5 Changes between 1.0.1 and 1.0.2 [xx XXX xxxx]
7 *) Transparently support X9.42 DH parameters when calling
8 PEM_read_bio_DHparameters. This means existing applications can handle
9 the new parameter format automatically.
12 *) Initial experimental support for X9.42 DH parameter format: mainly
13 to support use of 'q' parameter for RFC5114 parameters.
16 *) Add DH parameters from RFC5114 including test data to dhtest.
19 *) Support for automatic EC temporary key parameter selection. If enabled
20 the most preferred EC parameters are automatically used instead of
21 hardcoded fixed parameters. Now a server just has to call:
22 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
23 support ECDH and use the most appropriate parameters.
26 *) Enhance and tidy EC curve and point format TLS extension code. Use
27 static structures instead of allocation if default values are used.
28 New ctrls to set curves we wish to support and to retrieve shared curves.
29 Print out shared curves in s_server. New options to s_server and s_client
30 to set list of supported curves.
33 *) New ctrls to retrieve supported signature algorithms and
34 supported curve values as an array of NIDs. Extend openssl utility
35 to print out received values.
38 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
39 between NIDs and the more common NIST names such as "P-256". Enhance
40 ecparam utility and ECC method to recognise the NIST names for curves.
43 *) Enhance SSL/TLS certificate chain handling to support different
44 chains for each certificate instead of one chain in the parent SSL_CTX.
47 *) Support for fixed DH ciphersuite client authentication: where both
48 server and client use DH certificates with common parameters.
51 *) Support for fixed DH ciphersuites: those requiring DH server
55 Changes between 1.0.1 and 1.0.1a [xx XXX xxxx]
57 *) Workarounds for some broken servers that "hang" if a client hello
58 record length exceeds 255 bytes:
60 1. Do not use record version number > TLS 1.0 in initial client
61 hello: some (but not all) hanging servers will now work.
62 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
63 the number of ciphers sent in the client hello. This should be
64 set to an even number, such as 50, for example by passing:
65 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
66 Most broken servers should now work.
67 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
68 TLS 1.2 client support entirely.
70 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
73 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
75 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
76 STRING form instead of a DigestInfo.
79 *) The format used for MDC2 RSA signatures is inconsistent between EVP
80 and the RSA_sign/RSA_verify functions. This was made more apparent when
81 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
82 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
83 the correct format in RSA_verify so both forms transparently work.
86 *) Some servers which support TLS 1.0 can choke if we initially indicate
87 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
88 encrypted premaster secret. As a workaround use the maximum pemitted
89 client version in client hello, this should keep such servers happy
90 and still work with previous versions of OpenSSL.
93 *) Add support for TLS/DTLS heartbeats.
94 [Robin Seggelmann <seggelmann@fh-muenster.de>]
96 *) Add support for SCTP.
97 [Robin Seggelmann <seggelmann@fh-muenster.de>]
99 *) Improved PRNG seeding for VOS.
100 [Paul Green <Paul.Green@stratus.com>]
102 *) Extensive assembler packs updates, most notably:
104 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
105 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
106 - x86_64: bit-sliced AES implementation;
107 - ARM: NEON support, contemporary platforms optimizations;
108 - s390x: z196 support;
109 - *: GHASH and GF(2^m) multiplication implementations;
113 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
114 (removal of unnecessary code)
115 [Peter Sylvester <peter.sylvester@edelweb.fr>]
117 *) Add TLS key material exporter from RFC 5705.
120 *) Add DTLS-SRTP negotiation from RFC 5764.
123 *) Add Next Protocol Negotiation,
124 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
125 disabled with a no-npn flag to config or Configure. Code donated
127 [Adam Langley <agl@google.com> and Ben Laurie]
129 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
130 NIST-P256, NIST-P521, with constant-time single point multiplication on
131 typical inputs. Compiler support for the nonstandard type __uint128_t is
132 required to use this (present in gcc 4.4 and later, for 64-bit builds).
133 Code made available under Apache License version 2.0.
135 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
136 line to include this in your build of OpenSSL, and run "make depend" (or
137 "make update"). This enables the following EC_METHODs:
139 EC_GFp_nistp224_method()
140 EC_GFp_nistp256_method()
141 EC_GFp_nistp521_method()
143 EC_GROUP_new_by_curve_name() will automatically use these (while
144 EC_GROUP_new_curve_GFp() currently prefers the more flexible
146 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
148 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
149 all platforms. Move ssize_t definition from e_os.h to the public
150 header file e_os2.h as it now appears in public header file cms.h
153 *) New -sigopt option to the ca, req and x509 utilities. Additional
154 signature parameters can be passed using this option and in
158 *) Add RSA PSS signing function. This will generate and set the
159 appropriate AlgorithmIdentifiers for PSS based on those in the
160 corresponding EVP_MD_CTX structure. No application support yet.
163 *) Support for companion algorithm specific ASN1 signing routines.
164 New function ASN1_item_sign_ctx() signs a pre-initialised
165 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
166 the appropriate parameters.
169 *) Add new algorithm specific ASN1 verification initialisation function
170 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
171 handling will be the same no matter what EVP_PKEY_METHOD is used.
172 Add a PSS handler to support verification of PSS signatures: checked
173 against a number of sample certificates.
176 *) Add signature printing for PSS. Add PSS OIDs.
177 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
179 *) Add algorithm specific signature printing. An individual ASN1 method
180 can now print out signatures instead of the standard hex dump.
182 More complex signatures (e.g. PSS) can print out more meaningful
183 information. Include DSA version that prints out the signature
187 *) Password based recipient info support for CMS library: implementing
191 *) Split password based encryption into PBES2 and PBKDF2 functions. This
192 neatly separates the code into cipher and PBE sections and is required
193 for some algorithms that split PBES2 into separate pieces (such as
197 *) Session-handling fixes:
198 - Fix handling of connections that are resuming with a session ID,
199 but also support Session Tickets.
200 - Fix a bug that suppressed issuing of a new ticket if the client
201 presented a ticket with an expired session.
202 - Try to set the ticket lifetime hint to something reasonable.
203 - Make tickets shorter by excluding irrelevant information.
204 - On the client side, don't ignore renewed tickets.
205 [Adam Langley, Bodo Moeller (Google)]
207 *) Fix PSK session representation.
210 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
212 This work was sponsored by Intel.
215 *) Add GCM support to TLS library. Some custom code is needed to split
216 the IV between the fixed (from PRF) and explicit (from TLS record)
217 portions. This adds all GCM ciphersuites supported by RFC5288 and
218 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
219 add a special AESGCM string for GCM only.
222 *) Expand range of ctrls for AES GCM. Permit setting invocation
223 field on decrypt and retrieval of invocation field only on encrypt.
226 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
227 As required by RFC5289 these ciphersuites cannot be used if for
228 versions of TLS earlier than 1.2.
231 *) For FIPS capable OpenSSL interpret a NULL default public key method
232 as unset and return the appopriate default but do *not* set the default.
233 This means we can return the appopriate method in applications that
234 swicth between FIPS and non-FIPS modes.
237 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
238 ENGINE is used then we cannot handle that in the FIPS module so we
239 keep original code iff non-FIPS operations are allowed.
242 *) Add -attime option to openssl utilities.
243 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
245 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
248 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
249 FIPS EC methods unconditionally for now.
252 *) New build option no-ec2m to disable characteristic 2 code.
255 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
256 all cases can be covered as some introduce binary incompatibilities.
259 *) Redirect RSA operations to FIPS module including keygen,
260 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
263 *) Add similar low level API blocking to ciphers.
266 *) Low level digest APIs are not approved in FIPS mode: any attempt
267 to use these will cause a fatal error. Applications that *really* want
268 to use them can use the private_* version instead.
271 *) Redirect cipher operations to FIPS module for FIPS builds.
274 *) Redirect digest operations to FIPS module for FIPS builds.
277 *) Update build system to add "fips" flag which will link in fipscanister.o
278 for static and shared library builds embedding a signature if needed.
