml) and it can be hard to obtain.
9871 This is currently experimental but it seems to work OK and pass all
9872 the tests. Check out INSTALL.W32 for info.
9875 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9876 without temporary keys kept an extra copy of the server key,
9877 and connections with temporary keys did not free everything in case
9881 *) New function RSA_check_key and new openssl rsa option -check
9882 for verifying the consistency of RSA keys.
9883 [Ulf Moeller, Bodo Moeller]
9885 *) Various changes to make Win32 compile work:
9886 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9887 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9888 comparison" warnings.
9889 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9892 *) Add a debugging option to PKCS#5 v2 key generation function: when
9893 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9894 derived keys are printed to stderr.
9897 *) Copy the flags in ASN1_STRING_dup().
9898 [Roman E. Pavlov <pre@mo.msk.ru>]
9900 *) The x509 application mishandled signing requests containing DSA
9901 keys when the signing key was also DSA and the parameters didn't match.
9903 It was supposed to omit the parameters when they matched the signing key:
9904 the verifying software was then supposed to automatically use the CA's
9905 parameters if they were absent from the end user certificate.
9907 Omitting parameters is no longer recommended. The test was also
9908 the wrong way round! This was probably due to unusual behaviour in
9909 EVP_cmp_parameters() which returns 1 if the parameters match.
9910 This meant that parameters were omitted when they *didn't* match and
9911 the certificate was useless. Certificates signed with 'ca' didn't have
9913 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9915 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9916 The interface is as follows:
9917 Applications can use
9918 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9919 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9920 "off" is now the default.
9921 The library internally uses
9922 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9923 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9924 to disable memory-checking temporarily.
9926 Some inconsistent states that previously were possible (and were
9927 even the default) are now avoided.
9929 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9930 with each memory chunk allocated; this is occasionally more helpful
9931 than just having a counter.
9933 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9935 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9939 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9940 which largely parallels "options", but is for changing API behaviour,
9941 whereas "options" are about protocol behaviour.
9942 Initial "mode" flags are:
9944 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9945 a single record has been written.
9946 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9947 retries use the same buffer location.
9948 (But all of the contents must be
9952 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9955 *) Fix problems with no-hmac etc.
9956 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9958 *) New functions RSA_get_default_method(), RSA_set_method() and
9959 RSA_get_method(). These allows replacement of RSA_METHODs without having
9960 to mess around with the internals of an RSA structure.
9963 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9964 Also really enable memory leak checks in openssl.c and in some
9966 [Chad C. Mulligan, Bodo Moeller]
9968 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9969 up the length of negative integers. This has now been simplified to just
9970 store the length when it is first determined and use it later, rather
9971 than trying to keep track of where data is copied and updating it to
9973 [Steve Henson, reported by Brien Wheeler
9974 <bwheeler@authentica-security.com>]
9976 *) Add a new function PKCS7_signatureVerify. This allows the verification
9977 of a PKCS#7 signature but with the signing certificate passed to the
9978 function itself. This contrasts with PKCS7_dataVerify which assumes the
9979 certificate is present in the PKCS#7 structure. This isn't always the
9980 case: certificates can be omitted from a PKCS#7 structure and be
9981 distributed by "out of band" means (such as a certificate database).
9984 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9985 function prototypes in pem.h, also change util/mkdef.pl to add the
9986 necessary function names.
9989 *) mk1mf.pl (used by Windows builds) did not properly read the
9990 options set by Configure in the top level Makefile, and Configure
9991 was not even able to write more than one option correctly.
9992 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9995 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9996 file to be loaded from a BIO or FILE pointer. The BIO version will
9997 for example allow memory BIOs to contain config info.
10000 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10001 Whoever hopes to achieve shared-library compatibility across versions
10002 must use this, not the compile-time macro.
10003 (Exercise 0.9.4: Which is the minimum library version required by
10005 Note: All this applies only to multi-threaded programs, others don't
10009 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10010 through a BIO pair triggered the default case, i.e.
10011 SSLerr(...,SSL_R_UNKNOWN_STATE).
10014 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10015 can use the SSL library even if none of the specific BIOs is
10019 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10020 for the encoded length.
10021 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10023 *) Add initial documentation of the X509V3 functions.
10026 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10027 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10028 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10029 secure PKCS#8 private key format with a high iteration count.
10032 *) Fix determination of Perl interpreter: A perl or perl5
10033 _directory_ in $PATH was also accepted as the interpreter.
10034 [Ralf S. Engelschall]
10036 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10037 wrong with it but it was very old and did things like calling
10038 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10039 unusual formatting.
10042 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10043 to use the new extension code.
10046 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10047 with macros. This should make it easier to change their form, add extra
10048 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10052 *) Add to configuration table a new entry that can specify an alternative
10053 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10054 according to Mark Crispin <MRC@Panda.COM>.
10058 *) DES CBC did not update the IV. Weird.
10061 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10062 Changing the behaviour of the former might break existing programs --
10063 where IV updating is needed, des_ncbc_encrypt can be used.
