5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) State machine rewrite. The state machine code has been significantly
8 refactored in order to remove much duplication of code and solve issues
9 with the old code (see ssl/statem/README for further details). This change
10 does have some associated API changes. Notably the SSL_state() function
11 has been removed and replaced by SSL_get_state which now returns an
12 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
13 altogether. The previous handshake states defined in ssl.h and ssl3.h have
14 been redefined to be the nearest equivalent OSS_HANDSHAKE_STATE value. Not
15 all states have an equivalent value, (e.g. SSL_ST_CW_FLUSH). New
16 application code should not use the old handshake state values, but should
17 instead use OSSL_HANDSHAKE_STATE.
20 *) The demo files in crypto/threads were moved to demo/threads.
23 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron and sureware.
26 *) New ASN.1 embed macro.
28 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
29 structure is not allocated: it is part of the parent. That is instead of
37 This reduces memory fragmentation and make it impossible to accidentally
38 set a mandatory field to NULL.
40 This currently only works for some fields specifically a SEQUENCE, CHOICE,
41 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
42 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
46 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
49 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
50 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
51 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
52 DES and RC4 ciphersuites.
55 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
56 This changes the decoding behaviour for some invalid messages,
57 though the change is mostly in the more lenient direction, and
58 legacy behaviour is preserved as much as possible.
60 *) Fix no-stdio build.
61 [ David Woodhouse <David.Woodhouse@intel.com> and also
62 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
64 *) New testing framework
65 The testing framework has been largely rewritten and is now using
66 perl and the perl modules Test::Harness and an extended variant of
67 Test::More called OpenSSL::Test to do its work. All test scripts in
68 test/ have been rewritten into test recipes, and all direct calls to
69 executables in test/Makefile have become individual recipes using the
70 simplified testing OpenSSL::Test::Simple.
72 For documentation on our testing modules, do:
74 perldoc test/testlib/OpenSSL/Test/Simple.pm
75 perldoc test/testlib/OpenSSL/Test.pm
79 *) In DSA_generate_parameters_ex, if the provided seed is too short,
81 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
83 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
84 from RFC4279, RFC4785, RFC5487, RFC5489.
86 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
87 original RSA_PSK patch.
90 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
91 era flag was never set throughout the codebase (only read). Also removed
92 SSL3_FLAGS_POP_BUFFER which was only used if
93 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
96 *) Changed the default name options in the "ca", "crl", "req" and "x509"
97 to be "oneline" instead of "compat".
100 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
101 not aware of clients that still exhibit this bug, and the workaround
102 hasn't been working properly for a while.
105 *) The return type of BIO_number_read() and BIO_number_written() as well as
106 the corresponding num_read and num_write members in the BIO structure has
107 changed from unsigned long to uint64_t. On platforms where an unsigned
108 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
112 *) Given the pervasive nature of TLS extensions it is inadvisable to run
113 OpenSSL without support for them. It also means that maintaining
114 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
115 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
118 *) Removed support for the two export grade static DH ciphersuites
119 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
120 were newly added (along with a number of other static DH ciphersuites) to
121 1.0.2. However the two export ones have *never* worked since they were
122 introduced. It seems strange in any case to be adding new export
123 ciphersuites, and given "logjam" it also does not seem correct to fix them.
126 *) Version negotiation has been rewritten. In particular SSLv23_method(),
127 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
128 and turned into macros which simply call the new preferred function names
129 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
130 should use the new names instead. Also as part of this change the ssl23.h
131 header file has been removed.
134 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
135 code and the associated standard is no longer considered fit-for-purpose.
138 *) RT2547 was closed. When generating a private key, try to make the
139 output file readable only by the owner. This behavior change might
140 be noticeable when interacting with other software.
142 *) Added HTTP GET support to the ocsp command.
145 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
148 *) Added support for TLS extended master secret from
149 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
150 initial patch which was a great help during development.
153 *) All libssl internal structures have been removed from the public header
154 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
155 now redundant). Users should not attempt to access internal structures
156 directly. Instead they should use the provided API functions.
159 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
160 Access to deprecated functions can be re-enabled by running config with
161 "enable-deprecated". In addition applications wishing to use deprecated
162 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
163 will, by default, disable some transitive includes that previously existed
164 in the header files (e.g. ec.h will no longer, by default, include bn.h)
167 *) Added support for OCB mode. OpenSSL has been granted a patent license
168 compatible with the OpenSSL license for use of OCB. Details are available
169 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
170 for OCB can be removed by calling config with no-ocb.
173 *) SSLv2 support has been removed. It still supports receiving a SSLv2
174 compatible client hello.
177 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
178 done while fixing the error code for the key-too-small case.
179 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
181 *) CA.sh has been removmed; use CA.pl instead.
184 *) Removed old DES API.
187 *) Remove various unsupported platforms:
193 Sinix/ReliantUNIX RM400
198 16-bit platforms such as WIN16
201 *) Clean up OPENSSL_NO_xxx #define's
202 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
203 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
204 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
205 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
206 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
207 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
208 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
209 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
210 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
211 Remove MS_STATIC; it's a relic from platforms <32 bits.
214 *) Cleaned up dead code
215 Remove all but one '#ifdef undef' which is to be looked at.
218 *) Clean up calling of xxx_free routines.
219 Just like free(), fix most of the xxx_free routines to accept
220 NULL. Remove the non-null checks from callers. Save much code.
223 *) Add secure heap for storage of private keys (when possible).
224 Add BIO_s_secmem(), CBIGNUM, etc.
225 Contributed by Akamai Technologies under our Corporate CLA.
228 *) Experimental support for a new, fast, unbiased prime candidate generator,
229 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
230 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
232 *) New output format NSS in the sess_id command line tool. This allows
233 exporting the session id and the master key in NSS keylog format.
234 [Martin Kaiser <martin@kaiser.cx>]
236 *) Harmonize version and its documentation. -f flag is used to display
238 [mancha <mancha1@zoho.com>]
240 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
241 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
242 [mancha <mancha1@zoho.com>]
244 *) Fix some double frees. These are not thought to be exploitable.
245 [mancha <mancha1@zoho.com>]
247 *) A missing bounds check in the handling of the TLS heartbeat extension
248 can be used to reveal up to 64k of memory to a connected client or
251 Thanks for Neel Mehta of Google Security for discovering this bug and to
252 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
253 preparing the fix (CVE-2014-0160)
254 [Adam Langley, Bodo Moeller]
256 *) Fix for the attack described in the paper "Recovering OpenSSL
257 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
258 by Yuval Yarom and Naomi Benger. Details can be obtained from:
259 http://eprint.iacr.org/2014/140
261 Thanks to Yuval Yarom and Naomi Benger for discovering this
262 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
263 [Yuval Yarom and Naomi Benger]
265 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
266 this fixes a limitation in previous versions of OpenSSL.
269 *) Experimental encrypt-then-mac support.
271 Experimental support for encrypt then mac from
272 draft-gutmann-tls-encrypt-then-mac-02.txt
274 To enable it set the appropriate extension number (0x42 for the test
275 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
277 For non-compliant peers (i.e. just about everything) this should have no
280 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
284 *) Add EVP support for key wrapping algorithms, to avoid problems with
285 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
286 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
287 algorithms and include tests cases.
290 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
294 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
295 MGF1 digest and OAEP label.
298 *) Make openssl verify return errors.
299 [Chris Palmer <palmer@google.com> and Ben Laurie]
301 *) New function ASN1_TIME_diff to calculate the difference between two
302 ASN1_TIME structures or one structure and the current time.
305 *) Update fips_test_suite to support multiple command line options. New
306 test to induce all self test errors in sequence and check expected
310 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
311 sign or verify all in one operation.
314 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
315 test programs and fips_test_suite. Includes functionality to parse
316 the minimal script output of fipsalgest.pl directly.
319 *) Add authorisation parameter to FIPS_module_mode_set().
322 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
325 *) Use separate DRBG fields for internal and external flags. New function
326 FIPS_drbg_health_check() to perform on demand health checking. Add
327 generation tests to fips_test_suite with reduced health check interval to
328 demonstrate periodic health checking. Add "nodh" option to
329 fips_test_suite to skip very slow DH test.
332 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
336 *) More extensive health check for DRBG checking many more failure modes.
337 New function FIPS_selftest_drbg_all() to handle every possible DRBG
338 combination: call this in fips_test_suite.
341 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
342 and POST to handle Dual EC cases.
345 *) Add support for canonical generation of DSA parameter 'g'. See
348 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
349 POST to handle HMAC cases.
352 *) Add functions FIPS_module_version() and FIPS_module_version_text()
353 to return numerical and string versions of the FIPS module number.
356 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
357 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
358 outside the validated module in the FIPS capable OpenSSL.
361 *) Minor change to DRBG entropy callback semantics. In some cases
362 there is no multiple of the block length between min_len and
363 max_len. Allow the callback to return more than max_len bytes
364 of entropy but discard any extra: it is the callback's responsibility
365 to ensure that the extra data discarded does not impact the
366 requested amount of entropy.
369 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
370 information in FIPS186-3, SP800-57 and SP800-131A.
373 *) CCM support via EVP. Interface is very similar to GCM case except we
374 must supply all data in one chunk (i.e. no update, final) and the
375 message length must be supplied if AAD is used. Add algorithm test
379 *) Initial version of POST overhaul. Add POST callback to allow the status
380 of POST to be monitored and/or failures induced. Modify fips_test_suite
381 to use callback. Always run all selftests even if one fails.
384 *) XTS support including algorithm test driver in the fips_gcmtest program.
385 Note: this does increase the maximum key length from 32 to 64 bytes but
386 there should be no binary compatibility issues as existing applications
387 will never use XTS mode.
390 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
391 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
392 performs algorithm blocking for unapproved PRNG types. Also do not
393 set PRNG type in FIPS_mode_set(): leave this to the application.
394 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
395 the standard OpenSSL PRNG: set additional data to a date time vector.
398 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
399 This shouldn't present any incompatibility problems because applications
400 shouldn't be using these directly and any that are will need to rethink
401 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
404 *) Extensive self tests and health checking required by SP800-90 DRBG.
405 Remove strength parameter from FIPS_drbg_instantiate and always
406 instantiate at maximum supported strength.
409 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
412 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
415 *) New function DH_compute_key_padded() to compute a DH key and pad with
416 leading zeroes if needed: this complies with SP800-56A et al.
419 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
420 anything, incomplete, subject to change and largely untested at present.
423 *) Modify fipscanisteronly build option to only build the necessary object
424 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
427 *) Add experimental option FIPSSYMS to give all symbols in
428 fipscanister.o and FIPS or fips prefix. This will avoid
429 conflicts with future versions of OpenSSL. Add perl script
430 util/fipsas.pl to preprocess assembly language source files
431 and rename any affected symbols.
434 *) Add selftest checks and algorithm block of non-fips algorithms in
435 FIPS mode. Remove DES2 from selftests.
438 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
439 return internal method without any ENGINE dependencies. Add new
440 tiny fips sign and verify functions.
443 *) New build option no-ec2m to disable characteristic 2 code.
446 *) New build option "fipscanisteronly". This only builds fipscanister.o
447 and (currently) associated fips utilities. Uses the file Makefile.fips
448 instead of Makefile.org as the prototype.
451 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
452 Update fips_gcmtest to use IV generator.
455 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
456 setting output buffer to NULL. The *Final function must be
457 called although it will not retrieve any additional data. The tag
458 can be set or retrieved with a ctrl. The IV length is by default 12
459 bytes (96 bits) but can be set to an alternative value. If the IV
460 length exceeds the maximum IV length (currently 16 bytes) it cannot be
464 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
465 underlying do_cipher function handles all cipher semantics itself
466 including padding and finalisation. This is useful if (for example)
467 an ENGINE cipher handles block padding itself. The behaviour of
468 do_cipher is subtly changed if this flag is set: the return value
469 is the number of characters written to the output buffer (zero is
470 no longer an error code) or a negative error code. Also if the
471 input buffer is NULL and length 0 finalisation should be performed.
474 *) If a candidate issuer certificate is already part of the constructed
475 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
478 *) Improve forward-security support: add functions
480 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
481 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
483 for use by SSL/TLS servers; the callback function will be called whenever a
484 new session is created, and gets to decide whether the session may be
485 cached to make it resumable (return 0) or not (return 1). (As by the
486 SSL/TLS protocol specifications, the session_id sent by the server will be
487 empty to indicate that the session is not resumable; also, the server will
488 not generate RFC 4507 (RFC 5077) session tickets.)
490 A simple reasonable callback implementation is to return is_forward_secure.
491 This parameter will be set to 1 or 0 depending on the ciphersuite selected
492 by the SSL/TLS server library, indicating whether it can provide forward
494 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
496 *) New -verify_name option in command line utilities to set verification
500 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
501 Add CMAC pkey methods.
504 *) Experimental renegotiation in s_server -www mode. If the client
505 browses /reneg connection is renegotiated. If /renegcert it is
506 renegotiated requesting a certificate.
509 *) Add an "external" session cache for debugging purposes to s_server. This
510 should help trace issues which normally are only apparent in deployed
511 multi-process servers.
514 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
515 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
516 BIO_set_cipher() and some obscure PEM functions were changed so they
517 can now return an error. The RAND changes required a change to the
518 RAND_METHOD structure.
521 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
522 a gcc attribute to warn if the result of a function is ignored. This
523 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
524 whose return value is often ignored.
527 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
529 *) Alternate chains certificate forgery
531 During certificate verfification, OpenSSL will attempt to find an
532 alternative certificate chain if the first attempt to build such a chain
533 fails. An error in the implementation of this logic can mean that an
534 attacker could cause certain checks on untrusted certificates to be
535 bypassed, such as the CA flag, enabling them to use a valid leaf
536 certificate to act as a CA and "issue" an invalid certificate.
538 This issue was reported to OpenSSL by Adam Langley/David Benjamin
542 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
544 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
545 incompatibility in the handling of HMAC. The previous ABI has now been
549 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
551 *) Malformed ECParameters causes infinite loop
553 When processing an ECParameters structure OpenSSL enters an infinite loop
554 if the curve specified is over a specially malformed binary polynomial
557 This can be used to perform denial of service against any
558 system which processes public keys, certificate requests or
559 certificates. This includes TLS clients and TLS servers with
560 client authentication enabled.
562 This issue was reported to OpenSSL by Joseph Barr-Pixton.
566 *) Exploitable out-of-bounds read in X509_cmp_time
568 X509_cmp_time does not properly check the length of the ASN1_TIME
569 string and can read a few bytes out of bounds. In addition,
570 X509_cmp_time accepts an arbitrary number of fractional seconds in the
573 An attacker can use this to craft malformed certificates and CRLs of
574 various sizes and potentially cause a segmentation fault, resulting in
575 a DoS on applications that verify certificates or CRLs. TLS clients
576 that verify CRLs are affected. TLS clients and servers with client
577 authentication enabled may be affected if they use custom verification
580 This issue was reported to OpenSSL by Robert Swiecki (Google), and
581 independently by Hanno Böck.
585 *) PKCS7 crash with missing EnvelopedContent
587 The PKCS#7 parsing code does not handle missing inner EncryptedContent
588 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
589 with missing content and trigger a NULL pointer dereference on parsing.
591 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
592 structures from untrusted sources are affected. OpenSSL clients and
593 servers are not affected.
595 This issue was reported to OpenSSL by Michal Zalewski (Google).
599 *) CMS verify infinite loop with unknown hash function
601 When verifying a signedData message the CMS code can enter an infinite loop
602 if presented with an unknown hash function OID. This can be used to perform
603 denial of service against any system which verifies signedData messages using
605 This issue was reported to OpenSSL by Johannes Bauer.
609 *) Race condition handling NewSessionTicket
611 If a NewSessionTicket is received by a multi-threaded client when attempting to
612 reuse a previous ticket then a race condition can occur potentially leading to
613 a double free of the ticket data.
617 *) Only support 256-bit or stronger elliptic curves with the
618 'ecdh_auto' setting (server) or by default (client). Of supported
619 curves, prefer P-256 (both).
622 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
624 *) ClientHello sigalgs DoS fix
626 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
627 invalid signature algorithms extension a NULL pointer dereference will
628 occur. This can be exploited in a DoS attack against the server.
630 This issue was was reported to OpenSSL by David Ramos of Stanford
633 [Stephen Henson and Matt Caswell]
635 *) Multiblock corrupted pointer fix
637 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
638 feature only applies on 64 bit x86 architecture platforms that support AES
639 NI instructions. A defect in the implementation of "multiblock" can cause
640 OpenSSL's internal write buffer to become incorrectly set to NULL when
641 using non-blocking IO. Typically, when the user application is using a
642 socket BIO for writing, this will only result in a failed connection.
643 However if some other BIO is used then it is likely that a segmentation
644 fault will be triggered, thus enabling a potential DoS attack.
646 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
650 *) Segmentation fault in DTLSv1_listen fix
652 The DTLSv1_listen function is intended to be stateless and processes the
653 initial ClientHello from many peers. It is common for user code to loop
654 over the call to DTLSv1_listen until a valid ClientHello is received with
655 an associated cookie. A defect in the implementation of DTLSv1_listen means
656 that state is preserved in the SSL object from one invocation to the next
657 that can lead to a segmentation fault. Errors processing the initial
658 ClientHello can trigger this scenario. An example of such an error could be
659 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
662 This issue was reported to OpenSSL by Per Allansson.
666 *) Segmentation fault in ASN1_TYPE_cmp fix
668 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
669 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
670 certificate signature algorithm consistency this can be used to crash any
671 certificate verification operation and exploited in a DoS attack. Any
672 application which performs certificate verification is vulnerable including
673 OpenSSL clients and servers which enable client authentication.
677 *) Segmentation fault for invalid PSS parameters fix
679 The signature verification routines will crash with a NULL pointer
680 dereference if presented with an ASN.1 signature using the RSA PSS
681 algorithm and invalid parameters. Since these routines are used to verify
682 certificate signature algorithms this can be used to crash any
683 certificate verification operation and exploited in a DoS attack. Any
684 application which performs certificate verification is vulnerable including
685 OpenSSL clients and servers which enable client authentication.
687 This issue was was reported to OpenSSL by Brian Carpenter.
691 *) ASN.1 structure reuse memory corruption fix
693 Reusing a structure in ASN.1 parsing may allow an attacker to cause
694 memory corruption via an invalid write. Such reuse is and has been
695 strongly discouraged and is believed to be rare.
697 Applications that parse structures containing CHOICE or ANY DEFINED BY
698 components may be affected. Certificate parsing (d2i_X509 and related
699 functions) are however not affected. OpenSSL clients and servers are
704 *) PKCS7 NULL pointer dereferences fix
706 The PKCS#7 parsing code does not handle missing outer ContentInfo
707 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
708 missing content and trigger a NULL pointer dereference on parsing.
710 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
711 otherwise parse PKCS#7 structures from untrusted sources are
712 affected. OpenSSL clients and servers are not affected.
714 This issue was reported to OpenSSL by Michal Zalewski (Google).
718 *) DoS via reachable assert in SSLv2 servers fix
720 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
721 servers that both support SSLv2 and enable export cipher suites by sending
722 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
724 This issue was discovered by Sean Burford (Google) and Emilia Käsper
725 (OpenSSL development team).
729 *) Empty CKE with client auth and DHE fix
731 If client auth is used then a server can seg fault in the event of a DHE
732 ciphersuite being selected and a zero length ClientKeyExchange message
733 being sent by the client. This could be exploited in a DoS attack.
737 *) Handshake with unseeded PRNG fix
739 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
740 with an unseeded PRNG. The conditions are:
741 - The client is on a platform where the PRNG has not been seeded
742 automatically, and the user has not seeded manually
743 - A protocol specific client method version has been used (i.e. not
744 SSL_client_methodv23)
745 - A ciphersuite is used that does not require additional random data from
746 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
748 If the handshake succeeds then the client random that has been used will
749 have been generated from a PRNG with insufficient entropy and therefore the
750 output may be predictable.
752 For example using the following command with an unseeded openssl will
753 succeed on an unpatched platform:
755 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
759 *) Use After Free following d2i_ECPrivatekey error fix
761 A malformed EC private key file consumed via the d2i_ECPrivateKey function
762 could cause a use after free condition. This, in turn, could cause a double
763 free in several private key parsing functions (such as d2i_PrivateKey
764 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
765 for applications that receive EC private keys from untrusted
766 sources. This scenario is considered rare.
768 This issue was discovered by the BoringSSL project and fixed in their
773 *) X509_to_X509_REQ NULL pointer deref fix
775 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
776 the certificate key is invalid. This function is rarely used in practice.
778 This issue was discovered by Brian Carpenter.
782 *) Removed the export ciphers from the DEFAULT ciphers
785 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
787 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
788 ARMv5 through ARMv8, as opposite to "locking" it to single one.
789 So far those who have to target multiple plaforms would compromise
790 and argue that binary targeting say ARMv5 would still execute on
791 ARMv8. "Universal" build resolves this compromise by providing
792 near-optimal performance even on newer platforms.
795 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
796 (other platforms pending).
797 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
799 *) Add support for the SignedCertificateTimestampList certificate and
800 OCSP response extensions from RFC6962.
803 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
804 for corner cases. (Certain input points at infinity could lead to
805 bogus results, with non-infinity inputs mapped to infinity too.)
808 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
809 This covers AES, SHA256/512 and GHASH. "Initial" means that most
810 common cases are optimized and there still is room for further
811 improvements. Vector Permutation AES for Altivec is also added.
814 *) Add support for little-endian ppc64 Linux target.
815 [Marcelo Cerri (IBM)]
817 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
818 SHA1, SHA256 and GHASH. "Initial" means that most common cases
819 are optimized and there still is room for further improvements.
820 Both 32- and 64-bit modes are supported.
821 [Andy Polyakov, Ard Biesheuvel (Linaro)]
823 *) Improved ARMv7 NEON support.
826 *) Support for SPARC Architecture 2011 crypto extensions, first
827 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
828 SHA256/512, MD5, GHASH and modular exponentiation.
829 [Andy Polyakov, David Miller]
831 *) Accelerated modular exponentiation for Intel processors, a.k.a.
833 [Shay Gueron & Vlad Krasnov (Intel Corp)]
835 *) Support for new and upcoming Intel processors, including AVX2,
836 BMI and SHA ISA extensions. This includes additional "stitched"
837 implementations, AESNI-SHA256 and GCM, and multi-buffer support
840 This work was sponsored by Intel Corp.
843 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
844 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
845 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
848 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
849 this fixes a limiation in previous versions of OpenSSL.
852 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
853 MGF1 digest and OAEP label.
856 *) Add EVP support for key wrapping algorithms, to avoid problems with
857 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
858 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
859 algorithms and include tests cases.
862 *) Add functions to allocate and set the fields of an ECDSA_METHOD
864 [Douglas E. Engert, Steve Henson]
866 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
867 difference in days and seconds between two tm or ASN1_TIME structures.
870 *) Add -rev test option to s_server to just reverse order of characters
871 received by client and send back to server. Also prints an abbreviated
872 summary of the connection parameters.
875 *) New option -brief for s_client and s_server to print out a brief summary
876 of connection parameters.
879 *) Add callbacks for arbitrary TLS extensions.
880 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
882 *) New option -crl_download in several openssl utilities to download CRLs
883 from CRLDP extension in certificates.
886 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
889 *) New function X509_CRL_diff to generate a delta CRL from the difference
890 of two full CRLs. Add support to "crl" utility.
893 *) New functions to set lookup_crls function and to retrieve
894 X509_STORE from X509_STORE_CTX.
897 *) Print out deprecated issuer and subject unique ID fields in
901 *) Extend OCSP I/O functions so they can be used for simple general purpose
902 HTTP as well as OCSP. New wrapper function which can be used to download
903 CRLs using the OCSP API.
906 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
909 *) SSL_CONF* functions. These provide a common framework for application
910 configuration using configuration files or command lines.
913 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
914 message callback and prints the results. Needs compile time option
915 "enable-ssl-trace". New options to s_client and s_server to enable
919 *) New ctrl and macro to retrieve supported points extensions.
920 Print out extension in s_server and s_client.
923 *) New functions to retrieve certificate signature and signature
927 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
931 *) New Suite B modes for TLS code. These use and enforce the requirements
932 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
933 only use Suite B curves. The Suite B modes can be set by using the
934 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
937 *) New chain verification flags for Suite B levels of security. Check
938 algorithms are acceptable when flags are set in X509_verify_cert.
941 *) Make tls1_check_chain return a set of flags indicating checks passed
942 by a certificate chain. Add additional tests to handle client
943 certificates: checks for matching certificate type and issuer name
947 *) If an attempt is made to use a signature algorithm not in the peer
948 preference list abort the handshake. If client has no suitable
949 signature algorithms in response to a certificate request do not
953 *) If server EC tmp key is not in client preference list abort handshake.
956 *) Add support for certificate stores in CERT structure. This makes it
957 possible to have different stores per SSL structure or one store in
958 the parent SSL_CTX. Include distint stores for certificate chain
959 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
960 to build and store a certificate chain in CERT structure: returing
961 an error if the chain cannot be built: this will allow applications
962 to test if a chain is correctly configured.
964 Note: if the CERT based stores are not set then the parent SSL_CTX
965 store is used to retain compatibility with existing behaviour.
969 *) New function ssl_set_client_disabled to set a ciphersuite disabled
970 mask based on the current session, check mask when sending client
971 hello and checking the requested ciphersuite.
974 *) New ctrls to retrieve and set certificate types in a certificate
975 request message. Print out received values in s_client. If certificate
976 types is not set with custom values set sensible values based on
977 supported signature algorithms.
980 *) Support for distinct client and server supported signature algorithms.
983 *) Add certificate callback. If set this is called whenever a certificate
984 is required by client or server. An application can decide which
985 certificate chain to present based on arbitrary criteria: for example
986 supported signature algorithms. Add very simple example to s_server.
987 This fixes many of the problems and restrictions of the existing client
988 certificate callback: for example you can now clear an existing
989 certificate and specify the whole chain.
992 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
993 the certificate can be used for (if anything). Set valid_flags field
994 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
995 to have similar checks in it.
997 Add new "cert_flags" field to CERT structure and include a "strict mode".
998 This enforces some TLS certificate requirements (such as only permitting
999 certificate signature algorithms contained in the supported algorithms
1000 extension) which some implementations ignore: this option should be used
1001 with caution as it could cause interoperability issues.
1004 *) Update and tidy signature algorithm extension processing. Work out
1005 shared signature algorithms based on preferences and peer algorithms
1006 and print them out in s_client and s_server. Abort handshake if no
1007 shared signature algorithms.
1010 *) Add new functions to allow customised supported signature algorithms
1011 for SSL and SSL_CTX structures. Add options to s_client and s_server
1015 *) New function SSL_certs_clear() to delete all references to certificates
1016 from an SSL structure. Before this once a certificate had been added
1017 it couldn't be removed.
1020 *) Integrate hostname, email address and IP address checking with certificate
1021 verification. New verify options supporting checking in opensl utility.
1024 *) Fixes and wildcard matching support to hostname and email checking
1025 functions. Add manual page.
1026 [Florian Weimer (Red Hat Product Security Team)]
1028 *) New functions to check a hostname email or IP address against a
1029 certificate. Add options x509 utility to print results of checks against
1033 *) Fix OCSP checking.
1034 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1036 *) Initial experimental support for explicitly trusted non-root CAs.
1037 OpenSSL still tries to build a complete chain to a root but if an
1038 intermediate CA has a trust setting included that is used. The first
1039 setting is used: whether to trust (e.g., -addtrust option to the x509
1043 *) Add -trusted_first option which attempts to find certificates in the
1044 trusted store even if an untrusted chain is also supplied.
1047 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1048 platform support for Linux and Android.
1051 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1054 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1055 When in FIPS mode the approved implementations are used as normal,
1056 when not in FIPS mode the internal unapproved versions are used instead.
1057 This means that the FIPS capable OpenSSL isn't forced to use the
1058 (often lower performance) FIPS implementations outside FIPS mode.
1061 *) Transparently support X9.42 DH parameters when calling
1062 PEM_read_bio_DHparameters. This means existing applications can handle
1063 the new parameter format automatically.
1066 *) Initial experimental support for X9.42 DH parameter format: mainly
1067 to support use of 'q' parameter for RFC5114 parameters.
1070 *) Add DH parameters from RFC5114 including test data to dhtest.
1073 *) Support for automatic EC temporary key parameter selection. If enabled
1074 the most preferred EC parameters are automatically used instead of
1075 hardcoded fixed parameters. Now a server just has to call:
1076 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1077 support ECDH and use the most appropriate parameters.
1080 *) Enhance and tidy EC curve and point format TLS extension code. Use
1081 static structures instead of allocation if default values are used.
1082 New ctrls to set curves we wish to support and to retrieve shared curves.
1083 Print out shared curves in s_server. New options to s_server and s_client
1084 to set list of supported curves.
1087 *) New ctrls to retrieve supported signature algorithms and
1088 supported curve values as an array of NIDs. Extend openssl utility
1089 to print out received values.
1092 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1093 between NIDs and the more common NIST names such as "P-256". Enhance
1094 ecparam utility and ECC method to recognise the NIST names for curves.
1097 *) Enhance SSL/TLS certificate chain handling to support different
1098 chains for each certificate instead of one chain in the parent SSL_CTX.
1101 *) Support for fixed DH ciphersuite client authentication: where both
1102 server and client use DH certificates with common parameters.
1105 *) Support for fixed DH ciphersuites: those requiring DH server
1109 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1111 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1112 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1113 X509_CINF_get_signature were reverted post internal team review.
1115 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1117 *) Build fixes for the Windows and OpenVMS platforms
1118 [Matt Caswell and Richard Levitte]
1120 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1122 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1123 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1124 dereference. This could lead to a Denial Of Service attack. Thanks to
1125 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1129 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1130 dtls1_buffer_record function under certain conditions. In particular this
1131 could occur if an attacker sent repeated DTLS records with the same
1132 sequence number but for the next epoch. The memory leak could be exploited
1133 by an attacker in a Denial of Service attack through memory exhaustion.
1134 Thanks to Chris Mueller for reporting this issue.
1138 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1139 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1140 method would be set to NULL which could later result in a NULL pointer
1141 dereference. Thanks to Frank Schmirler for reporting this issue.
1145 *) Abort handshake if server key exchange message is omitted for ephemeral
1148 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1149 reporting this issue.
1153 *) Remove non-export ephemeral RSA code on client and server. This code
1154 violated the TLS standard by allowing the use of temporary RSA keys in
1155 non-export ciphersuites and could be used by a server to effectively
1156 downgrade the RSA key length used to a value smaller than the server
1157 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1158 INRIA or reporting this issue.
1162 *) Fixed issue where DH client certificates are accepted without verification.
1163 An OpenSSL server will accept a DH certificate for client authentication
1164 without the certificate verify message. This effectively allows a client to
1165 authenticate without the use of a private key. This only affects servers
1166 which trust a client certificate authority which issues certificates
1167 containing DH keys: these are extremely rare and hardly ever encountered.
1168 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1173 *) Ensure that the session ID context of an SSL is updated when its
1174 SSL_CTX is updated via SSL_set_SSL_CTX.
1176 The session ID context is typically set from the parent SSL_CTX,
1177 and can vary with the CTX.
1180 *) Fix various certificate fingerprint issues.
1182 By using non-DER or invalid encodings outside the signed portion of a
1183 certificate the fingerprint can be changed without breaking the signature.
1184 Although no details of the signed portion of the certificate can be changed
1185 this can cause problems with some applications: e.g. those using the
1186 certificate fingerprint for blacklists.
1188 1. Reject signatures with non zero unused bits.
1190 If the BIT STRING containing the signature has non zero unused bits reject
1191 the signature. All current signature algorithms require zero unused bits.
1193 2. Check certificate algorithm consistency.
1195 Check the AlgorithmIdentifier inside TBS matches the one in the
1196 certificate signature. NB: this will result in signature failure
1197 errors for some broken certificates.
1199 Thanks to Konrad Kraszewski from Google for reporting this issue.
1201 3. Check DSA/ECDSA signatures use DER.
1203 Reencode DSA/ECDSA signatures and compare with the original received
1204 signature. Return an error if there is a mismatch.
1206 This will reject various cases including garbage after signature
1207 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1208 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1209 (negative or with leading zeroes).
1211 Further analysis was conducted and fixes were developed by Stephen Henson
1212 of the OpenSSL core team.
1217 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1218 results on some platforms, including x86_64. This bug occurs at random
1219 with a very low probability, and is not known to be exploitable in any
1220 way, though its exact impact is difficult to determine. Thanks to Pieter
1221 Wuille (Blockstream) who reported this issue and also suggested an initial
1222 fix. Further analysis was conducted by the OpenSSL development team and
1223 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1224 the OpenSSL core team.
1228 *) Do not resume sessions on the server if the negotiated protocol
1229 version does not match the session's version. Resuming with a different
1230 version, while not strictly forbidden by the RFC, is of questionable
1231 sanity and breaks all known clients.
1232 [David Benjamin, Emilia Käsper]
1234 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1235 early CCS messages during renegotiation. (Note that because
1236 renegotiation is encrypted, this early CCS was not exploitable.)
1239 *) Tighten client-side session ticket handling during renegotiation:
1240 ensure that the client only accepts a session ticket if the server sends
1241 the extension anew in the ServerHello. Previously, a TLS client would
1242 reuse the old extension state and thus accept a session ticket if one was
1243 announced in the initial ServerHello.
1245 Similarly, ensure that the client requires a session ticket if one
1246 was advertised in the ServerHello. Previously, a TLS client would
1247 ignore a missing NewSessionTicket message.
1250 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1252 *) SRTP Memory Leak.
1254 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1255 sends a carefully crafted handshake message, to cause OpenSSL to fail
1256 to free up to 64k of memory causing a memory leak. This could be
1257 exploited in a Denial Of Service attack. This issue affects OpenSSL
1258 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1259 whether SRTP is used or configured. Implementations of OpenSSL that
1260 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1262 The fix was developed by the OpenSSL team.
1266 *) Session Ticket Memory Leak.
1268 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1269 integrity of that ticket is first verified. In the event of a session
1270 ticket integrity check failing, OpenSSL will fail to free memory
1271 causing a memory leak. By sending a large number of invalid session
1272 tickets an attacker could exploit this issue in a Denial Of Service
1277 *) Build option no-ssl3 is incomplete.
1279 When OpenSSL is configured with "no-ssl3" as a build option, servers
1280 could accept and complete a SSL 3.0 handshake, and clients could be
1281 configured to send them.
1283 [Akamai and the OpenSSL team]
1285 *) Add support for TLS_FALLBACK_SCSV.
1286 Client applications doing fallback retries should call
1287 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1289 [Adam Langley, Bodo Moeller]
1291 *) Add additional DigestInfo checks.
1293 Reencode DigestInto in DER and check against the original when
1294 verifying RSA signature: this will reject any improperly encoded
1295 DigestInfo structures.
1297 Note: this is a precautionary measure and no attacks are currently known.
1301 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1303 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1304 SRP code can be overrun an internal buffer. Add sanity check that
1305 g, A, B < N to SRP code.
1307 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1308 Group for discovering this issue.
1312 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1313 TLS 1.0 instead of higher protocol versions when the ClientHello message
1314 is badly fragmented. This allows a man-in-the-middle attacker to force a
1315 downgrade to TLS 1.0 even if both the server and the client support a
1316 higher protocol version, by modifying the client's TLS records.
1318 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1319 researching this issue.
1323 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1324 to a denial of service attack. A malicious server can crash the client
1325 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1326 ciphersuite and sending carefully crafted handshake messages.
1328 Thanks to Felix Gröbert (Google) for discovering and researching this
1333 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1334 to leak memory. This can be exploited through a Denial of Service attack.
1335 Thanks to Adam Langley for discovering and researching this issue.
1339 *) An attacker can force openssl to consume large amounts of memory whilst
1340 processing DTLS handshake messages. This can be exploited through a
1341 Denial of Service attack.
1342 Thanks to Adam Langley for discovering and researching this issue.
1346 *) An attacker can force an error condition which causes openssl to crash
1347 whilst processing DTLS packets due to memory being freed twice. This
1348 can be exploited through a Denial of Service attack.
1349 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1354 *) If a multithreaded client connects to a malicious server using a resumed
1355 session and the server sends an ec point format extension it could write
1356 up to 255 bytes to freed memory.
1358 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1363 *) A malicious server can crash an OpenSSL client with a null pointer
1364 dereference (read) by specifying an SRP ciphersuite even though it was not
1365 properly negotiated with the client. This can be exploited through a
1366 Denial of Service attack.
1368 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1369 discovering and researching this issue.
1373 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1374 X509_name_oneline, X509_name_print_ex et al. to leak some information
1375 from the stack. Applications may be affected if they echo pretty printing
1376 output to the attacker.
1378 Thanks to Ivan Fratric (Google) for discovering this issue.
1380 [Emilia Käsper, and Steve Henson]
1382 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1383 for corner cases. (Certain input points at infinity could lead to
1384 bogus results, with non-infinity inputs mapped to infinity too.)
1387 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1389 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1390 handshake can force the use of weak keying material in OpenSSL
1391 SSL/TLS clients and servers.
1393 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1394 researching this issue. (CVE-2014-0224)
1395 [KIKUCHI Masashi, Steve Henson]
1397 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1398 OpenSSL DTLS client the code can be made to recurse eventually crashing
1401 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1403 [Imre Rad, Steve Henson]
1405 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1406 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1407 client or server. This is potentially exploitable to run arbitrary
1408 code on a vulnerable client or server.
1410 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1411 [Jüri Aedla, Steve Henson]
1413 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1414 are subject to a denial of service attack.
1416 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1417 this issue. (CVE-2014-3470)
1418 [Felix Gröbert, Ivan Fratric, Steve Henson]
1420 *) Harmonize version and its documentation. -f flag is used to display
1422 [mancha <mancha1@zoho.com>]
1424 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1425 in i2d_ECPrivateKey.
1426 [mancha <mancha1@zoho.com>]
1428 *) Fix some double frees. These are not thought to be exploitable.
1429 [mancha <mancha1@zoho.com>]
1431 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1433 *) A missing bounds check in the handling of the TLS heartbeat extension
1434 can be used to reveal up to 64k of memory to a connected client or
1437 Thanks for Neel Mehta of Google Security for discovering this bug and to
1438 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1439 preparing the fix (CVE-2014-0160)
1440 [Adam Langley, Bodo Moeller]
1442 *) Fix for the attack described in the paper "Recovering OpenSSL
1443 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1444 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1445 http://eprint.iacr.org/2014/140
1447 Thanks to Yuval Yarom and Naomi Benger for discovering this
1448 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1449 [Yuval Yarom and Naomi Benger]
1451 *) TLS pad extension: draft-agl-tls-padding-03
1453 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1454 TLS client Hello record length value would otherwise be > 255 and
1455 less that 512 pad with a dummy extension containing zeroes so it
1456 is at least 512 bytes long.
1458 [Adam Langley, Steve Henson]
1460 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1462 *) Fix for TLS record tampering bug. A carefully crafted invalid
1463 handshake could crash OpenSSL with a NULL pointer exception.
1464 Thanks to Anton Johansson for reporting this issues.
1467 *) Keep original DTLS digest and encryption contexts in retransmission
1468 structures so we can use the previous session parameters if they need
1469 to be resent. (CVE-2013-6450)
1472 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1473 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1474 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1475 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1476 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1477 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1478 [Rob Stradling, Adam Langley]
1480 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1482 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1483 supporting platforms or when small records were transferred.
