5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
8 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
9 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
10 DES and RC4 ciphersuites.
13 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
14 This changes the decoding behaviour for some invalid messages,
15 though the change is mostly in the more lenient direction, and
16 legacy behaviour is preserved as much as possible.
18 *) Fix no-stdio build.
19 [ David Woodhouse <David.Woodhouse@intel.com> and also
20 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
21 *) New testing framework
22 The testing framework has been largely rewritten and is now using
23 perl and the perl modules Test::Harness and an extended variant of
24 Test::More called OpenSSL::Test to do its work. All test scripts in
25 test/ have been rewritten into test recipes, and all direct calls to
26 executables in test/Makefile have become individual recipes using the
27 simplified testing OpenSSL::Test::Simple.
29 For documentation on our testing modules, do:
31 perldoc test/testlib/OpenSSL/Test/Simple.pm
32 perldoc test/testlib/OpenSSL/Test.pm
36 *) In DSA_generate_parameters_ex, if the provided seed is too short,
38 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
40 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
41 from RFC4279, RFC4785, RFC5487, RFC5489.
43 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
44 original RSA_PSK patch.
47 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
48 era flag was never set throughout the codebase (only read). Also removed
49 SSL3_FLAGS_POP_BUFFER which was only used if
50 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
53 *) Changed the default name options in the "ca", "crl", "req" and "x509"
54 to be "oneline" instead of "compat".
57 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
58 not aware of clients that still exhibit this bug, and the workaround
59 hasn't been working properly for a while.
62 *) The return type of BIO_number_read() and BIO_number_written() as well as
63 the corresponding num_read and num_write members in the BIO structure has
64 changed from unsigned long to uint64_t. On platforms where an unsigned
65 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
69 *) Given the pervasive nature of TLS extensions it is inadvisable to run
70 OpenSSL without support for them. It also means that maintaining
71 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
72 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
75 *) Removed support for the two export grade static DH ciphersuites
76 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
77 were newly added (along with a number of other static DH ciphersuites) to
78 1.0.2. However the two export ones have *never* worked since they were
79 introduced. It seems strange in any case to be adding new export
80 ciphersuites, and given "logjam" it also does not seem correct to fix them.
83 *) Version negotiation has been rewritten. In particular SSLv23_method(),
84 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
85 and turned into macros which simply call the new preferred function names
86 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
87 should use the new names instead. Also as part of this change the ssl23.h
88 header file has been removed.
91 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
92 code and the associated standard is no longer considered fit-for-purpose.
95 *) RT2547 was closed. When generating a private key, try to make the
96 output file readable only by the owner. This behavior change might
97 be noticeable when interacting with other software.
99 *) Added HTTP GET support to the ocsp command.
102 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
105 *) Added support for TLS extended master secret from
106 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
107 initial patch which was a great help during development.
110 *) All libssl internal structures have been removed from the public header
111 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
112 now redundant). Users should not attempt to access internal structures
113 directly. Instead they should use the provided API functions.
116 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
117 Access to deprecated functions can be re-enabled by running config with
118 "enable-deprecated". In addition applications wishing to use deprecated
119 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
120 will, by default, disable some transitive includes that previously existed
121 in the header files (e.g. ec.h will no longer, by default, include bn.h)
124 *) Added support for OCB mode. OpenSSL has been granted a patent license
125 compatible with the OpenSSL license for use of OCB. Details are available
126 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
127 for OCB can be removed by calling config with no-ocb.
130 *) SSLv2 support has been removed. It still supports receiving a SSLv2
131 compatible client hello.
134 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
135 done while fixing the error code for the key-too-small case.
136 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
138 *) CA.sh has been removmed; use CA.pl instead.
141 *) Removed old DES API.
144 *) Remove various unsupported platforms:
150 Sinix/ReliantUNIX RM400
155 16-bit platforms such as WIN16
158 *) Clean up OPENSSL_NO_xxx #define's
159 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
160 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
161 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
162 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
163 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
164 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
165 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
166 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
167 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
168 Remove MS_STATIC; it's a relic from platforms <32 bits.
171 *) Cleaned up dead code
172 Remove all but one '#ifdef undef' which is to be looked at.
175 *) Clean up calling of xxx_free routines.
176 Just like free(), fix most of the xxx_free routines to accept
177 NULL. Remove the non-null checks from callers. Save much code.
180 *) Add secure heap for storage of private keys (when possible).
181 Add BIO_s_secmem(), CBIGNUM, etc.
182 Contributed by Akamai Technologies under our Corporate CLA.
185 *) Experimental support for a new, fast, unbiased prime candidate generator,
186 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
187 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
189 *) New output format NSS in the sess_id command line tool. This allows
190 exporting the session id and the master key in NSS keylog format.
191 [Martin Kaiser <martin@kaiser.cx>]
193 *) Harmonize version and its documentation. -f flag is used to display
195 [mancha <mancha1@zoho.com>]
197 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
198 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
199 [mancha <mancha1@zoho.com>]
201 *) Fix some double frees. These are not thought to be exploitable.
202 [mancha <mancha1@zoho.com>]
204 *) A missing bounds check in the handling of the TLS heartbeat extension
205 can be used to reveal up to 64k of memory to a connected client or
208 Thanks for Neel Mehta of Google Security for discovering this bug and to
209 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
210 preparing the fix (CVE-2014-0160)
211 [Adam Langley, Bodo Moeller]
213 *) Fix for the attack described in the paper "Recovering OpenSSL
214 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
215 by Yuval Yarom and Naomi Benger. Details can be obtained from:
216 http://eprint.iacr.org/2014/140
218 Thanks to Yuval Yarom and Naomi Benger for discovering this
219 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
220 [Yuval Yarom and Naomi Benger]
222 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
223 this fixes a limitation in previous versions of OpenSSL.
226 *) Experimental encrypt-then-mac support.
228 Experimental support for encrypt then mac from
229 draft-gutmann-tls-encrypt-then-mac-02.txt
231 To enable it set the appropriate extension number (0x42 for the test
232 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
234 For non-compliant peers (i.e. just about everything) this should have no
237 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
241 *) Add EVP support for key wrapping algorithms, to avoid problems with
242 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
243 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
244 algorithms and include tests cases.
247 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
251 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
252 MGF1 digest and OAEP label.
255 *) Make openssl verify return errors.
256 [Chris Palmer <palmer@google.com> and Ben Laurie]
258 *) New function ASN1_TIME_diff to calculate the difference between two
259 ASN1_TIME structures or one structure and the current time.
262 *) Update fips_test_suite to support multiple command line options. New
263 test to induce all self test errors in sequence and check expected
267 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
268 sign or verify all in one operation.
271 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
272 test programs and fips_test_suite. Includes functionality to parse
273 the minimal script output of fipsalgest.pl directly.
276 *) Add authorisation parameter to FIPS_module_mode_set().
279 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
282 *) Use separate DRBG fields for internal and external flags. New function
283 FIPS_drbg_health_check() to perform on demand health checking. Add
284 generation tests to fips_test_suite with reduced health check interval to
285 demonstrate periodic health checking. Add "nodh" option to
286 fips_test_suite to skip very slow DH test.
289 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
293 *) More extensive health check for DRBG checking many more failure modes.
294 New function FIPS_selftest_drbg_all() to handle every possible DRBG
295 combination: call this in fips_test_suite.
298 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
299 and POST to handle Dual EC cases.
302 *) Add support for canonical generation of DSA parameter 'g'. See
305 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
306 POST to handle HMAC cases.
309 *) Add functions FIPS_module_version() and FIPS_module_version_text()
310 to return numerical and string versions of the FIPS module number.
313 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
314 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
315 outside the validated module in the FIPS capable OpenSSL.
318 *) Minor change to DRBG entropy callback semantics. In some cases
319 there is no multiple of the block length between min_len and
320 max_len. Allow the callback to return more than max_len bytes
321 of entropy but discard any extra: it is the callback's responsibility
322 to ensure that the extra data discarded does not impact the
323 requested amount of entropy.
326 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
327 information in FIPS186-3, SP800-57 and SP800-131A.
330 *) CCM support via EVP. Interface is very similar to GCM case except we
331 must supply all data in one chunk (i.e. no update, final) and the
332 message length must be supplied if AAD is used. Add algorithm test
336 *) Initial version of POST overhaul. Add POST callback to allow the status
337 of POST to be monitored and/or failures induced. Modify fips_test_suite
338 to use callback. Always run all selftests even if one fails.
341 *) XTS support including algorithm test driver in the fips_gcmtest program.
342 Note: this does increase the maximum key length from 32 to 64 bytes but
343 there should be no binary compatibility issues as existing applications
344 will never use XTS mode.
347 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
348 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
349 performs algorithm blocking for unapproved PRNG types. Also do not
350 set PRNG type in FIPS_mode_set(): leave this to the application.
351 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
352 the standard OpenSSL PRNG: set additional data to a date time vector.
355 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
356 This shouldn't present any incompatibility problems because applications
357 shouldn't be using these directly and any that are will need to rethink
358 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
361 *) Extensive self tests and health checking required by SP800-90 DRBG.
362 Remove strength parameter from FIPS_drbg_instantiate and always
363 instantiate at maximum supported strength.
366 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
369 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
372 *) New function DH_compute_key_padded() to compute a DH key and pad with
373 leading zeroes if needed: this complies with SP800-56A et al.
376 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
377 anything, incomplete, subject to change and largely untested at present.
380 *) Modify fipscanisteronly build option to only build the necessary object
381 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
384 *) Add experimental option FIPSSYMS to give all symbols in
385 fipscanister.o and FIPS or fips prefix. This will avoid
386 conflicts with future versions of OpenSSL. Add perl script
387 util/fipsas.pl to preprocess assembly language source files
388 and rename any affected symbols.
391 *) Add selftest checks and algorithm block of non-fips algorithms in
392 FIPS mode. Remove DES2 from selftests.
395 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
396 return internal method without any ENGINE dependencies. Add new
397 tiny fips sign and verify functions.
400 *) New build option no-ec2m to disable characteristic 2 code.
403 *) New build option "fipscanisteronly". This only builds fipscanister.o
404 and (currently) associated fips utilities. Uses the file Makefile.fips
405 instead of Makefile.org as the prototype.
408 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
409 Update fips_gcmtest to use IV generator.
412 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
413 setting output buffer to NULL. The *Final function must be
414 called although it will not retrieve any additional data. The tag
415 can be set or retrieved with a ctrl. The IV length is by default 12
416 bytes (96 bits) but can be set to an alternative value. If the IV
417 length exceeds the maximum IV length (currently 16 bytes) it cannot be
421 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
422 underlying do_cipher function handles all cipher semantics itself
423 including padding and finalisation. This is useful if (for example)
424 an ENGINE cipher handles block padding itself. The behaviour of
425 do_cipher is subtly changed if this flag is set: the return value
426 is the number of characters written to the output buffer (zero is
427 no longer an error code) or a negative error code. Also if the
428 input buffer is NULL and length 0 finalisation should be performed.
431 *) If a candidate issuer certificate is already part of the constructed
432 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
435 *) Improve forward-security support: add functions
437 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
438 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
440 for use by SSL/TLS servers; the callback function will be called whenever a
441 new session is created, and gets to decide whether the session may be
442 cached to make it resumable (return 0) or not (return 1). (As by the
443 SSL/TLS protocol specifications, the session_id sent by the server will be
444 empty to indicate that the session is not resumable; also, the server will
445 not generate RFC 4507 (RFC 5077) session tickets.)
447 A simple reasonable callback implementation is to return is_forward_secure.
448 This parameter will be set to 1 or 0 depending on the ciphersuite selected
449 by the SSL/TLS server library, indicating whether it can provide forward
451 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
453 *) New -verify_name option in command line utilities to set verification
457 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
458 Add CMAC pkey methods.
461 *) Experimental renegotiation in s_server -www mode. If the client
462 browses /reneg connection is renegotiated. If /renegcert it is
463 renegotiated requesting a certificate.
466 *) Add an "external" session cache for debugging purposes to s_server. This
467 should help trace issues which normally are only apparent in deployed
468 multi-process servers.
471 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
472 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
473 BIO_set_cipher() and some obscure PEM functions were changed so they
474 can now return an error. The RAND changes required a change to the
475 RAND_METHOD structure.
478 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
479 a gcc attribute to warn if the result of a function is ignored. This
480 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
481 whose return value is often ignored.
484 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
486 *) Alternate chains certificate forgery
488 During certificate verfification, OpenSSL will attempt to find an
489 alternative certificate chain if the first attempt to build such a chain
490 fails. An error in the implementation of this logic can mean that an
491 attacker could cause certain checks on untrusted certificates to be
492 bypassed, such as the CA flag, enabling them to use a valid leaf
493 certificate to act as a CA and "issue" an invalid certificate.
495 This issue was reported to OpenSSL by Adam Langley/David Benjamin
499 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
501 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
502 incompatibility in the handling of HMAC. The previous ABI has now been
506 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
508 *) Malformed ECParameters causes infinite loop
510 When processing an ECParameters structure OpenSSL enters an infinite loop
511 if the curve specified is over a specially malformed binary polynomial
514 This can be used to perform denial of service against any
515 system which processes public keys, certificate requests or
516 certificates. This includes TLS clients and TLS servers with
517 client authentication enabled.
519 This issue was reported to OpenSSL by Joseph Barr-Pixton.
523 *) Exploitable out-of-bounds read in X509_cmp_time
525 X509_cmp_time does not properly check the length of the ASN1_TIME
526 string and can read a few bytes out of bounds. In addition,
527 X509_cmp_time accepts an arbitrary number of fractional seconds in the
530 An attacker can use this to craft malformed certificates and CRLs of
531 various sizes and potentially cause a segmentation fault, resulting in
532 a DoS on applications that verify certificates or CRLs. TLS clients
533 that verify CRLs are affected. TLS clients and servers with client
534 authentication enabled may be affected if they use custom verification
537 This issue was reported to OpenSSL by Robert Swiecki (Google), and
538 independently by Hanno Böck.
542 *) PKCS7 crash with missing EnvelopedContent
544 The PKCS#7 parsing code does not handle missing inner EncryptedContent
545 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
546 with missing content and trigger a NULL pointer dereference on parsing.
548 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
549 structures from untrusted sources are affected. OpenSSL clients and
550 servers are not affected.
552 This issue was reported to OpenSSL by Michal Zalewski (Google).
556 *) CMS verify infinite loop with unknown hash function
558 When verifying a signedData message the CMS code can enter an infinite loop
559 if presented with an unknown hash function OID. This can be used to perform
560 denial of service against any system which verifies signedData messages using
562 This issue was reported to OpenSSL by Johannes Bauer.
566 *) Race condition handling NewSessionTicket
568 If a NewSessionTicket is received by a multi-threaded client when attempting to
569 reuse a previous ticket then a race condition can occur potentially leading to
570 a double free of the ticket data.
574 *) Only support 256-bit or stronger elliptic curves with the
575 'ecdh_auto' setting (server) or by default (client). Of supported
576 curves, prefer P-256 (both).
579 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
581 *) ClientHello sigalgs DoS fix
583 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
584 invalid signature algorithms extension a NULL pointer dereference will
585 occur. This can be exploited in a DoS attack against the server.
587 This issue was was reported to OpenSSL by David Ramos of Stanford
590 [Stephen Henson and Matt Caswell]
592 *) Multiblock corrupted pointer fix
594 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
595 feature only applies on 64 bit x86 architecture platforms that support AES
596 NI instructions. A defect in the implementation of "multiblock" can cause
597 OpenSSL's internal write buffer to become incorrectly set to NULL when
598 using non-blocking IO. Typically, when the user application is using a
599 socket BIO for writing, this will only result in a failed connection.
600 However if some other BIO is used then it is likely that a segmentation
601 fault will be triggered, thus enabling a potential DoS attack.
603 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
607 *) Segmentation fault in DTLSv1_listen fix
609 The DTLSv1_listen function is intended to be stateless and processes the
610 initial ClientHello from many peers. It is common for user code to loop
611 over the call to DTLSv1_listen until a valid ClientHello is received with
612 an associated cookie. A defect in the implementation of DTLSv1_listen means
613 that state is preserved in the SSL object from one invocation to the next
614 that can lead to a segmentation fault. Errors processing the initial
615 ClientHello can trigger this scenario. An example of such an error could be
616 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
619 This issue was reported to OpenSSL by Per Allansson.
623 *) Segmentation fault in ASN1_TYPE_cmp fix
625 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
626 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
627 certificate signature algorithm consistency this can be used to crash any
628 certificate verification operation and exploited in a DoS attack. Any
629 application which performs certificate verification is vulnerable including
630 OpenSSL clients and servers which enable client authentication.
634 *) Segmentation fault for invalid PSS parameters fix
636 The signature verification routines will crash with a NULL pointer
637 dereference if presented with an ASN.1 signature using the RSA PSS
638 algorithm and invalid parameters. Since these routines are used to verify
639 certificate signature algorithms this can be used to crash any
640 certificate verification operation and exploited in a DoS attack. Any
641 application which performs certificate verification is vulnerable including
642 OpenSSL clients and servers which enable client authentication.
644 This issue was was reported to OpenSSL by Brian Carpenter.
648 *) ASN.1 structure reuse memory corruption fix
650 Reusing a structure in ASN.1 parsing may allow an attacker to cause
651 memory corruption via an invalid write. Such reuse is and has been
652 strongly discouraged and is believed to be rare.
654 Applications that parse structures containing CHOICE or ANY DEFINED BY
655 components may be affected. Certificate parsing (d2i_X509 and related
656 functions) are however not affected. OpenSSL clients and servers are
661 *) PKCS7 NULL pointer dereferences fix
663 The PKCS#7 parsing code does not handle missing outer ContentInfo
664 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
665 missing content and trigger a NULL pointer dereference on parsing.
667 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
668 otherwise parse PKCS#7 structures from untrusted sources are
669 affected. OpenSSL clients and servers are not affected.
671 This issue was reported to OpenSSL by Michal Zalewski (Google).
675 *) DoS via reachable assert in SSLv2 servers fix
677 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
678 servers that both support SSLv2 and enable export cipher suites by sending
679 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
681 This issue was discovered by Sean Burford (Google) and Emilia Käsper
682 (OpenSSL development team).
686 *) Empty CKE with client auth and DHE fix
688 If client auth is used then a server can seg fault in the event of a DHE
689 ciphersuite being selected and a zero length ClientKeyExchange message
690 being sent by the client. This could be exploited in a DoS attack.
694 *) Handshake with unseeded PRNG fix
696 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
697 with an unseeded PRNG. The conditions are:
698 - The client is on a platform where the PRNG has not been seeded
699 automatically, and the user has not seeded manually
700 - A protocol specific client method version has been used (i.e. not
701 SSL_client_methodv23)
702 - A ciphersuite is used that does not require additional random data from
703 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
705 If the handshake succeeds then the client random that has been used will
706 have been generated from a PRNG with insufficient entropy and therefore the
707 output may be predictable.
709 For example using the following command with an unseeded openssl will
710 succeed on an unpatched platform:
712 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
716 *) Use After Free following d2i_ECPrivatekey error fix
718 A malformed EC private key file consumed via the d2i_ECPrivateKey function
719 could cause a use after free condition. This, in turn, could cause a double
720 free in several private key parsing functions (such as d2i_PrivateKey
721 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
722 for applications that receive EC private keys from untrusted
723 sources. This scenario is considered rare.
725 This issue was discovered by the BoringSSL project and fixed in their
730 *) X509_to_X509_REQ NULL pointer deref fix
732 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
733 the certificate key is invalid. This function is rarely used in practice.
735 This issue was discovered by Brian Carpenter.
739 *) Removed the export ciphers from the DEFAULT ciphers
742 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
744 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
745 ARMv5 through ARMv8, as opposite to "locking" it to single one.
746 So far those who have to target multiple plaforms would compromise
747 and argue that binary targeting say ARMv5 would still execute on
748 ARMv8. "Universal" build resolves this compromise by providing
749 near-optimal performance even on newer platforms.
752 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
753 (other platforms pending).
754 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
756 *) Add support for the SignedCertificateTimestampList certificate and
757 OCSP response extensions from RFC6962.
760 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
761 for corner cases. (Certain input points at infinity could lead to
762 bogus results, with non-infinity inputs mapped to infinity too.)
765 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
766 This covers AES, SHA256/512 and GHASH. "Initial" means that most
767 common cases are optimized and there still is room for further
768 improvements. Vector Permutation AES for Altivec is also added.
771 *) Add support for little-endian ppc64 Linux target.
772 [Marcelo Cerri (IBM)]
774 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
775 SHA1, SHA256 and GHASH. "Initial" means that most common cases
776 are optimized and there still is room for further improvements.
777 Both 32- and 64-bit modes are supported.
778 [Andy Polyakov, Ard Biesheuvel (Linaro)]
780 *) Improved ARMv7 NEON support.
783 *) Support for SPARC Architecture 2011 crypto extensions, first
784 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
785 SHA256/512, MD5, GHASH and modular exponentiation.
786 [Andy Polyakov, David Miller]
788 *) Accelerated modular exponentiation for Intel processors, a.k.a.
790 [Shay Gueron & Vlad Krasnov (Intel Corp)]
792 *) Support for new and upcoming Intel processors, including AVX2,
793 BMI and SHA ISA extensions. This includes additional "stitched"
794 implementations, AESNI-SHA256 and GCM, and multi-buffer support
797 This work was sponsored by Intel Corp.
800 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
801 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
802 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
805 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
806 this fixes a limiation in previous versions of OpenSSL.
809 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
810 MGF1 digest and OAEP label.
813 *) Add EVP support for key wrapping algorithms, to avoid problems with
814 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
815 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
816 algorithms and include tests cases.
819 *) Add functions to allocate and set the fields of an ECDSA_METHOD
821 [Douglas E. Engert, Steve Henson]
823 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
824 difference in days and seconds between two tm or ASN1_TIME structures.
827 *) Add -rev test option to s_server to just reverse order of characters
828 received by client and send back to server. Also prints an abbreviated
829 summary of the connection parameters.
832 *) New option -brief for s_client and s_server to print out a brief summary
833 of connection parameters.
836 *) Add callbacks for arbitrary TLS extensions.
837 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
839 *) New option -crl_download in several openssl utilities to download CRLs
840 from CRLDP extension in certificates.
843 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
846 *) New function X509_CRL_diff to generate a delta CRL from the difference
847 of two full CRLs. Add support to "crl" utility.
850 *) New functions to set lookup_crls function and to retrieve
851 X509_STORE from X509_STORE_CTX.
854 *) Print out deprecated issuer and subject unique ID fields in
858 *) Extend OCSP I/O functions so they can be used for simple general purpose
859 HTTP as well as OCSP. New wrapper function which can be used to download
860 CRLs using the OCSP API.
863 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
866 *) SSL_CONF* functions. These provide a common framework for application
867 configuration using configuration files or command lines.
870 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
871 message callback and prints the results. Needs compile time option
872 "enable-ssl-trace". New options to s_client and s_server to enable
876 *) New ctrl and macro to retrieve supported points extensions.
877 Print out extension in s_server and s_client.
880 *) New functions to retrieve certificate signature and signature
884 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
888 *) New Suite B modes for TLS code. These use and enforce the requirements
889 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
890 only use Suite B curves. The Suite B modes can be set by using the
891 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
894 *) New chain verification flags for Suite B levels of security. Check
895 algorithms are acceptable when flags are set in X509_verify_cert.
898 *) Make tls1_check_chain return a set of flags indicating checks passed
899 by a certificate chain. Add additional tests to handle client
900 certificates: checks for matching certificate type and issuer name
904 *) If an attempt is made to use a signature algorithm not in the peer
905 preference list abort the handshake. If client has no suitable
906 signature algorithms in response to a certificate request do not
910 *) If server EC tmp key is not in client preference list abort handshake.
913 *) Add support for certificate stores in CERT structure. This makes it
914 possible to have different stores per SSL structure or one store in
915 the parent SSL_CTX. Include distint stores for certificate chain
916 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
917 to build and store a certificate chain in CERT structure: returing
918 an error if the chain cannot be built: this will allow applications
919 to test if a chain is correctly configured.
921 Note: if the CERT based stores are not set then the parent SSL_CTX
922 store is used to retain compatibility with existing behaviour.
926 *) New function ssl_set_client_disabled to set a ciphersuite disabled
927 mask based on the current session, check mask when sending client
928 hello and checking the requested ciphersuite.
931 *) New ctrls to retrieve and set certificate types in a certificate
932 request message. Print out received values in s_client. If certificate
933 types is not set with custom values set sensible values based on
934 supported signature algorithms.
937 *) Support for distinct client and server supported signature algorithms.
940 *) Add certificate callback. If set this is called whenever a certificate
941 is required by client or server. An application can decide which
942 certificate chain to present based on arbitrary criteria: for example
943 supported signature algorithms. Add very simple example to s_server.
944 This fixes many of the problems and restrictions of the existing client
945 certificate callback: for example you can now clear an existing
946 certificate and specify the whole chain.
949 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
950 the certificate can be used for (if anything). Set valid_flags field
951 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
952 to have similar checks in it.
954 Add new "cert_flags" field to CERT structure and include a "strict mode".
955 This enforces some TLS certificate requirements (such as only permitting
956 certificate signature algorithms contained in the supported algorithms
957 extension) which some implementations ignore: this option should be used
958 with caution as it could cause interoperability issues.
961 *) Update and tidy signature algorithm extension processing. Work out
962 shared signature algorithms based on preferences and peer algorithms
963 and print them out in s_client and s_server. Abort handshake if no
964 shared signature algorithms.
967 *) Add new functions to allow customised supported signature algorithms
968 for SSL and SSL_CTX structures. Add options to s_client and s_server
972 *) New function SSL_certs_clear() to delete all references to certificates
973 from an SSL structure. Before this once a certificate had been added
974 it couldn't be removed.
977 *) Integrate hostname, email address and IP address checking with certificate
978 verification. New verify options supporting checking in opensl utility.
981 *) Fixes and wildcard matching support to hostname and email checking
982 functions. Add manual page.
983 [Florian Weimer (Red Hat Product Security Team)]
985 *) New functions to check a hostname email or IP address against a
986 certificate. Add options x509 utility to print results of checks against
990 *) Fix OCSP checking.
991 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
993 *) Initial experimental support for explicitly trusted non-root CAs.
994 OpenSSL still tries to build a complete chain to a root but if an
995 intermediate CA has a trust setting included that is used. The first
996 setting is used: whether to trust (e.g., -addtrust option to the x509
1000 *) Add -trusted_first option which attempts to find certificates in the
1001 trusted store even if an untrusted chain is also supplied.
1004 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1005 platform support for Linux and Android.
1008 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1011 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1012 When in FIPS mode the approved implementations are used as normal,
1013 when not in FIPS mode the internal unapproved versions are used instead.
1014 This means that the FIPS capable OpenSSL isn't forced to use the
1015 (often lower performance) FIPS implementations outside FIPS mode.
1018 *) Transparently support X9.42 DH parameters when calling
1019 PEM_read_bio_DHparameters. This means existing applications can handle
1020 the new parameter format automatically.
1023 *) Initial experimental support for X9.42 DH parameter format: mainly
1024 to support use of 'q' parameter for RFC5114 parameters.
1027 *) Add DH parameters from RFC5114 including test data to dhtest.
1030 *) Support for automatic EC temporary key parameter selection. If enabled
1031 the most preferred EC parameters are automatically used instead of
1032 hardcoded fixed parameters. Now a server just has to call:
1033 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1034 support ECDH and use the most appropriate parameters.
1037 *) Enhance and tidy EC curve and point format TLS extension code. Use
1038 static structures instead of allocation if default values are used.
1039 New ctrls to set curves we wish to support and to retrieve shared curves.
1040 Print out shared curves in s_server. New options to s_server and s_client
1041 to set list of supported curves.
1044 *) New ctrls to retrieve supported signature algorithms and
1045 supported curve values as an array of NIDs. Extend openssl utility
1046 to print out received values.
1049 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1050 between NIDs and the more common NIST names such as "P-256". Enhance
1051 ecparam utility and ECC method to recognise the NIST names for curves.
1054 *) Enhance SSL/TLS certificate chain handling to support different
1055 chains for each certificate instead of one chain in the parent SSL_CTX.
1058 *) Support for fixed DH ciphersuite client authentication: where both
1059 server and client use DH certificates with common parameters.
1062 *) Support for fixed DH ciphersuites: those requiring DH server
1066 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1068 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1069 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1070 X509_CINF_get_signature were reverted post internal team review.
1072 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1074 *) Build fixes for the Windows and OpenVMS platforms
1075 [Matt Caswell and Richard Levitte]
1077 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1079 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1080 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1081 dereference. This could lead to a Denial Of Service attack. Thanks to
1082 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1086 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1087 dtls1_buffer_record function under certain conditions. In particular this
1088 could occur if an attacker sent repeated DTLS records with the same
1089 sequence number but for the next epoch. The memory leak could be exploited
1090 by an attacker in a Denial of Service attack through memory exhaustion.
1091 Thanks to Chris Mueller for reporting this issue.
1095 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1096 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1097 method would be set to NULL which could later result in a NULL pointer
1098 dereference. Thanks to Frank Schmirler for reporting this issue.
1102 *) Abort handshake if server key exchange message is omitted for ephemeral
1105 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1106 reporting this issue.
1110 *) Remove non-export ephemeral RSA code on client and server. This code
1111 violated the TLS standard by allowing the use of temporary RSA keys in
1112 non-export ciphersuites and could be used by a server to effectively
1113 downgrade the RSA key length used to a value smaller than the server
1114 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1115 INRIA or reporting this issue.
1119 *) Fixed issue where DH client certificates are accepted without verification.
1120 An OpenSSL server will accept a DH certificate for client authentication
1121 without the certificate verify message. This effectively allows a client to
1122 authenticate without the use of a private key. This only affects servers
1123 which trust a client certificate authority which issues certificates
1124 containing DH keys: these are extremely rare and hardly ever encountered.
1125 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1130 *) Ensure that the session ID context of an SSL is updated when its
1131 SSL_CTX is updated via SSL_set_SSL_CTX.
1133 The session ID context is typically set from the parent SSL_CTX,
1134 and can vary with the CTX.
1137 *) Fix various certificate fingerprint issues.
1139 By using non-DER or invalid encodings outside the signed portion of a
1140 certificate the fingerprint can be changed without breaking the signature.
1141 Although no details of the signed portion of the certificate can be changed
1142 this can cause problems with some applications: e.g. those using the
1143 certificate fingerprint for blacklists.
1145 1. Reject signatures with non zero unused bits.
1147 If the BIT STRING containing the signature has non zero unused bits reject
1148 the signature. All current signature algorithms require zero unused bits.
1150 2. Check certificate algorithm consistency.
1152 Check the AlgorithmIdentifier inside TBS matches the one in the
1153 certificate signature. NB: this will result in signature failure
1154 errors for some broken certificates.
1156 Thanks to Konrad Kraszewski from Google for reporting this issue.
1158 3. Check DSA/ECDSA signatures use DER.
1160 Reencode DSA/ECDSA signatures and compare with the original received
1161 signature. Return an error if there is a mismatch.
1163 This will reject various cases including garbage after signature
1164 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1165 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1166 (negative or with leading zeroes).
1168 Further analysis was conducted and fixes were developed by Stephen Henson
1169 of the OpenSSL core team.
1174 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1175 results on some platforms, including x86_64. This bug occurs at random
1176 with a very low probability, and is not known to be exploitable in any
1177 way, though its exact impact is difficult to determine. Thanks to Pieter
1178 Wuille (Blockstream) who reported this issue and also suggested an initial
1179 fix. Further analysis was conducted by the OpenSSL development team and
1180 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1181 the OpenSSL core team.
1185 *) Do not resume sessions on the server if the negotiated protocol
1186 version does not match the session's version. Resuming with a different
1187 version, while not strictly forbidden by the RFC, is of questionable
1188 sanity and breaks all known clients.
1189 [David Benjamin, Emilia Käsper]
1191 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1192 early CCS messages during renegotiation. (Note that because
1193 renegotiation is encrypted, this early CCS was not exploitable.)
1196 *) Tighten client-side session ticket handling during renegotiation:
1197 ensure that the client only accepts a session ticket if the server sends
1198 the extension anew in the ServerHello. Previously, a TLS client would
1199 reuse the old extension state and thus accept a session ticket if one was
1200 announced in the initial ServerHello.
1202 Similarly, ensure that the client requires a session ticket if one
1203 was advertised in the ServerHello. Previously, a TLS client would
1204 ignore a missing NewSessionTicket message.
1207 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1209 *) SRTP Memory Leak.
1211 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1212 sends a carefully crafted handshake message, to cause OpenSSL to fail
1213 to free up to 64k of memory causing a memory leak. This could be
1214 exploited in a Denial Of Service attack. This issue affects OpenSSL
1215 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1216 whether SRTP is used or configured. Implementations of OpenSSL that
1217 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1219 The fix was developed by the OpenSSL team.
1223 *) Session Ticket Memory Leak.
1225 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1226 integrity of that ticket is first verified. In the event of a session
1227 ticket integrity check failing, OpenSSL will fail to free memory
1228 causing a memory leak. By sending a large number of invalid session
1229 tickets an attacker could exploit this issue in a Denial Of Service
1234 *) Build option no-ssl3 is incomplete.
1236 When OpenSSL is configured with "no-ssl3" as a build option, servers
1237 could accept and complete a SSL 3.0 handshake, and clients could be
1238 configured to send them.
1240 [Akamai and the OpenSSL team]
1242 *) Add support for TLS_FALLBACK_SCSV.
1243 Client applications doing fallback retries should call
1244 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1246 [Adam Langley, Bodo Moeller]
1248 *) Add additional DigestInfo checks.
1250 Reencode DigestInto in DER and check against the original when
1251 verifying RSA signature: this will reject any improperly encoded
1252 DigestInfo structures.
1254 Note: this is a precautionary measure and no attacks are currently known.
1258 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1260 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1261 SRP code can be overrun an internal buffer. Add sanity check that
1262 g, A, B < N to SRP code.
1264 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1265 Group for discovering this issue.
1269 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1270 TLS 1.0 instead of higher protocol versions when the ClientHello message
1271 is badly fragmented. This allows a man-in-the-middle attacker to force a
1272 downgrade to TLS 1.0 even if both the server and the client support a
1273 higher protocol version, by modifying the client's TLS records.
1275 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1276 researching this issue.
1280 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1281 to a denial of service attack. A malicious server can crash the client
1282 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1283 ciphersuite and sending carefully crafted handshake messages.
1285 Thanks to Felix Gröbert (Google) for discovering and researching this
1290 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1291 to leak memory. This can be exploited through a Denial of Service attack.
1292 Thanks to Adam Langley for discovering and researching this issue.
1296 *) An attacker can force openssl to consume large amounts of memory whilst
1297 processing DTLS handshake messages. This can be exploited through a
1298 Denial of Service attack.
1299 Thanks to Adam Langley for discovering and researching this issue.
1303 *) An attacker can force an error condition which causes openssl to crash
1304 whilst processing DTLS packets due to memory being freed twice. This
1305 can be exploited through a Denial of Service attack.
1306 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1311 *) If a multithreaded client connects to a malicious server using a resumed
1312 session and the server sends an ec point format extension it could write
1313 up to 255 bytes to freed memory.
1315 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1320 *) A malicious server can crash an OpenSSL client with a null pointer
1321 dereference (read) by specifying an SRP ciphersuite even though it was not
1322 properly negotiated with the client. This can be exploited through a
1323 Denial of Service attack.
1325 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1326 discovering and researching this issue.
1330 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1331 X509_name_oneline, X509_name_print_ex et al. to leak some information
1332 from the stack. Applications may be affected if they echo pretty printing
1333 output to the attacker.
1335 Thanks to Ivan Fratric (Google) for discovering this issue.
1337 [Emilia Käsper, and Steve Henson]
1339 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1340 for corner cases. (Certain input points at infinity could lead to
1341 bogus results, with non-infinity inputs mapped to infinity too.)
1344 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1346 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1347 handshake can force the use of weak keying material in OpenSSL
1348 SSL/TLS clients and servers.
1350 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1351 researching this issue. (CVE-2014-0224)
1352 [KIKUCHI Masashi, Steve Henson]
1354 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1355 OpenSSL DTLS client the code can be made to recurse eventually crashing
1358 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1360 [Imre Rad, Steve Henson]
1362 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1363 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1364 client or server. This is potentially exploitable to run arbitrary
1365 code on a vulnerable client or server.
1367 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1368 [Jüri Aedla, Steve Henson]
1370 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1371 are subject to a denial of service attack.
1373 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1374 this issue. (CVE-2014-3470)
1375 [Felix Gröbert, Ivan Fratric, Steve Henson]
1377 *) Harmonize version and its documentation. -f flag is used to display
1379 [mancha <mancha1@zoho.com>]
1381 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1382 in i2d_ECPrivateKey.
1383 [mancha <mancha1@zoho.com>]
1385 *) Fix some double frees. These are not thought to be exploitable.
1386 [mancha <mancha1@zoho.com>]
1388 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1390 *) A missing bounds check in the handling of the TLS heartbeat extension
1391 can be used to reveal up to 64k of memory to a connected client or
1394 Thanks for Neel Mehta of Google Security for discovering this bug and to
1395 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1396 preparing the fix (CVE-2014-0160)
1397 [Adam Langley, Bodo Moeller]
1399 *) Fix for the attack described in the paper "Recovering OpenSSL
1400 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1401 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1402 http://eprint.iacr.org/2014/140
1404 Thanks to Yuval Yarom and Naomi Benger for discovering this
1405 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1406 [Yuval Yarom and Naomi Benger]
1408 *) TLS pad extension: draft-agl-tls-padding-03
1410 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1411 TLS client Hello record length value would otherwise be > 255 and
1412 less that 512 pad with a dummy extension containing zeroes so it
1413 is at least 512 bytes long.
1415 [Adam Langley, Steve Henson]
1417 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1419 *) Fix for TLS record tampering bug. A carefully crafted invalid
1420 handshake could crash OpenSSL with a NULL pointer exception.
1421 Thanks to Anton Johansson for reporting this issues.
1424 *) Keep original DTLS digest and encryption contexts in retransmission
1425 structures so we can use the previous session parameters if they need
1426 to be resent. (CVE-2013-6450)
1429 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1430 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1431 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1432 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1433 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1434 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1435 [Rob Stradling, Adam Langley]
1437 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1439 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1440 supporting platforms or when small records were transferred.
1441 [Andy Polyakov, Steve Henson]
1443 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1445 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1447 This addresses the flaw in CBC record processing discovered by
1448 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1449 at: http://www.isg.rhul.ac.uk/tls/
1451 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1452 Security Group at Royal Holloway, University of London
1453 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1454 Emilia Käsper for the initial patch.
1456 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1458 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1459 ciphersuites which can be exploited in a denial of service attack.
1460 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1461 and detecting this bug and to Wolfgang Ettlinger
1462 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1466 *) Return an error when checking OCSP signatures when key is NULL.
1467 This fixes a DoS attack. (CVE-2013-0166)
1470 *) Make openssl verify return errors.
1471 [Chris Palmer <palmer@google.com> and Ben Laurie]
1473 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1474 the right response is stapled. Also change SSL_get_certificate()
1475 so it returns the certificate actually sent.
1476 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1477 [Rob Stradling <rob.stradling@comodo.com>]
1479 *) Fix possible deadlock when decoding public keys.