10066 *) When bntest is run from "make test" it drives bc to check its
10067 calculations, as well as internally checking them. If an internal check
10068 fails, it needs to cause bc to give a non-zero result or make test carries
10069 on without noticing the failure. Fixed.
10072 *) DES library cleanups.
10075 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10076 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10077 ciphers. NOTE: although the key derivation function has been verified
10078 against some published test vectors it has not been extensively tested
10079 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10083 *) Instead of "mkdir -p", which is not fully portable, use new
10084 Perl script "util/mkdir-p.pl".
10087 *) Rewrite the way password based encryption (PBE) is handled. It used to
10088 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10089 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10090 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10091 the 'parameter' field of the AlgorithmIdentifier is passed to the
10092 underlying key generation function so it must do its own ASN1 parsing.
10093 This has also changed the EVP_PBE_CipherInit() function which now has a
10094 'parameter' argument instead of literal salt and iteration count values
10095 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10098 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10099 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10100 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10101 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10102 value was just used as a "magic string" and not used directly its
10103 value doesn't matter.
10106 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10110 *) "linux-sparc64" configuration (ultrapenguin).
10111 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10112 "linux-sparc" configuration.
10113 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10115 *) config now generates no-xxx options for missing ciphers.
10118 *) Support the EBCDIC character set (work in progress).
10119 File ebcdic.c not yet included because it has a different license.
10120 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10122 *) Support BS2000/OSD-POSIX.
10123 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10125 *) Make callbacks for key generation use void * instead of char *.
10128 *) Make S/MIME samples compile (not yet tested).
10131 *) Additional typesafe stacks.
10134 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10138 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10140 *) New configuration variant "sco5-gcc".
10142 *) Updated some demos.
10143 [Sean O Riordain, Wade Scholine]
10145 *) Add missing BIO_free at exit of pkcs12 application.
10148 *) Fix memory leak in conf.c.
10151 *) Updates for Win32 to assembler version of MD5.
10154 *) Set #! path to perl in apps/der_chop to where we found it
10155 instead of using a fixed path.
10158 *) SHA library changes for irix64-mips4-cc.
10161 *) Improvements for VMS support.
10165 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10167 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10168 This also avoids the problems with SC4.2 and unpatched SC5.
10169 [Andy Polyakov <appro@fy.chalmers.se>]
10171 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10172 These are required because of the typesafe stack would otherwise break
10173 existing code. If old code used a structure member which used to be STACK
10174 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10175 sk_num or sk_value it would produce an error because the num, data members
10176 are not present in STACK_OF. Now it just produces a warning. sk_set
10177 replaces the old method of assigning a value to sk_value
10178 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10179 that does this will no longer work (and should use sk_set instead) but
10180 this could be regarded as a "questionable" behaviour anyway.
10183 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10184 correctly handle encrypted S/MIME data.
10187 *) Change type of various DES function arguments from des_cblock
10188 (which means, in function argument declarations, pointer to char)
10189 to des_cblock * (meaning pointer to array with 8 char elements),
10190 which allows the compiler to do more typechecking; it was like
10191 that back in SSLeay, but with lots of ugly casts.
10193 Introduce new type const_des_cblock.
10196 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10197 problems: find RecipientInfo structure that matches recipient certificate
10198 and initialise the ASN1 structures properly based on passed cipher.
10201 *) Belatedly make the BN tests actually check the results.
10204 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10205 to and from BNs: it was completely broken. New compilation option
10206 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10207 key elements as negative integers.
10210 *) Reorganize and speed up MD5.
10211 [Andy Polyakov <appro@fy.chalmers.se>]
10214 [Richard Levitte <richard@levitte.org>]
10216 *) New option -out to asn1parse to allow the parsed structure to be
10217 output to a file. This is most useful when combined with the -strparse
10218 option to examine the output of things like OCTET STRINGS.
10221 *) Make SSL library a little more fool-proof by not requiring any longer
10222 that SSL_set_{accept,connect}_state be called before
10223 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10224 in many applications because usually everything *appeared* to work as
10225 intended anyway -- now it really works as intended).
10228 *) Move openssl.cnf out of lib/.
10231 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10232 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10233 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10234 [Ralf S. Engelschall]
10236 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10237 handle PKCS#7 enveloped data properly.
10238 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10240 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10241 copying pointers. The cert_st handling is changed by this in
10242 various ways (and thus what used to be known as ctx->default_cert
10243 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10244 any longer when s->cert does not give us what we need).
10245 ssl_cert_instantiate becomes obsolete by this change.
10246 As soon as we've got the new code right (possibly it already is?),
10247 we have solved a couple of bugs of the earlier code where s->cert
10248 was used as if it could not have been shared with other SSL structures.
10250 Note that using the SSL API in certain dirty ways now will result
10251 in different behaviour than observed with earlier library versions:
10252 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10253 does not influence s as it used to.