1484 [Andy Polyakov, Steve Henson]
1486 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1488 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1490 This addresses the flaw in CBC record processing discovered by
1491 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1492 at: http://www.isg.rhul.ac.uk/tls/
1494 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1495 Security Group at Royal Holloway, University of London
1496 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1497 Emilia Käsper for the initial patch.
1499 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1501 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1502 ciphersuites which can be exploited in a denial of service attack.
1503 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1504 and detecting this bug and to Wolfgang Ettlinger
1505 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1509 *) Return an error when checking OCSP signatures when key is NULL.
1510 This fixes a DoS attack. (CVE-2013-0166)
1513 *) Make openssl verify return errors.
1514 [Chris Palmer <palmer@google.com> and Ben Laurie]
1516 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1517 the right response is stapled. Also change SSL_get_certificate()
1518 so it returns the certificate actually sent.
1519 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1520 [Rob Stradling <rob.stradling@comodo.com>]
1522 *) Fix possible deadlock when decoding public keys.
1525 *) Don't use TLS 1.0 record version number in initial client hello
1529 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1531 *) Sanity check record length before skipping explicit IV in TLS
1532 1.2, 1.1 and DTLS to fix DoS attack.
1534 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1535 fuzzing as a service testing platform.
1539 *) Initialise tkeylen properly when encrypting CMS messages.
1540 Thanks to Solar Designer of Openwall for reporting this issue.
1543 *) In FIPS mode don't try to use composite ciphers as they are not
1547 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1549 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1550 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1551 mean any application compiled against OpenSSL 1.0.0 headers setting
1552 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1553 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1554 0x10000000L Any application which was previously compiled against
1555 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1556 will need to be recompiled as a result. Letting be results in
1557 inability to disable specifically TLS 1.1 and in client context,
1558 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1561 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1562 disable just protocol X, but all protocols above X *if* there are
1563 protocols *below* X still enabled. In more practical terms it means
1564 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1565 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1566 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1570 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1572 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1573 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1574 in CRYPTO_realloc_clean.
1576 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1577 issue and to Adam Langley <agl@chromium.org> for fixing it.
1579 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1581 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1584 *) Workarounds for some broken servers that "hang" if a client hello
1585 record length exceeds 255 bytes.
1587 1. Do not use record version number > TLS 1.0 in initial client
1588 hello: some (but not all) hanging servers will now work.
1589 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1590 the number of ciphers sent in the client hello. This should be
1591 set to an even number, such as 50, for example by passing:
1592 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1593 Most broken servers should now work.
1594 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1595 TLS 1.2 client support entirely.
1598 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1601 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1603 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1604 STRING form instead of a DigestInfo.
1607 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1608 and the RSA_sign/RSA_verify functions. This was made more apparent when
1609 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1610 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1611 the correct format in RSA_verify so both forms transparently work.
1614 *) Some servers which support TLS 1.0 can choke if we initially indicate
1615 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1616 encrypted premaster secret. As a workaround use the maximum permitted
1617 client version in client hello, this should keep such servers happy
1618 and still work with previous versions of OpenSSL.
1621 *) Add support for TLS/DTLS heartbeats.
1622 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1624 *) Add support for SCTP.
1625 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1627 *) Improved PRNG seeding for VOS.
1628 [Paul Green <Paul.Green@stratus.com>]
1630 *) Extensive assembler packs updates, most notably:
1632 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1633 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1634 - x86_64: bit-sliced AES implementation;
1635 - ARM: NEON support, contemporary platforms optimizations;
1636 - s390x: z196 support;
1637 - *: GHASH and GF(2^m) multiplication implementations;
1641 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1642 (removal of unnecessary code)
1643 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1645 *) Add TLS key material exporter from RFC 5705.
1648 *) Add DTLS-SRTP negotiation from RFC 5764.
1651 *) Add Next Protocol Negotiation,
1652 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1653 disabled with a no-npn flag to config or Configure. Code donated
1655 [Adam Langley <agl@google.com> and Ben Laurie]
1657 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1658 NIST-P256, NIST-P521, with constant-time single point multiplication on
1659 typical inputs. Compiler support for the nonstandard type __uint128_t is
1660 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1661 Code made available under Apache License version 2.0.
1663 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1664 line to include this in your build of OpenSSL, and run "make depend" (or
1665 "make update"). This enables the following EC_METHODs:
1667 EC_GFp_nistp224_method()
1668 EC_GFp_nistp256_method()
1669 EC_GFp_nistp521_method()
1671 EC_GROUP_new_by_curve_name() will automatically use these (while
1672 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1674 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1676 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1677 all platforms. Move ssize_t definition from e_os.h to the public
1678 header file e_os2.h as it now appears in public header file cms.h
1681 *) New -sigopt option to the ca, req and x509 utilities. Additional
1682 signature parameters can be passed using this option and in
1686 *) Add RSA PSS signing function. This will generate and set the
1687 appropriate AlgorithmIdentifiers for PSS based on those in the
1688 corresponding EVP_MD_CTX structure. No application support yet.
1691 *) Support for companion algorithm specific ASN1 signing routines.
1692 New function ASN1_item_sign_ctx() signs a pre-initialised
1693 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1694 the appropriate parameters.
1697 *) Add new algorithm specific ASN1 verification initialisation function
1698 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1699 handling will be the same no matter what EVP_PKEY_METHOD is used.
1700 Add a PSS handler to support verification of PSS signatures: checked
1701 against a number of sample certificates.
1704 *) Add signature printing for PSS. Add PSS OIDs.
1705 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1707 *) Add algorithm specific signature printing. An individual ASN1 method
1708 can now print out signatures instead of the standard hex dump.
1710 More complex signatures (e.g. PSS) can print out more meaningful
1711 information. Include DSA version that prints out the signature
1715 *) Password based recipient info support for CMS library: implementing
1719 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1720 neatly separates the code into cipher and PBE sections and is required
1721 for some algorithms that split PBES2 into separate pieces (such as
1722 password based CMS).
1725 *) Session-handling fixes:
1726 - Fix handling of connections that are resuming with a session ID,
1727 but also support Session Tickets.
1728 - Fix a bug that suppressed issuing of a new ticket if the client
1729 presented a ticket with an expired session.
1730 - Try to set the ticket lifetime hint to something reasonable.
1731 - Make tickets shorter by excluding irrelevant information.
1732 - On the client side, don't ignore renewed tickets.
1733 [Adam Langley, Bodo Moeller (Google)]
1735 *) Fix PSK session representation.
1738 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1740 This work was sponsored by Intel.
1743 *) Add GCM support to TLS library. Some custom code is needed to split
1744 the IV between the fixed (from PRF) and explicit (from TLS record)
1745 portions. This adds all GCM ciphersuites supported by RFC5288 and
1746 RFC5289. Generalise some AES* cipherstrings to include GCM and
1747 add a special AESGCM string for GCM only.
1750 *) Expand range of ctrls for AES GCM. Permit setting invocation
1751 field on decrypt and retrieval of invocation field only on encrypt.
1754 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1755 As required by RFC5289 these ciphersuites cannot be used if for
1756 versions of TLS earlier than 1.2.
1759 *) For FIPS capable OpenSSL interpret a NULL default public key method
1760 as unset and return the appropriate default but do *not* set the default.
1761 This means we can return the appropriate method in applications that
1762 switch between FIPS and non-FIPS modes.
1765 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1766 ENGINE is used then we cannot handle that in the FIPS module so we
1767 keep original code iff non-FIPS operations are allowed.
1770 *) Add -attime option to openssl utilities.
1771 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1773 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1776 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1777 FIPS EC methods unconditionally for now.
1780 *) New build option no-ec2m to disable characteristic 2 code.
1783 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1784 all cases can be covered as some introduce binary incompatibilities.
1787 *) Redirect RSA operations to FIPS module including keygen,
1788 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1791 *) Add similar low level API blocking to ciphers.
1794 *) Low level digest APIs are not approved in FIPS mode: any attempt
1795 to use these will cause a fatal error. Applications that *really* want
1796 to use them can use the private_* version instead.
1799 *) Redirect cipher operations to FIPS module for FIPS builds.
1802 *) Redirect digest operations to FIPS module for FIPS builds.
1805 *) Update build system to add "fips" flag which will link in fipscanister.o
1806 for static and shared library builds embedding a signature if needed.
1809 *) Output TLS supported curves in preference order instead of numerical
1810 order. This is currently hardcoded for the highest order curves first.
1811 This should be configurable so applications can judge speed vs strength.
1814 *) Add TLS v1.2 server support for client authentication.
1817 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1821 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1822 FIPS modules versions.
1825 *) Add TLS v1.2 client side support for client authentication. Keep cache
1826 of handshake records longer as we don't know the hash algorithm to use
1827 until after the certificate request message is received.
1830 *) Initial TLS v1.2 client support. Add a default signature algorithms
1831 extension including all the algorithms we support. Parse new signature
1832 format in client key exchange. Relax some ECC signing restrictions for
1833 TLS v1.2 as indicated in RFC5246.
1836 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1837 to new signature format when needed using client digest preference.
1838 All server ciphersuites should now work correctly in TLS v1.2. No client
1839 support yet and no support for client certificates.
1842 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1843 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1844 ciphersuites. At present only RSA key exchange ciphersuites work with
1845 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1846 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1847 and version checking.
1850 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1851 with this defined it will not be affected by any changes to ssl internal
1852 structures. Add several utility functions to allow openssl application
1853 to work with OPENSSL_NO_SSL_INTERN defined.
1857 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1859 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1862 *) Permit abbreviated handshakes when renegotiating using the function
1863 SSL_renegotiate_abbreviated().
1864 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1866 *) Add call to ENGINE_register_all_complete() to
1867 ENGINE_load_builtin_engines(), so some implementations get used
1868 automatically instead of needing explicit application support.
1871 *) Add support for TLS key exporter as described in RFC5705.
1872 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1874 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1875 a few changes are required:
1877 Add SSL_OP_NO_TLSv1_1 flag.
1878 Add TLSv1_1 methods.
1879 Update version checking logic to handle version 1.1.
1880 Add explicit IV handling (ported from DTLS code).
1881 Add command line options to s_client/s_server.
1884 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1886 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1887 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1888 content decryption and always return the same error. Note: this attack
1889 needs on average 2^20 messages so it only affects automated senders. The
1890 old behaviour can be reenabled in the CMS code by setting the
1891 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1892 an MMA defence is not necessary.
1893 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1894 this issue. (CVE-2012-0884)
1897 *) Fix CVE-2011-4619: make sure we really are receiving a
1898 client hello before rejecting multiple SGC restarts. Thanks to
1899 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1902 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1904 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1905 Thanks to Antonio Martin, Enterprise Secure Access Research and
1906 Development, Cisco Systems, Inc. for discovering this bug and
1907 preparing a fix. (CVE-2012-0050)
1910 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1912 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1913 of the Vaudenay padding oracle attack on CBC mode encryption
1914 which enables an efficient plaintext recovery attack against
1915 the OpenSSL implementation of DTLS. Their attack exploits timing
1916 differences arising during decryption processing. A research
1917 paper describing this attack can be found at:
1918 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1919 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1920 Security Group at Royal Holloway, University of London
1921 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1922 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1923 for preparing the fix. (CVE-2011-4108)
1924 [Robin Seggelmann, Michael Tuexen]
1926 *) Clear bytes used for block padding of SSL 3.0 records.
1928 [Adam Langley (Google)]
1930 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1931 Kadianakis <desnacked@gmail.com> for discovering this issue and
1932 Adam Langley for preparing the fix. (CVE-2011-4619)
1933 [Adam Langley (Google)]
1935 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1936 [Andrey Kulikov <amdeich@gmail.com>]
1938 *) Prevent malformed RFC3779 data triggering an assertion failure.
1939 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1940 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1941 [Rob Austein <sra@hactrn.net>]
1943 *) Improved PRNG seeding for VOS.
1944 [Paul Green <Paul.Green@stratus.com>]
1946 *) Fix ssl_ciph.c set-up race.
1947 [Adam Langley (Google)]
1949 *) Fix spurious failures in ecdsatest.c.
1950 [Emilia Käsper (Google)]
1952 *) Fix the BIO_f_buffer() implementation (which was mixing different
1953 interpretations of the '..._len' fields).
1954 [Adam Langley (Google)]
1956 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1957 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1958 threads won't reuse the same blinding coefficients.
1960 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1961 lock to call BN_BLINDING_invert_ex, and avoids one use of
1962 BN_BLINDING_update for each BN_BLINDING structure (previously,
1963 the last update always remained unused).
1964 [Emilia Käsper (Google)]
1966 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1967 [Bob Buckholz (Google)]
1969 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1971 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1972 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1973 [Kaspar Brand <ossl@velox.ch>]
1975 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1976 for multi-threaded use of ECDH. (CVE-2011-3210)
1977 [Adam Langley (Google)]
1979 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1982 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1983 signature public key algorithm by using OID xref utilities instead.
1984 Before this you could only use some ECC ciphersuites with SHA1 only.
1987 *) Add protection against ECDSA timing attacks as mentioned in the paper
1988 by Billy Bob Brumley and Nicola Tuveri, see:
1990 http://eprint.iacr.org/2011/232.pdf
1992 [Billy Bob Brumley and Nicola Tuveri]
1994 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1996 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1997 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1999 *) Fix bug in string printing code: if *any* escaping is enabled we must
2000 escape the escape character (backslash) or the resulting string is
2004 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2006 *) Disable code workaround for ancient and obsolete Netscape browsers
2007 and servers: an attacker can use it in a ciphersuite downgrade attack.
2008 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2011 *) Fixed J-PAKE implementation error, originally discovered by
2012 Sebastien Martini, further info and confirmation from Stefan
2013 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2016 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2018 *) Fix extension code to avoid race conditions which can result in a buffer
2019 overrun vulnerability: resumed sessions must not be modified as they can
2020 be shared by multiple threads. CVE-2010-3864
2023 *) Fix WIN32 build system to correctly link an ENGINE directory into
2027 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2029 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2031 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2033 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2035 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2036 context. The operation can be customised via the ctrl mechanism in
2037 case ENGINEs want to include additional functionality.
2040 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2043 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2044 output hashes compatible with older versions of OpenSSL.
2045 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2047 *) Fix compression algorithm handling: if resuming a session use the
2048 compression algorithm of the resumed session instead of determining
2049 it from client hello again. Don't allow server to change algorithm.
2052 *) Add load_crls() function to apps tidying load_certs() too. Add option
2053 to verify utility to allow additional CRLs to be included.
2056 *) Update OCSP request code to permit adding custom headers to the request:
2057 some responders need this.
2060 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2062 [Julia Lawall <julia@diku.dk>]
2064 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2065 needlessly dereferenced structures, used obsolete functions and
2066 didn't handle all updated verify codes correctly.
2069 *) Disable MD2 in the default configuration.
2072 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2073 indicate the initial BIO being pushed or popped. This makes it possible
2074 to determine whether the BIO is the one explicitly called or as a result
2075 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2076 it handles reference counts correctly and doesn't zero out the I/O bio
2077 when it is not being explicitly popped. WARNING: applications which
2078 included workarounds for the old buggy behaviour will need to be modified
2079 or they could free up already freed BIOs.
2082 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2083 renaming to all platforms (within the 0.9.8 branch, this was
2084 done conditionally on Netware platforms to avoid a name clash).
2085 [Guenter <lists@gknw.net>]
2087 *) Add ECDHE and PSK support to DTLS.
2088 [Michael Tuexen <tuexen@fh-muenster.de>]
2090 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2094 *) Add "missing" function EVP_MD_flags() (without this the only way to
2095 retrieve a digest flags is by accessing the structure directly. Update
2096 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2097 or cipher is registered as in the "from" argument. Print out all
2098 registered digests in the dgst usage message instead of manually
2099 attempting to work them out.
2102 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2103 this allows the use of compression and extensions. Change default cipher
2104 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2105 by default unless an application cipher string requests it.
2108 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2109 key ids to find matching certificates and keys but some PKCS#12 files
2110 don't follow the (somewhat unwritten) rules and this strategy fails.
2111 Now just gather all certificates together and the first private key
2112 then look for the first certificate that matches the key.
2115 *) Support use of registered digest and cipher names for dgst and cipher
2116 commands instead of having to add each one as a special case. So now
2123 openssl dgst -sha256 foo
2125 and this works for ENGINE based algorithms too.
2129 *) Update Gost ENGINE to support parameter files.
2130 [Victor B. Wagner <vitus@cryptocom.ru>]
2132 *) Support GeneralizedTime in ca utility.
2133 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2135 *) Enhance the hash format used for certificate directory links. The new
2136 form uses the canonical encoding (meaning equivalent names will work
2137 even if they aren't identical) and uses SHA1 instead of MD5. This form
2138 is incompatible with the older format and as a result c_rehash should
2139 be used to rebuild symbolic links.
2142 *) Make PKCS#8 the default write format for private keys, replacing the
2143 traditional format. This form is standardised, more secure and doesn't
2144 include an implicit MD5 dependency.
2147 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2148 committed to OpenSSL should pass this lot as a minimum.
2151 *) Add session ticket override functionality for use by EAP-FAST.
2152 [Jouni Malinen <j@w1.fi>]
2154 *) Modify HMAC functions to return a value. Since these can be implemented
2155 in an ENGINE errors can occur.
2158 *) Type-checked OBJ_bsearch_ex.
2161 *) Type-checked OBJ_bsearch. Also some constification necessitated
2162 by type-checking. Still to come: TXT_DB, bsearch(?),
2163 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2167 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2168 seconds to a tm structure directly, instead of going through OS
2169 specific date routines. This avoids any issues with OS routines such
2170 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2171 and X509_time_adj_ex() to cover the extended range. The existing
2172 X509_time_adj() is still usable and will no longer have any date issues.
2175 *) Delta CRL support. New use deltas option which will attempt to locate
2176 and search any appropriate delta CRLs available.
2178 This work was sponsored by Google.
2181 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2182 code and add additional score elements. Validate alternate CRL paths
2183 as part of the CRL checking and indicate a new error "CRL path validation
2184 error" in this case. Applications wanting additional details can use
2185 the verify callback and check the new "parent" field. If this is not
2186 NULL CRL path validation is taking place. Existing applications wont
2187 see this because it requires extended CRL support which is off by
2190 This work was sponsored by Google.
2193 *) Support for freshest CRL extension.
2195 This work was sponsored by Google.
2198 *) Initial indirect CRL support. Currently only supported in the CRLs
2199 passed directly and not via lookup. Process certificate issuer
2200 CRL entry extension and lookup CRL entries by bother issuer name
2201 and serial number. Check and process CRL issuer entry in IDP extension.
2203 This work was sponsored by Google.
2206 *) Add support for distinct certificate and CRL paths. The CRL issuer
2207 certificate is validated separately in this case. Only enabled if
2208 an extended CRL support flag is set: this flag will enable additional
2209 CRL functionality in future.
2211 This work was sponsored by Google.
2214 *) Add support for policy mappings extension.
2216 This work was sponsored by Google.
2219 *) Fixes to pathlength constraint, self issued certificate handling,
2220 policy processing to align with RFC3280 and PKITS tests.
2222 This work was sponsored by Google.
2225 *) Support for name constraints certificate extension. DN, email, DNS
2226 and URI types are currently supported.
2228 This work was sponsored by Google.
2231 *) To cater for systems that provide a pointer-based thread ID rather
2232 than numeric, deprecate the current numeric thread ID mechanism and
2233 replace it with a structure and associated callback type. This
2234 mechanism allows a numeric "hash" to be extracted from a thread ID in
2235 either case, and on platforms where pointers are larger than 'long',
2236 mixing is done to help ensure the numeric 'hash' is usable even if it
2237 can't be guaranteed unique. The default mechanism is to use "&errno"
2238 as a pointer-based thread ID to distinguish between threads.
2240 Applications that want to provide their own thread IDs should now use
2241 CRYPTO_THREADID_set_callback() to register a callback that will call
2242 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2244 Note that ERR_remove_state() is now deprecated, because it is tied
2245 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2246 to free the current thread's error state should be replaced by
2247 ERR_remove_thread_state(NULL).
2249 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2250 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2251 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2252 application was previously providing a numeric thread callback that
2253 was inappropriate for distinguishing threads, then uniqueness might
2254 have been obtained with &errno that happened immediately in the
2255 intermediate development versions of OpenSSL; this is no longer the
2256 case, the numeric thread callback will now override the automatic use
2258 [Geoff Thorpe, with help from Bodo Moeller]
2260 *) Initial support for different CRL issuing certificates. This covers a
2261 simple case where the self issued certificates in the chain exist and
2262 the real CRL issuer is higher in the existing chain.
2264 This work was sponsored by Google.
2267 *) Removed effectively defunct crypto/store from the build.
2270 *) Revamp of STACK to provide stronger type-checking. Still to come:
2271 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2272 ASN1_STRING, CONF_VALUE.
2275 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2276 RAM on SSL connections. This option can save about 34k per idle SSL.
2279 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2280 STACK, TXT_DB, bsearch, qsort.
2283 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2284 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2285 support for data, signedData, compressedData, digestedData and
2286 encryptedData, envelopedData types included. Scripts to check against
2287 RFC4134 examples draft and interop and consistency checks of many
2288 content types and variants.
2291 *) Add options to enc utility to support use of zlib compression BIO.
2294 *) Extend mk1mf to support importing of options and assembly language
2295 files from Configure script, currently only included in VC-WIN32.
2296 The assembly language rules can now optionally generate the source
2297 files from the associated perl scripts.
2300 *) Implement remaining functionality needed to support GOST ciphersuites.
2301 Interop testing has been performed using CryptoPro implementations.
2302 [Victor B. Wagner <vitus@cryptocom.ru>]
2304 *) s390x assembler pack.
2307 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2311 *) Implement Opaque PRF Input TLS extension as specified in
2312 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2313 official specification yet and no extension type assignment by
2314 IANA exists, this extension (for now) will have to be explicitly
2315 enabled when building OpenSSL by providing the extension number
2316 to use. For example, specify an option
2318 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2320 to the "config" or "Configure" script to enable the extension,
2321 assuming extension number 0x9527 (which is a completely arbitrary
2322 and unofficial assignment based on the MD5 hash of the Internet
2323 Draft). Note that by doing so, you potentially lose
2324 interoperability with other TLS implementations since these might
2325 be using the same extension number for other purposes.
2327 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2328 opaque PRF input value to use in the handshake. This will create
2329 an interal copy of the length-'len' string at 'src', and will
2330 return non-zero for success.
2332 To get more control and flexibility, provide a callback function
2335 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2336 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2340 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2343 Callback function 'cb' will be called in handshakes, and is
2344 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2345 Argument 'arg' is for application purposes (the value as given to
2346 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2347 be provided to the callback function). The callback function
2348 has to return non-zero to report success: usually 1 to use opaque
2349 PRF input just if possible, or 2 to enforce use of the opaque PRF
2350 input. In the latter case, the library will abort the handshake
2351 if opaque PRF input is not successfully negotiated.
2353 Arguments 'peerinput' and 'len' given to the callback function
2354 will always be NULL and 0 in the case of a client. A server will
2355 see the client's opaque PRF input through these variables if
2356 available (NULL and 0 otherwise). Note that if the server
2357 provides an opaque PRF input, the length must be the same as the
2358 length of the client's opaque PRF input.
2360 Note that the callback function will only be called when creating
2361 a new session (session resumption can resume whatever was
2362 previously negotiated), and will not be called in SSL 2.0
2363 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2364 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2365 for applications that need to enforce opaque PRF input.
2369 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2372 [Victor B. Wagner <vitus@cryptocom.ru>]
2374 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2375 RFC4507bis. The encrypted ticket format is an encrypted encoded
2376 SSL_SESSION structure, that way new session features are automatically
2379 If a client application caches session in an SSL_SESSION structure
2380 support is transparent because tickets are now stored in the encoded
2383 The SSL_CTX structure automatically generates keys for ticket
2384 protection in servers so again support should be possible
2385 with no application modification.
2387 If a client or server wishes to disable RFC4507 support then the option
2388 SSL_OP_NO_TICKET can be set.
2390 Add a TLS extension debugging callback to allow the contents of any client
2391 or server extensions to be examined.
2393 This work was sponsored by Google.
2396 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2397 OpenSSL should now compile cleanly on gcc 4.2
2398 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2400 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2401 support including streaming MAC support: this is required for GOST
2402 ciphersuite support.
2403 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2405 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2406 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2407 to output in BER and PEM format.
2410 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2411 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2412 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2413 ENGINE support for HMAC keys which are unextractable. New -mac and
2414 -macopt options to dgst utility.
2417 *) New option -sigopt to dgst utility. Update dgst to use
2418 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2419 alternative signing paramaters such as X9.31 or PSS in the dgst
2423 *) Change ssl_cipher_apply_rule(), the internal function that does
2424 the work each time a ciphersuite string requests enabling
2425 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2426 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2427 the order of disabled ciphersuites such that those ciphersuites
2428 that most recently went from enabled to disabled not only stay
2429 in order with respect to each other, but also have higher priority
2430 than other disabled ciphersuites the next time ciphersuites are
2433 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2434 the same ciphersuites as with "HIGH" alone, but in a specific
2435 order where the PSK ciphersuites come first (since they are the
2436 most recently disabled ciphersuites when "HIGH" is parsed).
2438 Also, change ssl_create_cipher_list() (using this new
2439 funcionality) such that between otherwise identical
2440 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2444 *) Change ssl_create_cipher_list() so that it automatically
2445 arranges the ciphersuites in reasonable order before starting
2446 to process the rule string. Thus, the definition for "DEFAULT"
2447 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2448 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2449 This makes it much easier to arrive at a reasonable default order
2450 in applications for which anonymous ciphers are OK (meaning
2451 that you can't actually use DEFAULT).
2452 [Bodo Moeller; suggested by Victor Duchovni]
2454 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2455 processing) into multiple integers instead of setting
2456 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2457 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2458 (These masks as well as the individual bit definitions are hidden
2459 away into the non-exported interface ssl/ssl_locl.h, so this
2460 change to the definition of the SSL_CIPHER structure shouldn't
2461 affect applications.) This give us more bits for each of these
2462 categories, so there is no longer a need to coagulate AES128 and
2463 AES256 into a single algorithm bit, and to coagulate Camellia128
2464 and Camellia256 into a single algorithm bit, which has led to all
2467 Thus, among other things, the kludge introduced in 0.9.7m and
2468 0.9.8e for masking out AES256 independently of AES128 or masking
2469 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2471 With the change, we also introduce new ciphersuite aliases that
2472 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2476 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2477 Use the leftmost N bytes of the signature input if the input is
2478 larger than the prime q (with N being the size in bytes of q).
2481 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2482 it yet and it is largely untested.
2485 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2488 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2489 some compilers (gcc 4.2 and later) reject their use. Safestack is
2490 reimplemented. Update ASN1 to avoid use of legacy functions.
2493 *) Win32/64 targets are linked with Winsock2.
2496 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2497 to external functions. This can be used to increase CRL handling
2498 efficiency especially when CRLs are very large by (for example) storing
2499 the CRL revoked certificates in a database.
2502 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2503 new CRLs added to a directory can be used. New command line option
2504 -verify_return_error to s_client and s_server. This causes real errors
2505 to be returned by the verify callback instead of carrying on no matter
2506 what. This reflects the way a "real world" verify callback would behave.
2509 *) GOST engine, supporting several GOST algorithms and public key formats.
2510 Kindly donated by Cryptocom.
2513 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2514 partitioned by DP are handled but no indirect CRL or reason partitioning
2515 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2516 selected via a scoring technique which handles IDP and AKID in CRLs.
2519 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2520 will ultimately be used for all verify operations: this will remove the
2521 X509_STORE dependency on certificate verification and allow alternative
2522 lookup methods. X509_STORE based implementations of these two callbacks.
2525 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2526 Modify get_crl() to find a valid (unexpired) CRL if possible.
2529 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2530 this would be called X509_CRL_cmp() but that name is already used by
2531 a function that just compares CRL issuer names. Cache several CRL
2532 extensions in X509_CRL structure and cache CRLDP in X509.
2535 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2536 this maps equivalent X509_NAME structures into a consistent structure.
2537 Name comparison can then be performed rapidly using memcmp().
2540 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2544 *) Allow digests to supply their own micalg string for S/MIME type using
2545 the ctrl EVP_MD_CTRL_MICALG.
2548 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2549 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2550 ctrl. It can then customise the structure before and/or after signing
2554 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2555 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2556 to free up any added signature OIDs.
2559 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2560 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2561 digest and cipher tables. New options added to openssl utility:
2562 list-message-digest-algorithms and list-cipher-algorithms.
2565 *) Change the array representation of binary polynomials: the list
2566 of degrees of non-zero coefficients is now terminated with -1.
2567 Previously it was terminated with 0, which was also part of the
2568 value; thus, the array representation was not applicable to
2569 polynomials where t^0 has coefficient zero. This change makes
2570 the array representation useful in a more general context.
2573 *) Various modifications and fixes to SSL/TLS cipher string
2574 handling. For ECC, the code now distinguishes between fixed ECDH
2575 with RSA certificates on the one hand and with ECDSA certificates
2576 on the other hand, since these are separate ciphersuites. The
2577 unused code for Fortezza ciphersuites has been removed.
2579 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2580 (not "ECDHE"). For consistency with the code for DH
2581 certificates, use of ECDH certificates is now considered ECDH
2582 authentication, not RSA or ECDSA authentication (the latter is
2583 merely the CA's signing algorithm and not actively used in the
2586 The temporary ciphersuite alias "ECCdraft" is no longer
2587 available, and ECC ciphersuites are no longer excluded from "ALL"
2588 and "DEFAULT". The following aliases now exist for RFC 4492
2589 ciphersuites, most of these by analogy with the DH case:
2591 kECDHr - ECDH cert, signed with RSA
2592 kECDHe - ECDH cert, signed with ECDSA
2593 kECDH - ECDH cert (signed with either RSA or ECDSA)
2594 kEECDH - ephemeral ECDH
2595 ECDH - ECDH cert or ephemeral ECDH
2601 AECDH - anonymous ECDH
2602 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2606 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2607 Use correct micalg parameters depending on digest(s) in signed message.
2610 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2611 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2614 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2615 an engine to register a method. Add ENGINE lookups for methods and
2616 functional reference processing.
2619 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2620 EVP_{Sign,Verify}* which allow an application to customise the signature
2624 *) New -resign option to smime utility. This adds one or more signers
2625 to an existing PKCS#7 signedData structure. Also -md option to use an
2626 alternative message digest algorithm for signing.
2629 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2630 create PKCS7 structures containing multiple signers. Update smime
2631 application to support multiple signers.
2634 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2638 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2639 Reorganize PBE internals to lookup from a static table using NIDs,
2640 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2641 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2642 PRF which will be automatically used with PBES2.
2645 *) Replace the algorithm specific calls to generate keys in "req" with the
2649 *) Update PKCS#7 enveloped data routines to use new API. This is now
2650 supported by any public key method supporting the encrypt operation. A
2651 ctrl is added to allow the public key algorithm to examine or modify
2652 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2656 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2657 a default digest type to use. In most cases this will be SHA1 but some
2658 algorithms (such as GOST) need to specify an alternative digest. The
2659 return value indicates how strong the preference is 1 means optional and
2660 2 is mandatory (that is it is the only supported type). Modify
2661 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2662 use the default md. Update openssl utilities to use the default digest
2663 type for signing if it is not explicitly indicated.
2666 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2667 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2668 signing method from the key type. This effectively removes the link
2669 between digests and public key types.
2672 *) Add an OID cross reference table and utility functions. Its purpose is to
2673 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2674 rsaEncryption. This will allow some of the algorithm specific hackery
2675 needed to use the correct OID to be removed.
2678 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2679 structures for PKCS7_sign(). They are now set up by the relevant public
2683 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2686 *) Add support for key derivation (agreement) in the API, DH method and
2690 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2691 public and private key formats. As a side effect these add additional
2692 command line functionality not previously available: DSA signatures can be
2693 generated and verified using pkeyutl and DH key support and generation in
2698 [Oliver Tappe <zooey@hirschkaefer.de>]
2700 *) New make target "install_html_docs" installs HTML renditions of the
2702 [Oliver Tappe <zooey@hirschkaefer.de>]
2704 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2705 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2706 support key and parameter generation and add initial key generation
2707 functionality for RSA.
2710 *) Add functions for main EVP_PKEY_method operations. The undocumented
2711 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2712 EVP_PKEY_{encrypt,decrypt}_old.
2715 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2716 key API, doesn't do much yet.
2719 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2720 public key algorithms. New option to openssl utility:
2721 "list-public-key-algorithms" to print out info.
2724 *) Implement the Supported Elliptic Curves Extension for
2725 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2728 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2729 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2732 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2733 utilities such as rsa, dsa, dsaparam etc except they process any key
2737 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2738 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2739 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2743 *) Initial support for pluggable public key ASN1.
2744 De-spaghettify the public key ASN1 handling. Move public and private
2745 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2746 algorithm specific handling to a single module within the relevant
2747 algorithm directory. Add functions to allow (near) opaque processing
2748 of public and private key structures.
2751 *) Implement the Supported Point Formats Extension for
2752 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2755 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2756 for the psk identity [hint] and the psk callback functions to the
2757 SSL_SESSION, SSL and SSL_CTX structure.
2760 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2764 SSL_CTX_use_psk_identity_hint
2765 SSL_get_psk_identity_hint
2766 SSL_get_psk_identity
2767 SSL_use_psk_identity_hint
2769 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2771 *) Add RFC 3161 compliant time stamp request creation, response generation
2772 and response verification functionality.
2773 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2775 *) Add initial support for TLS extensions, specifically for the server_name
2776 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2777 have new members for a host name. The SSL data structure has an
2778 additional member SSL_CTX *initial_ctx so that new sessions can be
2779 stored in that context to allow for session resumption, even after the
2780 SSL has been switched to a new SSL_CTX in reaction to a client's
2781 server_name extension.
2783 New functions (subject to change):
2785 SSL_get_servername()
2786 SSL_get_servername_type()
2789 New CTRL codes and macros (subject to change):
2791 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2792 - SSL_CTX_set_tlsext_servername_callback()
2793 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2794 - SSL_CTX_set_tlsext_servername_arg()
2795 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2797 openssl s_client has a new '-servername ...' option.
2799 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2800 '-key2 ...', '-servername_fatal' (subject to change). This allows
2801 testing the HostName extension for a specific single host name ('-cert'
2802 and '-key' remain fallbacks for handshakes without HostName
2803 negotiation). If the unrecognized_name alert has to be sent, this by
2804 default is a warning; it becomes fatal with the '-servername_fatal'
2807 [Peter Sylvester, Remy Allais, Christophe Renou]
2809 *) Whirlpool hash implementation is added.
2812 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2813 bn(64,32). Because of instruction set limitations it doesn't have
2814 any negative impact on performance. This was done mostly in order
2815 to make it possible to share assembler modules, such as bn_mul_mont
2816 implementations, between 32- and 64-bit builds without hassle.
2819 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2820 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2824 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2825 dedicated Montgomery multiplication procedure, is introduced.
2826 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2827 "64-bit" performance on certain 32-bit targets.
2830 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2831 in SSL structures. New SSL ctrl to set maximum send fragment size.
2832 Save memory by seeting the I/O buffer sizes dynamically instead of
2833 using the maximum available value.
2836 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2837 in addition to the text details.
2840 *) Very, very preliminary EXPERIMENTAL support for printing of general
2841 ASN1 structures. This currently produces rather ugly output and doesn't
2842 handle several customised structures at all.
2845 *) Integrated support for PVK file format and some related formats such
2846 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2847 these in the 'rsa' and 'dsa' utilities.
2850 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2853 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2854 place for the (very old) "NETSCAPE" format certificates which are now
2855 handled using new ASN1 code equivalents.
2858 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2859 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2860 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2863 *) Modify CRL distribution points extension code to print out previously
2864 unsupported fields. Enhance extension setting code to allow setting of
2868 *) Add print and set support for Issuing Distribution Point CRL extension.
2871 *) Change 'Configure' script to enable Camellia by default.
2874 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2876 *) When rejecting SSL/TLS records due to an incorrect version number, never
2877 update s->server with a new major version number. As of
2878 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2879 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2880 the previous behavior could result in a read attempt at NULL when
2881 receiving specific incorrect SSL/TLS records once record payload
2882 protection is active. (CVE-2010-0740)
2883 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2885 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2886 could be crashed if the relevant tables were not present (e.g. chrooted).
2887 [Tomas Hoger <thoger@redhat.com>]
2889 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2891 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2892 [Martin Olsson, Neel Mehta]
2894 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2895 accommodate for stack sorting, always a write lock!).
2898 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2899 excessive delays in the RAND_poll(): over a minute. As a workaround
2900 include a time check in the inner Heap32Next loop too.
2903 *) The code that handled flushing of data in SSL/TLS originally used the
2904 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2905 the problem outlined in PR#1949. The fix suggested there however can
2906 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2907 of Apache). So instead simplify the code to flush unconditionally.
2908 This should be fine since flushing with no data to flush is a no op.
2911 *) Handle TLS versions 2.0 and later properly and correctly use the
2912 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2913 off ancient servers have a habit of sticking around for a while...
2916 *) Modify compression code so it frees up structures without using the
2917 ex_data callbacks. This works around a problem where some applications
2918 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2919 restarting) then use compression (e.g. SSL with compression) later.
2920 This results in significant per-connection memory leaks and
2921 has caused some security issues including CVE-2008-1678 and
2925 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2926 change when encrypting or decrypting.
2929 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2930 connect and renegotiate with servers which do not support RI.
2931 Until RI is more widely deployed this option is enabled by default.
2934 *) Add "missing" ssl ctrls to clear options and mode.
2937 *) If client attempts to renegotiate and doesn't support RI respond with
2938 a no_renegotiation alert as required by RFC5746. Some renegotiating
2939 TLS clients will continue a connection gracefully when they receive
2940 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2941 waiting for a server hello which it will never receive. Now we treat a
2942 received no_renegotiation alert as a fatal error. This is because
2943 applications requesting a renegotiation might well expect it to succeed
2944 and would have no code in place to handle the server denying it so the
2945 only safe thing to do is to terminate the connection.
2948 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2949 peer supports secure renegotiation and 0 otherwise. Print out peer
2950 renegotiation support in s_client/s_server.
2953 *) Replace the highly broken and deprecated SPKAC certification method with
2954 the updated NID creation version. This should correctly handle UTF8.
2957 *) Implement RFC5746. Re-enable renegotiation but require the extension
2958 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2959 turns out to be a bad idea. It has been replaced by
2960 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2961 SSL_CTX_set_options(). This is really not recommended unless you
2962 know what you are doing.
2963 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2965 *) Fixes to stateless session resumption handling. Use initial_ctx when
2966 issuing and attempting to decrypt tickets in case it has changed during
2967 servername handling. Use a non-zero length session ID when attempting
2968 stateless session resumption: this makes it possible to determine if
2969 a resumption has occurred immediately after receiving server hello
2970 (several places in OpenSSL subtly assume this) instead of later in
2974 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2975 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2976 fixes for a few places where the return code is not checked
2978 [Julia Lawall <julia@diku.dk>]
2980 *) Add --strict-warnings option to Configure script to include devteam
2981 warnings in other configurations.
2984 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2985 makes it possible to install openssl libraries in locations which
2986 have names other than "lib", for example "/usr/lib64" which some
2988 [Steve Henson, based on patch from Jeremy Utley]
2990 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2991 X690 8.9.12 and can produce some misleading textual output of OIDs.