1482 *) Don't use TLS 1.0 record version number in initial client hello
1486 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1488 *) Sanity check record length before skipping explicit IV in TLS
1489 1.2, 1.1 and DTLS to fix DoS attack.
1491 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1492 fuzzing as a service testing platform.
1496 *) Initialise tkeylen properly when encrypting CMS messages.
1497 Thanks to Solar Designer of Openwall for reporting this issue.
1500 *) In FIPS mode don't try to use composite ciphers as they are not
1504 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1506 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1507 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1508 mean any application compiled against OpenSSL 1.0.0 headers setting
1509 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1510 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1511 0x10000000L Any application which was previously compiled against
1512 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1513 will need to be recompiled as a result. Letting be results in
1514 inability to disable specifically TLS 1.1 and in client context,
1515 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1518 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1519 disable just protocol X, but all protocols above X *if* there are
1520 protocols *below* X still enabled. In more practical terms it means
1521 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1522 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1523 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1527 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1529 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1530 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1531 in CRYPTO_realloc_clean.
1533 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1534 issue and to Adam Langley <agl@chromium.org> for fixing it.
1536 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1538 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1541 *) Workarounds for some broken servers that "hang" if a client hello
1542 record length exceeds 255 bytes.
1544 1. Do not use record version number > TLS 1.0 in initial client
1545 hello: some (but not all) hanging servers will now work.
1546 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1547 the number of ciphers sent in the client hello. This should be
1548 set to an even number, such as 50, for example by passing:
1549 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1550 Most broken servers should now work.
1551 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1552 TLS 1.2 client support entirely.
1555 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1558 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1560 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1561 STRING form instead of a DigestInfo.
1564 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1565 and the RSA_sign/RSA_verify functions. This was made more apparent when
1566 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1567 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1568 the correct format in RSA_verify so both forms transparently work.
1571 *) Some servers which support TLS 1.0 can choke if we initially indicate
1572 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1573 encrypted premaster secret. As a workaround use the maximum permitted
1574 client version in client hello, this should keep such servers happy
1575 and still work with previous versions of OpenSSL.
1578 *) Add support for TLS/DTLS heartbeats.
1579 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1581 *) Add support for SCTP.
1582 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1584 *) Improved PRNG seeding for VOS.
1585 [Paul Green <Paul.Green@stratus.com>]
1587 *) Extensive assembler packs updates, most notably:
1589 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1590 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1591 - x86_64: bit-sliced AES implementation;
1592 - ARM: NEON support, contemporary platforms optimizations;
1593 - s390x: z196 support;
1594 - *: GHASH and GF(2^m) multiplication implementations;
1598 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1599 (removal of unnecessary code)
1600 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1602 *) Add TLS key material exporter from RFC 5705.
1605 *) Add DTLS-SRTP negotiation from RFC 5764.
1608 *) Add Next Protocol Negotiation,
1609 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1610 disabled with a no-npn flag to config or Configure. Code donated
1612 [Adam Langley <agl@google.com> and Ben Laurie]
1614 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1615 NIST-P256, NIST-P521, with constant-time single point multiplication on
1616 typical inputs. Compiler support for the nonstandard type __uint128_t is
1617 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1618 Code made available under Apache License version 2.0.
1620 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1621 line to include this in your build of OpenSSL, and run "make depend" (or
1622 "make update"). This enables the following EC_METHODs:
1624 EC_GFp_nistp224_method()
1625 EC_GFp_nistp256_method()
1626 EC_GFp_nistp521_method()
1628 EC_GROUP_new_by_curve_name() will automatically use these (while
1629 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1631 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1633 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1634 all platforms. Move ssize_t definition from e_os.h to the public
1635 header file e_os2.h as it now appears in public header file cms.h
1638 *) New -sigopt option to the ca, req and x509 utilities. Additional
1639 signature parameters can be passed using this option and in
1643 *) Add RSA PSS signing function. This will generate and set the
1644 appropriate AlgorithmIdentifiers for PSS based on those in the
1645 corresponding EVP_MD_CTX structure. No application support yet.
1648 *) Support for companion algorithm specific ASN1 signing routines.
1649 New function ASN1_item_sign_ctx() signs a pre-initialised
1650 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1651 the appropriate parameters.
1654 *) Add new algorithm specific ASN1 verification initialisation function
1655 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1656 handling will be the same no matter what EVP_PKEY_METHOD is used.
1657 Add a PSS handler to support verification of PSS signatures: checked
1658 against a number of sample certificates.
1661 *) Add signature printing for PSS. Add PSS OIDs.
1662 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1664 *) Add algorithm specific signature printing. An individual ASN1 method
1665 can now print out signatures instead of the standard hex dump.
1667 More complex signatures (e.g. PSS) can print out more meaningful
1668 information. Include DSA version that prints out the signature
1672 *) Password based recipient info support for CMS library: implementing
1676 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1677 neatly separates the code into cipher and PBE sections and is required
1678 for some algorithms that split PBES2 into separate pieces (such as
1679 password based CMS).
1682 *) Session-handling fixes:
1683 - Fix handling of connections that are resuming with a session ID,
1684 but also support Session Tickets.
1685 - Fix a bug that suppressed issuing of a new ticket if the client
1686 presented a ticket with an expired session.
1687 - Try to set the ticket lifetime hint to something reasonable.
1688 - Make tickets shorter by excluding irrelevant information.
1689 - On the client side, don't ignore renewed tickets.
1690 [Adam Langley, Bodo Moeller (Google)]
1692 *) Fix PSK session representation.
1695 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1697 This work was sponsored by Intel.
1700 *) Add GCM support to TLS library. Some custom code is needed to split
1701 the IV between the fixed (from PRF) and explicit (from TLS record)
1702 portions. This adds all GCM ciphersuites supported by RFC5288 and
1703 RFC5289. Generalise some AES* cipherstrings to include GCM and
1704 add a special AESGCM string for GCM only.
1707 *) Expand range of ctrls for AES GCM. Permit setting invocation
1708 field on decrypt and retrieval of invocation field only on encrypt.
1711 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1712 As required by RFC5289 these ciphersuites cannot be used if for
1713 versions of TLS earlier than 1.2.
1716 *) For FIPS capable OpenSSL interpret a NULL default public key method
1717 as unset and return the appropriate default but do *not* set the default.
1718 This means we can return the appropriate method in applications that
1719 switch between FIPS and non-FIPS modes.
1722 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1723 ENGINE is used then we cannot handle that in the FIPS module so we
1724 keep original code iff non-FIPS operations are allowed.
1727 *) Add -attime option to openssl utilities.
1728 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1730 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1733 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1734 FIPS EC methods unconditionally for now.
1737 *) New build option no-ec2m to disable characteristic 2 code.
1740 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1741 all cases can be covered as some introduce binary incompatibilities.
1744 *) Redirect RSA operations to FIPS module including keygen,
1745 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1748 *) Add similar low level API blocking to ciphers.
1751 *) Low level digest APIs are not approved in FIPS mode: any attempt
1752 to use these will cause a fatal error. Applications that *really* want
1753 to use them can use the private_* version instead.
1756 *) Redirect cipher operations to FIPS module for FIPS builds.
1759 *) Redirect digest operations to FIPS module for FIPS builds.
1762 *) Update build system to add "fips" flag which will link in fipscanister.o
1763 for static and shared library builds embedding a signature if needed.
1766 *) Output TLS supported curves in preference order instead of numerical
1767 order. This is currently hardcoded for the highest order curves first.
1768 This should be configurable so applications can judge speed vs strength.
1771 *) Add TLS v1.2 server support for client authentication.
1774 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1778 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1779 FIPS modules versions.
1782 *) Add TLS v1.2 client side support for client authentication. Keep cache
1783 of handshake records longer as we don't know the hash algorithm to use
1784 until after the certificate request message is received.
1787 *) Initial TLS v1.2 client support. Add a default signature algorithms
1788 extension including all the algorithms we support. Parse new signature
1789 format in client key exchange. Relax some ECC signing restrictions for
1790 TLS v1.2 as indicated in RFC5246.
1793 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1794 to new signature format when needed using client digest preference.
1795 All server ciphersuites should now work correctly in TLS v1.2. No client
1796 support yet and no support for client certificates.
1799 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1800 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1801 ciphersuites. At present only RSA key exchange ciphersuites work with
1802 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1803 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1804 and version checking.
1807 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1808 with this defined it will not be affected by any changes to ssl internal
1809 structures. Add several utility functions to allow openssl application
1810 to work with OPENSSL_NO_SSL_INTERN defined.
1814 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1816 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1819 *) Permit abbreviated handshakes when renegotiating using the function
1820 SSL_renegotiate_abbreviated().
1821 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1823 *) Add call to ENGINE_register_all_complete() to
1824 ENGINE_load_builtin_engines(), so some implementations get used
1825 automatically instead of needing explicit application support.
1828 *) Add support for TLS key exporter as described in RFC5705.
1829 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1831 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1832 a few changes are required:
1834 Add SSL_OP_NO_TLSv1_1 flag.
1835 Add TLSv1_1 methods.
1836 Update version checking logic to handle version 1.1.
1837 Add explicit IV handling (ported from DTLS code).
1838 Add command line options to s_client/s_server.
1841 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1843 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1844 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1845 content decryption and always return the same error. Note: this attack
1846 needs on average 2^20 messages so it only affects automated senders. The
1847 old behaviour can be reenabled in the CMS code by setting the
1848 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1849 an MMA defence is not necessary.
1850 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1851 this issue. (CVE-2012-0884)
1854 *) Fix CVE-2011-4619: make sure we really are receiving a
1855 client hello before rejecting multiple SGC restarts. Thanks to
1856 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1859 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1861 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1862 Thanks to Antonio Martin, Enterprise Secure Access Research and
1863 Development, Cisco Systems, Inc. for discovering this bug and
1864 preparing a fix. (CVE-2012-0050)
1867 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1869 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1870 of the Vaudenay padding oracle attack on CBC mode encryption
1871 which enables an efficient plaintext recovery attack against
1872 the OpenSSL implementation of DTLS. Their attack exploits timing
1873 differences arising during decryption processing. A research
1874 paper describing this attack can be found at:
1875 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1876 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1877 Security Group at Royal Holloway, University of London
1878 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1879 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1880 for preparing the fix. (CVE-2011-4108)
1881 [Robin Seggelmann, Michael Tuexen]
1883 *) Clear bytes used for block padding of SSL 3.0 records.
1885 [Adam Langley (Google)]
1887 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1888 Kadianakis <desnacked@gmail.com> for discovering this issue and
1889 Adam Langley for preparing the fix. (CVE-2011-4619)
1890 [Adam Langley (Google)]
1892 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1893 [Andrey Kulikov <amdeich@gmail.com>]
1895 *) Prevent malformed RFC3779 data triggering an assertion failure.
1896 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1897 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1898 [Rob Austein <sra@hactrn.net>]
1900 *) Improved PRNG seeding for VOS.
1901 [Paul Green <Paul.Green@stratus.com>]
1903 *) Fix ssl_ciph.c set-up race.
1904 [Adam Langley (Google)]
1906 *) Fix spurious failures in ecdsatest.c.
1907 [Emilia Käsper (Google)]
1909 *) Fix the BIO_f_buffer() implementation (which was mixing different
1910 interpretations of the '..._len' fields).
1911 [Adam Langley (Google)]
1913 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1914 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1915 threads won't reuse the same blinding coefficients.
1917 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1918 lock to call BN_BLINDING_invert_ex, and avoids one use of
1919 BN_BLINDING_update for each BN_BLINDING structure (previously,
1920 the last update always remained unused).
1921 [Emilia Käsper (Google)]
1923 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1924 [Bob Buckholz (Google)]
1926 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1928 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1929 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1930 [Kaspar Brand <ossl@velox.ch>]
1932 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1933 for multi-threaded use of ECDH. (CVE-2011-3210)
1934 [Adam Langley (Google)]
1936 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1939 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1940 signature public key algorithm by using OID xref utilities instead.
1941 Before this you could only use some ECC ciphersuites with SHA1 only.
1944 *) Add protection against ECDSA timing attacks as mentioned in the paper
1945 by Billy Bob Brumley and Nicola Tuveri, see:
1947 http://eprint.iacr.org/2011/232.pdf
1949 [Billy Bob Brumley and Nicola Tuveri]
1951 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1953 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1954 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1956 *) Fix bug in string printing code: if *any* escaping is enabled we must
1957 escape the escape character (backslash) or the resulting string is
1961 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1963 *) Disable code workaround for ancient and obsolete Netscape browsers
1964 and servers: an attacker can use it in a ciphersuite downgrade attack.
1965 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1968 *) Fixed J-PAKE implementation error, originally discovered by
1969 Sebastien Martini, further info and confirmation from Stefan
1970 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1973 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1975 *) Fix extension code to avoid race conditions which can result in a buffer
1976 overrun vulnerability: resumed sessions must not be modified as they can
1977 be shared by multiple threads. CVE-2010-3864
1980 *) Fix WIN32 build system to correctly link an ENGINE directory into
1984 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1986 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1988 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1990 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1992 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1993 context. The operation can be customised via the ctrl mechanism in
1994 case ENGINEs want to include additional functionality.
1997 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2000 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2001 output hashes compatible with older versions of OpenSSL.
2002 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2004 *) Fix compression algorithm handling: if resuming a session use the
2005 compression algorithm of the resumed session instead of determining
2006 it from client hello again. Don't allow server to change algorithm.
2009 *) Add load_crls() function to apps tidying load_certs() too. Add option
2010 to verify utility to allow additional CRLs to be included.
2013 *) Update OCSP request code to permit adding custom headers to the request:
2014 some responders need this.
2017 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2019 [Julia Lawall <julia@diku.dk>]
2021 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2022 needlessly dereferenced structures, used obsolete functions and
2023 didn't handle all updated verify codes correctly.
2026 *) Disable MD2 in the default configuration.
2029 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2030 indicate the initial BIO being pushed or popped. This makes it possible
2031 to determine whether the BIO is the one explicitly called or as a result
2032 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2033 it handles reference counts correctly and doesn't zero out the I/O bio
2034 when it is not being explicitly popped. WARNING: applications which
2035 included workarounds for the old buggy behaviour will need to be modified
2036 or they could free up already freed BIOs.
2039 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2040 renaming to all platforms (within the 0.9.8 branch, this was
2041 done conditionally on Netware platforms to avoid a name clash).
2042 [Guenter <lists@gknw.net>]
2044 *) Add ECDHE and PSK support to DTLS.
2045 [Michael Tuexen <tuexen@fh-muenster.de>]
2047 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2051 *) Add "missing" function EVP_MD_flags() (without this the only way to
2052 retrieve a digest flags is by accessing the structure directly. Update
2053 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2054 or cipher is registered as in the "from" argument. Print out all
2055 registered digests in the dgst usage message instead of manually
2056 attempting to work them out.
2059 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2060 this allows the use of compression and extensions. Change default cipher
2061 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2062 by default unless an application cipher string requests it.
2065 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2066 key ids to find matching certificates and keys but some PKCS#12 files
2067 don't follow the (somewhat unwritten) rules and this strategy fails.
2068 Now just gather all certificates together and the first private key
2069 then look for the first certificate that matches the key.
2072 *) Support use of registered digest and cipher names for dgst and cipher
2073 commands instead of having to add each one as a special case. So now
2080 openssl dgst -sha256 foo
2082 and this works for ENGINE based algorithms too.
2086 *) Update Gost ENGINE to support parameter files.
2087 [Victor B. Wagner <vitus@cryptocom.ru>]
2089 *) Support GeneralizedTime in ca utility.
2090 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2092 *) Enhance the hash format used for certificate directory links. The new
2093 form uses the canonical encoding (meaning equivalent names will work
2094 even if they aren't identical) and uses SHA1 instead of MD5. This form
2095 is incompatible with the older format and as a result c_rehash should
2096 be used to rebuild symbolic links.
2099 *) Make PKCS#8 the default write format for private keys, replacing the
2100 traditional format. This form is standardised, more secure and doesn't
2101 include an implicit MD5 dependency.
2104 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2105 committed to OpenSSL should pass this lot as a minimum.
2108 *) Add session ticket override functionality for use by EAP-FAST.
2109 [Jouni Malinen <j@w1.fi>]
2111 *) Modify HMAC functions to return a value. Since these can be implemented
2112 in an ENGINE errors can occur.
2115 *) Type-checked OBJ_bsearch_ex.
2118 *) Type-checked OBJ_bsearch. Also some constification necessitated
2119 by type-checking. Still to come: TXT_DB, bsearch(?),
2120 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2124 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2125 seconds to a tm structure directly, instead of going through OS
2126 specific date routines. This avoids any issues with OS routines such
2127 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2128 and X509_time_adj_ex() to cover the extended range. The existing
2129 X509_time_adj() is still usable and will no longer have any date issues.
2132 *) Delta CRL support. New use deltas option which will attempt to locate
2133 and search any appropriate delta CRLs available.
2135 This work was sponsored by Google.
2138 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2139 code and add additional score elements. Validate alternate CRL paths
2140 as part of the CRL checking and indicate a new error "CRL path validation
2141 error" in this case. Applications wanting additional details can use
2142 the verify callback and check the new "parent" field. If this is not
2143 NULL CRL path validation is taking place. Existing applications wont
2144 see this because it requires extended CRL support which is off by
2147 This work was sponsored by Google.
2150 *) Support for freshest CRL extension.
2152 This work was sponsored by Google.
2155 *) Initial indirect CRL support. Currently only supported in the CRLs
2156 passed directly and not via lookup. Process certificate issuer
2157 CRL entry extension and lookup CRL entries by bother issuer name
2158 and serial number. Check and process CRL issuer entry in IDP extension.
2160 This work was sponsored by Google.
2163 *) Add support for distinct certificate and CRL paths. The CRL issuer
2164 certificate is validated separately in this case. Only enabled if
2165 an extended CRL support flag is set: this flag will enable additional
2166 CRL functionality in future.
2168 This work was sponsored by Google.
2171 *) Add support for policy mappings extension.
2173 This work was sponsored by Google.
2176 *) Fixes to pathlength constraint, self issued certificate handling,
2177 policy processing to align with RFC3280 and PKITS tests.
2179 This work was sponsored by Google.
2182 *) Support for name constraints certificate extension. DN, email, DNS
2183 and URI types are currently supported.
2185 This work was sponsored by Google.
2188 *) To cater for systems that provide a pointer-based thread ID rather
2189 than numeric, deprecate the current numeric thread ID mechanism and
2190 replace it with a structure and associated callback type. This
2191 mechanism allows a numeric "hash" to be extracted from a thread ID in
2192 either case, and on platforms where pointers are larger than 'long',
2193 mixing is done to help ensure the numeric 'hash' is usable even if it
2194 can't be guaranteed unique. The default mechanism is to use "&errno"
2195 as a pointer-based thread ID to distinguish between threads.
2197 Applications that want to provide their own thread IDs should now use
2198 CRYPTO_THREADID_set_callback() to register a callback that will call
2199 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2201 Note that ERR_remove_state() is now deprecated, because it is tied
2202 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2203 to free the current thread's error state should be replaced by
2204 ERR_remove_thread_state(NULL).
2206 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2207 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2208 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2209 application was previously providing a numeric thread callback that
2210 was inappropriate for distinguishing threads, then uniqueness might
2211 have been obtained with &errno that happened immediately in the
2212 intermediate development versions of OpenSSL; this is no longer the
2213 case, the numeric thread callback will now override the automatic use
2215 [Geoff Thorpe, with help from Bodo Moeller]
2217 *) Initial support for different CRL issuing certificates. This covers a
2218 simple case where the self issued certificates in the chain exist and
2219 the real CRL issuer is higher in the existing chain.
2221 This work was sponsored by Google.
2224 *) Removed effectively defunct crypto/store from the build.
2227 *) Revamp of STACK to provide stronger type-checking. Still to come:
2228 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2229 ASN1_STRING, CONF_VALUE.
2232 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2233 RAM on SSL connections. This option can save about 34k per idle SSL.
2236 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2237 STACK, TXT_DB, bsearch, qsort.
2240 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2241 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2242 support for data, signedData, compressedData, digestedData and
2243 encryptedData, envelopedData types included. Scripts to check against
2244 RFC4134 examples draft and interop and consistency checks of many
2245 content types and variants.
2248 *) Add options to enc utility to support use of zlib compression BIO.
2251 *) Extend mk1mf to support importing of options and assembly language
2252 files from Configure script, currently only included in VC-WIN32.
2253 The assembly language rules can now optionally generate the source
2254 files from the associated perl scripts.
2257 *) Implement remaining functionality needed to support GOST ciphersuites.
2258 Interop testing has been performed using CryptoPro implementations.
2259 [Victor B. Wagner <vitus@cryptocom.ru>]
2261 *) s390x assembler pack.
2264 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2268 *) Implement Opaque PRF Input TLS extension as specified in
2269 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2270 official specification yet and no extension type assignment by
2271 IANA exists, this extension (for now) will have to be explicitly
2272 enabled when building OpenSSL by providing the extension number
2273 to use. For example, specify an option
2275 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2277 to the "config" or "Configure" script to enable the extension,
2278 assuming extension number 0x9527 (which is a completely arbitrary
2279 and unofficial assignment based on the MD5 hash of the Internet
2280 Draft). Note that by doing so, you potentially lose
2281 interoperability with other TLS implementations since these might
2282 be using the same extension number for other purposes.
2284 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2285 opaque PRF input value to use in the handshake. This will create
2286 an interal copy of the length-'len' string at 'src', and will
2287 return non-zero for success.
2289 To get more control and flexibility, provide a callback function
2292 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2293 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2297 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2300 Callback function 'cb' will be called in handshakes, and is
2301 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2302 Argument 'arg' is for application purposes (the value as given to
2303 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2304 be provided to the callback function). The callback function
2305 has to return non-zero to report success: usually 1 to use opaque
2306 PRF input just if possible, or 2 to enforce use of the opaque PRF
2307 input. In the latter case, the library will abort the handshake
2308 if opaque PRF input is not successfully negotiated.
2310 Arguments 'peerinput' and 'len' given to the callback function
2311 will always be NULL and 0 in the case of a client. A server will
2312 see the client's opaque PRF input through these variables if
2313 available (NULL and 0 otherwise). Note that if the server
2314 provides an opaque PRF input, the length must be the same as the
2315 length of the client's opaque PRF input.
2317 Note that the callback function will only be called when creating
2318 a new session (session resumption can resume whatever was
2319 previously negotiated), and will not be called in SSL 2.0
2320 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2321 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2322 for applications that need to enforce opaque PRF input.
2326 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2329 [Victor B. Wagner <vitus@cryptocom.ru>]
2331 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2332 RFC4507bis. The encrypted ticket format is an encrypted encoded
2333 SSL_SESSION structure, that way new session features are automatically
2336 If a client application caches session in an SSL_SESSION structure
2337 support is transparent because tickets are now stored in the encoded
2340 The SSL_CTX structure automatically generates keys for ticket
2341 protection in servers so again support should be possible
2342 with no application modification.
2344 If a client or server wishes to disable RFC4507 support then the option
2345 SSL_OP_NO_TICKET can be set.
2347 Add a TLS extension debugging callback to allow the contents of any client
2348 or server extensions to be examined.
2350 This work was sponsored by Google.
2353 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2354 OpenSSL should now compile cleanly on gcc 4.2
2355 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2357 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2358 support including streaming MAC support: this is required for GOST
2359 ciphersuite support.
2360 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2362 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2363 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2364 to output in BER and PEM format.
2367 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2368 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2369 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2370 ENGINE support for HMAC keys which are unextractable. New -mac and
2371 -macopt options to dgst utility.
2374 *) New option -sigopt to dgst utility. Update dgst to use
2375 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2376 alternative signing paramaters such as X9.31 or PSS in the dgst
2380 *) Change ssl_cipher_apply_rule(), the internal function that does
2381 the work each time a ciphersuite string requests enabling
2382 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2383 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2384 the order of disabled ciphersuites such that those ciphersuites
2385 that most recently went from enabled to disabled not only stay
2386 in order with respect to each other, but also have higher priority
2387 than other disabled ciphersuites the next time ciphersuites are
2390 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2391 the same ciphersuites as with "HIGH" alone, but in a specific
2392 order where the PSK ciphersuites come first (since they are the
2393 most recently disabled ciphersuites when "HIGH" is parsed).
2395 Also, change ssl_create_cipher_list() (using this new
2396 funcionality) such that between otherwise identical
2397 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2401 *) Change ssl_create_cipher_list() so that it automatically
2402 arranges the ciphersuites in reasonable order before starting
2403 to process the rule string. Thus, the definition for "DEFAULT"
2404 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2405 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2406 This makes it much easier to arrive at a reasonable default order
2407 in applications for which anonymous ciphers are OK (meaning
2408 that you can't actually use DEFAULT).
2409 [Bodo Moeller; suggested by Victor Duchovni]
2411 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2412 processing) into multiple integers instead of setting
2413 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2414 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2415 (These masks as well as the individual bit definitions are hidden
2416 away into the non-exported interface ssl/ssl_locl.h, so this
2417 change to the definition of the SSL_CIPHER structure shouldn't
2418 affect applications.) This give us more bits for each of these
2419 categories, so there is no longer a need to coagulate AES128 and
2420 AES256 into a single algorithm bit, and to coagulate Camellia128
2421 and Camellia256 into a single algorithm bit, which has led to all
2424 Thus, among other things, the kludge introduced in 0.9.7m and
2425 0.9.8e for masking out AES256 independently of AES128 or masking
2426 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2428 With the change, we also introduce new ciphersuite aliases that
2429 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2433 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2434 Use the leftmost N bytes of the signature input if the input is
2435 larger than the prime q (with N being the size in bytes of q).
2438 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2439 it yet and it is largely untested.
2442 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2445 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2446 some compilers (gcc 4.2 and later) reject their use. Safestack is
2447 reimplemented. Update ASN1 to avoid use of legacy functions.
2450 *) Win32/64 targets are linked with Winsock2.
2453 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2454 to external functions. This can be used to increase CRL handling
2455 efficiency especially when CRLs are very large by (for example) storing
2456 the CRL revoked certificates in a database.
2459 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2460 new CRLs added to a directory can be used. New command line option
2461 -verify_return_error to s_client and s_server. This causes real errors
2462 to be returned by the verify callback instead of carrying on no matter
2463 what. This reflects the way a "real world" verify callback would behave.
2466 *) GOST engine, supporting several GOST algorithms and public key formats.
2467 Kindly donated by Cryptocom.
2470 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2471 partitioned by DP are handled but no indirect CRL or reason partitioning
2472 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2473 selected via a scoring technique which handles IDP and AKID in CRLs.
2476 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2477 will ultimately be used for all verify operations: this will remove the
2478 X509_STORE dependency on certificate verification and allow alternative
2479 lookup methods. X509_STORE based implementations of these two callbacks.
2482 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2483 Modify get_crl() to find a valid (unexpired) CRL if possible.
2486 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2487 this would be called X509_CRL_cmp() but that name is already used by
2488 a function that just compares CRL issuer names. Cache several CRL
2489 extensions in X509_CRL structure and cache CRLDP in X509.
2492 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2493 this maps equivalent X509_NAME structures into a consistent structure.
2494 Name comparison can then be performed rapidly using memcmp().
2497 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2501 *) Allow digests to supply their own micalg string for S/MIME type using
2502 the ctrl EVP_MD_CTRL_MICALG.
2505 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2506 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2507 ctrl. It can then customise the structure before and/or after signing
2511 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2512 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2513 to free up any added signature OIDs.
2516 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2517 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2518 digest and cipher tables. New options added to openssl utility:
2519 list-message-digest-algorithms and list-cipher-algorithms.
2522 *) Change the array representation of binary polynomials: the list
2523 of degrees of non-zero coefficients is now terminated with -1.
2524 Previously it was terminated with 0, which was also part of the
2525 value; thus, the array representation was not applicable to
2526 polynomials where t^0 has coefficient zero. This change makes
2527 the array representation useful in a more general context.
2530 *) Various modifications and fixes to SSL/TLS cipher string
2531 handling. For ECC, the code now distinguishes between fixed ECDH
2532 with RSA certificates on the one hand and with ECDSA certificates
2533 on the other hand, since these are separate ciphersuites. The
2534 unused code for Fortezza ciphersuites has been removed.
2536 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2537 (not "ECDHE"). For consistency with the code for DH
2538 certificates, use of ECDH certificates is now considered ECDH
2539 authentication, not RSA or ECDSA authentication (the latter is
2540 merely the CA's signing algorithm and not actively used in the
2543 The temporary ciphersuite alias "ECCdraft" is no longer
2544 available, and ECC ciphersuites are no longer excluded from "ALL"
2545 and "DEFAULT". The following aliases now exist for RFC 4492
2546 ciphersuites, most of these by analogy with the DH case:
2548 kECDHr - ECDH cert, signed with RSA
2549 kECDHe - ECDH cert, signed with ECDSA
2550 kECDH - ECDH cert (signed with either RSA or ECDSA)
2551 kEECDH - ephemeral ECDH
2552 ECDH - ECDH cert or ephemeral ECDH
2558 AECDH - anonymous ECDH
2559 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2563 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2564 Use correct micalg parameters depending on digest(s) in signed message.
2567 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2568 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2571 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2572 an engine to register a method. Add ENGINE lookups for methods and
2573 functional reference processing.
2576 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2577 EVP_{Sign,Verify}* which allow an application to customise the signature
2581 *) New -resign option to smime utility. This adds one or more signers
2582 to an existing PKCS#7 signedData structure. Also -md option to use an
2583 alternative message digest algorithm for signing.
2586 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2587 create PKCS7 structures containing multiple signers. Update smime
2588 application to support multiple signers.
2591 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2595 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2596 Reorganize PBE internals to lookup from a static table using NIDs,
2597 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2598 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2599 PRF which will be automatically used with PBES2.
2602 *) Replace the algorithm specific calls to generate keys in "req" with the
2606 *) Update PKCS#7 enveloped data routines to use new API. This is now
2607 supported by any public key method supporting the encrypt operation. A
2608 ctrl is added to allow the public key algorithm to examine or modify
2609 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2613 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2614 a default digest type to use. In most cases this will be SHA1 but some
2615 algorithms (such as GOST) need to specify an alternative digest. The
2616 return value indicates how strong the preference is 1 means optional and
2617 2 is mandatory (that is it is the only supported type). Modify
2618 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2619 use the default md. Update openssl utilities to use the default digest
2620 type for signing if it is not explicitly indicated.
2623 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2624 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2625 signing method from the key type. This effectively removes the link
2626 between digests and public key types.
2629 *) Add an OID cross reference table and utility functions. Its purpose is to
2630 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2631 rsaEncryption. This will allow some of the algorithm specific hackery
2632 needed to use the correct OID to be removed.
2635 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2636 structures for PKCS7_sign(). They are now set up by the relevant public
2640 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2643 *) Add support for key derivation (agreement) in the API, DH method and
2647 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2648 public and private key formats. As a side effect these add additional
2649 command line functionality not previously available: DSA signatures can be
2650 generated and verified using pkeyutl and DH key support and generation in
2655 [Oliver Tappe <zooey@hirschkaefer.de>]
2657 *) New make target "install_html_docs" installs HTML renditions of the
2659 [Oliver Tappe <zooey@hirschkaefer.de>]
2661 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2662 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2663 support key and parameter generation and add initial key generation
2664 functionality for RSA.
2667 *) Add functions for main EVP_PKEY_method operations. The undocumented
2668 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2669 EVP_PKEY_{encrypt,decrypt}_old.
2672 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2673 key API, doesn't do much yet.
2676 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2677 public key algorithms. New option to openssl utility:
2678 "list-public-key-algorithms" to print out info.
2681 *) Implement the Supported Elliptic Curves Extension for
2682 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2685 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2686 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2689 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2690 utilities such as rsa, dsa, dsaparam etc except they process any key
2694 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2695 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2696 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2700 *) Initial support for pluggable public key ASN1.
2701 De-spaghettify the public key ASN1 handling. Move public and private
2702 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2703 algorithm specific handling to a single module within the relevant
2704 algorithm directory. Add functions to allow (near) opaque processing
2705 of public and private key structures.
2708 *) Implement the Supported Point Formats Extension for
2709 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2712 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2713 for the psk identity [hint] and the psk callback functions to the
2714 SSL_SESSION, SSL and SSL_CTX structure.
2717 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2721 SSL_CTX_use_psk_identity_hint
2722 SSL_get_psk_identity_hint
2723 SSL_get_psk_identity
2724 SSL_use_psk_identity_hint
2726 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2728 *) Add RFC 3161 compliant time stamp request creation, response generation
2729 and response verification functionality.
2730 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2732 *) Add initial support for TLS extensions, specifically for the server_name
2733 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2734 have new members for a host name. The SSL data structure has an
2735 additional member SSL_CTX *initial_ctx so that new sessions can be
2736 stored in that context to allow for session resumption, even after the
2737 SSL has been switched to a new SSL_CTX in reaction to a client's
2738 server_name extension.
2740 New functions (subject to change):
2742 SSL_get_servername()
2743 SSL_get_servername_type()
2746 New CTRL codes and macros (subject to change):
2748 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2749 - SSL_CTX_set_tlsext_servername_callback()
2750 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2751 - SSL_CTX_set_tlsext_servername_arg()
2752 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2754 openssl s_client has a new '-servername ...' option.
2756 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2757 '-key2 ...', '-servername_fatal' (subject to change). This allows
2758 testing the HostName extension for a specific single host name ('-cert'
2759 and '-key' remain fallbacks for handshakes without HostName
2760 negotiation). If the unrecognized_name alert has to be sent, this by
2761 default is a warning; it becomes fatal with the '-servername_fatal'
2764 [Peter Sylvester, Remy Allais, Christophe Renou]
2766 *) Whirlpool hash implementation is added.
2769 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2770 bn(64,32). Because of instruction set limitations it doesn't have
2771 any negative impact on performance. This was done mostly in order
2772 to make it possible to share assembler modules, such as bn_mul_mont
2773 implementations, between 32- and 64-bit builds without hassle.
2776 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2777 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2781 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2782 dedicated Montgomery multiplication procedure, is introduced.
2783 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2784 "64-bit" performance on certain 32-bit targets.
2787 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2788 in SSL structures. New SSL ctrl to set maximum send fragment size.
2789 Save memory by seeting the I/O buffer sizes dynamically instead of
2790 using the maximum available value.
2793 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2794 in addition to the text details.
2797 *) Very, very preliminary EXPERIMENTAL support for printing of general
2798 ASN1 structures. This currently produces rather ugly output and doesn't
2799 handle several customised structures at all.
2802 *) Integrated support for PVK file format and some related formats such
2803 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2804 these in the 'rsa' and 'dsa' utilities.
2807 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2810 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2811 place for the (very old) "NETSCAPE" format certificates which are now
2812 handled using new ASN1 code equivalents.
2815 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2816 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2817 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2820 *) Modify CRL distribution points extension code to print out previously
2821 unsupported fields. Enhance extension setting code to allow setting of
2825 *) Add print and set support for Issuing Distribution Point CRL extension.
2828 *) Change 'Configure' script to enable Camellia by default.
2831 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2833 *) When rejecting SSL/TLS records due to an incorrect version number, never
2834 update s->server with a new major version number. As of
2835 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2836 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2837 the previous behavior could result in a read attempt at NULL when
2838 receiving specific incorrect SSL/TLS records once record payload
2839 protection is active. (CVE-2010-0740)
2840 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2842 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2843 could be crashed if the relevant tables were not present (e.g. chrooted).
2844 [Tomas Hoger <thoger@redhat.com>]
2846 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2848 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2849 [Martin Olsson, Neel Mehta]
2851 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2852 accommodate for stack sorting, always a write lock!).
2855 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2856 excessive delays in the RAND_poll(): over a minute. As a workaround
2857 include a time check in the inner Heap32Next loop too.
2860 *) The code that handled flushing of data in SSL/TLS originally used the
2861 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2862 the problem outlined in PR#1949. The fix suggested there however can
2863 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2864 of Apache). So instead simplify the code to flush unconditionally.
2865 This should be fine since flushing with no data to flush is a no op.
2868 *) Handle TLS versions 2.0 and later properly and correctly use the
2869 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2870 off ancient servers have a habit of sticking around for a while...
2873 *) Modify compression code so it frees up structures without using the
2874 ex_data callbacks. This works around a problem where some applications
2875 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2876 restarting) then use compression (e.g. SSL with compression) later.
2877 This results in significant per-connection memory leaks and
2878 has caused some security issues including CVE-2008-1678 and
2882 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2883 change when encrypting or decrypting.
2886 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2887 connect and renegotiate with servers which do not support RI.
2888 Until RI is more widely deployed this option is enabled by default.
2891 *) Add "missing" ssl ctrls to clear options and mode.
2894 *) If client attempts to renegotiate and doesn't support RI respond with
2895 a no_renegotiation alert as required by RFC5746. Some renegotiating
2896 TLS clients will continue a connection gracefully when they receive
2897 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2898 waiting for a server hello which it will never receive. Now we treat a
2899 received no_renegotiation alert as a fatal error. This is because
2900 applications requesting a renegotiation might well expect it to succeed
2901 and would have no code in place to handle the server denying it so the
2902 only safe thing to do is to terminate the connection.
2905 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2906 peer supports secure renegotiation and 0 otherwise. Print out peer
2907 renegotiation support in s_client/s_server.
2910 *) Replace the highly broken and deprecated SPKAC certification method with
2911 the updated NID creation version. This should correctly handle UTF8.
2914 *) Implement RFC5746. Re-enable renegotiation but require the extension
2915 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2916 turns out to be a bad idea. It has been replaced by
2917 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2918 SSL_CTX_set_options(). This is really not recommended unless you
2919 know what you are doing.
2920 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2922 *) Fixes to stateless session resumption handling. Use initial_ctx when
2923 issuing and attempting to decrypt tickets in case it has changed during
2924 servername handling. Use a non-zero length session ID when attempting
2925 stateless session resumption: this makes it possible to determine if
2926 a resumption has occurred immediately after receiving server hello
2927 (several places in OpenSSL subtly assume this) instead of later in
2931 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2932 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2933 fixes for a few places where the return code is not checked
2935 [Julia Lawall <julia@diku.dk>]
2937 *) Add --strict-warnings option to Configure script to include devteam
2938 warnings in other configurations.
2941 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2942 makes it possible to install openssl libraries in locations which
2943 have names other than "lib", for example "/usr/lib64" which some
2945 [Steve Henson, based on patch from Jeremy Utley]
2947 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2948 X690 8.9.12 and can produce some misleading textual output of OIDs.
2949 [Steve Henson, reported by Dan Kaminsky]
2951 *) Delete MD2 from algorithm tables. This follows the recommendation in
2952 several standards that it is not used in new applications due to
2953 several cryptographic weaknesses. For binary compatibility reasons
2954 the MD2 API is still compiled in by default.
2957 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2961 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2962 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2964 [Guenter <lists@gknw.net>]
2966 *) Fix the server certificate chain building code to use X509_verify_cert(),
2967 it used to have an ad-hoc builder which was unable to cope with anything
2968 other than a simple chain.
2969 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2971 *) Don't check self signed certificate signatures in X509_verify_cert()
2972 by default (a flag can override this): it just wastes time without
2973 adding any security. As a useful side effect self signed root CAs
2974 with non-FIPS digests are now usable in FIPS mode.
2977 *) In dtls1_process_out_of_seq_message() the check if the current message
2978 is already buffered was missing. For every new message was memory
2979 allocated, allowing an attacker to perform an denial of service attack
2980 with sending out of seq handshake messages until there is no memory
2981 left. Additionally every future messege was buffered, even if the
2982 sequence number made no sense and would be part of another handshake.