281 *) Output TLS supported curves in preference order instead of numerical
282 order. This is currently hardcoded for the highest order curves first.
283 This should be configurable so applications can judge speed vs strength.
286 *) Add TLS v1.2 server support for client authentication.
289 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
293 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
294 FIPS modules versions.
297 *) Add TLS v1.2 client side support for client authentication. Keep cache
298 of handshake records longer as we don't know the hash algorithm to use
299 until after the certificate request message is received.
302 *) Initial TLS v1.2 client support. Add a default signature algorithms
303 extension including all the algorithms we support. Parse new signature
304 format in client key exchange. Relax some ECC signing restrictions for
305 TLS v1.2 as indicated in RFC5246.
308 *) Add server support for TLS v1.2 signature algorithms extension. Switch
309 to new signature format when needed using client digest preference.
310 All server ciphersuites should now work correctly in TLS v1.2. No client
311 support yet and no support for client certificates.
314 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
315 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
316 ciphersuites. At present only RSA key exchange ciphersuites work with
317 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
318 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
319 and version checking.
322 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
323 with this defined it will not be affected by any changes to ssl internal
324 structures. Add several utility functions to allow openssl application
325 to work with OPENSSL_NO_SSL_INTERN defined.
329 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
331 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
334 *) Permit abbreviated handshakes when renegotiating using the function
335 SSL_renegotiate_abbreviated().
336 [Robin Seggelmann <seggelmann@fh-muenster.de>]
338 *) Add call to ENGINE_register_all_complete() to
339 ENGINE_load_builtin_engines(), so some implementations get used
340 automatically instead of needing explicit application support.
343 *) Add support for TLS key exporter as described in RFC5705.
344 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
346 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
347 a few changes are required:
349 Add SSL_OP_NO_TLSv1_1 flag.
351 Update version checking logic to handle version 1.1.
352 Add explicit IV handling (ported from DTLS code).
353 Add command line options to s_client/s_server.
356 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
358 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
359 in CMS and PKCS7 code. When RSA decryption fails use a random key for
360 content decryption and always return the same error. Note: this attack
361 needs on average 2^20 messages so it only affects automated senders. The
362 old behaviour can be reenabled in the CMS code by setting the
363 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
364 an MMA defence is not necessary.
365 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
366 this issue. (CVE-2012-0884)
369 *) Fix CVE-2011-4619: make sure we really are receiving a
370 client hello before rejecting multiple SGC restarts. Thanks to
371 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
374 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
376 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
377 Thanks to Antonio Martin, Enterprise Secure Access Research and
378 Development, Cisco Systems, Inc. for discovering this bug and
379 preparing a fix. (CVE-2012-0050)
382 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
384 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
385 of the Vaudenay padding oracle attack on CBC mode encryption
386 which enables an efficient plaintext recovery attack against
387 the OpenSSL implementation of DTLS. Their attack exploits timing
388 differences arising during decryption processing. A research
389 paper describing this attack can be found at:
390 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
391 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
392 Security Group at Royal Holloway, University of London
393 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
394 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
395 for preparing the fix. (CVE-2011-4108)
396 [Robin Seggelmann, Michael Tuexen]
398 *) Clear bytes used for block padding of SSL 3.0 records.
400 [Adam Langley (Google)]
402 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
403 Kadianakis <desnacked@gmail.com> for discovering this issue and
404 Adam Langley for preparing the fix. (CVE-2011-4619)
405 [Adam Langley (Google)]
407 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
408 [Andrey Kulikov <amdeich@gmail.com>]
410 *) Prevent malformed RFC3779 data triggering an assertion failure.
411 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
412 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
413 [Rob Austein <sra@hactrn.net>]
415 *) Improved PRNG seeding for VOS.
416 [Paul Green <Paul.Green@stratus.com>]
418 *) Fix ssl_ciph.c set-up race.
419 [Adam Langley (Google)]
421 *) Fix spurious failures in ecdsatest.c.
422 [Emilia Käsper (Google)]
424 *) Fix the BIO_f_buffer() implementation (which was mixing different
425 interpretations of the '..._len' fields).
426 [Adam Langley (Google)]
428 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
429 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
430 threads won't reuse the same blinding coefficients.
432 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
433 lock to call BN_BLINDING_invert_ex, and avoids one use of
434 BN_BLINDING_update for each BN_BLINDING structure (previously,
435 the last update always remained unused).
436 [Emilia Käsper (Google)]
438 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
439 [Bob Buckholz (Google)]
441 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
443 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
444 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
445 [Kaspar Brand <ossl@velox.ch>]
447 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
448 for multi-threaded use of ECDH. (CVE-2011-3210)
449 [Adam Langley (Google)]
451 *) Fix x509_name_ex_d2i memory leak on bad inputs.
454 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
455 signature public key algorithm by using OID xref utilities instead.
456 Before this you could only use some ECC ciphersuites with SHA1 only.
459 *) Add protection against ECDSA timing attacks as mentioned in the paper
460 by Billy Bob Brumley and Nicola Tuveri, see:
462 http://eprint.iacr.org/2011/232.pdf
464 [Billy Bob Brumley and Nicola Tuveri]
466 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
468 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
469 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
471 *) Fix bug in string printing code: if *any* escaping is enabled we must
472 escape the escape character (backslash) or the resulting string is
476 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
478 *) Disable code workaround for ancient and obsolete Netscape browsers
479 and servers: an attacker can use it in a ciphersuite downgrade attack.
480 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
483 *) Fixed J-PAKE implementation error, originally discovered by
484 Sebastien Martini, further info and confirmation from Stefan
485 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
488 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
490 *) Fix extension code to avoid race conditions which can result in a buffer
491 overrun vulnerability: resumed sessions must not be modified as they can
492 be shared by multiple threads. CVE-2010-3864
495 *) Fix WIN32 build system to correctly link an ENGINE directory into
499 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
501 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
503 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
505 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
507 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
508 context. The operation can be customised via the ctrl mechanism in
509 case ENGINEs want to include additional functionality.
512 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
515 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
516 output hashes compatible with older versions of OpenSSL.
517 [Willy Weisz <weisz@vcpc.univie.ac.at>]
519 *) Fix compression algorithm handling: if resuming a session use the
520 compression algorithm of the resumed session instead of determining
521 it from client hello again. Don't allow server to change algorithm.
524 *) Add load_crls() function to apps tidying load_certs() too. Add option
525 to verify utility to allow additional CRLs to be included.
528 *) Update OCSP request code to permit adding custom headers to the request:
529 some responders need this.
532 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
534 [Julia Lawall <julia@diku.dk>]
536 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
537 needlessly dereferenced structures, used obsolete functions and
538 didn't handle all updated verify codes correctly.
541 *) Disable MD2 in the default configuration.
544 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
545 indicate the initial BIO being pushed or popped. This makes it possible
546 to determine whether the BIO is the one explicitly called or as a result
547 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
548 it handles reference counts correctly and doesn't zero out the I/O bio
549 when it is not being explicitly popped. WARNING: applications which
550 included workarounds for the old buggy behaviour will need to be modified
551 or they could free up already freed BIOs.
554 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
555 renaming to all platforms (within the 0.9.8 branch, this was
556 done conditionally on Netware platforms to avoid a name clash).
557 [Guenter <lists@gknw.net>]
559 *) Add ECDHE and PSK support to DTLS.
560 [Michael Tuexen <tuexen@fh-muenster.de>]
562 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
566 *) Add "missing" function EVP_MD_flags() (without this the only way to
567 retrieve a digest flags is by accessing the structure directly. Update
568 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
569 or cipher is registered as in the "from" argument. Print out all
570 registered digests in the dgst usage message instead of manually
571 attempting to work them out.
574 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
575 this allows the use of compression and extensions. Change default cipher
576 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
577 by default unless an application cipher string requests it.
580 *) Alter match criteria in PKCS12_parse(). It used to try to use local
581 key ids to find matching certificates and keys but some PKCS#12 files
582 don't follow the (somewhat unwritten) rules and this strategy fails.
583 Now just gather all certificates together and the first private key
584 then look for the first certificate that matches the key.