2992 [Steve Henson, reported by Dan Kaminsky]
2994 *) Delete MD2 from algorithm tables. This follows the recommendation in
2995 several standards that it is not used in new applications due to
2996 several cryptographic weaknesses. For binary compatibility reasons
2997 the MD2 API is still compiled in by default.
3000 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3004 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3005 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3007 [Guenter <lists@gknw.net>]
3009 *) Fix the server certificate chain building code to use X509_verify_cert(),
3010 it used to have an ad-hoc builder which was unable to cope with anything
3011 other than a simple chain.
3012 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3014 *) Don't check self signed certificate signatures in X509_verify_cert()
3015 by default (a flag can override this): it just wastes time without
3016 adding any security. As a useful side effect self signed root CAs
3017 with non-FIPS digests are now usable in FIPS mode.
3020 *) In dtls1_process_out_of_seq_message() the check if the current message
3021 is already buffered was missing. For every new message was memory
3022 allocated, allowing an attacker to perform an denial of service attack
3023 with sending out of seq handshake messages until there is no memory
3024 left. Additionally every future messege was buffered, even if the
3025 sequence number made no sense and would be part of another handshake.
3026 So only messages with sequence numbers less than 10 in advance will be
3027 buffered. (CVE-2009-1378)
3028 [Robin Seggelmann, discovered by Daniel Mentz]
3030 *) Records are buffered if they arrive with a future epoch to be
3031 processed after finishing the corresponding handshake. There is
3032 currently no limitation to this buffer allowing an attacker to perform
3033 a DOS attack with sending records with future epochs until there is no
3034 memory left. This patch adds the pqueue_size() function to determine
3035 the size of a buffer and limits the record buffer to 100 entries.
3037 [Robin Seggelmann, discovered by Daniel Mentz]
3039 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3040 parent structure is freed. (CVE-2009-1379)
3043 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3044 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3047 [Ilya O. <vrghost@gmail.com>]
3049 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3051 *) Disable renegotiation completely - this fixes a severe security
3052 problem (CVE-2009-3555) at the cost of breaking all
3053 renegotiation. Renegotiation can be re-enabled by setting
3054 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3055 run-time. This is really not recommended unless you know what
3059 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3061 *) Don't set val to NULL when freeing up structures, it is freed up by
3062 underlying code. If sizeof(void *) > sizeof(long) this can result in
3063 zeroing past the valid field. (CVE-2009-0789)
3064 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3066 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3067 checked correctly. This would allow some invalid signed attributes to
3068 appear to verify correctly. (CVE-2009-0591)
3069 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3071 *) Reject UniversalString and BMPString types with invalid lengths. This
3072 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3073 a legal length. (CVE-2009-0590)
3076 *) Set S/MIME signing as the default purpose rather than setting it
3077 unconditionally. This allows applications to override it at the store
3081 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3082 to handle some structures.
3085 *) Improve efficiency of mem_gets: don't search whole buffer each time
3087 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3089 *) New -hex option for openssl rand.
3092 *) Print out UTF8String and NumericString when parsing ASN1.
3095 *) Support NumericString type for name components.
3098 *) Allow CC in the environment to override the automatically chosen
3099 compiler. Note that nothing is done to ensure flags work with the
3103 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3105 *) Properly check EVP_VerifyFinal() and similar return values
3107 [Ben Laurie, Bodo Moeller, Google Security Team]
3109 *) Enable TLS extensions by default.
3112 *) Allow the CHIL engine to be loaded, whether the application is
3113 multithreaded or not. (This does not release the developer from the
3114 obligation to set up the dynamic locking callbacks.)
3115 [Sander Temme <sander@temme.net>]
3117 *) Use correct exit code if there is an error in dgst command.
3118 [Steve Henson; problem pointed out by Roland Dirlewanger]
3120 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3121 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3124 *) Add experimental JPAKE support, including demo authentication in
3125 s_client and s_server.
3128 *) Set the comparison function in v3_addr_canonize().
3129 [Rob Austein <sra@hactrn.net>]
3131 *) Add support for XMPP STARTTLS in s_client.
3132 [Philip Paeps <philip@freebsd.org>]
3134 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3135 to ensure that even with this option, only ciphersuites in the
3136 server's preference list will be accepted. (Note that the option
3137 applies only when resuming a session, so the earlier behavior was
3138 just about the algorithm choice for symmetric cryptography.)
3141 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3143 *) Fix NULL pointer dereference if a DTLS server received
3144 ChangeCipherSpec as first record (CVE-2009-1386).
3147 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3148 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3151 *) The fix in 0.9.8c that supposedly got rid of unsafe
3152 double-checked locking was incomplete for RSA blinding,
3153 addressing just one layer of what turns out to have been
3154 doubly unsafe triple-checked locking.
3156 So now fix this for real by retiring the MONT_HELPER macro
3157 in crypto/rsa/rsa_eay.c.
3159 [Bodo Moeller; problem pointed out by Marius Schilder]
3161 *) Various precautionary measures:
3163 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3165 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3166 (NB: This would require knowledge of the secret session ticket key
3167 to exploit, in which case you'd be SOL either way.)
3169 - Change bn_nist.c so that it will properly handle input BIGNUMs
3170 outside the expected range.
3172 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3175 [Neel Mehta, Bodo Moeller]
3177 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3178 the load fails. Useful for distros.
3179 [Ben Laurie and the FreeBSD team]
3181 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3184 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3187 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3189 This work was sponsored by Logica.
3192 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3193 keystores. Support for SSL/TLS client authentication too.
3194 Not compiled unless enable-capieng specified to Configure.
3196 This work was sponsored by Logica.
3199 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3200 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3201 attribute creation routines such as certifcate requests and PKCS#12
3205 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3207 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3208 handshake which could lead to a cilent crash as found using the
3209 Codenomicon TLS test suite (CVE-2008-1672)
3210 [Steve Henson, Mark Cox]
3212 *) Fix double free in TLS server name extensions which could lead to
3213 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3216 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3218 Clear the error queue to ensure that error entries left from
3219 older function calls do not interfere with the correct operation.
3220 [Lutz Jaenicke, Erik de Castro Lopo]
3222 *) Remove root CA certificates of commercial CAs:
3224 The OpenSSL project does not recommend any specific CA and does not
3225 have any policy with respect to including or excluding any CA.
3226 Therefore it does not make any sense to ship an arbitrary selection
3227 of root CA certificates with the OpenSSL software.
3230 *) RSA OAEP patches to fix two separate invalid memory reads.
3231 The first one involves inputs when 'lzero' is greater than
3232 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3233 before the beginning of from). The second one involves inputs where
3234 the 'db' section contains nothing but zeroes (there is a one-byte
3235 invalid read after the end of 'db').
3236 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3238 *) Partial backport from 0.9.9-dev:
3240 Introduce bn_mul_mont (dedicated Montgomery multiplication
3241 procedure) as a candidate for BIGNUM assembler implementation.
3242 While 0.9.9-dev uses assembler for various architectures, only
3243 x86_64 is available by default here in the 0.9.8 branch, and
3244 32-bit x86 is available through a compile-time setting.
3246 To try the 32-bit x86 assembler implementation, use Configure
3247 option "enable-montasm" (which exists only for this backport).
3249 As "enable-montasm" for 32-bit x86 disclaims code stability
3250 anyway, in this constellation we activate additional code
3251 backported from 0.9.9-dev for further performance improvements,
3252 namely BN_from_montgomery_word. (To enable this otherwise,
3253 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3255 [Andy Polyakov (backport partially by Bodo Moeller)]
3257 *) Add TLS session ticket callback. This allows an application to set
3258 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3259 values. This is useful for key rollover for example where several key
3260 sets may exist with different names.
3263 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3264 This was broken until now in 0.9.8 releases, such that the only way
3265 a registered ENGINE could be used (assuming it initialises
3266 successfully on the host) was to explicitly set it as the default
3267 for the relevant algorithms. This is in contradiction with 0.9.7
3268 behaviour and the documentation. With this fix, when an ENGINE is
3269 registered into a given algorithm's table of implementations, the
3270 'uptodate' flag is reset so that auto-discovery will be used next
3271 time a new context for that algorithm attempts to select an
3273 [Ian Lister (tweaked by Geoff Thorpe)]
3275 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3276 implemention in the following ways:
3278 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3281 Lack of BER streaming support means one pass streaming processing is
3282 only supported if data is detached: setting the streaming flag is
3283 ignored for embedded content.
3285 CMS support is disabled by default and must be explicitly enabled
3286 with the enable-cms configuration option.
3289 *) Update the GMP engine glue to do direct copies between BIGNUM and
3290 mpz_t when openssl and GMP use the same limb size. Otherwise the
3291 existing "conversion via a text string export" trick is still used.
3292 [Paul Sheer <paulsheer@gmail.com>]
3294 *) Zlib compression BIO. This is a filter BIO which compressed and
3295 uncompresses any data passed through it.
3298 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3299 RFC3394 compatible AES key wrapping.
3302 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3303 sets string data without copying. X509_ALGOR_set0() and
3304 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3305 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3306 from an X509_ATTRIBUTE structure optionally checking it occurs only
3307 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3311 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3312 to get the expected BN_FLG_CONSTTIME behavior.
3313 [Bodo Moeller (Google)]
3317 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3318 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3319 - added some more tests to do_tests.pl
3320 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3321 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3322 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3323 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3324 - various changes to netware.pl to enable gcc-cross builds on Win32
3326 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3327 - various changes to fix missing prototype warnings
3328 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3329 - added AES, WHIRLPOOL and CPUID assembler code to build files
3330 - added missing AES assembler make rules to mk1mf.pl
3331 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3332 [Guenter Knauf <eflash@gmx.net>]
3334 *) Implement certificate status request TLS extension defined in RFC3546.
3335 A client can set the appropriate parameters and receive the encoded
3336 OCSP response via a callback. A server can query the supplied parameters
3337 and set the encoded OCSP response in the callback. Add simplified examples
3338 to s_client and s_server.
3341 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3343 *) Fix various bugs:
3344 + Binary incompatibility of ssl_ctx_st structure
3345 + DTLS interoperation with non-compliant servers
3346 + Don't call get_session_cb() without proposed session
3347 + Fix ia64 assembler code
3348 [Andy Polyakov, Steve Henson]
3350 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3352 *) DTLS Handshake overhaul. There were longstanding issues with
3353 OpenSSL DTLS implementation, which were making it impossible for
3354 RFC 4347 compliant client to communicate with OpenSSL server.
3355 Unfortunately just fixing these incompatibilities would "cut off"
3356 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3357 server keeps tolerating non RFC compliant syntax. The opposite is
3358 not true, 0.9.8f client can not communicate with earlier server.
3359 This update even addresses CVE-2007-4995.
3362 *) Changes to avoid need for function casts in OpenSSL: some compilers
3363 (gcc 4.2 and later) reject their use.
3364 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3367 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3368 RFC4507bis. The encrypted ticket format is an encrypted encoded
3369 SSL_SESSION structure, that way new session features are automatically
3372 If a client application caches session in an SSL_SESSION structure
3373 support is transparent because tickets are now stored in the encoded
3376 The SSL_CTX structure automatically generates keys for ticket
3377 protection in servers so again support should be possible
3378 with no application modification.
3380 If a client or server wishes to disable RFC4507 support then the option
3381 SSL_OP_NO_TICKET can be set.
3383 Add a TLS extension debugging callback to allow the contents of any client
3384 or server extensions to be examined.
3386 This work was sponsored by Google.
3389 *) Add initial support for TLS extensions, specifically for the server_name
3390 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3391 have new members for a host name. The SSL data structure has an
3392 additional member SSL_CTX *initial_ctx so that new sessions can be
3393 stored in that context to allow for session resumption, even after the
3394 SSL has been switched to a new SSL_CTX in reaction to a client's
3395 server_name extension.
3397 New functions (subject to change):
3399 SSL_get_servername()
3400 SSL_get_servername_type()
3403 New CTRL codes and macros (subject to change):
3405 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3406 - SSL_CTX_set_tlsext_servername_callback()
3407 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3408 - SSL_CTX_set_tlsext_servername_arg()
3409 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3411 openssl s_client has a new '-servername ...' option.
3413 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3414 '-key2 ...', '-servername_fatal' (subject to change). This allows
3415 testing the HostName extension for a specific single host name ('-cert'
3416 and '-key' remain fallbacks for handshakes without HostName
3417 negotiation). If the unrecognized_name alert has to be sent, this by
3418 default is a warning; it becomes fatal with the '-servername_fatal'
3421 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3423 *) Add AES and SSE2 assembly language support to VC++ build.
3426 *) Mitigate attack on final subtraction in Montgomery reduction.
3429 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3430 (which previously caused an internal error).
3433 *) Squeeze another 10% out of IGE mode when in != out.
3436 *) AES IGE mode speedup.
3437 [Dean Gaudet (Google)]
3439 *) Add the Korean symmetric 128-bit cipher SEED (see
3440 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3441 add SEED ciphersuites from RFC 4162:
3443 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3444 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3445 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3446 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3448 To minimize changes between patchlevels in the OpenSSL 0.9.8
3449 series, SEED remains excluded from compilation unless OpenSSL
3450 is configured with 'enable-seed'.
3451 [KISA, Bodo Moeller]
3453 *) Mitigate branch prediction attacks, which can be practical if a
3454 single processor is shared, allowing a spy process to extract
3455 information. For detailed background information, see
3456 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3457 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3458 and Necessary Software Countermeasures"). The core of the change
3459 are new versions BN_div_no_branch() and
3460 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3461 respectively, which are slower, but avoid the security-relevant
3462 conditional branches. These are automatically called by BN_div()
3463 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3464 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3465 remove a conditional branch.
3467 BN_FLG_CONSTTIME is the new name for the previous
3468 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3469 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3470 in the exponent causes BN_mod_exp_mont() to use the alternative
3471 implementation in BN_mod_exp_mont_consttime().) The old name
3472 remains as a deprecated alias.
3474 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3475 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3476 constant-time implementations for more than just exponentiation.
3477 Here too the old name is kept as a deprecated alias.
3479 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3480 the BN_BLINDING structure gets an independent copy of the
3481 modulus. This means that the previous "BIGNUM *m" argument to
3482 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3483 essentially becomes "const BIGNUM *m", although we can't actually
3484 change this in the header file before 0.9.9. It allows
3485 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3486 enable BN_FLG_CONSTTIME.
3488 [Matthew D Wood (Intel Corp)]
3490 *) In the SSL/TLS server implementation, be strict about session ID
3491 context matching (which matters if an application uses a single
3492 external cache for different purposes). Previously,
3493 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3494 set. This did ensure strict client verification, but meant that,
3495 with applications using a single external cache for quite
3496 different requirements, clients could circumvent ciphersuite
3497 restrictions for a given session ID context by starting a session
3498 in a different context.
3501 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3502 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3503 authentication-only ciphersuites.
3506 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3507 not complete and could lead to a possible single byte overflow
3508 (CVE-2007-5135) [Ben Laurie]
3510 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3512 *) Since AES128 and AES256 (and similarly Camellia128 and
3513 Camellia256) share a single mask bit in the logic of
3514 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3515 kludge to work properly if AES128 is available and AES256 isn't
3516 (or if Camellia128 is available and Camellia256 isn't).
3519 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3520 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3521 When a point or a seed is encoded in a BIT STRING, we need to
3522 prevent the removal of trailing zero bits to get the proper DER
3523 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3524 of a NamedBitList, for which trailing 0 bits need to be removed.)
3527 *) Have SSL/TLS server implementation tolerate "mismatched" record
3528 protocol version while receiving ClientHello even if the
3529 ClientHello is fragmented. (The server can't insist on the
3530 particular protocol version it has chosen before the ServerHello
3531 message has informed the client about his choice.)
3534 *) Add RFC 3779 support.
3535 [Rob Austein for ARIN, Ben Laurie]
3537 *) Load error codes if they are not already present instead of using a
3538 static variable. This allows them to be cleanly unloaded and reloaded.
3539 Improve header file function name parsing.
3542 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3543 or CAPABILITY handshake as required by RFCs.
3546 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3548 *) Introduce limits to prevent malicious keys being able to
3549 cause a denial of service. (CVE-2006-2940)
3550 [Steve Henson, Bodo Moeller]
3552 *) Fix ASN.1 parsing of certain invalid structures that can result
3553 in a denial of service. (CVE-2006-2937) [Steve Henson]
3555 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3556 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3558 *) Fix SSL client code which could crash if connecting to a
3559 malicious SSLv2 server. (CVE-2006-4343)
3560 [Tavis Ormandy and Will Drewry, Google Security Team]
3562 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3563 match only those. Before that, "AES256-SHA" would be interpreted
3564 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3565 the same strength classification in 0.9.7h) as we currently only
3566 have a single AES bit in the ciphersuite description bitmap.
3567 That change, however, also applied to ciphersuite strings such as
3568 "RC4-MD5" that intentionally matched multiple ciphersuites --
3569 namely, SSL 2.0 ciphersuites in addition to the more common ones
3570 from SSL 3.0/TLS 1.0.
3572 So we change the selection algorithm again: Naming an explicit
3573 ciphersuite selects this one ciphersuite, and any other similar
3574 ciphersuite (same bitmap) from *other* protocol versions.
3575 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3576 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3578 Since SSL 2.0 does not have any ciphersuites for which the
3579 128/256 bit distinction would be relevant, this works for now.
3580 The proper fix will be to use different bits for AES128 and
3581 AES256, which would have avoided the problems from the beginning;
3582 however, bits are scarce, so we can only do this in a new release
3583 (not just a patchlevel) when we can change the SSL_CIPHER
3584 definition to split the single 'unsigned long mask' bitmap into
3585 multiple values to extend the available space.
3589 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3591 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3592 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3594 *) Add AES IGE and biIGE modes.
3597 *) Change the Unix randomness entropy gathering to use poll() when
3598 possible instead of select(), since the latter has some
3599 undesirable limitations.
3600 [Darryl Miles via Richard Levitte and Bodo Moeller]
3602 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3603 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3604 cannot be implicitly activated as part of, e.g., the "AES" alias.
3605 However, please upgrade to OpenSSL 0.9.9[-dev] for
3606 non-experimental use of the ECC ciphersuites to get TLS extension
3607 support, which is required for curve and point format negotiation
3608 to avoid potential handshake problems.
3611 *) Disable rogue ciphersuites:
3613 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3614 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3615 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3617 The latter two were purportedly from
3618 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3621 Also deactivate the remaining ciphersuites from
3622 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3623 unofficial, and the ID has long expired.
3626 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3627 dual-core machines) and other potential thread-safety issues.
3630 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3631 versions), which is now available for royalty-free use
3632 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3633 Also, add Camellia TLS ciphersuites from RFC 4132.
3635 To minimize changes between patchlevels in the OpenSSL 0.9.8
3636 series, Camellia remains excluded from compilation unless OpenSSL
3637 is configured with 'enable-camellia'.
3640 *) Disable the padding bug check when compression is in use. The padding
3641 bug check assumes the first packet is of even length, this is not
3642 necessarily true if compresssion is enabled and can result in false
3643 positives causing handshake failure. The actual bug test is ancient
3644 code so it is hoped that implementations will either have fixed it by
3645 now or any which still have the bug do not support compression.
3648 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3650 *) When applying a cipher rule check to see if string match is an explicit
3651 cipher suite and only match that one cipher suite if it is.
3654 *) Link in manifests for VC++ if needed.
3655 [Austin Ziegler <halostatue@gmail.com>]
3657 *) Update support for ECC-based TLS ciphersuites according to
3658 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3659 TLS extensions, which are supported starting with the 0.9.9
3660 branch, not in the OpenSSL 0.9.8 branch).
3663 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3664 opaque EVP_CIPHER_CTX handling.
3667 *) Fixes and enhancements to zlib compression code. We now only use
3668 "zlib1.dll" and use the default __cdecl calling convention on Win32
3669 to conform with the standards mentioned here:
3670 http://www.zlib.net/DLL_FAQ.txt
3671 Static zlib linking now works on Windows and the new --with-zlib-include
3672 --with-zlib-lib options to Configure can be used to supply the location
3673 of the headers and library. Gracefully handle case where zlib library
3677 *) Several fixes and enhancements to the OID generation code. The old code
3678 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3679 handle numbers larger than ULONG_MAX, truncated printing and had a
3680 non standard OBJ_obj2txt() behaviour.
3683 *) Add support for building of engines under engine/ as shared libraries
3684 under VC++ build system.
3687 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3688 Hopefully, we will not see any false combination of paths any more.
3691 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3693 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3694 (part of SSL_OP_ALL). This option used to disable the
3695 countermeasure against man-in-the-middle protocol-version
3696 rollback in the SSL 2.0 server implementation, which is a bad
3697 idea. (CVE-2005-2969)
3699 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3700 for Information Security, National Institute of Advanced Industrial
3701 Science and Technology [AIST], Japan)]
3703 *) Add two function to clear and return the verify parameter flags.
3706 *) Keep cipherlists sorted in the source instead of sorting them at
3707 runtime, thus removing the need for a lock.
3710 *) Avoid some small subgroup attacks in Diffie-Hellman.
3711 [Nick Mathewson and Ben Laurie]
3713 *) Add functions for well-known primes.
3716 *) Extended Windows CE support.
3717 [Satoshi Nakamura and Andy Polyakov]
3719 *) Initialize SSL_METHOD structures at compile time instead of during
3720 runtime, thus removing the need for a lock.
3723 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3724 attempting to decrypt each encrypted key in turn. Add support to
3728 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3730 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3733 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3736 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3737 key into the same file any more.
3740 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3743 *) Add -utf8 command line and config file option to 'ca'.
3744 [Stefan <stf@udoma.org]
3746 *) Removed the macro des_crypt(), as it seems to conflict with some
3747 libraries. Use DES_crypt().
3750 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3751 involves renaming the source and generated shared-libs for
3752 both. The engines will accept the corrected or legacy ids
3753 ('ncipher' and '4758_cca' respectively) when binding. NB,
3754 this only applies when building 'shared'.
3755 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3757 *) Add attribute functions to EVP_PKEY structure. Modify
3758 PKCS12_create() to recognize a CSP name attribute and
3759 use it. Make -CSP option work again in pkcs12 utility.
3762 *) Add new functionality to the bn blinding code:
3763 - automatic re-creation of the BN_BLINDING parameters after
3764 a fixed number of uses (currently 32)
3765 - add new function for parameter creation
3766 - introduce flags to control the update behaviour of the
3767 BN_BLINDING parameters
3768 - hide BN_BLINDING structure
3769 Add a second BN_BLINDING slot to the RSA structure to improve
3770 performance when a single RSA object is shared among several
3774 *) Add support for DTLS.
3775 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3777 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3778 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3781 *) Remove buggy and incomplete DH cert support from
3782 ssl/ssl_rsa.c and ssl/s3_both.c
3785 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3786 the apps/openssl applications.
3789 *) Compile clean with "-Wall -Wmissing-prototypes
3790 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3791 DEBUG_SAFESTACK must also be set.
3794 *) Change ./Configure so that certain algorithms can be disabled by default.
3795 The new counterpiece to "no-xxx" is "enable-xxx".
3797 The patented RC5 and MDC2 algorithms will now be disabled unless
3798 "enable-rc5" and "enable-mdc2", respectively, are specified.
3800 (IDEA remains enabled despite being patented. This is because IDEA
3801 is frequently required for interoperability, and there is no license
3802 fee for non-commercial use. As before, "no-idea" can be used to
3803 avoid this algorithm.)
3807 *) Add processing of proxy certificates (see RFC 3820). This work was
3808 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3809 EGEE (Enabling Grids for E-science in Europe).
3812 *) RC4 performance overhaul on modern architectures/implementations, such
3813 as Intel P4, IA-64 and AMD64.
3816 *) New utility extract-section.pl. This can be used specify an alternative
3817 section number in a pod file instead of having to treat each file as
3818 a separate case in Makefile. This can be done by adding two lines to the
3821 =for comment openssl_section:XXX
3823 The blank line is mandatory.
3827 *) New arguments -certform, -keyform and -pass for s_client and s_server
3828 to allow alternative format key and certificate files and passphrase
3832 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3833 update associated structures and add various utility functions.
3835 Add new policy related verify parameters, include policy checking in
3836 standard verify code. Enhance 'smime' application with extra parameters
3837 to support policy checking and print out.
3840 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3841 Nehemiah processors. These extensions support AES encryption in hardware
3842 as well as RNG (though RNG support is currently disabled).
3843 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3845 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3848 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3849 [Andy Polyakov and a number of other people]
3851 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3852 implementation contributed by IBM.
3853 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3855 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3856 exponent rather than 'unsigned long'. There is a corresponding change to
3857 the new 'rsa_keygen' element of the RSA_METHOD structure.
3858 [Jelte Jansen, Geoff Thorpe]
3860 *) Functionality for creating the initial serial number file is now
3861 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3863 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3864 number file to 1, which is bound to cause problems. To avoid
3865 the problems while respecting compatibility between different 0.9.7
3866 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3867 CA.pl for serial number initialization. With the new release 0.9.8,
3868 we can fix the problem directly in the 'ca' utility.)
3871 *) Reduced header interdepencies by declaring more opaque objects in
3872 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3873 give fewer recursive includes, which could break lazy source code - so
3874 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3875 developers should define this symbol when building and using openssl to
3876 ensure they track the recommended behaviour, interfaces, [etc], but
3877 backwards-compatible behaviour prevails when this isn't defined.
3880 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3883 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3884 This will generate a random key of the appropriate length based on the
3885 cipher context. The EVP_CIPHER can provide its own random key generation
3886 routine to support keys of a specific form. This is used in the des and
3887 3des routines to generate a key of the correct parity. Update S/MIME
3888 code to use new functions and hence generate correct parity DES keys.
3889 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3890 valid (weak or incorrect parity).
3893 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3894 as looking them up. This is useful when the verified structure may contain
3895 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3896 present unless the new PKCS7_NO_CRL flag is asserted.
3899 *) Extend ASN1 oid configuration module. It now additionally accepts the
3902 shortName = some long name, 1.2.3.4
3905 *) Reimplemented the BN_CTX implementation. There is now no more static
3906 limitation on the number of variables it can handle nor the depth of the
3907 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3908 information can now expand as required, and rather than having a single
3909 static array of bignums, BN_CTX now uses a linked-list of such arrays
3910 allowing it to expand on demand whilst maintaining the usefulness of
3911 BN_CTX's "bundling".
3914 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3915 to allow all RSA operations to function using a single BN_CTX.
3918 *) Preliminary support for certificate policy evaluation and checking. This
3919 is initially intended to pass the tests outlined in "Conformance Testing
3920 of Relying Party Client Certificate Path Processing Logic" v1.07.
3923 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3924 remained unused and not that useful. A variety of other little bignum
3925 tweaks and fixes have also been made continuing on from the audit (see
3929 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3930 associated ASN1, EVP and SSL functions and old ASN1 macros.
3933 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3934 and this should never fail. So the return value from the use of
3935 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3936 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3939 *) BN_CTX_get() should return zero-valued bignums, providing the same
3940 initialised value as BN_new().
3941 [Geoff Thorpe, suggested by Ulf Möller]
3943 *) Support for inhibitAnyPolicy certificate extension.
3946 *) An audit of the BIGNUM code is underway, for which debugging code is
3947 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3948 is considered valid when processing BIGNUMs, and causes execution to
3949 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3950 further steps are taken to deliberately pollute unused data in BIGNUM
3951 structures to try and expose faulty code further on. For now, openssl will
3952 (in its default mode of operation) continue to tolerate the inconsistent
3953 forms that it has tolerated in the past, but authors and packagers should
3954 consider trying openssl and their own applications when compiled with
3955 these debugging symbols defined. It will help highlight potential bugs in
3956 their own code, and will improve the test coverage for OpenSSL itself. At
3957 some point, these tighter rules will become openssl's default to improve
3958 maintainability, though the assert()s and other overheads will remain only
3959 in debugging configurations. See bn.h for more details.
3960 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3962 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3963 that can only be obtained through BN_CTX_new() (which implicitly
3964 initialises it). The presence of this function only made it possible
3965 to overwrite an existing structure (and cause memory leaks).
3968 *) Because of the callback-based approach for implementing LHASH as a
3969 template type, lh_insert() adds opaque objects to hash-tables and
3970 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3971 to clean up those corresponding objects before destroying the hash table
3972 (and losing the object pointers). So some over-zealous constifications in
3973 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3974 objects as "const" and the lh_doall[_arg] callback wrappers are not
3975 prototyped to have "const" restrictions on the object pointers they are
3976 given (and so aren't required to cast them away any more).
3979 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3980 (speed) prefers to use its own implementation. The two implementations
3981 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3982 its object type properly exposed (MS_TM) instead of casting to/from "char
3983 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3984 aren't necessarily the greatest nomenclatures - but this is what was used
3985 internally to the implementation so I've used that for now.
3988 *) Ensure that deprecated functions do not get compiled when
3989 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3990 the self-tests were still using deprecated key-generation functions so
3991 these have been updated also.
3994 *) Reorganise PKCS#7 code to separate the digest location functionality
3995 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3996 New function PKCS7_set_digest() to set the digest type for PKCS#7
3997 digestedData type. Add additional code to correctly generate the
3998 digestedData type and add support for this type in PKCS7 initialization
4002 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4003 structure of type "other".
4006 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4007 sure the loop does correctly stop and breaking ("division by zero")
4008 modulus operations are not performed. The (pre-generated) prime
4009 table crypto/bn/bn_prime.h was already correct, but it could not be
4010 re-generated on some platforms because of the "division by zero"
4011 situation in the script.
4012 [Ralf S. Engelschall]
4014 *) Update support for ECC-based TLS ciphersuites according to
4015 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4016 SHA-1 now is only used for "small" curves (where the
4017 representation of a field element takes up to 24 bytes); for
4018 larger curves, the field element resulting from ECDH is directly
4019 used as premaster secret.
4020 [Douglas Stebila (Sun Microsystems Laboratories)]
4022 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4023 curve secp160r1 to the tests.
4024 [Douglas Stebila (Sun Microsystems Laboratories)]
4026 *) Add the possibility to load symbols globally with DSO.
4027 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4029 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4030 control of the error stack.
4033 *) Add support for STORE in ENGINE.
4036 *) Add the STORE type. The intention is to provide a common interface
4037 to certificate and key stores, be they simple file-based stores, or
4038 HSM-type store, or LDAP stores, or...
4039 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4042 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4043 pass a list of arguments to any function as well as provide a way
4044 for a function to pass data back to the caller.
4047 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4048 works like BUF_strdup() but can be used to duplicate a portion of
4049 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4053 *) Add the function sk_find_ex() which works like sk_find(), but will
4054 return an index to an element even if an exact match couldn't be
4055 found. The index is guaranteed to point at the element where the
4056 searched-for key would be inserted to preserve sorting order.
4059 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4060 takes an extra flags argument for optional functionality. Currently,
4061 the following flags are defined:
4063 OBJ_BSEARCH_VALUE_ON_NOMATCH
4064 This one gets OBJ_bsearch_ex() to return a pointer to the first
4065 element where the comparing function returns a negative or zero
4068 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4069 This one gets OBJ_bsearch_ex() to return a pointer to the first
4070 element where the comparing function returns zero. This is useful
4071 if there are more than one element where the comparing function
4075 *) Make it possible to create self-signed certificates with 'openssl ca'
4076 in such a way that the self-signed certificate becomes part of the
4077 CA database and uses the same mechanisms for serial number generation
4078 as all other certificate signing. The new flag '-selfsign' enables
4079 this functionality. Adapt CA.sh and CA.pl.in.
4082 *) Add functionality to check the public key of a certificate request
4083 against a given private. This is useful to check that a certificate
4084 request can be signed by that key (self-signing).
4087 *) Make it possible to have multiple active certificates with the same
4088 subject in the CA index file. This is done only if the keyword
4089 'unique_subject' is set to 'no' in the main CA section (default
4090 if 'CA_default') of the configuration file. The value is saved
4091 with the database itself in a separate index attribute file,
4092 named like the index file with '.attr' appended to the name.
4095 *) Generate muti valued AVAs using '+' notation in config files for
4099 *) Support for nameConstraints certificate extension.
4102 *) Support for policyConstraints certificate extension.
4105 *) Support for policyMappings certificate extension.
4108 *) Make sure the default DSA_METHOD implementation only uses its
4109 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4110 and change its own handlers to be NULL so as to remove unnecessary
4111 indirection. This lets alternative implementations fallback to the
4112 default implementation more easily.
4115 *) Support for directoryName in GeneralName related extensions
4119 *) Make it possible to link applications using Makefile.shared.
4120 Make that possible even when linking against static libraries!
4123 *) Support for single pass processing for S/MIME signing. This now
4124 means that S/MIME signing can be done from a pipe, in addition
4125 cleartext signing (multipart/signed type) is effectively streaming
4126 and the signed data does not need to be all held in memory.
4128 This is done with a new flag PKCS7_STREAM. When this flag is set
4129 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4130 is done after the data is output (and digests calculated) in
4131 SMIME_write_PKCS7().
4134 *) Add full support for -rpath/-R, both in shared libraries and
4135 applications, at least on the platforms where it's known how
4139 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4140 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4141 will now compute a table of multiples of the generator that
4142 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4143 faster (notably in the case of a single point multiplication,
4144 scalar * generator).
4145 [Nils Larsch, Bodo Moeller]
4147 *) IPv6 support for certificate extensions. The various extensions
4148 which use the IP:a.b.c.d can now take IPv6 addresses using the
4149 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4153 *) Added an ENGINE that implements RSA by performing private key
4154 exponentiations with the GMP library. The conversions to and from
4155 GMP's mpz_t format aren't optimised nor are any montgomery forms
4156 cached, and on x86 it appears OpenSSL's own performance has caught up.
4157 However there are likely to be other architectures where GMP could
4158 provide a boost. This ENGINE is not built in by default, but it can be
4159 specified at Configure time and should be accompanied by the necessary
4160 linker additions, eg;
4161 ./config -DOPENSSL_USE_GMP -lgmp
4164 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4165 testing availability of engines with "-t" - the old behaviour is
4166 produced by increasing the feature's verbosity with "-tt".
4169 *) ECDSA routines: under certain error conditions uninitialized BN objects
4170 could be freed. Solution: make sure initialization is performed early
4171 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4175 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4176 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4177 software implementations. For DSA and DH, parameter generation can
4178 also be overriden by providing the appropriate method callbacks.
4181 *) Change the "progress" mechanism used in key-generation and
4182 primality testing to functions that take a new BN_GENCB pointer in
4183 place of callback/argument pairs. The new API functions have "_ex"
4184 postfixes and the older functions are reimplemented as wrappers for
4185 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4186 declarations of the old functions to help (graceful) attempts to
4187 migrate to the new functions. Also, the new key-generation API
4188 functions operate on a caller-supplied key-structure and return
4189 success/failure rather than returning a key or NULL - this is to
4190 help make "keygen" another member function of RSA_METHOD etc.
4192 Example for using the new callback interface:
4194 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4198 BN_GENCB_set(&my_cb, my_callback, my_arg);
4200 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4201 /* For the meaning of a, b in calls to my_callback(), see the
4202 * documentation of the function that calls the callback.
4203 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4204 * my_callback should return 1 if it wants BN_is_prime_ex()
4205 * to continue, or 0 to stop.
4210 *) Change the ZLIB compression method to be stateful, and make it
4211 available to TLS with the number defined in
4212 draft-ietf-tls-compression-04.txt.
4215 *) Add the ASN.1 structures and functions for CertificatePair, which
4216 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4218 CertificatePair ::= SEQUENCE {
4219 forward [0] Certificate OPTIONAL,
4220 reverse [1] Certificate OPTIONAL,
4221 -- at least one of the pair shall be present -- }
4223 Also implement the PEM functions to read and write certificate
4224 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4226 This needed to be defined, mostly for the sake of the LDAP
4227 attribute crossCertificatePair, but may prove useful elsewhere as
4231 *) Make it possible to inhibit symlinking of shared libraries in
4232 Makefile.shared, for Cygwin's sake.
4235 *) Extend the BIGNUM API by creating a function
4236 void BN_set_negative(BIGNUM *a, int neg);
4237 and a macro that behave like
4238 int BN_is_negative(const BIGNUM *a);
4240 to avoid the need to access 'a->neg' directly in applications.
4243 *) Implement fast modular reduction for pseudo-Mersenne primes
4244 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4245 EC_GROUP_new_curve_GFp() will now automatically use this
4247 [Nils Larsch <nla@trustcenter.de>]
4249 *) Add new lock type (CRYPTO_LOCK_BN).
4252 *) Change the ENGINE framework to automatically load engines
4253 dynamically from specific directories unless they could be
4254 found to already be built in or loaded. Move all the
4255 current engines except for the cryptodev one to a new
4257 The engines in engines/ are built as shared libraries if
4258 the "shared" options was given to ./Configure or ./config.
4259 Otherwise, they are inserted in libcrypto.a.
4260 /usr/local/ssl/engines is the default directory for dynamic
4261 engines, but that can be overriden at configure time through
4262 the usual use of --prefix and/or --openssldir, and at run
4263 time with the environment variable OPENSSL_ENGINES.
4264 [Geoff Thorpe and Richard Levitte]
4266 *) Add Makefile.shared, a helper makefile to build shared
4267 libraries. Adapt Makefile.org.
4270 *) Add version info to Win32 DLLs.
4271 [Peter 'Luna' Runestig" <peter@runestig.com>]
4273 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4274 can be added using this API to created arbitrary PKCS#12
4275 files while avoiding the low level API.
4277 New options to PKCS12_create(), key or cert can be NULL and
4278 will then be omitted from the output file. The encryption
4279 algorithm NIDs can be set to -1 for no encryption, the mac
4280 iteration count can be set to 0 to omit the mac.
4282 Enhance pkcs12 utility by making the -nokeys and -nocerts
4283 options work when creating a PKCS#12 file. New option -nomac
4284 to omit the mac, NONE can be set for an encryption algorithm.
4285 New code is modified to use the enhanced PKCS12_create()
4286 instead of the low level API.
4289 *) Extend ASN1 encoder to support indefinite length constructed
4290 encoding. This can output sequences tags and octet strings in
4291 this form. Modify pk7_asn1.c to support indefinite length
4292 encoding. This is experimental and needs additional code to
4293 be useful, such as an ASN1 bio and some enhanced streaming
4296 Extend template encode functionality so that tagging is passed
4297 down to the template encoder.
4300 *) Let 'openssl req' fail if an argument to '-newkey' is not
4301 recognized instead of using RSA as a default.
4304 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4305 As these are not official, they are not included in "ALL";
4306 the "ECCdraft" ciphersuite group alias can be used to select them.
4307 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4309 *) Add ECDH engine support.
4310 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4312 *) Add ECDH in new directory crypto/ecdh/.
4313 [Douglas Stebila (Sun Microsystems Laboratories)]
4315 *) Let BN_rand_range() abort with an error after 100 iterations
4316 without success (which indicates a broken PRNG).
4319 *) Change BN_mod_sqrt() so that it verifies that the input value
4320 is really the square of the return value. (Previously,
4321 BN_mod_sqrt would show GIGO behaviour.)
4324 *) Add named elliptic curves over binary fields from X9.62, SECG,
4325 and WAP/WTLS; add OIDs that were still missing.
4327 [Sheueling Chang Shantz and Douglas Stebila
4328 (Sun Microsystems Laboratories)]
4330 *) Extend the EC library for elliptic curves over binary fields
4331 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4334 EC_GF2m_simple_method
4338 EC_GROUP_new_curve_GF2m
4339 EC_GROUP_set_curve_GF2m
4340 EC_GROUP_get_curve_GF2m
4341 EC_POINT_set_affine_coordinates_GF2m
4342 EC_POINT_get_affine_coordinates_GF2m
4343 EC_POINT_set_compressed_coordinates_GF2m
4345 Point compression for binary fields is disabled by default for
4346 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4349 As binary polynomials are represented as BIGNUMs, various members
4350 of the EC_GROUP and EC_POINT data structures can be shared
4351 between the implementations for prime fields and binary fields;
4352 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4353 are essentially identical to their ..._GFp counterparts.