2983 So only messages with sequence numbers less than 10 in advance will be
2984 buffered. (CVE-2009-1378)
2985 [Robin Seggelmann, discovered by Daniel Mentz]
2987 *) Records are buffered if they arrive with a future epoch to be
2988 processed after finishing the corresponding handshake. There is
2989 currently no limitation to this buffer allowing an attacker to perform
2990 a DOS attack with sending records with future epochs until there is no
2991 memory left. This patch adds the pqueue_size() function to determine
2992 the size of a buffer and limits the record buffer to 100 entries.
2994 [Robin Seggelmann, discovered by Daniel Mentz]
2996 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2997 parent structure is freed. (CVE-2009-1379)
3000 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3001 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3004 [Ilya O. <vrghost@gmail.com>]
3006 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3008 *) Disable renegotiation completely - this fixes a severe security
3009 problem (CVE-2009-3555) at the cost of breaking all
3010 renegotiation. Renegotiation can be re-enabled by setting
3011 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3012 run-time. This is really not recommended unless you know what
3016 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3018 *) Don't set val to NULL when freeing up structures, it is freed up by
3019 underlying code. If sizeof(void *) > sizeof(long) this can result in
3020 zeroing past the valid field. (CVE-2009-0789)
3021 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3023 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3024 checked correctly. This would allow some invalid signed attributes to
3025 appear to verify correctly. (CVE-2009-0591)
3026 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3028 *) Reject UniversalString and BMPString types with invalid lengths. This
3029 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3030 a legal length. (CVE-2009-0590)
3033 *) Set S/MIME signing as the default purpose rather than setting it
3034 unconditionally. This allows applications to override it at the store
3038 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3039 to handle some structures.
3042 *) Improve efficiency of mem_gets: don't search whole buffer each time
3044 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3046 *) New -hex option for openssl rand.
3049 *) Print out UTF8String and NumericString when parsing ASN1.
3052 *) Support NumericString type for name components.
3055 *) Allow CC in the environment to override the automatically chosen
3056 compiler. Note that nothing is done to ensure flags work with the
3060 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3062 *) Properly check EVP_VerifyFinal() and similar return values
3064 [Ben Laurie, Bodo Moeller, Google Security Team]
3066 *) Enable TLS extensions by default.
3069 *) Allow the CHIL engine to be loaded, whether the application is
3070 multithreaded or not. (This does not release the developer from the
3071 obligation to set up the dynamic locking callbacks.)
3072 [Sander Temme <sander@temme.net>]
3074 *) Use correct exit code if there is an error in dgst command.
3075 [Steve Henson; problem pointed out by Roland Dirlewanger]
3077 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3078 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3081 *) Add experimental JPAKE support, including demo authentication in
3082 s_client and s_server.
3085 *) Set the comparison function in v3_addr_canonize().
3086 [Rob Austein <sra@hactrn.net>]
3088 *) Add support for XMPP STARTTLS in s_client.
3089 [Philip Paeps <philip@freebsd.org>]
3091 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3092 to ensure that even with this option, only ciphersuites in the
3093 server's preference list will be accepted. (Note that the option
3094 applies only when resuming a session, so the earlier behavior was
3095 just about the algorithm choice for symmetric cryptography.)
3098 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3100 *) Fix NULL pointer dereference if a DTLS server received
3101 ChangeCipherSpec as first record (CVE-2009-1386).
3104 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3105 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3108 *) The fix in 0.9.8c that supposedly got rid of unsafe
3109 double-checked locking was incomplete for RSA blinding,
3110 addressing just one layer of what turns out to have been
3111 doubly unsafe triple-checked locking.
3113 So now fix this for real by retiring the MONT_HELPER macro
3114 in crypto/rsa/rsa_eay.c.
3116 [Bodo Moeller; problem pointed out by Marius Schilder]
3118 *) Various precautionary measures:
3120 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3122 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3123 (NB: This would require knowledge of the secret session ticket key
3124 to exploit, in which case you'd be SOL either way.)
3126 - Change bn_nist.c so that it will properly handle input BIGNUMs
3127 outside the expected range.
3129 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3132 [Neel Mehta, Bodo Moeller]
3134 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3135 the load fails. Useful for distros.
3136 [Ben Laurie and the FreeBSD team]
3138 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3141 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3144 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3146 This work was sponsored by Logica.
3149 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3150 keystores. Support for SSL/TLS client authentication too.
3151 Not compiled unless enable-capieng specified to Configure.
3153 This work was sponsored by Logica.
3156 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3157 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3158 attribute creation routines such as certifcate requests and PKCS#12
3162 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3164 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3165 handshake which could lead to a cilent crash as found using the
3166 Codenomicon TLS test suite (CVE-2008-1672)
3167 [Steve Henson, Mark Cox]
3169 *) Fix double free in TLS server name extensions which could lead to
3170 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3173 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3175 Clear the error queue to ensure that error entries left from
3176 older function calls do not interfere with the correct operation.
3177 [Lutz Jaenicke, Erik de Castro Lopo]
3179 *) Remove root CA certificates of commercial CAs:
3181 The OpenSSL project does not recommend any specific CA and does not
3182 have any policy with respect to including or excluding any CA.
3183 Therefore it does not make any sense to ship an arbitrary selection
3184 of root CA certificates with the OpenSSL software.
3187 *) RSA OAEP patches to fix two separate invalid memory reads.
3188 The first one involves inputs when 'lzero' is greater than
3189 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3190 before the beginning of from). The second one involves inputs where
3191 the 'db' section contains nothing but zeroes (there is a one-byte
3192 invalid read after the end of 'db').
3193 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3195 *) Partial backport from 0.9.9-dev:
3197 Introduce bn_mul_mont (dedicated Montgomery multiplication
3198 procedure) as a candidate for BIGNUM assembler implementation.
3199 While 0.9.9-dev uses assembler for various architectures, only
3200 x86_64 is available by default here in the 0.9.8 branch, and
3201 32-bit x86 is available through a compile-time setting.
3203 To try the 32-bit x86 assembler implementation, use Configure
3204 option "enable-montasm" (which exists only for this backport).
3206 As "enable-montasm" for 32-bit x86 disclaims code stability
3207 anyway, in this constellation we activate additional code
3208 backported from 0.9.9-dev for further performance improvements,
3209 namely BN_from_montgomery_word. (To enable this otherwise,
3210 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3212 [Andy Polyakov (backport partially by Bodo Moeller)]
3214 *) Add TLS session ticket callback. This allows an application to set
3215 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3216 values. This is useful for key rollover for example where several key
3217 sets may exist with different names.
3220 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3221 This was broken until now in 0.9.8 releases, such that the only way
3222 a registered ENGINE could be used (assuming it initialises
3223 successfully on the host) was to explicitly set it as the default
3224 for the relevant algorithms. This is in contradiction with 0.9.7
3225 behaviour and the documentation. With this fix, when an ENGINE is
3226 registered into a given algorithm's table of implementations, the
3227 'uptodate' flag is reset so that auto-discovery will be used next
3228 time a new context for that algorithm attempts to select an
3230 [Ian Lister (tweaked by Geoff Thorpe)]
3232 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3233 implemention in the following ways:
3235 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3238 Lack of BER streaming support means one pass streaming processing is
3239 only supported if data is detached: setting the streaming flag is
3240 ignored for embedded content.
3242 CMS support is disabled by default and must be explicitly enabled
3243 with the enable-cms configuration option.
3246 *) Update the GMP engine glue to do direct copies between BIGNUM and
3247 mpz_t when openssl and GMP use the same limb size. Otherwise the
3248 existing "conversion via a text string export" trick is still used.
3249 [Paul Sheer <paulsheer@gmail.com>]
3251 *) Zlib compression BIO. This is a filter BIO which compressed and
3252 uncompresses any data passed through it.
3255 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3256 RFC3394 compatible AES key wrapping.
3259 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3260 sets string data without copying. X509_ALGOR_set0() and
3261 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3262 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3263 from an X509_ATTRIBUTE structure optionally checking it occurs only
3264 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3268 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3269 to get the expected BN_FLG_CONSTTIME behavior.
3270 [Bodo Moeller (Google)]
3274 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3275 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3276 - added some more tests to do_tests.pl
3277 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3278 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3279 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3280 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3281 - various changes to netware.pl to enable gcc-cross builds on Win32
3283 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3284 - various changes to fix missing prototype warnings
3285 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3286 - added AES, WHIRLPOOL and CPUID assembler code to build files
3287 - added missing AES assembler make rules to mk1mf.pl
3288 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3289 [Guenter Knauf <eflash@gmx.net>]
3291 *) Implement certificate status request TLS extension defined in RFC3546.
3292 A client can set the appropriate parameters and receive the encoded
3293 OCSP response via a callback. A server can query the supplied parameters
3294 and set the encoded OCSP response in the callback. Add simplified examples
3295 to s_client and s_server.
3298 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3300 *) Fix various bugs:
3301 + Binary incompatibility of ssl_ctx_st structure
3302 + DTLS interoperation with non-compliant servers
3303 + Don't call get_session_cb() without proposed session
3304 + Fix ia64 assembler code
3305 [Andy Polyakov, Steve Henson]
3307 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3309 *) DTLS Handshake overhaul. There were longstanding issues with
3310 OpenSSL DTLS implementation, which were making it impossible for
3311 RFC 4347 compliant client to communicate with OpenSSL server.
3312 Unfortunately just fixing these incompatibilities would "cut off"
3313 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3314 server keeps tolerating non RFC compliant syntax. The opposite is
3315 not true, 0.9.8f client can not communicate with earlier server.
3316 This update even addresses CVE-2007-4995.
3319 *) Changes to avoid need for function casts in OpenSSL: some compilers
3320 (gcc 4.2 and later) reject their use.
3321 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3324 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3325 RFC4507bis. The encrypted ticket format is an encrypted encoded
3326 SSL_SESSION structure, that way new session features are automatically
3329 If a client application caches session in an SSL_SESSION structure
3330 support is transparent because tickets are now stored in the encoded
3333 The SSL_CTX structure automatically generates keys for ticket
3334 protection in servers so again support should be possible
3335 with no application modification.
3337 If a client or server wishes to disable RFC4507 support then the option
3338 SSL_OP_NO_TICKET can be set.
3340 Add a TLS extension debugging callback to allow the contents of any client
3341 or server extensions to be examined.
3343 This work was sponsored by Google.
3346 *) Add initial support for TLS extensions, specifically for the server_name
3347 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3348 have new members for a host name. The SSL data structure has an
3349 additional member SSL_CTX *initial_ctx so that new sessions can be
3350 stored in that context to allow for session resumption, even after the
3351 SSL has been switched to a new SSL_CTX in reaction to a client's
3352 server_name extension.
3354 New functions (subject to change):
3356 SSL_get_servername()
3357 SSL_get_servername_type()
3360 New CTRL codes and macros (subject to change):
3362 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3363 - SSL_CTX_set_tlsext_servername_callback()
3364 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3365 - SSL_CTX_set_tlsext_servername_arg()
3366 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3368 openssl s_client has a new '-servername ...' option.
3370 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3371 '-key2 ...', '-servername_fatal' (subject to change). This allows
3372 testing the HostName extension for a specific single host name ('-cert'
3373 and '-key' remain fallbacks for handshakes without HostName
3374 negotiation). If the unrecognized_name alert has to be sent, this by
3375 default is a warning; it becomes fatal with the '-servername_fatal'
3378 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3380 *) Add AES and SSE2 assembly language support to VC++ build.
3383 *) Mitigate attack on final subtraction in Montgomery reduction.
3386 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3387 (which previously caused an internal error).
3390 *) Squeeze another 10% out of IGE mode when in != out.
3393 *) AES IGE mode speedup.
3394 [Dean Gaudet (Google)]
3396 *) Add the Korean symmetric 128-bit cipher SEED (see
3397 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3398 add SEED ciphersuites from RFC 4162:
3400 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3401 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3402 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3403 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3405 To minimize changes between patchlevels in the OpenSSL 0.9.8
3406 series, SEED remains excluded from compilation unless OpenSSL
3407 is configured with 'enable-seed'.
3408 [KISA, Bodo Moeller]
3410 *) Mitigate branch prediction attacks, which can be practical if a
3411 single processor is shared, allowing a spy process to extract
3412 information. For detailed background information, see
3413 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3414 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3415 and Necessary Software Countermeasures"). The core of the change
3416 are new versions BN_div_no_branch() and
3417 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3418 respectively, which are slower, but avoid the security-relevant
3419 conditional branches. These are automatically called by BN_div()
3420 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3421 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3422 remove a conditional branch.
3424 BN_FLG_CONSTTIME is the new name for the previous
3425 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3426 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3427 in the exponent causes BN_mod_exp_mont() to use the alternative
3428 implementation in BN_mod_exp_mont_consttime().) The old name
3429 remains as a deprecated alias.
3431 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3432 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3433 constant-time implementations for more than just exponentiation.
3434 Here too the old name is kept as a deprecated alias.
3436 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3437 the BN_BLINDING structure gets an independent copy of the
3438 modulus. This means that the previous "BIGNUM *m" argument to
3439 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3440 essentially becomes "const BIGNUM *m", although we can't actually
3441 change this in the header file before 0.9.9. It allows
3442 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3443 enable BN_FLG_CONSTTIME.
3445 [Matthew D Wood (Intel Corp)]
3447 *) In the SSL/TLS server implementation, be strict about session ID
3448 context matching (which matters if an application uses a single
3449 external cache for different purposes). Previously,
3450 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3451 set. This did ensure strict client verification, but meant that,
3452 with applications using a single external cache for quite
3453 different requirements, clients could circumvent ciphersuite
3454 restrictions for a given session ID context by starting a session
3455 in a different context.
3458 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3459 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3460 authentication-only ciphersuites.
3463 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3464 not complete and could lead to a possible single byte overflow
3465 (CVE-2007-5135) [Ben Laurie]
3467 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3469 *) Since AES128 and AES256 (and similarly Camellia128 and
3470 Camellia256) share a single mask bit in the logic of
3471 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3472 kludge to work properly if AES128 is available and AES256 isn't
3473 (or if Camellia128 is available and Camellia256 isn't).
3476 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3477 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3478 When a point or a seed is encoded in a BIT STRING, we need to
3479 prevent the removal of trailing zero bits to get the proper DER
3480 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3481 of a NamedBitList, for which trailing 0 bits need to be removed.)
3484 *) Have SSL/TLS server implementation tolerate "mismatched" record
3485 protocol version while receiving ClientHello even if the
3486 ClientHello is fragmented. (The server can't insist on the
3487 particular protocol version it has chosen before the ServerHello
3488 message has informed the client about his choice.)
3491 *) Add RFC 3779 support.
3492 [Rob Austein for ARIN, Ben Laurie]
3494 *) Load error codes if they are not already present instead of using a
3495 static variable. This allows them to be cleanly unloaded and reloaded.
3496 Improve header file function name parsing.
3499 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3500 or CAPABILITY handshake as required by RFCs.
3503 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3505 *) Introduce limits to prevent malicious keys being able to
3506 cause a denial of service. (CVE-2006-2940)
3507 [Steve Henson, Bodo Moeller]
3509 *) Fix ASN.1 parsing of certain invalid structures that can result
3510 in a denial of service. (CVE-2006-2937) [Steve Henson]
3512 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3513 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3515 *) Fix SSL client code which could crash if connecting to a
3516 malicious SSLv2 server. (CVE-2006-4343)
3517 [Tavis Ormandy and Will Drewry, Google Security Team]
3519 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3520 match only those. Before that, "AES256-SHA" would be interpreted
3521 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3522 the same strength classification in 0.9.7h) as we currently only
3523 have a single AES bit in the ciphersuite description bitmap.
3524 That change, however, also applied to ciphersuite strings such as
3525 "RC4-MD5" that intentionally matched multiple ciphersuites --
3526 namely, SSL 2.0 ciphersuites in addition to the more common ones
3527 from SSL 3.0/TLS 1.0.
3529 So we change the selection algorithm again: Naming an explicit
3530 ciphersuite selects this one ciphersuite, and any other similar
3531 ciphersuite (same bitmap) from *other* protocol versions.
3532 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3533 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3535 Since SSL 2.0 does not have any ciphersuites for which the
3536 128/256 bit distinction would be relevant, this works for now.
3537 The proper fix will be to use different bits for AES128 and
3538 AES256, which would have avoided the problems from the beginning;
3539 however, bits are scarce, so we can only do this in a new release
3540 (not just a patchlevel) when we can change the SSL_CIPHER
3541 definition to split the single 'unsigned long mask' bitmap into
3542 multiple values to extend the available space.
3546 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3548 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3549 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3551 *) Add AES IGE and biIGE modes.
3554 *) Change the Unix randomness entropy gathering to use poll() when
3555 possible instead of select(), since the latter has some
3556 undesirable limitations.
3557 [Darryl Miles via Richard Levitte and Bodo Moeller]
3559 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3560 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3561 cannot be implicitly activated as part of, e.g., the "AES" alias.
3562 However, please upgrade to OpenSSL 0.9.9[-dev] for
3563 non-experimental use of the ECC ciphersuites to get TLS extension
3564 support, which is required for curve and point format negotiation
3565 to avoid potential handshake problems.
3568 *) Disable rogue ciphersuites:
3570 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3571 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3572 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3574 The latter two were purportedly from
3575 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3578 Also deactivate the remaining ciphersuites from
3579 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3580 unofficial, and the ID has long expired.
3583 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3584 dual-core machines) and other potential thread-safety issues.
3587 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3588 versions), which is now available for royalty-free use
3589 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3590 Also, add Camellia TLS ciphersuites from RFC 4132.
3592 To minimize changes between patchlevels in the OpenSSL 0.9.8
3593 series, Camellia remains excluded from compilation unless OpenSSL
3594 is configured with 'enable-camellia'.
3597 *) Disable the padding bug check when compression is in use. The padding
3598 bug check assumes the first packet is of even length, this is not
3599 necessarily true if compresssion is enabled and can result in false
3600 positives causing handshake failure. The actual bug test is ancient
3601 code so it is hoped that implementations will either have fixed it by
3602 now or any which still have the bug do not support compression.
3605 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3607 *) When applying a cipher rule check to see if string match is an explicit
3608 cipher suite and only match that one cipher suite if it is.
3611 *) Link in manifests for VC++ if needed.
3612 [Austin Ziegler <halostatue@gmail.com>]
3614 *) Update support for ECC-based TLS ciphersuites according to
3615 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3616 TLS extensions, which are supported starting with the 0.9.9
3617 branch, not in the OpenSSL 0.9.8 branch).
3620 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3621 opaque EVP_CIPHER_CTX handling.
3624 *) Fixes and enhancements to zlib compression code. We now only use
3625 "zlib1.dll" and use the default __cdecl calling convention on Win32
3626 to conform with the standards mentioned here:
3627 http://www.zlib.net/DLL_FAQ.txt
3628 Static zlib linking now works on Windows and the new --with-zlib-include
3629 --with-zlib-lib options to Configure can be used to supply the location
3630 of the headers and library. Gracefully handle case where zlib library
3634 *) Several fixes and enhancements to the OID generation code. The old code
3635 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3636 handle numbers larger than ULONG_MAX, truncated printing and had a
3637 non standard OBJ_obj2txt() behaviour.
3640 *) Add support for building of engines under engine/ as shared libraries
3641 under VC++ build system.
3644 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3645 Hopefully, we will not see any false combination of paths any more.
3648 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3650 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3651 (part of SSL_OP_ALL). This option used to disable the
3652 countermeasure against man-in-the-middle protocol-version
3653 rollback in the SSL 2.0 server implementation, which is a bad
3654 idea. (CVE-2005-2969)
3656 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3657 for Information Security, National Institute of Advanced Industrial
3658 Science and Technology [AIST], Japan)]
3660 *) Add two function to clear and return the verify parameter flags.
3663 *) Keep cipherlists sorted in the source instead of sorting them at
3664 runtime, thus removing the need for a lock.
3667 *) Avoid some small subgroup attacks in Diffie-Hellman.
3668 [Nick Mathewson and Ben Laurie]
3670 *) Add functions for well-known primes.
3673 *) Extended Windows CE support.
3674 [Satoshi Nakamura and Andy Polyakov]
3676 *) Initialize SSL_METHOD structures at compile time instead of during
3677 runtime, thus removing the need for a lock.
3680 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3681 attempting to decrypt each encrypted key in turn. Add support to
3685 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3687 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3690 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3693 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3694 key into the same file any more.
3697 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3700 *) Add -utf8 command line and config file option to 'ca'.
3701 [Stefan <stf@udoma.org]
3703 *) Removed the macro des_crypt(), as it seems to conflict with some
3704 libraries. Use DES_crypt().
3707 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3708 involves renaming the source and generated shared-libs for
3709 both. The engines will accept the corrected or legacy ids
3710 ('ncipher' and '4758_cca' respectively) when binding. NB,
3711 this only applies when building 'shared'.
3712 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3714 *) Add attribute functions to EVP_PKEY structure. Modify
3715 PKCS12_create() to recognize a CSP name attribute and
3716 use it. Make -CSP option work again in pkcs12 utility.
3719 *) Add new functionality to the bn blinding code:
3720 - automatic re-creation of the BN_BLINDING parameters after
3721 a fixed number of uses (currently 32)
3722 - add new function for parameter creation
3723 - introduce flags to control the update behaviour of the
3724 BN_BLINDING parameters
3725 - hide BN_BLINDING structure
3726 Add a second BN_BLINDING slot to the RSA structure to improve
3727 performance when a single RSA object is shared among several
3731 *) Add support for DTLS.
3732 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3734 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3735 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3738 *) Remove buggy and incomplete DH cert support from
3739 ssl/ssl_rsa.c and ssl/s3_both.c
3742 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3743 the apps/openssl applications.
3746 *) Compile clean with "-Wall -Wmissing-prototypes
3747 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3748 DEBUG_SAFESTACK must also be set.
3751 *) Change ./Configure so that certain algorithms can be disabled by default.
3752 The new counterpiece to "no-xxx" is "enable-xxx".
3754 The patented RC5 and MDC2 algorithms will now be disabled unless
3755 "enable-rc5" and "enable-mdc2", respectively, are specified.
3757 (IDEA remains enabled despite being patented. This is because IDEA
3758 is frequently required for interoperability, and there is no license
3759 fee for non-commercial use. As before, "no-idea" can be used to
3760 avoid this algorithm.)
3764 *) Add processing of proxy certificates (see RFC 3820). This work was
3765 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3766 EGEE (Enabling Grids for E-science in Europe).
3769 *) RC4 performance overhaul on modern architectures/implementations, such
3770 as Intel P4, IA-64 and AMD64.
3773 *) New utility extract-section.pl. This can be used specify an alternative
3774 section number in a pod file instead of having to treat each file as
3775 a separate case in Makefile. This can be done by adding two lines to the
3778 =for comment openssl_section:XXX
3780 The blank line is mandatory.
3784 *) New arguments -certform, -keyform and -pass for s_client and s_server
3785 to allow alternative format key and certificate files and passphrase
3789 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3790 update associated structures and add various utility functions.
3792 Add new policy related verify parameters, include policy checking in
3793 standard verify code. Enhance 'smime' application with extra parameters
3794 to support policy checking and print out.
3797 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3798 Nehemiah processors. These extensions support AES encryption in hardware
3799 as well as RNG (though RNG support is currently disabled).
3800 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3802 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3805 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3806 [Andy Polyakov and a number of other people]
3808 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3809 implementation contributed by IBM.
3810 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3812 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3813 exponent rather than 'unsigned long'. There is a corresponding change to
3814 the new 'rsa_keygen' element of the RSA_METHOD structure.
3815 [Jelte Jansen, Geoff Thorpe]
3817 *) Functionality for creating the initial serial number file is now
3818 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3820 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3821 number file to 1, which is bound to cause problems. To avoid
3822 the problems while respecting compatibility between different 0.9.7
3823 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3824 CA.pl for serial number initialization. With the new release 0.9.8,
3825 we can fix the problem directly in the 'ca' utility.)
3828 *) Reduced header interdepencies by declaring more opaque objects in
3829 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3830 give fewer recursive includes, which could break lazy source code - so
3831 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3832 developers should define this symbol when building and using openssl to
3833 ensure they track the recommended behaviour, interfaces, [etc], but
3834 backwards-compatible behaviour prevails when this isn't defined.
3837 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3840 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3841 This will generate a random key of the appropriate length based on the
3842 cipher context. The EVP_CIPHER can provide its own random key generation
3843 routine to support keys of a specific form. This is used in the des and
3844 3des routines to generate a key of the correct parity. Update S/MIME
3845 code to use new functions and hence generate correct parity DES keys.
3846 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3847 valid (weak or incorrect parity).
3850 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3851 as looking them up. This is useful when the verified structure may contain
3852 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3853 present unless the new PKCS7_NO_CRL flag is asserted.
3856 *) Extend ASN1 oid configuration module. It now additionally accepts the
3859 shortName = some long name, 1.2.3.4
3862 *) Reimplemented the BN_CTX implementation. There is now no more static
3863 limitation on the number of variables it can handle nor the depth of the
3864 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3865 information can now expand as required, and rather than having a single
3866 static array of bignums, BN_CTX now uses a linked-list of such arrays
3867 allowing it to expand on demand whilst maintaining the usefulness of
3868 BN_CTX's "bundling".
3871 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3872 to allow all RSA operations to function using a single BN_CTX.
3875 *) Preliminary support for certificate policy evaluation and checking. This
3876 is initially intended to pass the tests outlined in "Conformance Testing
3877 of Relying Party Client Certificate Path Processing Logic" v1.07.
3880 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3881 remained unused and not that useful. A variety of other little bignum
3882 tweaks and fixes have also been made continuing on from the audit (see
3886 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3887 associated ASN1, EVP and SSL functions and old ASN1 macros.
3890 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3891 and this should never fail. So the return value from the use of
3892 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3893 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3896 *) BN_CTX_get() should return zero-valued bignums, providing the same
3897 initialised value as BN_new().
3898 [Geoff Thorpe, suggested by Ulf Möller]
3900 *) Support for inhibitAnyPolicy certificate extension.
3903 *) An audit of the BIGNUM code is underway, for which debugging code is
3904 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3905 is considered valid when processing BIGNUMs, and causes execution to
3906 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3907 further steps are taken to deliberately pollute unused data in BIGNUM
3908 structures to try and expose faulty code further on. For now, openssl will
3909 (in its default mode of operation) continue to tolerate the inconsistent
3910 forms that it has tolerated in the past, but authors and packagers should
3911 consider trying openssl and their own applications when compiled with
3912 these debugging symbols defined. It will help highlight potential bugs in
3913 their own code, and will improve the test coverage for OpenSSL itself. At
3914 some point, these tighter rules will become openssl's default to improve
3915 maintainability, though the assert()s and other overheads will remain only
3916 in debugging configurations. See bn.h for more details.
3917 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3919 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3920 that can only be obtained through BN_CTX_new() (which implicitly
3921 initialises it). The presence of this function only made it possible
3922 to overwrite an existing structure (and cause memory leaks).
3925 *) Because of the callback-based approach for implementing LHASH as a
3926 template type, lh_insert() adds opaque objects to hash-tables and
3927 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3928 to clean up those corresponding objects before destroying the hash table
3929 (and losing the object pointers). So some over-zealous constifications in
3930 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3931 objects as "const" and the lh_doall[_arg] callback wrappers are not
3932 prototyped to have "const" restrictions on the object pointers they are
3933 given (and so aren't required to cast them away any more).
3936 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3937 (speed) prefers to use its own implementation. The two implementations
3938 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3939 its object type properly exposed (MS_TM) instead of casting to/from "char
3940 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3941 aren't necessarily the greatest nomenclatures - but this is what was used
3942 internally to the implementation so I've used that for now.
3945 *) Ensure that deprecated functions do not get compiled when
3946 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3947 the self-tests were still using deprecated key-generation functions so
3948 these have been updated also.
3951 *) Reorganise PKCS#7 code to separate the digest location functionality
3952 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3953 New function PKCS7_set_digest() to set the digest type for PKCS#7
3954 digestedData type. Add additional code to correctly generate the
3955 digestedData type and add support for this type in PKCS7 initialization
3959 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3960 structure of type "other".
3963 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3964 sure the loop does correctly stop and breaking ("division by zero")
3965 modulus operations are not performed. The (pre-generated) prime
3966 table crypto/bn/bn_prime.h was already correct, but it could not be
3967 re-generated on some platforms because of the "division by zero"
3968 situation in the script.
3969 [Ralf S. Engelschall]
3971 *) Update support for ECC-based TLS ciphersuites according to
3972 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3973 SHA-1 now is only used for "small" curves (where the
3974 representation of a field element takes up to 24 bytes); for
3975 larger curves, the field element resulting from ECDH is directly
3976 used as premaster secret.
3977 [Douglas Stebila (Sun Microsystems Laboratories)]
3979 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3980 curve secp160r1 to the tests.
3981 [Douglas Stebila (Sun Microsystems Laboratories)]
3983 *) Add the possibility to load symbols globally with DSO.
3984 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3986 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3987 control of the error stack.
3990 *) Add support for STORE in ENGINE.
3993 *) Add the STORE type. The intention is to provide a common interface
3994 to certificate and key stores, be they simple file-based stores, or
3995 HSM-type store, or LDAP stores, or...
3996 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3999 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4000 pass a list of arguments to any function as well as provide a way
4001 for a function to pass data back to the caller.
4004 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4005 works like BUF_strdup() but can be used to duplicate a portion of
4006 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4010 *) Add the function sk_find_ex() which works like sk_find(), but will
4011 return an index to an element even if an exact match couldn't be
4012 found. The index is guaranteed to point at the element where the
4013 searched-for key would be inserted to preserve sorting order.
4016 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4017 takes an extra flags argument for optional functionality. Currently,
4018 the following flags are defined:
4020 OBJ_BSEARCH_VALUE_ON_NOMATCH
4021 This one gets OBJ_bsearch_ex() to return a pointer to the first
4022 element where the comparing function returns a negative or zero
4025 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4026 This one gets OBJ_bsearch_ex() to return a pointer to the first
4027 element where the comparing function returns zero. This is useful
4028 if there are more than one element where the comparing function
4032 *) Make it possible to create self-signed certificates with 'openssl ca'
4033 in such a way that the self-signed certificate becomes part of the
4034 CA database and uses the same mechanisms for serial number generation
4035 as all other certificate signing. The new flag '-selfsign' enables
4036 this functionality. Adapt CA.sh and CA.pl.in.
4039 *) Add functionality to check the public key of a certificate request
4040 against a given private. This is useful to check that a certificate
4041 request can be signed by that key (self-signing).
4044 *) Make it possible to have multiple active certificates with the same
4045 subject in the CA index file. This is done only if the keyword
4046 'unique_subject' is set to 'no' in the main CA section (default
4047 if 'CA_default') of the configuration file. The value is saved
4048 with the database itself in a separate index attribute file,
4049 named like the index file with '.attr' appended to the name.
4052 *) Generate muti valued AVAs using '+' notation in config files for
4056 *) Support for nameConstraints certificate extension.
4059 *) Support for policyConstraints certificate extension.
4062 *) Support for policyMappings certificate extension.
4065 *) Make sure the default DSA_METHOD implementation only uses its
4066 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4067 and change its own handlers to be NULL so as to remove unnecessary
4068 indirection. This lets alternative implementations fallback to the
4069 default implementation more easily.
4072 *) Support for directoryName in GeneralName related extensions
4076 *) Make it possible to link applications using Makefile.shared.
4077 Make that possible even when linking against static libraries!
4080 *) Support for single pass processing for S/MIME signing. This now
4081 means that S/MIME signing can be done from a pipe, in addition
4082 cleartext signing (multipart/signed type) is effectively streaming
4083 and the signed data does not need to be all held in memory.
4085 This is done with a new flag PKCS7_STREAM. When this flag is set
4086 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4087 is done after the data is output (and digests calculated) in
4088 SMIME_write_PKCS7().
4091 *) Add full support for -rpath/-R, both in shared libraries and
4092 applications, at least on the platforms where it's known how
4096 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4097 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4098 will now compute a table of multiples of the generator that
4099 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4100 faster (notably in the case of a single point multiplication,
4101 scalar * generator).
4102 [Nils Larsch, Bodo Moeller]
4104 *) IPv6 support for certificate extensions. The various extensions
4105 which use the IP:a.b.c.d can now take IPv6 addresses using the
4106 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4110 *) Added an ENGINE that implements RSA by performing private key
4111 exponentiations with the GMP library. The conversions to and from
4112 GMP's mpz_t format aren't optimised nor are any montgomery forms
4113 cached, and on x86 it appears OpenSSL's own performance has caught up.
4114 However there are likely to be other architectures where GMP could
4115 provide a boost. This ENGINE is not built in by default, but it can be
4116 specified at Configure time and should be accompanied by the necessary
4117 linker additions, eg;
4118 ./config -DOPENSSL_USE_GMP -lgmp
4121 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4122 testing availability of engines with "-t" - the old behaviour is
4123 produced by increasing the feature's verbosity with "-tt".
4126 *) ECDSA routines: under certain error conditions uninitialized BN objects
4127 could be freed. Solution: make sure initialization is performed early
4128 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4132 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4133 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4134 software implementations. For DSA and DH, parameter generation can
4135 also be overriden by providing the appropriate method callbacks.
4138 *) Change the "progress" mechanism used in key-generation and
4139 primality testing to functions that take a new BN_GENCB pointer in
4140 place of callback/argument pairs. The new API functions have "_ex"
4141 postfixes and the older functions are reimplemented as wrappers for
4142 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4143 declarations of the old functions to help (graceful) attempts to
4144 migrate to the new functions. Also, the new key-generation API
4145 functions operate on a caller-supplied key-structure and return
4146 success/failure rather than returning a key or NULL - this is to
4147 help make "keygen" another member function of RSA_METHOD etc.
4149 Example for using the new callback interface:
4151 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4155 BN_GENCB_set(&my_cb, my_callback, my_arg);
4157 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4158 /* For the meaning of a, b in calls to my_callback(), see the
4159 * documentation of the function that calls the callback.
4160 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4161 * my_callback should return 1 if it wants BN_is_prime_ex()
4162 * to continue, or 0 to stop.
4167 *) Change the ZLIB compression method to be stateful, and make it
4168 available to TLS with the number defined in
4169 draft-ietf-tls-compression-04.txt.
4172 *) Add the ASN.1 structures and functions for CertificatePair, which
4173 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4175 CertificatePair ::= SEQUENCE {
4176 forward [0] Certificate OPTIONAL,
4177 reverse [1] Certificate OPTIONAL,
4178 -- at least one of the pair shall be present -- }
4180 Also implement the PEM functions to read and write certificate
4181 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4183 This needed to be defined, mostly for the sake of the LDAP
4184 attribute crossCertificatePair, but may prove useful elsewhere as
4188 *) Make it possible to inhibit symlinking of shared libraries in
4189 Makefile.shared, for Cygwin's sake.
4192 *) Extend the BIGNUM API by creating a function
4193 void BN_set_negative(BIGNUM *a, int neg);
4194 and a macro that behave like
4195 int BN_is_negative(const BIGNUM *a);
4197 to avoid the need to access 'a->neg' directly in applications.
4200 *) Implement fast modular reduction for pseudo-Mersenne primes
4201 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4202 EC_GROUP_new_curve_GFp() will now automatically use this
4204 [Nils Larsch <nla@trustcenter.de>]
4206 *) Add new lock type (CRYPTO_LOCK_BN).
4209 *) Change the ENGINE framework to automatically load engines
4210 dynamically from specific directories unless they could be
4211 found to already be built in or loaded. Move all the
4212 current engines except for the cryptodev one to a new
4214 The engines in engines/ are built as shared libraries if
4215 the "shared" options was given to ./Configure or ./config.
4216 Otherwise, they are inserted in libcrypto.a.
4217 /usr/local/ssl/engines is the default directory for dynamic
4218 engines, but that can be overriden at configure time through
4219 the usual use of --prefix and/or --openssldir, and at run
4220 time with the environment variable OPENSSL_ENGINES.
4221 [Geoff Thorpe and Richard Levitte]
4223 *) Add Makefile.shared, a helper makefile to build shared
4224 libraries. Adapt Makefile.org.
4227 *) Add version info to Win32 DLLs.
4228 [Peter 'Luna' Runestig" <peter@runestig.com>]
4230 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4231 can be added using this API to created arbitrary PKCS#12
4232 files while avoiding the low level API.
4234 New options to PKCS12_create(), key or cert can be NULL and
4235 will then be omitted from the output file. The encryption
4236 algorithm NIDs can be set to -1 for no encryption, the mac
4237 iteration count can be set to 0 to omit the mac.
4239 Enhance pkcs12 utility by making the -nokeys and -nocerts
4240 options work when creating a PKCS#12 file. New option -nomac
4241 to omit the mac, NONE can be set for an encryption algorithm.
4242 New code is modified to use the enhanced PKCS12_create()
4243 instead of the low level API.
4246 *) Extend ASN1 encoder to support indefinite length constructed
4247 encoding. This can output sequences tags and octet strings in
4248 this form. Modify pk7_asn1.c to support indefinite length
4249 encoding. This is experimental and needs additional code to
4250 be useful, such as an ASN1 bio and some enhanced streaming
4253 Extend template encode functionality so that tagging is passed
4254 down to the template encoder.
4257 *) Let 'openssl req' fail if an argument to '-newkey' is not
4258 recognized instead of using RSA as a default.
4261 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4262 As these are not official, they are not included in "ALL";
4263 the "ECCdraft" ciphersuite group alias can be used to select them.
4264 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4266 *) Add ECDH engine support.
4267 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4269 *) Add ECDH in new directory crypto/ecdh/.
4270 [Douglas Stebila (Sun Microsystems Laboratories)]
4272 *) Let BN_rand_range() abort with an error after 100 iterations
4273 without success (which indicates a broken PRNG).
4276 *) Change BN_mod_sqrt() so that it verifies that the input value
4277 is really the square of the return value. (Previously,
4278 BN_mod_sqrt would show GIGO behaviour.)
4281 *) Add named elliptic curves over binary fields from X9.62, SECG,
4282 and WAP/WTLS; add OIDs that were still missing.
4284 [Sheueling Chang Shantz and Douglas Stebila
4285 (Sun Microsystems Laboratories)]
4287 *) Extend the EC library for elliptic curves over binary fields
4288 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4291 EC_GF2m_simple_method
4295 EC_GROUP_new_curve_GF2m
4296 EC_GROUP_set_curve_GF2m
4297 EC_GROUP_get_curve_GF2m
4298 EC_POINT_set_affine_coordinates_GF2m
4299 EC_POINT_get_affine_coordinates_GF2m
4300 EC_POINT_set_compressed_coordinates_GF2m
4302 Point compression for binary fields is disabled by default for
4303 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4306 As binary polynomials are represented as BIGNUMs, various members
4307 of the EC_GROUP and EC_POINT data structures can be shared
4308 between the implementations for prime fields and binary fields;
4309 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4310 are essentially identical to their ..._GFp counterparts.
4311 (For simplicity, the '..._GFp' prefix has been dropped from
4312 various internal method names.)
4314 An internal 'field_div' method (similar to 'field_mul' and
4315 'field_sqr') has been added; this is used only for binary fields.