10255 In order to clean up things more thoroughly, inside SSL_SESSION
10256 we don't use CERT any longer, but a new structure SESS_CERT
10257 that holds per-session data (if available); currently, this is
10258 the peer's certificate chain and, for clients, the server's certificate
10259 and temporary key. CERT holds only those values that can have
10260 meaningful defaults in an SSL_CTX.
10263 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10264 from the internal representation. Various PKCS#7 fixes: remove some
10265 evil casts and set the enc_dig_alg field properly based on the signing
10269 *) Allow PKCS#12 password to be set from the command line or the
10270 environment. Let 'ca' get its config file name from the environment
10271 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10275 *) Allow certificate policies extension to use an IA5STRING for the
10276 organization field. This is contrary to the PKIX definition but
10277 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10281 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10282 without disallowing inline assembler and the like for non-pedantic builds.
10285 *) Support Borland C++ builder.
10286 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10288 *) Support Mingw32.
10291 *) SHA-1 cleanups and performance enhancements.
10292 [Andy Polyakov <appro@fy.chalmers.se>]
10294 *) Sparc v8plus assembler for the bignum library.
10295 [Andy Polyakov <appro@fy.chalmers.se>]
10297 *) Accept any -xxx and +xxx compiler options in Configure.
10300 *) Update HPUX configuration.
10303 *) Add missing sk_<type>_unshift() function to safestack.h
10304 [Ralf S. Engelschall]
10306 *) New function SSL_CTX_use_certificate_chain_file that sets the
10307 "extra_cert"s in addition to the certificate. (This makes sense
10308 only for "PEM" format files, as chains as a whole are not
10312 *) Support verify_depth from the SSL API.
10313 x509_vfy.c had what can be considered an off-by-one-error:
10314 Its depth (which was not part of the external interface)
10315 was actually counting the number of certificates in a chain;
10316 now it really counts the depth.
10319 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10320 instead of X509err, which often resulted in confusing error
10321 messages since the error codes are not globally unique
10322 (e.g. an alleged error in ssl3_accept when a certificate
10323 didn't match the private key).
10325 *) New function SSL_CTX_set_session_id_context that allows to set a default
10326 value (so that you don't need SSL_set_session_id_context for each
10327 connection using the SSL_CTX).
10330 *) OAEP decoding bug fix.
10333 *) Support INSTALL_PREFIX for package builders, as proposed by
10337 *) New Configure options "threads" and "no-threads". For systems
10338 where the proper compiler options are known (currently Solaris
10339 and Linux), "threads" is the default.
10342 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10345 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10346 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10347 such as /usr/local/bin.
10350 *) "make linux-shared" to build shared libraries.
10351 [Niels Poppe <niels@netbox.org>]
10353 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10356 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10357 extension adding in x509 utility.
10360 *) Remove NOPROTO sections and error code comments.
10363 *) Partial rewrite of the DEF file generator to now parse the ANSI
10367 *) New Configure options --prefix=DIR and --openssldir=DIR.
10370 *) Complete rewrite of the error code script(s). It is all now handled
10371 by one script at the top level which handles error code gathering,
10372 header rewriting and C source file generation. It should be much better
10373 than the old method: it now uses a modified version of Ulf's parser to
10374 read the ANSI prototypes in all header files (thus the old K&R definitions
10375 aren't needed for error creation any more) and do a better job of
10376 translating function codes into names. The old 'ASN1 error code imbedded
10377 in a comment' is no longer necessary and it doesn't use .err files which
10378 have now been deleted. Also the error code call doesn't have to appear all
10379 on one line (which resulted in some large lines...).
10382 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10385 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10386 0 (which usually indicates a closed connection), but continue reading.
10389 *) Fix some race conditions.
10392 *) Add support for CRL distribution points extension. Add Certificate
10393 Policies and CRL distribution points documentation.
10396 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10399 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10400 8 of keying material. Merlin has also confirmed interop with this fix
10401 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10402 [Merlin Hughes <merlin@baltimore.ie>]
10404 *) Fix lots of warnings.
10405 [Richard Levitte <levitte@stacken.kth.se>]
10407 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10408 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10409 [Richard Levitte <levitte@stacken.kth.se>]
10411 *) Fix problems with sizeof(long) == 8.
10412 [Andy Polyakov <appro@fy.chalmers.se>]
10414 *) Change functions to ANSI C.
10417 *) Fix typos in error codes.
10418 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10420 *) Remove defunct assembler files from Configure.
10423 *) SPARC v8 assembler BIGNUM implementation.
10424 [Andy Polyakov <appro@fy.chalmers.se>]
10426 *) Support for Certificate Policies extension: both print and set.
10427 Various additions to support the r2i method this uses.
10430 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10431 return a const string when you are expecting an allocated buffer.
10434 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10435 types DirectoryString and DisplayText.
10438 *) Add code to allow r2i extensions to access the configuration database,
10439 add an LHASH database driver and add several ctx helper functions.
10442 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10443 fail when they extended the size of a BIGNUM.