587 *) Support use of registered digest and cipher names for dgst and cipher
588 commands instead of having to add each one as a special case. So now
595 openssl dgst -sha256 foo
597 and this works for ENGINE based algorithms too.
601 *) Update Gost ENGINE to support parameter files.
602 [Victor B. Wagner <vitus@cryptocom.ru>]
604 *) Support GeneralizedTime in ca utility.
605 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
607 *) Enhance the hash format used for certificate directory links. The new
608 form uses the canonical encoding (meaning equivalent names will work
609 even if they aren't identical) and uses SHA1 instead of MD5. This form
610 is incompatible with the older format and as a result c_rehash should
611 be used to rebuild symbolic links.
614 *) Make PKCS#8 the default write format for private keys, replacing the
615 traditional format. This form is standardised, more secure and doesn't
616 include an implicit MD5 dependency.
619 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
620 committed to OpenSSL should pass this lot as a minimum.
623 *) Add session ticket override functionality for use by EAP-FAST.
624 [Jouni Malinen <j@w1.fi>]
626 *) Modify HMAC functions to return a value. Since these can be implemented
627 in an ENGINE errors can occur.
630 *) Type-checked OBJ_bsearch_ex.
633 *) Type-checked OBJ_bsearch. Also some constification necessitated
634 by type-checking. Still to come: TXT_DB, bsearch(?),
635 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
639 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
640 seconds to a tm structure directly, instead of going through OS
641 specific date routines. This avoids any issues with OS routines such
642 as the year 2038 bug. New *_adj() functions for ASN1 time structures
643 and X509_time_adj_ex() to cover the extended range. The existing
644 X509_time_adj() is still usable and will no longer have any date issues.
647 *) Delta CRL support. New use deltas option which will attempt to locate
648 and search any appropriate delta CRLs available.
650 This work was sponsored by Google.
653 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
654 code and add additional score elements. Validate alternate CRL paths
655 as part of the CRL checking and indicate a new error "CRL path validation
656 error" in this case. Applications wanting additional details can use
657 the verify callback and check the new "parent" field. If this is not
658 NULL CRL path validation is taking place. Existing applications wont
659 see this because it requires extended CRL support which is off by
662 This work was sponsored by Google.
665 *) Support for freshest CRL extension.
667 This work was sponsored by Google.
670 *) Initial indirect CRL support. Currently only supported in the CRLs
671 passed directly and not via lookup. Process certificate issuer
672 CRL entry extension and lookup CRL entries by bother issuer name
673 and serial number. Check and process CRL issuer entry in IDP extension.
675 This work was sponsored by Google.
678 *) Add support for distinct certificate and CRL paths. The CRL issuer
679 certificate is validated separately in this case. Only enabled if
680 an extended CRL support flag is set: this flag will enable additional
681 CRL functionality in future.
683 This work was sponsored by Google.
686 *) Add support for policy mappings extension.
688 This work was sponsored by Google.
691 *) Fixes to pathlength constraint, self issued certificate handling,
692 policy processing to align with RFC3280 and PKITS tests.
694 This work was sponsored by Google.
697 *) Support for name constraints certificate extension. DN, email, DNS
698 and URI types are currently supported.
700 This work was sponsored by Google.
703 *) To cater for systems that provide a pointer-based thread ID rather
704 than numeric, deprecate the current numeric thread ID mechanism and
705 replace it with a structure and associated callback type. This
706 mechanism allows a numeric "hash" to be extracted from a thread ID in
707 either case, and on platforms where pointers are larger than 'long',
708 mixing is done to help ensure the numeric 'hash' is usable even if it
709 can't be guaranteed unique. The default mechanism is to use "&errno"
710 as a pointer-based thread ID to distinguish between threads.
712 Applications that want to provide their own thread IDs should now use
713 CRYPTO_THREADID_set_callback() to register a callback that will call
714 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
716 Note that ERR_remove_state() is now deprecated, because it is tied
717 to the assumption that thread IDs are numeric. ERR_remove_state(0)
718 to free the current thread's error state should be replaced by
719 ERR_remove_thread_state(NULL).
721 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
722 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
723 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
724 application was previously providing a numeric thread callback that
725 was inappropriate for distinguishing threads, then uniqueness might
726 have been obtained with &errno that happened immediately in the
727 intermediate development versions of OpenSSL; this is no longer the
728 case, the numeric thread callback will now override the automatic use
730 [Geoff Thorpe, with help from Bodo Moeller]
732 *) Initial support for different CRL issuing certificates. This covers a
733 simple case where the self issued certificates in the chain exist and
734 the real CRL issuer is higher in the existing chain.
736 This work was sponsored by Google.
739 *) Removed effectively defunct crypto/store from the build.
742 *) Revamp of STACK to provide stronger type-checking. Still to come:
743 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
744 ASN1_STRING, CONF_VALUE.
747 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
748 RAM on SSL connections. This option can save about 34k per idle SSL.
751 *) Revamp of LHASH to provide stronger type-checking. Still to come:
752 STACK, TXT_DB, bsearch, qsort.
755 *) Initial support for Cryptographic Message Syntax (aka CMS) based
756 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
757 support for data, signedData, compressedData, digestedData and
758 encryptedData, envelopedData types included. Scripts to check against
759 RFC4134 examples draft and interop and consistency checks of many
760 content types and variants.
763 *) Add options to enc utility to support use of zlib compression BIO.
766 *) Extend mk1mf to support importing of options and assembly language
767 files from Configure script, currently only included in VC-WIN32.
768 The assembly language rules can now optionally generate the source
769 files from the associated perl scripts.
772 *) Implement remaining functionality needed to support GOST ciphersuites.
773 Interop testing has been performed using CryptoPro implementations.
774 [Victor B. Wagner <vitus@cryptocom.ru>]
776 *) s390x assembler pack.
779 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
783 *) Implement Opaque PRF Input TLS extension as specified in
784 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
785 official specification yet and no extension type assignment by
786 IANA exists, this extension (for now) will have to be explicitly
787 enabled when building OpenSSL by providing the extension number
788 to use. For example, specify an option
790 -DTLSEXT_TYPE_opaque_prf_input=0x9527
792 to the "config" or "Configure" script to enable the extension,
793 assuming extension number 0x9527 (which is a completely arbitrary
794 and unofficial assignment based on the MD5 hash of the Internet
795 Draft). Note that by doing so, you potentially lose
796 interoperability with other TLS implementations since these might
797 be using the same extension number for other purposes.
799 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
800 opaque PRF input value to use in the handshake. This will create
801 an interal copy of the length-'len' string at 'src', and will
802 return non-zero for success.
804 To get more control and flexibility, provide a callback function
807 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
808 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
812 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
815 Callback function 'cb' will be called in handshakes, and is
816 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
817 Argument 'arg' is for application purposes (the value as given to
818 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
819 be provided to the callback function). The callback function
820 has to return non-zero to report success: usually 1 to use opaque
821 PRF input just if possible, or 2 to enforce use of the opaque PRF
822 input. In the latter case, the library will abort the handshake
823 if opaque PRF input is not successfully negotiated.
825 Arguments 'peerinput' and 'len' given to the callback function
826 will always be NULL and 0 in the case of a client. A server will
827 see the client's opaque PRF input through these variables if
828 available (NULL and 0 otherwise). Note that if the server
829 provides an opaque PRF input, the length must be the same as the
830 length of the client's opaque PRF input.
832 Note that the callback function will only be called when creating
833 a new session (session resumption can resume whatever was
834 previously negotiated), and will not be called in SSL 2.0
835 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
836 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
837 for applications that need to enforce opaque PRF input.
841 *) Update ssl code to support digests other than SHA1+MD5 for handshake
844 [Victor B. Wagner <vitus@cryptocom.ru>]
846 *) Add RFC4507 support to OpenSSL. This includes the corrections in
847 RFC4507bis. The encrypted ticket format is an encrypted encoded
848 SSL_SESSION structure, that way new session features are automatically
851 If a client application caches session in an SSL_SESSION structure
852 support is transparent because tickets are now stored in the encoded
855 The SSL_CTX structure automatically generates keys for ticket
856 protection in servers so again support should be possible
857 with no application modification.