4317 [Sheueling Chang Shantz and Douglas Stebila
4318 (Sun Microsystems Laboratories)]
4320 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4321 through methods ('mul', 'precompute_mult').
4323 The generic implementations (now internally called 'ec_wNAF_mul'
4324 and 'ec_wNAF_precomputed_mult') remain the default if these
4325 methods are undefined.
4327 [Sheueling Chang Shantz and Douglas Stebila
4328 (Sun Microsystems Laboratories)]
4330 *) New function EC_GROUP_get_degree, which is defined through
4331 EC_METHOD. For curves over prime fields, this returns the bit
4332 length of the modulus.
4334 [Sheueling Chang Shantz and Douglas Stebila
4335 (Sun Microsystems Laboratories)]
4337 *) New functions EC_GROUP_dup, EC_POINT_dup.
4338 (These simply call ..._new and ..._copy).
4340 [Sheueling Chang Shantz and Douglas Stebila
4341 (Sun Microsystems Laboratories)]
4343 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4344 Polynomials are represented as BIGNUMs (where the sign bit is not
4345 used) in the following functions [macros]:
4348 BN_GF2m_sub [= BN_GF2m_add]
4349 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4350 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4351 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4353 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4354 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4355 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4356 BN_GF2m_cmp [= BN_ucmp]
4358 (Note that only the 'mod' functions are actually for fields GF(2^m).
4359 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4361 For some functions, an the irreducible polynomial defining a
4362 field can be given as an 'unsigned int[]' with strictly
4363 decreasing elements giving the indices of those bits that are set;
4364 i.e., p[] represents the polynomial
4365 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4367 p[0] > p[1] > ... > p[k] = 0.
4368 This applies to the following functions:
4373 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4374 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4376 BN_GF2m_mod_sqrt_arr
4377 BN_GF2m_mod_solve_quad_arr
4381 Conversion can be performed by the following functions:
4386 bntest.c has additional tests for binary polynomial arithmetic.
4388 Two implementations for BN_GF2m_mod_div() are available.
4389 The default algorithm simply uses BN_GF2m_mod_inv() and
4390 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4391 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4392 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4394 [Sheueling Chang Shantz and Douglas Stebila
4395 (Sun Microsystems Laboratories)]
4397 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4398 functionality is disabled at compile-time.
4399 [Douglas Stebila <douglas.stebila@sun.com>]
4401 *) Change default behaviour of 'openssl asn1parse' so that more
4402 information is visible when viewing, e.g., a certificate:
4404 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4405 mode the content of non-printable OCTET STRINGs is output in a
4406 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4407 avoid the appearance of a printable string.
4408 [Nils Larsch <nla@trustcenter.de>]
4410 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4412 EC_GROUP_set_asn1_flag()
4413 EC_GROUP_get_asn1_flag()
4414 EC_GROUP_set_point_conversion_form()
4415 EC_GROUP_get_point_conversion_form()
4416 These control ASN1 encoding details:
4417 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4418 has been set to OPENSSL_EC_NAMED_CURVE.
4419 - Points are encoded in uncompressed form by default; options for
4420 asn1_for are as for point2oct, namely
4421 POINT_CONVERSION_COMPRESSED
4422 POINT_CONVERSION_UNCOMPRESSED
4423 POINT_CONVERSION_HYBRID
4425 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4428 EC_GROUP_get0_seed()
4429 EC_GROUP_get_seed_len()
4430 This is used only for ASN1 purposes (so far).
4431 [Nils Larsch <nla@trustcenter.de>]
4433 *) Add 'field_type' member to EC_METHOD, which holds the NID
4434 of the appropriate field type OID. The new function
4435 EC_METHOD_get_field_type() returns this value.
4436 [Nils Larsch <nla@trustcenter.de>]
4441 EC_POINT_point2hex()
4442 EC_POINT_hex2point()
4443 providing useful interfaces to EC_POINT_point2oct() and
4444 EC_POINT_oct2point().
4445 [Nils Larsch <nla@trustcenter.de>]
4447 *) Change internals of the EC library so that the functions
4448 EC_GROUP_set_generator()
4449 EC_GROUP_get_generator()
4450 EC_GROUP_get_order()
4451 EC_GROUP_get_cofactor()
4452 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4453 to methods, which would lead to unnecessary code duplication when
4454 adding different types of curves.
4455 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4457 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4458 arithmetic, and such that modified wNAFs are generated
4459 (which avoid length expansion in many cases).
4462 *) Add a function EC_GROUP_check_discriminant() (defined via
4463 EC_METHOD) that verifies that the curve discriminant is non-zero.
4465 Add a function EC_GROUP_check() that makes some sanity tests
4466 on a EC_GROUP, its generator and order. This includes
4467 EC_GROUP_check_discriminant().
4468 [Nils Larsch <nla@trustcenter.de>]
4470 *) Add ECDSA in new directory crypto/ecdsa/.
4472 Add applications 'openssl ecparam' and 'openssl ecdsa'
4473 (these are based on 'openssl dsaparam' and 'openssl dsa').
4475 ECDSA support is also included in various other files across the
4476 library. Most notably,
4477 - 'openssl req' now has a '-newkey ecdsa:file' option;
4478 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4479 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4480 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4481 them suitable for ECDSA where domain parameters must be
4482 extracted before the specific public key;
4483 - ECDSA engine support has been added.
4484 [Nils Larsch <nla@trustcenter.de>]
4486 *) Include some named elliptic curves, and add OIDs from X9.62,
4487 SECG, and WAP/WTLS. Each curve can be obtained from the new
4489 EC_GROUP_new_by_curve_name(),
4490 and the list of available named curves can be obtained with
4491 EC_get_builtin_curves().
4492 Also add a 'curve_name' member to EC_GROUP objects, which can be
4494 EC_GROUP_set_curve_name()
4495 EC_GROUP_get_curve_name()
4496 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4498 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4499 was actually never needed) and in BN_mul(). The removal in BN_mul()
4500 required a small change in bn_mul_part_recursive() and the addition
4501 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4502 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4503 bn_sub_words() and bn_add_words() except they take arrays with
4507 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4509 *) Cleanse PEM buffers before freeing them since they may contain
4511 [Benjamin Bennett <ben@psc.edu>]
4513 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4514 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4515 authentication-only ciphersuites.
4518 *) Since AES128 and AES256 share a single mask bit in the logic of
4519 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4520 kludge to work properly if AES128 is available and AES256 isn't.
4523 *) Expand security boundary to match 1.1.1 module.
4526 *) Remove redundant features: hash file source, editing of test vectors
4527 modify fipsld to use external fips_premain.c signature.
4530 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4531 run algorithm test programs.
4534 *) Make algorithm test programs more tolerant of whitespace.
4537 *) Have SSL/TLS server implementation tolerate "mismatched" record
4538 protocol version while receiving ClientHello even if the
4539 ClientHello is fragmented. (The server can't insist on the
4540 particular protocol version it has chosen before the ServerHello
4541 message has informed the client about his choice.)
4544 *) Load error codes if they are not already present instead of using a
4545 static variable. This allows them to be cleanly unloaded and reloaded.
4548 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4550 *) Introduce limits to prevent malicious keys being able to
4551 cause a denial of service. (CVE-2006-2940)
4552 [Steve Henson, Bodo Moeller]
4554 *) Fix ASN.1 parsing of certain invalid structures that can result
4555 in a denial of service. (CVE-2006-2937) [Steve Henson]
4557 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4558 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4560 *) Fix SSL client code which could crash if connecting to a
4561 malicious SSLv2 server. (CVE-2006-4343)
4562 [Tavis Ormandy and Will Drewry, Google Security Team]
4564 *) Change ciphersuite string processing so that an explicit
4565 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4566 will no longer include "AES128-SHA"), and any other similar
4567 ciphersuite (same bitmap) from *other* protocol versions (so that
4568 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4569 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4570 changes from 0.9.8b and 0.9.8d.
4573 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4575 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4576 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4578 *) Change the Unix randomness entropy gathering to use poll() when
4579 possible instead of select(), since the latter has some
4580 undesirable limitations.
4581 [Darryl Miles via Richard Levitte and Bodo Moeller]
4583 *) Disable rogue ciphersuites:
4585 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4586 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4587 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4589 The latter two were purportedly from
4590 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4593 Also deactive the remaining ciphersuites from
4594 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4595 unofficial, and the ID has long expired.
4598 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4599 dual-core machines) and other potential thread-safety issues.
4602 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4604 *) Adapt fipsld and the build system to link against the validated FIPS
4605 module in FIPS mode.
4608 *) Fixes for VC++ 2005 build under Windows.
4611 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4612 from a Windows bash shell such as MSYS. It is autodetected from the
4613 "config" script when run from a VC++ environment. Modify standard VC++
4614 build to use fipscanister.o from the GNU make build.
4617 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4619 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4620 The value now differs depending on if you build for FIPS or not.
4621 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4622 safely run with a non-FIPSed libcrypto, as it may crash because of
4623 the difference induced by this change.
4626 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4628 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4629 (part of SSL_OP_ALL). This option used to disable the
4630 countermeasure against man-in-the-middle protocol-version
4631 rollback in the SSL 2.0 server implementation, which is a bad
4632 idea. (CVE-2005-2969)
4634 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4635 for Information Security, National Institute of Advanced Industrial
4636 Science and Technology [AIST], Japan)]
4638 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4639 mainly for FIPS compliance and not fully integrated at this stage.
4642 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4643 the exponentiation using a fixed-length exponent. (Otherwise,
4644 the information leaked through timing could expose the secret key
4645 after many signatures; cf. Bleichenbacher's attack on DSA with
4649 *) Make a new fixed-window mod_exp implementation the default for
4650 RSA, DSA, and DH private-key operations so that the sequence of
4651 squares and multiplies and the memory access pattern are
4652 independent of the particular secret key. This will mitigate
4653 cache-timing and potential related attacks.
4655 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4656 and this is automatically used by BN_mod_exp_mont() if the new flag
4657 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4658 will use this BN flag for private exponents unless the flag
4659 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4660 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4662 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4664 *) Change the client implementation for SSLv23_method() and
4665 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4666 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4667 (Previously, the SSL 2.0 backwards compatible Client Hello
4668 message format would be used even with SSL_OP_NO_SSLv2.)
4671 *) Add support for smime-type MIME parameter in S/MIME messages which some
4675 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4676 a threadsafe manner. Modify rsa code to use new function and add calls
4677 to dsa and dh code (which had race conditions before).
4680 *) Include the fixed error library code in the C error file definitions
4681 instead of fixing them up at runtime. This keeps the error code
4682 structures constant.
4685 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4687 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4690 *) Fixes for newer kerberos headers. NB: the casts are needed because
4691 the 'length' field is signed on one version and unsigned on another
4692 with no (?) obvious way to tell the difference, without these VC++
4693 complains. Also the "definition" of FAR (blank) is no longer included
4694 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4695 some needed definitions.
4698 *) Undo Cygwin change.
4701 *) Added support for proxy certificates according to RFC 3820.
4702 Because they may be a security thread to unaware applications,
4703 they must be explicitely allowed in run-time. See
4704 docs/HOWTO/proxy_certificates.txt for further information.
4707 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4709 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4710 server and client random values. Previously
4711 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4712 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4714 This change has negligible security impact because:
4716 1. Server and client random values still have 24 bytes of pseudo random
4719 2. Server and client random values are sent in the clear in the initial
4722 3. The master secret is derived using the premaster secret (48 bytes in
4723 size for static RSA ciphersuites) as well as client server and random
4726 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4729 [Stephen Henson, reported by UK NISCC]
4731 *) Use Windows randomness collection on Cygwin.
4734 *) Fix hang in EGD/PRNGD query when communication socket is closed
4735 prematurely by EGD/PRNGD.
4736 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4738 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4741 *) Back-port of selected performance improvements from development
4742 branch, as well as improved support for PowerPC platforms.
4745 *) Add lots of checks for memory allocation failure, error codes to indicate
4746 failure and freeing up memory if a failure occurs.
4747 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4749 *) Add new -passin argument to dgst.
4752 *) Perform some character comparisons of different types in X509_NAME_cmp:
4753 this is needed for some certificates that reencode DNs into UTF8Strings
4754 (in violation of RFC3280) and can't or wont issue name rollover
4758 *) Make an explicit check during certificate validation to see that
4759 the CA setting in each certificate on the chain is correct. As a
4760 side effect always do the following basic checks on extensions,
4761 not just when there's an associated purpose to the check:
4763 - if there is an unhandled critical extension (unless the user
4764 has chosen to ignore this fault)
4765 - if the path length has been exceeded (if one is set at all)
4766 - that certain extensions fit the associated purpose (if one has
4770 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4772 *) Avoid a race condition when CRLs are checked in a multi threaded
4773 environment. This would happen due to the reordering of the revoked
4774 entries during signature checking and serial number lookup. Now the
4775 encoding is cached and the serial number sort performed under a lock.
4776 Add new STACK function sk_is_sorted().
4779 *) Add Delta CRL to the extension code.
4782 *) Various fixes to s3_pkt.c so alerts are sent properly.
4783 [David Holmes <d.holmes@f5.com>]
4785 *) Reduce the chances of duplicate issuer name and serial numbers (in
4786 violation of RFC3280) using the OpenSSL certificate creation utilities.
4787 This is done by creating a random 64 bit value for the initial serial
4788 number when a serial number file is created or when a self signed
4789 certificate is created using 'openssl req -x509'. The initial serial
4790 number file is created using 'openssl x509 -next_serial' in CA.pl
4791 rather than being initialized to 1.
4794 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4796 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4797 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4798 [Joe Orton, Steve Henson]
4800 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4802 [Joe Orton, Steve Henson]
4804 *) Make it possible to have multiple active certificates with the same
4805 subject in the CA index file. This is done only if the keyword
4806 'unique_subject' is set to 'no' in the main CA section (default
4807 if 'CA_default') of the configuration file. The value is saved
4808 with the database itself in a separate index attribute file,
4809 named like the index file with '.attr' appended to the name.
4812 *) X509 verify fixes. Disable broken certificate workarounds when
4813 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4814 keyUsage extension present. Don't accept CRLs with unhandled critical
4815 extensions: since verify currently doesn't process CRL extensions this
4816 rejects a CRL with *any* critical extensions. Add new verify error codes
4820 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4821 A clarification of RFC2560 will require the use of OCTET STRINGs and
4822 some implementations cannot handle the current raw format. Since OpenSSL
4823 copies and compares OCSP nonces as opaque blobs without any attempt at
4824 parsing them this should not create any compatibility issues.
4827 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4828 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4829 this HMAC (and other) operations are several times slower than OpenSSL
4833 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4834 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4836 *) Use the correct content when signing type "other".
4839 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4841 *) Fix various bugs revealed by running the NISCC test suite:
4843 Stop out of bounds reads in the ASN1 code when presented with
4844 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4846 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4848 If verify callback ignores invalid public key errors don't try to check
4849 certificate signature with the NULL public key.
4853 *) New -ignore_err option in ocsp application to stop the server
4854 exiting on the first error in a request.
4857 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4858 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4862 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4863 extra data after the compression methods not only for TLS 1.0
4864 but also for SSL 3.0 (as required by the specification).
4865 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4867 *) Change X509_certificate_type() to mark the key as exported/exportable
4868 when it's 512 *bits* long, not 512 bytes.
4871 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4872 blocks during encryption.
4875 *) Various fixes to base64 BIO and non blocking I/O. On write
4876 flushes were not handled properly if the BIO retried. On read
4877 data was not being buffered properly and had various logic bugs.
4878 This also affects blocking I/O when the data being decoded is a
4882 *) Various S/MIME bugfixes and compatibility changes:
4883 output correct application/pkcs7 MIME type if
4884 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4885 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4886 of files as .eml work). Correctly handle very long lines in MIME
4890 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4892 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4893 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4894 a protocol version number mismatch like a decryption error
4895 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4898 *) Turn on RSA blinding by default in the default implementation
4899 to avoid a timing attack. Applications that don't want it can call
4900 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4901 They would be ill-advised to do so in most cases.
4902 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4904 *) Change RSA blinding code so that it works when the PRNG is not
4905 seeded (in this case, the secret RSA exponent is abused as
4906 an unpredictable seed -- if it is not unpredictable, there
4907 is no point in blinding anyway). Make RSA blinding thread-safe
4908 by remembering the creator's thread ID in rsa->blinding and
4909 having all other threads use local one-time blinding factors
4910 (this requires more computation than sharing rsa->blinding, but
4911 avoids excessive locking; and if an RSA object is not shared
4912 between threads, blinding will still be very fast).
4915 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4916 ENGINE as defaults for all supported algorithms irrespective of
4917 the 'flags' parameter. 'flags' is now honoured, so applications
4918 should make sure they are passing it correctly.
4921 *) Target "mingw" now allows native Windows code to be generated in
4922 the Cygwin environment as well as with the MinGW compiler.
4925 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4927 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4928 via timing by performing a MAC computation even if incorrect
4929 block cipher padding has been found. This is a countermeasure
4930 against active attacks where the attacker has to distinguish
4931 between bad padding and a MAC verification error. (CVE-2003-0078)
4933 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4934 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4935 Martin Vuagnoux (EPFL, Ilion)]
4937 *) Make the no-err option work as intended. The intention with no-err
4938 is not to have the whole error stack handling routines removed from
4939 libcrypto, it's only intended to remove all the function name and
4940 reason texts, thereby removing some of the footprint that may not
4941 be interesting if those errors aren't displayed anyway.
4943 NOTE: it's still possible for any application or module to have it's
4944 own set of error texts inserted. The routines are there, just not
4945 used by default when no-err is given.
4948 *) Add support for FreeBSD on IA64.
4949 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4951 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4952 Kerberos function mit_des_cbc_cksum(). Before this change,
4953 the value returned by DES_cbc_cksum() was like the one from
4954 mit_des_cbc_cksum(), except the bytes were swapped.
4955 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4957 *) Allow an application to disable the automatic SSL chain building.
4958 Before this a rather primitive chain build was always performed in
4959 ssl3_output_cert_chain(): an application had no way to send the
4960 correct chain if the automatic operation produced an incorrect result.
4962 Now the chain builder is disabled if either:
4964 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4966 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4968 The reasoning behind this is that an application would not want the
4969 auto chain building to take place if extra chain certificates are
4970 present and it might also want a means of sending no additional
4971 certificates (for example the chain has two certificates and the
4975 *) Add the possibility to build without the ENGINE framework.
4976 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4978 *) Under Win32 gmtime() can return NULL: check return value in
4979 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4982 *) DSA routines: under certain error conditions uninitialized BN objects
4983 could be freed. Solution: make sure initialization is performed early
4984 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4985 Nils Larsch <nla@trustcenter.de> via PR#459)
4988 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4989 checked on reconnect on the client side, therefore session resumption
4990 could still fail with a "ssl session id is different" error. This
4991 behaviour is masked when SSL_OP_ALL is used due to
4992 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4993 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4994 followup to PR #377.
4997 *) IA-32 assembler support enhancements: unified ELF targets, support
4998 for SCO/Caldera platforms, fix for Cygwin shared build.
5001 *) Add support for FreeBSD on sparc64. As a consequence, support for
5002 FreeBSD on non-x86 processors is separate from x86 processors on
5003 the config script, much like the NetBSD support.
5004 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5006 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5008 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5011 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5012 code (06) was taken as the first octet of the session ID and the last
5013 octet was ignored consequently. As a result SSLv2 client side session
5014 caching could not have worked due to the session ID mismatch between
5016 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5020 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5021 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5025 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5026 seems that in spite of existing for more than a year, many application
5027 author have done nothing to provide the necessary callbacks, which
5028 means that this particular engine will not work properly anywhere.
5029 This is a very unfortunate situation which forces us, in the name
5030 of usability, to give the hw_ncipher.c a static lock, which is part
5032 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5033 appear in 0.9.8 or later. We EXPECT application authors to have
5034 dealt properly with this when 0.9.8 is released (unless we actually
5035 make such changes in the libcrypto locking code that changes will
5036 have to be made anyway).
5039 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5040 octets have been read, EOF or an error occurs. Without this change
5041 some truncated ASN1 structures will not produce an error.
5044 *) Disable Heimdal support, since it hasn't been fully implemented.
5045 Still give the possibility to force the use of Heimdal, but with
5046 warnings and a request that patches get sent to openssl-dev.
5049 *) Add the VC-CE target, introduce the WINCE sysname, and add
5050 INSTALL.WCE and appropriate conditionals to make it build.
5051 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5053 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5054 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5055 edit numbers of the version.
5056 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5058 *) Introduce safe string copy and catenation functions
5059 (BUF_strlcpy() and BUF_strlcat()).
5060 [Ben Laurie (CHATS) and Richard Levitte]
5062 *) Avoid using fixed-size buffers for one-line DNs.
5063 [Ben Laurie (CHATS)]
5065 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5066 resizing buffers containing secrets, and use where appropriate.
5067 [Ben Laurie (CHATS)]
5069 *) Avoid using fixed size buffers for configuration file location.
5070 [Ben Laurie (CHATS)]
5072 *) Avoid filename truncation for various CA files.
5073 [Ben Laurie (CHATS)]
5075 *) Use sizeof in preference to magic numbers.
5076 [Ben Laurie (CHATS)]
5078 *) Avoid filename truncation in cert requests.
5079 [Ben Laurie (CHATS)]
5081 *) Add assertions to check for (supposedly impossible) buffer
5083 [Ben Laurie (CHATS)]
5085 *) Don't cache truncated DNS entries in the local cache (this could
5086 potentially lead to a spoofing attack).
5087 [Ben Laurie (CHATS)]
5089 *) Fix various buffers to be large enough for hex/decimal
5090 representations in a platform independent manner.
5091 [Ben Laurie (CHATS)]
5093 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5094 resizing buffers containing secrets, and use where appropriate.
5095 [Ben Laurie (CHATS)]
5097 *) Add BIO_indent() to avoid much slightly worrying code to do
5099 [Ben Laurie (CHATS)]
5101 *) Convert sprintf()/BIO_puts() to BIO_printf().
5102 [Ben Laurie (CHATS)]
5104 *) buffer_gets() could terminate with the buffer only half
5106 [Ben Laurie (CHATS)]
5108 *) Add assertions to prevent user-supplied crypto functions from
5109 overflowing internal buffers by having large block sizes, etc.
5110 [Ben Laurie (CHATS)]
5112 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5114 [Ben Laurie (CHATS)]
5116 *) Eliminate unused copy of key in RC4.
5117 [Ben Laurie (CHATS)]
5119 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5120 [Ben Laurie (CHATS)]
5122 *) Fix off-by-one error in EGD path.
5123 [Ben Laurie (CHATS)]
5125 *) If RANDFILE path is too long, ignore instead of truncating.
5126 [Ben Laurie (CHATS)]
5128 *) Eliminate unused and incorrectly sized X.509 structure
5130 [Ben Laurie (CHATS)]
5132 *) Eliminate unused and dangerous function knumber().
5133 [Ben Laurie (CHATS)]
5135 *) Eliminate unused and dangerous structure, KSSL_ERR.
5136 [Ben Laurie (CHATS)]
5138 *) Protect against overlong session ID context length in an encoded
5139 session object. Since these are local, this does not appear to be
5141 [Ben Laurie (CHATS)]
5143 *) Change from security patch (see 0.9.6e below) that did not affect
5144 the 0.9.6 release series:
5146 Remote buffer overflow in SSL3 protocol - an attacker could
5147 supply an oversized master key in Kerberos-enabled versions.
5149 [Ben Laurie (CHATS)]
5151 *) Change the SSL kerb5 codes to match RFC 2712.
5154 *) Make -nameopt work fully for req and add -reqopt switch.
5155 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5157 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5158 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5160 *) Make sure tests can be performed even if the corresponding algorithms
5161 have been removed entirely. This was also the last step to make
5162 OpenSSL compilable with DJGPP under all reasonable conditions.
5163 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5165 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5166 to allow version independent disabling of normally unselected ciphers,
5167 which may be activated as a side-effect of selecting a single cipher.
5169 (E.g., cipher list string "RSA" enables ciphersuites that are left
5170 out of "ALL" because they do not provide symmetric encryption.
5171 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5172 [Lutz Jaenicke, Bodo Moeller]
5174 *) Add appropriate support for separate platform-dependent build
5175 directories. The recommended way to make a platform-dependent
5176 build directory is the following (tested on Linux), maybe with
5179 # Place yourself outside of the OpenSSL source tree. In
5180 # this example, the environment variable OPENSSL_SOURCE
5181 # is assumed to contain the absolute OpenSSL source directory.
5182 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5183 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5184 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5185 mkdir -p `dirname $F`
5186 ln -s $OPENSSL_SOURCE/$F $F
5189 To be absolutely sure not to disturb the source tree, a "make clean"
5190 is a good thing. If it isn't successful, don't worry about it,
5191 it probably means the source directory is very clean.
5194 *) Make sure any ENGINE control commands make local copies of string
5195 pointers passed to them whenever necessary. Otherwise it is possible
5196 the caller may have overwritten (or deallocated) the original string
5197 data when a later ENGINE operation tries to use the stored values.
5198 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5200 *) Improve diagnostics in file reading and command-line digests.
5201 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5203 *) Add AES modes CFB and OFB to the object database. Correct an
5204 error in AES-CFB decryption.
5207 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5208 allows existing EVP_CIPHER_CTX structures to be reused after
5209 calling EVP_*Final(). This behaviour is used by encryption
5210 BIOs and some applications. This has the side effect that
5211 applications must explicitly clean up cipher contexts with
5212 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5215 *) Check the values of dna and dnb in bn_mul_recursive before calling
5216 bn_mul_comba (a non zero value means the a or b arrays do not contain
5217 n2 elements) and fallback to bn_mul_normal if either is not zero.
5220 *) Fix escaping of non-ASCII characters when using the -subj option
5221 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5224 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5225 form for "surname", serialNumber has no short form.
5226 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5227 therefore remove "mail" short name for "internet 7".
5228 The OID for unique identifiers in X509 certificates is
5229 x500UniqueIdentifier, not uniqueIdentifier.
5230 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5233 *) Add an "init" command to the ENGINE config module and auto initialize
5234 ENGINEs. Without any "init" command the ENGINE will be initialized
5235 after all ctrl commands have been executed on it. If init=1 the
5236 ENGINE is initailized at that point (ctrls before that point are run
5237 on the uninitialized ENGINE and after on the initialized one). If
5238 init=0 then the ENGINE will not be iniatialized at all.
5241 *) Fix the 'app_verify_callback' interface so that the user-defined
5242 argument is actually passed to the callback: In the
5243 SSL_CTX_set_cert_verify_callback() prototype, the callback
5244 declaration has been changed from
5247 int (*cb)(X509_STORE_CTX *,void *);
5248 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5249 i=s->ctx->app_verify_callback(&ctx)
5250 has been changed into
5251 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5253 To update applications using SSL_CTX_set_cert_verify_callback(),
5254 a dummy argument can be added to their callback functions.
5255 [D. K. Smetters <smetters@parc.xerox.com>]
5257 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5258 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5260 *) Add and OPENSSL_LOAD_CONF define which will cause
5261 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5262 This allows older applications to transparently support certain
5263 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5264 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5265 load the config file and OPENSSL_add_all_algorithms_conf() which will
5266 always load it have also been added.
5269 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5270 Adjust NIDs and EVP layer.
5271 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5273 *) Config modules support in openssl utility.
5275 Most commands now load modules from the config file,
5276 though in a few (such as version) this isn't done
5277 because it couldn't be used for anything.
5279 In the case of ca and req the config file used is
5280 the same as the utility itself: that is the -config
5281 command line option can be used to specify an
5285 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5286 use "openssl_conf" if filename is NULL use default openssl config file.
5289 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5290 config section name. Add a new flag to tolerate a missing config file
5291 and move code to CONF_modules_load_file().
5294 *) Support for crypto accelerator cards from Accelerated Encryption
5295 Processing, www.aep.ie. (Use engine 'aep')
5296 The support was copied from 0.9.6c [engine] and adapted/corrected
5297 to work with the new engine framework.
5298 [AEP Inc. and Richard Levitte]
5300 *) Support for SureWare crypto accelerator cards from Baltimore
5301 Technologies. (Use engine 'sureware')
5302 The support was copied from 0.9.6c [engine] and adapted
5303 to work with the new engine framework.
5306 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5307 make the newer ENGINE framework commands for the CHIL engine work.
5308 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5310 *) Make it possible to produce shared libraries on ReliantUNIX.
5311 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5313 *) Add the configuration target debug-linux-ppro.
5314 Make 'openssl rsa' use the general key loading routines
5315 implemented in apps.c, and make those routines able to
5316 handle the key format FORMAT_NETSCAPE and the variant
5318 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5320 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5321 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5323 *) Add -keyform to rsautl, and document -engine.
5324 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5326 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5327 BIO_R_NO_SUCH_FILE error code rather than the generic
5328 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5331 *) Add new functions
5333 ERR_peek_last_error_line
5334 ERR_peek_last_error_line_data.
5335 These are similar to
5338 ERR_peek_error_line_data,
5339 but report on the latest error recorded rather than the first one
5340 still in the error queue.
5341 [Ben Laurie, Bodo Moeller]
5343 *) default_algorithms option in ENGINE config module. This allows things
5345 default_algorithms = ALL
5346 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5349 *) Preliminary ENGINE config module.
5352 *) New experimental application configuration code.
5355 *) Change the AES code to follow the same name structure as all other
5356 symmetric ciphers, and behave the same way. Move everything to
5357 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5358 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5360 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5361 [Ben Laurie and Theo de Raadt]
5363 *) Add option to output public keys in req command.
5364 [Massimiliano Pala madwolf@openca.org]
5366 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5367 (up to about 10% better than before for P-192 and P-224).
5370 *) New functions/macros
5372 SSL_CTX_set_msg_callback(ctx, cb)
5373 SSL_CTX_set_msg_callback_arg(ctx, arg)
5374 SSL_set_msg_callback(ssl, cb)
5375 SSL_set_msg_callback_arg(ssl, arg)
5377 to request calling a callback function
5379 void cb(int write_p, int version, int content_type,
5380 const void *buf, size_t len, SSL *ssl, void *arg)
5382 whenever a protocol message has been completely received
5383 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5384 protocol version according to which the SSL library interprets
5385 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5386 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5387 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5388 specification (change_cipher_spec(20), alert(21), handshake(22)).
5389 'buf' and 'len' point to the actual message, 'ssl' to the
5390 SSL object, and 'arg' is the application-defined value set by
5391 SSL[_CTX]_set_msg_callback_arg().
5393 'openssl s_client' and 'openssl s_server' have new '-msg' options
5394 to enable a callback that displays all protocol messages.
5397 *) Change the shared library support so shared libraries are built as
5398 soon as the corresponding static library is finished, and thereby get
5399 openssl and the test programs linked against the shared library.
5400 This still only happens when the keyword "shard" has been given to
5401 the configuration scripts.
5403 NOTE: shared library support is still an experimental thing, and
5404 backward binary compatibility is still not guaranteed.
5405 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5407 *) Add support for Subject Information Access extension.
5408 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5410 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5411 additional bytes when new memory had to be allocated, not just
5412 when reusing an existing buffer.
5415 *) New command line and configuration option 'utf8' for the req command.
5416 This allows field values to be specified as UTF8 strings.
5419 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5420 runs for the former and machine-readable output for the latter.
5423 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5424 of the e-mail address in the DN (i.e., it will go into a certificate
5425 extension only). The new configuration file option 'email_in_dn = no'
5426 has the same effect.
5427 [Massimiliano Pala madwolf@openca.org]
5429 *) Change all functions with names starting with des_ to be starting
5430 with DES_ instead. Add wrappers that are compatible with libdes,
5431 but are named _ossl_old_des_*. Finally, add macros that map the
5432 des_* symbols to the corresponding _ossl_old_des_* if libdes
5433 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5434 desired, the des_* symbols will be mapped to DES_*, with one
5437 Since we provide two compatibility mappings, the user needs to
5438 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5439 compatibility is desired. The default (i.e., when that macro
5440 isn't defined) is OpenSSL 0.9.6c compatibility.
5442 There are also macros that enable and disable the support of old
5443 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5444 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5445 are defined, the default will apply: to support the old des routines.
5447 In either case, one must include openssl/des.h to get the correct
5448 definitions. Do not try to just include openssl/des_old.h, that
5451 NOTE: This is a major break of an old API into a new one. Software
5452 authors are encouraged to switch to the DES_ style functions. Some
5453 time in the future, des_old.h and the libdes compatibility functions
5454 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5455 default), and then completely removed.
5458 *) Test for certificates which contain unsupported critical extensions.
5459 If such a certificate is found during a verify operation it is
5460 rejected by default: this behaviour can be overridden by either
5461 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5462 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5463 X509_supported_extension() has also been added which returns 1 if a
5464 particular extension is supported.
5467 *) Modify the behaviour of EVP cipher functions in similar way to digests
5468 to retain compatibility with existing code.
5471 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5472 compatibility with existing code. In particular the 'ctx' parameter does
5473 not have to be to be initialized before the call to EVP_DigestInit() and
5474 it is tidied up after a call to EVP_DigestFinal(). New function
5475 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5476 EVP_MD_CTX_copy() changed to not require the destination to be
5477 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5478 requires the destination to be valid.
5480 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5481 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5484 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5485 so that complete 'Handshake' protocol structures are kept in memory
5486 instead of overwriting 'msg_type' and 'length' with 'body' data.
5489 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5490 [Massimo Santin via Richard Levitte]
5492 *) Major restructuring to the underlying ENGINE code. This includes
5493 reduction of linker bloat, separation of pure "ENGINE" manipulation
5494 (initialisation, etc) from functionality dealing with implementations
5495 of specific crypto iterfaces. This change also introduces integrated
5496 support for symmetric ciphers and digest implementations - so ENGINEs
5497 can now accelerate these by providing EVP_CIPHER and EVP_MD
5498 implementations of their own. This is detailed in crypto/engine/README
5499 as it couldn't be adequately described here. However, there are a few
5500 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5501 were changed in the original introduction of ENGINE code have now
5502 reverted back - the hooking from this code to ENGINE is now a good
5503 deal more passive and at run-time, operations deal directly with
5504 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5505 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5506 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5507 they were not being used by the framework as there is no concept of a
5508 BIGNUM_METHOD and they could not be generalised to the new
5509 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5510 ENGINE_cpy() has been removed as it cannot be consistently defined in
5514 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5517 *) Change mkdef.pl to sort symbols that get the same entry number,
5518 and make sure the automatically generated functions ERR_load_*
5519 become part of libeay.num as well.
5522 *) New function SSL_renegotiate_pending(). This returns true once
5523 renegotiation has been requested (either SSL_renegotiate() call
5524 or HelloRequest/ClientHello received from the peer) and becomes
5525 false once a handshake has been completed.
5526 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5527 sends a HelloRequest, but does not ensure that a handshake takes
5528 place. SSL_renegotiate_pending() is useful for checking if the
5529 client has followed the request.)
5532 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5533 By default, clients may request session resumption even during
5534 renegotiation (if session ID contexts permit); with this option,
5535 session resumption is possible only in the first handshake.
5537 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5538 more bits available for options that should not be part of
5539 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5542 *) Add some demos for certificate and certificate request creation.
5545 *) Make maximum certificate chain size accepted from the peer application
5546 settable (SSL*_get/set_max_cert_list()), as proposed by
5547 "Douglas E. Engert" <deengert@anl.gov>.
5550 *) Add support for shared libraries for Unixware-7
5551 (Boyd Lynn Gerber <gerberb@zenez.com>).
5554 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5555 be done prior to destruction. Use this to unload error strings from
5556 ENGINEs that load their own error strings. NB: This adds two new API
5557 functions to "get" and "set" this destroy handler in an ENGINE.
5560 *) Alter all existing ENGINE implementations (except "openssl" and
5561 "openbsd") to dynamically instantiate their own error strings. This
5562 makes them more flexible to be built both as statically-linked ENGINEs
5563 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5564 Also, add stub code to each that makes building them as self-contained
5565 shared-libraries easier (see README.ENGINE).
5568 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5569 implementations into applications that are completely implemented in
5570 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5571 commands that can be used to configure what shared-library to load and
5572 to control aspects of the way it is handled. Also, made an update to
5573 the README.ENGINE file that brings its information up-to-date and
5574 provides some information and instructions on the "dynamic" ENGINE
5575 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5578 *) Make it possible to unload ranges of ERR strings with a new
5579 "ERR_unload_strings" function.
5582 *) Add a copy() function to EVP_MD.
5585 *) Make EVP_MD routines take a context pointer instead of just the
5586 md_data void pointer.
5589 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5590 that the digest can only process a single chunk of data
5591 (typically because it is provided by a piece of
5592 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5593 is only going to provide a single chunk of data, and hence the
5594 framework needn't accumulate the data for oneshot drivers.
5597 *) As with "ERR", make it possible to replace the underlying "ex_data"
5598 functions. This change also alters the storage and management of global
5599 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5600 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5601 index counters. The API functions that use this state have been changed
5602 to take a "class_index" rather than pointers to the class's local STACK
5603 and counter, and there is now an API function to dynamically create new
5604 classes. This centralisation allows us to (a) plug a lot of the
5605 thread-safety problems that existed, and (b) makes it possible to clean
5606 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5607 such data would previously have always leaked in application code and
5608 workarounds were in place to make the memory debugging turn a blind eye
5609 to it. Application code that doesn't use this new function will still
5610 leak as before, but their memory debugging output will announce it now
5611 rather than letting it slide.
5613 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5614 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5615 has a return value to indicate success or failure.
5618 *) Make it possible to replace the underlying "ERR" functions such that the
5619 global state (2 LHASH tables and 2 locks) is only used by the "default"
5620 implementation. This change also adds two functions to "get" and "set"
5621 the implementation prior to it being automatically set the first time
5622 any other ERR function takes place. Ie. an application can call "get",
5623 pass the return value to a module it has just loaded, and that module
5624 can call its own "set" function using that value. This means the
5625 module's "ERR" operations will use (and modify) the error state in the
5626 application and not in its own statically linked copy of OpenSSL code.
5629 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5630 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5631 the operation, and provides a more encapsulated way for external code
5632 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5633 to use these functions rather than manually incrementing the counts.
5635 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5638 *) Add EVP test program.
5641 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5644 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5645 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5646 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5647 These allow a CRL to be built without having to access X509_CRL fields
5648 directly. Modify 'ca' application to use new functions.
5651 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5652 bug workarounds. Rollback attack detection is a security feature.
5653 The problem will only arise on OpenSSL servers when TLSv1 is not
5654 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5655 Software authors not wanting to support TLSv1 will have special reasons
5656 for their choice and can explicitly enable this option.
5657 [Bodo Moeller, Lutz Jaenicke]
5659 *) Rationalise EVP so it can be extended: don't include a union of
5660 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5661 (similar to those existing for EVP_CIPHER_CTX).
5666 EVP_MD_CTX_init(&md); /* new function call */
5667 EVP_DigestInit(&md, EVP_sha1());
5668 EVP_DigestUpdate(&md, in, len);
5669 EVP_DigestFinal(&md, out, NULL);
5670 EVP_MD_CTX_cleanup(&md); /* new function call */
5674 *) Make DES key schedule conform to the usual scheme, as well as
5675 correcting its structure. This means that calls to DES functions
5676 now have to pass a pointer to a des_key_schedule instead of a
5677 plain des_key_schedule (which was actually always a pointer
5680 des_key_schedule ks;
5682 des_set_key_checked(..., &ks);
5683 des_ncbc_encrypt(..., &ks, ...);
5685 (Note that a later change renames 'des_...' into 'DES_...'.)