10446 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10447 support typesafe stack.
10450 *) Fix typo in SSL_[gs]et_options().
10451 [Nils Frostberg <nils@medcom.se>]
10453 *) Delete various functions and files that belonged to the (now obsolete)
10454 old X509V3 handling code.
10457 *) New Configure option "rsaref".
10460 *) Don't auto-generate pem.h.
10463 *) Introduce type-safe ASN.1 SETs.
10466 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10467 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10469 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10470 that links with OpenSSL (well at least cause lots of warnings), but fear
10471 not: the conversion is trivial, and it eliminates loads of evil casts. A
10472 few STACKed things have been converted already. Feel free to convert more.
10473 In the fullness of time, I'll do away with the STACK type altogether.
10476 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10477 specified in <certfile> by updating the entry in the index.txt file.
10478 This way one no longer has to edit the index.txt file manually for
10479 revoking a certificate. The -revoke option does the gory details now.
10480 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10482 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10483 `-text' option at all and this way the `-noout -text' combination was
10484 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10485 [Ralf S. Engelschall]
10487 *) Make sure a corresponding plain text error message exists for the
10488 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10489 verify callback function determined that a certificate was revoked.
10490 [Ralf S. Engelschall]
10492 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10493 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10494 all available cipers including rc5, which was forgotten until now.
10495 In order to let the testing shell script know which algorithms
10496 are available, a new (up to now undocumented) command
10497 "openssl list-cipher-commands" is used.
10500 *) Bugfix: s_client occasionally would sleep in select() when
10501 it should have checked SSL_pending() first.
10504 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10505 the raw DSA values prior to ASN.1 encoding.
10508 *) Tweaks to Configure
10509 [Niels Poppe <niels@netbox.org>]
10511 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10515 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10518 *) New config option to avoid instructions that are illegal on the 80386.
10519 The default code is faster, but requires at least a 486.
10522 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10523 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10524 same as SSL2_VERSION anyway.
10527 *) New "-showcerts" option for s_client.
10530 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10531 application. Various cleanups and fixes.
10534 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10535 modify error routines to work internally. Add error codes and PBE init
10536 to library startup routines.
10539 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10540 packing functions to asn1 and evp. Changed function names and error
10541 codes along the way.
10544 *) PKCS12 integration: and so it begins... First of several patches to
10545 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10546 objects to objects.h
10549 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10550 and display support for Thawte strong extranet extension.
10553 *) Add LinuxPPC support.
10554 [Jeff Dubrule <igor@pobox.org>]
10556 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10557 bn_div_words in alpha.s.
10558 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10560 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10561 OAEP isn't supported when OpenSSL is built with RSAref.
10562 [Ulf Moeller <ulf@fitug.de>]
10564 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10565 so they no longer are missing under -DNOPROTO.
10566 [Soren S. Jorvang <soren@t.dk>]
10569 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10571 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10572 doesn't work when the session is reused. Coming soon!
10575 *) Fix a security hole, that allows sessions to be reused in the wrong
10576 context thus bypassing client cert protection! All software that uses
10577 client certs and session caches in multiple contexts NEEDS PATCHING to
10578 allow session reuse! A fuller solution is in the works.
10579 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10581 *) Some more source tree cleanups (removed obsolete files
10582 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10583 permission on "config" script to be executable) and a fix for the INSTALL
10585 [Ulf Moeller <ulf@fitug.de>]
10587 *) Remove some legacy and erroneous uses of malloc, free instead of
10589 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10591 *) Make rsa_oaep_test return non-zero on error.
10592 [Ulf Moeller <ulf@fitug.de>]
10594 *) Add support for native Solaris shared libraries. Configure
10595 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10596 if someone would make that last step automatic.
10597 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10599 *) ctx_size was not built with the right compiler during "make links". Fixed.
10602 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10603 except NULL ciphers". This means the default cipher list will no longer
10604 enable NULL ciphers. They need to be specifically enabled e.g. with
10605 the string "DEFAULT:eNULL".
10608 *) Fix to RSA private encryption routines: if p < q then it would
10609 occasionally produce an invalid result. This will only happen with
10610 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10613 *) Be less restrictive and allow also `perl util/perlpath.pl
10614 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10615 because this way one can also use an interpreter named `perl5' (which is
10616 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10617 installed as `perl').
10618 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10620 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10621 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10623 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10624 advapi32.lib to Win32 build and change the pem test comparison
10625 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10626 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10627 and crypto/des/ede_cbcm_enc.c.
10630 *) DES quad checksum was broken on big-endian architectures. Fixed.
10633 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10634 Win32 test batch file so it (might) work again. The Win32 test batch file
10635 is horrible: I feel ill....
10638 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10639 in e_os.h. Audit of header files to check ANSI and non ANSI
10640 sections: 10 functions were absent from non ANSI section and not exported
10641 from Windows DLLs. Fixed up libeay.num for new functions.
10644 *) Make `openssl version' output lines consistent.