859 If a client or server wishes to disable RFC4507 support then the option
860 SSL_OP_NO_TICKET can be set.
862 Add a TLS extension debugging callback to allow the contents of any client
863 or server extensions to be examined.
865 This work was sponsored by Google.
868 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
869 OpenSSL should now compile cleanly on gcc 4.2
870 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
872 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
873 support including streaming MAC support: this is required for GOST
875 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
877 *) Add option -stream to use PKCS#7 streaming in smime utility. New
878 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
879 to output in BER and PEM format.
882 *) Experimental support for use of HMAC via EVP_PKEY interface. This
883 allows HMAC to be handled via the EVP_DigestSign*() interface. The
884 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
885 ENGINE support for HMAC keys which are unextractable. New -mac and
886 -macopt options to dgst utility.
889 *) New option -sigopt to dgst utility. Update dgst to use
890 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
891 alternative signing paramaters such as X9.31 or PSS in the dgst
895 *) Change ssl_cipher_apply_rule(), the internal function that does
896 the work each time a ciphersuite string requests enabling
897 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
898 removing ("!foo+bar") a class of ciphersuites: Now it maintains
899 the order of disabled ciphersuites such that those ciphersuites
900 that most recently went from enabled to disabled not only stay
901 in order with respect to each other, but also have higher priority
902 than other disabled ciphersuites the next time ciphersuites are
905 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
906 the same ciphersuites as with "HIGH" alone, but in a specific
907 order where the PSK ciphersuites come first (since they are the
908 most recently disabled ciphersuites when "HIGH" is parsed).
910 Also, change ssl_create_cipher_list() (using this new
911 funcionality) such that between otherwise identical
912 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
916 *) Change ssl_create_cipher_list() so that it automatically
917 arranges the ciphersuites in reasonable order before starting
918 to process the rule string. Thus, the definition for "DEFAULT"
919 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
920 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
921 This makes it much easier to arrive at a reasonable default order
922 in applications for which anonymous ciphers are OK (meaning
923 that you can't actually use DEFAULT).
924 [Bodo Moeller; suggested by Victor Duchovni]
926 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
927 processing) into multiple integers instead of setting
928 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
929 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
930 (These masks as well as the individual bit definitions are hidden
931 away into the non-exported interface ssl/ssl_locl.h, so this
932 change to the definition of the SSL_CIPHER structure shouldn't
933 affect applications.) This give us more bits for each of these
934 categories, so there is no longer a need to coagulate AES128 and
935 AES256 into a single algorithm bit, and to coagulate Camellia128
936 and Camellia256 into a single algorithm bit, which has led to all
939 Thus, among other things, the kludge introduced in 0.9.7m and
940 0.9.8e for masking out AES256 independently of AES128 or masking
941 out Camellia256 independently of AES256 is not needed here in 0.9.9.
943 With the change, we also introduce new ciphersuite aliases that
944 so far were missing: "AES128", "AES256", "CAMELLIA128", and
948 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
949 Use the leftmost N bytes of the signature input if the input is
950 larger than the prime q (with N being the size in bytes of q).
953 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
954 it yet and it is largely untested.
957 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
960 *) Initial incomplete changes to avoid need for function casts in OpenSSL
961 some compilers (gcc 4.2 and later) reject their use. Safestack is
962 reimplemented. Update ASN1 to avoid use of legacy functions.
965 *) Win32/64 targets are linked with Winsock2.
968 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
969 to external functions. This can be used to increase CRL handling
970 efficiency especially when CRLs are very large by (for example) storing
971 the CRL revoked certificates in a database.
974 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
975 new CRLs added to a directory can be used. New command line option
976 -verify_return_error to s_client and s_server. This causes real errors
977 to be returned by the verify callback instead of carrying on no matter
978 what. This reflects the way a "real world" verify callback would behave.
981 *) GOST engine, supporting several GOST algorithms and public key formats.
982 Kindly donated by Cryptocom.
985 *) Partial support for Issuing Distribution Point CRL extension. CRLs
986 partitioned by DP are handled but no indirect CRL or reason partitioning
987 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
988 selected via a scoring technique which handles IDP and AKID in CRLs.
991 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
992 will ultimately be used for all verify operations: this will remove the
993 X509_STORE dependency on certificate verification and allow alternative
994 lookup methods. X509_STORE based implementations of these two callbacks.
997 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
998 Modify get_crl() to find a valid (unexpired) CRL if possible.
1001 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1002 this would be called X509_CRL_cmp() but that name is already used by
1003 a function that just compares CRL issuer names. Cache several CRL
1004 extensions in X509_CRL structure and cache CRLDP in X509.
1007 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1008 this maps equivalent X509_NAME structures into a consistent structure.
1009 Name comparison can then be performed rapidly using memcmp().
1012 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1016 *) Allow digests to supply their own micalg string for S/MIME type using
1017 the ctrl EVP_MD_CTRL_MICALG.
1020 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1021 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1022 ctrl. It can then customise the structure before and/or after signing
1026 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1027 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1028 to free up any added signature OIDs.
1031 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1032 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1033 digest and cipher tables. New options added to openssl utility:
1034 list-message-digest-algorithms and list-cipher-algorithms.
1037 *) Change the array representation of binary polynomials: the list
1038 of degrees of non-zero coefficients is now terminated with -1.
1039 Previously it was terminated with 0, which was also part of the
1040 value; thus, the array representation was not applicable to
1041 polynomials where t^0 has coefficient zero. This change makes
1042 the array representation useful in a more general context.
1045 *) Various modifications and fixes to SSL/TLS cipher string
1046 handling. For ECC, the code now distinguishes between fixed ECDH
1047 with RSA certificates on the one hand and with ECDSA certificates
1048 on the other hand, since these are separate ciphersuites. The
1049 unused code for Fortezza ciphersuites has been removed.
1051 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1052 (not "ECDHE"). For consistency with the code for DH
1053 certificates, use of ECDH certificates is now considered ECDH
1054 authentication, not RSA or ECDSA authentication (the latter is
1055 merely the CA's signing algorithm and not actively used in the
1058 The temporary ciphersuite alias "ECCdraft" is no longer
1059 available, and ECC ciphersuites are no longer excluded from "ALL"
1060 and "DEFAULT". The following aliases now exist for RFC 4492
1061 ciphersuites, most of these by analogy with the DH case:
1063 kECDHr - ECDH cert, signed with RSA
1064 kECDHe - ECDH cert, signed with ECDSA
1065 kECDH - ECDH cert (signed with either RSA or ECDSA)
1066 kEECDH - ephemeral ECDH
1067 ECDH - ECDH cert or ephemeral ECDH
1073 AECDH - anonymous ECDH
1074 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1078 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1079 Use correct micalg parameters depending on digest(s) in signed message.
1082 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1083 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1086 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1087 an engine to register a method. Add ENGINE lookups for methods and
1088 functional reference processing.
1091 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1092 EVP_{Sign,Verify}* which allow an application to customise the signature
1096 *) New -resign option to smime utility. This adds one or more signers
1097 to an existing PKCS#7 signedData structure. Also -md option to use an
1098 alternative message digest algorithm for signing.
1101 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1102 create PKCS7 structures containing multiple signers. Update smime
1103 application to support multiple signers.
1106 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1110 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1111 Reorganize PBE internals to lookup from a static table using NIDs,
1112 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1113 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1114 PRF which will be automatically used with PBES2.
1117 *) Replace the algorithm specific calls to generate keys in "req" with the
1121 *) Update PKCS#7 enveloped data routines to use new API. This is now
1122 supported by any public key method supporting the encrypt operation. A
1123 ctrl is added to allow the public key algorithm to examine or modify
1124 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1128 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1129 a default digest type to use. In most cases this will be SHA1 but some
1130 algorithms (such as GOST) need to specify an alternative digest. The
1131 return value indicates how strong the prefernce is 1 means optional and
1132 2 is mandatory (that is it is the only supported type). Modify
1133 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1134 use the default md. Update openssl utilities to use the default digest
1135 type for signing if it is not explicitly indicated.