5688 *) Initial reduction of linker bloat: the use of some functions, such as
5689 PEM causes large amounts of unused functions to be linked in due to
5690 poor organisation. For example pem_all.c contains every PEM function
5691 which has a knock on effect of linking in large amounts of (unused)
5692 ASN1 code. Grouping together similar functions and splitting unrelated
5693 functions prevents this.
5696 *) Cleanup of EVP macros.
5699 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5700 correct _ecb suffix.
5703 *) Add initial OCSP responder support to ocsp application. The
5704 revocation information is handled using the text based index
5705 use by the ca application. The responder can either handle
5706 requests generated internally, supplied in files (for example
5707 via a CGI script) or using an internal minimal server.
5710 *) Add configuration choices to get zlib compression for TLS.
5713 *) Changes to Kerberos SSL for RFC 2712 compliance:
5714 1. Implemented real KerberosWrapper, instead of just using
5715 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5716 2. Implemented optional authenticator field of KerberosWrapper.
5718 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5719 and authenticator structs; see crypto/krb5/.
5721 Generalized Kerberos calls to support multiple Kerberos libraries.
5722 [Vern Staats <staatsvr@asc.hpc.mil>,
5723 Jeffrey Altman <jaltman@columbia.edu>
5724 via Richard Levitte]
5726 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5727 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5728 values for each of the key sizes rather than having just
5729 parameters (and 'speed' generating keys each time).
5732 *) Speed up EVP routines.
5735 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5736 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5737 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5738 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5740 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5741 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5742 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5745 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5747 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5750 *) Added the OS2-EMX target.
5751 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5753 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5754 to support NCONF routines in extension code. New function CONF_set_nconf()
5755 to allow functions which take an NCONF to also handle the old LHASH
5756 structure: this means that the old CONF compatible routines can be
5757 retained (in particular wrt extensions) without having to duplicate the
5758 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5761 *) Enhance the general user interface with mechanisms for inner control
5762 and with possibilities to have yes/no kind of prompts.
5765 *) Change all calls to low level digest routines in the library and
5766 applications to use EVP. Add missing calls to HMAC_cleanup() and
5767 don't assume HMAC_CTX can be copied using memcpy().
5768 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5770 *) Add the possibility to control engines through control names but with
5771 arbitrary arguments instead of just a string.
5772 Change the key loaders to take a UI_METHOD instead of a callback
5773 function pointer. NOTE: this breaks binary compatibility with earlier
5774 versions of OpenSSL [engine].
5775 Adapt the nCipher code for these new conditions and add a card insertion
5779 *) Enhance the general user interface with mechanisms to better support
5780 dialog box interfaces, application-defined prompts, the possibility
5781 to use defaults (for example default passwords from somewhere else)
5782 and interrupts/cancellations.
5785 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5786 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5789 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5790 tidy up some unnecessarily weird code in 'sk_new()').
5791 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5793 *) Change the key loading routines for ENGINEs to use the same kind
5794 callback (pem_password_cb) as all other routines that need this
5798 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5799 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5800 than this minimum value is recommended.
5803 *) New random seeder for OpenVMS, using the system process statistics
5804 that are easily reachable.
5807 *) Windows apparently can't transparently handle global
5808 variables defined in DLLs. Initialisations such as:
5810 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5812 wont compile. This is used by the any applications that need to
5813 declare their own ASN1 modules. This was fixed by adding the option
5814 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5815 needed for static libraries under Win32.
5818 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5819 setting of purpose and trust fields. New X509_STORE trust and
5820 purpose functions and tidy up setting in other SSL functions.
5823 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5824 structure. These are inherited by X509_STORE_CTX when it is
5825 initialised. This allows various defaults to be set in the
5826 X509_STORE structure (such as flags for CRL checking and custom
5827 purpose or trust settings) for functions which only use X509_STORE_CTX
5828 internally such as S/MIME.
5830 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5831 trust settings if they are not set in X509_STORE. This allows X509_STORE
5832 purposes and trust (in S/MIME for example) to override any set by default.
5834 Add command line options for CRL checking to smime, s_client and s_server
5838 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5839 are set then the CRL is looked up in the X509_STORE structure and
5840 its validity and signature checked, then if the certificate is found
5841 in the CRL the verify fails with a revoked error.
5843 Various new CRL related callbacks added to X509_STORE_CTX structure.
5845 Command line options added to 'verify' application to support this.
5847 This needs some additional work, such as being able to handle multiple
5848 CRLs with different times, extension based lookup (rather than just
5849 by subject name) and ultimately more complete V2 CRL extension
5853 *) Add a general user interface API (crypto/ui/). This is designed
5854 to replace things like des_read_password and friends (backward
5855 compatibility functions using this new API are provided).
5856 The purpose is to remove prompting functions from the DES code
5857 section as well as provide for prompting through dialog boxes in
5858 a window system and the like.
5861 *) Add "ex_data" support to ENGINE so implementations can add state at a
5862 per-structure level rather than having to store it globally.
5865 *) Make it possible for ENGINE structures to be copied when retrieved by
5866 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5867 This causes the "original" ENGINE structure to act like a template,
5868 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5869 operational state can be localised to each ENGINE structure, despite the
5870 fact they all share the same "methods". New ENGINE structures returned in
5871 this case have no functional references and the return value is the single
5872 structural reference. This matches the single structural reference returned
5873 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5877 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5878 needs to match any other type at all we need to manually clear the
5882 *) Changes to the "openssl engine" utility to include;
5883 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5884 about an ENGINE's available control commands.
5885 - executing control commands from command line arguments using the
5886 '-pre' and '-post' switches. '-post' is only used if '-t' is
5887 specified and the ENGINE is successfully initialised. The syntax for
5888 the individual commands are colon-separated, for example;
5889 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5892 *) New dynamic control command support for ENGINEs. ENGINEs can now
5893 declare their own commands (numbers), names (strings), descriptions,
5894 and input types for run-time discovery by calling applications. A
5895 subset of these commands are implicitly classed as "executable"
5896 depending on their input type, and only these can be invoked through
5897 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5898 can be based on user input, config files, etc). The distinction is
5899 that "executable" commands cannot return anything other than a boolean
5900 result and can only support numeric or string input, whereas some
5901 discoverable commands may only be for direct use through
5902 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5903 pointers, or other custom uses. The "executable" commands are to
5904 support parameterisations of ENGINE behaviour that can be
5905 unambiguously defined by ENGINEs and used consistently across any
5906 OpenSSL-based application. Commands have been added to all the
5907 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5908 control over shared-library paths without source code alterations.
5911 *) Changed all ENGINE implementations to dynamically allocate their
5912 ENGINEs rather than declaring them statically. Apart from this being
5913 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5914 this also allows the implementations to compile without using the
5915 internal engine_int.h header.
5918 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5919 'const' value. Any code that should be able to modify a RAND_METHOD
5920 should already have non-const pointers to it (ie. they should only
5921 modify their own ones).
5924 *) Made a variety of little tweaks to the ENGINE code.
5925 - "atalla" and "ubsec" string definitions were moved from header files
5926 to C code. "nuron" string definitions were placed in variables
5927 rather than hard-coded - allowing parameterisation of these values
5928 later on via ctrl() commands.
5929 - Removed unused "#if 0"'d code.
5930 - Fixed engine list iteration code so it uses ENGINE_free() to release
5931 structural references.
5932 - Constified the RAND_METHOD element of ENGINE structures.
5933 - Constified various get/set functions as appropriate and added
5934 missing functions (including a catch-all ENGINE_cpy that duplicates
5935 all ENGINE values onto a new ENGINE except reference counts/state).
5936 - Removed NULL parameter checks in get/set functions. Setting a method
5937 or function to NULL is a way of cancelling out a previously set
5938 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5939 and doesn't justify the extra error symbols and code.
5940 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5941 flags from engine_int.h to engine.h.
5942 - Changed prototypes for ENGINE handler functions (init(), finish(),
5943 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5946 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5947 to the algorithm using long division. The binary algorithm can be
5948 used only if the modulus is odd. On 32-bit systems, it is faster
5949 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5950 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5951 up to 450 bits. In 64-bit environments, the binary algorithm
5952 appears to be advantageous for much longer moduli; here we use it
5953 for moduli up to 2048 bits.
5956 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5957 could not support the combine flag in choice fields.
5960 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5961 extensions from a certificate request to the certificate.
5964 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5965 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5966 file: this allows the display of the certificate about to be
5967 signed to be customised, to allow certain fields to be included
5968 or excluded and extension details. The old system didn't display
5969 multicharacter strings properly, omitted fields not in the policy
5970 and couldn't display additional details such as extensions.
5973 *) Function EC_POINTs_mul for multiple scalar multiplication
5974 of an arbitrary number of elliptic curve points
5975 \sum scalars[i]*points[i],
5976 optionally including the generator defined for the EC_GROUP:
5977 scalar*generator + \sum scalars[i]*points[i].
5979 EC_POINT_mul is a simple wrapper function for the typical case
5980 that the point list has just one item (besides the optional
5984 *) First EC_METHODs for curves over GF(p):
5986 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5987 operations and provides various method functions that can also
5988 operate with faster implementations of modular arithmetic.
5990 EC_GFp_mont_method() reuses most functions that are part of
5991 EC_GFp_simple_method, but uses Montgomery arithmetic.
5993 [Bodo Moeller; point addition and point doubling
5994 implementation directly derived from source code provided by
5995 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5997 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5998 crypto/ec/ec_lib.c):
6000 Curves are EC_GROUP objects (with an optional group generator)
6001 based on EC_METHODs that are built into the library.
6003 Points are EC_POINT objects based on EC_GROUP objects.
6005 Most of the framework would be able to handle curves over arbitrary
6006 finite fields, but as there are no obvious types for fields other
6007 than GF(p), some functions are limited to that for now.
6010 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6011 that the file contains a complete HTTP response.
6014 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6015 change the def and num file printf format specifier from "%-40sXXX"
6016 to "%-39s XXX". The latter will always guarantee a space after the
6017 field while the former will cause them to run together if the field
6018 is 40 of more characters long.
6021 *) Constify the cipher and digest 'method' functions and structures
6022 and modify related functions to take constant EVP_MD and EVP_CIPHER
6026 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6027 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6030 *) Modify EVP_Digest*() routines so they now return values. Although the
6031 internal software routines can never fail additional hardware versions
6035 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6037 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6038 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6041 ERR_R_NESTED_ASN1_ERROR
6043 ERR_R_MISSING_ASN1_EOS
6044 were 4 .. 9, conflicting with
6045 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6047 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6048 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6050 Add new error code 'ERR_R_INTERNAL_ERROR'.
6053 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6057 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6058 sets the subject name for a new request or supersedes the
6059 subject name in a given request. Formats that can be parsed are
6060 'CN=Some Name, OU=myOU, C=IT'
6062 'CN=Some Name/OU=myOU/C=IT'.
6064 Add options '-batch' and '-verbose' to 'openssl req'.
6065 [Massimiliano Pala <madwolf@hackmasters.net>]
6067 *) Introduce the possibility to access global variables through
6068 functions on platform were that's the best way to handle exporting
6069 global variables in shared libraries. To enable this functionality,
6070 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6071 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6072 is normally done by Configure or something similar).
6074 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6075 in the source file (foo.c) like this:
6077 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6078 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6080 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6081 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6083 OPENSSL_DECLARE_GLOBAL(int,foo);
6084 #define foo OPENSSL_GLOBAL_REF(foo)
6085 OPENSSL_DECLARE_GLOBAL(double,bar);
6086 #define bar OPENSSL_GLOBAL_REF(bar)
6088 The #defines are very important, and therefore so is including the
6089 header file everywhere where the defined globals are used.
6091 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6092 of ASN.1 items, but that structure is a bit different.
6094 The largest change is in util/mkdef.pl which has been enhanced with
6095 better and easier to understand logic to choose which symbols should
6096 go into the Windows .def files as well as a number of fixes and code
6097 cleanup (among others, algorithm keywords are now sorted
6098 lexicographically to avoid constant rewrites).
6101 *) In BN_div() keep a copy of the sign of 'num' before writing the
6102 result to 'rm' because if rm==num the value will be overwritten
6103 and produce the wrong result if 'num' is negative: this caused
6104 problems with BN_mod() and BN_nnmod().
6107 *) Function OCSP_request_verify(). This checks the signature on an
6108 OCSP request and verifies the signer certificate. The signer
6109 certificate is just checked for a generic purpose and OCSP request
6113 *) Add OCSP_check_validity() function to check the validity of OCSP
6114 responses. OCSP responses are prepared in real time and may only
6115 be a few seconds old. Simply checking that the current time lies
6116 between thisUpdate and nextUpdate max reject otherwise valid responses
6117 caused by either OCSP responder or client clock inaccuracy. Instead
6118 we allow thisUpdate and nextUpdate to fall within a certain period of
6119 the current time. The age of the response can also optionally be
6120 checked. Two new options -validity_period and -status_age added to
6124 *) If signature or public key algorithm is unrecognized print out its
6125 OID rather that just UNKNOWN.
6128 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6129 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6130 ID to be generated from the issuer certificate alone which can then be
6131 passed to OCSP_id_issuer_cmp().
6134 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6135 ASN1 modules to export functions returning ASN1_ITEM pointers
6136 instead of the ASN1_ITEM structures themselves. This adds several
6137 new macros which allow the underlying ASN1 function/structure to
6138 be accessed transparently. As a result code should not use ASN1_ITEM
6139 references directly (such as &X509_it) but instead use the relevant
6140 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6141 use of the new ASN1 code on platforms where exporting structures
6142 is problematical (for example in shared libraries) but exporting
6143 functions returning pointers to structures is not.
6146 *) Add support for overriding the generation of SSL/TLS session IDs.
6147 These callbacks can be registered either in an SSL_CTX or per SSL.
6148 The purpose of this is to allow applications to control, if they wish,
6149 the arbitrary values chosen for use as session IDs, particularly as it
6150 can be useful for session caching in multiple-server environments. A
6151 command-line switch for testing this (and any client code that wishes
6152 to use such a feature) has been added to "s_server".
6153 [Geoff Thorpe, Lutz Jaenicke]
6155 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6156 of the form '#if defined(...) || defined(...) || ...' and
6157 '#if !defined(...) && !defined(...) && ...'. This also avoids
6158 the growing number of special cases it was previously handling.
6161 *) Make all configuration macros available for application by making
6162 sure they are available in opensslconf.h, by giving them names starting
6163 with "OPENSSL_" to avoid conflicts with other packages and by making
6164 sure e_os2.h will cover all platform-specific cases together with
6166 Additionally, it is now possible to define configuration/platform-
6167 specific names (called "system identities"). In the C code, these
6168 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6169 macro with the name beginning with "OPENSSL_SYS_", which is determined
6170 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6174 *) New option -set_serial to 'req' and 'x509' this allows the serial
6175 number to use to be specified on the command line. Previously self
6176 signed certificates were hard coded with serial number 0 and the
6177 CA options of 'x509' had to use a serial number in a file which was
6181 *) New options to 'ca' utility to support V2 CRL entry extensions.
6182 Currently CRL reason, invalidity date and hold instruction are
6183 supported. Add new CRL extensions to V3 code and some new objects.
6186 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6187 disable standard block padding (aka PKCS#5 padding) in the EVP
6188 API, which was previously mandatory. This means that the data is
6189 not padded in any way and so the total length much be a multiple
6190 of the block size, otherwise an error occurs.
6193 *) Initial (incomplete) OCSP SSL support.
6196 *) New function OCSP_parse_url(). This splits up a URL into its host,
6197 port and path components: primarily to parse OCSP URLs. New -url
6198 option to ocsp utility.
6201 *) New nonce behavior. The return value of OCSP_check_nonce() now
6202 reflects the various checks performed. Applications can decide
6203 whether to tolerate certain situations such as an absent nonce
6204 in a response when one was present in a request: the ocsp application
6205 just prints out a warning. New function OCSP_add1_basic_nonce()
6206 this is to allow responders to include a nonce in a response even if
6207 the request is nonce-less.
6210 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6211 skipped when using openssl x509 multiple times on a single input file,
6212 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6215 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6216 set string type: to handle setting ASN1_TIME structures. Fix ca
6217 utility to correctly initialize revocation date of CRLs.
6220 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6221 the clients preferred ciphersuites and rather use its own preferences.
6222 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6223 Internet Explorer by ensuring unchanged hash method during stepup.
6224 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6227 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6228 to aes and add a new 'exist' option to print out symbols that don't
6232 *) Additional options to ocsp utility to allow flags to be set and
6233 additional certificates supplied.
6236 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6237 OCSP client a number of certificate to only verify the response
6241 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6242 handle the new API. Currently only ECB, CBC modes supported. Add new
6245 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6246 Encryption Standard (AES) Ciphersuites for Transport Layer
6247 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6248 not enabled by default and were not part of the "ALL" ciphersuite
6249 alias because they were not yet official; they could be
6250 explicitly requested by specifying the "AESdraft" ciphersuite
6251 group alias. In the final release of OpenSSL 0.9.7, the group
6252 alias is called "AES" and is part of "ALL".)
6253 [Ben Laurie, Steve Henson, Bodo Moeller]
6255 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6256 request to response.
6259 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6260 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6261 extract information from a certificate request. OCSP_response_create()
6262 creates a response and optionally adds a basic response structure.
6263 OCSP_basic_add1_status() adds a complete single response to a basic
6264 response and returns the OCSP_SINGLERESP structure just added (to allow
6265 extensions to be included for example). OCSP_basic_add1_cert() adds a
6266 certificate to a basic response and OCSP_basic_sign() signs a basic
6267 response with various flags. New helper functions ASN1_TIME_check()
6268 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6269 (converts ASN1_TIME to GeneralizedTime).
6272 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6273 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6274 structure from a certificate. X509_pubkey_digest() digests the public_key
6275 contents: this is used in various key identifiers.
6278 *) Make sk_sort() tolerate a NULL argument.
6279 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6281 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6282 passed by the function are trusted implicitly. If any of them signed the
6283 response then it is assumed to be valid and is not verified.
6286 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6287 to data. This was previously part of the PKCS7 ASN1 code. This
6288 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6289 [Steve Henson, reported by Kenneth R. Robinette
6290 <support@securenetterm.com>]
6292 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6293 routines: without these tracing memory leaks is very painful.
6294 Fix leaks in PKCS12 and PKCS7 routines.
6297 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6298 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6299 effectively meant GeneralizedTime would never be used. Now it
6300 is initialised to -1 but X509_time_adj() now has to check the value
6301 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6302 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6303 [Steve Henson, reported by Kenneth R. Robinette
6304 <support@securenetterm.com>]
6306 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6307 result in a zero length in the ASN1_INTEGER structure which was
6308 not consistent with the structure when d2i_ASN1_INTEGER() was used
6309 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6310 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6311 where it did not print out a minus for negative ASN1_INTEGER.
6314 *) Add summary printout to ocsp utility. The various functions which
6315 convert status values to strings have been renamed to:
6316 OCSP_response_status_str(), OCSP_cert_status_str() and
6317 OCSP_crl_reason_str() and are no longer static. New options
6318 to verify nonce values and to disable verification. OCSP response
6319 printout format cleaned up.
6322 *) Add additional OCSP certificate checks. These are those specified
6323 in RFC2560. This consists of two separate checks: the CA of the
6324 certificate being checked must either be the OCSP signer certificate
6325 or the issuer of the OCSP signer certificate. In the latter case the
6326 OCSP signer certificate must contain the OCSP signing extended key
6327 usage. This check is performed by attempting to match the OCSP
6328 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6329 in the OCSP_CERTID structures of the response.
6332 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6333 and related routines. This uses the standard OpenSSL certificate
6334 verify routines to perform initial checks (just CA validity) and
6335 to obtain the certificate chain. Then additional checks will be
6336 performed on the chain. Currently the root CA is checked to see
6337 if it is explicitly trusted for OCSP signing. This is used to set
6338 a root CA as a global signing root: that is any certificate that
6339 chains to that CA is an acceptable OCSP signing certificate.
6342 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6343 extensions from a separate configuration file.
6344 As when reading extensions from the main configuration file,
6345 the '-extensions ...' option may be used for specifying the
6347 [Massimiliano Pala <madwolf@comune.modena.it>]
6349 *) New OCSP utility. Allows OCSP requests to be generated or
6350 read. The request can be sent to a responder and the output
6351 parsed, outputed or printed in text form. Not complete yet:
6352 still needs to check the OCSP response validity.
6355 *) New subcommands for 'openssl ca':
6356 'openssl ca -status <serial>' prints the status of the cert with
6357 the given serial number (according to the index file).
6358 'openssl ca -updatedb' updates the expiry status of certificates
6360 [Massimiliano Pala <madwolf@comune.modena.it>]
6362 *) New '-newreq-nodes' command option to CA.pl. This is like
6363 '-newreq', but calls 'openssl req' with the '-nodes' option
6364 so that the resulting key is not encrypted.
6365 [Damien Miller <djm@mindrot.org>]
6367 *) New configuration for the GNU Hurd.
6368 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6370 *) Initial code to implement OCSP basic response verify. This
6371 is currently incomplete. Currently just finds the signer's
6372 certificate and verifies the signature on the response.
6375 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6376 value of OPENSSLDIR. This is available via the new '-d' option
6377 to 'openssl version', and is also included in 'openssl version -a'.
6380 *) Allowing defining memory allocation callbacks that will be given
6381 file name and line number information in additional arguments
6382 (a const char* and an int). The basic functionality remains, as
6383 well as the original possibility to just replace malloc(),
6384 realloc() and free() by functions that do not know about these
6385 additional arguments. To register and find out the current
6386 settings for extended allocation functions, the following
6387 functions are provided:
6389 CRYPTO_set_mem_ex_functions
6390 CRYPTO_set_locked_mem_ex_functions
6391 CRYPTO_get_mem_ex_functions
6392 CRYPTO_get_locked_mem_ex_functions
6394 These work the same way as CRYPTO_set_mem_functions and friends.
6395 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6396 extended allocation function is enabled.
6397 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6398 a conventional allocation function is enabled.
6399 [Richard Levitte, Bodo Moeller]
6401 *) Finish off removing the remaining LHASH function pointer casts.
6402 There should no longer be any prototype-casting required when using
6403 the LHASH abstraction, and any casts that remain are "bugs". See
6404 the callback types and macros at the head of lhash.h for details
6405 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6408 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6409 If /dev/[u]random devices are not available or do not return enough
6410 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6412 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6413 /etc/entropy will be queried once each in this sequence, quering stops
6414 when enough entropy was collected without querying more sockets.
6417 *) Change the Unix RAND_poll() variant to be able to poll several
6418 random devices, as specified by DEVRANDOM, until a sufficient amount
6419 of data has been collected. We spend at most 10 ms on each file
6420 (select timeout) and read in non-blocking mode. DEVRANDOM now
6421 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6422 (previously it was just the string "/dev/urandom"), so on typical
6423 platforms the 10 ms delay will never occur.
6424 Also separate out the Unix variant to its own file, rand_unix.c.
6425 For VMS, there's a currently-empty rand_vms.c.
6428 *) Move OCSP client related routines to ocsp_cl.c. These
6429 provide utility functions which an application needing
6430 to issue a request to an OCSP responder and analyse the
6431 response will typically need: as opposed to those which an
6432 OCSP responder itself would need which will be added later.
6434 OCSP_request_sign() signs an OCSP request with an API similar
6435 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6436 response. OCSP_response_get1_basic() extracts basic response
6437 from response. OCSP_resp_find_status(): finds and extracts status
6438 information from an OCSP_CERTID structure (which will be created
6439 when the request structure is built). These are built from lower
6440 level functions which work on OCSP_SINGLERESP structures but
6441 wont normally be used unless the application wishes to examine
6442 extensions in the OCSP response for example.
6444 Replace nonce routines with a pair of functions.
6445 OCSP_request_add1_nonce() adds a nonce value and optionally
6446 generates a random value. OCSP_check_nonce() checks the
6447 validity of the nonce in an OCSP response.
6450 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6451 This doesn't copy the supplied OCSP_CERTID and avoids the
6452 need to free up the newly created id. Change return type
6453 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6454 This can then be used to add extensions to the request.
6455 Deleted OCSP_request_new(), since most of its functionality
6456 is now in OCSP_REQUEST_new() (and the case insensitive name
6457 clash) apart from the ability to set the request name which
6458 will be added elsewhere.
6461 *) Update OCSP API. Remove obsolete extensions argument from
6462 various functions. Extensions are now handled using the new
6463 OCSP extension code. New simple OCSP HTTP function which
6464 can be used to send requests and parse the response.
6467 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6468 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6469 uses the special reorder version of SET OF to sort the attributes
6470 and reorder them to match the encoded order. This resolves a long
6471 standing problem: a verify on a PKCS7 structure just after signing
6472 it used to fail because the attribute order did not match the
6473 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6474 it uses the received order. This is necessary to tolerate some broken
6475 software that does not order SET OF. This is handled by encoding
6476 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6477 to produce the required SET OF.
6480 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6481 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6482 files to get correct declarations of the ASN.1 item variables.
6485 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6486 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6487 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6488 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6489 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6490 ASN1_ITEM and no wrapper functions.
6493 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6494 replace the old function pointer based I/O routines. Change most of
6495 the *_d2i_bio() and *_d2i_fp() functions to use these.
6498 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6499 lines, recognice more "algorithms" that can be deselected, and make
6500 it complain about algorithm deselection that isn't recognised.
6503 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6504 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6505 to use new functions. Add NO_ASN1_OLD which can be set to remove
6506 some old style ASN1 functions: this can be used to determine if old
6507 code will still work when these eventually go away.
6510 *) New extension functions for OCSP structures, these follow the
6511 same conventions as certificates and CRLs.
6514 *) New function X509V3_add1_i2d(). This automatically encodes and
6515 adds an extension. Its behaviour can be customised with various
6516 flags to append, replace or delete. Various wrappers added for
6517 certifcates and CRLs.
6520 *) Fix to avoid calling the underlying ASN1 print routine when
6521 an extension cannot be parsed. Correct a typo in the
6522 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6525 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6526 entries for variables.
6529 *) Add functionality to apps/openssl.c for detecting locking
6530 problems: As the program is single-threaded, all we have
6531 to do is register a locking callback using an array for
6532 storing which locks are currently held by the program.
6535 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6536 SSL_get_ex_data_X509_STORE_idx(), which is used in
6537 ssl_verify_cert_chain() and thus can be called at any time
6538 during TLS/SSL handshakes so that thread-safety is essential.
6539 Unfortunately, the ex_data design is not at all suited
6540 for multi-threaded use, so it probably should be abolished.
6543 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6544 [Broadcom, tweaked and integrated by Geoff Thorpe]
6546 *) Move common extension printing code to new function
6547 X509V3_print_extensions(). Reorganise OCSP print routines and
6548 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6551 *) New function X509_signature_print() to remove duplication in some
6555 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6556 set (this was treated exactly the same as SET OF previously). This
6557 is used to reorder the STACK representing the structure to match the
6558 encoding. This will be used to get round a problem where a PKCS7
6559 structure which was signed could not be verified because the STACK
6560 order did not reflect the encoded order.
6563 *) Reimplement the OCSP ASN1 module using the new code.
6566 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6567 for its ASN1 operations. The old style function pointers still exist
6568 for now but they will eventually go away.
6571 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6572 completely replaces the old ASN1 functionality with a table driven
6573 encoder and decoder which interprets an ASN1_ITEM structure describing
6574 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6575 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6576 has also been converted to the new form.
6579 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6580 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6581 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6582 for negative moduli.
6585 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6586 of not touching the result's sign bit.
6589 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6593 *) Changed the LHASH code to use prototypes for callbacks, and created
6594 macros to declare and implement thin (optionally static) functions
6595 that provide type-safety and avoid function pointer casting for the
6596 type-specific callbacks.
6599 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6601 [Veers Staats <staatsvr@asc.hpc.mil>,
6602 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6604 *) Reformat the FAQ so the different questions and answers can be divided
6605 in sections depending on the subject.
6608 *) Have the zlib compression code load ZLIB.DLL dynamically under
6612 *) New function BN_mod_sqrt for computing square roots modulo a prime
6613 (using the probabilistic Tonelli-Shanks algorithm unless
6614 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6615 be handled deterministically).
6616 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6618 *) Make BN_mod_inverse faster by explicitly handling small quotients
6619 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6620 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6623 *) New function BN_kronecker.
6626 *) Fix BN_gcd so that it works on negative inputs; the result is
6627 positive unless both parameters are zero.
6628 Previously something reasonably close to an infinite loop was
6629 possible because numbers could be growing instead of shrinking
6630 in the implementation of Euclid's algorithm.
6633 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6634 sign of the number in question.
6636 Fix BN_is_word(a,w) to work correctly for w == 0.
6638 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6639 because its test if the absolute value of 'a' equals 'w'.
6640 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6641 it exists mostly for use in the implementations of BN_is_zero(),
6642 BN_is_one(), and BN_is_word().
6645 *) New function BN_swap.
6648 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6649 the exponentiation functions are more likely to produce reasonable
6650 results on negative inputs.
6653 *) Change BN_mod_mul so that the result is always non-negative.
6654 Previously, it could be negative if one of the factors was negative;
6655 I don't think anyone really wanted that behaviour.
6658 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6659 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6660 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6661 and add new functions:
6670 BN_mod_lshift1_quick
6674 These functions always generate non-negative results.
6676 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6677 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6679 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6680 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6681 be reduced modulo m.
6682 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6685 The following entry accidentally appeared in the CHANGES file
6686 distributed with OpenSSL 0.9.7. The modifications described in
6687 it do *not* apply to OpenSSL 0.9.7.
6689 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6690 was actually never needed) and in BN_mul(). The removal in BN_mul()
6691 required a small change in bn_mul_part_recursive() and the addition
6692 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6693 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6694 bn_sub_words() and bn_add_words() except they take arrays with
6699 *) In 'openssl passwd', verify passwords read from the terminal
6700 unless the '-salt' option is used (which usually means that
6701 verification would just waste user's time since the resulting
6702 hash is going to be compared with some given password hash)
6703 or the new '-noverify' option is used.
6705 This is an incompatible change, but it does not affect
6706 non-interactive use of 'openssl passwd' (passwords on the command
6707 line, '-stdin' option, '-in ...' option) and thus should not
6711 *) Remove all references to RSAref, since there's no more need for it.
6714 *) Make DSO load along a path given through an environment variable
6715 (SHLIB_PATH) with shl_load().
6718 *) Constify the ENGINE code as a result of BIGNUM constification.
6719 Also constify the RSA code and most things related to it. In a
6720 few places, most notable in the depth of the ASN.1 code, ugly
6721 casts back to non-const were required (to be solved at a later
6725 *) Make it so the openssl application has all engines loaded by default.
6728 *) Constify the BIGNUM routines a little more.
6731 *) Add the following functions:
6733 ENGINE_load_cswift()
6735 ENGINE_load_atalla()
6737 ENGINE_load_builtin_engines()
6739 That way, an application can itself choose if external engines that
6740 are built-in in OpenSSL shall ever be used or not. The benefit is
6741 that applications won't have to be linked with libdl or other dso
6742 libraries unless it's really needed.
6744 Changed 'openssl engine' to load all engines on demand.
6745 Changed the engine header files to avoid the duplication of some
6746 declarations (they differed!).
6749 *) 'openssl engine' can now list capabilities.
6752 *) Better error reporting in 'openssl engine'.
6755 *) Never call load_dh_param(NULL) in s_server.
6758 *) Add engine application. It can currently list engines by name and
6759 identity, and test if they are actually available.
6762 *) Improve RPM specification file by forcing symbolic linking and making
6763 sure the installed documentation is also owned by root.root.
6764 [Damien Miller <djm@mindrot.org>]
6766 *) Give the OpenSSL applications more possibilities to make use of
6767 keys (public as well as private) handled by engines.
6770 *) Add OCSP code that comes from CertCo.
6773 *) Add VMS support for the Rijndael code.
6776 *) Added untested support for Nuron crypto accelerator.
6779 *) Add support for external cryptographic devices. This code was
6780 previously distributed separately as the "engine" branch.
6781 [Geoff Thorpe, Richard Levitte]
6783 *) Rework the filename-translation in the DSO code. It is now possible to
6784 have far greater control over how a "name" is turned into a filename
6785 depending on the operating environment and any oddities about the
6786 different shared library filenames on each system.
6789 *) Support threads on FreeBSD-elf in Configure.
6792 *) Fix for SHA1 assembly problem with MASM: it produces
6793 warnings about corrupt line number information when assembling
6794 with debugging information. This is caused by the overlapping
6796 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6799 NCONF_get_number() has no error checking at all. As a replacement,
6800 NCONF_get_number_e() is defined (_e for "error checking") and is
6801 promoted strongly. The old NCONF_get_number is kept around for
6802 binary backward compatibility.
6803 Make it possible for methods to load from something other than a BIO,
6804 by providing a function pointer that is given a name instead of a BIO.
6805 For example, this could be used to load configuration data from an
6809 *) Fix for non blocking accept BIOs. Added new I/O special reason
6810 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6811 with non blocking I/O was not possible because no retry code was
6812 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6816 *) Added the beginnings of Rijndael support.
6819 *) Fix for bug in DirectoryString mask setting. Add support for
6820 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6821 to allow certificate printing to more controllable, additional
6822 'certopt' option to 'x509' to allow new printing options to be
6826 *) Clean old EAY MD5 hack from e_os.h.
6829 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6831 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6832 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6833 [Joe Orton, Steve Henson]
6835 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6837 *) Fix additional bug revealed by the NISCC test suite:
6839 Stop bug triggering large recursion when presented with
6840 certain ASN.1 tags (CVE-2003-0851)
6843 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6845 *) Fix various bugs revealed by running the NISCC test suite:
6847 Stop out of bounds reads in the ASN1 code when presented with
6848 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6850 If verify callback ignores invalid public key errors don't try to check
6851 certificate signature with the NULL public key.
6855 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6856 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6860 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6861 extra data after the compression methods not only for TLS 1.0
6862 but also for SSL 3.0 (as required by the specification).
6863 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6865 *) Change X509_certificate_type() to mark the key as exported/exportable
6866 when it's 512 *bits* long, not 512 bytes.
6869 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6871 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6872 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6873 a protocol version number mismatch like a decryption error
6874 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6877 *) Turn on RSA blinding by default in the default implementation
6878 to avoid a timing attack. Applications that don't want it can call
6879 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6880 They would be ill-advised to do so in most cases.
6881 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6883 *) Change RSA blinding code so that it works when the PRNG is not
6884 seeded (in this case, the secret RSA exponent is abused as
6885 an unpredictable seed -- if it is not unpredictable, there
6886 is no point in blinding anyway). Make RSA blinding thread-safe
6887 by remembering the creator's thread ID in rsa->blinding and
6888 having all other threads use local one-time blinding factors
6889 (this requires more computation than sharing rsa->blinding, but
6890 avoids excessive locking; and if an RSA object is not shared
6891 between threads, blinding will still be very fast).
6894 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6896 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6897 via timing by performing a MAC computation even if incorrrect
6898 block cipher padding has been found. This is a countermeasure
6899 against active attacks where the attacker has to distinguish
6900 between bad padding and a MAC verification error. (CVE-2003-0078)
6902 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6903 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6904 Martin Vuagnoux (EPFL, Ilion)]
6906 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6908 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6909 memory from it's contents. This is done with a counter that will
6910 place alternating values in each byte. This can be used to solve
6911 two issues: 1) the removal of calls to memset() by highly optimizing
6912 compilers, and 2) cleansing with other values than 0, since those can
6913 be read through on certain media, for example a swap space on disk.
6916 *) Bugfix: client side session caching did not work with external caching,
6917 because the session->cipher setting was not restored when reloading
6918 from the external cache. This problem was masked, when
6919 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6920 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6923 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6924 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6925 [Zeev Lieber <zeev-l@yahoo.com>]
6927 *) Undo an undocumented change introduced in 0.9.6e which caused
6928 repeated calls to OpenSSL_add_all_ciphers() and
6929 OpenSSL_add_all_digests() to be ignored, even after calling
6933 *) Change the default configuration reader to deal with last line not
6934 being properly terminated.
6937 *) Change X509_NAME_cmp() so it applies the special rules on handling
6938 DN values that are of type PrintableString, as well as RDNs of type
6939 emailAddress where the value has the type ia5String.
6940 [stefank@valicert.com via Richard Levitte]
6942 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6943 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6944 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6945 the bitwise-OR of the two for use by the majority of applications
6946 wanting this behaviour, and update the docs. The documented
6947 behaviour and actual behaviour were inconsistent and had been
6948 changing anyway, so this is more a bug-fix than a behavioural
6950 [Geoff Thorpe, diagnosed by Nadav Har'El]
6952 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6953 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6956 *) Fix initialization code race conditions in
6957 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6958 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6959 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6960 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6961 ssl2_get_cipher_by_char(),
6962 ssl3_get_cipher_by_char().
6963 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6965 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6966 the cached sessions are flushed, as the remove_cb() might use ex_data
6967 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6968 (see [openssl.org #212]).
6969 [Geoff Thorpe, Lutz Jaenicke]
6971 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6972 length, instead of the encoding length to d2i_ASN1_OBJECT.
6975 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6977 *) [In 0.9.6g-engine release:]
6978 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6979 [Lynn Gazis <lgazis@rainbow.com>]
6981 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6983 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6984 and get fix the header length calculation.
6985 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6986 Alon Kantor <alonk@checkpoint.com> (and others),
6989 *) Use proper error handling instead of 'assertions' in buffer
6990 overflow checks added in 0.9.6e. This prevents DoS (the
6991 assertions could call abort()).
6992 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6994 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6996 *) Add various sanity checks to asn1_get_length() to reject
6997 the ASN1 length bytes if they exceed sizeof(long), will appear
6998 negative or the content length exceeds the length of the
7000 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7002 *) Fix cipher selection routines: ciphers without encryption had no flags
7003 for the cipher strength set and where therefore not handled correctly
7004 by the selection routines (PR #130).
7007 *) Fix EVP_dsa_sha macro.
7011 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7012 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7013 that was added in OpenSSL 0.9.6d.
7015 As the countermeasure turned out to be incompatible with some
7016 broken SSL implementations, the new option is part of SSL_OP_ALL.
7017 SSL_OP_ALL is usually employed when compatibility with weird SSL
7018 implementations is desired (e.g. '-bugs' option to 's_client' and
7019 's_server'), so the new option is automatically set in many
7023 *) Changes in security patch:
7025 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7026 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7027 Air Force Materiel Command, USAF, under agreement number
7030 *) Add various sanity checks to asn1_get_length() to reject
7031 the ASN1 length bytes if they exceed sizeof(long), will appear
7032 negative or the content length exceeds the length of the
7033 supplied buffer. (CVE-2002-0659)
7034 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7036 *) Assertions for various potential buffer overflows, not known to
7038 [Ben Laurie (CHATS)]
7040 *) Various temporary buffers to hold ASCII versions of integers were
7041 too small for 64 bit platforms. (CVE-2002-0655)
7042 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7044 *) Remote buffer overflow in SSL3 protocol - an attacker could
7045 supply an oversized session ID to a client. (CVE-2002-0656)
7046 [Ben Laurie (CHATS)]
7048 *) Remote buffer overflow in SSL2 protocol - an attacker could
7049 supply an oversized client master key. (CVE-2002-0656)
7050 [Ben Laurie (CHATS)]
7052 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7054 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7055 encoded as NULL) with id-dsa-with-sha1.