10645 [Ralf S. Engelschall]
10647 *) Fix Win32 symbol export lists for BIO functions: Added
10648 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10649 to ms/libeay{16,32}.def.
10650 [Ralf S. Engelschall]
10652 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10653 fine under Unix and passes some trivial tests I've now added. But the
10654 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10655 added to make sure no one expects that this stuff really works in the
10656 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10657 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10659 [Ralf S. Engelschall]
10661 *) Fix the generation of two part addresses in perl.
10662 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10664 *) Add config entry for Linux on MIPS.
10665 [John Tobey <jtobey@channel1.com>]
10667 *) Make links whenever Configure is run, unless we are on Windoze.
10670 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10671 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10675 *) Add a useful kludge to allow package maintainers to specify compiler and
10676 other platforms details on the command line without having to patch the
10677 Configure script everytime: One now can use ``perl Configure
10678 <id>:<details>'', i.e. platform ids are allowed to have details appended
10679 to them (separated by colons). This is treated as there would be a static
10680 pre-configured entry in Configure's %table under key <id> with value
10681 <details> and ``perl Configure <id>'' is called. So, when you want to
10682 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10683 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10684 now, which overrides the FreeBSD-elf entry on-the-fly.
10685 [Ralf S. Engelschall]
10687 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10690 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10691 on the `perl Configure ...' command line. This way one can compile
10692 OpenSSL libraries with Position Independent Code (PIC) which is needed
10693 for linking it into DSOs.
10694 [Ralf S. Engelschall]
10696 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10700 *) Cleaned up the LICENSE document: The official contact for any license
10701 questions now is the OpenSSL core team under openssl-core@openssl.org.
10702 And add a paragraph about the dual-license situation to make sure people
10703 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10704 to the OpenSSL toolkit.
10705 [Ralf S. Engelschall]
10707 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10708 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10709 Additionally cleaned up the `make links' target: Remove unnecessary
10710 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10711 to speed processing and no longer clutter the display with confusing
10712 stuff. Instead only the actually done links are displayed.
10713 [Ralf S. Engelschall]
10715 *) Permit null encryption ciphersuites, used for authentication only. It used
10716 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10717 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10721 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10722 signed attributes when verifying signatures (this would break them),
10723 the detached data encoding was wrong and public keys obtained using
10724 X509_get_pubkey() weren't freed.
10727 *) Add text documentation for the BUFFER functions. Also added a work around
10728 to a Win95 console bug. This was triggered by the password read stuff: the
10729 last character typed gets carried over to the next fread(). If you were
10730 generating a new cert request using 'req' for example then the last
10731 character of the passphrase would be CR which would then enter the first
10735 *) Added the new `Includes OpenSSL Cryptography Software' button as
10736 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10737 button and can be used by applications based on OpenSSL to show the
10738 relationship to the OpenSSL project.
10739 [Ralf S. Engelschall]
10741 *) Remove confusing variables in function signatures in files
10742 ssl/ssl_lib.c and ssl/ssl.h.
10743 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10745 *) Don't install bss_file.c under PREFIX/include/
10746 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10748 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10749 functions that return function pointers and has support for NT specific
10750 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10751 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10752 unsigned to signed types: this was killing the Win32 compile.
10755 *) Add new certificate file to stack functions,
10756 SSL_add_dir_cert_subjects_to_stack() and
10757 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10758 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10759 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10760 This means that Apache-SSL and similar packages don't have to mess around
10761 to add as many CAs as they want to the preferred list.
10764 *) Experiment with doxygen documentation. Currently only partially applied to
10766 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10767 openssl.doxy as the configuration file.
10770 *) Get rid of remaining C++-style comments which strict C compilers hate.
10771 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10773 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10774 compiled in by default: it has problems with large keys.
10777 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10778 DH private keys and/or callback functions which directly correspond to
10779 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10780 is needed for applications which have to configure certificates on a
10781 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10783 For the RSA certificate situation is makes no difference, but
10784 for the DSA certificate situation this fixes the "no shared cipher"
10785 problem where the OpenSSL cipher selection procedure failed because the
10786 temporary keys were not overtaken from the context and the API provided
10787 no way to reconfigure them.
10788 The new functions now let applications reconfigure the stuff and they
10789 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10790 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10791 non-public-API function ssl_cert_instantiate() is used as a helper
10792 function and also to reduce code redundancy inside ssl_rsa.c.
10793 [Ralf S. Engelschall]
10795 *) Move s_server -dcert and -dkey options out of the undocumented feature
10796 area because they are useful for the DSA situation and should be
10797 recognized by the users.
10798 [Ralf S. Engelschall]
10800 *) Fix the cipher decision scheme for export ciphers: the export bits are
10801 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10802 SSL_EXP_MASK. So, the original variable has to be used instead of the
10803 already masked variable.