4354 (For simplicity, the '..._GFp' prefix has been dropped from
4355 various internal method names.)
4357 An internal 'field_div' method (similar to 'field_mul' and
4358 'field_sqr') has been added; this is used only for binary fields.
4360 [Sheueling Chang Shantz and Douglas Stebila
4361 (Sun Microsystems Laboratories)]
4363 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4364 through methods ('mul', 'precompute_mult').
4366 The generic implementations (now internally called 'ec_wNAF_mul'
4367 and 'ec_wNAF_precomputed_mult') remain the default if these
4368 methods are undefined.
4370 [Sheueling Chang Shantz and Douglas Stebila
4371 (Sun Microsystems Laboratories)]
4373 *) New function EC_GROUP_get_degree, which is defined through
4374 EC_METHOD. For curves over prime fields, this returns the bit
4375 length of the modulus.
4377 [Sheueling Chang Shantz and Douglas Stebila
4378 (Sun Microsystems Laboratories)]
4380 *) New functions EC_GROUP_dup, EC_POINT_dup.
4381 (These simply call ..._new and ..._copy).
4383 [Sheueling Chang Shantz and Douglas Stebila
4384 (Sun Microsystems Laboratories)]
4386 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4387 Polynomials are represented as BIGNUMs (where the sign bit is not
4388 used) in the following functions [macros]:
4391 BN_GF2m_sub [= BN_GF2m_add]
4392 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4393 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4394 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4396 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4397 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4398 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4399 BN_GF2m_cmp [= BN_ucmp]
4401 (Note that only the 'mod' functions are actually for fields GF(2^m).
4402 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4404 For some functions, an the irreducible polynomial defining a
4405 field can be given as an 'unsigned int[]' with strictly
4406 decreasing elements giving the indices of those bits that are set;
4407 i.e., p[] represents the polynomial
4408 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4410 p[0] > p[1] > ... > p[k] = 0.
4411 This applies to the following functions:
4416 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4417 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4419 BN_GF2m_mod_sqrt_arr
4420 BN_GF2m_mod_solve_quad_arr
4424 Conversion can be performed by the following functions:
4429 bntest.c has additional tests for binary polynomial arithmetic.
4431 Two implementations for BN_GF2m_mod_div() are available.
4432 The default algorithm simply uses BN_GF2m_mod_inv() and
4433 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4434 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4435 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4437 [Sheueling Chang Shantz and Douglas Stebila
4438 (Sun Microsystems Laboratories)]
4440 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4441 functionality is disabled at compile-time.
4442 [Douglas Stebila <douglas.stebila@sun.com>]
4444 *) Change default behaviour of 'openssl asn1parse' so that more
4445 information is visible when viewing, e.g., a certificate:
4447 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4448 mode the content of non-printable OCTET STRINGs is output in a
4449 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4450 avoid the appearance of a printable string.
4451 [Nils Larsch <nla@trustcenter.de>]
4453 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4455 EC_GROUP_set_asn1_flag()
4456 EC_GROUP_get_asn1_flag()
4457 EC_GROUP_set_point_conversion_form()
4458 EC_GROUP_get_point_conversion_form()
4459 These control ASN1 encoding details:
4460 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4461 has been set to OPENSSL_EC_NAMED_CURVE.
4462 - Points are encoded in uncompressed form by default; options for
4463 asn1_for are as for point2oct, namely
4464 POINT_CONVERSION_COMPRESSED
4465 POINT_CONVERSION_UNCOMPRESSED
4466 POINT_CONVERSION_HYBRID
4468 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4471 EC_GROUP_get0_seed()
4472 EC_GROUP_get_seed_len()
4473 This is used only for ASN1 purposes (so far).
4474 [Nils Larsch <nla@trustcenter.de>]
4476 *) Add 'field_type' member to EC_METHOD, which holds the NID
4477 of the appropriate field type OID. The new function
4478 EC_METHOD_get_field_type() returns this value.
4479 [Nils Larsch <nla@trustcenter.de>]
4484 EC_POINT_point2hex()
4485 EC_POINT_hex2point()
4486 providing useful interfaces to EC_POINT_point2oct() and
4487 EC_POINT_oct2point().
4488 [Nils Larsch <nla@trustcenter.de>]
4490 *) Change internals of the EC library so that the functions
4491 EC_GROUP_set_generator()
4492 EC_GROUP_get_generator()
4493 EC_GROUP_get_order()
4494 EC_GROUP_get_cofactor()
4495 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4496 to methods, which would lead to unnecessary code duplication when
4497 adding different types of curves.
4498 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4500 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4501 arithmetic, and such that modified wNAFs are generated
4502 (which avoid length expansion in many cases).
4505 *) Add a function EC_GROUP_check_discriminant() (defined via
4506 EC_METHOD) that verifies that the curve discriminant is non-zero.
4508 Add a function EC_GROUP_check() that makes some sanity tests
4509 on a EC_GROUP, its generator and order. This includes
4510 EC_GROUP_check_discriminant().
4511 [Nils Larsch <nla@trustcenter.de>]
4513 *) Add ECDSA in new directory crypto/ecdsa/.
4515 Add applications 'openssl ecparam' and 'openssl ecdsa'
4516 (these are based on 'openssl dsaparam' and 'openssl dsa').
4518 ECDSA support is also included in various other files across the
4519 library. Most notably,
4520 - 'openssl req' now has a '-newkey ecdsa:file' option;
4521 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4522 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4523 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4524 them suitable for ECDSA where domain parameters must be
4525 extracted before the specific public key;
4526 - ECDSA engine support has been added.
4527 [Nils Larsch <nla@trustcenter.de>]
4529 *) Include some named elliptic curves, and add OIDs from X9.62,
4530 SECG, and WAP/WTLS. Each curve can be obtained from the new
4532 EC_GROUP_new_by_curve_name(),
4533 and the list of available named curves can be obtained with
4534 EC_get_builtin_curves().
4535 Also add a 'curve_name' member to EC_GROUP objects, which can be
4537 EC_GROUP_set_curve_name()
4538 EC_GROUP_get_curve_name()
4539 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4541 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4542 was actually never needed) and in BN_mul(). The removal in BN_mul()
4543 required a small change in bn_mul_part_recursive() and the addition
4544 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4545 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4546 bn_sub_words() and bn_add_words() except they take arrays with
4550 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4552 *) Cleanse PEM buffers before freeing them since they may contain
4554 [Benjamin Bennett <ben@psc.edu>]
4556 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4557 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4558 authentication-only ciphersuites.
4561 *) Since AES128 and AES256 share a single mask bit in the logic of
4562 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4563 kludge to work properly if AES128 is available and AES256 isn't.
4566 *) Expand security boundary to match 1.1.1 module.
4569 *) Remove redundant features: hash file source, editing of test vectors
4570 modify fipsld to use external fips_premain.c signature.
4573 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4574 run algorithm test programs.
4577 *) Make algorithm test programs more tolerant of whitespace.
4580 *) Have SSL/TLS server implementation tolerate "mismatched" record
4581 protocol version while receiving ClientHello even if the
4582 ClientHello is fragmented. (The server can't insist on the
4583 particular protocol version it has chosen before the ServerHello
4584 message has informed the client about his choice.)
4587 *) Load error codes if they are not already present instead of using a
4588 static variable. This allows them to be cleanly unloaded and reloaded.
4591 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4593 *) Introduce limits to prevent malicious keys being able to
4594 cause a denial of service. (CVE-2006-2940)
4595 [Steve Henson, Bodo Moeller]
4597 *) Fix ASN.1 parsing of certain invalid structures that can result
4598 in a denial of service. (CVE-2006-2937) [Steve Henson]
4600 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4601 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4603 *) Fix SSL client code which could crash if connecting to a
4604 malicious SSLv2 server. (CVE-2006-4343)
4605 [Tavis Ormandy and Will Drewry, Google Security Team]
4607 *) Change ciphersuite string processing so that an explicit
4608 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4609 will no longer include "AES128-SHA"), and any other similar
4610 ciphersuite (same bitmap) from *other* protocol versions (so that
4611 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4612 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4613 changes from 0.9.8b and 0.9.8d.
4616 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4618 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4619 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4621 *) Change the Unix randomness entropy gathering to use poll() when
4622 possible instead of select(), since the latter has some
4623 undesirable limitations.
4624 [Darryl Miles via Richard Levitte and Bodo Moeller]
4626 *) Disable rogue ciphersuites:
4628 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4629 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4630 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4632 The latter two were purportedly from
4633 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4636 Also deactive the remaining ciphersuites from
4637 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4638 unofficial, and the ID has long expired.
4641 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4642 dual-core machines) and other potential thread-safety issues.
4645 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4647 *) Adapt fipsld and the build system to link against the validated FIPS
4648 module in FIPS mode.
4651 *) Fixes for VC++ 2005 build under Windows.
4654 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4655 from a Windows bash shell such as MSYS. It is autodetected from the
4656 "config" script when run from a VC++ environment. Modify standard VC++
4657 build to use fipscanister.o from the GNU make build.
4660 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4662 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4663 The value now differs depending on if you build for FIPS or not.
4664 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4665 safely run with a non-FIPSed libcrypto, as it may crash because of
4666 the difference induced by this change.
4669 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4671 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4672 (part of SSL_OP_ALL). This option used to disable the
4673 countermeasure against man-in-the-middle protocol-version
4674 rollback in the SSL 2.0 server implementation, which is a bad
4675 idea. (CVE-2005-2969)
4677 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4678 for Information Security, National Institute of Advanced Industrial
4679 Science and Technology [AIST], Japan)]
4681 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4682 mainly for FIPS compliance and not fully integrated at this stage.
4685 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4686 the exponentiation using a fixed-length exponent. (Otherwise,
4687 the information leaked through timing could expose the secret key
4688 after many signatures; cf. Bleichenbacher's attack on DSA with
4692 *) Make a new fixed-window mod_exp implementation the default for
4693 RSA, DSA, and DH private-key operations so that the sequence of
4694 squares and multiplies and the memory access pattern are
4695 independent of the particular secret key. This will mitigate
4696 cache-timing and potential related attacks.
4698 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4699 and this is automatically used by BN_mod_exp_mont() if the new flag
4700 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4701 will use this BN flag for private exponents unless the flag
4702 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4703 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4705 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4707 *) Change the client implementation for SSLv23_method() and
4708 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4709 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4710 (Previously, the SSL 2.0 backwards compatible Client Hello
4711 message format would be used even with SSL_OP_NO_SSLv2.)
4714 *) Add support for smime-type MIME parameter in S/MIME messages which some
4718 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4719 a threadsafe manner. Modify rsa code to use new function and add calls
4720 to dsa and dh code (which had race conditions before).
4723 *) Include the fixed error library code in the C error file definitions
4724 instead of fixing them up at runtime. This keeps the error code
4725 structures constant.
4728 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4730 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4733 *) Fixes for newer kerberos headers. NB: the casts are needed because
4734 the 'length' field is signed on one version and unsigned on another
4735 with no (?) obvious way to tell the difference, without these VC++
4736 complains. Also the "definition" of FAR (blank) is no longer included
4737 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4738 some needed definitions.
4741 *) Undo Cygwin change.
4744 *) Added support for proxy certificates according to RFC 3820.
4745 Because they may be a security thread to unaware applications,
4746 they must be explicitely allowed in run-time. See
4747 docs/HOWTO/proxy_certificates.txt for further information.
4750 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4752 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4753 server and client random values. Previously
4754 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4755 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4757 This change has negligible security impact because:
4759 1. Server and client random values still have 24 bytes of pseudo random
4762 2. Server and client random values are sent in the clear in the initial
4765 3. The master secret is derived using the premaster secret (48 bytes in
4766 size for static RSA ciphersuites) as well as client server and random
4769 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4772 [Stephen Henson, reported by UK NISCC]
4774 *) Use Windows randomness collection on Cygwin.
4777 *) Fix hang in EGD/PRNGD query when communication socket is closed
4778 prematurely by EGD/PRNGD.
4779 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4781 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4784 *) Back-port of selected performance improvements from development
4785 branch, as well as improved support for PowerPC platforms.
4788 *) Add lots of checks for memory allocation failure, error codes to indicate
4789 failure and freeing up memory if a failure occurs.
4790 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4792 *) Add new -passin argument to dgst.
4795 *) Perform some character comparisons of different types in X509_NAME_cmp:
4796 this is needed for some certificates that reencode DNs into UTF8Strings
4797 (in violation of RFC3280) and can't or wont issue name rollover
4801 *) Make an explicit check during certificate validation to see that
4802 the CA setting in each certificate on the chain is correct. As a
4803 side effect always do the following basic checks on extensions,
4804 not just when there's an associated purpose to the check:
4806 - if there is an unhandled critical extension (unless the user
4807 has chosen to ignore this fault)
4808 - if the path length has been exceeded (if one is set at all)
4809 - that certain extensions fit the associated purpose (if one has
4813 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4815 *) Avoid a race condition when CRLs are checked in a multi threaded
4816 environment. This would happen due to the reordering of the revoked
4817 entries during signature checking and serial number lookup. Now the
4818 encoding is cached and the serial number sort performed under a lock.
4819 Add new STACK function sk_is_sorted().
4822 *) Add Delta CRL to the extension code.
4825 *) Various fixes to s3_pkt.c so alerts are sent properly.
4826 [David Holmes <d.holmes@f5.com>]
4828 *) Reduce the chances of duplicate issuer name and serial numbers (in
4829 violation of RFC3280) using the OpenSSL certificate creation utilities.
4830 This is done by creating a random 64 bit value for the initial serial
4831 number when a serial number file is created or when a self signed
4832 certificate is created using 'openssl req -x509'. The initial serial
4833 number file is created using 'openssl x509 -next_serial' in CA.pl
4834 rather than being initialized to 1.
4837 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4839 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4840 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4841 [Joe Orton, Steve Henson]
4843 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4845 [Joe Orton, Steve Henson]
4847 *) Make it possible to have multiple active certificates with the same
4848 subject in the CA index file. This is done only if the keyword
4849 'unique_subject' is set to 'no' in the main CA section (default
4850 if 'CA_default') of the configuration file. The value is saved
4851 with the database itself in a separate index attribute file,
4852 named like the index file with '.attr' appended to the name.
4855 *) X509 verify fixes. Disable broken certificate workarounds when
4856 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4857 keyUsage extension present. Don't accept CRLs with unhandled critical
4858 extensions: since verify currently doesn't process CRL extensions this
4859 rejects a CRL with *any* critical extensions. Add new verify error codes
4863 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4864 A clarification of RFC2560 will require the use of OCTET STRINGs and
4865 some implementations cannot handle the current raw format. Since OpenSSL
4866 copies and compares OCSP nonces as opaque blobs without any attempt at
4867 parsing them this should not create any compatibility issues.
4870 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4871 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4872 this HMAC (and other) operations are several times slower than OpenSSL
4876 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4877 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4879 *) Use the correct content when signing type "other".
4882 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4884 *) Fix various bugs revealed by running the NISCC test suite:
4886 Stop out of bounds reads in the ASN1 code when presented with
4887 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4889 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4891 If verify callback ignores invalid public key errors don't try to check
4892 certificate signature with the NULL public key.
4896 *) New -ignore_err option in ocsp application to stop the server
4897 exiting on the first error in a request.
4900 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4901 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4905 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4906 extra data after the compression methods not only for TLS 1.0
4907 but also for SSL 3.0 (as required by the specification).
4908 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4910 *) Change X509_certificate_type() to mark the key as exported/exportable
4911 when it's 512 *bits* long, not 512 bytes.
4914 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4915 blocks during encryption.
4918 *) Various fixes to base64 BIO and non blocking I/O. On write
4919 flushes were not handled properly if the BIO retried. On read
4920 data was not being buffered properly and had various logic bugs.
4921 This also affects blocking I/O when the data being decoded is a
4925 *) Various S/MIME bugfixes and compatibility changes:
4926 output correct application/pkcs7 MIME type if
4927 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4928 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4929 of files as .eml work). Correctly handle very long lines in MIME
4933 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4935 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4936 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4937 a protocol version number mismatch like a decryption error
4938 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4941 *) Turn on RSA blinding by default in the default implementation
4942 to avoid a timing attack. Applications that don't want it can call
4943 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4944 They would be ill-advised to do so in most cases.
4945 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4947 *) Change RSA blinding code so that it works when the PRNG is not
4948 seeded (in this case, the secret RSA exponent is abused as
4949 an unpredictable seed -- if it is not unpredictable, there
4950 is no point in blinding anyway). Make RSA blinding thread-safe
4951 by remembering the creator's thread ID in rsa->blinding and
4952 having all other threads use local one-time blinding factors
4953 (this requires more computation than sharing rsa->blinding, but
4954 avoids excessive locking; and if an RSA object is not shared
4955 between threads, blinding will still be very fast).
4958 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4959 ENGINE as defaults for all supported algorithms irrespective of
4960 the 'flags' parameter. 'flags' is now honoured, so applications
4961 should make sure they are passing it correctly.
4964 *) Target "mingw" now allows native Windows code to be generated in
4965 the Cygwin environment as well as with the MinGW compiler.
4968 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4970 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4971 via timing by performing a MAC computation even if incorrect
4972 block cipher padding has been found. This is a countermeasure
4973 against active attacks where the attacker has to distinguish
4974 between bad padding and a MAC verification error. (CVE-2003-0078)
4976 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4977 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4978 Martin Vuagnoux (EPFL, Ilion)]
4980 *) Make the no-err option work as intended. The intention with no-err
4981 is not to have the whole error stack handling routines removed from
4982 libcrypto, it's only intended to remove all the function name and
4983 reason texts, thereby removing some of the footprint that may not
4984 be interesting if those errors aren't displayed anyway.
4986 NOTE: it's still possible for any application or module to have it's
4987 own set of error texts inserted. The routines are there, just not
4988 used by default when no-err is given.
4991 *) Add support for FreeBSD on IA64.
4992 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4994 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4995 Kerberos function mit_des_cbc_cksum(). Before this change,
4996 the value returned by DES_cbc_cksum() was like the one from
4997 mit_des_cbc_cksum(), except the bytes were swapped.
4998 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5000 *) Allow an application to disable the automatic SSL chain building.
5001 Before this a rather primitive chain build was always performed in
5002 ssl3_output_cert_chain(): an application had no way to send the
5003 correct chain if the automatic operation produced an incorrect result.
5005 Now the chain builder is disabled if either:
5007 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5009 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5011 The reasoning behind this is that an application would not want the
5012 auto chain building to take place if extra chain certificates are
5013 present and it might also want a means of sending no additional
5014 certificates (for example the chain has two certificates and the
5018 *) Add the possibility to build without the ENGINE framework.
5019 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5021 *) Under Win32 gmtime() can return NULL: check return value in
5022 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5025 *) DSA routines: under certain error conditions uninitialized BN objects
5026 could be freed. Solution: make sure initialization is performed early
5027 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5028 Nils Larsch <nla@trustcenter.de> via PR#459)
5031 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5032 checked on reconnect on the client side, therefore session resumption
5033 could still fail with a "ssl session id is different" error. This
5034 behaviour is masked when SSL_OP_ALL is used due to
5035 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5036 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5037 followup to PR #377.
5040 *) IA-32 assembler support enhancements: unified ELF targets, support
5041 for SCO/Caldera platforms, fix for Cygwin shared build.
5044 *) Add support for FreeBSD on sparc64. As a consequence, support for
5045 FreeBSD on non-x86 processors is separate from x86 processors on
5046 the config script, much like the NetBSD support.
5047 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5049 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5051 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5054 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5055 code (06) was taken as the first octet of the session ID and the last
5056 octet was ignored consequently. As a result SSLv2 client side session
5057 caching could not have worked due to the session ID mismatch between
5059 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5063 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5064 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5068 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5069 seems that in spite of existing for more than a year, many application
5070 author have done nothing to provide the necessary callbacks, which
5071 means that this particular engine will not work properly anywhere.
5072 This is a very unfortunate situation which forces us, in the name
5073 of usability, to give the hw_ncipher.c a static lock, which is part
5075 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5076 appear in 0.9.8 or later. We EXPECT application authors to have
5077 dealt properly with this when 0.9.8 is released (unless we actually
5078 make such changes in the libcrypto locking code that changes will
5079 have to be made anyway).
5082 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5083 octets have been read, EOF or an error occurs. Without this change
5084 some truncated ASN1 structures will not produce an error.
5087 *) Disable Heimdal support, since it hasn't been fully implemented.
5088 Still give the possibility to force the use of Heimdal, but with
5089 warnings and a request that patches get sent to openssl-dev.
5092 *) Add the VC-CE target, introduce the WINCE sysname, and add
5093 INSTALL.WCE and appropriate conditionals to make it build.
5094 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5096 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5097 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5098 edit numbers of the version.
5099 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5101 *) Introduce safe string copy and catenation functions
5102 (BUF_strlcpy() and BUF_strlcat()).
5103 [Ben Laurie (CHATS) and Richard Levitte]
5105 *) Avoid using fixed-size buffers for one-line DNs.
5106 [Ben Laurie (CHATS)]
5108 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5109 resizing buffers containing secrets, and use where appropriate.
5110 [Ben Laurie (CHATS)]
5112 *) Avoid using fixed size buffers for configuration file location.
5113 [Ben Laurie (CHATS)]
5115 *) Avoid filename truncation for various CA files.
5116 [Ben Laurie (CHATS)]
5118 *) Use sizeof in preference to magic numbers.
5119 [Ben Laurie (CHATS)]
5121 *) Avoid filename truncation in cert requests.
5122 [Ben Laurie (CHATS)]
5124 *) Add assertions to check for (supposedly impossible) buffer
5126 [Ben Laurie (CHATS)]
5128 *) Don't cache truncated DNS entries in the local cache (this could
5129 potentially lead to a spoofing attack).
5130 [Ben Laurie (CHATS)]
5132 *) Fix various buffers to be large enough for hex/decimal
5133 representations in a platform independent manner.
5134 [Ben Laurie (CHATS)]
5136 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5137 resizing buffers containing secrets, and use where appropriate.
5138 [Ben Laurie (CHATS)]
5140 *) Add BIO_indent() to avoid much slightly worrying code to do
5142 [Ben Laurie (CHATS)]
5144 *) Convert sprintf()/BIO_puts() to BIO_printf().
5145 [Ben Laurie (CHATS)]
5147 *) buffer_gets() could terminate with the buffer only half
5149 [Ben Laurie (CHATS)]
5151 *) Add assertions to prevent user-supplied crypto functions from
5152 overflowing internal buffers by having large block sizes, etc.
5153 [Ben Laurie (CHATS)]
5155 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5157 [Ben Laurie (CHATS)]
5159 *) Eliminate unused copy of key in RC4.
5160 [Ben Laurie (CHATS)]
5162 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5163 [Ben Laurie (CHATS)]
5165 *) Fix off-by-one error in EGD path.
5166 [Ben Laurie (CHATS)]
5168 *) If RANDFILE path is too long, ignore instead of truncating.
5169 [Ben Laurie (CHATS)]
5171 *) Eliminate unused and incorrectly sized X.509 structure
5173 [Ben Laurie (CHATS)]
5175 *) Eliminate unused and dangerous function knumber().
5176 [Ben Laurie (CHATS)]
5178 *) Eliminate unused and dangerous structure, KSSL_ERR.
5179 [Ben Laurie (CHATS)]
5181 *) Protect against overlong session ID context length in an encoded
5182 session object. Since these are local, this does not appear to be
5184 [Ben Laurie (CHATS)]
5186 *) Change from security patch (see 0.9.6e below) that did not affect
5187 the 0.9.6 release series:
5189 Remote buffer overflow in SSL3 protocol - an attacker could
5190 supply an oversized master key in Kerberos-enabled versions.
5192 [Ben Laurie (CHATS)]
5194 *) Change the SSL kerb5 codes to match RFC 2712.
5197 *) Make -nameopt work fully for req and add -reqopt switch.
5198 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5200 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5201 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5203 *) Make sure tests can be performed even if the corresponding algorithms
5204 have been removed entirely. This was also the last step to make
5205 OpenSSL compilable with DJGPP under all reasonable conditions.
5206 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5208 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5209 to allow version independent disabling of normally unselected ciphers,
5210 which may be activated as a side-effect of selecting a single cipher.
5212 (E.g., cipher list string "RSA" enables ciphersuites that are left
5213 out of "ALL" because they do not provide symmetric encryption.
5214 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5215 [Lutz Jaenicke, Bodo Moeller]
5217 *) Add appropriate support for separate platform-dependent build
5218 directories. The recommended way to make a platform-dependent
5219 build directory is the following (tested on Linux), maybe with
5222 # Place yourself outside of the OpenSSL source tree. In
5223 # this example, the environment variable OPENSSL_SOURCE
5224 # is assumed to contain the absolute OpenSSL source directory.
5225 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5226 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5227 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5228 mkdir -p `dirname $F`
5229 ln -s $OPENSSL_SOURCE/$F $F
5232 To be absolutely sure not to disturb the source tree, a "make clean"
5233 is a good thing. If it isn't successful, don't worry about it,
5234 it probably means the source directory is very clean.
5237 *) Make sure any ENGINE control commands make local copies of string
5238 pointers passed to them whenever necessary. Otherwise it is possible
5239 the caller may have overwritten (or deallocated) the original string
5240 data when a later ENGINE operation tries to use the stored values.
5241 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5243 *) Improve diagnostics in file reading and command-line digests.
5244 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5246 *) Add AES modes CFB and OFB to the object database. Correct an
5247 error in AES-CFB decryption.
5250 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5251 allows existing EVP_CIPHER_CTX structures to be reused after
5252 calling EVP_*Final(). This behaviour is used by encryption
5253 BIOs and some applications. This has the side effect that
5254 applications must explicitly clean up cipher contexts with
5255 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5258 *) Check the values of dna and dnb in bn_mul_recursive before calling
5259 bn_mul_comba (a non zero value means the a or b arrays do not contain
5260 n2 elements) and fallback to bn_mul_normal if either is not zero.
5263 *) Fix escaping of non-ASCII characters when using the -subj option
5264 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5267 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5268 form for "surname", serialNumber has no short form.
5269 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5270 therefore remove "mail" short name for "internet 7".
5271 The OID for unique identifiers in X509 certificates is
5272 x500UniqueIdentifier, not uniqueIdentifier.
5273 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5276 *) Add an "init" command to the ENGINE config module and auto initialize
5277 ENGINEs. Without any "init" command the ENGINE will be initialized
5278 after all ctrl commands have been executed on it. If init=1 the
5279 ENGINE is initailized at that point (ctrls before that point are run
5280 on the uninitialized ENGINE and after on the initialized one). If
5281 init=0 then the ENGINE will not be iniatialized at all.
5284 *) Fix the 'app_verify_callback' interface so that the user-defined
5285 argument is actually passed to the callback: In the
5286 SSL_CTX_set_cert_verify_callback() prototype, the callback
5287 declaration has been changed from
5290 int (*cb)(X509_STORE_CTX *,void *);
5291 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5292 i=s->ctx->app_verify_callback(&ctx)
5293 has been changed into
5294 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5296 To update applications using SSL_CTX_set_cert_verify_callback(),
5297 a dummy argument can be added to their callback functions.
5298 [D. K. Smetters <smetters@parc.xerox.com>]
5300 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5301 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5303 *) Add and OPENSSL_LOAD_CONF define which will cause
5304 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5305 This allows older applications to transparently support certain
5306 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5307 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5308 load the config file and OPENSSL_add_all_algorithms_conf() which will
5309 always load it have also been added.
5312 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5313 Adjust NIDs and EVP layer.
5314 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5316 *) Config modules support in openssl utility.
5318 Most commands now load modules from the config file,
5319 though in a few (such as version) this isn't done
5320 because it couldn't be used for anything.
5322 In the case of ca and req the config file used is
5323 the same as the utility itself: that is the -config
5324 command line option can be used to specify an
5328 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5329 use "openssl_conf" if filename is NULL use default openssl config file.
5332 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5333 config section name. Add a new flag to tolerate a missing config file
5334 and move code to CONF_modules_load_file().
5337 *) Support for crypto accelerator cards from Accelerated Encryption
5338 Processing, www.aep.ie. (Use engine 'aep')
5339 The support was copied from 0.9.6c [engine] and adapted/corrected
5340 to work with the new engine framework.
5341 [AEP Inc. and Richard Levitte]
5343 *) Support for SureWare crypto accelerator cards from Baltimore
5344 Technologies. (Use engine 'sureware')
5345 The support was copied from 0.9.6c [engine] and adapted
5346 to work with the new engine framework.
5349 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5350 make the newer ENGINE framework commands for the CHIL engine work.
5351 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5353 *) Make it possible to produce shared libraries on ReliantUNIX.
5354 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5356 *) Add the configuration target debug-linux-ppro.
5357 Make 'openssl rsa' use the general key loading routines
5358 implemented in apps.c, and make those routines able to
5359 handle the key format FORMAT_NETSCAPE and the variant
5361 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5363 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5364 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5366 *) Add -keyform to rsautl, and document -engine.
5367 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5369 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5370 BIO_R_NO_SUCH_FILE error code rather than the generic
5371 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5374 *) Add new functions
5376 ERR_peek_last_error_line
5377 ERR_peek_last_error_line_data.
5378 These are similar to
5381 ERR_peek_error_line_data,
5382 but report on the latest error recorded rather than the first one
5383 still in the error queue.
5384 [Ben Laurie, Bodo Moeller]
5386 *) default_algorithms option in ENGINE config module. This allows things
5388 default_algorithms = ALL
5389 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5392 *) Preliminary ENGINE config module.
5395 *) New experimental application configuration code.
5398 *) Change the AES code to follow the same name structure as all other
5399 symmetric ciphers, and behave the same way. Move everything to
5400 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5401 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5403 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5404 [Ben Laurie and Theo de Raadt]
5406 *) Add option to output public keys in req command.
5407 [Massimiliano Pala madwolf@openca.org]
5409 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5410 (up to about 10% better than before for P-192 and P-224).
5413 *) New functions/macros
5415 SSL_CTX_set_msg_callback(ctx, cb)
5416 SSL_CTX_set_msg_callback_arg(ctx, arg)
5417 SSL_set_msg_callback(ssl, cb)
5418 SSL_set_msg_callback_arg(ssl, arg)
5420 to request calling a callback function
5422 void cb(int write_p, int version, int content_type,
5423 const void *buf, size_t len, SSL *ssl, void *arg)
5425 whenever a protocol message has been completely received
5426 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5427 protocol version according to which the SSL library interprets
5428 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5429 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5430 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5431 specification (change_cipher_spec(20), alert(21), handshake(22)).
5432 'buf' and 'len' point to the actual message, 'ssl' to the
5433 SSL object, and 'arg' is the application-defined value set by
5434 SSL[_CTX]_set_msg_callback_arg().
5436 'openssl s_client' and 'openssl s_server' have new '-msg' options
5437 to enable a callback that displays all protocol messages.
5440 *) Change the shared library support so shared libraries are built as
5441 soon as the corresponding static library is finished, and thereby get
5442 openssl and the test programs linked against the shared library.
5443 This still only happens when the keyword "shard" has been given to
5444 the configuration scripts.
5446 NOTE: shared library support is still an experimental thing, and
5447 backward binary compatibility is still not guaranteed.
5448 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5450 *) Add support for Subject Information Access extension.
5451 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5453 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5454 additional bytes when new memory had to be allocated, not just
5455 when reusing an existing buffer.
5458 *) New command line and configuration option 'utf8' for the req command.
5459 This allows field values to be specified as UTF8 strings.
5462 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5463 runs for the former and machine-readable output for the latter.
5466 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5467 of the e-mail address in the DN (i.e., it will go into a certificate
5468 extension only). The new configuration file option 'email_in_dn = no'
5469 has the same effect.
5470 [Massimiliano Pala madwolf@openca.org]
5472 *) Change all functions with names starting with des_ to be starting
5473 with DES_ instead. Add wrappers that are compatible with libdes,
5474 but are named _ossl_old_des_*. Finally, add macros that map the
5475 des_* symbols to the corresponding _ossl_old_des_* if libdes
5476 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5477 desired, the des_* symbols will be mapped to DES_*, with one
5480 Since we provide two compatibility mappings, the user needs to
5481 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5482 compatibility is desired. The default (i.e., when that macro
5483 isn't defined) is OpenSSL 0.9.6c compatibility.
5485 There are also macros that enable and disable the support of old
5486 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5487 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5488 are defined, the default will apply: to support the old des routines.
5490 In either case, one must include openssl/des.h to get the correct
5491 definitions. Do not try to just include openssl/des_old.h, that
5494 NOTE: This is a major break of an old API into a new one. Software
5495 authors are encouraged to switch to the DES_ style functions. Some
5496 time in the future, des_old.h and the libdes compatibility functions
5497 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5498 default), and then completely removed.
5501 *) Test for certificates which contain unsupported critical extensions.
5502 If such a certificate is found during a verify operation it is
5503 rejected by default: this behaviour can be overridden by either
5504 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5505 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5506 X509_supported_extension() has also been added which returns 1 if a
5507 particular extension is supported.
5510 *) Modify the behaviour of EVP cipher functions in similar way to digests
5511 to retain compatibility with existing code.
5514 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5515 compatibility with existing code. In particular the 'ctx' parameter does
5516 not have to be to be initialized before the call to EVP_DigestInit() and
5517 it is tidied up after a call to EVP_DigestFinal(). New function
5518 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5519 EVP_MD_CTX_copy() changed to not require the destination to be
5520 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5521 requires the destination to be valid.
5523 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5524 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5527 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5528 so that complete 'Handshake' protocol structures are kept in memory
5529 instead of overwriting 'msg_type' and 'length' with 'body' data.
5532 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5533 [Massimo Santin via Richard Levitte]
5535 *) Major restructuring to the underlying ENGINE code. This includes
5536 reduction of linker bloat, separation of pure "ENGINE" manipulation
5537 (initialisation, etc) from functionality dealing with implementations
5538 of specific crypto iterfaces. This change also introduces integrated
5539 support for symmetric ciphers and digest implementations - so ENGINEs
5540 can now accelerate these by providing EVP_CIPHER and EVP_MD
5541 implementations of their own. This is detailed in crypto/engine/README
5542 as it couldn't be adequately described here. However, there are a few
5543 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5544 were changed in the original introduction of ENGINE code have now
5545 reverted back - the hooking from this code to ENGINE is now a good
5546 deal more passive and at run-time, operations deal directly with
5547 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5548 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5549 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5550 they were not being used by the framework as there is no concept of a
5551 BIGNUM_METHOD and they could not be generalised to the new
5552 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5553 ENGINE_cpy() has been removed as it cannot be consistently defined in
5557 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5560 *) Change mkdef.pl to sort symbols that get the same entry number,
5561 and make sure the automatically generated functions ERR_load_*
5562 become part of libeay.num as well.
5565 *) New function SSL_renegotiate_pending(). This returns true once
5566 renegotiation has been requested (either SSL_renegotiate() call
5567 or HelloRequest/ClientHello received from the peer) and becomes
5568 false once a handshake has been completed.
5569 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5570 sends a HelloRequest, but does not ensure that a handshake takes
5571 place. SSL_renegotiate_pending() is useful for checking if the
5572 client has followed the request.)
5575 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5576 By default, clients may request session resumption even during
5577 renegotiation (if session ID contexts permit); with this option,
5578 session resumption is possible only in the first handshake.
5580 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5581 more bits available for options that should not be part of
5582 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5585 *) Add some demos for certificate and certificate request creation.
5588 *) Make maximum certificate chain size accepted from the peer application
5589 settable (SSL*_get/set_max_cert_list()), as proposed by
5590 "Douglas E. Engert" <deengert@anl.gov>.
5593 *) Add support for shared libraries for Unixware-7
5594 (Boyd Lynn Gerber <gerberb@zenez.com>).
5597 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5598 be done prior to destruction. Use this to unload error strings from
5599 ENGINEs that load their own error strings. NB: This adds two new API
5600 functions to "get" and "set" this destroy handler in an ENGINE.
5603 *) Alter all existing ENGINE implementations (except "openssl" and
5604 "openbsd") to dynamically instantiate their own error strings. This
5605 makes them more flexible to be built both as statically-linked ENGINEs
5606 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5607 Also, add stub code to each that makes building them as self-contained
5608 shared-libraries easier (see README.ENGINE).
5611 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5612 implementations into applications that are completely implemented in
5613 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5614 commands that can be used to configure what shared-library to load and
5615 to control aspects of the way it is handled. Also, made an update to
5616 the README.ENGINE file that brings its information up-to-date and
5617 provides some information and instructions on the "dynamic" ENGINE
5618 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5621 *) Make it possible to unload ranges of ERR strings with a new
5622 "ERR_unload_strings" function.
5625 *) Add a copy() function to EVP_MD.
5628 *) Make EVP_MD routines take a context pointer instead of just the
5629 md_data void pointer.
5632 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5633 that the digest can only process a single chunk of data
5634 (typically because it is provided by a piece of
5635 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5636 is only going to provide a single chunk of data, and hence the
5637 framework needn't accumulate the data for oneshot drivers.
5640 *) As with "ERR", make it possible to replace the underlying "ex_data"
5641 functions. This change also alters the storage and management of global
5642 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5643 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5644 index counters. The API functions that use this state have been changed
5645 to take a "class_index" rather than pointers to the class's local STACK
5646 and counter, and there is now an API function to dynamically create new
5647 classes. This centralisation allows us to (a) plug a lot of the
5648 thread-safety problems that existed, and (b) makes it possible to clean
5649 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5650 such data would previously have always leaked in application code and
5651 workarounds were in place to make the memory debugging turn a blind eye
5652 to it. Application code that doesn't use this new function will still
5653 leak as before, but their memory debugging output will announce it now
5654 rather than letting it slide.
5656 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5657 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5658 has a return value to indicate success or failure.
5661 *) Make it possible to replace the underlying "ERR" functions such that the
5662 global state (2 LHASH tables and 2 locks) is only used by the "default"
5663 implementation. This change also adds two functions to "get" and "set"
5664 the implementation prior to it being automatically set the first time
5665 any other ERR function takes place. Ie. an application can call "get",
5666 pass the return value to a module it has just loaded, and that module
5667 can call its own "set" function using that value. This means the
5668 module's "ERR" operations will use (and modify) the error state in the
5669 application and not in its own statically linked copy of OpenSSL code.
5672 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5673 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5674 the operation, and provides a more encapsulated way for external code
5675 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5676 to use these functions rather than manually incrementing the counts.
5678 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5681 *) Add EVP test program.
5684 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5687 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5688 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5689 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5690 These allow a CRL to be built without having to access X509_CRL fields
5691 directly. Modify 'ca' application to use new functions.
5694 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5695 bug workarounds. Rollback attack detection is a security feature.
5696 The problem will only arise on OpenSSL servers when TLSv1 is not
5697 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5698 Software authors not wanting to support TLSv1 will have special reasons
5699 for their choice and can explicitly enable this option.
5700 [Bodo Moeller, Lutz Jaenicke]
5702 *) Rationalise EVP so it can be extended: don't include a union of
5703 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5704 (similar to those existing for EVP_CIPHER_CTX).