7056 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7058 *) Check various X509_...() return values in apps/req.c.
7059 [Nils Larsch <nla@trustcenter.de>]
7061 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7062 an end-of-file condition would erronously be flagged, when the CRLF
7063 was just at the end of a processed block. The bug was discovered when
7064 processing data through a buffering memory BIO handing the data to a
7065 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7066 <ptsekov@syntrex.com> and Nedelcho Stanev.
7069 *) Implement a countermeasure against a vulnerability recently found
7070 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7071 before application data chunks to avoid the use of known IVs
7072 with data potentially chosen by the attacker.
7075 *) Fix length checks in ssl3_get_client_hello().
7078 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7079 to prevent ssl3_read_internal() from incorrectly assuming that
7080 ssl3_read_bytes() found application data while handshake
7081 processing was enabled when in fact s->s3->in_read_app_data was
7082 merely automatically cleared during the initial handshake.
7083 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7085 *) Fix object definitions for Private and Enterprise: they were not
7086 recognized in their shortname (=lowercase) representation. Extend
7087 obj_dat.pl to issue an error when using undefined keywords instead
7088 of silently ignoring the problem (Svenning Sorensen
7089 <sss@sss.dnsalias.net>).
7092 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7093 generators, i.e. generators other than 2 and 5. (Previously, the
7094 code did not properly initialise the 'add' and 'rem' values to
7095 BN_generate_prime().)
7097 In the new general case, we do not insist that 'generator' is
7098 actually a primitive root: This requirement is rather pointless;
7099 a generator of the order-q subgroup is just as good, if not
7103 *) Map new X509 verification errors to alerts. Discovered and submitted by
7104 Tom Wu <tom@arcot.com>.
7107 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7108 returning non-zero before the data has been completely received
7109 when using non-blocking I/O.
7110 [Bodo Moeller; problem pointed out by John Hughes]
7112 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7113 [Ben Laurie, Lutz Jaenicke]
7115 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7116 Yoram Zahavi <YoramZ@gilian.com>).
7119 *) Add information about CygWin 1.3 and on, and preserve proper
7120 configuration for the versions before that.
7121 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7123 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7124 check whether we deal with a copy of a session and do not delete from
7125 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7126 <izhar@checkpoint.com>.
7129 *) Do not store session data into the internal session cache, if it
7130 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7131 flag is set). Proposed by Aslam <aslam@funk.com>.
7134 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7138 *) [In 0.9.6d-engine release:]
7139 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7140 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7142 *) Add the configuration target linux-s390x.
7143 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7145 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7146 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7147 variable as an indication that a ClientHello message has been
7148 received. As the flag value will be lost between multiple
7149 invocations of ssl3_accept when using non-blocking I/O, the
7150 function may not be aware that a handshake has actually taken
7151 place, thus preventing a new session from being added to the
7154 To avoid this problem, we now set s->new_session to 2 instead of
7155 using a local variable.
7156 [Lutz Jaenicke, Bodo Moeller]
7158 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7159 if the SSL_R_LENGTH_MISMATCH error is detected.
7160 [Geoff Thorpe, Bodo Moeller]
7162 *) New 'shared_ldflag' column in Configure platform table.
7165 *) Fix EVP_CIPHER_mode macro.
7166 ["Dan S. Camper" <dan@bti.net>]
7168 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7169 type, we must throw them away by setting rr->length to 0.
7170 [D P Chang <dpc@qualys.com>]
7172 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7174 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7175 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7176 worked incorrectly for those cases where range = 10..._2 and
7177 3*range is two bits longer than range.)
7180 *) Only add signing time to PKCS7 structures if it is not already
7184 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7185 OBJ_ld_ce should be OBJ_id_ce.
7186 Also some ip-pda OIDs in crypto/objects/objects.txt were
7187 incorrect (cf. RFC 3039).
7188 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7190 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7191 returns early because it has nothing to do.
7192 [Andy Schneider <andy.schneider@bjss.co.uk>]
7194 *) [In 0.9.6c-engine release:]
7195 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7196 [Andy Schneider <andy.schneider@bjss.co.uk>]
7198 *) [In 0.9.6c-engine release:]
7199 Add support for Cryptographic Appliance's keyserver technology.
7200 (Use engine 'keyclient')
7201 [Cryptographic Appliances and Geoff Thorpe]
7203 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7204 is called via tools/c89.sh because arguments have to be
7205 rearranged (all '-L' options must appear before the first object
7207 [Richard Shapiro <rshapiro@abinitio.com>]
7209 *) [In 0.9.6c-engine release:]
7210 Add support for Broadcom crypto accelerator cards, backported
7212 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7214 *) [In 0.9.6c-engine release:]
7215 Add support for SureWare crypto accelerator cards from
7216 Baltimore Technologies. (Use engine 'sureware')
7217 [Baltimore Technologies and Mark Cox]
7219 *) [In 0.9.6c-engine release:]
7220 Add support for crypto accelerator cards from Accelerated
7221 Encryption Processing, www.aep.ie. (Use engine 'aep')
7222 [AEP Inc. and Mark Cox]
7224 *) Add a configuration entry for gcc on UnixWare.
7225 [Gary Benson <gbenson@redhat.com>]
7227 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7228 messages are stored in a single piece (fixed-length part and
7229 variable-length part combined) and fix various bugs found on the way.
7232 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7233 instead. BIO_gethostbyname() does not know what timeouts are
7234 appropriate, so entries would stay in cache even when they have
7236 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7238 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7239 faced with a pathologically small ClientHello fragment that does
7240 not contain client_version: Instead of aborting with an error,
7241 simply choose the highest available protocol version (i.e.,
7242 TLS 1.0 unless it is disabled). In practice, ClientHello
7243 messages are never sent like this, but this change gives us
7244 strictly correct behaviour at least for TLS.
7247 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7248 never resets s->method to s->ctx->method when called from within
7249 one of the SSL handshake functions.
7250 [Bodo Moeller; problem pointed out by Niko Baric]
7252 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7253 (sent using the client's version number) if client_version is
7254 smaller than the protocol version in use. Also change
7255 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7256 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7257 the client will at least see that alert.
7260 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7264 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7265 client receives HelloRequest while in a handshake.
7266 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7268 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7269 should end in 'break', not 'goto end' which circumvents various
7270 cleanups done in state SSL_ST_OK. But session related stuff
7271 must be disabled for SSL_ST_OK in the case that we just sent a
7274 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7275 before just sending a HelloRequest.
7276 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7278 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7279 reveal whether illegal block cipher padding was found or a MAC
7280 verification error occured. (Neither SSLerr() codes nor alerts
7281 are directly visible to potential attackers, but the information
7282 may leak via logfiles.)
7284 Similar changes are not required for the SSL 2.0 implementation
7285 because the number of padding bytes is sent in clear for SSL 2.0,
7286 and the extra bytes are just ignored. However ssl/s2_pkt.c
7287 failed to verify that the purported number of padding bytes is in
7291 *) Add OpenUNIX-8 support including shared libraries
7292 (Boyd Lynn Gerber <gerberb@zenez.com>).
7295 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7296 'wristwatch attack' using huge encoding parameters (cf.
7297 James H. Manger's CRYPTO 2001 paper). Note that the
7298 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7299 encoding parameters and hence was not vulnerable.
7302 *) BN_sqr() bug fix.
7303 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7305 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7306 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7307 followed by modular reduction.
7308 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7310 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7311 equivalent based on BN_pseudo_rand() instead of BN_rand().
7314 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7315 This function was broken, as the check for a new client hello message
7316 to handle SGC did not allow these large messages.
7317 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7320 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7323 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7324 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7327 *) Rework the configuration and shared library support for Tru64 Unix.
7328 The configuration part makes use of modern compiler features and
7329 still retains old compiler behavior for those that run older versions
7330 of the OS. The shared library support part includes a variant that
7331 uses the RPATH feature, and is available through the special
7332 configuration target "alpha-cc-rpath", which will never be selected
7334 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7336 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7337 with the same message size as in ssl3_get_certificate_request().
7338 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7339 messages might inadvertently be reject as too long.
7340 [Petr Lampa <lampa@fee.vutbr.cz>]
7342 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7345 *) Modified SSL library such that the verify_callback that has been set
7346 specificly for an SSL object with SSL_set_verify() is actually being
7347 used. Before the change, a verify_callback set with this function was
7348 ignored and the verify_callback() set in the SSL_CTX at the time of
7349 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7350 to allow the necessary settings.
7353 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7354 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7355 done automatically (in contradiction to the requirements of the C
7356 standard). This made problems when used from OpenSSH.
7359 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7360 dh->length and always used
7362 BN_rand_range(priv_key, dh->p).
7364 BN_rand_range() is not necessary for Diffie-Hellman, and this
7365 specific range makes Diffie-Hellman unnecessarily inefficient if
7366 dh->length (recommended exponent length) is much smaller than the
7367 length of dh->p. We could use BN_rand_range() if the order of
7368 the subgroup was stored in the DH structure, but we only have
7373 BN_rand(priv_key, l, ...)
7375 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7381 RSA_eay_public_encrypt
7382 RSA_eay_private_decrypt
7383 RSA_eay_private_encrypt (signing)
7384 RSA_eay_public_decrypt (signature verification)
7386 (default implementations for RSA_public_encrypt,
7387 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7388 always reject numbers >= n.
7391 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7392 to synchronize access to 'locking_thread'. This is necessary on
7393 systems where access to 'locking_thread' (an 'unsigned long'
7394 variable) is not atomic.
7397 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7398 *before* setting the 'crypto_lock_rand' flag. The previous code had
7399 a race condition if 0 is a valid thread ID.
7400 [Travis Vitek <vitek@roguewave.com>]
7402 *) Add support for shared libraries under Irix.
7403 [Albert Chin-A-Young <china@thewrittenword.com>]
7405 *) Add configuration option to build on Linux on both big-endian and
7407 [Ralf Baechle <ralf@uni-koblenz.de>]
7409 *) Add the possibility to create shared libraries on HP-UX.
7412 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7414 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7415 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7416 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7417 PRNG state recovery was possible based on the output of
7418 one PRNG request appropriately sized to gain knowledge on
7419 'md' followed by enough consecutive 1-byte PRNG requests
7420 to traverse all of 'state'.
7422 1. When updating 'md_local' (the current thread's copy of 'md')
7423 during PRNG output generation, hash all of the previous
7424 'md_local' value, not just the half used for PRNG output.
7426 2. Make the number of bytes from 'state' included into the hash
7427 independent from the number of PRNG bytes requested.
7429 The first measure alone would be sufficient to avoid
7430 Markku-Juhani's attack. (Actually it had never occurred
7431 to me that the half of 'md_local' used for chaining was the
7432 half from which PRNG output bytes were taken -- I had always
7433 assumed that the secret half would be used.) The second
7434 measure makes sure that additional data from 'state' is never
7435 mixed into 'md_local' in small portions; this heuristically
7436 further strengthens the PRNG.
7439 *) Fix crypto/bn/asm/mips3.s.
7442 *) When only the key is given to "enc", the IV is undefined. Print out
7443 an error message in this case.
7446 *) Handle special case when X509_NAME is empty in X509 printing routines.
7449 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7450 positive and less than q.
7453 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7454 used: it isn't thread safe and the add_lock_callback should handle
7456 [Paul Rose <Paul.Rose@bridge.com>]
7458 *) Verify that incoming data obeys the block size in
7459 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7463 [Ulf Möller, Bodo Möller]
7465 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7466 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7467 when fixing the server behaviour for backwards-compatible 'client
7468 hello' messages. (Note that the attack is impractical against
7469 SSL 3.0 and TLS 1.0 anyway because length and version checking
7470 means that the probability of guessing a valid ciphertext is
7471 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7474 Before 0.9.5, the countermeasure (hide the error by generating a
7475 random 'decryption result') did not work properly because
7476 ERR_clear_error() was missing, meaning that SSL_get_error() would
7477 detect the supposedly ignored error.
7479 Both problems are now fixed.
7482 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7483 (previously it was 1024).
7486 *) Fix for compatibility mode trust settings: ignore trust settings
7487 unless some valid trust or reject settings are present.
7490 *) Fix for blowfish EVP: its a variable length cipher.
7493 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7494 parameters in DSA public key structures and return an error in the
7495 DSA routines if parameters are absent.
7498 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7499 in the current directory if neither $RANDFILE nor $HOME was set.
7500 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7501 caused some confusion to Windows users who haven't defined $HOME.
7502 Thus RAND_file_name() is changed again: e_os.h can define a
7503 DEFAULT_HOME, which will be used if $HOME is not set.
7504 For Windows, we use "C:"; on other platforms, we still require
7505 environment variables.
7507 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7508 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7509 having multiple threads call RAND_poll() concurrently.
7512 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7513 combination of a flag and a thread ID variable.
7514 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7515 flag), *other* threads can enter ssleay_add_bytes without obeying
7516 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7517 that they do not hold after the first thread unsets add_do_not_lock).
7520 *) Change bctest again: '-x' expressions are not available in all
7524 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7526 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7527 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7529 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7530 the default extension for executables, if any. Also, make the perl
7531 scripts that use symlink() to test if it really exists and use "cp"
7532 if it doesn't. All this made OpenSSL compilable and installable in
7536 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7537 If SEQUENCE is length is indefinite just set c->slen to the total
7538 amount of data available.
7539 [Steve Henson, reported by shige@FreeBSD.org]
7540 [This change does not apply to 0.9.7.]
7542 *) Change bctest to avoid here-documents inside command substitution
7543 (workaround for FreeBSD /bin/sh bug).
7544 For compatibility with Ultrix, avoid shell functions (introduced
7545 in the bctest version that searches along $PATH).
7548 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7549 with des_encrypt() defined on some operating systems, like Solaris
7553 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7554 On the Importance of Eliminating Errors in Cryptographic
7555 Computations, J. Cryptology 14 (2001) 2, 101-119,
7556 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7559 *) MIPS assembler BIGNUM division bug fix.
7562 *) Disabled incorrect Alpha assembler code.
7565 *) Fix PKCS#7 decode routines so they correctly update the length
7566 after reading an EOC for the EXPLICIT tag.
7568 [This change does not apply to 0.9.7.]
7570 *) Fix bug in PKCS#12 key generation routines. This was triggered
7571 if a 3DES key was generated with a 0 initial byte. Include
7572 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7573 (but broken) behaviour.
7576 *) Enhance bctest to search for a working bc along $PATH and print
7578 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7580 *) Fix memory leaks in err.c: free err_data string if necessary;
7581 don't write to the wrong index in ERR_set_error_data.
7584 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7588 *) Replace rdtsc with _emit statements for VC++ version 5.
7589 [Jeremy Cooper <jeremy@baymoo.org>]
7591 *) Make it possible to reuse SSLv2 sessions.
7594 *) In copy_email() check for >= 0 as a return value for
7595 X509_NAME_get_index_by_NID() since 0 is a valid index.
7596 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7598 *) Avoid coredump with unsupported or invalid public keys by checking if
7599 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7600 PKCS7_verify() fails with non detached data.
7603 *) Don't use getenv in library functions when run as setuid/setgid.
7604 New function OPENSSL_issetugid().
7607 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7608 due to incorrect handling of multi-threading:
7610 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7612 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7614 3. Count how many times MemCheck_off() has been called so that
7615 nested use can be treated correctly. This also avoids
7616 inband-signalling in the previous code (which relied on the
7617 assumption that thread ID 0 is impossible).
7620 *) Add "-rand" option also to s_client and s_server.
7623 *) Fix CPU detection on Irix 6.x.
7624 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7625 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7627 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7630 [This change does not apply to 0.9.7.]
7632 *) Use the cached encoding of an X509_NAME structure rather than
7633 copying it. This is apparently the reason for the libsafe "errors"
7634 but the code is actually correct.
7637 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7638 Bleichenbacher's DSA attack.
7639 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7640 to be set and top=0 forces the highest bit to be set; top=-1 is new
7641 and leaves the highest bit random.
7642 [Ulf Moeller, Bodo Moeller]
7644 *) In the NCONF_...-based implementations for CONF_... queries
7645 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7646 a temporary CONF structure with the data component set to NULL
7647 (which gives segmentation faults in lh_retrieve).
7648 Instead, use NULL for the CONF pointer in CONF_get_string and
7649 CONF_get_number (which may use environment variables) and directly
7650 return NULL from CONF_get_section.
7653 *) Fix potential buffer overrun for EBCDIC.
7656 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7657 keyUsage if basicConstraints absent for a CA.
7660 *) Make SMIME_write_PKCS7() write mail header values with a format that
7661 is more generally accepted (no spaces before the semicolon), since
7662 some programs can't parse those values properly otherwise. Also make
7663 sure BIO's that break lines after each write do not create invalid
7667 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7668 macros previously used would not encode an empty SEQUENCE OF
7669 and break the signature.
7671 [This change does not apply to 0.9.7.]
7673 *) Zero the premaster secret after deriving the master secret in
7677 *) Add some EVP_add_digest_alias registrations (as found in
7678 OpenSSL_add_all_digests()) to SSL_library_init()
7679 aka OpenSSL_add_ssl_algorithms(). This provides improved
7680 compatibility with peers using X.509 certificates
7681 with unconventional AlgorithmIdentifier OIDs.
7684 *) Fix for Irix with NO_ASM.
7685 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7687 *) ./config script fixes.
7688 [Ulf Moeller, Richard Levitte]
7690 *) Fix 'openssl passwd -1'.
7693 *) Change PKCS12_key_gen_asc() so it can cope with non null
7694 terminated strings whose length is passed in the passlen
7695 parameter, for example from PEM callbacks. This was done
7696 by adding an extra length parameter to asc2uni().
7697 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7699 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7700 call failed, free the DSA structure.
7703 *) Fix to uni2asc() to cope with zero length Unicode strings.
7704 These are present in some PKCS#12 files.
7707 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7708 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7709 when writing a 32767 byte record.
7710 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7712 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7713 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7715 (RSA objects have a reference count access to which is protected
7716 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7717 so they are meant to be shared between threads.)
7718 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7719 "Reddie, Steven" <Steven.Reddie@ca.com>]
7721 *) Fix a deadlock in CRYPTO_mem_leaks().
7724 *) Use better test patterns in bntest.
7727 *) rand_win.c fix for Borland C.
7730 *) BN_rshift bugfix for n == 0.
7733 *) Add a 'bctest' script that checks for some known 'bc' bugs
7734 so that 'make test' does not abort just because 'bc' is broken.
7737 *) Store verify_result within SSL_SESSION also for client side to
7738 avoid potential security hole. (Re-used sessions on the client side
7739 always resulted in verify_result==X509_V_OK, not using the original
7740 result of the server certificate verification.)
7743 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7744 SSL3_RT_APPLICATION_DATA, return 0.
7745 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7749 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7750 releases, have been re-implemented by renaming the previous
7751 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7752 and ssl3_read_internal, respectively, and adding 'peek' parameters
7753 to them. The new ssl[23]_{read,peek} functions are calls to
7754 ssl[23]_read_internal with the 'peek' flag set appropriately.
7755 A 'peek' parameter has also been added to ssl3_read_bytes, which
7756 does the actual work for ssl3_read_internal.
7759 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7760 the method-specific "init()" handler. Also clean up ex_data after
7761 calling the method-specific "finish()" handler. Previously, this was
7762 happening the other way round.
7765 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7766 The previous value, 12, was not always sufficient for BN_mod_exp().
7769 *) Make sure that shared libraries get the internal name engine with
7770 the full version number and not just 0. This should mark the
7771 shared libraries as not backward compatible. Of course, this should
7772 be changed again when we can guarantee backward binary compatibility.
7775 *) Fix typo in get_cert_by_subject() in by_dir.c
7776 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7778 *) Rework the system to generate shared libraries:
7780 - Make note of the expected extension for the shared libraries and
7781 if there is a need for symbolic links from for example libcrypto.so.0
7782 to libcrypto.so.0.9.7. There is extended info in Configure for
7785 - Make as few rebuilds of the shared libraries as possible.
7787 - Still avoid linking the OpenSSL programs with the shared libraries.
7789 - When installing, install the shared libraries separately from the
7793 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7795 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7796 and not in SSL_clear because the latter is also used by the
7797 accept/connect functions; previously, the settings made by
7798 SSL_set_read_ahead would be lost during the handshake.
7799 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7801 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7802 Previously, it would create entries for disabled algorithms no
7806 *) Added several new manual pages for SSL_* function.
7809 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7811 *) In ssl23_get_client_hello, generate an error message when faced
7812 with an initial SSL 3.0/TLS record that is too small to contain the
7813 first two bytes of the ClientHello message, i.e. client_version.
7814 (Note that this is a pathologic case that probably has never happened
7815 in real life.) The previous approach was to use the version number
7816 from the record header as a substitute; but our protocol choice
7817 should not depend on that one because it is not authenticated
7818 by the Finished messages.
7821 *) More robust randomness gathering functions for Windows.
7822 [Jeffrey Altman <jaltman@columbia.edu>]
7824 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7825 not set then we don't setup the error code for issuer check errors
7826 to avoid possibly overwriting other errors which the callback does
7827 handle. If an application does set the flag then we assume it knows
7828 what it is doing and can handle the new informational codes
7832 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7833 a general "ANY" type, as such it should be able to decode anything
7834 including tagged types. However it didn't check the class so it would
7835 wrongly interpret tagged types in the same way as their universal
7836 counterpart and unknown types were just rejected. Changed so that the
7837 tagged and unknown types are handled in the same way as a SEQUENCE:
7838 that is the encoding is stored intact. There is also a new type
7839 "V_ASN1_OTHER" which is used when the class is not universal, in this
7840 case we have no idea what the actual type is so we just lump them all
7844 *) On VMS, stdout may very well lead to a file that is written to
7845 in a record-oriented fashion. That means that every write() will
7846 write a separate record, which will be read separately by the
7847 programs trying to read from it. This can be very confusing.
7849 The solution is to put a BIO filter in the way that will buffer
7850 text until a linefeed is reached, and then write everything a
7851 line at a time, so every record written will be an actual line,
7852 not chunks of lines and not (usually doesn't happen, but I've
7853 seen it once) several lines in one record. BIO_f_linebuffer() is
7856 Currently, it's a VMS-only method, because that's where it has
7857 been tested well enough.
7860 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7861 it can return incorrect results.
7862 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7863 but it was in 0.9.6-beta[12].)
7866 *) Disable the check for content being present when verifying detached
7867 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7868 include zero length content when signing messages.
7871 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7872 BIO_ctrl (for BIO pairs).
7875 *) Add DSO method for VMS.
7878 *) Bug fix: Montgomery multiplication could produce results with the
7882 *) Add RPM specification openssl.spec and modify it to build three
7883 packages. The default package contains applications, application
7884 documentation and run-time libraries. The devel package contains
7885 include files, static libraries and function documentation. The
7886 doc package contains the contents of the doc directory. The original
7887 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7890 *) Add a large number of documentation files for many SSL routines.
7891 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7893 *) Add a configuration entry for Sony News 4.
7894 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7896 *) Don't set the two most significant bits to one when generating a
7897 random number < q in the DSA library.
7900 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7901 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7902 the underlying transport is blocking) if a handshake took place.
7903 (The default behaviour is needed by applications such as s_client
7904 and s_server that use select() to determine when to use SSL_read;
7905 but for applications that know in advance when to expect data, it
7906 just makes things more complicated.)
7909 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7913 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7914 work better on such systems.
7915 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7917 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7918 Update PKCS12_parse() so it copies the friendlyName and the
7919 keyid to the certificates aux info.
7922 *) Fix bug in PKCS7_verify() which caused an infinite loop
7923 if there was more than one signature.
7924 [Sven Uszpelkat <su@celocom.de>]
7926 *) Major change in util/mkdef.pl to include extra information
7927 about each symbol, as well as presenting variables as well
7928 as functions. This change means that there's n more need
7929 to rebuild the .num files when some algorithms are excluded.
7932 *) Allow the verify time to be set by an application,
7933 rather than always using the current time.
7936 *) Phase 2 verify code reorganisation. The certificate
7937 verify code now looks up an issuer certificate by a
7938 number of criteria: subject name, authority key id
7939 and key usage. It also verifies self signed certificates
7940 by the same criteria. The main comparison function is
7941 X509_check_issued() which performs these checks.
7943 Lot of changes were necessary in order to support this
7944 without completely rewriting the lookup code.
7946 Authority and subject key identifier are now cached.
7948 The LHASH 'certs' is X509_STORE has now been replaced
7949 by a STACK_OF(X509_OBJECT). This is mainly because an
7950 LHASH can't store or retrieve multiple objects with
7951 the same hash value.
7953 As a result various functions (which were all internal
7954 use only) have changed to handle the new X509_STORE
7955 structure. This will break anything that messed round
7956 with X509_STORE internally.
7958 The functions X509_STORE_add_cert() now checks for an
7959 exact match, rather than just subject name.
7961 The X509_STORE API doesn't directly support the retrieval
7962 of multiple certificates matching a given criteria, however
7963 this can be worked round by performing a lookup first
7964 (which will fill the cache with candidate certificates)
7965 and then examining the cache for matches. This is probably
7966 the best we can do without throwing out X509_LOOKUP
7967 entirely (maybe later...).
7969 The X509_VERIFY_CTX structure has been enhanced considerably.
7971 All certificate lookup operations now go via a get_issuer()
7972 callback. Although this currently uses an X509_STORE it
7973 can be replaced by custom lookups. This is a simple way
7974 to bypass the X509_STORE hackery necessary to make this
7975 work and makes it possible to use more efficient techniques
7976 in future. A very simple version which uses a simple
7977 STACK for its trusted certificate store is also provided
7978 using X509_STORE_CTX_trusted_stack().
7980 The verify_cb() and verify() callbacks now have equivalents
7981 in the X509_STORE_CTX structure.
7983 X509_STORE_CTX also has a 'flags' field which can be used
7984 to customise the verify behaviour.
7987 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7988 excludes S/MIME capabilities.
7991 *) When a certificate request is read in keep a copy of the
7992 original encoding of the signed data and use it when outputing
7993 again. Signatures then use the original encoding rather than
7994 a decoded, encoded version which may cause problems if the
7995 request is improperly encoded.
7998 *) For consistency with other BIO_puts implementations, call
7999 buffer_write(b, ...) directly in buffer_puts instead of calling
8002 In BIO_puts, increment b->num_write as in BIO_write.
8003 [Peter.Sylvester@EdelWeb.fr]
8005 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8006 BN_zero, we may not return a BIGNUM with an array consisting of
8010 *) Avoid calling abort() from within the library when problems are
8011 detected, except if preprocessor symbols have been defined
8012 (such as REF_CHECK, BN_DEBUG etc.).
8015 *) New openssl application 'rsautl'. This utility can be
8016 used for low level RSA operations. DER public key
8017 BIO/fp routines also added.
8020 *) New Configure entry and patches for compiling on QNX 4.
8021 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8023 *) A demo state-machine implementation was sponsored by
8024 Nuron (http://www.nuron.com/) and is now available in
8025 demos/state_machine.
8028 *) New options added to the 'dgst' utility for signature
8029 generation and verification.
8032 *) Unrecognized PKCS#7 content types are now handled via a
8033 catch all ASN1_TYPE structure. This allows unsupported
8034 types to be stored as a "blob" and an application can
8035 encode and decode it manually.
8038 *) Fix various signed/unsigned issues to make a_strex.c
8040 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8042 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8043 length if passed a buffer. ASN1_INTEGER_to_BN failed
8044 if passed a NULL BN and its argument was negative.
8045 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8047 *) Modification to PKCS#7 encoding routines to output definite
8048 length encoding. Since currently the whole structures are in
8049 memory there's not real point in using indefinite length
8050 constructed encoding. However if OpenSSL is compiled with
8051 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8054 *) Added BIO_vprintf() and BIO_vsnprintf().
8057 *) Added more prefixes to parse for in the the strings written
8058 through a logging bio, to cover all the levels that are available
8059 through syslog. The prefixes are now:
8061 PANIC, EMERG, EMR => LOG_EMERG
8062 ALERT, ALR => LOG_ALERT
8063 CRIT, CRI => LOG_CRIT
8064 ERROR, ERR => LOG_ERR
8065 WARNING, WARN, WAR => LOG_WARNING
8066 NOTICE, NOTE, NOT => LOG_NOTICE
8067 INFO, INF => LOG_INFO
8068 DEBUG, DBG => LOG_DEBUG
8070 and as before, if none of those prefixes are present at the
8071 beginning of the string, LOG_ERR is chosen.
8073 On Win32, the LOG_* levels are mapped according to this:
8075 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8076 LOG_WARNING => EVENTLOG_WARNING_TYPE
8077 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8081 *) Made it possible to reconfigure with just the configuration
8082 argument "reconf" or "reconfigure". The command line arguments
8083 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8084 and are retrieved from there when reconfiguring.
8088 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8090 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8093 *) The obj_dat.pl script was messing up the sorting of object
8094 names. The reason was that it compared the quoted version
8095 of strings as a result "OCSP" > "OCSP Signing" because
8096 " > SPACE. Changed script to store unquoted versions of
8097 names and add quotes on output. It was also omitting some
8098 names from the lookup table if they were given a default
8099 value (that is if SN is missing it is given the same
8100 value as LN and vice versa), these are now added on the
8101 grounds that if an object has a name we should be able to
8102 look it up. Finally added warning output when duplicate
8103 short or long names are found.
8106 *) Changes needed for Tandem NSK.
8107 [Scott Uroff <scott@xypro.com>]
8109 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8110 RSA_padding_check_SSLv23(), special padding was never detected
8111 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8112 version rollback attacks was not effective.
8114 In s23_clnt.c, don't use special rollback-attack detection padding
8115 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8116 client; similarly, in s23_srvr.c, don't do the rollback check if
8117 SSL 2.0 is the only protocol enabled in the server.
8120 *) Make it possible to get hexdumps of unprintable data with 'openssl
8121 asn1parse'. By implication, the functions ASN1_parse_dump() and
8122 BIO_dump_indent() are added.
8125 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8126 these print out strings and name structures based on various
8127 flags including RFC2253 support and proper handling of
8128 multibyte characters. Added options to the 'x509' utility
8129 to allow the various flags to be set.
8132 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8133 Also change the functions X509_cmp_current_time() and
8134 X509_gmtime_adj() work with an ASN1_TIME structure,
8135 this will enable certificates using GeneralizedTime in validity
8136 dates to be checked.
8139 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8140 negative public key encodings) on by default,
8141 NO_NEG_PUBKEY_BUG can be set to disable it.
8144 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8145 content octets. An i2c_ASN1_OBJECT is unnecessary because
8146 the encoding can be trivially obtained from the structure.
8149 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8150 not read locks (CRYPTO_r_[un]lock).
8153 *) A first attempt at creating official support for shared
8154 libraries through configuration. I've kept it so the
8155 default is static libraries only, and the OpenSSL programs
8156 are always statically linked for now, but there are
8157 preparations for dynamic linking in place.
8158 This has been tested on Linux and Tru64.
8161 *) Randomness polling function for Win9x, as described in:
8162 Peter Gutmann, Software Generation of Practically Strong
8166 *) Fix so PRNG is seeded in req if using an already existing
8170 *) New options to smime application. -inform and -outform
8171 allow alternative formats for the S/MIME message including
8172 PEM and DER. The -content option allows the content to be
8173 specified separately. This should allow things like Netscape
8174 form signing output easier to verify.
8177 *) Fix the ASN1 encoding of tags using the 'long form'.
8180 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8181 STRING types. These convert content octets to and from the
8182 underlying type. The actual tag and length octets are
8183 already assumed to have been read in and checked. These
8184 are needed because all other string types have virtually
8185 identical handling apart from the tag. By having versions
8186 of the ASN1 functions that just operate on content octets
8187 IMPLICIT tagging can be handled properly. It also allows
8188 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8189 and ASN1_INTEGER are identical apart from the tag.
8192 *) Change the handling of OID objects as follows:
8194 - New object identifiers are inserted in objects.txt, following
8195 the syntax given in objects.README.
8196 - objects.pl is used to process obj_mac.num and create a new
8198 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8201 This is currently kind of a hack, and the perl code in objects.pl
8202 isn't very elegant, but it works as I intended. The simplest way
8203 to check that it worked correctly is to look in obj_dat.h and
8204 check the array nid_objs and make sure the objects haven't moved
8205 around (this is important!). Additions are OK, as well as
8206 consistent name changes.
8209 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8212 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8213 The given file adds to whatever has already been seeded into the
8214 random pool through the RANDFILE configuration file option or
8215 environment variable, or the default random state file.
8218 *) mkstack.pl now sorts each macro group into lexical order.
8219 Previously the output order depended on the order the files
8220 appeared in the directory, resulting in needless rewriting
8224 *) Patches to make OpenSSL compile under Win32 again. Mostly
8225 work arounds for the VC++ problem that it treats func() as
8226 func(void). Also stripped out the parts of mkdef.pl that
8227 added extra typesafe functions: these no longer exist.
8230 *) Reorganisation of the stack code. The macros are now all
8231 collected in safestack.h . Each macro is defined in terms of
8232 a "stack macro" of the form SKM_<name>(type, a, b). The
8233 DEBUG_SAFESTACK is now handled in terms of function casts,
8234 this has the advantage of retaining type safety without the
8235 use of additional functions. If DEBUG_SAFESTACK is not defined
8236 then the non typesafe macros are used instead. Also modified the
8237 mkstack.pl script to handle the new form. Needs testing to see
8238 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8239 the default if no major problems. Similar behaviour for ASN1_SET_OF
8240 and PKCS12_STACK_OF.
8243 *) When some versions of IIS use the 'NET' form of private key the
8244 key derivation algorithm is different. Normally MD5(password) is
8245 used as a 128 bit RC4 key. In the modified case
8246 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8247 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8248 as the old Netscape_RSA functions except they have an additional
8249 'sgckey' parameter which uses the modified algorithm. Also added
8250 an -sgckey command line option to the rsa utility. Thanks to
8251 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8252 algorithm to openssl-dev.
8255 *) The evp_local.h macros were using 'c.##kname' which resulted in
8256 invalid expansion on some systems (SCO 5.0.5 for example).
8257 Corrected to 'c.kname'.
8258 [Phillip Porch <root@theporch.com>]
8260 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8261 a STACK of email addresses from a certificate or request, these look
8262 in the subject name and the subject alternative name extensions and
8263 omit any duplicate addresses.
8266 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8267 This makes DSA verification about 2 % faster.
8270 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8271 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8272 plus overhead for 1024 bit moduli).
8273 This makes exponentiations about 0.5 % faster for 1024 bit
8274 exponents (as measured by "openssl speed rsa2048").
8277 *) Rename memory handling macros to avoid conflicts with other
8279 Malloc => OPENSSL_malloc
8280 Malloc_locked => OPENSSL_malloc_locked
8281 Realloc => OPENSSL_realloc
8282 Free => OPENSSL_free
8285 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8286 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8289 *) CygWin32 support.
8290 [John Jarvie <jjarvie@newsguy.com>]
8292 *) The type-safe stack code has been rejigged. It is now only compiled
8293 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8294 by default all type-specific stack functions are "#define"d back to
8295 standard stack functions. This results in more streamlined output
8296 but retains the type-safety checking possibilities of the original
8300 *) The STACK code has been cleaned up, and certain type declarations
8301 that didn't make a lot of sense have been brought in line. This has
8302 also involved a cleanup of sorts in safestack.h to more correctly
8303 map type-safe stack functions onto their plain stack counterparts.
8304 This work has also resulted in a variety of "const"ifications of
8305 lots of the code, especially "_cmp" operations which should normally
8306 be prototyped with "const" parameters anyway.
8309 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8310 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8311 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8312 where all of 'md' is used each time the PRNG is used, but 'state'
8313 is used only indexed by a cyclic counter. As entropy may not be
8314 well distributed from the beginning, 'md' is important as a
8315 chaining variable. However, the output function chains only half
8316 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8317 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8318 in all of 'state' being rewritten, with the new values depending
8319 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8322 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8323 the handshake is continued after ssl_verify_cert_chain();
8324 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8325 can lead to 'unexplainable' connection aborts later.
8326 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8328 *) Major EVP API cipher revision.
8329 Add hooks for extra EVP features. This allows various cipher
8330 parameters to be set in the EVP interface. Support added for variable
8331 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8332 setting of RC2 and RC5 parameters.
8334 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8337 Remove lots of duplicated code from the EVP library. For example *every*
8338 cipher init() function handles the 'iv' in the same way according to the
8339 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8340 for CFB and OFB modes they zero ctx->num.
8342 New functionality allows removal of S/MIME code RC2 hack.
8344 Most of the routines have the same form and so can be declared in terms
8347 By shifting this to the top level EVP_CipherInit() it can be removed from
8348 all individual ciphers. If the cipher wants to handle IVs or keys
8349 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8352 Change lots of functions like EVP_EncryptUpdate() to now return a
8353 value: although software versions of the algorithms cannot fail
8354 any installed hardware versions can.
8357 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8358 this option is set, tolerate broken clients that send the negotiated
8359 protocol version number instead of the requested protocol version
8363 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8364 i.e. non-zero for export ciphersuites, zero otherwise.
8365 Previous versions had this flag inverted, inconsistent with
8366 rsa_tmp_cb (..._TMP_RSA_CB).
8367 [Bodo Moeller; problem reported by Amit Chopra]
8369 *) Add missing DSA library text string. Work around for some IIS
8370 key files with invalid SEQUENCE encoding.
8373 *) Add a document (doc/standards.txt) that list all kinds of standards
8374 and so on that are implemented in OpenSSL.
8377 *) Enhance c_rehash script. Old version would mishandle certificates
8378 with the same subject name hash and wouldn't handle CRLs at all.
8379 Added -fingerprint option to crl utility, to support new c_rehash
8383 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8386 *) Fix for SSL server purpose checking. Server checking was
8387 rejecting certificates which had extended key usage present
8388 but no ssl client purpose.
8389 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8391 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8392 is a little unclear about how a blank password is handled.
8393 Since the password in encoded as a BMPString with terminating
8394 double NULL a zero length password would end up as just the
8395 double NULL. However no password at all is different and is
8396 handled differently in the PKCS#12 key generation code. NS
8397 treats a blank password as zero length. MSIE treats it as no
8398 password on export: but it will try both on import. We now do
8399 the same: PKCS12_parse() tries zero length and no password if
8400 the password is set to "" or NULL (NULL is now a valid password:
8401 it wasn't before) as does the pkcs12 application.
8404 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8405 perror when PEM_read_bio_X509_REQ fails, the error message must
8406 be obtained from the error queue.
8409 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8410 it in ERR_remove_state if appropriate, and change ERR_get_state
8411 accordingly to avoid race conditions (this is necessary because
8412 thread_hash is no longer constant once set).
8415 *) Bugfix for linux-elf makefile.one.
8418 *) RSA_get_default_method() will now cause a default
8419 RSA_METHOD to be chosen if one doesn't exist already.
8420 Previously this was only set during a call to RSA_new()
8421 or RSA_new_method(NULL) meaning it was possible for
8422 RSA_get_default_method() to return NULL.
8425 *) Added native name translation to the existing DSO code
8426 that will convert (if the flag to do so is set) filenames
8427 that are sufficiently small and have no path information
8428 into a canonical native form. Eg. "blah" converted to
8429 "libblah.so" or "blah.dll" etc.