1138 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1139 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1140 signing method from the key type. This effectively removes the link
1141 between digests and public key types.
1144 *) Add an OID cross reference table and utility functions. Its purpose is to
1145 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1146 rsaEncryption. This will allow some of the algorithm specific hackery
1147 needed to use the correct OID to be removed.
1150 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1151 structures for PKCS7_sign(). They are now set up by the relevant public
1155 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1158 *) Add support for key derivation (agreement) in the API, DH method and
1162 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1163 public and private key formats. As a side effect these add additional
1164 command line functionality not previously available: DSA signatures can be
1165 generated and verified using pkeyutl and DH key support and generation in
1170 [Oliver Tappe <zooey@hirschkaefer.de>]
1172 *) New make target "install_html_docs" installs HTML renditions of the
1174 [Oliver Tappe <zooey@hirschkaefer.de>]
1176 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1177 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1178 support key and parameter generation and add initial key generation
1179 functionality for RSA.
1182 *) Add functions for main EVP_PKEY_method operations. The undocumented
1183 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1184 EVP_PKEY_{encrypt,decrypt}_old.
1187 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1188 key API, doesn't do much yet.
1191 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1192 public key algorithms. New option to openssl utility:
1193 "list-public-key-algorithms" to print out info.
1196 *) Implement the Supported Elliptic Curves Extension for
1197 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1200 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1201 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1204 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1205 utilities such as rsa, dsa, dsaparam etc except they process any key
1209 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1210 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1211 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1215 *) Initial support for pluggable public key ASN1.
1216 De-spaghettify the public key ASN1 handling. Move public and private
1217 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1218 algorithm specific handling to a single module within the relevant
1219 algorithm directory. Add functions to allow (near) opaque processing
1220 of public and private key structures.
1223 *) Implement the Supported Point Formats Extension for
1224 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1227 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1228 for the psk identity [hint] and the psk callback functions to the
1229 SSL_SESSION, SSL and SSL_CTX structure.
1232 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1236 SSL_CTX_use_psk_identity_hint
1237 SSL_get_psk_identity_hint
1238 SSL_get_psk_identity
1239 SSL_use_psk_identity_hint
1241 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1243 *) Add RFC 3161 compliant time stamp request creation, response generation
1244 and response verification functionality.
1245 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1247 *) Add initial support for TLS extensions, specifically for the server_name
1248 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1249 have new members for a host name. The SSL data structure has an
1250 additional member SSL_CTX *initial_ctx so that new sessions can be
1251 stored in that context to allow for session resumption, even after the
1252 SSL has been switched to a new SSL_CTX in reaction to a client's
1253 server_name extension.
1255 New functions (subject to change):
1257 SSL_get_servername()
1258 SSL_get_servername_type()
1261 New CTRL codes and macros (subject to change):
1263 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1264 - SSL_CTX_set_tlsext_servername_callback()
1265 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1266 - SSL_CTX_set_tlsext_servername_arg()
1267 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1269 openssl s_client has a new '-servername ...' option.
1271 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1272 '-key2 ...', '-servername_fatal' (subject to change). This allows
1273 testing the HostName extension for a specific single host name ('-cert'
1274 and '-key' remain fallbacks for handshakes without HostName
1275 negotiation). If the unrecogninzed_name alert has to be sent, this by
1276 default is a warning; it becomes fatal with the '-servername_fatal'
1279 [Peter Sylvester, Remy Allais, Christophe Renou]
1281 *) Whirlpool hash implementation is added.
1284 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1285 bn(64,32). Because of instruction set limitations it doesn't have
1286 any negative impact on performance. This was done mostly in order
1287 to make it possible to share assembler modules, such as bn_mul_mont
1288 implementations, between 32- and 64-bit builds without hassle.
1291 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1292 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1296 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1297 dedicated Montgomery multiplication procedure, is introduced.
1298 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1299 "64-bit" performance on certain 32-bit targets.
1302 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1303 in SSL structures. New SSL ctrl to set maximum send fragment size.
1304 Save memory by seeting the I/O buffer sizes dynamically instead of
1305 using the maximum available value.
1308 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1309 in addition to the text details.
1312 *) Very, very preliminary EXPERIMENTAL support for printing of general
1313 ASN1 structures. This currently produces rather ugly output and doesn't
1314 handle several customised structures at all.
1317 *) Integrated support for PVK file format and some related formats such
1318 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1319 these in the 'rsa' and 'dsa' utilities.
1322 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1325 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1326 place for the (very old) "NETSCAPE" format certificates which are now
1327 handled using new ASN1 code equivalents.
1330 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1331 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1332 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1335 *) Modify CRL distribution points extension code to print out previously
1336 unsupported fields. Enhance extension setting code to allow setting of
1340 *) Add print and set support for Issuing Distribution Point CRL extension.
1343 *) Change 'Configure' script to enable Camellia by default.
1346 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1348 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1349 Thanks to Antonio Martin, Enterprise Secure Access Research and
1350 Development, Cisco Systems, Inc. for discovering this bug and
1351 preparing a fix. (CVE-2012-0050)
1354 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1356 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1357 of the Vaudenay padding oracle attack on CBC mode encryption
1358 which enables an efficient plaintext recovery attack against
1359 the OpenSSL implementation of DTLS. Their attack exploits timing
1360 differences arising during decryption processing. A research
1361 paper describing this attack can be found at:
1362 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1363 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1364 Security Group at Royal Holloway, University of London
1365 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1366 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1367 for preparing the fix. (CVE-2011-4108)
1368 [Robin Seggelmann, Michael Tuexen]
1370 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1371 [Ben Laurie, Kasper <ekasper@google.com>]
1373 *) Clear bytes used for block padding of SSL 3.0 records.
1375 [Adam Langley (Google)]
1377 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1378 Kadianakis <desnacked@gmail.com> for discovering this issue and
1379 Adam Langley for preparing the fix. (CVE-2011-4619)
1380 [Adam Langley (Google)]
1382 *) Prevent malformed RFC3779 data triggering an assertion failure.
1383 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1384 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1385 [Rob Austein <sra@hactrn.net>]
1387 *) Fix ssl_ciph.c set-up race.
1388 [Adam Langley (Google)]
1390 *) Fix spurious failures in ecdsatest.c.
1391 [Emilia Käsper (Google)]
1393 *) Fix the BIO_f_buffer() implementation (which was mixing different
1394 interpretations of the '..._len' fields).
1395 [Adam Langley (Google)]
1397 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1398 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1399 threads won't reuse the same blinding coefficients.
1401 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1402 lock to call BN_BLINDING_invert_ex, and avoids one use of
1403 BN_BLINDING_update for each BN_BLINDING structure (previously,
1404 the last update always remained unused).
1405 [Emilia Käsper (Google)]
1407 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1408 for multi-threaded use of ECDH.
1409 [Adam Langley (Google)]
1411 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1414 *) Add protection against ECDSA timing attacks as mentioned in the paper
1415 by Billy Bob Brumley and Nicola Tuveri, see:
1417 http://eprint.iacr.org/2011/232.pdf
1419 [Billy Bob Brumley and Nicola Tuveri]
1421 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1423 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1424 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1426 *) Fix bug in string printing code: if *any* escaping is enabled we must
1427 escape the escape character (backslash) or the resulting string is
1431 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1433 *) Disable code workaround for ancient and obsolete Netscape browsers
1434 and servers: an attacker can use it in a ciphersuite downgrade attack.
1435 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1438 *) Fixed J-PAKE implementation error, originally discovered by
1439 Sebastien Martini, further info and confirmation from Stefan
1440 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1443 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1445 *) Fix extension code to avoid race conditions which can result in a buffer
1446 overrun vulnerability: resumed sessions must not be modified as they can
1447 be shared by multiple threads. CVE-2010-3864
1450 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1453 *) Don't reencode certificate when calculating signature: cache and use
1454 the original encoding instead. This makes signature verification of
1455 some broken encodings work correctly.