5709 EVP_MD_CTX_init(&md); /* new function call */
5710 EVP_DigestInit(&md, EVP_sha1());
5711 EVP_DigestUpdate(&md, in, len);
5712 EVP_DigestFinal(&md, out, NULL);
5713 EVP_MD_CTX_cleanup(&md); /* new function call */
5717 *) Make DES key schedule conform to the usual scheme, as well as
5718 correcting its structure. This means that calls to DES functions
5719 now have to pass a pointer to a des_key_schedule instead of a
5720 plain des_key_schedule (which was actually always a pointer
5723 des_key_schedule ks;
5725 des_set_key_checked(..., &ks);
5726 des_ncbc_encrypt(..., &ks, ...);
5728 (Note that a later change renames 'des_...' into 'DES_...'.)
5731 *) Initial reduction of linker bloat: the use of some functions, such as
5732 PEM causes large amounts of unused functions to be linked in due to
5733 poor organisation. For example pem_all.c contains every PEM function
5734 which has a knock on effect of linking in large amounts of (unused)
5735 ASN1 code. Grouping together similar functions and splitting unrelated
5736 functions prevents this.
5739 *) Cleanup of EVP macros.
5742 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5743 correct _ecb suffix.
5746 *) Add initial OCSP responder support to ocsp application. The
5747 revocation information is handled using the text based index
5748 use by the ca application. The responder can either handle
5749 requests generated internally, supplied in files (for example
5750 via a CGI script) or using an internal minimal server.
5753 *) Add configuration choices to get zlib compression for TLS.
5756 *) Changes to Kerberos SSL for RFC 2712 compliance:
5757 1. Implemented real KerberosWrapper, instead of just using
5758 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5759 2. Implemented optional authenticator field of KerberosWrapper.
5761 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5762 and authenticator structs; see crypto/krb5/.
5764 Generalized Kerberos calls to support multiple Kerberos libraries.
5765 [Vern Staats <staatsvr@asc.hpc.mil>,
5766 Jeffrey Altman <jaltman@columbia.edu>
5767 via Richard Levitte]
5769 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5770 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5771 values for each of the key sizes rather than having just
5772 parameters (and 'speed' generating keys each time).
5775 *) Speed up EVP routines.
5778 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5779 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5780 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5781 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5783 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5784 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5785 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5788 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5790 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5793 *) Added the OS2-EMX target.
5794 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5796 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5797 to support NCONF routines in extension code. New function CONF_set_nconf()
5798 to allow functions which take an NCONF to also handle the old LHASH
5799 structure: this means that the old CONF compatible routines can be
5800 retained (in particular wrt extensions) without having to duplicate the
5801 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5804 *) Enhance the general user interface with mechanisms for inner control
5805 and with possibilities to have yes/no kind of prompts.
5808 *) Change all calls to low level digest routines in the library and
5809 applications to use EVP. Add missing calls to HMAC_cleanup() and
5810 don't assume HMAC_CTX can be copied using memcpy().
5811 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5813 *) Add the possibility to control engines through control names but with
5814 arbitrary arguments instead of just a string.
5815 Change the key loaders to take a UI_METHOD instead of a callback
5816 function pointer. NOTE: this breaks binary compatibility with earlier
5817 versions of OpenSSL [engine].
5818 Adapt the nCipher code for these new conditions and add a card insertion
5822 *) Enhance the general user interface with mechanisms to better support
5823 dialog box interfaces, application-defined prompts, the possibility
5824 to use defaults (for example default passwords from somewhere else)
5825 and interrupts/cancellations.
5828 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5829 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5832 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5833 tidy up some unnecessarily weird code in 'sk_new()').
5834 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5836 *) Change the key loading routines for ENGINEs to use the same kind
5837 callback (pem_password_cb) as all other routines that need this
5841 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5842 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5843 than this minimum value is recommended.
5846 *) New random seeder for OpenVMS, using the system process statistics
5847 that are easily reachable.
5850 *) Windows apparently can't transparently handle global
5851 variables defined in DLLs. Initialisations such as:
5853 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5855 wont compile. This is used by the any applications that need to
5856 declare their own ASN1 modules. This was fixed by adding the option
5857 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5858 needed for static libraries under Win32.
5861 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5862 setting of purpose and trust fields. New X509_STORE trust and
5863 purpose functions and tidy up setting in other SSL functions.
5866 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5867 structure. These are inherited by X509_STORE_CTX when it is
5868 initialised. This allows various defaults to be set in the
5869 X509_STORE structure (such as flags for CRL checking and custom
5870 purpose or trust settings) for functions which only use X509_STORE_CTX
5871 internally such as S/MIME.
5873 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5874 trust settings if they are not set in X509_STORE. This allows X509_STORE
5875 purposes and trust (in S/MIME for example) to override any set by default.
5877 Add command line options for CRL checking to smime, s_client and s_server
5881 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5882 are set then the CRL is looked up in the X509_STORE structure and
5883 its validity and signature checked, then if the certificate is found
5884 in the CRL the verify fails with a revoked error.
5886 Various new CRL related callbacks added to X509_STORE_CTX structure.
5888 Command line options added to 'verify' application to support this.
5890 This needs some additional work, such as being able to handle multiple
5891 CRLs with different times, extension based lookup (rather than just
5892 by subject name) and ultimately more complete V2 CRL extension
5896 *) Add a general user interface API (crypto/ui/). This is designed
5897 to replace things like des_read_password and friends (backward
5898 compatibility functions using this new API are provided).
5899 The purpose is to remove prompting functions from the DES code
5900 section as well as provide for prompting through dialog boxes in
5901 a window system and the like.
5904 *) Add "ex_data" support to ENGINE so implementations can add state at a
5905 per-structure level rather than having to store it globally.
5908 *) Make it possible for ENGINE structures to be copied when retrieved by
5909 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5910 This causes the "original" ENGINE structure to act like a template,
5911 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5912 operational state can be localised to each ENGINE structure, despite the
5913 fact they all share the same "methods". New ENGINE structures returned in
5914 this case have no functional references and the return value is the single
5915 structural reference. This matches the single structural reference returned
5916 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5920 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5921 needs to match any other type at all we need to manually clear the
5925 *) Changes to the "openssl engine" utility to include;
5926 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5927 about an ENGINE's available control commands.
5928 - executing control commands from command line arguments using the
5929 '-pre' and '-post' switches. '-post' is only used if '-t' is
5930 specified and the ENGINE is successfully initialised. The syntax for
5931 the individual commands are colon-separated, for example;
5932 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5935 *) New dynamic control command support for ENGINEs. ENGINEs can now
5936 declare their own commands (numbers), names (strings), descriptions,
5937 and input types for run-time discovery by calling applications. A
5938 subset of these commands are implicitly classed as "executable"
5939 depending on their input type, and only these can be invoked through
5940 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5941 can be based on user input, config files, etc). The distinction is
5942 that "executable" commands cannot return anything other than a boolean
5943 result and can only support numeric or string input, whereas some
5944 discoverable commands may only be for direct use through
5945 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5946 pointers, or other custom uses. The "executable" commands are to
5947 support parameterisations of ENGINE behaviour that can be
5948 unambiguously defined by ENGINEs and used consistently across any
5949 OpenSSL-based application. Commands have been added to all the
5950 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5951 control over shared-library paths without source code alterations.
5954 *) Changed all ENGINE implementations to dynamically allocate their
5955 ENGINEs rather than declaring them statically. Apart from this being
5956 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5957 this also allows the implementations to compile without using the
5958 internal engine_int.h header.
5961 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5962 'const' value. Any code that should be able to modify a RAND_METHOD
5963 should already have non-const pointers to it (ie. they should only
5964 modify their own ones).
5967 *) Made a variety of little tweaks to the ENGINE code.
5968 - "atalla" and "ubsec" string definitions were moved from header files
5969 to C code. "nuron" string definitions were placed in variables
5970 rather than hard-coded - allowing parameterisation of these values
5971 later on via ctrl() commands.
5972 - Removed unused "#if 0"'d code.
5973 - Fixed engine list iteration code so it uses ENGINE_free() to release
5974 structural references.
5975 - Constified the RAND_METHOD element of ENGINE structures.
5976 - Constified various get/set functions as appropriate and added
5977 missing functions (including a catch-all ENGINE_cpy that duplicates
5978 all ENGINE values onto a new ENGINE except reference counts/state).
5979 - Removed NULL parameter checks in get/set functions. Setting a method
5980 or function to NULL is a way of cancelling out a previously set
5981 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5982 and doesn't justify the extra error symbols and code.
5983 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5984 flags from engine_int.h to engine.h.
5985 - Changed prototypes for ENGINE handler functions (init(), finish(),
5986 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5989 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5990 to the algorithm using long division. The binary algorithm can be
5991 used only if the modulus is odd. On 32-bit systems, it is faster
5992 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5993 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5994 up to 450 bits. In 64-bit environments, the binary algorithm
5995 appears to be advantageous for much longer moduli; here we use it
5996 for moduli up to 2048 bits.
5999 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6000 could not support the combine flag in choice fields.
6003 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6004 extensions from a certificate request to the certificate.
6007 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6008 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6009 file: this allows the display of the certificate about to be
6010 signed to be customised, to allow certain fields to be included
6011 or excluded and extension details. The old system didn't display
6012 multicharacter strings properly, omitted fields not in the policy
6013 and couldn't display additional details such as extensions.
6016 *) Function EC_POINTs_mul for multiple scalar multiplication
6017 of an arbitrary number of elliptic curve points
6018 \sum scalars[i]*points[i],
6019 optionally including the generator defined for the EC_GROUP:
6020 scalar*generator + \sum scalars[i]*points[i].
6022 EC_POINT_mul is a simple wrapper function for the typical case
6023 that the point list has just one item (besides the optional
6027 *) First EC_METHODs for curves over GF(p):
6029 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6030 operations and provides various method functions that can also
6031 operate with faster implementations of modular arithmetic.
6033 EC_GFp_mont_method() reuses most functions that are part of
6034 EC_GFp_simple_method, but uses Montgomery arithmetic.
6036 [Bodo Moeller; point addition and point doubling
6037 implementation directly derived from source code provided by
6038 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6040 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6041 crypto/ec/ec_lib.c):
6043 Curves are EC_GROUP objects (with an optional group generator)
6044 based on EC_METHODs that are built into the library.
6046 Points are EC_POINT objects based on EC_GROUP objects.
6048 Most of the framework would be able to handle curves over arbitrary
6049 finite fields, but as there are no obvious types for fields other
6050 than GF(p), some functions are limited to that for now.
6053 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6054 that the file contains a complete HTTP response.
6057 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6058 change the def and num file printf format specifier from "%-40sXXX"
6059 to "%-39s XXX". The latter will always guarantee a space after the
6060 field while the former will cause them to run together if the field
6061 is 40 of more characters long.
6064 *) Constify the cipher and digest 'method' functions and structures
6065 and modify related functions to take constant EVP_MD and EVP_CIPHER
6069 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6070 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6073 *) Modify EVP_Digest*() routines so they now return values. Although the
6074 internal software routines can never fail additional hardware versions
6078 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6080 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6081 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6084 ERR_R_NESTED_ASN1_ERROR
6086 ERR_R_MISSING_ASN1_EOS
6087 were 4 .. 9, conflicting with
6088 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6090 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6091 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6093 Add new error code 'ERR_R_INTERNAL_ERROR'.
6096 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6100 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6101 sets the subject name for a new request or supersedes the
6102 subject name in a given request. Formats that can be parsed are
6103 'CN=Some Name, OU=myOU, C=IT'
6105 'CN=Some Name/OU=myOU/C=IT'.
6107 Add options '-batch' and '-verbose' to 'openssl req'.
6108 [Massimiliano Pala <madwolf@hackmasters.net>]
6110 *) Introduce the possibility to access global variables through
6111 functions on platform were that's the best way to handle exporting
6112 global variables in shared libraries. To enable this functionality,
6113 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6114 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6115 is normally done by Configure or something similar).
6117 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6118 in the source file (foo.c) like this:
6120 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6121 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6123 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6124 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6126 OPENSSL_DECLARE_GLOBAL(int,foo);
6127 #define foo OPENSSL_GLOBAL_REF(foo)
6128 OPENSSL_DECLARE_GLOBAL(double,bar);
6129 #define bar OPENSSL_GLOBAL_REF(bar)
6131 The #defines are very important, and therefore so is including the
6132 header file everywhere where the defined globals are used.
6134 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6135 of ASN.1 items, but that structure is a bit different.
6137 The largest change is in util/mkdef.pl which has been enhanced with
6138 better and easier to understand logic to choose which symbols should
6139 go into the Windows .def files as well as a number of fixes and code
6140 cleanup (among others, algorithm keywords are now sorted
6141 lexicographically to avoid constant rewrites).
6144 *) In BN_div() keep a copy of the sign of 'num' before writing the
6145 result to 'rm' because if rm==num the value will be overwritten
6146 and produce the wrong result if 'num' is negative: this caused
6147 problems with BN_mod() and BN_nnmod().
6150 *) Function OCSP_request_verify(). This checks the signature on an
6151 OCSP request and verifies the signer certificate. The signer
6152 certificate is just checked for a generic purpose and OCSP request
6156 *) Add OCSP_check_validity() function to check the validity of OCSP
6157 responses. OCSP responses are prepared in real time and may only
6158 be a few seconds old. Simply checking that the current time lies
6159 between thisUpdate and nextUpdate max reject otherwise valid responses
6160 caused by either OCSP responder or client clock inaccuracy. Instead
6161 we allow thisUpdate and nextUpdate to fall within a certain period of
6162 the current time. The age of the response can also optionally be
6163 checked. Two new options -validity_period and -status_age added to
6167 *) If signature or public key algorithm is unrecognized print out its
6168 OID rather that just UNKNOWN.
6171 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6172 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6173 ID to be generated from the issuer certificate alone which can then be
6174 passed to OCSP_id_issuer_cmp().
6177 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6178 ASN1 modules to export functions returning ASN1_ITEM pointers
6179 instead of the ASN1_ITEM structures themselves. This adds several
6180 new macros which allow the underlying ASN1 function/structure to
6181 be accessed transparently. As a result code should not use ASN1_ITEM
6182 references directly (such as &X509_it) but instead use the relevant
6183 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6184 use of the new ASN1 code on platforms where exporting structures
6185 is problematical (for example in shared libraries) but exporting
6186 functions returning pointers to structures is not.
6189 *) Add support for overriding the generation of SSL/TLS session IDs.
6190 These callbacks can be registered either in an SSL_CTX or per SSL.
6191 The purpose of this is to allow applications to control, if they wish,
6192 the arbitrary values chosen for use as session IDs, particularly as it
6193 can be useful for session caching in multiple-server environments. A
6194 command-line switch for testing this (and any client code that wishes
6195 to use such a feature) has been added to "s_server".
6196 [Geoff Thorpe, Lutz Jaenicke]
6198 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6199 of the form '#if defined(...) || defined(...) || ...' and
6200 '#if !defined(...) && !defined(...) && ...'. This also avoids
6201 the growing number of special cases it was previously handling.
6204 *) Make all configuration macros available for application by making
6205 sure they are available in opensslconf.h, by giving them names starting
6206 with "OPENSSL_" to avoid conflicts with other packages and by making
6207 sure e_os2.h will cover all platform-specific cases together with
6209 Additionally, it is now possible to define configuration/platform-
6210 specific names (called "system identities"). In the C code, these
6211 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6212 macro with the name beginning with "OPENSSL_SYS_", which is determined
6213 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6217 *) New option -set_serial to 'req' and 'x509' this allows the serial
6218 number to use to be specified on the command line. Previously self
6219 signed certificates were hard coded with serial number 0 and the
6220 CA options of 'x509' had to use a serial number in a file which was
6224 *) New options to 'ca' utility to support V2 CRL entry extensions.
6225 Currently CRL reason, invalidity date and hold instruction are
6226 supported. Add new CRL extensions to V3 code and some new objects.
6229 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6230 disable standard block padding (aka PKCS#5 padding) in the EVP
6231 API, which was previously mandatory. This means that the data is
6232 not padded in any way and so the total length much be a multiple
6233 of the block size, otherwise an error occurs.
6236 *) Initial (incomplete) OCSP SSL support.
6239 *) New function OCSP_parse_url(). This splits up a URL into its host,
6240 port and path components: primarily to parse OCSP URLs. New -url
6241 option to ocsp utility.
6244 *) New nonce behavior. The return value of OCSP_check_nonce() now
6245 reflects the various checks performed. Applications can decide
6246 whether to tolerate certain situations such as an absent nonce
6247 in a response when one was present in a request: the ocsp application
6248 just prints out a warning. New function OCSP_add1_basic_nonce()
6249 this is to allow responders to include a nonce in a response even if
6250 the request is nonce-less.
6253 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6254 skipped when using openssl x509 multiple times on a single input file,
6255 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6258 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6259 set string type: to handle setting ASN1_TIME structures. Fix ca
6260 utility to correctly initialize revocation date of CRLs.
6263 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6264 the clients preferred ciphersuites and rather use its own preferences.
6265 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6266 Internet Explorer by ensuring unchanged hash method during stepup.
6267 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6270 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6271 to aes and add a new 'exist' option to print out symbols that don't
6275 *) Additional options to ocsp utility to allow flags to be set and
6276 additional certificates supplied.
6279 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6280 OCSP client a number of certificate to only verify the response
6284 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6285 handle the new API. Currently only ECB, CBC modes supported. Add new
6288 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6289 Encryption Standard (AES) Ciphersuites for Transport Layer
6290 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6291 not enabled by default and were not part of the "ALL" ciphersuite
6292 alias because they were not yet official; they could be
6293 explicitly requested by specifying the "AESdraft" ciphersuite
6294 group alias. In the final release of OpenSSL 0.9.7, the group
6295 alias is called "AES" and is part of "ALL".)
6296 [Ben Laurie, Steve Henson, Bodo Moeller]
6298 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6299 request to response.
6302 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6303 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6304 extract information from a certificate request. OCSP_response_create()
6305 creates a response and optionally adds a basic response structure.
6306 OCSP_basic_add1_status() adds a complete single response to a basic
6307 response and returns the OCSP_SINGLERESP structure just added (to allow
6308 extensions to be included for example). OCSP_basic_add1_cert() adds a
6309 certificate to a basic response and OCSP_basic_sign() signs a basic
6310 response with various flags. New helper functions ASN1_TIME_check()
6311 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6312 (converts ASN1_TIME to GeneralizedTime).
6315 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6316 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6317 structure from a certificate. X509_pubkey_digest() digests the public_key
6318 contents: this is used in various key identifiers.
6321 *) Make sk_sort() tolerate a NULL argument.
6322 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6324 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6325 passed by the function are trusted implicitly. If any of them signed the
6326 response then it is assumed to be valid and is not verified.
6329 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6330 to data. This was previously part of the PKCS7 ASN1 code. This
6331 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6332 [Steve Henson, reported by Kenneth R. Robinette
6333 <support@securenetterm.com>]
6335 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6336 routines: without these tracing memory leaks is very painful.
6337 Fix leaks in PKCS12 and PKCS7 routines.
6340 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6341 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6342 effectively meant GeneralizedTime would never be used. Now it
6343 is initialised to -1 but X509_time_adj() now has to check the value
6344 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6345 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6346 [Steve Henson, reported by Kenneth R. Robinette
6347 <support@securenetterm.com>]
6349 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6350 result in a zero length in the ASN1_INTEGER structure which was
6351 not consistent with the structure when d2i_ASN1_INTEGER() was used
6352 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6353 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6354 where it did not print out a minus for negative ASN1_INTEGER.
6357 *) Add summary printout to ocsp utility. The various functions which
6358 convert status values to strings have been renamed to:
6359 OCSP_response_status_str(), OCSP_cert_status_str() and
6360 OCSP_crl_reason_str() and are no longer static. New options
6361 to verify nonce values and to disable verification. OCSP response
6362 printout format cleaned up.
6365 *) Add additional OCSP certificate checks. These are those specified
6366 in RFC2560. This consists of two separate checks: the CA of the
6367 certificate being checked must either be the OCSP signer certificate
6368 or the issuer of the OCSP signer certificate. In the latter case the
6369 OCSP signer certificate must contain the OCSP signing extended key
6370 usage. This check is performed by attempting to match the OCSP
6371 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6372 in the OCSP_CERTID structures of the response.
6375 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6376 and related routines. This uses the standard OpenSSL certificate
6377 verify routines to perform initial checks (just CA validity) and
6378 to obtain the certificate chain. Then additional checks will be
6379 performed on the chain. Currently the root CA is checked to see
6380 if it is explicitly trusted for OCSP signing. This is used to set
6381 a root CA as a global signing root: that is any certificate that
6382 chains to that CA is an acceptable OCSP signing certificate.
6385 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6386 extensions from a separate configuration file.
6387 As when reading extensions from the main configuration file,
6388 the '-extensions ...' option may be used for specifying the
6390 [Massimiliano Pala <madwolf@comune.modena.it>]
6392 *) New OCSP utility. Allows OCSP requests to be generated or
6393 read. The request can be sent to a responder and the output
6394 parsed, outputed or printed in text form. Not complete yet:
6395 still needs to check the OCSP response validity.
6398 *) New subcommands for 'openssl ca':
6399 'openssl ca -status <serial>' prints the status of the cert with
6400 the given serial number (according to the index file).
6401 'openssl ca -updatedb' updates the expiry status of certificates
6403 [Massimiliano Pala <madwolf@comune.modena.it>]
6405 *) New '-newreq-nodes' command option to CA.pl. This is like
6406 '-newreq', but calls 'openssl req' with the '-nodes' option
6407 so that the resulting key is not encrypted.
6408 [Damien Miller <djm@mindrot.org>]
6410 *) New configuration for the GNU Hurd.
6411 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6413 *) Initial code to implement OCSP basic response verify. This
6414 is currently incomplete. Currently just finds the signer's
6415 certificate and verifies the signature on the response.
6418 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6419 value of OPENSSLDIR. This is available via the new '-d' option
6420 to 'openssl version', and is also included in 'openssl version -a'.
6423 *) Allowing defining memory allocation callbacks that will be given
6424 file name and line number information in additional arguments
6425 (a const char* and an int). The basic functionality remains, as
6426 well as the original possibility to just replace malloc(),
6427 realloc() and free() by functions that do not know about these
6428 additional arguments. To register and find out the current
6429 settings for extended allocation functions, the following
6430 functions are provided:
6432 CRYPTO_set_mem_ex_functions
6433 CRYPTO_set_locked_mem_ex_functions
6434 CRYPTO_get_mem_ex_functions
6435 CRYPTO_get_locked_mem_ex_functions
6437 These work the same way as CRYPTO_set_mem_functions and friends.
6438 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6439 extended allocation function is enabled.
6440 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6441 a conventional allocation function is enabled.
6442 [Richard Levitte, Bodo Moeller]
6444 *) Finish off removing the remaining LHASH function pointer casts.
6445 There should no longer be any prototype-casting required when using
6446 the LHASH abstraction, and any casts that remain are "bugs". See
6447 the callback types and macros at the head of lhash.h for details
6448 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6451 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6452 If /dev/[u]random devices are not available or do not return enough
6453 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6455 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6456 /etc/entropy will be queried once each in this sequence, quering stops
6457 when enough entropy was collected without querying more sockets.
6460 *) Change the Unix RAND_poll() variant to be able to poll several
6461 random devices, as specified by DEVRANDOM, until a sufficient amount
6462 of data has been collected. We spend at most 10 ms on each file
6463 (select timeout) and read in non-blocking mode. DEVRANDOM now
6464 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6465 (previously it was just the string "/dev/urandom"), so on typical
6466 platforms the 10 ms delay will never occur.
6467 Also separate out the Unix variant to its own file, rand_unix.c.
6468 For VMS, there's a currently-empty rand_vms.c.
6471 *) Move OCSP client related routines to ocsp_cl.c. These
6472 provide utility functions which an application needing
6473 to issue a request to an OCSP responder and analyse the
6474 response will typically need: as opposed to those which an
6475 OCSP responder itself would need which will be added later.
6477 OCSP_request_sign() signs an OCSP request with an API similar
6478 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6479 response. OCSP_response_get1_basic() extracts basic response
6480 from response. OCSP_resp_find_status(): finds and extracts status
6481 information from an OCSP_CERTID structure (which will be created
6482 when the request structure is built). These are built from lower
6483 level functions which work on OCSP_SINGLERESP structures but
6484 wont normally be used unless the application wishes to examine
6485 extensions in the OCSP response for example.
6487 Replace nonce routines with a pair of functions.
6488 OCSP_request_add1_nonce() adds a nonce value and optionally
6489 generates a random value. OCSP_check_nonce() checks the
6490 validity of the nonce in an OCSP response.
6493 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6494 This doesn't copy the supplied OCSP_CERTID and avoids the
6495 need to free up the newly created id. Change return type
6496 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6497 This can then be used to add extensions to the request.
6498 Deleted OCSP_request_new(), since most of its functionality
6499 is now in OCSP_REQUEST_new() (and the case insensitive name
6500 clash) apart from the ability to set the request name which
6501 will be added elsewhere.
6504 *) Update OCSP API. Remove obsolete extensions argument from
6505 various functions. Extensions are now handled using the new
6506 OCSP extension code. New simple OCSP HTTP function which
6507 can be used to send requests and parse the response.
6510 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6511 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6512 uses the special reorder version of SET OF to sort the attributes
6513 and reorder them to match the encoded order. This resolves a long
6514 standing problem: a verify on a PKCS7 structure just after signing
6515 it used to fail because the attribute order did not match the
6516 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6517 it uses the received order. This is necessary to tolerate some broken
6518 software that does not order SET OF. This is handled by encoding
6519 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6520 to produce the required SET OF.
6523 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6524 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6525 files to get correct declarations of the ASN.1 item variables.
6528 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6529 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6530 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6531 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6532 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6533 ASN1_ITEM and no wrapper functions.
6536 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6537 replace the old function pointer based I/O routines. Change most of
6538 the *_d2i_bio() and *_d2i_fp() functions to use these.
6541 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6542 lines, recognice more "algorithms" that can be deselected, and make
6543 it complain about algorithm deselection that isn't recognised.
6546 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6547 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6548 to use new functions. Add NO_ASN1_OLD which can be set to remove
6549 some old style ASN1 functions: this can be used to determine if old
6550 code will still work when these eventually go away.
6553 *) New extension functions for OCSP structures, these follow the
6554 same conventions as certificates and CRLs.
6557 *) New function X509V3_add1_i2d(). This automatically encodes and
6558 adds an extension. Its behaviour can be customised with various
6559 flags to append, replace or delete. Various wrappers added for
6560 certifcates and CRLs.
6563 *) Fix to avoid calling the underlying ASN1 print routine when
6564 an extension cannot be parsed. Correct a typo in the
6565 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6568 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6569 entries for variables.
6572 *) Add functionality to apps/openssl.c for detecting locking
6573 problems: As the program is single-threaded, all we have
6574 to do is register a locking callback using an array for
6575 storing which locks are currently held by the program.
6578 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6579 SSL_get_ex_data_X509_STORE_idx(), which is used in
6580 ssl_verify_cert_chain() and thus can be called at any time
6581 during TLS/SSL handshakes so that thread-safety is essential.
6582 Unfortunately, the ex_data design is not at all suited
6583 for multi-threaded use, so it probably should be abolished.
6586 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6587 [Broadcom, tweaked and integrated by Geoff Thorpe]
6589 *) Move common extension printing code to new function
6590 X509V3_print_extensions(). Reorganise OCSP print routines and
6591 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6594 *) New function X509_signature_print() to remove duplication in some
6598 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6599 set (this was treated exactly the same as SET OF previously). This
6600 is used to reorder the STACK representing the structure to match the
6601 encoding. This will be used to get round a problem where a PKCS7
6602 structure which was signed could not be verified because the STACK
6603 order did not reflect the encoded order.
6606 *) Reimplement the OCSP ASN1 module using the new code.
6609 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6610 for its ASN1 operations. The old style function pointers still exist
6611 for now but they will eventually go away.
6614 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6615 completely replaces the old ASN1 functionality with a table driven
6616 encoder and decoder which interprets an ASN1_ITEM structure describing
6617 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6618 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6619 has also been converted to the new form.
6622 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6623 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6624 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6625 for negative moduli.
6628 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6629 of not touching the result's sign bit.
6632 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6636 *) Changed the LHASH code to use prototypes for callbacks, and created
6637 macros to declare and implement thin (optionally static) functions
6638 that provide type-safety and avoid function pointer casting for the
6639 type-specific callbacks.
6642 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6644 [Veers Staats <staatsvr@asc.hpc.mil>,
6645 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6647 *) Reformat the FAQ so the different questions and answers can be divided
6648 in sections depending on the subject.
6651 *) Have the zlib compression code load ZLIB.DLL dynamically under
6655 *) New function BN_mod_sqrt for computing square roots modulo a prime
6656 (using the probabilistic Tonelli-Shanks algorithm unless
6657 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6658 be handled deterministically).
6659 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6661 *) Make BN_mod_inverse faster by explicitly handling small quotients
6662 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6663 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6666 *) New function BN_kronecker.
6669 *) Fix BN_gcd so that it works on negative inputs; the result is
6670 positive unless both parameters are zero.
6671 Previously something reasonably close to an infinite loop was
6672 possible because numbers could be growing instead of shrinking
6673 in the implementation of Euclid's algorithm.
6676 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6677 sign of the number in question.
6679 Fix BN_is_word(a,w) to work correctly for w == 0.
6681 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6682 because its test if the absolute value of 'a' equals 'w'.
6683 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6684 it exists mostly for use in the implementations of BN_is_zero(),
6685 BN_is_one(), and BN_is_word().
6688 *) New function BN_swap.
6691 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6692 the exponentiation functions are more likely to produce reasonable
6693 results on negative inputs.
6696 *) Change BN_mod_mul so that the result is always non-negative.
6697 Previously, it could be negative if one of the factors was negative;
6698 I don't think anyone really wanted that behaviour.
6701 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6702 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6703 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6704 and add new functions:
6713 BN_mod_lshift1_quick
6717 These functions always generate non-negative results.
6719 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6720 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6722 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6723 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6724 be reduced modulo m.
6725 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6728 The following entry accidentally appeared in the CHANGES file
6729 distributed with OpenSSL 0.9.7. The modifications described in
6730 it do *not* apply to OpenSSL 0.9.7.
6732 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6733 was actually never needed) and in BN_mul(). The removal in BN_mul()
6734 required a small change in bn_mul_part_recursive() and the addition
6735 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6736 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6737 bn_sub_words() and bn_add_words() except they take arrays with
6742 *) In 'openssl passwd', verify passwords read from the terminal
6743 unless the '-salt' option is used (which usually means that
6744 verification would just waste user's time since the resulting
6745 hash is going to be compared with some given password hash)
6746 or the new '-noverify' option is used.
6748 This is an incompatible change, but it does not affect
6749 non-interactive use of 'openssl passwd' (passwords on the command
6750 line, '-stdin' option, '-in ...' option) and thus should not
6754 *) Remove all references to RSAref, since there's no more need for it.
6757 *) Make DSO load along a path given through an environment variable
6758 (SHLIB_PATH) with shl_load().
6761 *) Constify the ENGINE code as a result of BIGNUM constification.
6762 Also constify the RSA code and most things related to it. In a
6763 few places, most notable in the depth of the ASN.1 code, ugly
6764 casts back to non-const were required (to be solved at a later
6768 *) Make it so the openssl application has all engines loaded by default.
6771 *) Constify the BIGNUM routines a little more.
6774 *) Add the following functions:
6776 ENGINE_load_cswift()
6778 ENGINE_load_atalla()
6780 ENGINE_load_builtin_engines()
6782 That way, an application can itself choose if external engines that
6783 are built-in in OpenSSL shall ever be used or not. The benefit is
6784 that applications won't have to be linked with libdl or other dso
6785 libraries unless it's really needed.
6787 Changed 'openssl engine' to load all engines on demand.
6788 Changed the engine header files to avoid the duplication of some
6789 declarations (they differed!).
6792 *) 'openssl engine' can now list capabilities.
6795 *) Better error reporting in 'openssl engine'.
6798 *) Never call load_dh_param(NULL) in s_server.
6801 *) Add engine application. It can currently list engines by name and
6802 identity, and test if they are actually available.
6805 *) Improve RPM specification file by forcing symbolic linking and making
6806 sure the installed documentation is also owned by root.root.
6807 [Damien Miller <djm@mindrot.org>]
6809 *) Give the OpenSSL applications more possibilities to make use of
6810 keys (public as well as private) handled by engines.
6813 *) Add OCSP code that comes from CertCo.
6816 *) Add VMS support for the Rijndael code.
6819 *) Added untested support for Nuron crypto accelerator.
6822 *) Add support for external cryptographic devices. This code was
6823 previously distributed separately as the "engine" branch.
6824 [Geoff Thorpe, Richard Levitte]
6826 *) Rework the filename-translation in the DSO code. It is now possible to
6827 have far greater control over how a "name" is turned into a filename
6828 depending on the operating environment and any oddities about the
6829 different shared library filenames on each system.
6832 *) Support threads on FreeBSD-elf in Configure.
6835 *) Fix for SHA1 assembly problem with MASM: it produces
6836 warnings about corrupt line number information when assembling
6837 with debugging information. This is caused by the overlapping
6839 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6842 NCONF_get_number() has no error checking at all. As a replacement,
6843 NCONF_get_number_e() is defined (_e for "error checking") and is
6844 promoted strongly. The old NCONF_get_number is kept around for
6845 binary backward compatibility.
6846 Make it possible for methods to load from something other than a BIO,
6847 by providing a function pointer that is given a name instead of a BIO.
6848 For example, this could be used to load configuration data from an
6852 *) Fix for non blocking accept BIOs. Added new I/O special reason
6853 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6854 with non blocking I/O was not possible because no retry code was
6855 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6859 *) Added the beginnings of Rijndael support.
6862 *) Fix for bug in DirectoryString mask setting. Add support for
6863 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6864 to allow certificate printing to more controllable, additional
6865 'certopt' option to 'x509' to allow new printing options to be
6869 *) Clean old EAY MD5 hack from e_os.h.
6872 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6874 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6875 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6876 [Joe Orton, Steve Henson]
6878 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6880 *) Fix additional bug revealed by the NISCC test suite:
6882 Stop bug triggering large recursion when presented with
6883 certain ASN.1 tags (CVE-2003-0851)
6886 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6888 *) Fix various bugs revealed by running the NISCC test suite:
6890 Stop out of bounds reads in the ASN1 code when presented with
6891 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6893 If verify callback ignores invalid public key errors don't try to check
6894 certificate signature with the NULL public key.
6898 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6899 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6903 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6904 extra data after the compression methods not only for TLS 1.0
6905 but also for SSL 3.0 (as required by the specification).
6906 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6908 *) Change X509_certificate_type() to mark the key as exported/exportable
6909 when it's 512 *bits* long, not 512 bytes.
6912 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6914 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6915 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6916 a protocol version number mismatch like a decryption error
6917 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6920 *) Turn on RSA blinding by default in the default implementation
6921 to avoid a timing attack. Applications that don't want it can call
6922 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6923 They would be ill-advised to do so in most cases.
6924 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6926 *) Change RSA blinding code so that it works when the PRNG is not
6927 seeded (in this case, the secret RSA exponent is abused as
6928 an unpredictable seed -- if it is not unpredictable, there
6929 is no point in blinding anyway). Make RSA blinding thread-safe
6930 by remembering the creator's thread ID in rsa->blinding and
6931 having all other threads use local one-time blinding factors
6932 (this requires more computation than sharing rsa->blinding, but
6933 avoids excessive locking; and if an RSA object is not shared
6934 between threads, blinding will still be very fast).
6937 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6939 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6940 via timing by performing a MAC computation even if incorrrect
6941 block cipher padding has been found. This is a countermeasure
6942 against active attacks where the attacker has to distinguish
6943 between bad padding and a MAC verification error. (CVE-2003-0078)
6945 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6946 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6947 Martin Vuagnoux (EPFL, Ilion)]
6949 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6951 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6952 memory from it's contents. This is done with a counter that will
6953 place alternating values in each byte. This can be used to solve
6954 two issues: 1) the removal of calls to memset() by highly optimizing
6955 compilers, and 2) cleansing with other values than 0, since those can
6956 be read through on certain media, for example a swap space on disk.
6959 *) Bugfix: client side session caching did not work with external caching,
6960 because the session->cipher setting was not restored when reloading
6961 from the external cache. This problem was masked, when
6962 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6963 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6966 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6967 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6968 [Zeev Lieber <zeev-l@yahoo.com>]
6970 *) Undo an undocumented change introduced in 0.9.6e which caused
6971 repeated calls to OpenSSL_add_all_ciphers() and
6972 OpenSSL_add_all_digests() to be ignored, even after calling
6976 *) Change the default configuration reader to deal with last line not
6977 being properly terminated.
6980 *) Change X509_NAME_cmp() so it applies the special rules on handling
6981 DN values that are of type PrintableString, as well as RDNs of type
6982 emailAddress where the value has the type ia5String.
6983 [stefank@valicert.com via Richard Levitte]
6985 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6986 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6987 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6988 the bitwise-OR of the two for use by the majority of applications
6989 wanting this behaviour, and update the docs. The documented
6990 behaviour and actual behaviour were inconsistent and had been
6991 changing anyway, so this is more a bug-fix than a behavioural
6993 [Geoff Thorpe, diagnosed by Nadav Har'El]
6995 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6996 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6999 *) Fix initialization code race conditions in
7000 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7001 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7002 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7003 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7004 ssl2_get_cipher_by_char(),
7005 ssl3_get_cipher_by_char().
7006 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7008 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7009 the cached sessions are flushed, as the remove_cb() might use ex_data
7010 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7011 (see [openssl.org #212]).
7012 [Geoff Thorpe, Lutz Jaenicke]
7014 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7015 length, instead of the encoding length to d2i_ASN1_OBJECT.
7018 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7020 *) [In 0.9.6g-engine release:]
7021 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7022 [Lynn Gazis <lgazis@rainbow.com>]
7024 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7026 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7027 and get fix the header length calculation.
7028 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7029 Alon Kantor <alonk@checkpoint.com> (and others),
7032 *) Use proper error handling instead of 'assertions' in buffer
7033 overflow checks added in 0.9.6e. This prevents DoS (the
7034 assertions could call abort()).
7035 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7037 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7039 *) Add various sanity checks to asn1_get_length() to reject
7040 the ASN1 length bytes if they exceed sizeof(long), will appear
7041 negative or the content length exceeds the length of the
7043 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7045 *) Fix cipher selection routines: ciphers without encryption had no flags
7046 for the cipher strength set and where therefore not handled correctly
7047 by the selection routines (PR #130).
7050 *) Fix EVP_dsa_sha macro.
7054 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7055 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7056 that was added in OpenSSL 0.9.6d.
7058 As the countermeasure turned out to be incompatible with some
7059 broken SSL implementations, the new option is part of SSL_OP_ALL.
7060 SSL_OP_ALL is usually employed when compatibility with weird SSL
7061 implementations is desired (e.g. '-bugs' option to 's_client' and
7062 's_server'), so the new option is automatically set in many
7066 *) Changes in security patch:
7068 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7069 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7070 Air Force Materiel Command, USAF, under agreement number
7073 *) Add various sanity checks to asn1_get_length() to reject
7074 the ASN1 length bytes if they exceed sizeof(long), will appear
7075 negative or the content length exceeds the length of the
7076 supplied buffer. (CVE-2002-0659)
7077 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7079 *) Assertions for various potential buffer overflows, not known to
7081 [Ben Laurie (CHATS)]
7083 *) Various temporary buffers to hold ASCII versions of integers were
7084 too small for 64 bit platforms. (CVE-2002-0655)
7085 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7087 *) Remote buffer overflow in SSL3 protocol - an attacker could
7088 supply an oversized session ID to a client. (CVE-2002-0656)
7089 [Ben Laurie (CHATS)]
7091 *) Remote buffer overflow in SSL2 protocol - an attacker could
7092 supply an oversized client master key. (CVE-2002-0656)
7093 [Ben Laurie (CHATS)]
7095 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7097 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7098 encoded as NULL) with id-dsa-with-sha1.