10804 [Richard Levitte <levitte@stacken.kth.se>]
10806 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10807 [Richard Levitte <levitte@stacken.kth.se>]
10809 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10810 from `int' to `unsigned int' because it's a length and initialized by
10811 EVP_DigestFinal() which expects an `unsigned int *'.
10812 [Richard Levitte <levitte@stacken.kth.se>]
10814 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10815 script. Instead use the usual Shell->Perl transition trick.
10816 [Ralf S. Engelschall]
10818 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10819 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10820 -noout -modulus' as it's already the case for `openssl rsa -noout
10821 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10822 currently the public key is printed (a decision which was already done by
10823 `openssl dsa -modulus' in the past) which serves a similar purpose.
10824 Additionally the NO_RSA no longer completely removes the whole -modulus
10825 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10827 [Ralf S. Engelschall]
10829 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10830 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10831 [Arne Ansper <arne@ats.cyber.ee>]
10833 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10834 to be added. Now both 'req' and 'ca' can use new objects defined in the
10838 *) Add cool BIO that does syslog (or event log on NT).
10839 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10841 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10842 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10843 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10844 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10847 *) Add preliminary config info for new extension code.
10850 *) Make RSA_NO_PADDING really use no padding.
10851 [Ulf Moeller <ulf@fitug.de>]
10853 *) Generate errors when private/public key check is done.
10856 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10857 for some CRL extensions and new objects added.
10860 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10861 key usage extension and fuller support for authority key id.
10864 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10865 padding method for RSA, which is recommended for new applications in PKCS
10866 #1 v2.0 (RFC 2437, October 1998).
10867 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10868 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10869 against Bleichbacher's attack on RSA.
10870 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10873 *) Updates to the new SSL compression code
10874 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10876 *) Fix so that the version number in the master secret, when passed
10877 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10878 (because the server will not accept higher), that the version number
10879 is 0x03,0x01, not 0x03,0x00
10880 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10882 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10883 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10884 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10887 *) Support for RAW extensions where an arbitrary extension can be
10888 created by including its DER encoding. See apps/openssl.cnf for
10892 *) Make sure latest Perl versions don't interpret some generated C array
10893 code as Perl array code in the crypto/err/err_genc.pl script.
10894 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10896 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10897 not many people have the assembler. Various Win32 compilation fixes and
10898 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10899 build instructions.
10902 *) Modify configure script 'Configure' to automatically create crypto/date.h
10903 file under Win32 and also build pem.h from pem.org. New script
10904 util/mkfiles.pl to create the MINFO file on environments that can't do a
10905 'make files': perl util/mkfiles.pl >MINFO should work.
10908 *) Major rework of DES function declarations, in the pursuit of correctness
10909 and purity. As a result, many evil casts evaporated, and some weirdness,
10910 too. You may find this causes warnings in your code. Zapping your evil
10911 casts will probably fix them. Mostly.
10914 *) Fix for a typo in asn1.h. Bug fix to object creation script
10915 obj_dat.pl. It considered a zero in an object definition to mean
10916 "end of object": none of the objects in objects.h have any zeros
10917 so it wasn't spotted.
10918 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10920 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10921 Masking (CBCM). In the absence of test vectors, the best I have been able
10922 to do is check that the decrypt undoes the encrypt, so far. Send me test
10923 vectors if you have them.
10926 *) Correct calculation of key length for export ciphers (too much space was
10927 allocated for null ciphers). This has not been tested!
10930 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10931 message is now correct (it understands "crypto" and "ssl" on its
10932 command line). There is also now an "update" option. This will update
10933 the util/ssleay.num and util/libeay.num files with any new functions.
10935 perl util/mkdef.pl crypto ssl update
10936 it will update them.
10939 *) Overhauled the Perl interface (perl/*):
10940 - ported BN stuff to OpenSSL's different BN library
10941 - made the perl/ source tree CVS-aware
10942 - renamed the package from SSLeay to OpenSSL (the files still contain
10943 their history because I've copied them in the repository)
10944 - removed obsolete files (the test scripts will be replaced
10945 by better Test::Harness variants in the future)
10946 [Ralf S. Engelschall]
10948 *) First cut for a very conservative source tree cleanup:
10949 1. merge various obsolete readme texts into doc/ssleay.txt
10950 where we collect the old documents and readme texts.
10951 2. remove the first part of files where I'm already sure that we no
10952 longer need them because of three reasons: either they are just temporary
10953 files which were left by Eric or they are preserved original files where
10954 I've verified that the diff is also available in the CVS via "cvs diff
10955 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10956 the crypto/md/ stuff).
10957 [Ralf S. Engelschall]
10959 *) More extension code. Incomplete support for subject and issuer alt
10960 name, issuer and authority key id. Change the i2v function parameters
10961 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10962 what that's for :-) Fix to ASN1 macro which messed up
10963 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10966 *) Preliminary support for ENUMERATED type. This is largely copied from the
10970 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10971 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10973 *) Make sure `make rehash' target really finds the `openssl' program.
10974 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10976 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10977 like to hear about it if this slows down other processors.