1458 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1459 is also one of the inputs.
1460 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1462 *) Don't repeatedly append PBE algorithms to table if they already exist.
1463 Sort table on each new add. This effectively makes the table read only
1464 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1468 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1470 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1473 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1474 access or freeing data twice (CVE-2010-0742)
1475 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1477 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1478 common in certificates and some applications which only call
1479 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1483 Reduce copying into .apps and .test in makevms.com
1484 Don't try to use blank CA certificate in CA.com
1485 Allow use of C files from original directories in maketests.com
1486 [Steven M. Schweda" <sms@antinode.info>]
1488 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1490 *) When rejecting SSL/TLS records due to an incorrect version number, never
1491 update s->server with a new major version number. As of
1492 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1493 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1494 the previous behavior could result in a read attempt at NULL when
1495 receiving specific incorrect SSL/TLS records once record payload
1496 protection is active. (CVE-2010-0740)
1497 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1499 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1500 could be crashed if the relevant tables were not present (e.g. chrooted).
1501 [Tomas Hoger <thoger@redhat.com>]
1503 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1505 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1506 [Martin Olsson, Neel Mehta]
1508 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1509 accommodate for stack sorting, always a write lock!).
1512 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1513 excessive delays in the RAND_poll(): over a minute. As a workaround
1514 include a time check in the inner Heap32Next loop too.
1517 *) The code that handled flushing of data in SSL/TLS originally used the
1518 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1519 the problem outlined in PR#1949. The fix suggested there however can
1520 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1521 of Apache). So instead simplify the code to flush unconditionally.
1522 This should be fine since flushing with no data to flush is a no op.
1525 *) Handle TLS versions 2.0 and later properly and correctly use the
1526 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1527 off ancient servers have a habit of sticking around for a while...
1530 *) Modify compression code so it frees up structures without using the
1531 ex_data callbacks. This works around a problem where some applications
1532 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1533 restarting) then use compression (e.g. SSL with compression) later.
1534 This results in significant per-connection memory leaks and
1535 has caused some security issues including CVE-2008-1678 and
1539 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1540 change when encrypting or decrypting.
1543 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1544 connect and renegotiate with servers which do not support RI.
1545 Until RI is more widely deployed this option is enabled by default.
1548 *) Add "missing" ssl ctrls to clear options and mode.
1551 *) If client attempts to renegotiate and doesn't support RI respond with
1552 a no_renegotiation alert as required by RFC5746. Some renegotiating
1553 TLS clients will continue a connection gracefully when they receive
1554 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1555 waiting for a server hello which it will never receive. Now we treat a
1556 received no_renegotiation alert as a fatal error. This is because
1557 applications requesting a renegotiation might well expect it to succeed
1558 and would have no code in place to handle the server denying it so the
1559 only safe thing to do is to terminate the connection.
1562 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1563 peer supports secure renegotiation and 0 otherwise. Print out peer
1564 renegotiation support in s_client/s_server.
1567 *) Replace the highly broken and deprecated SPKAC certification method with
1568 the updated NID creation version. This should correctly handle UTF8.
1571 *) Implement RFC5746. Re-enable renegotiation but require the extension
1572 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1573 turns out to be a bad idea. It has been replaced by
1574 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1575 SSL_CTX_set_options(). This is really not recommended unless you
1576 know what you are doing.
1577 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1579 *) Fixes to stateless session resumption handling. Use initial_ctx when
1580 issuing and attempting to decrypt tickets in case it has changed during
1581 servername handling. Use a non-zero length session ID when attempting
1582 stateless session resumption: this makes it possible to determine if
1583 a resumption has occurred immediately after receiving server hello
1584 (several places in OpenSSL subtly assume this) instead of later in
1588 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1589 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1590 fixes for a few places where the return code is not checked
1592 [Julia Lawall <julia@diku.dk>]
1594 *) Add --strict-warnings option to Configure script to include devteam
1595 warnings in other configurations.
1598 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1599 makes it possible to install openssl libraries in locations which
1600 have names other than "lib", for example "/usr/lib64" which some
1602 [Steve Henson, based on patch from Jeremy Utley]
1604 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1605 X690 8.9.12 and can produce some misleading textual output of OIDs.
1606 [Steve Henson, reported by Dan Kaminsky]
1608 *) Delete MD2 from algorithm tables. This follows the recommendation in
1609 several standards that it is not used in new applications due to
1610 several cryptographic weaknesses. For binary compatibility reasons
1611 the MD2 API is still compiled in by default.
1614 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1618 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1619 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1621 [Guenter <lists@gknw.net>]
1623 *) Fix the server certificate chain building code to use X509_verify_cert(),
1624 it used to have an ad-hoc builder which was unable to cope with anything
1625 other than a simple chain.
1626 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1628 *) Don't check self signed certificate signatures in X509_verify_cert()
1629 by default (a flag can override this): it just wastes time without
1630 adding any security. As a useful side effect self signed root CAs
1631 with non-FIPS digests are now usable in FIPS mode.
1634 *) In dtls1_process_out_of_seq_message() the check if the current message
1635 is already buffered was missing. For every new message was memory
1636 allocated, allowing an attacker to perform an denial of service attack
1637 with sending out of seq handshake messages until there is no memory
1638 left. Additionally every future messege was buffered, even if the
1639 sequence number made no sense and would be part of another handshake.
1640 So only messages with sequence numbers less than 10 in advance will be
1641 buffered. (CVE-2009-1378)
1642 [Robin Seggelmann, discovered by Daniel Mentz]
1644 *) Records are buffered if they arrive with a future epoch to be
1645 processed after finishing the corresponding handshake. There is
1646 currently no limitation to this buffer allowing an attacker to perform
1647 a DOS attack with sending records with future epochs until there is no
1648 memory left. This patch adds the pqueue_size() function to detemine
1649 the size of a buffer and limits the record buffer to 100 entries.
1651 [Robin Seggelmann, discovered by Daniel Mentz]
1653 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1654 parent structure is freed. (CVE-2009-1379)
1657 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1658 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1661 [Ilya O. <vrghost@gmail.com>]
1663 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1665 *) Disable renegotiation completely - this fixes a severe security
1666 problem (CVE-2009-3555) at the cost of breaking all
1667 renegotiation. Renegotiation can be re-enabled by setting
1668 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1669 run-time. This is really not recommended unless you know what
1673 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1675 *) Don't set val to NULL when freeing up structures, it is freed up by
1676 underlying code. If sizeof(void *) > sizeof(long) this can result in
1677 zeroing past the valid field. (CVE-2009-0789)
1678 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1680 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1681 checked correctly. This would allow some invalid signed attributes to
1682 appear to verify correctly. (CVE-2009-0591)
1683 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1685 *) Reject UniversalString and BMPString types with invalid lengths. This
1686 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1687 a legal length. (CVE-2009-0590)
1690 *) Set S/MIME signing as the default purpose rather than setting it
1691 unconditionally. This allows applications to override it at the store
1695 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1696 to handle some structures.
1699 *) Improve efficiency of mem_gets: don't search whole buffer each time
1701 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1703 *) New -hex option for openssl rand.
1706 *) Print out UTF8String and NumericString when parsing ASN1.
1709 *) Support NumericString type for name components.
1712 *) Allow CC in the environment to override the automatically chosen
1713 compiler. Note that nothing is done to ensure flags work with the
1717 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1719 *) Properly check EVP_VerifyFinal() and similar return values
1721 [Ben Laurie, Bodo Moeller, Google Security Team]
1723 *) Enable TLS extensions by default.
1726 *) Allow the CHIL engine to be loaded, whether the application is
1727 multithreaded or not. (This does not release the developer from the
1728 obligation to set up the dynamic locking callbacks.)
1729 [Sander Temme <sander@temme.net>]
1731 *) Use correct exit code if there is an error in dgst command.
1732 [Steve Henson; problem pointed out by Roland Dirlewanger]
1734 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1735 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1738 *) Add experimental JPAKE support, including demo authentication in
1739 s_client and s_server.
1742 *) Set the comparison function in v3_addr_canonize().