7099 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7101 *) Check various X509_...() return values in apps/req.c.
7102 [Nils Larsch <nla@trustcenter.de>]
7104 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7105 an end-of-file condition would erronously be flagged, when the CRLF
7106 was just at the end of a processed block. The bug was discovered when
7107 processing data through a buffering memory BIO handing the data to a
7108 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7109 <ptsekov@syntrex.com> and Nedelcho Stanev.
7112 *) Implement a countermeasure against a vulnerability recently found
7113 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7114 before application data chunks to avoid the use of known IVs
7115 with data potentially chosen by the attacker.
7118 *) Fix length checks in ssl3_get_client_hello().
7121 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7122 to prevent ssl3_read_internal() from incorrectly assuming that
7123 ssl3_read_bytes() found application data while handshake
7124 processing was enabled when in fact s->s3->in_read_app_data was
7125 merely automatically cleared during the initial handshake.
7126 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7128 *) Fix object definitions for Private and Enterprise: they were not
7129 recognized in their shortname (=lowercase) representation. Extend
7130 obj_dat.pl to issue an error when using undefined keywords instead
7131 of silently ignoring the problem (Svenning Sorensen
7132 <sss@sss.dnsalias.net>).
7135 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7136 generators, i.e. generators other than 2 and 5. (Previously, the
7137 code did not properly initialise the 'add' and 'rem' values to
7138 BN_generate_prime().)
7140 In the new general case, we do not insist that 'generator' is
7141 actually a primitive root: This requirement is rather pointless;
7142 a generator of the order-q subgroup is just as good, if not
7146 *) Map new X509 verification errors to alerts. Discovered and submitted by
7147 Tom Wu <tom@arcot.com>.
7150 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7151 returning non-zero before the data has been completely received
7152 when using non-blocking I/O.
7153 [Bodo Moeller; problem pointed out by John Hughes]
7155 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7156 [Ben Laurie, Lutz Jaenicke]
7158 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7159 Yoram Zahavi <YoramZ@gilian.com>).
7162 *) Add information about CygWin 1.3 and on, and preserve proper
7163 configuration for the versions before that.
7164 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7166 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7167 check whether we deal with a copy of a session and do not delete from
7168 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7169 <izhar@checkpoint.com>.
7172 *) Do not store session data into the internal session cache, if it
7173 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7174 flag is set). Proposed by Aslam <aslam@funk.com>.
7177 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7181 *) [In 0.9.6d-engine release:]
7182 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7183 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7185 *) Add the configuration target linux-s390x.
7186 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7188 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7189 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7190 variable as an indication that a ClientHello message has been
7191 received. As the flag value will be lost between multiple
7192 invocations of ssl3_accept when using non-blocking I/O, the
7193 function may not be aware that a handshake has actually taken
7194 place, thus preventing a new session from being added to the
7197 To avoid this problem, we now set s->new_session to 2 instead of
7198 using a local variable.
7199 [Lutz Jaenicke, Bodo Moeller]
7201 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7202 if the SSL_R_LENGTH_MISMATCH error is detected.
7203 [Geoff Thorpe, Bodo Moeller]
7205 *) New 'shared_ldflag' column in Configure platform table.
7208 *) Fix EVP_CIPHER_mode macro.
7209 ["Dan S. Camper" <dan@bti.net>]
7211 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7212 type, we must throw them away by setting rr->length to 0.
7213 [D P Chang <dpc@qualys.com>]
7215 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7217 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7218 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7219 worked incorrectly for those cases where range = 10..._2 and
7220 3*range is two bits longer than range.)
7223 *) Only add signing time to PKCS7 structures if it is not already
7227 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7228 OBJ_ld_ce should be OBJ_id_ce.
7229 Also some ip-pda OIDs in crypto/objects/objects.txt were
7230 incorrect (cf. RFC 3039).
7231 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7233 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7234 returns early because it has nothing to do.
7235 [Andy Schneider <andy.schneider@bjss.co.uk>]
7237 *) [In 0.9.6c-engine release:]
7238 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7239 [Andy Schneider <andy.schneider@bjss.co.uk>]
7241 *) [In 0.9.6c-engine release:]
7242 Add support for Cryptographic Appliance's keyserver technology.
7243 (Use engine 'keyclient')
7244 [Cryptographic Appliances and Geoff Thorpe]
7246 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7247 is called via tools/c89.sh because arguments have to be
7248 rearranged (all '-L' options must appear before the first object
7250 [Richard Shapiro <rshapiro@abinitio.com>]
7252 *) [In 0.9.6c-engine release:]
7253 Add support for Broadcom crypto accelerator cards, backported
7255 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7257 *) [In 0.9.6c-engine release:]
7258 Add support for SureWare crypto accelerator cards from
7259 Baltimore Technologies. (Use engine 'sureware')
7260 [Baltimore Technologies and Mark Cox]
7262 *) [In 0.9.6c-engine release:]
7263 Add support for crypto accelerator cards from Accelerated
7264 Encryption Processing, www.aep.ie. (Use engine 'aep')
7265 [AEP Inc. and Mark Cox]
7267 *) Add a configuration entry for gcc on UnixWare.
7268 [Gary Benson <gbenson@redhat.com>]
7270 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7271 messages are stored in a single piece (fixed-length part and
7272 variable-length part combined) and fix various bugs found on the way.
7275 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7276 instead. BIO_gethostbyname() does not know what timeouts are
7277 appropriate, so entries would stay in cache even when they have
7279 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7281 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7282 faced with a pathologically small ClientHello fragment that does
7283 not contain client_version: Instead of aborting with an error,
7284 simply choose the highest available protocol version (i.e.,
7285 TLS 1.0 unless it is disabled). In practice, ClientHello
7286 messages are never sent like this, but this change gives us
7287 strictly correct behaviour at least for TLS.
7290 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7291 never resets s->method to s->ctx->method when called from within
7292 one of the SSL handshake functions.
7293 [Bodo Moeller; problem pointed out by Niko Baric]
7295 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7296 (sent using the client's version number) if client_version is
7297 smaller than the protocol version in use. Also change
7298 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7299 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7300 the client will at least see that alert.
7303 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7307 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7308 client receives HelloRequest while in a handshake.
7309 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7311 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7312 should end in 'break', not 'goto end' which circumvents various
7313 cleanups done in state SSL_ST_OK. But session related stuff
7314 must be disabled for SSL_ST_OK in the case that we just sent a
7317 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7318 before just sending a HelloRequest.
7319 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7321 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7322 reveal whether illegal block cipher padding was found or a MAC
7323 verification error occured. (Neither SSLerr() codes nor alerts
7324 are directly visible to potential attackers, but the information
7325 may leak via logfiles.)
7327 Similar changes are not required for the SSL 2.0 implementation
7328 because the number of padding bytes is sent in clear for SSL 2.0,
7329 and the extra bytes are just ignored. However ssl/s2_pkt.c
7330 failed to verify that the purported number of padding bytes is in
7334 *) Add OpenUNIX-8 support including shared libraries
7335 (Boyd Lynn Gerber <gerberb@zenez.com>).
7338 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7339 'wristwatch attack' using huge encoding parameters (cf.
7340 James H. Manger's CRYPTO 2001 paper). Note that the
7341 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7342 encoding parameters and hence was not vulnerable.
7345 *) BN_sqr() bug fix.
7346 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7348 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7349 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7350 followed by modular reduction.
7351 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7353 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7354 equivalent based on BN_pseudo_rand() instead of BN_rand().
7357 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7358 This function was broken, as the check for a new client hello message
7359 to handle SGC did not allow these large messages.
7360 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7363 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7366 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7367 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7370 *) Rework the configuration and shared library support for Tru64 Unix.
7371 The configuration part makes use of modern compiler features and
7372 still retains old compiler behavior for those that run older versions
7373 of the OS. The shared library support part includes a variant that
7374 uses the RPATH feature, and is available through the special
7375 configuration target "alpha-cc-rpath", which will never be selected
7377 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7379 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7380 with the same message size as in ssl3_get_certificate_request().
7381 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7382 messages might inadvertently be reject as too long.
7383 [Petr Lampa <lampa@fee.vutbr.cz>]
7385 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7388 *) Modified SSL library such that the verify_callback that has been set
7389 specificly for an SSL object with SSL_set_verify() is actually being
7390 used. Before the change, a verify_callback set with this function was
7391 ignored and the verify_callback() set in the SSL_CTX at the time of
7392 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7393 to allow the necessary settings.
7396 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7397 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7398 done automatically (in contradiction to the requirements of the C
7399 standard). This made problems when used from OpenSSH.
7402 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7403 dh->length and always used
7405 BN_rand_range(priv_key, dh->p).
7407 BN_rand_range() is not necessary for Diffie-Hellman, and this
7408 specific range makes Diffie-Hellman unnecessarily inefficient if
7409 dh->length (recommended exponent length) is much smaller than the
7410 length of dh->p. We could use BN_rand_range() if the order of
7411 the subgroup was stored in the DH structure, but we only have
7416 BN_rand(priv_key, l, ...)
7418 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7424 RSA_eay_public_encrypt
7425 RSA_eay_private_decrypt
7426 RSA_eay_private_encrypt (signing)
7427 RSA_eay_public_decrypt (signature verification)
7429 (default implementations for RSA_public_encrypt,
7430 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7431 always reject numbers >= n.
7434 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7435 to synchronize access to 'locking_thread'. This is necessary on
7436 systems where access to 'locking_thread' (an 'unsigned long'
7437 variable) is not atomic.
7440 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7441 *before* setting the 'crypto_lock_rand' flag. The previous code had
7442 a race condition if 0 is a valid thread ID.
7443 [Travis Vitek <vitek@roguewave.com>]
7445 *) Add support for shared libraries under Irix.
7446 [Albert Chin-A-Young <china@thewrittenword.com>]
7448 *) Add configuration option to build on Linux on both big-endian and
7450 [Ralf Baechle <ralf@uni-koblenz.de>]
7452 *) Add the possibility to create shared libraries on HP-UX.
7455 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7457 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7458 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7459 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7460 PRNG state recovery was possible based on the output of
7461 one PRNG request appropriately sized to gain knowledge on
7462 'md' followed by enough consecutive 1-byte PRNG requests
7463 to traverse all of 'state'.
7465 1. When updating 'md_local' (the current thread's copy of 'md')
7466 during PRNG output generation, hash all of the previous
7467 'md_local' value, not just the half used for PRNG output.
7469 2. Make the number of bytes from 'state' included into the hash
7470 independent from the number of PRNG bytes requested.
7472 The first measure alone would be sufficient to avoid
7473 Markku-Juhani's attack. (Actually it had never occurred
7474 to me that the half of 'md_local' used for chaining was the
7475 half from which PRNG output bytes were taken -- I had always
7476 assumed that the secret half would be used.) The second
7477 measure makes sure that additional data from 'state' is never
7478 mixed into 'md_local' in small portions; this heuristically
7479 further strengthens the PRNG.
7482 *) Fix crypto/bn/asm/mips3.s.
7485 *) When only the key is given to "enc", the IV is undefined. Print out
7486 an error message in this case.
7489 *) Handle special case when X509_NAME is empty in X509 printing routines.
7492 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7493 positive and less than q.
7496 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7497 used: it isn't thread safe and the add_lock_callback should handle
7499 [Paul Rose <Paul.Rose@bridge.com>]
7501 *) Verify that incoming data obeys the block size in
7502 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7506 [Ulf Möller, Bodo Möller]
7508 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7509 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7510 when fixing the server behaviour for backwards-compatible 'client
7511 hello' messages. (Note that the attack is impractical against
7512 SSL 3.0 and TLS 1.0 anyway because length and version checking
7513 means that the probability of guessing a valid ciphertext is
7514 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7517 Before 0.9.5, the countermeasure (hide the error by generating a
7518 random 'decryption result') did not work properly because
7519 ERR_clear_error() was missing, meaning that SSL_get_error() would
7520 detect the supposedly ignored error.
7522 Both problems are now fixed.
7525 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7526 (previously it was 1024).
7529 *) Fix for compatibility mode trust settings: ignore trust settings
7530 unless some valid trust or reject settings are present.
7533 *) Fix for blowfish EVP: its a variable length cipher.
7536 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7537 parameters in DSA public key structures and return an error in the
7538 DSA routines if parameters are absent.
7541 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7542 in the current directory if neither $RANDFILE nor $HOME was set.
7543 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7544 caused some confusion to Windows users who haven't defined $HOME.
7545 Thus RAND_file_name() is changed again: e_os.h can define a
7546 DEFAULT_HOME, which will be used if $HOME is not set.
7547 For Windows, we use "C:"; on other platforms, we still require
7548 environment variables.
7550 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7551 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7552 having multiple threads call RAND_poll() concurrently.
7555 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7556 combination of a flag and a thread ID variable.
7557 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7558 flag), *other* threads can enter ssleay_add_bytes without obeying
7559 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7560 that they do not hold after the first thread unsets add_do_not_lock).
7563 *) Change bctest again: '-x' expressions are not available in all
7567 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7569 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7570 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7572 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7573 the default extension for executables, if any. Also, make the perl
7574 scripts that use symlink() to test if it really exists and use "cp"
7575 if it doesn't. All this made OpenSSL compilable and installable in
7579 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7580 If SEQUENCE is length is indefinite just set c->slen to the total
7581 amount of data available.
7582 [Steve Henson, reported by shige@FreeBSD.org]
7583 [This change does not apply to 0.9.7.]
7585 *) Change bctest to avoid here-documents inside command substitution
7586 (workaround for FreeBSD /bin/sh bug).
7587 For compatibility with Ultrix, avoid shell functions (introduced
7588 in the bctest version that searches along $PATH).
7591 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7592 with des_encrypt() defined on some operating systems, like Solaris
7596 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7597 On the Importance of Eliminating Errors in Cryptographic
7598 Computations, J. Cryptology 14 (2001) 2, 101-119,
7599 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7602 *) MIPS assembler BIGNUM division bug fix.
7605 *) Disabled incorrect Alpha assembler code.
7608 *) Fix PKCS#7 decode routines so they correctly update the length
7609 after reading an EOC for the EXPLICIT tag.
7611 [This change does not apply to 0.9.7.]
7613 *) Fix bug in PKCS#12 key generation routines. This was triggered
7614 if a 3DES key was generated with a 0 initial byte. Include
7615 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7616 (but broken) behaviour.
7619 *) Enhance bctest to search for a working bc along $PATH and print
7621 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7623 *) Fix memory leaks in err.c: free err_data string if necessary;
7624 don't write to the wrong index in ERR_set_error_data.
7627 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7631 *) Replace rdtsc with _emit statements for VC++ version 5.
7632 [Jeremy Cooper <jeremy@baymoo.org>]
7634 *) Make it possible to reuse SSLv2 sessions.
7637 *) In copy_email() check for >= 0 as a return value for
7638 X509_NAME_get_index_by_NID() since 0 is a valid index.
7639 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7641 *) Avoid coredump with unsupported or invalid public keys by checking if
7642 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7643 PKCS7_verify() fails with non detached data.
7646 *) Don't use getenv in library functions when run as setuid/setgid.
7647 New function OPENSSL_issetugid().
7650 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7651 due to incorrect handling of multi-threading:
7653 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7655 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7657 3. Count how many times MemCheck_off() has been called so that
7658 nested use can be treated correctly. This also avoids
7659 inband-signalling in the previous code (which relied on the
7660 assumption that thread ID 0 is impossible).
7663 *) Add "-rand" option also to s_client and s_server.
7666 *) Fix CPU detection on Irix 6.x.
7667 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7668 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7670 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7673 [This change does not apply to 0.9.7.]
7675 *) Use the cached encoding of an X509_NAME structure rather than
7676 copying it. This is apparently the reason for the libsafe "errors"
7677 but the code is actually correct.
7680 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7681 Bleichenbacher's DSA attack.
7682 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7683 to be set and top=0 forces the highest bit to be set; top=-1 is new
7684 and leaves the highest bit random.
7685 [Ulf Moeller, Bodo Moeller]
7687 *) In the NCONF_...-based implementations for CONF_... queries
7688 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7689 a temporary CONF structure with the data component set to NULL
7690 (which gives segmentation faults in lh_retrieve).
7691 Instead, use NULL for the CONF pointer in CONF_get_string and
7692 CONF_get_number (which may use environment variables) and directly
7693 return NULL from CONF_get_section.
7696 *) Fix potential buffer overrun for EBCDIC.
7699 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7700 keyUsage if basicConstraints absent for a CA.
7703 *) Make SMIME_write_PKCS7() write mail header values with a format that
7704 is more generally accepted (no spaces before the semicolon), since
7705 some programs can't parse those values properly otherwise. Also make
7706 sure BIO's that break lines after each write do not create invalid
7710 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7711 macros previously used would not encode an empty SEQUENCE OF
7712 and break the signature.
7714 [This change does not apply to 0.9.7.]
7716 *) Zero the premaster secret after deriving the master secret in
7720 *) Add some EVP_add_digest_alias registrations (as found in
7721 OpenSSL_add_all_digests()) to SSL_library_init()
7722 aka OpenSSL_add_ssl_algorithms(). This provides improved
7723 compatibility with peers using X.509 certificates
7724 with unconventional AlgorithmIdentifier OIDs.
7727 *) Fix for Irix with NO_ASM.
7728 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7730 *) ./config script fixes.
7731 [Ulf Moeller, Richard Levitte]
7733 *) Fix 'openssl passwd -1'.
7736 *) Change PKCS12_key_gen_asc() so it can cope with non null
7737 terminated strings whose length is passed in the passlen
7738 parameter, for example from PEM callbacks. This was done
7739 by adding an extra length parameter to asc2uni().
7740 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7742 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7743 call failed, free the DSA structure.
7746 *) Fix to uni2asc() to cope with zero length Unicode strings.
7747 These are present in some PKCS#12 files.
7750 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7751 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7752 when writing a 32767 byte record.
7753 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7755 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7756 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7758 (RSA objects have a reference count access to which is protected
7759 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7760 so they are meant to be shared between threads.)
7761 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7762 "Reddie, Steven" <Steven.Reddie@ca.com>]
7764 *) Fix a deadlock in CRYPTO_mem_leaks().
7767 *) Use better test patterns in bntest.
7770 *) rand_win.c fix for Borland C.
7773 *) BN_rshift bugfix for n == 0.
7776 *) Add a 'bctest' script that checks for some known 'bc' bugs
7777 so that 'make test' does not abort just because 'bc' is broken.
7780 *) Store verify_result within SSL_SESSION also for client side to
7781 avoid potential security hole. (Re-used sessions on the client side
7782 always resulted in verify_result==X509_V_OK, not using the original
7783 result of the server certificate verification.)
7786 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7787 SSL3_RT_APPLICATION_DATA, return 0.
7788 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7792 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7793 releases, have been re-implemented by renaming the previous
7794 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7795 and ssl3_read_internal, respectively, and adding 'peek' parameters
7796 to them. The new ssl[23]_{read,peek} functions are calls to
7797 ssl[23]_read_internal with the 'peek' flag set appropriately.
7798 A 'peek' parameter has also been added to ssl3_read_bytes, which
7799 does the actual work for ssl3_read_internal.
7802 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7803 the method-specific "init()" handler. Also clean up ex_data after
7804 calling the method-specific "finish()" handler. Previously, this was
7805 happening the other way round.
7808 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7809 The previous value, 12, was not always sufficient for BN_mod_exp().
7812 *) Make sure that shared libraries get the internal name engine with
7813 the full version number and not just 0. This should mark the
7814 shared libraries as not backward compatible. Of course, this should
7815 be changed again when we can guarantee backward binary compatibility.
7818 *) Fix typo in get_cert_by_subject() in by_dir.c
7819 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7821 *) Rework the system to generate shared libraries:
7823 - Make note of the expected extension for the shared libraries and
7824 if there is a need for symbolic links from for example libcrypto.so.0
7825 to libcrypto.so.0.9.7. There is extended info in Configure for
7828 - Make as few rebuilds of the shared libraries as possible.
7830 - Still avoid linking the OpenSSL programs with the shared libraries.
7832 - When installing, install the shared libraries separately from the
7836 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7838 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7839 and not in SSL_clear because the latter is also used by the
7840 accept/connect functions; previously, the settings made by
7841 SSL_set_read_ahead would be lost during the handshake.
7842 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7844 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7845 Previously, it would create entries for disabled algorithms no
7849 *) Added several new manual pages for SSL_* function.
7852 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7854 *) In ssl23_get_client_hello, generate an error message when faced
7855 with an initial SSL 3.0/TLS record that is too small to contain the
7856 first two bytes of the ClientHello message, i.e. client_version.
7857 (Note that this is a pathologic case that probably has never happened
7858 in real life.) The previous approach was to use the version number
7859 from the record header as a substitute; but our protocol choice
7860 should not depend on that one because it is not authenticated
7861 by the Finished messages.
7864 *) More robust randomness gathering functions for Windows.
7865 [Jeffrey Altman <jaltman@columbia.edu>]
7867 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7868 not set then we don't setup the error code for issuer check errors
7869 to avoid possibly overwriting other errors which the callback does
7870 handle. If an application does set the flag then we assume it knows
7871 what it is doing and can handle the new informational codes
7875 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7876 a general "ANY" type, as such it should be able to decode anything
7877 including tagged types. However it didn't check the class so it would
7878 wrongly interpret tagged types in the same way as their universal
7879 counterpart and unknown types were just rejected. Changed so that the
7880 tagged and unknown types are handled in the same way as a SEQUENCE:
7881 that is the encoding is stored intact. There is also a new type
7882 "V_ASN1_OTHER" which is used when the class is not universal, in this
7883 case we have no idea what the actual type is so we just lump them all
7887 *) On VMS, stdout may very well lead to a file that is written to
7888 in a record-oriented fashion. That means that every write() will
7889 write a separate record, which will be read separately by the
7890 programs trying to read from it. This can be very confusing.
7892 The solution is to put a BIO filter in the way that will buffer
7893 text until a linefeed is reached, and then write everything a
7894 line at a time, so every record written will be an actual line,
7895 not chunks of lines and not (usually doesn't happen, but I've
7896 seen it once) several lines in one record. BIO_f_linebuffer() is
7899 Currently, it's a VMS-only method, because that's where it has
7900 been tested well enough.
7903 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7904 it can return incorrect results.
7905 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7906 but it was in 0.9.6-beta[12].)
7909 *) Disable the check for content being present when verifying detached
7910 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7911 include zero length content when signing messages.
7914 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7915 BIO_ctrl (for BIO pairs).
7918 *) Add DSO method for VMS.
7921 *) Bug fix: Montgomery multiplication could produce results with the
7925 *) Add RPM specification openssl.spec and modify it to build three
7926 packages. The default package contains applications, application
7927 documentation and run-time libraries. The devel package contains
7928 include files, static libraries and function documentation. The
7929 doc package contains the contents of the doc directory. The original
7930 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7933 *) Add a large number of documentation files for many SSL routines.
7934 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7936 *) Add a configuration entry for Sony News 4.
7937 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7939 *) Don't set the two most significant bits to one when generating a
7940 random number < q in the DSA library.
7943 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7944 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7945 the underlying transport is blocking) if a handshake took place.
7946 (The default behaviour is needed by applications such as s_client
7947 and s_server that use select() to determine when to use SSL_read;
7948 but for applications that know in advance when to expect data, it
7949 just makes things more complicated.)
7952 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7956 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7957 work better on such systems.
7958 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7960 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7961 Update PKCS12_parse() so it copies the friendlyName and the
7962 keyid to the certificates aux info.
7965 *) Fix bug in PKCS7_verify() which caused an infinite loop
7966 if there was more than one signature.
7967 [Sven Uszpelkat <su@celocom.de>]
7969 *) Major change in util/mkdef.pl to include extra information
7970 about each symbol, as well as presenting variables as well
7971 as functions. This change means that there's n more need
7972 to rebuild the .num files when some algorithms are excluded.
7975 *) Allow the verify time to be set by an application,
7976 rather than always using the current time.
7979 *) Phase 2 verify code reorganisation. The certificate
7980 verify code now looks up an issuer certificate by a
7981 number of criteria: subject name, authority key id
7982 and key usage. It also verifies self signed certificates
7983 by the same criteria. The main comparison function is
7984 X509_check_issued() which performs these checks.
7986 Lot of changes were necessary in order to support this
7987 without completely rewriting the lookup code.
7989 Authority and subject key identifier are now cached.
7991 The LHASH 'certs' is X509_STORE has now been replaced
7992 by a STACK_OF(X509_OBJECT). This is mainly because an
7993 LHASH can't store or retrieve multiple objects with
7994 the same hash value.
7996 As a result various functions (which were all internal
7997 use only) have changed to handle the new X509_STORE
7998 structure. This will break anything that messed round
7999 with X509_STORE internally.
8001 The functions X509_STORE_add_cert() now checks for an
8002 exact match, rather than just subject name.
8004 The X509_STORE API doesn't directly support the retrieval
8005 of multiple certificates matching a given criteria, however
8006 this can be worked round by performing a lookup first
8007 (which will fill the cache with candidate certificates)
8008 and then examining the cache for matches. This is probably
8009 the best we can do without throwing out X509_LOOKUP
8010 entirely (maybe later...).
8012 The X509_VERIFY_CTX structure has been enhanced considerably.
8014 All certificate lookup operations now go via a get_issuer()
8015 callback. Although this currently uses an X509_STORE it
8016 can be replaced by custom lookups. This is a simple way
8017 to bypass the X509_STORE hackery necessary to make this
8018 work and makes it possible to use more efficient techniques
8019 in future. A very simple version which uses a simple
8020 STACK for its trusted certificate store is also provided
8021 using X509_STORE_CTX_trusted_stack().
8023 The verify_cb() and verify() callbacks now have equivalents
8024 in the X509_STORE_CTX structure.
8026 X509_STORE_CTX also has a 'flags' field which can be used
8027 to customise the verify behaviour.
8030 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8031 excludes S/MIME capabilities.
8034 *) When a certificate request is read in keep a copy of the
8035 original encoding of the signed data and use it when outputing
8036 again. Signatures then use the original encoding rather than
8037 a decoded, encoded version which may cause problems if the
8038 request is improperly encoded.
8041 *) For consistency with other BIO_puts implementations, call
8042 buffer_write(b, ...) directly in buffer_puts instead of calling
8045 In BIO_puts, increment b->num_write as in BIO_write.
8046 [Peter.Sylvester@EdelWeb.fr]
8048 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8049 BN_zero, we may not return a BIGNUM with an array consisting of
8053 *) Avoid calling abort() from within the library when problems are
8054 detected, except if preprocessor symbols have been defined
8055 (such as REF_CHECK, BN_DEBUG etc.).
8058 *) New openssl application 'rsautl'. This utility can be
8059 used for low level RSA operations. DER public key
8060 BIO/fp routines also added.
8063 *) New Configure entry and patches for compiling on QNX 4.
8064 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8066 *) A demo state-machine implementation was sponsored by
8067 Nuron (http://www.nuron.com/) and is now available in
8068 demos/state_machine.
8071 *) New options added to the 'dgst' utility for signature
8072 generation and verification.
8075 *) Unrecognized PKCS#7 content types are now handled via a
8076 catch all ASN1_TYPE structure. This allows unsupported
8077 types to be stored as a "blob" and an application can
8078 encode and decode it manually.
8081 *) Fix various signed/unsigned issues to make a_strex.c
8083 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8085 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8086 length if passed a buffer. ASN1_INTEGER_to_BN failed
8087 if passed a NULL BN and its argument was negative.
8088 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8090 *) Modification to PKCS#7 encoding routines to output definite
8091 length encoding. Since currently the whole structures are in
8092 memory there's not real point in using indefinite length
8093 constructed encoding. However if OpenSSL is compiled with
8094 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8097 *) Added BIO_vprintf() and BIO_vsnprintf().
8100 *) Added more prefixes to parse for in the the strings written
8101 through a logging bio, to cover all the levels that are available
8102 through syslog. The prefixes are now:
8104 PANIC, EMERG, EMR => LOG_EMERG
8105 ALERT, ALR => LOG_ALERT
8106 CRIT, CRI => LOG_CRIT
8107 ERROR, ERR => LOG_ERR
8108 WARNING, WARN, WAR => LOG_WARNING
8109 NOTICE, NOTE, NOT => LOG_NOTICE
8110 INFO, INF => LOG_INFO
8111 DEBUG, DBG => LOG_DEBUG
8113 and as before, if none of those prefixes are present at the
8114 beginning of the string, LOG_ERR is chosen.
8116 On Win32, the LOG_* levels are mapped according to this:
8118 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8119 LOG_WARNING => EVENTLOG_WARNING_TYPE
8120 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8124 *) Made it possible to reconfigure with just the configuration
8125 argument "reconf" or "reconfigure". The command line arguments
8126 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8127 and are retrieved from there when reconfiguring.
8131 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8133 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8136 *) The obj_dat.pl script was messing up the sorting of object
8137 names. The reason was that it compared the quoted version
8138 of strings as a result "OCSP" > "OCSP Signing" because
8139 " > SPACE. Changed script to store unquoted versions of
8140 names and add quotes on output. It was also omitting some
8141 names from the lookup table if they were given a default
8142 value (that is if SN is missing it is given the same
8143 value as LN and vice versa), these are now added on the
8144 grounds that if an object has a name we should be able to
8145 look it up. Finally added warning output when duplicate
8146 short or long names are found.
8149 *) Changes needed for Tandem NSK.
8150 [Scott Uroff <scott@xypro.com>]
8152 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8153 RSA_padding_check_SSLv23(), special padding was never detected
8154 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8155 version rollback attacks was not effective.
8157 In s23_clnt.c, don't use special rollback-attack detection padding
8158 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8159 client; similarly, in s23_srvr.c, don't do the rollback check if
8160 SSL 2.0 is the only protocol enabled in the server.
8163 *) Make it possible to get hexdumps of unprintable data with 'openssl
8164 asn1parse'. By implication, the functions ASN1_parse_dump() and
8165 BIO_dump_indent() are added.
8168 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8169 these print out strings and name structures based on various
8170 flags including RFC2253 support and proper handling of
8171 multibyte characters. Added options to the 'x509' utility
8172 to allow the various flags to be set.
8175 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8176 Also change the functions X509_cmp_current_time() and
8177 X509_gmtime_adj() work with an ASN1_TIME structure,
8178 this will enable certificates using GeneralizedTime in validity
8179 dates to be checked.
8182 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8183 negative public key encodings) on by default,
8184 NO_NEG_PUBKEY_BUG can be set to disable it.
8187 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8188 content octets. An i2c_ASN1_OBJECT is unnecessary because
8189 the encoding can be trivially obtained from the structure.
8192 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8193 not read locks (CRYPTO_r_[un]lock).
8196 *) A first attempt at creating official support for shared
8197 libraries through configuration. I've kept it so the
8198 default is static libraries only, and the OpenSSL programs
8199 are always statically linked for now, but there are
8200 preparations for dynamic linking in place.
8201 This has been tested on Linux and Tru64.
8204 *) Randomness polling function for Win9x, as described in:
8205 Peter Gutmann, Software Generation of Practically Strong
8209 *) Fix so PRNG is seeded in req if using an already existing
8213 *) New options to smime application. -inform and -outform
8214 allow alternative formats for the S/MIME message including
8215 PEM and DER. The -content option allows the content to be
8216 specified separately. This should allow things like Netscape
8217 form signing output easier to verify.
8220 *) Fix the ASN1 encoding of tags using the 'long form'.
8223 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8224 STRING types. These convert content octets to and from the
8225 underlying type. The actual tag and length octets are
8226 already assumed to have been read in and checked. These
8227 are needed because all other string types have virtually
8228 identical handling apart from the tag. By having versions
8229 of the ASN1 functions that just operate on content octets
8230 IMPLICIT tagging can be handled properly. It also allows
8231 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8232 and ASN1_INTEGER are identical apart from the tag.
8235 *) Change the handling of OID objects as follows:
8237 - New object identifiers are inserted in objects.txt, following
8238 the syntax given in objects.README.
8239 - objects.pl is used to process obj_mac.num and create a new
8241 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8244 This is currently kind of a hack, and the perl code in objects.pl
8245 isn't very elegant, but it works as I intended. The simplest way
8246 to check that it worked correctly is to look in obj_dat.h and
8247 check the array nid_objs and make sure the objects haven't moved
8248 around (this is important!). Additions are OK, as well as
8249 consistent name changes.
8252 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8255 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8256 The given file adds to whatever has already been seeded into the
8257 random pool through the RANDFILE configuration file option or
8258 environment variable, or the default random state file.
8261 *) mkstack.pl now sorts each macro group into lexical order.
8262 Previously the output order depended on the order the files
8263 appeared in the directory, resulting in needless rewriting
8267 *) Patches to make OpenSSL compile under Win32 again. Mostly
8268 work arounds for the VC++ problem that it treats func() as
8269 func(void). Also stripped out the parts of mkdef.pl that
8270 added extra typesafe functions: these no longer exist.
8273 *) Reorganisation of the stack code. The macros are now all
8274 collected in safestack.h . Each macro is defined in terms of
8275 a "stack macro" of the form SKM_<name>(type, a, b). The
8276 DEBUG_SAFESTACK is now handled in terms of function casts,
8277 this has the advantage of retaining type safety without the
8278 use of additional functions. If DEBUG_SAFESTACK is not defined
8279 then the non typesafe macros are used instead. Also modified the
8280 mkstack.pl script to handle the new form. Needs testing to see
8281 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8282 the default if no major problems. Similar behaviour for ASN1_SET_OF
8283 and PKCS12_STACK_OF.
8286 *) When some versions of IIS use the 'NET' form of private key the
8287 key derivation algorithm is different. Normally MD5(password) is
8288 used as a 128 bit RC4 key. In the modified case
8289 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8290 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8291 as the old Netscape_RSA functions except they have an additional
8292 'sgckey' parameter which uses the modified algorithm. Also added
8293 an -sgckey command line option to the rsa utility. Thanks to
8294 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8295 algorithm to openssl-dev.
8298 *) The evp_local.h macros were using 'c.##kname' which resulted in
8299 invalid expansion on some systems (SCO 5.0.5 for example).
8300 Corrected to 'c.kname'.
8301 [Phillip Porch <root@theporch.com>]
8303 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8304 a STACK of email addresses from a certificate or request, these look
8305 in the subject name and the subject alternative name extensions and
8306 omit any duplicate addresses.
8309 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8310 This makes DSA verification about 2 % faster.
8313 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8314 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8315 plus overhead for 1024 bit moduli).
8316 This makes exponentiations about 0.5 % faster for 1024 bit
8317 exponents (as measured by "openssl speed rsa2048").
8320 *) Rename memory handling macros to avoid conflicts with other
8322 Malloc => OPENSSL_malloc
8323 Malloc_locked => OPENSSL_malloc_locked
8324 Realloc => OPENSSL_realloc
8325 Free => OPENSSL_free
8328 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8329 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8332 *) CygWin32 support.
8333 [John Jarvie <jjarvie@newsguy.com>]
8335 *) The type-safe stack code has been rejigged. It is now only compiled
8336 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8337 by default all type-specific stack functions are "#define"d back to
8338 standard stack functions. This results in more streamlined output
8339 but retains the type-safety checking possibilities of the original
8343 *) The STACK code has been cleaned up, and certain type declarations
8344 that didn't make a lot of sense have been brought in line. This has
8345 also involved a cleanup of sorts in safestack.h to more correctly
8346 map type-safe stack functions onto their plain stack counterparts.
8347 This work has also resulted in a variety of "const"ifications of
8348 lots of the code, especially "_cmp" operations which should normally
8349 be prototyped with "const" parameters anyway.
8352 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8353 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8354 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8355 where all of 'md' is used each time the PRNG is used, but 'state'
8356 is used only indexed by a cyclic counter. As entropy may not be
8357 well distributed from the beginning, 'md' is important as a
8358 chaining variable. However, the output function chains only half
8359 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8360 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8361 in all of 'state' being rewritten, with the new values depending
8362 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8365 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8366 the handshake is continued after ssl_verify_cert_chain();
8367 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8368 can lead to 'unexplainable' connection aborts later.
8369 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8371 *) Major EVP API cipher revision.
8372 Add hooks for extra EVP features. This allows various cipher
8373 parameters to be set in the EVP interface. Support added for variable
8374 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8375 setting of RC2 and RC5 parameters.
8377 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8380 Remove lots of duplicated code from the EVP library. For example *every*
8381 cipher init() function handles the 'iv' in the same way according to the
8382 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8383 for CFB and OFB modes they zero ctx->num.
8385 New functionality allows removal of S/MIME code RC2 hack.
8387 Most of the routines have the same form and so can be declared in terms
8390 By shifting this to the top level EVP_CipherInit() it can be removed from
8391 all individual ciphers. If the cipher wants to handle IVs or keys
8392 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8395 Change lots of functions like EVP_EncryptUpdate() to now return a
8396 value: although software versions of the algorithms cannot fail
8397 any installed hardware versions can.
8400 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8401 this option is set, tolerate broken clients that send the negotiated
8402 protocol version number instead of the requested protocol version
8406 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8407 i.e. non-zero for export ciphersuites, zero otherwise.
8408 Previous versions had this flag inverted, inconsistent with
8409 rsa_tmp_cb (..._TMP_RSA_CB).
8410 [Bodo Moeller; problem reported by Amit Chopra]
8412 *) Add missing DSA library text string. Work around for some IIS
8413 key files with invalid SEQUENCE encoding.
8416 *) Add a document (doc/standards.txt) that list all kinds of standards
8417 and so on that are implemented in OpenSSL.
8420 *) Enhance c_rehash script. Old version would mishandle certificates
8421 with the same subject name hash and wouldn't handle CRLs at all.
8422 Added -fingerprint option to crl utility, to support new c_rehash
8426 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8429 *) Fix for SSL server purpose checking. Server checking was
8430 rejecting certificates which had extended key usage present
8431 but no ssl client purpose.
8432 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8434 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8435 is a little unclear about how a blank password is handled.
8436 Since the password in encoded as a BMPString with terminating
8437 double NULL a zero length password would end up as just the
8438 double NULL. However no password at all is different and is
8439 handled differently in the PKCS#12 key generation code. NS
8440 treats a blank password as zero length. MSIE treats it as no
8441 password on export: but it will try both on import. We now do
8442 the same: PKCS12_parse() tries zero length and no password if
8443 the password is set to "" or NULL (NULL is now a valid password:
8444 it wasn't before) as does the pkcs12 application.
8447 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8448 perror when PEM_read_bio_X509_REQ fails, the error message must
8449 be obtained from the error queue.
8452 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8453 it in ERR_remove_state if appropriate, and change ERR_get_state
8454 accordingly to avoid race conditions (this is necessary because
8455 thread_hash is no longer constant once set).
8458 *) Bugfix for linux-elf makefile.one.
8461 *) RSA_get_default_method() will now cause a default
8462 RSA_METHOD to be chosen if one doesn't exist already.
8463 Previously this was only set during a call to RSA_new()
8464 or RSA_new_method(NULL) meaning it was possible for
8465 RSA_get_default_method() to return NULL.
8468 *) Added native name translation to the existing DSO code
8469 that will convert (if the flag to do so is set) filenames
8470 that are sufficiently small and have no path information
8471 into a canonical native form. Eg. "blah" converted to
8472 "libblah.so" or "blah.dll" etc.