10980 *) Add CygWin32 platform information to Configure script.
10981 [Alan Batie <batie@aahz.jf.intel.com>]
10983 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10984 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10986 *) New program nseq to manipulate netscape certificate sequences
10989 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10993 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10994 but the BN code had some problems that would cause failures when
10995 doing certificate verification and some other functions.
10996 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10998 *) Add ASN1 and PEM code to support netscape certificate sequences.
11001 *) Add ASN1 and PEM code to support netscape certificate sequences.
11004 *) Add several PKIX and private extended key usage OIDs.
11007 *) Modify the 'ca' program to handle the new extension code. Modify
11008 openssl.cnf for new extension format, add comments.
11011 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11012 and add a sample to openssl.cnf so req -x509 now adds appropriate
11016 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11017 error code, add initial support to X509_print() and x509 application.
11020 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11021 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11022 stuff is currently isolated and isn't even compiled yet.
11025 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11026 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11027 Removed the versions check from X509 routines when loading extensions:
11028 this allows certain broken certificates that don't set the version
11029 properly to be processed.
11032 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11033 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11034 can still be regenerated with "make depend".
11037 *) Spelling mistake in C version of CAST-128.
11038 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11040 *) Changes to the error generation code. The perl script err-code.pl
11041 now reads in the old error codes and retains the old numbers, only
11042 adding new ones if necessary. It also only changes the .err files if new
11043 codes are added. The makefiles have been modified to only insert errors
11044 when needed (to avoid needlessly modifying header files). This is done
11045 by only inserting errors if the .err file is newer than the auto generated
11046 C file. To rebuild all the error codes from scratch (the old behaviour)
11047 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11048 or delete all the .err files.
11051 *) CAST-128 was incorrectly implemented for short keys. The C version has
11052 been fixed, but is untested. The assembler versions are also fixed, but
11053 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11054 to regenerate it if needed.
11055 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11056 Hagino <itojun@kame.net>]
11058 *) File was opened incorrectly in randfile.c.
11059 [Ulf Möller <ulf@fitug.de>]
11061 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11062 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11063 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11064 al: it's just almost always a UTCTime. Note this patch adds new error
11065 codes so do a "make errors" if there are problems.
11068 *) Correct Linux 1 recognition in config.
11069 [Ulf Möller <ulf@fitug.de>]
11071 *) Remove pointless MD5 hash when using DSA keys in ca.
11072 [Anonymous <nobody@replay.com>]
11074 *) Generate an error if given an empty string as a cert directory. Also
11075 generate an error if handed NULL (previously returned 0 to indicate an
11076 error, but didn't set one).
11077 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11079 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11082 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11083 parameters. This was causing a warning which killed off the Win32 compile.
11086 *) Remove C++ style comments from crypto/bn/bn_local.h.
11087 [Neil Costigan <neil.costigan@celocom.com>]
11089 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11090 based on a text string, looking up short and long names and finally
11091 "dot" format. The "dot" format stuff didn't work. Added new function
11092 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11093 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11094 OID is not part of the table.
11097 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11098 X509_LOOKUP_by_alias().
11101 *) Sort openssl functions by name.
11104 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11105 encryption from sample DSA keys (in case anyone is interested the password
11109 *) Make _all_ *_free functions accept a NULL pointer.
11110 [Frans Heymans <fheymans@isaserver.be>]
11112 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11114 [Anonymous <nobody@replay.com>]
11116 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11117 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11119 *) Don't blow it for numeric -newkey arguments to apps/req.
11120 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11122 *) Temp key "for export" tests were wrong in s3_srvr.c.
11123 [Anonymous <nobody@replay.com>]
11125 *) Add prototype for temp key callback functions
11126 SSL_CTX_set_tmp_{rsa,dh}_callback().
11129 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11130 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11133 *) X509_name_add_entry() freed the wrong thing after an error.
11134 [Arne Ansper <arne@ats.cyber.ee>]
11136 *) rsa_eay.c would attempt to free a NULL context.
11137 [Arne Ansper <arne@ats.cyber.ee>]
11139 *) BIO_s_socket() had a broken should_retry() on Windoze.
11140 [Arne Ansper <arne@ats.cyber.ee>]
11142 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11143 [Arne Ansper <arne@ats.cyber.ee>]
11145 *) Make sure the already existing X509_STORE->depth variable is initialized
11146 in X509_STORE_new(), but document the fact that this variable is still
11147 unused in the certificate verification process.
11148 [Ralf S. Engelschall]
11150 *) Fix the various library and apps files to free up pkeys obtained from
11151 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11154 *) Fix reference counting in X509_PUBKEY_get(). This makes
11155 demos/maurice/example2.c work, amongst others, probably.
11156 [Steve Henson and Ben Laurie]
11158 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11159 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11160 are no longer created. This way we have a single and consistent command
11161 line interface `openssl <command>', similar to `cvs <command>'.
11162 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11164 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11165 BIT STRING wrapper always have zero unused bits.