1743 [Rob Austein <sra@hactrn.net>]
1745 *) Add support for XMPP STARTTLS in s_client.
1746 [Philip Paeps <philip@freebsd.org>]
1748 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1749 to ensure that even with this option, only ciphersuites in the
1750 server's preference list will be accepted. (Note that the option
1751 applies only when resuming a session, so the earlier behavior was
1752 just about the algorithm choice for symmetric cryptography.)
1755 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1757 *) Fix NULL pointer dereference if a DTLS server received
1758 ChangeCipherSpec as first record (CVE-2009-1386).
1761 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1762 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1765 *) The fix in 0.9.8c that supposedly got rid of unsafe
1766 double-checked locking was incomplete for RSA blinding,
1767 addressing just one layer of what turns out to have been
1768 doubly unsafe triple-checked locking.
1770 So now fix this for real by retiring the MONT_HELPER macro
1771 in crypto/rsa/rsa_eay.c.
1773 [Bodo Moeller; problem pointed out by Marius Schilder]
1775 *) Various precautionary measures:
1777 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1779 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1780 (NB: This would require knowledge of the secret session ticket key
1781 to exploit, in which case you'd be SOL either way.)
1783 - Change bn_nist.c so that it will properly handle input BIGNUMs
1784 outside the expected range.
1786 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1789 [Neel Mehta, Bodo Moeller]
1791 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1792 the load fails. Useful for distros.
1793 [Ben Laurie and the FreeBSD team]
1795 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1798 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1801 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1803 This work was sponsored by Logica.
1806 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1807 keystores. Support for SSL/TLS client authentication too.
1808 Not compiled unless enable-capieng specified to Configure.
1810 This work was sponsored by Logica.
1813 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1814 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1815 attribute creation routines such as certifcate requests and PKCS#12
1819 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1821 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1822 handshake which could lead to a cilent crash as found using the
1823 Codenomicon TLS test suite (CVE-2008-1672)
1824 [Steve Henson, Mark Cox]
1826 *) Fix double free in TLS server name extensions which could lead to
1827 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1830 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1832 Clear the error queue to ensure that error entries left from
1833 older function calls do not interfere with the correct operation.
1834 [Lutz Jaenicke, Erik de Castro Lopo]
1836 *) Remove root CA certificates of commercial CAs:
1838 The OpenSSL project does not recommend any specific CA and does not
1839 have any policy with respect to including or excluding any CA.
1840 Therefore it does not make any sense to ship an arbitrary selection
1841 of root CA certificates with the OpenSSL software.
1844 *) RSA OAEP patches to fix two separate invalid memory reads.
1845 The first one involves inputs when 'lzero' is greater than
1846 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1847 before the beginning of from). The second one involves inputs where
1848 the 'db' section contains nothing but zeroes (there is a one-byte
1849 invalid read after the end of 'db').
1850 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1852 *) Partial backport from 0.9.9-dev:
1854 Introduce bn_mul_mont (dedicated Montgomery multiplication
1855 procedure) as a candidate for BIGNUM assembler implementation.
1856 While 0.9.9-dev uses assembler for various architectures, only
1857 x86_64 is available by default here in the 0.9.8 branch, and
1858 32-bit x86 is available through a compile-time setting.
1860 To try the 32-bit x86 assembler implementation, use Configure
1861 option "enable-montasm" (which exists only for this backport).
1863 As "enable-montasm" for 32-bit x86 disclaims code stability
1864 anyway, in this constellation we activate additional code
1865 backported from 0.9.9-dev for further performance improvements,
1866 namely BN_from_montgomery_word. (To enable this otherwise,
1867 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1869 [Andy Polyakov (backport partially by Bodo Moeller)]
1871 *) Add TLS session ticket callback. This allows an application to set
1872 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1873 values. This is useful for key rollover for example where several key
1874 sets may exist with different names.
1877 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1878 This was broken until now in 0.9.8 releases, such that the only way
1879 a registered ENGINE could be used (assuming it initialises
1880 successfully on the host) was to explicitly set it as the default
1881 for the relevant algorithms. This is in contradiction with 0.9.7
1882 behaviour and the documentation. With this fix, when an ENGINE is
1883 registered into a given algorithm's table of implementations, the
1884 'uptodate' flag is reset so that auto-discovery will be used next
1885 time a new context for that algorithm attempts to select an
1887 [Ian Lister (tweaked by Geoff Thorpe)]
1889 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1890 implemention in the following ways:
1892 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1895 Lack of BER streaming support means one pass streaming processing is
1896 only supported if data is detached: setting the streaming flag is
1897 ignored for embedded content.
1899 CMS support is disabled by default and must be explicitly enabled
1900 with the enable-cms configuration option.
1903 *) Update the GMP engine glue to do direct copies between BIGNUM and
1904 mpz_t when openssl and GMP use the same limb size. Otherwise the
1905 existing "conversion via a text string export" trick is still used.
1906 [Paul Sheer <paulsheer@gmail.com>]
1908 *) Zlib compression BIO. This is a filter BIO which compressed and
1909 uncompresses any data passed through it.
1912 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1913 RFC3394 compatible AES key wrapping.
1916 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1917 sets string data without copying. X509_ALGOR_set0() and
1918 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1919 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1920 from an X509_ATTRIBUTE structure optionally checking it occurs only
1921 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1925 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1926 to get the expected BN_FLG_CONSTTIME behavior.
1927 [Bodo Moeller (Google)]
1931 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1932 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1933 - added some more tests to do_tests.pl
1934 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1935 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1936 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1937 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1938 - various changes to netware.pl to enable gcc-cross builds on Win32
1940 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1941 - various changes to fix missing prototype warnings
1942 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1943 - added AES, WHIRLPOOL and CPUID assembler code to build files
1944 - added missing AES assembler make rules to mk1mf.pl
1945 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1946 [Guenter Knauf <eflash@gmx.net>]
1948 *) Implement certificate status request TLS extension defined in RFC3546.
1949 A client can set the appropriate parameters and receive the encoded
1950 OCSP response via a callback. A server can query the supplied parameters
1951 and set the encoded OCSP response in the callback. Add simplified examples
1952 to s_client and s_server.
1955 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1957 *) Fix various bugs:
1958 + Binary incompatibility of ssl_ctx_st structure
1959 + DTLS interoperation with non-compliant servers
1960 + Don't call get_session_cb() without proposed session
1961 + Fix ia64 assembler code
1962 [Andy Polyakov, Steve Henson]
1964 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1966 *) DTLS Handshake overhaul. There were longstanding issues with
1967 OpenSSL DTLS implementation, which were making it impossible for
1968 RFC 4347 compliant client to communicate with OpenSSL server.
1969 Unfortunately just fixing these incompatibilities would "cut off"
1970 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1971 server keeps tolerating non RFC compliant syntax. The opposite is
1972 not true, 0.9.8f client can not communicate with earlier server.
1973 This update even addresses CVE-2007-4995.
1976 *) Changes to avoid need for function casts in OpenSSL: some compilers
1977 (gcc 4.2 and later) reject their use.
1978 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1981 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1982 RFC4507bis. The encrypted ticket format is an encrypted encoded
1983 SSL_SESSION structure, that way new session features are automatically
1986 If a client application caches session in an SSL_SESSION structure
1987 support is transparent because tickets are now stored in the encoded
1990 The SSL_CTX structure automatically generates keys for ticket
1991 protection in servers so again support should be possible
1992 with no application modification.
1994 If a client or server wishes to disable RFC4507 support then the option
1995 SSL_OP_NO_TICKET can be set.
1997 Add a TLS extension debugging callback to allow the contents of any client
1998 or server extensions to be examined.
2000 This work was sponsored by Google.
2003 *) Add initial support for TLS extensions, specifically for the server_name
2004 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2005 have new members for a host name. The SSL data structure has an
2006 additional member SSL_CTX *initial_ctx so that new sessions can be
2007 stored in that context to allow for session resumption, even after the
2008 SSL has been switched to a new SSL_CTX in reaction to a client's
2009 server_name extension.