8475 *) New function ERR_error_string_n(e, buf, len) which is like
8476 ERR_error_string(e, buf), but writes at most 'len' bytes
8477 including the 0 terminator. For ERR_error_string_n, 'buf'
8479 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8481 *) CONF library reworked to become more general. A new CONF
8482 configuration file reader "class" is implemented as well as a
8483 new functions (NCONF_*, for "New CONF") to handle it. The now
8484 old CONF_* functions are still there, but are reimplemented to
8485 work in terms of the new functions. Also, a set of functions
8486 to handle the internal storage of the configuration data is
8487 provided to make it easier to write new configuration file
8488 reader "classes" (I can definitely see something reading a
8489 configuration file in XML format, for example), called _CONF_*,
8490 or "the configuration storage API"...
8492 The new configuration file reading functions are:
8494 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8495 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8497 NCONF_default, NCONF_WIN32
8499 NCONF_dump_fp, NCONF_dump_bio
8501 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8502 NCONF_new creates a new CONF object. This works in the same way
8503 as other interfaces in OpenSSL, like the BIO interface.
8504 NCONF_dump_* dump the internal storage of the configuration file,
8505 which is useful for debugging. All other functions take the same
8506 arguments as the old CONF_* functions wth the exception of the
8507 first that must be a `CONF *' instead of a `LHASH *'.
8509 To make it easer to use the new classes with the old CONF_* functions,
8510 the function CONF_set_default_method is provided.
8513 *) Add '-tls1' option to 'openssl ciphers', which was already
8514 mentioned in the documentation but had not been implemented.
8515 (This option is not yet really useful because even the additional
8516 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8519 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8520 OpenSSL-based applications) load shared libraries and bind to
8521 them in a portable way.
8522 [Geoff Thorpe, with contributions from Richard Levitte]
8524 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8526 *) Make sure _lrotl and _lrotr are only used with MSVC.
8528 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8529 (the default implementation of RAND_status).
8531 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8532 to '-clrext' (= clear extensions), as intended and documented.
8533 [Bodo Moeller; inconsistency pointed out by Michael Attili
8536 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8537 was larger than the MD block size.
8538 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8540 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8541 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8542 using the passed key: if the passed key was a private key the result
8543 of X509_print(), for example, would be to print out all the private key
8547 *) des_quad_cksum() byte order bug fix.
8548 [Ulf Möller, using the problem description in krb4-0.9.7, where
8549 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8551 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8553 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8555 *) For easily testing in shell scripts whether some command
8556 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8557 returns with exit code 0 iff no command of the given name is available.
8558 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8559 the output goes to stdout and nothing is printed to stderr.
8560 Additional arguments are always ignored.
8562 Since for each cipher there is a command of the same name,
8563 the 'no-cipher' compilation switches can be tested this way.
8565 ('openssl no-XXX' is not able to detect pseudo-commands such
8566 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8569 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8572 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8573 is set; it will be thrown away anyway because each handshake creates
8575 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8576 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8577 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8578 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8581 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8582 'Q' and 'R' lose their special meanings (quit/renegotiate).
8583 This is part of what -quiet does; unlike -quiet, -ign_eof
8584 does not suppress any output.
8587 *) Add compatibility options to the purpose and trust code. The
8588 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8589 accepts a certificate or CA, this was the previous behaviour,
8590 with all the associated security issues.
8592 X509_TRUST_COMPAT is the old trust behaviour: only and
8593 automatically trust self signed roots in certificate store. A
8594 new trust setting X509_TRUST_DEFAULT is used to specify that
8595 a purpose has no associated trust setting and it should instead
8596 use the value in the default purpose.
8599 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8600 and fix a memory leak.
8603 *) In util/mkerr.pl (which implements 'make errors'), preserve
8604 reason strings from the previous version of the .c file, as
8605 the default to have only downcase letters (and digits) in
8606 automatically generated reasons codes is not always appropriate.
8609 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8610 using strerror. Previously, ERR_reason_error_string() returned
8611 library names as reason strings for SYSerr; but SYSerr is a special
8612 case where small numbers are errno values, not library numbers.
8615 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8616 converts DSA parameters into DH parameters. (When creating parameters,
8617 DSA_generate_parameters is used.)
8620 *) Include 'length' (recommended exponent length) in C code generated
8621 by 'openssl dhparam -C'.
8624 *) The second argument to set_label in perlasm was already being used
8625 so couldn't be used as a "file scope" flag. Moved to third argument
8629 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8630 instead of RAND_bytes for encryption IVs and salts.
8633 *) Include RAND_status() into RAND_METHOD instead of implementing
8634 it only for md_rand.c Otherwise replacing the PRNG by calling
8635 RAND_set_rand_method would be impossible.
8638 *) Don't let DSA_generate_key() enter an infinite loop if the random
8639 number generation fails.
8642 *) New 'rand' application for creating pseudo-random output.
8645 *) Added configuration support for Linux/IA64
8646 [Rolf Haberrecker <rolf@suse.de>]
8648 *) Assembler module support for Mingw32.
8651 *) Shared library support for HPUX (in shlib/).
8652 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8654 *) Shared library support for Solaris gcc.
8655 [Lutz Behnke <behnke@trustcenter.de>]
8657 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8659 *) PKCS7_encrypt() was adding text MIME headers twice because they
8660 were added manually and by SMIME_crlf_copy().
8663 *) In bntest.c don't call BN_rand with zero bits argument.
8664 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8666 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8667 case was implemented. This caused BN_div_recp() to fail occasionally.
8670 *) Add an optional second argument to the set_label() in the perl
8671 assembly language builder. If this argument exists and is set
8672 to 1 it signals that the assembler should use a symbol whose
8673 scope is the entire file, not just the current function. This
8674 is needed with MASM which uses the format label:: for this scope.
8675 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8677 *) Change the ASN1 types so they are typedefs by default. Before
8678 almost all types were #define'd to ASN1_STRING which was causing
8679 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8683 *) Change names of new functions to the new get1/get0 naming
8684 convention: After 'get1', the caller owns a reference count
8685 and has to call ..._free; 'get0' returns a pointer to some
8686 data structure without incrementing reference counters.
8687 (Some of the existing 'get' functions increment a reference
8688 counter, some don't.)
8689 Similarly, 'set1' and 'add1' functions increase reference
8690 counters or duplicate objects.
8693 *) Allow for the possibility of temp RSA key generation failure:
8694 the code used to assume it always worked and crashed on failure.
8697 *) Fix potential buffer overrun problem in BIO_printf().
8698 [Ulf Möller, using public domain code by Patrick Powell; problem
8699 pointed out by David Sacerdote <das33@cornell.edu>]
8701 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8702 RAND_egd() and RAND_status(). In the command line application,
8703 the EGD socket can be specified like a seed file using RANDFILE
8707 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8708 Some CAs (e.g. Verisign) distribute certificates in this form.
8711 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8712 list to exclude them. This means that no special compilation option
8713 is needed to use anonymous DH: it just needs to be included in the
8717 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8718 EVP_MD_type. The old functionality is available in a new macro called
8719 EVP_MD_md(). Change code that uses it and update docs.
8722 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8723 where the 'void *' argument is replaced by a function pointer argument.
8724 Previously 'void *' was abused to point to functions, which works on
8725 many platforms, but is not correct. As these functions are usually
8726 called by macros defined in OpenSSL header files, most source code
8727 should work without changes.
8730 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8731 sections with information on -D... compiler switches used for
8732 compiling the library so that applications can see them. To enable
8733 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8734 must be defined. E.g.,
8735 #define OPENSSL_ALGORITHM_DEFINES
8736 #include <openssl/opensslconf.h>
8737 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8738 [Richard Levitte, Ulf and Bodo Möller]
8740 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8744 *) Change the 'other' type in certificate aux info to a STACK_OF
8745 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8746 the required ASN1 format: arbitrary types determined by an OID.
8749 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8750 argument to 'req'. This is not because the function is newer or
8751 better than others it just uses the work 'NEW' in the certificate
8752 request header lines. Some software needs this.
8755 *) Reorganise password command line arguments: now passwords can be
8756 obtained from various sources. Delete the PEM_cb function and make
8757 it the default behaviour: i.e. if the callback is NULL and the
8758 usrdata argument is not NULL interpret it as a null terminated pass
8759 phrase. If usrdata and the callback are NULL then the pass phrase
8760 is prompted for as usual.
8763 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8764 the support is automatically enabled. The resulting binaries will
8765 autodetect the card and use it if present.
8766 [Ben Laurie and Compaq Inc.]
8768 *) Work around for Netscape hang bug. This sends certificate request
8769 and server done in one record. Since this is perfectly legal in the
8770 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8771 the bugs/SSLv3 entry for more info.
8774 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8777 *) Add -rand argument to smime and pkcs12 applications and read/write
8781 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8784 *) Add command line password options to the remaining applications.
8787 *) Bug fix for BN_div_recp() for numerators with an even number of
8791 *) More tests in bntest.c, and changed test_bn output.
8794 *) ./config recognizes MacOS X now.
8797 *) Bug fix for BN_div() when the first words of num and divsor are
8798 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8801 *) Add support for various broken PKCS#8 formats, and command line
8802 options to produce them.
8805 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8806 get temporary BIGNUMs from a BN_CTX.
8809 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8813 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8814 include a #define from the old name to the new. The original intent
8815 was that statically linked binaries could for example just call
8816 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8817 link with digests. This never worked because SSLeay_add_all_digests()
8818 and SSLeay_add_all_ciphers() were in the same source file so calling
8819 one would link with the other. They are now in separate source files.
8822 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8825 *) Use a less unusual form of the Miller-Rabin primality test (it used
8826 a binary algorithm for exponentiation integrated into the Miller-Rabin
8827 loop, our standard modexp algorithms are faster).
8830 *) Support for the EBCDIC character set completed.
8831 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8833 *) Source code cleanups: use const where appropriate, eliminate casts,
8834 use void * instead of char * in lhash.
8837 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8838 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8839 this the server could overwrite ephemeral keys that the client
8843 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8844 using 50 iterations of the Rabin-Miller test.
8846 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8847 iterations of the Rabin-Miller test as required by the appendix
8848 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8849 As BN_is_prime_fasttest includes trial division, DSA parameter
8850 generation becomes much faster.
8852 This implies a change for the callback functions in DSA_is_prime
8853 and DSA_generate_parameters: The callback function is called once
8854 for each positive witness in the Rabin-Miller test, not just
8855 occasionally in the inner loop; and the parameters to the
8856 callback function now provide an iteration count for the outer
8857 loop rather than for the current invocation of the inner loop.
8858 DSA_generate_parameters additionally can call the callback
8859 function with an 'iteration count' of -1, meaning that a
8860 candidate has passed the trial division test (when q is generated
8861 from an application-provided seed, trial division is skipped).
8864 *) New function BN_is_prime_fasttest that optionally does trial
8865 division before starting the Rabin-Miller test and has
8866 an additional BN_CTX * argument (whereas BN_is_prime always
8867 has to allocate at least one BN_CTX).
8868 'callback(1, -1, cb_arg)' is called when a number has passed the
8869 trial division stage.
8872 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8876 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8879 *) New function BN_pseudo_rand().
8882 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8883 bignum version of BN_from_montgomery() with the working code from
8884 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8888 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8889 made it impossible to use the same SSL_SESSION data structure in
8890 SSL2 clients in multiple threads.
8893 *) The return value of RAND_load_file() no longer counts bytes obtained
8894 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8895 to seed the PRNG (previously an explicit byte count was required).
8896 [Ulf Möller, Bodo Möller]
8898 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8899 used (char *) instead of (void *) and had casts all over the place.
8902 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8905 *) Retain source code compatibility for BN_prime_checks macro:
8906 BN_is_prime(..., BN_prime_checks, ...) now uses
8907 BN_prime_checks_for_size to determine the appropriate number of
8908 Rabin-Miller iterations.
8911 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8912 DH_CHECK_P_NOT_SAFE_PRIME.
8913 (Check if this is true? OpenPGP calls them "strong".)
8916 *) Merge the functionality of "dh" and "gendh" programs into a new program
8917 "dhparam". The old programs are retained for now but will handle DH keys
8918 (instead of parameters) in future.
8921 *) Make the ciphers, s_server and s_client programs check the return values
8922 when a new cipher list is set.
8925 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8926 ciphers. Before when the 56bit ciphers were enabled the sorting was
8929 The syntax for the cipher sorting has been extended to support sorting by
8930 cipher-strength (using the strength_bits hard coded in the tables).
8931 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8933 Fix a bug in the cipher-command parser: when supplying a cipher command
8934 string with an "undefined" symbol (neither command nor alphanumeric
8935 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8936 an error is flagged.
8938 Due to the strength-sorting extension, the code of the
8939 ssl_create_cipher_list() function was completely rearranged. I hope that
8940 the readability was also increased :-)
8941 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8943 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8944 for the first serial number and places 2 in the serial number file. This
8945 avoids problems when the root CA is created with serial number zero and
8946 the first user certificate has the same issuer name and serial number
8950 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8951 the new code. Add documentation for this stuff.
8954 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8955 X509_*() to X509at_*() on the grounds that they don't handle X509
8956 structures and behave in an analogous way to the X509v3 functions:
8957 they shouldn't be called directly but wrapper functions should be used
8960 So we also now have some wrapper functions that call the X509at functions
8961 when passed certificate requests. (TO DO: similar things can be done with
8962 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8963 things. Some of these need some d2i or i2d and print functionality
8964 because they handle more complex structures.)
8967 *) Add missing #ifndefs that caused missing symbols when building libssl
8968 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8969 NO_RSA in ssl/s2*.c.
8970 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8972 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8973 has a return value which indicates the quality of the random data
8974 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8975 error queue. New function RAND_pseudo_bytes() generates output that is
8976 guaranteed to be unique but not unpredictable. RAND_add is like
8977 RAND_seed, but takes an extra argument for an entropy estimate
8978 (RAND_seed always assumes full entropy).
8981 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8982 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8983 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8984 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8985 false-positive rate of at most 2^-80 for random input.
8988 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8991 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8992 in the 0.9.5 release), this returns the chain
8993 from an X509_CTX structure with a dup of the stack and all
8994 the X509 reference counts upped: so the stack will exist
8995 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8998 Also make SSL_SESSION_print() print out the verify return
9002 *) Add manpage for the pkcs12 command. Also change the default
9003 behaviour so MAC iteration counts are used unless the new
9004 -nomaciter option is used. This improves file security and
9005 only older versions of MSIE (4.0 for example) need it.
9008 *) Honor the no-xxx Configure options when creating .DEF files.
9011 *) Add PKCS#10 attributes to field table: challengePassword,
9012 unstructuredName and unstructuredAddress. These are taken from
9013 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9014 international characters are used.
9016 More changes to X509_ATTRIBUTE code: allow the setting of types
9017 based on strings. Remove the 'loc' parameter when adding
9018 attributes because these will be a SET OF encoding which is sorted
9022 *) Initial changes to the 'req' utility to allow request generation
9023 automation. This will allow an application to just generate a template
9024 file containing all the field values and have req construct the
9027 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9028 used all over the place including certificate requests and PKCS#7
9029 structures. They are currently handled manually where necessary with
9030 some primitive wrappers for PKCS#7. The new functions behave in a
9031 manner analogous to the X509 extension functions: they allow
9032 attributes to be looked up by NID and added.
9034 Later something similar to the X509V3 code would be desirable to
9035 automatically handle the encoding, decoding and printing of the
9036 more complex types. The string types like challengePassword can
9037 be handled by the string table functions.
9039 Also modified the multi byte string table handling. Now there is
9040 a 'global mask' which masks out certain types. The table itself
9041 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9042 is useful when for example there is only one permissible type
9043 (as in countryName) and using the mask might result in no valid
9047 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9048 SSL_get_peer_finished to allow applications to obtain the latest
9049 Finished messages sent to the peer or expected from the peer,
9050 respectively. (SSL_get_peer_finished is usually the Finished message
9051 actually received from the peer, otherwise the protocol will be aborted.)
9053 As the Finished message are message digests of the complete handshake
9054 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9055 be used for external authentication procedures when the authentication
9056 provided by SSL/TLS is not desired or is not enough.
9059 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9060 the host supports BWX extension and if Compaq C is present on the
9061 $PATH. Just exploiting of the BWX extension results in 20-30%
9062 performance kick for some algorithms, e.g. DES and RC4 to mention
9063 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9067 *) Add support for MS "fast SGC". This is arguably a violation of the
9068 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9069 weak crypto and after checking the certificate is SGC a second one
9070 with strong crypto. MS SGC stops the first handshake after receiving
9071 the server certificate message and sends a second client hello. Since
9072 a server will typically do all the time consuming operations before
9073 expecting any further messages from the client (server key exchange
9074 is the most expensive) there is little difference between the two.
9076 To get OpenSSL to support MS SGC we have to permit a second client
9077 hello message after we have sent server done. In addition we have to
9078 reset the MAC if we do get this second client hello.
9081 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9082 if a DER encoded private key is RSA or DSA traditional format. Changed
9083 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9084 format DER encoded private key. Newer code should use PKCS#8 format which
9085 has the key type encoded in the ASN1 structure. Added DER private key
9086 support to pkcs8 application.
9089 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9090 ciphersuites has been selected (as required by the SSL 3/TLS 1
9091 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9092 is set, we interpret this as a request to violate the specification
9093 (the worst that can happen is a handshake failure, and 'correct'
9094 behaviour would result in a handshake failure anyway).
9097 *) In SSL_CTX_add_session, take into account that there might be multiple
9098 SSL_SESSION structures with the same session ID (e.g. when two threads
9099 concurrently obtain them from an external cache).
9100 The internal cache can handle only one SSL_SESSION with a given ID,
9101 so if there's a conflict, we now throw out the old one to achieve
9105 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9106 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9107 some routines that use cipher OIDs: some ciphers do not have OIDs
9108 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9112 *) Simplify the trust setting structure and code. Now we just have
9113 two sequences of OIDs for trusted and rejected settings. These will
9114 typically have values the same as the extended key usage extension
9115 and any application specific purposes.
9117 The trust checking code now has a default behaviour: it will just
9118 check for an object with the same NID as the passed id. Functions can
9119 be provided to override either the default behaviour or the behaviour
9120 for a given id. SSL client, server and email already have functions
9121 in place for compatibility: they check the NID and also return "trusted"
9122 if the certificate is self signed.
9125 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9126 traditional format into an EVP_PKEY structure.
9129 *) Add a password callback function PEM_cb() which either prompts for
9130 a password if usr_data is NULL or otherwise assumes it is a null
9131 terminated password. Allow passwords to be passed on command line
9132 environment or config files in a few more utilities.
9135 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9136 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9137 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9138 Update documentation.
9141 *) Support for ASN1 "NULL" type. This could be handled before by using
9142 ASN1_TYPE but there wasn't any function that would try to read a NULL
9143 and produce an error if it couldn't. For compatibility we also have
9144 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9145 don't allocate anything because they don't need to.
9148 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9150 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9152 *) Rebuild of the memory allocation routines used by OpenSSL code and
9153 possibly others as well. The purpose is to make an interface that
9154 provide hooks so anyone can build a separate set of allocation and
9155 deallocation routines to be used by OpenSSL, for example memory
9156 pool implementations, or something else, which was previously hard
9157 since Malloc(), Realloc() and Free() were defined as macros having
9158 the values malloc, realloc and free, respectively (except for Win32
9159 compilations). The same is provided for memory debugging code.
9160 OpenSSL already comes with functionality to find memory leaks, but
9161 this gives people a chance to debug other memory problems.
9163 With these changes, a new set of functions and macros have appeared:
9165 CRYPTO_set_mem_debug_functions() [F]
9166 CRYPTO_get_mem_debug_functions() [F]
9167 CRYPTO_dbg_set_options() [F]
9168 CRYPTO_dbg_get_options() [F]
9169 CRYPTO_malloc_debug_init() [M]
9171 The memory debug functions are NULL by default, unless the library
9172 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9173 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9174 gives the standard debugging functions that come with OpenSSL) or
9175 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9176 provided by the library user) must be used. When the standard
9177 debugging functions are used, CRYPTO_dbg_set_options can be used to
9178 request additional information:
9179 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9180 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9182 Also, things like CRYPTO_set_mem_functions will always give the
9183 expected result (the new set of functions is used for allocation
9184 and deallocation) at all times, regardless of platform and compiler
9187 To finish it up, some functions that were never use in any other
9188 way than through macros have a new API and new semantic:
9191 CRYPTO_dbg_realloc()
9194 All macros of value have retained their old syntax.
9195 [Richard Levitte and Bodo Moeller]
9197 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9198 ordering of SMIMECapabilities wasn't in "strength order" and there
9199 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9203 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9204 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9205 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9207 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9208 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9209 functionality to handle multipart/signed properly) and a utility
9210 called 'smime' to call all this stuff. This is based on code I
9211 originally wrote for Celo who have kindly allowed it to be
9212 included in OpenSSL.
9215 *) Add variants des_set_key_checked and des_set_key_unchecked of
9216 des_set_key (aka des_key_sched). Global variable des_check_key
9217 decides which of these is called by des_set_key; this way
9218 des_check_key behaves as it always did, but applications and
9219 the library itself, which was buggy for des_check_key == 1,
9220 have a cleaner way to pick the version they need.
9223 *) New function PKCS12_newpass() which changes the password of a
9227 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9228 dynamic mix. In both cases the ids can be used as an index into the
9229 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9230 functions so they accept a list of the field values and the
9231 application doesn't need to directly manipulate the X509_TRUST
9235 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9239 *) Modify the way the V3 extension code looks up extensions. This now
9240 works in a similar way to the object code: we have some "standard"
9241 extensions in a static table which is searched with OBJ_bsearch()
9242 and the application can add dynamic ones if needed. The file
9243 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9244 updated whenever a new extension is added to the core code and kept
9245 in ext_nid order. There is a simple program 'tabtest.c' which checks
9246 this. New extensions are not added too often so this file can readily
9247 be maintained manually.
9249 There are two big advantages in doing things this way. The extensions
9250 can be looked up immediately and no longer need to be "added" using
9251 X509V3_add_standard_extensions(): this function now does nothing.
9252 [Side note: I get *lots* of email saying the extension code doesn't
9253 work because people forget to call this function]
9254 Also no dynamic allocation is done unless new extensions are added:
9255 so if we don't add custom extensions there is no need to call
9256 X509V3_EXT_cleanup().
9259 *) Modify enc utility's salting as follows: make salting the default. Add a
9260 magic header, so unsalted files fail gracefully instead of just decrypting
9261 to garbage. This is because not salting is a big security hole, so people
9262 should be discouraged from doing it.
9265 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9266 digest to be passed on the command line but it only used this
9267 parameter when signing a certificate. Modified so all relevant
9268 operations are affected by the digest parameter including the
9269 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9270 DSA key was used because it didn't fix the digest.
9273 *) Initial certificate chain verify code. Currently tests the untrusted
9274 certificates for consistency with the verify purpose (which is set
9275 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9277 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9278 this is because it will reject chains with invalid extensions whereas
9279 every previous version of OpenSSL and SSLeay made no checks at all.
9281 Trust code: checks the root CA for the relevant trust settings. Trust
9282 settings have an initial value consistent with the verify purpose: e.g.
9283 if the verify purpose is for SSL client use it expects the CA to be
9284 trusted for SSL client use. However the default value can be changed to
9285 permit custom trust settings: one example of this would be to only trust
9286 certificates from a specific "secure" set of CAs.
9288 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9289 which should be used for version portability: especially since the
9290 verify structure is likely to change more often now.
9292 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9293 to set them. If not set then assume SSL clients will verify SSL servers
9296 Two new options to the verify program: -untrusted allows a set of
9297 untrusted certificates to be passed in and -purpose which sets the
9298 intended purpose of the certificate. If a purpose is set then the
9299 new chain verify code is used to check extension consistency.
9302 *) Support for the authority information access extension.
9305 *) Modify RSA and DSA PEM read routines to transparently handle
9306 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9307 public keys in a format compatible with certificate
9308 SubjectPublicKeyInfo structures. Unfortunately there were already
9309 functions called *_PublicKey_* which used various odd formats so
9310 these are retained for compatibility: however the DSA variants were
9311 never in a public release so they have been deleted. Changed dsa/rsa
9312 utilities to handle the new format: note no releases ever handled public
9313 keys so we should be OK.
9315 The primary motivation for this change is to avoid the same fiasco
9316 that dogs private keys: there are several incompatible private key
9317 formats some of which are standard and some OpenSSL specific and
9318 require various evil hacks to allow partial transparent handling and
9319 even then it doesn't work with DER formats. Given the option anything
9320 other than PKCS#8 should be dumped: but the other formats have to
9321 stay in the name of compatibility.
9323 With public keys and the benefit of hindsight one standard format
9324 is used which works with EVP_PKEY, RSA or DSA structures: though
9325 it clearly returns an error if you try to read the wrong kind of key.
9327 Added a -pubkey option to the 'x509' utility to output the public key.
9328 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9329 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9330 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9331 that do the same as the EVP_PKEY_assign_*() except they up the
9332 reference count of the added key (they don't "swallow" the
9336 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9337 CRLs would fail if the file contained no certificates or no CRLs:
9338 added a new function to read in both types and return the number
9339 read: this means that if none are read it will be an error. The
9340 DER versions of the certificate and CRL reader would always fail
9341 because it isn't possible to mix certificates and CRLs in DER format
9342 without choking one or the other routine. Changed this to just read
9343 a certificate: this is the best we can do. Also modified the code
9344 in apps/verify.c to take notice of return codes: it was previously
9345 attempting to read in certificates from NULL pointers and ignoring
9346 any errors: this is one reason why the cert and CRL reader seemed
9347 to work. It doesn't check return codes from the default certificate
9348 routines: these may well fail if the certificates aren't installed.
9351 *) Code to support otherName option in GeneralName.
9354 *) First update to verify code. Change the verify utility
9355 so it warns if it is passed a self signed certificate:
9356 for consistency with the normal behaviour. X509_verify
9357 has been modified to it will now verify a self signed
9358 certificate if *exactly* the same certificate appears
9359 in the store: it was previously impossible to trust a
9360 single self signed certificate. This means that:
9361 openssl verify ss.pem
9362 now gives a warning about a self signed certificate but
9363 openssl verify -CAfile ss.pem ss.pem
9367 *) For servers, store verify_result in SSL_SESSION data structure
9368 (and add it to external session representation).
9369 This is needed when client certificate verifications fails,
9370 but an application-provided verification callback (set by
9371 SSL_CTX_set_cert_verify_callback) allows accepting the session
9372 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9373 but returns 1): When the session is reused, we have to set
9374 ssl->verify_result to the appropriate error code to avoid
9376 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9378 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9379 case in PKCS7_dataInit() where the signed PKCS7 structure
9380 didn't contain any existing data because it was being created.
9381 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9383 *) Add a salt to the key derivation routines in enc.c. This
9384 forms the first 8 bytes of the encrypted file. Also add a
9385 -S option to allow a salt to be input on the command line.
9388 *) New function X509_cmp(). Oddly enough there wasn't a function
9389 to compare two certificates. We do this by working out the SHA1
9390 hash and comparing that. X509_cmp() will be needed by the trust
9394 *) SSL_get1_session() is like SSL_get_session(), but increments
9395 the reference count in the SSL_SESSION returned.
9396 [Geoff Thorpe <geoff@eu.c2.net>]
9398 *) Fix for 'req': it was adding a null to request attributes.
9399 Also change the X509_LOOKUP and X509_INFO code to handle
9400 certificate auxiliary information.
9403 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9407 *) Add the possibility to add extra information to the memory leak
9408 detecting output, to form tracebacks, showing from where each
9409 allocation was originated: CRYPTO_push_info("constant string") adds
9410 the string plus current file name and line number to a per-thread
9411 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9412 is like calling CYRPTO_pop_info() until the stack is empty.
9413 Also updated memory leak detection code to be multi-thread-safe.
9416 *) Add options -text and -noout to pkcs7 utility and delete the
9417 encryption options which never did anything. Update docs.
9420 *) Add options to some of the utilities to allow the pass phrase
9421 to be included on either the command line (not recommended on
9422 OSes like Unix) or read from the environment. Update the
9423 manpages and fix a few bugs.
9426 *) Add a few manpages for some of the openssl commands.
9429 *) Fix the -revoke option in ca. It was freeing up memory twice,
9430 leaking and not finding already revoked certificates.
9433 *) Extensive changes to support certificate auxiliary information.
9434 This involves the use of X509_CERT_AUX structure and X509_AUX
9435 functions. An X509_AUX function such as PEM_read_X509_AUX()
9436 can still read in a certificate file in the usual way but it
9437 will also read in any additional "auxiliary information". By
9438 doing things this way a fair degree of compatibility can be
9439 retained: existing certificates can have this information added
9440 using the new 'x509' options.
9442 Current auxiliary information includes an "alias" and some trust
9443 settings. The trust settings will ultimately be used in enhanced
9444 certificate chain verification routines: currently a certificate
9445 can only be trusted if it is self signed and then it is trusted
9449 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9450 The problem was that one of the replacement routines had not been working
9451 since SSLeay releases. For now the offending routine has been replaced
9452 with non-optimised assembler. Even so, this now gives around 95%
9453 performance improvement for 1024 bit RSA signs.
9456 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9457 handling. Most clients have the effective key size in bits equal to
9458 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9459 A few however don't do this and instead use the size of the decrypted key
9460 to determine the RC2 key length and the AlgorithmIdentifier to determine
9461 the effective key length. In this case the effective key length can still
9462 be 40 bits but the key length can be 168 bits for example. This is fixed
9463 by manually forcing an RC2 key into the EVP_PKEY structure because the
9464 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9465 the key length and effective key length are equal.
9468 *) Add a bunch of functions that should simplify the creation of
9469 X509_NAME structures. Now you should be able to do:
9470 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9471 and have it automatically work out the correct field type and fill in
9472 the structures. The more adventurous can try:
9473 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9474 and it will (hopefully) work out the correct multibyte encoding.
9477 *) Change the 'req' utility to use the new field handling and multibyte
9478 copy routines. Before the DN field creation was handled in an ad hoc
9479 way in req, ca, and x509 which was rather broken and didn't support
9480 BMPStrings or UTF8Strings. Since some software doesn't implement
9481 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9482 using the dirstring_type option. See the new comment in the default
9483 openssl.cnf for more info.
9486 *) Make crypto/rand/md_rand.c more robust:
9487 - Assure unique random numbers after fork().
9488 - Make sure that concurrent threads access the global counter and
9489 md serializably so that we never lose entropy in them
9490 or use exactly the same state in multiple threads.
9491 Access to the large state is not always serializable because
9492 the additional locking could be a performance killer, and
9493 md should be large enough anyway.
9496 *) New file apps/app_rand.c with commonly needed functionality
9497 for handling the random seed file.
9499 Use the random seed file in some applications that previously did not:
9501 dsaparam -genkey (which also ignored its '-rand' option),
9504 x509 (when signing).
9505 Except on systems with /dev/urandom, it is crucial to have a random
9506 seed file at least for key creation, DSA signing, and for DH exchanges;
9507 for RSA signatures we could do without one.
9509 gendh and gendsa (unlike genrsa) used to read only the first byte
9510 of each file listed in the '-rand' option. The function as previously
9511 found in genrsa is now in app_rand.c and is used by all programs
9512 that support '-rand'.
9515 *) In RAND_write_file, use mode 0600 for creating files;
9516 don't just chmod when it may be too late.
9519 *) Report an error from X509_STORE_load_locations
9520 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9523 *) New function ASN1_mbstring_copy() this copies a string in either
9524 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9525 into an ASN1_STRING type. A mask of permissible types is passed
9526 and it chooses the "minimal" type to use or an error if not type
9530 *) Add function equivalents to the various macros in asn1.h. The old
9531 macros are retained with an M_ prefix. Code inside the library can
9532 use the M_ macros. External code (including the openssl utility)
9533 should *NOT* in order to be "shared library friendly".
9536 *) Add various functions that can check a certificate's extensions
9537 to see if it usable for various purposes such as SSL client,
9538 server or S/MIME and CAs of these types. This is currently
9539 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9540 verification. Also added a -purpose flag to x509 utility to
9541 print out all the purposes.
9544 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9548 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9549 for, obtain and decode and extension and obtain its critical flag.
9550 This allows all the necessary extension code to be handled in a
9551 single function call.
9554 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9555 platforms. See crypto/rc4/rc4_enc.c for further details.
9558 *) New -noout option to asn1parse. This causes no output to be produced
9559 its main use is when combined with -strparse and -out to extract data
9560 from a file (which may not be in ASN.1 format).
9563 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9564 when producing the local key id.
9565 [Richard Levitte <levitte@stacken.kth.se>]
9567 *) New option -dhparam in s_server. This allows a DH parameter file to be
9568 stated explicitly. If it is not stated then it tries the first server
9569 certificate file. The previous behaviour hard coded the filename
9573 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9574 a public key to be input or output. For example:
9575 openssl rsa -in key.pem -pubout -out pubkey.pem
9576 Also added necessary DSA public key functions to handle this.
9579 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9580 in the message. This was handled by allowing
9581 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9582 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9584 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9585 to the end of the strings whereas this didn't. This would cause problems
9586 if strings read with d2i_ASN1_bytes() were later modified.
9587 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9589 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9590 data and it contains EOF it will end up returning an error. This is
9591 caused by input 46 bytes long. The cause is due to the way base64
9592 BIOs find the start of base64 encoded data. They do this by trying a
9593 trial decode on each line until they find one that works. When they
9594 do a flag is set and it starts again knowing it can pass all the
9595 data directly through the decoder. Unfortunately it doesn't reset
9596 the context it uses. This means that if EOF is reached an attempt
9597 is made to pass two EOFs through the context and this causes the
9598 resulting error. This can also cause other problems as well. As is
9599 usual with these problems it takes *ages* to find and the fix is
9600 trivial: move one line.
9601 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9603 *) Ugly workaround to get s_client and s_server working under Windows. The
9604 old code wouldn't work because it needed to select() on sockets and the
9605 tty (for keypresses and to see if data could be written). Win32 only
9606 supports select() on sockets so we select() with a 1s timeout on the
9607 sockets and then see if any characters are waiting to be read, if none
9608 are present then we retry, we also assume we can always write data to
9609 the tty. This isn't nice because the code then blocks until we've
9610 received a complete line of data and it is effectively polling the
9611 keyboard at 1s intervals: however it's quite a bit better than not
9612 working at all :-) A dedicated Windows application might handle this
9613 with an event loop for example.
9616 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9617 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9618 will be called when RSA_sign() and RSA_verify() are used. This is useful
9619 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9620 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9621 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9622 This necessitated the support of an extra signature type NID_md5_sha1
9623 for SSL signatures and modifications to the SSL library to use it instead
9624 of calling RSA_public_decrypt() and RSA_private_encrypt().
9627 *) Add new -verify -CAfile and -CApath options to the crl program, these
9628 will lookup a CRL issuers certificate and verify the signature in a
9629 similar way to the verify program. Tidy up the crl program so it
9630 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9631 less strict. It will now permit CRL extensions even if it is not
9632 a V2 CRL: this will allow it to tolerate some broken CRLs.
9635 *) Initialize all non-automatic variables each time one of the openssl
9636 sub-programs is started (this is necessary as they may be started
9637 multiple times from the "OpenSSL>" prompt).
9638 [Lennart Bang, Bodo Moeller]
9640 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9641 removing all other RSA functionality (this is what NO_RSA does). This
9642 is so (for example) those in the US can disable those operations covered
9643 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9647 *) Non-copying interface to BIO pairs.
9648 (still largely untested)
9651 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9652 ASCII string. This was handled independently in various places before.
9655 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9656 UTF8 strings a character at a time.
9659 *) Use client_version from client hello to select the protocol
9660 (s23_srvr.c) and for RSA client key exchange verification
9661 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9664 *) Add various utility functions to handle SPKACs, these were previously
9665 handled by poking round in the structure internals. Added new function
9666 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9667 print, verify and generate SPKACs. Based on an original idea from
9668 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9671 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9674 *) Allow the config file extension section to be overwritten on the
9675 command line. Based on an original idea from Massimiliano Pala
9676 <madwolf@comune.modena.it>. The new option is called -extensions
9677 and can be applied to ca, req and x509. Also -reqexts to override
9678 the request extensions in req and -crlexts to override the crl extensions
9682 *) Add new feature to the SPKAC handling in ca. Now you can include
9683 the same field multiple times by preceding it by "XXXX." for example:
9686 this is the same syntax as used in the req config file.
9689 *) Allow certificate extensions to be added to certificate requests. These
9690 are specified in a 'req_extensions' option of the req section of the
9691 config file. They can be printed out with the -text option to req but
9692 are otherwise ignored at present.
9695 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9696 data read consists of only the final block it would not decrypted because
9697 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9698 A misplaced 'break' also meant the decrypted final block might not be
9699 copied until the next read.
9702 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9703 a few extra parameters to the DH structure: these will be useful if
9704 for example we want the value of 'q' or implement X9.42 DH.
9707 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9708 provides hooks that allow the default DSA functions or functions on a
9709 "per key" basis to be replaced. This allows hardware acceleration and
9710 hardware key storage to be handled without major modification to the
9711 library. Also added low level modexp hooks and CRYPTO_EX structure and
9712 associated functions.
9715 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9716 as "read only": it can't be written to and the buffer it points to will
9717 not be freed. Reading from a read only BIO is much more efficient than
9718 a normal memory BIO. This was added because there are several times when
9719 an area of memory needs to be read from a BIO. The previous method was
9720 to create a memory BIO and write the data to it, this results in two
9721 copies of the data and an O(n^2) reading algorithm. There is a new
9722 function BIO_new_mem_buf() which creates a read only memory BIO from
9723 an area of memory. Also modified the PKCS#7 routines to use read only
9727 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9728 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9729 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9730 but a retry condition occured while trying to read the rest.
9733 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9734 NID_pkcs7_encrypted by default: this was wrong since this should almost
9735 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9736 the encrypted data type: this is a more sensible place to put it and it
9737 allows the PKCS#12 code to be tidied up that duplicated this
9741 *) Changed obj_dat.pl script so it takes its input and output files on
9742 the command line. This should avoid shell escape redirection problems
9746 *) Initial support for certificate extension requests, these are included
9747 in things like Xenroll certificate requests. Included functions to allow
9748 extensions to be obtained and added.
9751 *) -crlf option to s_client and s_server for sending newlines as
9752 CRLF (as required by many protocols).
9755 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9757 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9758 [Ralf S. Engelschall]
9760 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9761 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9763 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9767 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9768 DH parameters/keys (q is lost during that conversion, but the resulting
9769 DH parameters contain its length).
9771 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9772 much faster than DH_generate_parameters (which creates parameters
9773 where p = 2*q + 1), and also the smaller q makes DH computations
9774 much more efficient (160-bit exponentiation instead of 1024-bit
9775 exponentiation); so this provides a convenient way to support DHE
9776 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9777 utter importance to use
9778 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9780 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9781 when such DH parameters are used, because otherwise small subgroup
9782 attacks may become possible!
9785 *) Avoid memory leak in i2d_DHparams.
9788 *) Allow the -k option to be used more than once in the enc program:
9789 this allows the same encrypted message to be read by multiple recipients.
9792 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9793 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9794 it will always use the numerical form of the OID, even if it has a short
9798 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9799 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9800 otherwise bn_mod_exp was called. In the case of hardware keys for example
9801 no private key components need be present and it might store extra data
9802 in the RSA structure, which cannot be accessed from bn_mod_exp.