8432 *) New function ERR_error_string_n(e, buf, len) which is like
8433 ERR_error_string(e, buf), but writes at most 'len' bytes
8434 including the 0 terminator. For ERR_error_string_n, 'buf'
8436 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8438 *) CONF library reworked to become more general. A new CONF
8439 configuration file reader "class" is implemented as well as a
8440 new functions (NCONF_*, for "New CONF") to handle it. The now
8441 old CONF_* functions are still there, but are reimplemented to
8442 work in terms of the new functions. Also, a set of functions
8443 to handle the internal storage of the configuration data is
8444 provided to make it easier to write new configuration file
8445 reader "classes" (I can definitely see something reading a
8446 configuration file in XML format, for example), called _CONF_*,
8447 or "the configuration storage API"...
8449 The new configuration file reading functions are:
8451 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8452 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8454 NCONF_default, NCONF_WIN32
8456 NCONF_dump_fp, NCONF_dump_bio
8458 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8459 NCONF_new creates a new CONF object. This works in the same way
8460 as other interfaces in OpenSSL, like the BIO interface.
8461 NCONF_dump_* dump the internal storage of the configuration file,
8462 which is useful for debugging. All other functions take the same
8463 arguments as the old CONF_* functions wth the exception of the
8464 first that must be a `CONF *' instead of a `LHASH *'.
8466 To make it easer to use the new classes with the old CONF_* functions,
8467 the function CONF_set_default_method is provided.
8470 *) Add '-tls1' option to 'openssl ciphers', which was already
8471 mentioned in the documentation but had not been implemented.
8472 (This option is not yet really useful because even the additional
8473 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8476 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8477 OpenSSL-based applications) load shared libraries and bind to
8478 them in a portable way.
8479 [Geoff Thorpe, with contributions from Richard Levitte]
8481 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8483 *) Make sure _lrotl and _lrotr are only used with MSVC.
8485 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8486 (the default implementation of RAND_status).
8488 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8489 to '-clrext' (= clear extensions), as intended and documented.
8490 [Bodo Moeller; inconsistency pointed out by Michael Attili
8493 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8494 was larger than the MD block size.
8495 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8497 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8498 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8499 using the passed key: if the passed key was a private key the result
8500 of X509_print(), for example, would be to print out all the private key
8504 *) des_quad_cksum() byte order bug fix.
8505 [Ulf Möller, using the problem description in krb4-0.9.7, where
8506 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8508 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8510 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8512 *) For easily testing in shell scripts whether some command
8513 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8514 returns with exit code 0 iff no command of the given name is available.
8515 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8516 the output goes to stdout and nothing is printed to stderr.
8517 Additional arguments are always ignored.
8519 Since for each cipher there is a command of the same name,
8520 the 'no-cipher' compilation switches can be tested this way.
8522 ('openssl no-XXX' is not able to detect pseudo-commands such
8523 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8526 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8529 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8530 is set; it will be thrown away anyway because each handshake creates
8532 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8533 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8534 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8535 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8538 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8539 'Q' and 'R' lose their special meanings (quit/renegotiate).
8540 This is part of what -quiet does; unlike -quiet, -ign_eof
8541 does not suppress any output.
8544 *) Add compatibility options to the purpose and trust code. The
8545 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8546 accepts a certificate or CA, this was the previous behaviour,
8547 with all the associated security issues.
8549 X509_TRUST_COMPAT is the old trust behaviour: only and
8550 automatically trust self signed roots in certificate store. A
8551 new trust setting X509_TRUST_DEFAULT is used to specify that
8552 a purpose has no associated trust setting and it should instead
8553 use the value in the default purpose.
8556 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8557 and fix a memory leak.
8560 *) In util/mkerr.pl (which implements 'make errors'), preserve
8561 reason strings from the previous version of the .c file, as
8562 the default to have only downcase letters (and digits) in
8563 automatically generated reasons codes is not always appropriate.
8566 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8567 using strerror. Previously, ERR_reason_error_string() returned
8568 library names as reason strings for SYSerr; but SYSerr is a special
8569 case where small numbers are errno values, not library numbers.
8572 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8573 converts DSA parameters into DH parameters. (When creating parameters,
8574 DSA_generate_parameters is used.)
8577 *) Include 'length' (recommended exponent length) in C code generated
8578 by 'openssl dhparam -C'.
8581 *) The second argument to set_label in perlasm was already being used
8582 so couldn't be used as a "file scope" flag. Moved to third argument
8586 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8587 instead of RAND_bytes for encryption IVs and salts.
8590 *) Include RAND_status() into RAND_METHOD instead of implementing
8591 it only for md_rand.c Otherwise replacing the PRNG by calling
8592 RAND_set_rand_method would be impossible.
8595 *) Don't let DSA_generate_key() enter an infinite loop if the random
8596 number generation fails.
8599 *) New 'rand' application for creating pseudo-random output.
8602 *) Added configuration support for Linux/IA64
8603 [Rolf Haberrecker <rolf@suse.de>]
8605 *) Assembler module support for Mingw32.
8608 *) Shared library support for HPUX (in shlib/).
8609 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8611 *) Shared library support for Solaris gcc.
8612 [Lutz Behnke <behnke@trustcenter.de>]
8614 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8616 *) PKCS7_encrypt() was adding text MIME headers twice because they
8617 were added manually and by SMIME_crlf_copy().
8620 *) In bntest.c don't call BN_rand with zero bits argument.
8621 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8623 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8624 case was implemented. This caused BN_div_recp() to fail occasionally.
8627 *) Add an optional second argument to the set_label() in the perl
8628 assembly language builder. If this argument exists and is set
8629 to 1 it signals that the assembler should use a symbol whose
8630 scope is the entire file, not just the current function. This
8631 is needed with MASM which uses the format label:: for this scope.
8632 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8634 *) Change the ASN1 types so they are typedefs by default. Before
8635 almost all types were #define'd to ASN1_STRING which was causing
8636 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8640 *) Change names of new functions to the new get1/get0 naming
8641 convention: After 'get1', the caller owns a reference count
8642 and has to call ..._free; 'get0' returns a pointer to some
8643 data structure without incrementing reference counters.
8644 (Some of the existing 'get' functions increment a reference
8645 counter, some don't.)
8646 Similarly, 'set1' and 'add1' functions increase reference
8647 counters or duplicate objects.
8650 *) Allow for the possibility of temp RSA key generation failure:
8651 the code used to assume it always worked and crashed on failure.
8654 *) Fix potential buffer overrun problem in BIO_printf().
8655 [Ulf Möller, using public domain code by Patrick Powell; problem
8656 pointed out by David Sacerdote <das33@cornell.edu>]
8658 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8659 RAND_egd() and RAND_status(). In the command line application,
8660 the EGD socket can be specified like a seed file using RANDFILE
8664 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8665 Some CAs (e.g. Verisign) distribute certificates in this form.
8668 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8669 list to exclude them. This means that no special compilation option
8670 is needed to use anonymous DH: it just needs to be included in the
8674 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8675 EVP_MD_type. The old functionality is available in a new macro called
8676 EVP_MD_md(). Change code that uses it and update docs.
8679 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8680 where the 'void *' argument is replaced by a function pointer argument.
8681 Previously 'void *' was abused to point to functions, which works on
8682 many platforms, but is not correct. As these functions are usually
8683 called by macros defined in OpenSSL header files, most source code
8684 should work without changes.
8687 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8688 sections with information on -D... compiler switches used for
8689 compiling the library so that applications can see them. To enable
8690 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8691 must be defined. E.g.,
8692 #define OPENSSL_ALGORITHM_DEFINES
8693 #include <openssl/opensslconf.h>
8694 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8695 [Richard Levitte, Ulf and Bodo Möller]
8697 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8701 *) Change the 'other' type in certificate aux info to a STACK_OF
8702 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8703 the required ASN1 format: arbitrary types determined by an OID.
8706 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8707 argument to 'req'. This is not because the function is newer or
8708 better than others it just uses the work 'NEW' in the certificate
8709 request header lines. Some software needs this.
8712 *) Reorganise password command line arguments: now passwords can be
8713 obtained from various sources. Delete the PEM_cb function and make
8714 it the default behaviour: i.e. if the callback is NULL and the
8715 usrdata argument is not NULL interpret it as a null terminated pass
8716 phrase. If usrdata and the callback are NULL then the pass phrase
8717 is prompted for as usual.
8720 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8721 the support is automatically enabled. The resulting binaries will
8722 autodetect the card and use it if present.
8723 [Ben Laurie and Compaq Inc.]
8725 *) Work around for Netscape hang bug. This sends certificate request
8726 and server done in one record. Since this is perfectly legal in the
8727 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8728 the bugs/SSLv3 entry for more info.
8731 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8734 *) Add -rand argument to smime and pkcs12 applications and read/write
8738 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8741 *) Add command line password options to the remaining applications.
8744 *) Bug fix for BN_div_recp() for numerators with an even number of
8748 *) More tests in bntest.c, and changed test_bn output.
8751 *) ./config recognizes MacOS X now.
8754 *) Bug fix for BN_div() when the first words of num and divsor are
8755 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8758 *) Add support for various broken PKCS#8 formats, and command line
8759 options to produce them.
8762 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8763 get temporary BIGNUMs from a BN_CTX.
8766 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8770 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8771 include a #define from the old name to the new. The original intent
8772 was that statically linked binaries could for example just call
8773 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8774 link with digests. This never worked because SSLeay_add_all_digests()
8775 and SSLeay_add_all_ciphers() were in the same source file so calling
8776 one would link with the other. They are now in separate source files.
8779 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8782 *) Use a less unusual form of the Miller-Rabin primality test (it used
8783 a binary algorithm for exponentiation integrated into the Miller-Rabin
8784 loop, our standard modexp algorithms are faster).
8787 *) Support for the EBCDIC character set completed.
8788 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8790 *) Source code cleanups: use const where appropriate, eliminate casts,
8791 use void * instead of char * in lhash.
8794 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8795 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8796 this the server could overwrite ephemeral keys that the client
8800 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8801 using 50 iterations of the Rabin-Miller test.
8803 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8804 iterations of the Rabin-Miller test as required by the appendix
8805 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8806 As BN_is_prime_fasttest includes trial division, DSA parameter
8807 generation becomes much faster.
8809 This implies a change for the callback functions in DSA_is_prime
8810 and DSA_generate_parameters: The callback function is called once
8811 for each positive witness in the Rabin-Miller test, not just
8812 occasionally in the inner loop; and the parameters to the
8813 callback function now provide an iteration count for the outer
8814 loop rather than for the current invocation of the inner loop.
8815 DSA_generate_parameters additionally can call the callback
8816 function with an 'iteration count' of -1, meaning that a
8817 candidate has passed the trial division test (when q is generated
8818 from an application-provided seed, trial division is skipped).
8821 *) New function BN_is_prime_fasttest that optionally does trial
8822 division before starting the Rabin-Miller test and has
8823 an additional BN_CTX * argument (whereas BN_is_prime always
8824 has to allocate at least one BN_CTX).
8825 'callback(1, -1, cb_arg)' is called when a number has passed the
8826 trial division stage.
8829 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8833 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8836 *) New function BN_pseudo_rand().
8839 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8840 bignum version of BN_from_montgomery() with the working code from
8841 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8845 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8846 made it impossible to use the same SSL_SESSION data structure in
8847 SSL2 clients in multiple threads.
8850 *) The return value of RAND_load_file() no longer counts bytes obtained
8851 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8852 to seed the PRNG (previously an explicit byte count was required).
8853 [Ulf Möller, Bodo Möller]
8855 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8856 used (char *) instead of (void *) and had casts all over the place.
8859 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8862 *) Retain source code compatibility for BN_prime_checks macro:
8863 BN_is_prime(..., BN_prime_checks, ...) now uses
8864 BN_prime_checks_for_size to determine the appropriate number of
8865 Rabin-Miller iterations.
8868 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8869 DH_CHECK_P_NOT_SAFE_PRIME.
8870 (Check if this is true? OpenPGP calls them "strong".)
8873 *) Merge the functionality of "dh" and "gendh" programs into a new program
8874 "dhparam". The old programs are retained for now but will handle DH keys
8875 (instead of parameters) in future.
8878 *) Make the ciphers, s_server and s_client programs check the return values
8879 when a new cipher list is set.
8882 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8883 ciphers. Before when the 56bit ciphers were enabled the sorting was
8886 The syntax for the cipher sorting has been extended to support sorting by
8887 cipher-strength (using the strength_bits hard coded in the tables).
8888 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8890 Fix a bug in the cipher-command parser: when supplying a cipher command
8891 string with an "undefined" symbol (neither command nor alphanumeric
8892 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8893 an error is flagged.
8895 Due to the strength-sorting extension, the code of the
8896 ssl_create_cipher_list() function was completely rearranged. I hope that
8897 the readability was also increased :-)
8898 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8900 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8901 for the first serial number and places 2 in the serial number file. This
8902 avoids problems when the root CA is created with serial number zero and
8903 the first user certificate has the same issuer name and serial number
8907 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8908 the new code. Add documentation for this stuff.
8911 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8912 X509_*() to X509at_*() on the grounds that they don't handle X509
8913 structures and behave in an analogous way to the X509v3 functions:
8914 they shouldn't be called directly but wrapper functions should be used
8917 So we also now have some wrapper functions that call the X509at functions
8918 when passed certificate requests. (TO DO: similar things can be done with
8919 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8920 things. Some of these need some d2i or i2d and print functionality
8921 because they handle more complex structures.)
8924 *) Add missing #ifndefs that caused missing symbols when building libssl
8925 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8926 NO_RSA in ssl/s2*.c.
8927 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8929 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8930 has a return value which indicates the quality of the random data
8931 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8932 error queue. New function RAND_pseudo_bytes() generates output that is
8933 guaranteed to be unique but not unpredictable. RAND_add is like
8934 RAND_seed, but takes an extra argument for an entropy estimate
8935 (RAND_seed always assumes full entropy).
8938 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8939 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8940 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8941 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8942 false-positive rate of at most 2^-80 for random input.
8945 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8948 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8949 in the 0.9.5 release), this returns the chain
8950 from an X509_CTX structure with a dup of the stack and all
8951 the X509 reference counts upped: so the stack will exist
8952 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8955 Also make SSL_SESSION_print() print out the verify return
8959 *) Add manpage for the pkcs12 command. Also change the default
8960 behaviour so MAC iteration counts are used unless the new
8961 -nomaciter option is used. This improves file security and
8962 only older versions of MSIE (4.0 for example) need it.
8965 *) Honor the no-xxx Configure options when creating .DEF files.
8968 *) Add PKCS#10 attributes to field table: challengePassword,
8969 unstructuredName and unstructuredAddress. These are taken from
8970 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8971 international characters are used.
8973 More changes to X509_ATTRIBUTE code: allow the setting of types
8974 based on strings. Remove the 'loc' parameter when adding
8975 attributes because these will be a SET OF encoding which is sorted
8979 *) Initial changes to the 'req' utility to allow request generation
8980 automation. This will allow an application to just generate a template
8981 file containing all the field values and have req construct the
8984 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8985 used all over the place including certificate requests and PKCS#7
8986 structures. They are currently handled manually where necessary with
8987 some primitive wrappers for PKCS#7. The new functions behave in a
8988 manner analogous to the X509 extension functions: they allow
8989 attributes to be looked up by NID and added.
8991 Later something similar to the X509V3 code would be desirable to
8992 automatically handle the encoding, decoding and printing of the
8993 more complex types. The string types like challengePassword can
8994 be handled by the string table functions.
8996 Also modified the multi byte string table handling. Now there is
8997 a 'global mask' which masks out certain types. The table itself
8998 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8999 is useful when for example there is only one permissible type
9000 (as in countryName) and using the mask might result in no valid
9004 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9005 SSL_get_peer_finished to allow applications to obtain the latest
9006 Finished messages sent to the peer or expected from the peer,
9007 respectively. (SSL_get_peer_finished is usually the Finished message
9008 actually received from the peer, otherwise the protocol will be aborted.)
9010 As the Finished message are message digests of the complete handshake
9011 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9012 be used for external authentication procedures when the authentication
9013 provided by SSL/TLS is not desired or is not enough.
9016 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9017 the host supports BWX extension and if Compaq C is present on the
9018 $PATH. Just exploiting of the BWX extension results in 20-30%
9019 performance kick for some algorithms, e.g. DES and RC4 to mention
9020 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9024 *) Add support for MS "fast SGC". This is arguably a violation of the
9025 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9026 weak crypto and after checking the certificate is SGC a second one
9027 with strong crypto. MS SGC stops the first handshake after receiving
9028 the server certificate message and sends a second client hello. Since
9029 a server will typically do all the time consuming operations before
9030 expecting any further messages from the client (server key exchange
9031 is the most expensive) there is little difference between the two.
9033 To get OpenSSL to support MS SGC we have to permit a second client
9034 hello message after we have sent server done. In addition we have to
9035 reset the MAC if we do get this second client hello.
9038 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9039 if a DER encoded private key is RSA or DSA traditional format. Changed
9040 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9041 format DER encoded private key. Newer code should use PKCS#8 format which
9042 has the key type encoded in the ASN1 structure. Added DER private key
9043 support to pkcs8 application.
9046 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9047 ciphersuites has been selected (as required by the SSL 3/TLS 1
9048 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9049 is set, we interpret this as a request to violate the specification
9050 (the worst that can happen is a handshake failure, and 'correct'
9051 behaviour would result in a handshake failure anyway).
9054 *) In SSL_CTX_add_session, take into account that there might be multiple
9055 SSL_SESSION structures with the same session ID (e.g. when two threads
9056 concurrently obtain them from an external cache).
9057 The internal cache can handle only one SSL_SESSION with a given ID,
9058 so if there's a conflict, we now throw out the old one to achieve
9062 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9063 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9064 some routines that use cipher OIDs: some ciphers do not have OIDs
9065 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9069 *) Simplify the trust setting structure and code. Now we just have
9070 two sequences of OIDs for trusted and rejected settings. These will
9071 typically have values the same as the extended key usage extension
9072 and any application specific purposes.
9074 The trust checking code now has a default behaviour: it will just
9075 check for an object with the same NID as the passed id. Functions can
9076 be provided to override either the default behaviour or the behaviour
9077 for a given id. SSL client, server and email already have functions
9078 in place for compatibility: they check the NID and also return "trusted"
9079 if the certificate is self signed.
9082 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9083 traditional format into an EVP_PKEY structure.
9086 *) Add a password callback function PEM_cb() which either prompts for
9087 a password if usr_data is NULL or otherwise assumes it is a null
9088 terminated password. Allow passwords to be passed on command line
9089 environment or config files in a few more utilities.
9092 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9093 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9094 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9095 Update documentation.
9098 *) Support for ASN1 "NULL" type. This could be handled before by using
9099 ASN1_TYPE but there wasn't any function that would try to read a NULL
9100 and produce an error if it couldn't. For compatibility we also have
9101 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9102 don't allocate anything because they don't need to.
9105 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9107 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9109 *) Rebuild of the memory allocation routines used by OpenSSL code and
9110 possibly others as well. The purpose is to make an interface that
9111 provide hooks so anyone can build a separate set of allocation and
9112 deallocation routines to be used by OpenSSL, for example memory
9113 pool implementations, or something else, which was previously hard
9114 since Malloc(), Realloc() and Free() were defined as macros having
9115 the values malloc, realloc and free, respectively (except for Win32
9116 compilations). The same is provided for memory debugging code.
9117 OpenSSL already comes with functionality to find memory leaks, but
9118 this gives people a chance to debug other memory problems.
9120 With these changes, a new set of functions and macros have appeared:
9122 CRYPTO_set_mem_debug_functions() [F]
9123 CRYPTO_get_mem_debug_functions() [F]
9124 CRYPTO_dbg_set_options() [F]
9125 CRYPTO_dbg_get_options() [F]
9126 CRYPTO_malloc_debug_init() [M]
9128 The memory debug functions are NULL by default, unless the library
9129 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9130 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9131 gives the standard debugging functions that come with OpenSSL) or
9132 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9133 provided by the library user) must be used. When the standard
9134 debugging functions are used, CRYPTO_dbg_set_options can be used to
9135 request additional information:
9136 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9137 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9139 Also, things like CRYPTO_set_mem_functions will always give the
9140 expected result (the new set of functions is used for allocation
9141 and deallocation) at all times, regardless of platform and compiler
9144 To finish it up, some functions that were never use in any other
9145 way than through macros have a new API and new semantic:
9148 CRYPTO_dbg_realloc()
9151 All macros of value have retained their old syntax.
9152 [Richard Levitte and Bodo Moeller]
9154 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9155 ordering of SMIMECapabilities wasn't in "strength order" and there
9156 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9160 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9161 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9162 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9164 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9165 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9166 functionality to handle multipart/signed properly) and a utility
9167 called 'smime' to call all this stuff. This is based on code I
9168 originally wrote for Celo who have kindly allowed it to be
9169 included in OpenSSL.
9172 *) Add variants des_set_key_checked and des_set_key_unchecked of
9173 des_set_key (aka des_key_sched). Global variable des_check_key
9174 decides which of these is called by des_set_key; this way
9175 des_check_key behaves as it always did, but applications and
9176 the library itself, which was buggy for des_check_key == 1,
9177 have a cleaner way to pick the version they need.
9180 *) New function PKCS12_newpass() which changes the password of a
9184 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9185 dynamic mix. In both cases the ids can be used as an index into the
9186 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9187 functions so they accept a list of the field values and the
9188 application doesn't need to directly manipulate the X509_TRUST
9192 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9196 *) Modify the way the V3 extension code looks up extensions. This now
9197 works in a similar way to the object code: we have some "standard"
9198 extensions in a static table which is searched with OBJ_bsearch()
9199 and the application can add dynamic ones if needed. The file
9200 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9201 updated whenever a new extension is added to the core code and kept
9202 in ext_nid order. There is a simple program 'tabtest.c' which checks
9203 this. New extensions are not added too often so this file can readily
9204 be maintained manually.
9206 There are two big advantages in doing things this way. The extensions
9207 can be looked up immediately and no longer need to be "added" using
9208 X509V3_add_standard_extensions(): this function now does nothing.
9209 [Side note: I get *lots* of email saying the extension code doesn't
9210 work because people forget to call this function]
9211 Also no dynamic allocation is done unless new extensions are added:
9212 so if we don't add custom extensions there is no need to call
9213 X509V3_EXT_cleanup().
9216 *) Modify enc utility's salting as follows: make salting the default. Add a
9217 magic header, so unsalted files fail gracefully instead of just decrypting
9218 to garbage. This is because not salting is a big security hole, so people
9219 should be discouraged from doing it.
9222 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9223 digest to be passed on the command line but it only used this
9224 parameter when signing a certificate. Modified so all relevant
9225 operations are affected by the digest parameter including the
9226 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9227 DSA key was used because it didn't fix the digest.
9230 *) Initial certificate chain verify code. Currently tests the untrusted
9231 certificates for consistency with the verify purpose (which is set
9232 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9234 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9235 this is because it will reject chains with invalid extensions whereas
9236 every previous version of OpenSSL and SSLeay made no checks at all.
9238 Trust code: checks the root CA for the relevant trust settings. Trust
9239 settings have an initial value consistent with the verify purpose: e.g.
9240 if the verify purpose is for SSL client use it expects the CA to be
9241 trusted for SSL client use. However the default value can be changed to
9242 permit custom trust settings: one example of this would be to only trust
9243 certificates from a specific "secure" set of CAs.
9245 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9246 which should be used for version portability: especially since the
9247 verify structure is likely to change more often now.
9249 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9250 to set them. If not set then assume SSL clients will verify SSL servers
9253 Two new options to the verify program: -untrusted allows a set of
9254 untrusted certificates to be passed in and -purpose which sets the
9255 intended purpose of the certificate. If a purpose is set then the
9256 new chain verify code is used to check extension consistency.
9259 *) Support for the authority information access extension.
9262 *) Modify RSA and DSA PEM read routines to transparently handle
9263 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9264 public keys in a format compatible with certificate
9265 SubjectPublicKeyInfo structures. Unfortunately there were already
9266 functions called *_PublicKey_* which used various odd formats so
9267 these are retained for compatibility: however the DSA variants were
9268 never in a public release so they have been deleted. Changed dsa/rsa
9269 utilities to handle the new format: note no releases ever handled public
9270 keys so we should be OK.
9272 The primary motivation for this change is to avoid the same fiasco
9273 that dogs private keys: there are several incompatible private key
9274 formats some of which are standard and some OpenSSL specific and
9275 require various evil hacks to allow partial transparent handling and
9276 even then it doesn't work with DER formats. Given the option anything
9277 other than PKCS#8 should be dumped: but the other formats have to
9278 stay in the name of compatibility.
9280 With public keys and the benefit of hindsight one standard format
9281 is used which works with EVP_PKEY, RSA or DSA structures: though
9282 it clearly returns an error if you try to read the wrong kind of key.
9284 Added a -pubkey option to the 'x509' utility to output the public key.
9285 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9286 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9287 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9288 that do the same as the EVP_PKEY_assign_*() except they up the
9289 reference count of the added key (they don't "swallow" the
9293 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9294 CRLs would fail if the file contained no certificates or no CRLs:
9295 added a new function to read in both types and return the number
9296 read: this means that if none are read it will be an error. The
9297 DER versions of the certificate and CRL reader would always fail
9298 because it isn't possible to mix certificates and CRLs in DER format
9299 without choking one or the other routine. Changed this to just read
9300 a certificate: this is the best we can do. Also modified the code
9301 in apps/verify.c to take notice of return codes: it was previously
9302 attempting to read in certificates from NULL pointers and ignoring
9303 any errors: this is one reason why the cert and CRL reader seemed
9304 to work. It doesn't check return codes from the default certificate
9305 routines: these may well fail if the certificates aren't installed.
9308 *) Code to support otherName option in GeneralName.
9311 *) First update to verify code. Change the verify utility
9312 so it warns if it is passed a self signed certificate:
9313 for consistency with the normal behaviour. X509_verify
9314 has been modified to it will now verify a self signed
9315 certificate if *exactly* the same certificate appears
9316 in the store: it was previously impossible to trust a
9317 single self signed certificate. This means that:
9318 openssl verify ss.pem
9319 now gives a warning about a self signed certificate but
9320 openssl verify -CAfile ss.pem ss.pem
9324 *) For servers, store verify_result in SSL_SESSION data structure
9325 (and add it to external session representation).
9326 This is needed when client certificate verifications fails,
9327 but an application-provided verification callback (set by
9328 SSL_CTX_set_cert_verify_callback) allows accepting the session
9329 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9330 but returns 1): When the session is reused, we have to set
9331 ssl->verify_result to the appropriate error code to avoid
9333 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9335 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9336 case in PKCS7_dataInit() where the signed PKCS7 structure
9337 didn't contain any existing data because it was being created.
9338 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9340 *) Add a salt to the key derivation routines in enc.c. This
9341 forms the first 8 bytes of the encrypted file. Also add a
9342 -S option to allow a salt to be input on the command line.
9345 *) New function X509_cmp(). Oddly enough there wasn't a function
9346 to compare two certificates. We do this by working out the SHA1
9347 hash and comparing that. X509_cmp() will be needed by the trust
9351 *) SSL_get1_session() is like SSL_get_session(), but increments
9352 the reference count in the SSL_SESSION returned.
9353 [Geoff Thorpe <geoff@eu.c2.net>]
9355 *) Fix for 'req': it was adding a null to request attributes.
9356 Also change the X509_LOOKUP and X509_INFO code to handle
9357 certificate auxiliary information.
9360 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9364 *) Add the possibility to add extra information to the memory leak
9365 detecting output, to form tracebacks, showing from where each
9366 allocation was originated: CRYPTO_push_info("constant string") adds
9367 the string plus current file name and line number to a per-thread
9368 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9369 is like calling CYRPTO_pop_info() until the stack is empty.
9370 Also updated memory leak detection code to be multi-thread-safe.
9373 *) Add options -text and -noout to pkcs7 utility and delete the
9374 encryption options which never did anything. Update docs.
9377 *) Add options to some of the utilities to allow the pass phrase
9378 to be included on either the command line (not recommended on
9379 OSes like Unix) or read from the environment. Update the
9380 manpages and fix a few bugs.
9383 *) Add a few manpages for some of the openssl commands.
9386 *) Fix the -revoke option in ca. It was freeing up memory twice,
9387 leaking and not finding already revoked certificates.
9390 *) Extensive changes to support certificate auxiliary information.
9391 This involves the use of X509_CERT_AUX structure and X509_AUX
9392 functions. An X509_AUX function such as PEM_read_X509_AUX()
9393 can still read in a certificate file in the usual way but it
9394 will also read in any additional "auxiliary information". By
9395 doing things this way a fair degree of compatibility can be
9396 retained: existing certificates can have this information added
9397 using the new 'x509' options.
9399 Current auxiliary information includes an "alias" and some trust
9400 settings. The trust settings will ultimately be used in enhanced
9401 certificate chain verification routines: currently a certificate
9402 can only be trusted if it is self signed and then it is trusted
9406 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9407 The problem was that one of the replacement routines had not been working
9408 since SSLeay releases. For now the offending routine has been replaced
9409 with non-optimised assembler. Even so, this now gives around 95%
9410 performance improvement for 1024 bit RSA signs.
9413 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9414 handling. Most clients have the effective key size in bits equal to
9415 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9416 A few however don't do this and instead use the size of the decrypted key
9417 to determine the RC2 key length and the AlgorithmIdentifier to determine
9418 the effective key length. In this case the effective key length can still
9419 be 40 bits but the key length can be 168 bits for example. This is fixed
9420 by manually forcing an RC2 key into the EVP_PKEY structure because the
9421 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9422 the key length and effective key length are equal.
9425 *) Add a bunch of functions that should simplify the creation of
9426 X509_NAME structures. Now you should be able to do:
9427 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9428 and have it automatically work out the correct field type and fill in
9429 the structures. The more adventurous can try:
9430 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9431 and it will (hopefully) work out the correct multibyte encoding.
9434 *) Change the 'req' utility to use the new field handling and multibyte
9435 copy routines. Before the DN field creation was handled in an ad hoc
9436 way in req, ca, and x509 which was rather broken and didn't support
9437 BMPStrings or UTF8Strings. Since some software doesn't implement
9438 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9439 using the dirstring_type option. See the new comment in the default
9440 openssl.cnf for more info.
9443 *) Make crypto/rand/md_rand.c more robust:
9444 - Assure unique random numbers after fork().
9445 - Make sure that concurrent threads access the global counter and
9446 md serializably so that we never lose entropy in them
9447 or use exactly the same state in multiple threads.
9448 Access to the large state is not always serializable because
9449 the additional locking could be a performance killer, and
9450 md should be large enough anyway.
9453 *) New file apps/app_rand.c with commonly needed functionality
9454 for handling the random seed file.
9456 Use the random seed file in some applications that previously did not:
9458 dsaparam -genkey (which also ignored its '-rand' option),
9461 x509 (when signing).
9462 Except on systems with /dev/urandom, it is crucial to have a random
9463 seed file at least for key creation, DSA signing, and for DH exchanges;
9464 for RSA signatures we could do without one.
9466 gendh and gendsa (unlike genrsa) used to read only the first byte
9467 of each file listed in the '-rand' option. The function as previously
9468 found in genrsa is now in app_rand.c and is used by all programs
9469 that support '-rand'.
9472 *) In RAND_write_file, use mode 0600 for creating files;
9473 don't just chmod when it may be too late.
9476 *) Report an error from X509_STORE_load_locations
9477 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9480 *) New function ASN1_mbstring_copy() this copies a string in either
9481 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9482 into an ASN1_STRING type. A mask of permissible types is passed
9483 and it chooses the "minimal" type to use or an error if not type
9487 *) Add function equivalents to the various macros in asn1.h. The old
9488 macros are retained with an M_ prefix. Code inside the library can
9489 use the M_ macros. External code (including the openssl utility)
9490 should *NOT* in order to be "shared library friendly".
9493 *) Add various functions that can check a certificate's extensions
9494 to see if it usable for various purposes such as SSL client,
9495 server or S/MIME and CAs of these types. This is currently
9496 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9497 verification. Also added a -purpose flag to x509 utility to
9498 print out all the purposes.
9501 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9505 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9506 for, obtain and decode and extension and obtain its critical flag.
9507 This allows all the necessary extension code to be handled in a
9508 single function call.
9511 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9512 platforms. See crypto/rc4/rc4_enc.c for further details.
9515 *) New -noout option to asn1parse. This causes no output to be produced
9516 its main use is when combined with -strparse and -out to extract data
9517 from a file (which may not be in ASN.1 format).
9520 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9521 when producing the local key id.
9522 [Richard Levitte <levitte@stacken.kth.se>]
9524 *) New option -dhparam in s_server. This allows a DH parameter file to be
9525 stated explicitly. If it is not stated then it tries the first server
9526 certificate file. The previous behaviour hard coded the filename
9530 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9531 a public key to be input or output. For example:
9532 openssl rsa -in key.pem -pubout -out pubkey.pem
9533 Also added necessary DSA public key functions to handle this.
9536 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9537 in the message. This was handled by allowing
9538 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9539 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9541 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9542 to the end of the strings whereas this didn't. This would cause problems
9543 if strings read with d2i_ASN1_bytes() were later modified.
9544 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9546 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9547 data and it contains EOF it will end up returning an error. This is
9548 caused by input 46 bytes long. The cause is due to the way base64
9549 BIOs find the start of base64 encoded data. They do this by trying a
9550 trial decode on each line until they find one that works. When they
9551 do a flag is set and it starts again knowing it can pass all the
9552 data directly through the decoder. Unfortunately it doesn't reset
9553 the context it uses. This means that if EOF is reached an attempt
9554 is made to pass two EOFs through the context and this causes the
9555 resulting error. This can also cause other problems as well. As is
9556 usual with these problems it takes *ages* to find and the fix is
9557 trivial: move one line.
9558 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9560 *) Ugly workaround to get s_client and s_server working under Windows. The
9561 old code wouldn't work because it needed to select() on sockets and the
9562 tty (for keypresses and to see if data could be written). Win32 only
9563 supports select() on sockets so we select() with a 1s timeout on the
9564 sockets and then see if any characters are waiting to be read, if none
9565 are present then we retry, we also assume we can always write data to
9566 the tty. This isn't nice because the code then blocks until we've
9567 received a complete line of data and it is effectively polling the
9568 keyboard at 1s intervals: however it's quite a bit better than not
9569 working at all :-) A dedicated Windows application might handle this
9570 with an event loop for example.
9573 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9574 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9575 will be called when RSA_sign() and RSA_verify() are used. This is useful
9576 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9577 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9578 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9579 This necessitated the support of an extra signature type NID_md5_sha1
9580 for SSL signatures and modifications to the SSL library to use it instead
9581 of calling RSA_public_decrypt() and RSA_private_encrypt().
9584 *) Add new -verify -CAfile and -CApath options to the crl program, these
9585 will lookup a CRL issuers certificate and verify the signature in a
9586 similar way to the verify program. Tidy up the crl program so it
9587 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9588 less strict. It will now permit CRL extensions even if it is not
9589 a V2 CRL: this will allow it to tolerate some broken CRLs.
9592 *) Initialize all non-automatic variables each time one of the openssl
9593 sub-programs is started (this is necessary as they may be started
9594 multiple times from the "OpenSSL>" prompt).
9595 [Lennart Bang, Bodo Moeller]
9597 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9598 removing all other RSA functionality (this is what NO_RSA does). This
9599 is so (for example) those in the US can disable those operations covered
9600 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9604 *) Non-copying interface to BIO pairs.
9605 (still largely untested)
9608 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9609 ASCII string. This was handled independently in various places before.
9612 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9613 UTF8 strings a character at a time.
9616 *) Use client_version from client hello to select the protocol
9617 (s23_srvr.c) and for RSA client key exchange verification
9618 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9621 *) Add various utility functions to handle SPKACs, these were previously
9622 handled by poking round in the structure internals. Added new function
9623 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9624 print, verify and generate SPKACs. Based on an original idea from
9625 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9628 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9631 *) Allow the config file extension section to be overwritten on the
9632 command line. Based on an original idea from Massimiliano Pala
9633 <madwolf@comune.modena.it>. The new option is called -extensions
9634 and can be applied to ca, req and x509. Also -reqexts to override
9635 the request extensions in req and -crlexts to override the crl extensions
9639 *) Add new feature to the SPKAC handling in ca. Now you can include
9640 the same field multiple times by preceding it by "XXXX." for example:
9643 this is the same syntax as used in the req config file.
9646 *) Allow certificate extensions to be added to certificate requests. These
9647 are specified in a 'req_extensions' option of the req section of the
9648 config file. They can be printed out with the -text option to req but
9649 are otherwise ignored at present.
9652 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9653 data read consists of only the final block it would not decrypted because
9654 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9655 A misplaced 'break' also meant the decrypted final block might not be
9656 copied until the next read.
9659 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9660 a few extra parameters to the DH structure: these will be useful if
9661 for example we want the value of 'q' or implement X9.42 DH.
9664 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9665 provides hooks that allow the default DSA functions or functions on a
9666 "per key" basis to be replaced. This allows hardware acceleration and
9667 hardware key storage to be handled without major modification to the
9668 library. Also added low level modexp hooks and CRYPTO_EX structure and
9669 associated functions.
9672 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9673 as "read only": it can't be written to and the buffer it points to will
9674 not be freed. Reading from a read only BIO is much more efficient than
9675 a normal memory BIO. This was added because there are several times when
9676 an area of memory needs to be read from a BIO. The previous method was
9677 to create a memory BIO and write the data to it, this results in two
9678 copies of the data and an O(n^2) reading algorithm. There is a new
9679 function BIO_new_mem_buf() which creates a read only memory BIO from
9680 an area of memory. Also modified the PKCS#7 routines to use read only
9684 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9685 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9686 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9687 but a retry condition occured while trying to read the rest.
9690 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9691 NID_pkcs7_encrypted by default: this was wrong since this should almost
9692 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9693 the encrypted data type: this is a more sensible place to put it and it
9694 allows the PKCS#12 code to be tidied up that duplicated this
9698 *) Changed obj_dat.pl script so it takes its input and output files on
9699 the command line. This should avoid shell escape redirection problems
9703 *) Initial support for certificate extension requests, these are included
9704 in things like Xenroll certificate requests. Included functions to allow
9705 extensions to be obtained and added.
9708 *) -crlf option to s_client and s_server for sending newlines as
9709 CRLF (as required by many protocols).
9712 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9714 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9715 [Ralf S. Engelschall]
9717 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9718 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9720 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9724 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9725 DH parameters/keys (q is lost during that conversion, but the resulting
9726 DH parameters contain its length).
9728 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9729 much faster than DH_generate_parameters (which creates parameters
9730 where p = 2*q + 1), and also the smaller q makes DH computations
9731 much more efficient (160-bit exponentiation instead of 1024-bit
9732 exponentiation); so this provides a convenient way to support DHE
9733 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9734 utter importance to use
9735 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9737 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9738 when such DH parameters are used, because otherwise small subgroup
9739 attacks may become possible!
9742 *) Avoid memory leak in i2d_DHparams.
9745 *) Allow the -k option to be used more than once in the enc program:
9746 this allows the same encrypted message to be read by multiple recipients.
9749 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9750 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9751 it will always use the numerical form of the OID, even if it has a short
9755 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9756 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9757 otherwise bn_mod_exp was called. In the case of hardware keys for example
9758 no private key components need be present and it might store extra data
9759 in the RSA structure, which cannot be accessed from bn_mod_exp.