11168 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11171 *) Make the top-level INSTALL documentation easier to understand.
11174 *) Makefiles updated to exit if an error occurs in a sub-directory
11175 make (including if user presses ^C) [Paul Sutton]
11177 *) Make Montgomery context stuff explicit in RSA data structure.
11180 *) Fix build order of pem and err to allow for generated pem.h.
11183 *) Fix renumbering bug in X509_NAME_delete_entry().
11186 *) Enhanced the err-ins.pl script so it makes the error library number
11187 global and can add a library name. This is needed for external ASN1 and
11188 other error libraries.
11191 *) Fixed sk_insert which never worked properly.
11194 *) Fix ASN1 macros so they can handle indefinite length constructed
11195 EXPLICIT tags. Some non standard certificates use these: they can now
11199 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11200 into a single doc/ssleay.txt bundle. This way the information is still
11201 preserved but no longer messes up this directory. Now it's new room for
11202 the new set of documentation files.
11203 [Ralf S. Engelschall]
11205 *) SETs were incorrectly DER encoded. This was a major pain, because they
11206 shared code with SEQUENCEs, which aren't coded the same. This means that
11207 almost everything to do with SETs or SEQUENCEs has either changed name or
11208 number of arguments.
11209 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11211 *) Fix test data to work with the above.
11214 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11215 was already fixed by Eric for 0.9.1 it seems.
11216 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11218 *) Autodetect FreeBSD3.
11221 *) Fix various bugs in Configure. This affects the following platforms:
11225 unixware-2.0-pentium
11229 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11230 before they are needed.
11233 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11237 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11239 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11240 changed SSLeay to OpenSSL in version strings.
11241 [Ralf S. Engelschall]
11243 *) Some fixups to the top-level documents.
11246 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11247 because the symlink to include/ was missing.
11248 [Ralf S. Engelschall]
11250 *) Incorporated the popular no-RSA/DSA-only patches
11251 which allow to compile a RSA-free SSLeay.
11252 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11254 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11255 when "ssleay" is still not found.
11256 [Ralf S. Engelschall]
11258 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11259 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11261 *) Updated the README file.
11262 [Ralf S. Engelschall]
11264 *) Added various .cvsignore files in the CVS repository subdirs
11265 to make a "cvs update" really silent.
11266 [Ralf S. Engelschall]
11268 *) Recompiled the error-definition header files and added
11269 missing symbols to the Win32 linker tables.
11270 [Ralf S. Engelschall]
11272 *) Cleaned up the top-level documents;
11273 o new files: CHANGES and LICENSE
11274 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11275 o merged COPYRIGHT into LICENSE
11276 o removed obsolete TODO file
11277 o renamed MICROSOFT to INSTALL.W32
11278 [Ralf S. Engelschall]
11280 *) Removed dummy files from the 0.9.1b source tree:
11281 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11282 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11283 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11284 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11285 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11286 [Ralf S. Engelschall]
11288 *) Added various platform portability fixes.
11291 *) The Genesis of the OpenSSL rpject:
11292 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11293 Young and Tim J. Hudson created while they were working for C2Net until
11295 [The OpenSSL Project]
11298 Changes between 0.9.0b and 0.9.1b [not released]
11300 *) Updated a few CA certificates under certs/
11303 *) Changed some BIGNUM api stuff.
11306 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11307 DGUX x86, Linux Alpha, etc.
11310 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11311 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11315 *) Add -strparse option to asn1pars program which parses nested
11317 [Dr Stephen Henson <shenson@bigfoot.com>]
11319 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11322 *) DSA fix for "ca" program.
11325 *) Added "-genkey" option to "dsaparam" program.
11328 *) Added RIPE MD160 (rmd160) message digest.
11331 *) Added -a (all) option to "ssleay version" command.
11334 *) Added PLATFORM define which is the id given to Configure.
11337 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11340 *) Extended the ASN.1 parser routines.
11343 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11346 *) Added a BN_CTX to the BN library.
11349 *) Fixed the weak key values in DES library
11352 *) Changed API in EVP library for cipher aliases.
11355 *) Added support for RC2/64bit cipher.
11358 *) Converted the lhash library to the crypto/mem.c functions.
11361 *) Added more recognized ASN.1 object ids.
11364 *) Added more RSA padding checks for SSL/TLS.
11367 *) Added BIO proxy/filter functionality.
11370 *) Added extra_certs to SSL_CTX which can be used
11371 send extra CA certificates to the client in the CA cert chain sending
11372 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11375 *) Now Fortezza is denied in the authentication phase because
11376 this is key exchange mechanism is not supported by SSLeay at all.
11379 *) Additional PKCS1 checks.
11382 *) Support the string "TLSv1" for all TLS v1 ciphers.
11385 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11386 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11389 *) Fixed a few memory leaks.
11392 *) Fixed various code and comment typos.
11395 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11396 bytes sent in the client random.
11397 [Edward Bishop <ebishop@spyglass.com>]