9803 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9804 private key operations.
9807 *) Added support for SPARC Linux.
9810 *) pem_password_cb function type incompatibly changed from
9811 typedef int pem_password_cb(char *buf, int size, int rwflag);
9813 ....(char *buf, int size, int rwflag, void *userdata);
9814 so that applications can pass data to their callbacks:
9815 The PEM[_ASN1]_{read,write}... functions and macros now take an
9816 additional void * argument, which is just handed through whenever
9817 the password callback is called.
9818 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9820 New function SSL_CTX_set_default_passwd_cb_userdata.
9822 Compatibility note: As many C implementations push function arguments
9823 onto the stack in reverse order, the new library version is likely to
9824 interoperate with programs that have been compiled with the old
9825 pem_password_cb definition (PEM_whatever takes some data that
9826 happens to be on the stack as its last argument, and the callback
9827 just ignores this garbage); but there is no guarantee whatsoever that
9830 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9831 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9832 problems not only on Windows, but also on some Unix platforms.
9833 To avoid problematic command lines, these definitions are now in an
9834 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9835 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9838 *) MIPS III/IV assembler module is reimplemented.
9841 *) More DES library cleanups: remove references to srand/rand and
9842 delete an unused file.
9845 *) Add support for the the free Netwide assembler (NASM) under Win32,
9846 since not many people have MASM (ml) and it can be hard to obtain.
9847 This is currently experimental but it seems to work OK and pass all
9848 the tests. Check out INSTALL.W32 for info.
9851 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9852 without temporary keys kept an extra copy of the server key,
9853 and connections with temporary keys did not free everything in case
9857 *) New function RSA_check_key and new openssl rsa option -check
9858 for verifying the consistency of RSA keys.
9859 [Ulf Moeller, Bodo Moeller]
9861 *) Various changes to make Win32 compile work:
9862 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9863 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9864 comparison" warnings.
9865 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9868 *) Add a debugging option to PKCS#5 v2 key generation function: when
9869 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9870 derived keys are printed to stderr.
9873 *) Copy the flags in ASN1_STRING_dup().
9874 [Roman E. Pavlov <pre@mo.msk.ru>]
9876 *) The x509 application mishandled signing requests containing DSA
9877 keys when the signing key was also DSA and the parameters didn't match.
9879 It was supposed to omit the parameters when they matched the signing key:
9880 the verifying software was then supposed to automatically use the CA's
9881 parameters if they were absent from the end user certificate.
9883 Omitting parameters is no longer recommended. The test was also
9884 the wrong way round! This was probably due to unusual behaviour in
9885 EVP_cmp_parameters() which returns 1 if the parameters match.
9886 This meant that parameters were omitted when they *didn't* match and
9887 the certificate was useless. Certificates signed with 'ca' didn't have
9889 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9891 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9892 The interface is as follows:
9893 Applications can use
9894 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9895 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9896 "off" is now the default.
9897 The library internally uses
9898 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9899 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9900 to disable memory-checking temporarily.
9902 Some inconsistent states that previously were possible (and were
9903 even the default) are now avoided.
9905 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9906 with each memory chunk allocated; this is occasionally more helpful
9907 than just having a counter.
9909 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9911 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9915 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9916 which largely parallels "options", but is for changing API behaviour,
9917 whereas "options" are about protocol behaviour.
9918 Initial "mode" flags are:
9920 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9921 a single record has been written.
9922 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9923 retries use the same buffer location.
9924 (But all of the contents must be
9928 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9931 *) Fix problems with no-hmac etc.
9932 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9934 *) New functions RSA_get_default_method(), RSA_set_method() and
9935 RSA_get_method(). These allows replacement of RSA_METHODs without having
9936 to mess around with the internals of an RSA structure.
9939 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9940 Also really enable memory leak checks in openssl.c and in some
9942 [Chad C. Mulligan, Bodo Moeller]
9944 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9945 up the length of negative integers. This has now been simplified to just
9946 store the length when it is first determined and use it later, rather
9947 than trying to keep track of where data is copied and updating it to
9949 [Steve Henson, reported by Brien Wheeler
9950 <bwheeler@authentica-security.com>]
9952 *) Add a new function PKCS7_signatureVerify. This allows the verification
9953 of a PKCS#7 signature but with the signing certificate passed to the
9954 function itself. This contrasts with PKCS7_dataVerify which assumes the
9955 certificate is present in the PKCS#7 structure. This isn't always the
9956 case: certificates can be omitted from a PKCS#7 structure and be
9957 distributed by "out of band" means (such as a certificate database).
9960 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9961 function prototypes in pem.h, also change util/mkdef.pl to add the
9962 necessary function names.
9965 *) mk1mf.pl (used by Windows builds) did not properly read the
9966 options set by Configure in the top level Makefile, and Configure
9967 was not even able to write more than one option correctly.
9968 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9971 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9972 file to be loaded from a BIO or FILE pointer. The BIO version will
9973 for example allow memory BIOs to contain config info.
9976 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9977 Whoever hopes to achieve shared-library compatibility across versions
9978 must use this, not the compile-time macro.
9979 (Exercise 0.9.4: Which is the minimum library version required by
9981 Note: All this applies only to multi-threaded programs, others don't
9985 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9986 through a BIO pair triggered the default case, i.e.
9987 SSLerr(...,SSL_R_UNKNOWN_STATE).
9990 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9991 can use the SSL library even if none of the specific BIOs is
9995 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9996 for the encoded length.
9997 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9999 *) Add initial documentation of the X509V3 functions.
10002 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10003 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10004 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10005 secure PKCS#8 private key format with a high iteration count.
10008 *) Fix determination of Perl interpreter: A perl or perl5
10009 _directory_ in $PATH was also accepted as the interpreter.
10010 [Ralf S. Engelschall]
10012 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10013 wrong with it but it was very old and did things like calling
10014 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10015 unusual formatting.
10018 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10019 to use the new extension code.
10022 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10023 with macros. This should make it easier to change their form, add extra
10024 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10028 *) Add to configuration table a new entry that can specify an alternative
10029 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10030 according to Mark Crispin <MRC@Panda.COM>.
10034 *) DES CBC did not update the IV. Weird.
10037 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10038 Changing the behaviour of the former might break existing programs --
10039 where IV updating is needed, des_ncbc_encrypt can be used.
10042 *) When bntest is run from "make test" it drives bc to check its
10043 calculations, as well as internally checking them. If an internal check
10044 fails, it needs to cause bc to give a non-zero result or make test carries
10045 on without noticing the failure. Fixed.
10048 *) DES library cleanups.
10051 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10052 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10053 ciphers. NOTE: although the key derivation function has been verified
10054 against some published test vectors it has not been extensively tested
10055 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10059 *) Instead of "mkdir -p", which is not fully portable, use new
10060 Perl script "util/mkdir-p.pl".
10063 *) Rewrite the way password based encryption (PBE) is handled. It used to
10064 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10065 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10066 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10067 the 'parameter' field of the AlgorithmIdentifier is passed to the
10068 underlying key generation function so it must do its own ASN1 parsing.
10069 This has also changed the EVP_PBE_CipherInit() function which now has a
10070 'parameter' argument instead of literal salt and iteration count values
10071 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10074 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10075 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10076 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10077 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10078 value was just used as a "magic string" and not used directly its
10079 value doesn't matter.
10082 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10086 *) "linux-sparc64" configuration (ultrapenguin).
10087 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10088 "linux-sparc" configuration.
10089 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10091 *) config now generates no-xxx options for missing ciphers.
10094 *) Support the EBCDIC character set (work in progress).
10095 File ebcdic.c not yet included because it has a different license.
10096 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10098 *) Support BS2000/OSD-POSIX.
10099 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10101 *) Make callbacks for key generation use void * instead of char *.
10104 *) Make S/MIME samples compile (not yet tested).
10107 *) Additional typesafe stacks.
10110 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10114 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10116 *) New configuration variant "sco5-gcc".
10118 *) Updated some demos.
10119 [Sean O Riordain, Wade Scholine]
10121 *) Add missing BIO_free at exit of pkcs12 application.
10124 *) Fix memory leak in conf.c.
10127 *) Updates for Win32 to assembler version of MD5.
10130 *) Set #! path to perl in apps/der_chop to where we found it
10131 instead of using a fixed path.
10134 *) SHA library changes for irix64-mips4-cc.
10137 *) Improvements for VMS support.
10141 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10143 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10144 This also avoids the problems with SC4.2 and unpatched SC5.
10145 [Andy Polyakov <appro@fy.chalmers.se>]
10147 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10148 These are required because of the typesafe stack would otherwise break
10149 existing code. If old code used a structure member which used to be STACK
10150 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10151 sk_num or sk_value it would produce an error because the num, data members
10152 are not present in STACK_OF. Now it just produces a warning. sk_set
10153 replaces the old method of assigning a value to sk_value
10154 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10155 that does this will no longer work (and should use sk_set instead) but
10156 this could be regarded as a "questionable" behaviour anyway.
10159 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10160 correctly handle encrypted S/MIME data.
10163 *) Change type of various DES function arguments from des_cblock
10164 (which means, in function argument declarations, pointer to char)
10165 to des_cblock * (meaning pointer to array with 8 char elements),
10166 which allows the compiler to do more typechecking; it was like
10167 that back in SSLeay, but with lots of ugly casts.
10169 Introduce new type const_des_cblock.
10172 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10173 problems: find RecipientInfo structure that matches recipient certificate
10174 and initialise the ASN1 structures properly based on passed cipher.
10177 *) Belatedly make the BN tests actually check the results.
10180 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10181 to and from BNs: it was completely broken. New compilation option
10182 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10183 key elements as negative integers.
10186 *) Reorganize and speed up MD5.
10187 [Andy Polyakov <appro@fy.chalmers.se>]
10190 [Richard Levitte <richard@levitte.org>]
10192 *) New option -out to asn1parse to allow the parsed structure to be
10193 output to a file. This is most useful when combined with the -strparse
10194 option to examine the output of things like OCTET STRINGS.
10197 *) Make SSL library a little more fool-proof by not requiring any longer
10198 that SSL_set_{accept,connect}_state be called before
10199 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10200 in many applications because usually everything *appeared* to work as
10201 intended anyway -- now it really works as intended).
10204 *) Move openssl.cnf out of lib/.
10207 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10208 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10209 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10210 [Ralf S. Engelschall]
10212 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10213 handle PKCS#7 enveloped data properly.
10214 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10216 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10217 copying pointers. The cert_st handling is changed by this in
10218 various ways (and thus what used to be known as ctx->default_cert
10219 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10220 any longer when s->cert does not give us what we need).
10221 ssl_cert_instantiate becomes obsolete by this change.
10222 As soon as we've got the new code right (possibly it already is?),
10223 we have solved a couple of bugs of the earlier code where s->cert
10224 was used as if it could not have been shared with other SSL structures.
10226 Note that using the SSL API in certain dirty ways now will result
10227 in different behaviour than observed with earlier library versions:
10228 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10229 does not influence s as it used to.
10231 In order to clean up things more thoroughly, inside SSL_SESSION
10232 we don't use CERT any longer, but a new structure SESS_CERT
10233 that holds per-session data (if available); currently, this is
10234 the peer's certificate chain and, for clients, the server's certificate
10235 and temporary key. CERT holds only those values that can have
10236 meaningful defaults in an SSL_CTX.
10239 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10240 from the internal representation. Various PKCS#7 fixes: remove some
10241 evil casts and set the enc_dig_alg field properly based on the signing
10245 *) Allow PKCS#12 password to be set from the command line or the
10246 environment. Let 'ca' get its config file name from the environment
10247 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10251 *) Allow certificate policies extension to use an IA5STRING for the
10252 organization field. This is contrary to the PKIX definition but
10253 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10257 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10258 without disallowing inline assembler and the like for non-pedantic builds.
10261 *) Support Borland C++ builder.
10262 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10264 *) Support Mingw32.
10267 *) SHA-1 cleanups and performance enhancements.
10268 [Andy Polyakov <appro@fy.chalmers.se>]
10270 *) Sparc v8plus assembler for the bignum library.
10271 [Andy Polyakov <appro@fy.chalmers.se>]
10273 *) Accept any -xxx and +xxx compiler options in Configure.
10276 *) Update HPUX configuration.
10279 *) Add missing sk_<type>_unshift() function to safestack.h
10280 [Ralf S. Engelschall]
10282 *) New function SSL_CTX_use_certificate_chain_file that sets the
10283 "extra_cert"s in addition to the certificate. (This makes sense
10284 only for "PEM" format files, as chains as a whole are not
10288 *) Support verify_depth from the SSL API.
10289 x509_vfy.c had what can be considered an off-by-one-error:
10290 Its depth (which was not part of the external interface)
10291 was actually counting the number of certificates in a chain;
10292 now it really counts the depth.
10295 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10296 instead of X509err, which often resulted in confusing error
10297 messages since the error codes are not globally unique
10298 (e.g. an alleged error in ssl3_accept when a certificate
10299 didn't match the private key).
10301 *) New function SSL_CTX_set_session_id_context that allows to set a default
10302 value (so that you don't need SSL_set_session_id_context for each
10303 connection using the SSL_CTX).
10306 *) OAEP decoding bug fix.
10309 *) Support INSTALL_PREFIX for package builders, as proposed by
10313 *) New Configure options "threads" and "no-threads". For systems
10314 where the proper compiler options are known (currently Solaris
10315 and Linux), "threads" is the default.
10318 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10321 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10322 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10323 such as /usr/local/bin.
10326 *) "make linux-shared" to build shared libraries.
10327 [Niels Poppe <niels@netbox.org>]
10329 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10332 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10333 extension adding in x509 utility.
10336 *) Remove NOPROTO sections and error code comments.
10339 *) Partial rewrite of the DEF file generator to now parse the ANSI
10343 *) New Configure options --prefix=DIR and --openssldir=DIR.
10346 *) Complete rewrite of the error code script(s). It is all now handled
10347 by one script at the top level which handles error code gathering,
10348 header rewriting and C source file generation. It should be much better
10349 than the old method: it now uses a modified version of Ulf's parser to
10350 read the ANSI prototypes in all header files (thus the old K&R definitions
10351 aren't needed for error creation any more) and do a better job of
10352 translating function codes into names. The old 'ASN1 error code imbedded
10353 in a comment' is no longer necessary and it doesn't use .err files which
10354 have now been deleted. Also the error code call doesn't have to appear all
10355 on one line (which resulted in some large lines...).
10358 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10361 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10362 0 (which usually indicates a closed connection), but continue reading.
10365 *) Fix some race conditions.
10368 *) Add support for CRL distribution points extension. Add Certificate
10369 Policies and CRL distribution points documentation.
10372 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10375 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10376 8 of keying material. Merlin has also confirmed interop with this fix
10377 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10378 [Merlin Hughes <merlin@baltimore.ie>]
10380 *) Fix lots of warnings.
10381 [Richard Levitte <levitte@stacken.kth.se>]
10383 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10384 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10385 [Richard Levitte <levitte@stacken.kth.se>]
10387 *) Fix problems with sizeof(long) == 8.
10388 [Andy Polyakov <appro@fy.chalmers.se>]
10390 *) Change functions to ANSI C.
10393 *) Fix typos in error codes.
10394 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10396 *) Remove defunct assembler files from Configure.
10399 *) SPARC v8 assembler BIGNUM implementation.
10400 [Andy Polyakov <appro@fy.chalmers.se>]
10402 *) Support for Certificate Policies extension: both print and set.
10403 Various additions to support the r2i method this uses.
10406 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10407 return a const string when you are expecting an allocated buffer.
10410 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10411 types DirectoryString and DisplayText.
10414 *) Add code to allow r2i extensions to access the configuration database,
10415 add an LHASH database driver and add several ctx helper functions.
10418 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10419 fail when they extended the size of a BIGNUM.
10422 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10423 support typesafe stack.
10426 *) Fix typo in SSL_[gs]et_options().
10427 [Nils Frostberg <nils@medcom.se>]
10429 *) Delete various functions and files that belonged to the (now obsolete)
10430 old X509V3 handling code.
10433 *) New Configure option "rsaref".
10436 *) Don't auto-generate pem.h.
10439 *) Introduce type-safe ASN.1 SETs.
10442 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10443 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10445 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10446 that links with OpenSSL (well at least cause lots of warnings), but fear
10447 not: the conversion is trivial, and it eliminates loads of evil casts. A
10448 few STACKed things have been converted already. Feel free to convert more.
10449 In the fullness of time, I'll do away with the STACK type altogether.
10452 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10453 specified in <certfile> by updating the entry in the index.txt file.
10454 This way one no longer has to edit the index.txt file manually for
10455 revoking a certificate. The -revoke option does the gory details now.
10456 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10458 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10459 `-text' option at all and this way the `-noout -text' combination was
10460 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10461 [Ralf S. Engelschall]
10463 *) Make sure a corresponding plain text error message exists for the
10464 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10465 verify callback function determined that a certificate was revoked.
10466 [Ralf S. Engelschall]
10468 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10469 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10470 all available cipers including rc5, which was forgotten until now.
10471 In order to let the testing shell script know which algorithms
10472 are available, a new (up to now undocumented) command
10473 "openssl list-cipher-commands" is used.
10476 *) Bugfix: s_client occasionally would sleep in select() when
10477 it should have checked SSL_pending() first.
10480 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10481 the raw DSA values prior to ASN.1 encoding.
10484 *) Tweaks to Configure
10485 [Niels Poppe <niels@netbox.org>]
10487 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10491 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10494 *) New config option to avoid instructions that are illegal on the 80386.
10495 The default code is faster, but requires at least a 486.
10498 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10499 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10500 same as SSL2_VERSION anyway.
10503 *) New "-showcerts" option for s_client.
10506 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10507 application. Various cleanups and fixes.
10510 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10511 modify error routines to work internally. Add error codes and PBE init
10512 to library startup routines.
10515 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10516 packing functions to asn1 and evp. Changed function names and error
10517 codes along the way.
10520 *) PKCS12 integration: and so it begins... First of several patches to
10521 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10522 objects to objects.h
10525 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10526 and display support for Thawte strong extranet extension.
10529 *) Add LinuxPPC support.
10530 [Jeff Dubrule <igor@pobox.org>]
10532 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10533 bn_div_words in alpha.s.
10534 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10536 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10537 OAEP isn't supported when OpenSSL is built with RSAref.
10538 [Ulf Moeller <ulf@fitug.de>]
10540 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10541 so they no longer are missing under -DNOPROTO.
10542 [Soren S. Jorvang <soren@t.dk>]
10545 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10547 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10548 doesn't work when the session is reused. Coming soon!
10551 *) Fix a security hole, that allows sessions to be reused in the wrong
10552 context thus bypassing client cert protection! All software that uses
10553 client certs and session caches in multiple contexts NEEDS PATCHING to
10554 allow session reuse! A fuller solution is in the works.
10555 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10557 *) Some more source tree cleanups (removed obsolete files
10558 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10559 permission on "config" script to be executable) and a fix for the INSTALL
10561 [Ulf Moeller <ulf@fitug.de>]
10563 *) Remove some legacy and erroneous uses of malloc, free instead of
10565 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10567 *) Make rsa_oaep_test return non-zero on error.
10568 [Ulf Moeller <ulf@fitug.de>]
10570 *) Add support for native Solaris shared libraries. Configure
10571 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10572 if someone would make that last step automatic.
10573 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10575 *) ctx_size was not built with the right compiler during "make links". Fixed.
10578 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10579 except NULL ciphers". This means the default cipher list will no longer
10580 enable NULL ciphers. They need to be specifically enabled e.g. with
10581 the string "DEFAULT:eNULL".
10584 *) Fix to RSA private encryption routines: if p < q then it would
10585 occasionally produce an invalid result. This will only happen with
10586 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10589 *) Be less restrictive and allow also `perl util/perlpath.pl
10590 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10591 because this way one can also use an interpreter named `perl5' (which is
10592 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10593 installed as `perl').
10594 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10596 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10597 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10599 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10600 advapi32.lib to Win32 build and change the pem test comparison
10601 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10602 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10603 and crypto/des/ede_cbcm_enc.c.
10606 *) DES quad checksum was broken on big-endian architectures. Fixed.
10609 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10610 Win32 test batch file so it (might) work again. The Win32 test batch file
10611 is horrible: I feel ill....
10614 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10615 in e_os.h. Audit of header files to check ANSI and non ANSI
10616 sections: 10 functions were absent from non ANSI section and not exported
10617 from Windows DLLs. Fixed up libeay.num for new functions.
10620 *) Make `openssl version' output lines consistent.
10621 [Ralf S. Engelschall]
10623 *) Fix Win32 symbol export lists for BIO functions: Added
10624 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10625 to ms/libeay{16,32}.def.
10626 [Ralf S. Engelschall]
10628 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10629 fine under Unix and passes some trivial tests I've now added. But the
10630 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10631 added to make sure no one expects that this stuff really works in the
10632 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10633 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10635 [Ralf S. Engelschall]
10637 *) Fix the generation of two part addresses in perl.
10638 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10640 *) Add config entry for Linux on MIPS.
10641 [John Tobey <jtobey@channel1.com>]
10643 *) Make links whenever Configure is run, unless we are on Windoze.
10646 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10647 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10651 *) Add a useful kludge to allow package maintainers to specify compiler and
10652 other platforms details on the command line without having to patch the
10653 Configure script everytime: One now can use ``perl Configure
10654 <id>:<details>'', i.e. platform ids are allowed to have details appended
10655 to them (separated by colons). This is treated as there would be a static
10656 pre-configured entry in Configure's %table under key <id> with value
10657 <details> and ``perl Configure <id>'' is called. So, when you want to
10658 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10659 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10660 now, which overrides the FreeBSD-elf entry on-the-fly.
10661 [Ralf S. Engelschall]
10663 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10666 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10667 on the `perl Configure ...' command line. This way one can compile
10668 OpenSSL libraries with Position Independent Code (PIC) which is needed
10669 for linking it into DSOs.
10670 [Ralf S. Engelschall]
10672 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10676 *) Cleaned up the LICENSE document: The official contact for any license
10677 questions now is the OpenSSL core team under openssl-core@openssl.org.
10678 And add a paragraph about the dual-license situation to make sure people
10679 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10680 to the OpenSSL toolkit.
10681 [Ralf S. Engelschall]
10683 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10684 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10685 Additionally cleaned up the `make links' target: Remove unnecessary
10686 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10687 to speed processing and no longer clutter the display with confusing
10688 stuff. Instead only the actually done links are displayed.
10689 [Ralf S. Engelschall]
10691 *) Permit null encryption ciphersuites, used for authentication only. It used
10692 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10693 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10697 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10698 signed attributes when verifying signatures (this would break them),
10699 the detached data encoding was wrong and public keys obtained using
10700 X509_get_pubkey() weren't freed.
10703 *) Add text documentation for the BUFFER functions. Also added a work around
10704 to a Win95 console bug. This was triggered by the password read stuff: the
10705 last character typed gets carried over to the next fread(). If you were
10706 generating a new cert request using 'req' for example then the last
10707 character of the passphrase would be CR which would then enter the first
10711 *) Added the new `Includes OpenSSL Cryptography Software' button as
10712 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10713 button and can be used by applications based on OpenSSL to show the
10714 relationship to the OpenSSL project.
10715 [Ralf S. Engelschall]
10717 *) Remove confusing variables in function signatures in files
10718 ssl/ssl_lib.c and ssl/ssl.h.
10719 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10721 *) Don't install bss_file.c under PREFIX/include/
10722 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10724 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10725 functions that return function pointers and has support for NT specific
10726 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10727 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10728 unsigned to signed types: this was killing the Win32 compile.
10731 *) Add new certificate file to stack functions,
10732 SSL_add_dir_cert_subjects_to_stack() and
10733 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10734 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10735 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10736 This means that Apache-SSL and similar packages don't have to mess around
10737 to add as many CAs as they want to the preferred list.
10740 *) Experiment with doxygen documentation. Currently only partially applied to
10742 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10743 openssl.doxy as the configuration file.
10746 *) Get rid of remaining C++-style comments which strict C compilers hate.
10747 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10749 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10750 compiled in by default: it has problems with large keys.
10753 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10754 DH private keys and/or callback functions which directly correspond to
10755 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10756 is needed for applications which have to configure certificates on a
10757 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10759 For the RSA certificate situation is makes no difference, but
10760 for the DSA certificate situation this fixes the "no shared cipher"
10761 problem where the OpenSSL cipher selection procedure failed because the
10762 temporary keys were not overtaken from the context and the API provided
10763 no way to reconfigure them.
10764 The new functions now let applications reconfigure the stuff and they
10765 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10766 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10767 non-public-API function ssl_cert_instantiate() is used as a helper
10768 function and also to reduce code redundancy inside ssl_rsa.c.
10769 [Ralf S. Engelschall]
10771 *) Move s_server -dcert and -dkey options out of the undocumented feature
10772 area because they are useful for the DSA situation and should be
10773 recognized by the users.
10774 [Ralf S. Engelschall]
10776 *) Fix the cipher decision scheme for export ciphers: the export bits are
10777 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10778 SSL_EXP_MASK. So, the original variable has to be used instead of the
10779 already masked variable.
10780 [Richard Levitte <levitte@stacken.kth.se>]
10782 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10783 [Richard Levitte <levitte@stacken.kth.se>]
10785 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10786 from `int' to `unsigned int' because it's a length and initialized by
10787 EVP_DigestFinal() which expects an `unsigned int *'.
10788 [Richard Levitte <levitte@stacken.kth.se>]
10790 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10791 script. Instead use the usual Shell->Perl transition trick.
10792 [Ralf S. Engelschall]
10794 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10795 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10796 -noout -modulus' as it's already the case for `openssl rsa -noout
10797 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10798 currently the public key is printed (a decision which was already done by
10799 `openssl dsa -modulus' in the past) which serves a similar purpose.
10800 Additionally the NO_RSA no longer completely removes the whole -modulus
10801 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10803 [Ralf S. Engelschall]
10805 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10806 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10807 [Arne Ansper <arne@ats.cyber.ee>]
10809 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10810 to be added. Now both 'req' and 'ca' can use new objects defined in the
10814 *) Add cool BIO that does syslog (or event log on NT).
10815 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10817 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10818 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10819 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10820 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10823 *) Add preliminary config info for new extension code.
10826 *) Make RSA_NO_PADDING really use no padding.
10827 [Ulf Moeller <ulf@fitug.de>]
10829 *) Generate errors when private/public key check is done.
10832 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10833 for some CRL extensions and new objects added.
10836 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10837 key usage extension and fuller support for authority key id.
10840 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10841 padding method for RSA, which is recommended for new applications in PKCS
10842 #1 v2.0 (RFC 2437, October 1998).
10843 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10844 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10845 against Bleichbacher's attack on RSA.
10846 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10849 *) Updates to the new SSL compression code
10850 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10852 *) Fix so that the version number in the master secret, when passed
10853 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10854 (because the server will not accept higher), that the version number
10855 is 0x03,0x01, not 0x03,0x00
10856 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10858 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10859 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10860 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10863 *) Support for RAW extensions where an arbitrary extension can be
10864 created by including its DER encoding. See apps/openssl.cnf for
10868 *) Make sure latest Perl versions don't interpret some generated C array
10869 code as Perl array code in the crypto/err/err_genc.pl script.
10870 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10872 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10873 not many people have the assembler. Various Win32 compilation fixes and
10874 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10875 build instructions.
10878 *) Modify configure script 'Configure' to automatically create crypto/date.h
10879 file under Win32 and also build pem.h from pem.org. New script
10880 util/mkfiles.pl to create the MINFO file on environments that can't do a
10881 'make files': perl util/mkfiles.pl >MINFO should work.
10884 *) Major rework of DES function declarations, in the pursuit of correctness
10885 and purity. As a result, many evil casts evaporated, and some weirdness,
10886 too. You may find this causes warnings in your code. Zapping your evil
10887 casts will probably fix them. Mostly.
10890 *) Fix for a typo in asn1.h. Bug fix to object creation script
10891 obj_dat.pl. It considered a zero in an object definition to mean
10892 "end of object": none of the objects in objects.h have any zeros
10893 so it wasn't spotted.
10894 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10896 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10897 Masking (CBCM). In the absence of test vectors, the best I have been able
10898 to do is check that the decrypt undoes the encrypt, so far. Send me test
10899 vectors if you have them.
10902 *) Correct calculation of key length for export ciphers (too much space was
10903 allocated for null ciphers). This has not been tested!
10906 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10907 message is now correct (it understands "crypto" and "ssl" on its
10908 command line). There is also now an "update" option. This will update
10909 the util/ssleay.num and util/libeay.num files with any new functions.
10911 perl util/mkdef.pl crypto ssl update
10912 it will update them.
10915 *) Overhauled the Perl interface (perl/*):
10916 - ported BN stuff to OpenSSL's different BN library
10917 - made the perl/ source tree CVS-aware
10918 - renamed the package from SSLeay to OpenSSL (the files still contain
10919 their history because I've copied them in the repository)
10920 - removed obsolete files (the test scripts will be replaced
10921 by better Test::Harness variants in the future)
10922 [Ralf S. Engelschall]
10924 *) First cut for a very conservative source tree cleanup:
10925 1. merge various obsolete readme texts into doc/ssleay.txt
10926 where we collect the old documents and readme texts.
10927 2. remove the first part of files where I'm already sure that we no
10928 longer need them because of three reasons: either they are just temporary
10929 files which were left by Eric or they are preserved original files where
10930 I've verified that the diff is also available in the CVS via "cvs diff
10931 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10932 the crypto/md/ stuff).
10933 [Ralf S. Engelschall]
10935 *) More extension code. Incomplete support for subject and issuer alt
10936 name, issuer and authority key id. Change the i2v function parameters
10937 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10938 what that's for :-) Fix to ASN1 macro which messed up
10939 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10942 *) Preliminary support for ENUMERATED type. This is largely copied from the
10946 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10947 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10949 *) Make sure `make rehash' target really finds the `openssl' program.
10950 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10952 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10953 like to hear about it if this slows down other processors.
10956 *) Add CygWin32 platform information to Configure script.
10957 [Alan Batie <batie@aahz.jf.intel.com>]
10959 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10960 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10962 *) New program nseq to manipulate netscape certificate sequences
10965 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10969 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10970 but the BN code had some problems that would cause failures when
10971 doing certificate verification and some other functions.
10972 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10974 *) Add ASN1 and PEM code to support netscape certificate sequences.
10977 *) Add ASN1 and PEM code to support netscape certificate sequences.
10980 *) Add several PKIX and private extended key usage OIDs.
10983 *) Modify the 'ca' program to handle the new extension code. Modify
10984 openssl.cnf for new extension format, add comments.
10987 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10988 and add a sample to openssl.cnf so req -x509 now adds appropriate
10992 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10993 error code, add initial support to X509_print() and x509 application.
10996 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10997 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10998 stuff is currently isolated and isn't even compiled yet.
11001 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11002 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11003 Removed the versions check from X509 routines when loading extensions:
11004 this allows certain broken certificates that don't set the version
11005 properly to be processed.
11008 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11009 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11010 can still be regenerated with "make depend".
11013 *) Spelling mistake in C version of CAST-128.
11014 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11016 *) Changes to the error generation code. The perl script err-code.pl
11017 now reads in the old error codes and retains the old numbers, only
11018 adding new ones if necessary. It also only changes the .err files if new
11019 codes are added. The makefiles have been modified to only insert errors
11020 when needed (to avoid needlessly modifying header files). This is done
11021 by only inserting errors if the .err file is newer than the auto generated
11022 C file. To rebuild all the error codes from scratch (the old behaviour)
11023 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11024 or delete all the .err files.
11027 *) CAST-128 was incorrectly implemented for short keys. The C version has
11028 been fixed, but is untested. The assembler versions are also fixed, but
11029 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11030 to regenerate it if needed.
11031 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11032 Hagino <itojun@kame.net>]
11034 *) File was opened incorrectly in randfile.c.
11035 [Ulf Möller <ulf@fitug.de>]
11037 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11038 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11039 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11040 al: it's just almost always a UTCTime. Note this patch adds new error
11041 codes so do a "make errors" if there are problems.
11044 *) Correct Linux 1 recognition in config.
11045 [Ulf Möller <ulf@fitug.de>]
11047 *) Remove pointless MD5 hash when using DSA keys in ca.
11048 [Anonymous <nobody@replay.com>]
11050 *) Generate an error if given an empty string as a cert directory. Also
11051 generate an error if handed NULL (previously returned 0 to indicate an
11052 error, but didn't set one).
11053 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11055 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11058 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11059 parameters. This was causing a warning which killed off the Win32 compile.
11062 *) Remove C++ style comments from crypto/bn/bn_local.h.
11063 [Neil Costigan <neil.costigan@celocom.com>]
11065 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11066 based on a text string, looking up short and long names and finally
11067 "dot" format. The "dot" format stuff didn't work. Added new function
11068 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11069 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11070 OID is not part of the table.
11073 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11074 X509_LOOKUP_by_alias().
11077 *) Sort openssl functions by name.
11080 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11081 encryption from sample DSA keys (in case anyone is interested the password
11085 *) Make _all_ *_free functions accept a NULL pointer.
11086 [Frans Heymans <fheymans@isaserver.be>]
11088 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11090 [Anonymous <nobody@replay.com>]
11092 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11093 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11095 *) Don't blow it for numeric -newkey arguments to apps/req.
11096 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11098 *) Temp key "for export" tests were wrong in s3_srvr.c.
11099 [Anonymous <nobody@replay.com>]
11101 *) Add prototype for temp key callback functions
11102 SSL_CTX_set_tmp_{rsa,dh}_callback().
11105 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11106 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11109 *) X509_name_add_entry() freed the wrong thing after an error.
11110 [Arne Ansper <arne@ats.cyber.ee>]
11112 *) rsa_eay.c would attempt to free a NULL context.
11113 [Arne Ansper <arne@ats.cyber.ee>]
11115 *) BIO_s_socket() had a broken should_retry() on Windoze.
11116 [Arne Ansper <arne@ats.cyber.ee>]
11118 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11119 [Arne Ansper <arne@ats.cyber.ee>]
11121 *) Make sure the already existing X509_STORE->depth variable is initialized
11122 in X509_STORE_new(), but document the fact that this variable is still
11123 unused in the certificate verification process.
11124 [Ralf S. Engelschall]
11126 *) Fix the various library and apps files to free up pkeys obtained from
11127 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11130 *) Fix reference counting in X509_PUBKEY_get(). This makes
11131 demos/maurice/example2.c work, amongst others, probably.
11132 [Steve Henson and Ben Laurie]
11134 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11135 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11136 are no longer created. This way we have a single and consistent command
11137 line interface `openssl <command>', similar to `cvs <command>'.
11138 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11140 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11141 BIT STRING wrapper always have zero unused bits.
11144 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11147 *) Make the top-level INSTALL documentation easier to understand.
11150 *) Makefiles updated to exit if an error occurs in a sub-directory
11151 make (including if user presses ^C) [Paul Sutton]
11153 *) Make Montgomery context stuff explicit in RSA data structure.
11156 *) Fix build order of pem and err to allow for generated pem.h.
11159 *) Fix renumbering bug in X509_NAME_delete_entry().
11162 *) Enhanced the err-ins.pl script so it makes the error library number
11163 global and can add a library name. This is needed for external ASN1 and
11164 other error libraries.
11167 *) Fixed sk_insert which never worked properly.
11170 *) Fix ASN1 macros so they can handle indefinite length constructed
11171 EXPLICIT tags. Some non standard certificates use these: they can now
11175 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11176 into a single doc/ssleay.txt bundle. This way the information is still
11177 preserved but no longer messes up this directory. Now it's new room for
11178 the new set of documentation files.
11179 [Ralf S. Engelschall]
11181 *) SETs were incorrectly DER encoded. This was a major pain, because they
11182 shared code with SEQUENCEs, which aren't coded the same. This means that
11183 almost everything to do with SETs or SEQUENCEs has either changed name or
11184 number of arguments.
11185 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11187 *) Fix test data to work with the above.
11190 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11191 was already fixed by Eric for 0.9.1 it seems.
11192 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11194 *) Autodetect FreeBSD3.
11197 *) Fix various bugs in Configure. This affects the following platforms:
11201 unixware-2.0-pentium
11205 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11206 before they are needed.
11209 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11213 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11215 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11216 changed SSLeay to OpenSSL in version strings.
11217 [Ralf S. Engelschall]
11219 *) Some fixups to the top-level documents.
11222 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11223 because the symlink to include/ was missing.
11224 [Ralf S. Engelschall]
11226 *) Incorporated the popular no-RSA/DSA-only patches
11227 which allow to compile a RSA-free SSLeay.
11228 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11230 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11231 when "ssleay" is still not found.
11232 [Ralf S. Engelschall]
11234 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11235 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11237 *) Updated the README file.
11238 [Ralf S. Engelschall]
11240 *) Added various .cvsignore files in the CVS repository subdirs
11241 to make a "cvs update" really silent.
11242 [Ralf S. Engelschall]
11244 *) Recompiled the error-definition header files and added
11245 missing symbols to the Win32 linker tables.
11246 [Ralf S. Engelschall]
11248 *) Cleaned up the top-level documents;
11249 o new files: CHANGES and LICENSE
11250 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11251 o merged COPYRIGHT into LICENSE
11252 o removed obsolete TODO file
11253 o renamed MICROSOFT to INSTALL.W32
11254 [Ralf S. Engelschall]
11256 *) Removed dummy files from the 0.9.1b source tree:
11257 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11258 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11259 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11260 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11261 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11262 [Ralf S. Engelschall]
11264 *) Added various platform portability fixes.
11267 *) The Genesis of the OpenSSL rpject:
11268 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11269 Young and Tim J. Hudson created while they were working for C2Net until
11271 [The OpenSSL Project]
11274 Changes between 0.9.0b and 0.9.1b [not released]
11276 *) Updated a few CA certificates under certs/
11279 *) Changed some BIGNUM api stuff.
11282 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11283 DGUX x86, Linux Alpha, etc.
11286 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11287 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11291 *) Add -strparse option to asn1pars program which parses nested
11293 [Dr Stephen Henson <shenson@bigfoot.com>]
11295 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11298 *) DSA fix for "ca" program.
11301 *) Added "-genkey" option to "dsaparam" program.
11304 *) Added RIPE MD160 (rmd160) message digest.
11307 *) Added -a (all) option to "ssleay version" command.
11310 *) Added PLATFORM define which is the id given to Configure.
11313 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11316 *) Extended the ASN.1 parser routines.
11319 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11322 *) Added a BN_CTX to the BN library.
11325 *) Fixed the weak key values in DES library
11328 *) Changed API in EVP library for cipher aliases.
11331 *) Added support for RC2/64bit cipher.
11334 *) Converted the lhash library to the crypto/mem.c functions.
11337 *) Added more recognized ASN.1 object ids.
11340 *) Added more RSA padding checks for SSL/TLS.
11343 *) Added BIO proxy/filter functionality.
11346 *) Added extra_certs to SSL_CTX which can be used
11347 send extra CA certificates to the client in the CA cert chain sending
11348 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11351 *) Now Fortezza is denied in the authentication phase because
11352 this is key exchange mechanism is not supported by SSLeay at all.
11355 *) Additional PKCS1 checks.
11358 *) Support the string "TLSv1" for all TLS v1 ciphers.
11361 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11362 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11365 *) Fixed a few memory leaks.
11368 *) Fixed various code and comment typos.
11371 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11372 bytes sent in the client random.
11373 [Edward Bishop <ebishop@spyglass.com>]