9760 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9761 private key operations.
9764 *) Added support for SPARC Linux.
9767 *) pem_password_cb function type incompatibly changed from
9768 typedef int pem_password_cb(char *buf, int size, int rwflag);
9770 ....(char *buf, int size, int rwflag, void *userdata);
9771 so that applications can pass data to their callbacks:
9772 The PEM[_ASN1]_{read,write}... functions and macros now take an
9773 additional void * argument, which is just handed through whenever
9774 the password callback is called.
9775 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9777 New function SSL_CTX_set_default_passwd_cb_userdata.
9779 Compatibility note: As many C implementations push function arguments
9780 onto the stack in reverse order, the new library version is likely to
9781 interoperate with programs that have been compiled with the old
9782 pem_password_cb definition (PEM_whatever takes some data that
9783 happens to be on the stack as its last argument, and the callback
9784 just ignores this garbage); but there is no guarantee whatsoever that
9787 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9788 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9789 problems not only on Windows, but also on some Unix platforms.
9790 To avoid problematic command lines, these definitions are now in an
9791 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9792 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9795 *) MIPS III/IV assembler module is reimplemented.
9798 *) More DES library cleanups: remove references to srand/rand and
9799 delete an unused file.
9802 *) Add support for the the free Netwide assembler (NASM) under Win32,
9803 since not many people have MASM (ml) and it can be hard to obtain.
9804 This is currently experimental but it seems to work OK and pass all
9805 the tests. Check out INSTALL.W32 for info.
9808 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9809 without temporary keys kept an extra copy of the server key,
9810 and connections with temporary keys did not free everything in case
9814 *) New function RSA_check_key and new openssl rsa option -check
9815 for verifying the consistency of RSA keys.
9816 [Ulf Moeller, Bodo Moeller]
9818 *) Various changes to make Win32 compile work:
9819 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9820 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9821 comparison" warnings.
9822 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9825 *) Add a debugging option to PKCS#5 v2 key generation function: when
9826 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9827 derived keys are printed to stderr.
9830 *) Copy the flags in ASN1_STRING_dup().
9831 [Roman E. Pavlov <pre@mo.msk.ru>]
9833 *) The x509 application mishandled signing requests containing DSA
9834 keys when the signing key was also DSA and the parameters didn't match.
9836 It was supposed to omit the parameters when they matched the signing key:
9837 the verifying software was then supposed to automatically use the CA's
9838 parameters if they were absent from the end user certificate.
9840 Omitting parameters is no longer recommended. The test was also
9841 the wrong way round! This was probably due to unusual behaviour in
9842 EVP_cmp_parameters() which returns 1 if the parameters match.
9843 This meant that parameters were omitted when they *didn't* match and
9844 the certificate was useless. Certificates signed with 'ca' didn't have
9846 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9848 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9849 The interface is as follows:
9850 Applications can use
9851 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9852 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9853 "off" is now the default.
9854 The library internally uses
9855 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9856 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9857 to disable memory-checking temporarily.
9859 Some inconsistent states that previously were possible (and were
9860 even the default) are now avoided.
9862 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9863 with each memory chunk allocated; this is occasionally more helpful
9864 than just having a counter.
9866 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9868 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9872 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9873 which largely parallels "options", but is for changing API behaviour,
9874 whereas "options" are about protocol behaviour.
9875 Initial "mode" flags are:
9877 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9878 a single record has been written.
9879 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9880 retries use the same buffer location.
9881 (But all of the contents must be
9885 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9888 *) Fix problems with no-hmac etc.
9889 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9891 *) New functions RSA_get_default_method(), RSA_set_method() and
9892 RSA_get_method(). These allows replacement of RSA_METHODs without having
9893 to mess around with the internals of an RSA structure.
9896 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9897 Also really enable memory leak checks in openssl.c and in some
9899 [Chad C. Mulligan, Bodo Moeller]
9901 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9902 up the length of negative integers. This has now been simplified to just
9903 store the length when it is first determined and use it later, rather
9904 than trying to keep track of where data is copied and updating it to
9906 [Steve Henson, reported by Brien Wheeler
9907 <bwheeler@authentica-security.com>]
9909 *) Add a new function PKCS7_signatureVerify. This allows the verification
9910 of a PKCS#7 signature but with the signing certificate passed to the
9911 function itself. This contrasts with PKCS7_dataVerify which assumes the
9912 certificate is present in the PKCS#7 structure. This isn't always the
9913 case: certificates can be omitted from a PKCS#7 structure and be
9914 distributed by "out of band" means (such as a certificate database).
9917 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9918 function prototypes in pem.h, also change util/mkdef.pl to add the
9919 necessary function names.
9922 *) mk1mf.pl (used by Windows builds) did not properly read the
9923 options set by Configure in the top level Makefile, and Configure
9924 was not even able to write more than one option correctly.
9925 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9928 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9929 file to be loaded from a BIO or FILE pointer. The BIO version will
9930 for example allow memory BIOs to contain config info.
9933 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9934 Whoever hopes to achieve shared-library compatibility across versions
9935 must use this, not the compile-time macro.
9936 (Exercise 0.9.4: Which is the minimum library version required by
9938 Note: All this applies only to multi-threaded programs, others don't
9942 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9943 through a BIO pair triggered the default case, i.e.
9944 SSLerr(...,SSL_R_UNKNOWN_STATE).
9947 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9948 can use the SSL library even if none of the specific BIOs is
9952 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9953 for the encoded length.
9954 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9956 *) Add initial documentation of the X509V3 functions.
9959 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9960 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9961 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9962 secure PKCS#8 private key format with a high iteration count.
9965 *) Fix determination of Perl interpreter: A perl or perl5
9966 _directory_ in $PATH was also accepted as the interpreter.
9967 [Ralf S. Engelschall]
9969 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9970 wrong with it but it was very old and did things like calling
9971 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9975 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9976 to use the new extension code.
9979 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9980 with macros. This should make it easier to change their form, add extra
9981 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9985 *) Add to configuration table a new entry that can specify an alternative
9986 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9987 according to Mark Crispin <MRC@Panda.COM>.
9991 *) DES CBC did not update the IV. Weird.
9994 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9995 Changing the behaviour of the former might break existing programs --
9996 where IV updating is needed, des_ncbc_encrypt can be used.
9999 *) When bntest is run from "make test" it drives bc to check its
10000 calculations, as well as internally checking them. If an internal check
10001 fails, it needs to cause bc to give a non-zero result or make test carries
10002 on without noticing the failure. Fixed.
10005 *) DES library cleanups.
10008 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10009 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10010 ciphers. NOTE: although the key derivation function has been verified
10011 against some published test vectors it has not been extensively tested
10012 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10016 *) Instead of "mkdir -p", which is not fully portable, use new
10017 Perl script "util/mkdir-p.pl".
10020 *) Rewrite the way password based encryption (PBE) is handled. It used to
10021 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10022 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10023 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10024 the 'parameter' field of the AlgorithmIdentifier is passed to the
10025 underlying key generation function so it must do its own ASN1 parsing.
10026 This has also changed the EVP_PBE_CipherInit() function which now has a
10027 'parameter' argument instead of literal salt and iteration count values
10028 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10031 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10032 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10033 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10034 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10035 value was just used as a "magic string" and not used directly its
10036 value doesn't matter.
10039 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10043 *) "linux-sparc64" configuration (ultrapenguin).
10044 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10045 "linux-sparc" configuration.
10046 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10048 *) config now generates no-xxx options for missing ciphers.
10051 *) Support the EBCDIC character set (work in progress).
10052 File ebcdic.c not yet included because it has a different license.
10053 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10055 *) Support BS2000/OSD-POSIX.
10056 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10058 *) Make callbacks for key generation use void * instead of char *.
10061 *) Make S/MIME samples compile (not yet tested).
10064 *) Additional typesafe stacks.
10067 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10071 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10073 *) New configuration variant "sco5-gcc".
10075 *) Updated some demos.
10076 [Sean O Riordain, Wade Scholine]
10078 *) Add missing BIO_free at exit of pkcs12 application.
10081 *) Fix memory leak in conf.c.
10084 *) Updates for Win32 to assembler version of MD5.
10087 *) Set #! path to perl in apps/der_chop to where we found it
10088 instead of using a fixed path.
10091 *) SHA library changes for irix64-mips4-cc.
10094 *) Improvements for VMS support.
10098 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10100 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10101 This also avoids the problems with SC4.2 and unpatched SC5.
10102 [Andy Polyakov <appro@fy.chalmers.se>]
10104 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10105 These are required because of the typesafe stack would otherwise break
10106 existing code. If old code used a structure member which used to be STACK
10107 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10108 sk_num or sk_value it would produce an error because the num, data members
10109 are not present in STACK_OF. Now it just produces a warning. sk_set
10110 replaces the old method of assigning a value to sk_value
10111 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10112 that does this will no longer work (and should use sk_set instead) but
10113 this could be regarded as a "questionable" behaviour anyway.
10116 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10117 correctly handle encrypted S/MIME data.
10120 *) Change type of various DES function arguments from des_cblock
10121 (which means, in function argument declarations, pointer to char)
10122 to des_cblock * (meaning pointer to array with 8 char elements),
10123 which allows the compiler to do more typechecking; it was like
10124 that back in SSLeay, but with lots of ugly casts.
10126 Introduce new type const_des_cblock.
10129 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10130 problems: find RecipientInfo structure that matches recipient certificate
10131 and initialise the ASN1 structures properly based on passed cipher.
10134 *) Belatedly make the BN tests actually check the results.
10137 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10138 to and from BNs: it was completely broken. New compilation option
10139 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10140 key elements as negative integers.
10143 *) Reorganize and speed up MD5.
10144 [Andy Polyakov <appro@fy.chalmers.se>]
10147 [Richard Levitte <richard@levitte.org>]
10149 *) New option -out to asn1parse to allow the parsed structure to be
10150 output to a file. This is most useful when combined with the -strparse
10151 option to examine the output of things like OCTET STRINGS.
10154 *) Make SSL library a little more fool-proof by not requiring any longer
10155 that SSL_set_{accept,connect}_state be called before
10156 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10157 in many applications because usually everything *appeared* to work as
10158 intended anyway -- now it really works as intended).
10161 *) Move openssl.cnf out of lib/.
10164 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10165 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10166 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10167 [Ralf S. Engelschall]
10169 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10170 handle PKCS#7 enveloped data properly.
10171 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10173 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10174 copying pointers. The cert_st handling is changed by this in
10175 various ways (and thus what used to be known as ctx->default_cert
10176 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10177 any longer when s->cert does not give us what we need).
10178 ssl_cert_instantiate becomes obsolete by this change.
10179 As soon as we've got the new code right (possibly it already is?),
10180 we have solved a couple of bugs of the earlier code where s->cert
10181 was used as if it could not have been shared with other SSL structures.
10183 Note that using the SSL API in certain dirty ways now will result
10184 in different behaviour than observed with earlier library versions:
10185 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10186 does not influence s as it used to.
10188 In order to clean up things more thoroughly, inside SSL_SESSION
10189 we don't use CERT any longer, but a new structure SESS_CERT
10190 that holds per-session data (if available); currently, this is
10191 the peer's certificate chain and, for clients, the server's certificate
10192 and temporary key. CERT holds only those values that can have
10193 meaningful defaults in an SSL_CTX.
10196 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10197 from the internal representation. Various PKCS#7 fixes: remove some
10198 evil casts and set the enc_dig_alg field properly based on the signing
10202 *) Allow PKCS#12 password to be set from the command line or the
10203 environment. Let 'ca' get its config file name from the environment
10204 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10208 *) Allow certificate policies extension to use an IA5STRING for the
10209 organization field. This is contrary to the PKIX definition but
10210 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10214 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10215 without disallowing inline assembler and the like for non-pedantic builds.
10218 *) Support Borland C++ builder.
10219 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10221 *) Support Mingw32.
10224 *) SHA-1 cleanups and performance enhancements.
10225 [Andy Polyakov <appro@fy.chalmers.se>]
10227 *) Sparc v8plus assembler for the bignum library.
10228 [Andy Polyakov <appro@fy.chalmers.se>]
10230 *) Accept any -xxx and +xxx compiler options in Configure.
10233 *) Update HPUX configuration.
10236 *) Add missing sk_<type>_unshift() function to safestack.h
10237 [Ralf S. Engelschall]
10239 *) New function SSL_CTX_use_certificate_chain_file that sets the
10240 "extra_cert"s in addition to the certificate. (This makes sense
10241 only for "PEM" format files, as chains as a whole are not
10245 *) Support verify_depth from the SSL API.
10246 x509_vfy.c had what can be considered an off-by-one-error:
10247 Its depth (which was not part of the external interface)
10248 was actually counting the number of certificates in a chain;
10249 now it really counts the depth.
10252 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10253 instead of X509err, which often resulted in confusing error
10254 messages since the error codes are not globally unique
10255 (e.g. an alleged error in ssl3_accept when a certificate
10256 didn't match the private key).
10258 *) New function SSL_CTX_set_session_id_context that allows to set a default
10259 value (so that you don't need SSL_set_session_id_context for each
10260 connection using the SSL_CTX).
10263 *) OAEP decoding bug fix.
10266 *) Support INSTALL_PREFIX for package builders, as proposed by
10270 *) New Configure options "threads" and "no-threads". For systems
10271 where the proper compiler options are known (currently Solaris
10272 and Linux), "threads" is the default.
10275 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10278 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10279 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10280 such as /usr/local/bin.
10283 *) "make linux-shared" to build shared libraries.
10284 [Niels Poppe <niels@netbox.org>]
10286 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10289 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10290 extension adding in x509 utility.
10293 *) Remove NOPROTO sections and error code comments.
10296 *) Partial rewrite of the DEF file generator to now parse the ANSI
10300 *) New Configure options --prefix=DIR and --openssldir=DIR.
10303 *) Complete rewrite of the error code script(s). It is all now handled
10304 by one script at the top level which handles error code gathering,
10305 header rewriting and C source file generation. It should be much better
10306 than the old method: it now uses a modified version of Ulf's parser to
10307 read the ANSI prototypes in all header files (thus the old K&R definitions
10308 aren't needed for error creation any more) and do a better job of
10309 translating function codes into names. The old 'ASN1 error code imbedded
10310 in a comment' is no longer necessary and it doesn't use .err files which
10311 have now been deleted. Also the error code call doesn't have to appear all
10312 on one line (which resulted in some large lines...).
10315 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10318 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10319 0 (which usually indicates a closed connection), but continue reading.
10322 *) Fix some race conditions.
10325 *) Add support for CRL distribution points extension. Add Certificate
10326 Policies and CRL distribution points documentation.
10329 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10332 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10333 8 of keying material. Merlin has also confirmed interop with this fix
10334 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10335 [Merlin Hughes <merlin@baltimore.ie>]
10337 *) Fix lots of warnings.
10338 [Richard Levitte <levitte@stacken.kth.se>]
10340 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10341 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10342 [Richard Levitte <levitte@stacken.kth.se>]
10344 *) Fix problems with sizeof(long) == 8.
10345 [Andy Polyakov <appro@fy.chalmers.se>]
10347 *) Change functions to ANSI C.
10350 *) Fix typos in error codes.
10351 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10353 *) Remove defunct assembler files from Configure.
10356 *) SPARC v8 assembler BIGNUM implementation.
10357 [Andy Polyakov <appro@fy.chalmers.se>]
10359 *) Support for Certificate Policies extension: both print and set.
10360 Various additions to support the r2i method this uses.
10363 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10364 return a const string when you are expecting an allocated buffer.
10367 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10368 types DirectoryString and DisplayText.
10371 *) Add code to allow r2i extensions to access the configuration database,
10372 add an LHASH database driver and add several ctx helper functions.
10375 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10376 fail when they extended the size of a BIGNUM.
10379 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10380 support typesafe stack.
10383 *) Fix typo in SSL_[gs]et_options().
10384 [Nils Frostberg <nils@medcom.se>]
10386 *) Delete various functions and files that belonged to the (now obsolete)
10387 old X509V3 handling code.
10390 *) New Configure option "rsaref".
10393 *) Don't auto-generate pem.h.
10396 *) Introduce type-safe ASN.1 SETs.
10399 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10400 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10402 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10403 that links with OpenSSL (well at least cause lots of warnings), but fear
10404 not: the conversion is trivial, and it eliminates loads of evil casts. A
10405 few STACKed things have been converted already. Feel free to convert more.
10406 In the fullness of time, I'll do away with the STACK type altogether.
10409 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10410 specified in <certfile> by updating the entry in the index.txt file.
10411 This way one no longer has to edit the index.txt file manually for
10412 revoking a certificate. The -revoke option does the gory details now.
10413 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10415 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10416 `-text' option at all and this way the `-noout -text' combination was
10417 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10418 [Ralf S. Engelschall]
10420 *) Make sure a corresponding plain text error message exists for the
10421 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10422 verify callback function determined that a certificate was revoked.
10423 [Ralf S. Engelschall]
10425 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10426 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10427 all available cipers including rc5, which was forgotten until now.
10428 In order to let the testing shell script know which algorithms
10429 are available, a new (up to now undocumented) command
10430 "openssl list-cipher-commands" is used.
10433 *) Bugfix: s_client occasionally would sleep in select() when
10434 it should have checked SSL_pending() first.
10437 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10438 the raw DSA values prior to ASN.1 encoding.
10441 *) Tweaks to Configure
10442 [Niels Poppe <niels@netbox.org>]
10444 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10448 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10451 *) New config option to avoid instructions that are illegal on the 80386.
10452 The default code is faster, but requires at least a 486.
10455 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10456 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10457 same as SSL2_VERSION anyway.
10460 *) New "-showcerts" option for s_client.
10463 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10464 application. Various cleanups and fixes.
10467 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10468 modify error routines to work internally. Add error codes and PBE init
10469 to library startup routines.
10472 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10473 packing functions to asn1 and evp. Changed function names and error
10474 codes along the way.
10477 *) PKCS12 integration: and so it begins... First of several patches to
10478 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10479 objects to objects.h
10482 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10483 and display support for Thawte strong extranet extension.
10486 *) Add LinuxPPC support.
10487 [Jeff Dubrule <igor@pobox.org>]
10489 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10490 bn_div_words in alpha.s.
10491 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10493 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10494 OAEP isn't supported when OpenSSL is built with RSAref.
10495 [Ulf Moeller <ulf@fitug.de>]
10497 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10498 so they no longer are missing under -DNOPROTO.
10499 [Soren S. Jorvang <soren@t.dk>]
10502 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10504 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10505 doesn't work when the session is reused. Coming soon!
10508 *) Fix a security hole, that allows sessions to be reused in the wrong
10509 context thus bypassing client cert protection! All software that uses
10510 client certs and session caches in multiple contexts NEEDS PATCHING to
10511 allow session reuse! A fuller solution is in the works.
10512 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10514 *) Some more source tree cleanups (removed obsolete files
10515 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10516 permission on "config" script to be executable) and a fix for the INSTALL
10518 [Ulf Moeller <ulf@fitug.de>]
10520 *) Remove some legacy and erroneous uses of malloc, free instead of
10522 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10524 *) Make rsa_oaep_test return non-zero on error.
10525 [Ulf Moeller <ulf@fitug.de>]
10527 *) Add support for native Solaris shared libraries. Configure
10528 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10529 if someone would make that last step automatic.
10530 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10532 *) ctx_size was not built with the right compiler during "make links". Fixed.
10535 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10536 except NULL ciphers". This means the default cipher list will no longer
10537 enable NULL ciphers. They need to be specifically enabled e.g. with
10538 the string "DEFAULT:eNULL".
10541 *) Fix to RSA private encryption routines: if p < q then it would
10542 occasionally produce an invalid result. This will only happen with
10543 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10546 *) Be less restrictive and allow also `perl util/perlpath.pl
10547 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10548 because this way one can also use an interpreter named `perl5' (which is
10549 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10550 installed as `perl').
10551 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10553 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10554 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10556 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10557 advapi32.lib to Win32 build and change the pem test comparison
10558 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10559 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10560 and crypto/des/ede_cbcm_enc.c.
10563 *) DES quad checksum was broken on big-endian architectures. Fixed.
10566 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10567 Win32 test batch file so it (might) work again. The Win32 test batch file
10568 is horrible: I feel ill....
10571 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10572 in e_os.h. Audit of header files to check ANSI and non ANSI
10573 sections: 10 functions were absent from non ANSI section and not exported
10574 from Windows DLLs. Fixed up libeay.num for new functions.
10577 *) Make `openssl version' output lines consistent.
10578 [Ralf S. Engelschall]
10580 *) Fix Win32 symbol export lists for BIO functions: Added
10581 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10582 to ms/libeay{16,32}.def.
10583 [Ralf S. Engelschall]
10585 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10586 fine under Unix and passes some trivial tests I've now added. But the
10587 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10588 added to make sure no one expects that this stuff really works in the
10589 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10590 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10592 [Ralf S. Engelschall]
10594 *) Fix the generation of two part addresses in perl.
10595 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10597 *) Add config entry for Linux on MIPS.
10598 [John Tobey <jtobey@channel1.com>]
10600 *) Make links whenever Configure is run, unless we are on Windoze.
10603 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10604 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10608 *) Add a useful kludge to allow package maintainers to specify compiler and
10609 other platforms details on the command line without having to patch the
10610 Configure script everytime: One now can use ``perl Configure
10611 <id>:<details>'', i.e. platform ids are allowed to have details appended
10612 to them (separated by colons). This is treated as there would be a static
10613 pre-configured entry in Configure's %table under key <id> with value
10614 <details> and ``perl Configure <id>'' is called. So, when you want to
10615 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10616 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10617 now, which overrides the FreeBSD-elf entry on-the-fly.
10618 [Ralf S. Engelschall]
10620 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10623 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10624 on the `perl Configure ...' command line. This way one can compile
10625 OpenSSL libraries with Position Independent Code (PIC) which is needed
10626 for linking it into DSOs.
10627 [Ralf S. Engelschall]
10629 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10633 *) Cleaned up the LICENSE document: The official contact for any license
10634 questions now is the OpenSSL core team under openssl-core@openssl.org.
10635 And add a paragraph about the dual-license situation to make sure people
10636 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10637 to the OpenSSL toolkit.
10638 [Ralf S. Engelschall]
10640 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10641 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10642 Additionally cleaned up the `make links' target: Remove unnecessary
10643 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10644 to speed processing and no longer clutter the display with confusing
10645 stuff. Instead only the actually done links are displayed.
10646 [Ralf S. Engelschall]
10648 *) Permit null encryption ciphersuites, used for authentication only. It used
10649 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10650 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10654 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10655 signed attributes when verifying signatures (this would break them),
10656 the detached data encoding was wrong and public keys obtained using
10657 X509_get_pubkey() weren't freed.
10660 *) Add text documentation for the BUFFER functions. Also added a work around
10661 to a Win95 console bug. This was triggered by the password read stuff: the
10662 last character typed gets carried over to the next fread(). If you were
10663 generating a new cert request using 'req' for example then the last
10664 character of the passphrase would be CR which would then enter the first
10668 *) Added the new `Includes OpenSSL Cryptography Software' button as
10669 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10670 button and can be used by applications based on OpenSSL to show the
10671 relationship to the OpenSSL project.
10672 [Ralf S. Engelschall]
10674 *) Remove confusing variables in function signatures in files
10675 ssl/ssl_lib.c and ssl/ssl.h.
10676 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10678 *) Don't install bss_file.c under PREFIX/include/
10679 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10681 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10682 functions that return function pointers and has support for NT specific
10683 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10684 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10685 unsigned to signed types: this was killing the Win32 compile.
10688 *) Add new certificate file to stack functions,
10689 SSL_add_dir_cert_subjects_to_stack() and
10690 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10691 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10692 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10693 This means that Apache-SSL and similar packages don't have to mess around
10694 to add as many CAs as they want to the preferred list.
10697 *) Experiment with doxygen documentation. Currently only partially applied to
10699 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10700 openssl.doxy as the configuration file.
10703 *) Get rid of remaining C++-style comments which strict C compilers hate.
10704 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10706 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10707 compiled in by default: it has problems with large keys.
10710 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10711 DH private keys and/or callback functions which directly correspond to
10712 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10713 is needed for applications which have to configure certificates on a
10714 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10716 For the RSA certificate situation is makes no difference, but
10717 for the DSA certificate situation this fixes the "no shared cipher"
10718 problem where the OpenSSL cipher selection procedure failed because the
10719 temporary keys were not overtaken from the context and the API provided
10720 no way to reconfigure them.
10721 The new functions now let applications reconfigure the stuff and they
10722 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10723 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10724 non-public-API function ssl_cert_instantiate() is used as a helper
10725 function and also to reduce code redundancy inside ssl_rsa.c.
10726 [Ralf S. Engelschall]
10728 *) Move s_server -dcert and -dkey options out of the undocumented feature
10729 area because they are useful for the DSA situation and should be
10730 recognized by the users.
10731 [Ralf S. Engelschall]
10733 *) Fix the cipher decision scheme for export ciphers: the export bits are
10734 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10735 SSL_EXP_MASK. So, the original variable has to be used instead of the
10736 already masked variable.
10737 [Richard Levitte <levitte@stacken.kth.se>]
10739 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10740 [Richard Levitte <levitte@stacken.kth.se>]
10742 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10743 from `int' to `unsigned int' because it's a length and initialized by
10744 EVP_DigestFinal() which expects an `unsigned int *'.
10745 [Richard Levitte <levitte@stacken.kth.se>]
10747 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10748 script. Instead use the usual Shell->Perl transition trick.
10749 [Ralf S. Engelschall]
10751 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10752 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10753 -noout -modulus' as it's already the case for `openssl rsa -noout
10754 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10755 currently the public key is printed (a decision which was already done by
10756 `openssl dsa -modulus' in the past) which serves a similar purpose.
10757 Additionally the NO_RSA no longer completely removes the whole -modulus
10758 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10760 [Ralf S. Engelschall]
10762 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10763 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10764 [Arne Ansper <arne@ats.cyber.ee>]
10766 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10767 to be added. Now both 'req' and 'ca' can use new objects defined in the
10771 *) Add cool BIO that does syslog (or event log on NT).
10772 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10774 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10775 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10776 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10777 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10780 *) Add preliminary config info for new extension code.
10783 *) Make RSA_NO_PADDING really use no padding.
10784 [Ulf Moeller <ulf@fitug.de>]
10786 *) Generate errors when private/public key check is done.
10789 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10790 for some CRL extensions and new objects added.
10793 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10794 key usage extension and fuller support for authority key id.
10797 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10798 padding method for RSA, which is recommended for new applications in PKCS
10799 #1 v2.0 (RFC 2437, October 1998).
10800 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10801 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10802 against Bleichbacher's attack on RSA.
10803 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10806 *) Updates to the new SSL compression code
10807 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10809 *) Fix so that the version number in the master secret, when passed
10810 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10811 (because the server will not accept higher), that the version number
10812 is 0x03,0x01, not 0x03,0x00
10813 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10815 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10816 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10817 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10820 *) Support for RAW extensions where an arbitrary extension can be
10821 created by including its DER encoding. See apps/openssl.cnf for
10825 *) Make sure latest Perl versions don't interpret some generated C array
10826 code as Perl array code in the crypto/err/err_genc.pl script.
10827 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10829 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10830 not many people have the assembler. Various Win32 compilation fixes and
10831 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10832 build instructions.
10835 *) Modify configure script 'Configure' to automatically create crypto/date.h
10836 file under Win32 and also build pem.h from pem.org. New script
10837 util/mkfiles.pl to create the MINFO file on environments that can't do a
10838 'make files': perl util/mkfiles.pl >MINFO should work.
10841 *) Major rework of DES function declarations, in the pursuit of correctness
10842 and purity. As a result, many evil casts evaporated, and some weirdness,
10843 too. You may find this causes warnings in your code. Zapping your evil
10844 casts will probably fix them. Mostly.
10847 *) Fix for a typo in asn1.h. Bug fix to object creation script
10848 obj_dat.pl. It considered a zero in an object definition to mean
10849 "end of object": none of the objects in objects.h have any zeros
10850 so it wasn't spotted.
10851 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10853 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10854 Masking (CBCM). In the absence of test vectors, the best I have been able
10855 to do is check that the decrypt undoes the encrypt, so far. Send me test
10856 vectors if you have them.
10859 *) Correct calculation of key length for export ciphers (too much space was
10860 allocated for null ciphers). This has not been tested!
10863 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10864 message is now correct (it understands "crypto" and "ssl" on its
10865 command line). There is also now an "update" option. This will update
10866 the util/ssleay.num and util/libeay.num files with any new functions.
10868 perl util/mkdef.pl crypto ssl update
10869 it will update them.
10872 *) Overhauled the Perl interface (perl/*):
10873 - ported BN stuff to OpenSSL's different BN library
10874 - made the perl/ source tree CVS-aware
10875 - renamed the package from SSLeay to OpenSSL (the files still contain
10876 their history because I've copied them in the repository)
10877 - removed obsolete files (the test scripts will be replaced
10878 by better Test::Harness variants in the future)
10879 [Ralf S. Engelschall]
10881 *) First cut for a very conservative source tree cleanup:
10882 1. merge various obsolete readme texts into doc/ssleay.txt
10883 where we collect the old documents and readme texts.
10884 2. remove the first part of files where I'm already sure that we no
10885 longer need them because of three reasons: either they are just temporary
10886 files which were left by Eric or they are preserved original files where
10887 I've verified that the diff is also available in the CVS via "cvs diff
10888 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10889 the crypto/md/ stuff).
10890 [Ralf S. Engelschall]
10892 *) More extension code. Incomplete support for subject and issuer alt
10893 name, issuer and authority key id. Change the i2v function parameters
10894 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10895 what that's for :-) Fix to ASN1 macro which messed up
10896 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10899 *) Preliminary support for ENUMERATED type. This is largely copied from the
10903 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10904 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10906 *) Make sure `make rehash' target really finds the `openssl' program.
10907 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10909 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10910 like to hear about it if this slows down other processors.
10913 *) Add CygWin32 platform information to Configure script.
10914 [Alan Batie <batie@aahz.jf.intel.com>]
10916 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10917 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10919 *) New program nseq to manipulate netscape certificate sequences
10922 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10926 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10927 but the BN code had some problems that would cause failures when
10928 doing certificate verification and some other functions.
10929 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10931 *) Add ASN1 and PEM code to support netscape certificate sequences.
10934 *) Add ASN1 and PEM code to support netscape certificate sequences.
10937 *) Add several PKIX and private extended key usage OIDs.
10940 *) Modify the 'ca' program to handle the new extension code. Modify
10941 openssl.cnf for new extension format, add comments.
10944 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10945 and add a sample to openssl.cnf so req -x509 now adds appropriate
10949 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10950 error code, add initial support to X509_print() and x509 application.
10953 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10954 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10955 stuff is currently isolated and isn't even compiled yet.
10958 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10959 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10960 Removed the versions check from X509 routines when loading extensions:
10961 this allows certain broken certificates that don't set the version
10962 properly to be processed.
10965 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10966 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10967 can still be regenerated with "make depend".
10970 *) Spelling mistake in C version of CAST-128.
10971 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10973 *) Changes to the error generation code. The perl script err-code.pl
10974 now reads in the old error codes and retains the old numbers, only
10975 adding new ones if necessary. It also only changes the .err files if new
10976 codes are added. The makefiles have been modified to only insert errors
10977 when needed (to avoid needlessly modifying header files). This is done
10978 by only inserting errors if the .err file is newer than the auto generated
10979 C file. To rebuild all the error codes from scratch (the old behaviour)
10980 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10981 or delete all the .err files.
10984 *) CAST-128 was incorrectly implemented for short keys. The C version has
10985 been fixed, but is untested. The assembler versions are also fixed, but
10986 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10987 to regenerate it if needed.
10988 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10989 Hagino <itojun@kame.net>]
10991 *) File was opened incorrectly in randfile.c.
10992 [Ulf Möller <ulf@fitug.de>]
10994 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10995 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10996 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10997 al: it's just almost always a UTCTime. Note this patch adds new error
10998 codes so do a "make errors" if there are problems.
11001 *) Correct Linux 1 recognition in config.
11002 [Ulf Möller <ulf@fitug.de>]
11004 *) Remove pointless MD5 hash when using DSA keys in ca.
11005 [Anonymous <nobody@replay.com>]
11007 *) Generate an error if given an empty string as a cert directory. Also
11008 generate an error if handed NULL (previously returned 0 to indicate an
11009 error, but didn't set one).
11010 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11012 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11015 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11016 parameters. This was causing a warning which killed off the Win32 compile.
11019 *) Remove C++ style comments from crypto/bn/bn_local.h.
11020 [Neil Costigan <neil.costigan@celocom.com>]
11022 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11023 based on a text string, looking up short and long names and finally
11024 "dot" format. The "dot" format stuff didn't work. Added new function
11025 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11026 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11027 OID is not part of the table.
11030 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11031 X509_LOOKUP_by_alias().
11034 *) Sort openssl functions by name.
11037 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11038 encryption from sample DSA keys (in case anyone is interested the password
11042 *) Make _all_ *_free functions accept a NULL pointer.
11043 [Frans Heymans <fheymans@isaserver.be>]
11045 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11047 [Anonymous <nobody@replay.com>]
11049 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11050 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11052 *) Don't blow it for numeric -newkey arguments to apps/req.
11053 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11055 *) Temp key "for export" tests were wrong in s3_srvr.c.
11056 [Anonymous <nobody@replay.com>]
11058 *) Add prototype for temp key callback functions
11059 SSL_CTX_set_tmp_{rsa,dh}_callback().
11062 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11063 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11066 *) X509_name_add_entry() freed the wrong thing after an error.
11067 [Arne Ansper <arne@ats.cyber.ee>]
11069 *) rsa_eay.c would attempt to free a NULL context.
11070 [Arne Ansper <arne@ats.cyber.ee>]
11072 *) BIO_s_socket() had a broken should_retry() on Windoze.
11073 [Arne Ansper <arne@ats.cyber.ee>]
11075 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11076 [Arne Ansper <arne@ats.cyber.ee>]
11078 *) Make sure the already existing X509_STORE->depth variable is initialized
11079 in X509_STORE_new(), but document the fact that this variable is still
11080 unused in the certificate verification process.
11081 [Ralf S. Engelschall]
11083 *) Fix the various library and apps files to free up pkeys obtained from
11084 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11087 *) Fix reference counting in X509_PUBKEY_get(). This makes
11088 demos/maurice/example2.c work, amongst others, probably.
11089 [Steve Henson and Ben Laurie]
11091 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11092 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11093 are no longer created. This way we have a single and consistent command
11094 line interface `openssl <command>', similar to `cvs <command>'.
11095 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11097 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11098 BIT STRING wrapper always have zero unused bits.
11101 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11104 *) Make the top-level INSTALL documentation easier to understand.
11107 *) Makefiles updated to exit if an error occurs in a sub-directory
11108 make (including if user presses ^C) [Paul Sutton]
11110 *) Make Montgomery context stuff explicit in RSA data structure.
11113 *) Fix build order of pem and err to allow for generated pem.h.
11116 *) Fix renumbering bug in X509_NAME_delete_entry().
11119 *) Enhanced the err-ins.pl script so it makes the error library number
11120 global and can add a library name. This is needed for external ASN1 and
11121 other error libraries.
11124 *) Fixed sk_insert which never worked properly.
11127 *) Fix ASN1 macros so they can handle indefinite length constructed
11128 EXPLICIT tags. Some non standard certificates use these: they can now
11132 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11133 into a single doc/ssleay.txt bundle. This way the information is still
11134 preserved but no longer messes up this directory. Now it's new room for
11135 the new set of documentation files.
11136 [Ralf S. Engelschall]
11138 *) SETs were incorrectly DER encoded. This was a major pain, because they
11139 shared code with SEQUENCEs, which aren't coded the same. This means that
11140 almost everything to do with SETs or SEQUENCEs has either changed name or
11141 number of arguments.
11142 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11144 *) Fix test data to work with the above.
11147 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11148 was already fixed by Eric for 0.9.1 it seems.
11149 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11151 *) Autodetect FreeBSD3.
11154 *) Fix various bugs in Configure. This affects the following platforms:
11158 unixware-2.0-pentium
11162 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11163 before they are needed.
11166 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11170 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11172 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11173 changed SSLeay to OpenSSL in version strings.
11174 [Ralf S. Engelschall]
11176 *) Some fixups to the top-level documents.
11179 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11180 because the symlink to include/ was missing.
11181 [Ralf S. Engelschall]
11183 *) Incorporated the popular no-RSA/DSA-only patches
11184 which allow to compile a RSA-free SSLeay.
11185 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11187 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11188 when "ssleay" is still not found.
11189 [Ralf S. Engelschall]
11191 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11192 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11194 *) Updated the README file.
11195 [Ralf S. Engelschall]
11197 *) Added various .cvsignore files in the CVS repository subdirs
11198 to make a "cvs update" really silent.
11199 [Ralf S. Engelschall]
11201 *) Recompiled the error-definition header files and added
11202 missing symbols to the Win32 linker tables.
11203 [Ralf S. Engelschall]
11205 *) Cleaned up the top-level documents;
11206 o new files: CHANGES and LICENSE
11207 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11208 o merged COPYRIGHT into LICENSE
11209 o removed obsolete TODO file
11210 o renamed MICROSOFT to INSTALL.W32
11211 [Ralf S. Engelschall]
11213 *) Removed dummy files from the 0.9.1b source tree:
11214 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11215 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11216 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11217 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11218 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11219 [Ralf S. Engelschall]
11221 *) Added various platform portability fixes.
11224 *) The Genesis of the OpenSSL rpject:
11225 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11226 Young and Tim J. Hudson created while they were working for C2Net until
11228 [The OpenSSL Project]
11231 Changes between 0.9.0b and 0.9.1b [not released]
11233 *) Updated a few CA certificates under certs/
11236 *) Changed some BIGNUM api stuff.
11239 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11240 DGUX x86, Linux Alpha, etc.
11243 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11244 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11248 *) Add -strparse option to asn1pars program which parses nested
11250 [Dr Stephen Henson <shenson@bigfoot.com>]
11252 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11255 *) DSA fix for "ca" program.
11258 *) Added "-genkey" option to "dsaparam" program.
11261 *) Added RIPE MD160 (rmd160) message digest.
11264 *) Added -a (all) option to "ssleay version" command.
11267 *) Added PLATFORM define which is the id given to Configure.
11270 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11273 *) Extended the ASN.1 parser routines.
11276 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11279 *) Added a BN_CTX to the BN library.
11282 *) Fixed the weak key values in DES library
11285 *) Changed API in EVP library for cipher aliases.
11288 *) Added support for RC2/64bit cipher.
11291 *) Converted the lhash library to the crypto/mem.c functions.
11294 *) Added more recognized ASN.1 object ids.
11297 *) Added more RSA padding checks for SSL/TLS.
11300 *) Added BIO proxy/filter functionality.
11303 *) Added extra_certs to SSL_CTX which can be used
11304 send extra CA certificates to the client in the CA cert chain sending
11305 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11308 *) Now Fortezza is denied in the authentication phase because
11309 this is key exchange mechanism is not supported by SSLeay at all.
11312 *) Additional PKCS1 checks.
11315 *) Support the string "TLSv1" for all TLS v1 ciphers.
11318 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11319 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11322 *) Fixed a few memory leaks.
11325 *) Fixed various code and comment typos.
11328 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11329 bytes sent in the client random.
11330 [Edward Bishop <ebishop@spyglass.com>]