2011 New functions (subject to change):
2013 SSL_get_servername()
2014 SSL_get_servername_type()
2017 New CTRL codes and macros (subject to change):
2019 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2020 - SSL_CTX_set_tlsext_servername_callback()
2021 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2022 - SSL_CTX_set_tlsext_servername_arg()
2023 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2025 openssl s_client has a new '-servername ...' option.
2027 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2028 '-key2 ...', '-servername_fatal' (subject to change). This allows
2029 testing the HostName extension for a specific single host name ('-cert'
2030 and '-key' remain fallbacks for handshakes without HostName
2031 negotiation). If the unrecogninzed_name alert has to be sent, this by
2032 default is a warning; it becomes fatal with the '-servername_fatal'
2035 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2037 *) Add AES and SSE2 assembly language support to VC++ build.
2040 *) Mitigate attack on final subtraction in Montgomery reduction.
2043 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2044 (which previously caused an internal error).
2047 *) Squeeze another 10% out of IGE mode when in != out.
2050 *) AES IGE mode speedup.
2051 [Dean Gaudet (Google)]
2053 *) Add the Korean symmetric 128-bit cipher SEED (see
2054 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2055 add SEED ciphersuites from RFC 4162:
2057 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2058 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2059 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2060 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2062 To minimize changes between patchlevels in the OpenSSL 0.9.8
2063 series, SEED remains excluded from compilation unless OpenSSL
2064 is configured with 'enable-seed'.
2065 [KISA, Bodo Moeller]
2067 *) Mitigate branch prediction attacks, which can be practical if a
2068 single processor is shared, allowing a spy process to extract
2069 information. For detailed background information, see
2070 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2071 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2072 and Necessary Software Countermeasures"). The core of the change
2073 are new versions BN_div_no_branch() and
2074 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2075 respectively, which are slower, but avoid the security-relevant
2076 conditional branches. These are automatically called by BN_div()
2077 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2078 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2079 remove a conditional branch.
2081 BN_FLG_CONSTTIME is the new name for the previous
2082 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2083 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2084 in the exponent causes BN_mod_exp_mont() to use the alternative
2085 implementation in BN_mod_exp_mont_consttime().) The old name
2086 remains as a deprecated alias.
2088 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2089 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2090 constant-time implementations for more than just exponentiation.
2091 Here too the old name is kept as a deprecated alias.
2093 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2094 the BN_BLINDING structure gets an independent copy of the
2095 modulus. This means that the previous "BIGNUM *m" argument to
2096 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2097 essentially becomes "const BIGNUM *m", although we can't actually
2098 change this in the header file before 0.9.9. It allows
2099 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2100 enable BN_FLG_CONSTTIME.
2102 [Matthew D Wood (Intel Corp)]
2104 *) In the SSL/TLS server implementation, be strict about session ID
2105 context matching (which matters if an application uses a single
2106 external cache for different purposes). Previously,
2107 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2108 set. This did ensure strict client verification, but meant that,
2109 with applications using a single external cache for quite
2110 different requirements, clients could circumvent ciphersuite
2111 restrictions for a given session ID context by starting a session
2112 in a different context.
2115 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2116 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2117 authentication-only ciphersuites.
2120 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2121 not complete and could lead to a possible single byte overflow
2122 (CVE-2007-5135) [Ben Laurie]
2124 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2126 *) Since AES128 and AES256 (and similarly Camellia128 and
2127 Camellia256) share a single mask bit in the logic of
2128 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2129 kludge to work properly if AES128 is available and AES256 isn't
2130 (or if Camellia128 is available and Camellia256 isn't).
2133 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2134 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2135 When a point or a seed is encoded in a BIT STRING, we need to
2136 prevent the removal of trailing zero bits to get the proper DER
2137 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2138 of a NamedBitList, for which trailing 0 bits need to be removed.)
2141 *) Have SSL/TLS server implementation tolerate "mismatched" record
2142 protocol version while receiving ClientHello even if the
2143 ClientHello is fragmented. (The server can't insist on the
2144 particular protocol version it has chosen before the ServerHello
2145 message has informed the client about his choice.)
2148 *) Add RFC 3779 support.
2149 [Rob Austein for ARIN, Ben Laurie]
2151 *) Load error codes if they are not already present instead of using a
2152 static variable. This allows them to be cleanly unloaded and reloaded.
2153 Improve header file function name parsing.
2156 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2157 or CAPABILITY handshake as required by RFCs.
2160 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2162 *) Introduce limits to prevent malicious keys being able to
2163 cause a denial of service. (CVE-2006-2940)
2164 [Steve Henson, Bodo Moeller]
2166 *) Fix ASN.1 parsing of certain invalid structures that can result
2167 in a denial of service. (CVE-2006-2937) [Steve Henson]
2169 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2170 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2172 *) Fix SSL client code which could crash if connecting to a
2173 malicious SSLv2 server. (CVE-2006-4343)
2174 [Tavis Ormandy and Will Drewry, Google Security Team]
2176 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2177 match only those. Before that, "AES256-SHA" would be interpreted
2178 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2179 the same strength classification in 0.9.7h) as we currently only
2180 have a single AES bit in the ciphersuite description bitmap.
2181 That change, however, also applied to ciphersuite strings such as
2182 "RC4-MD5" that intentionally matched multiple ciphersuites --
2183 namely, SSL 2.0 ciphersuites in addition to the more common ones
2184 from SSL 3.0/TLS 1.0.
2186 So we change the selection algorithm again: Naming an explicit
2187 ciphersuite selects this one ciphersuite, and any other similar
2188 ciphersuite (same bitmap) from *other* protocol versions.
2189 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2190 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2192 Since SSL 2.0 does not have any ciphersuites for which the
2193 128/256 bit distinction would be relevant, this works for now.
2194 The proper fix will be to use different bits for AES128 and
2195 AES256, which would have avoided the problems from the beginning;
2196 however, bits are scarce, so we can only do this in a new release
2197 (not just a patchlevel) when we can change the SSL_CIPHER
2198 definition to split the single 'unsigned long mask' bitmap into
2199 multiple values to extend the available space.
2203 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2205 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2206 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2208 *) Add AES IGE and biIGE modes.
2211 *) Change the Unix randomness entropy gathering to use poll() when
2212 possible instead of select(), since the latter has some
2213 undesirable limitations.
2214 [Darryl Miles via Richard Levitte and Bodo Moeller]
2216 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2217 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2218 cannot be implicitly activated as part of, e.g., the "AES" alias.
2219 However, please upgrade to OpenSSL 0.9.9[-dev] for
2220 non-experimental use of the ECC ciphersuites to get TLS extension
2221 support, which is required for curve and point format negotiation
2222 to avoid potential handshake problems.
2225 *) Disable rogue ciphersuites:
2227 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2228 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2229 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2231 The latter two were purportedly from
2232 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2235 Also deactivate the remaining ciphersuites from
2236 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2237 unofficial, and the ID has long expired.
2240 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2241 dual-core machines) and other potential thread-safety issues.
2244 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2245 versions), which is now available for royalty-free use
2246 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2247 Also, add Camellia TLS ciphersuites from RFC 4132.
2249 To minimize changes between patchlevels in the OpenSSL 0.9.8
2250 series, Camellia remains excluded from compilation unless OpenSSL
2251 is configured with 'enable-camellia'.
2254 *) Disable the padding bug check when compression is in use. The padding
2255 bug check assumes the first packet is of even length, this is not
2256 necessarily true if compresssion is enabled and can result in false
2257 positives causing handshake failure. The actual bug test is ancient
2258 code so it is hoped that implementations will either have fixed it by
2259 now or any which still have the bug do not support compression.
2262 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2264 *) When applying a cipher rule check to see if string match is an explicit
2265 cipher suite and only match that one cipher suite if it is.
2268 *) Link in manifests for VC++ if needed.
2269 [Austin Ziegler <halostatue@gmail.com>]
2271 *) Update support for ECC-based TLS ciphersuites according to
2272 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2273 TLS extensions, which are supported starting with the 0.9.9
2274 branch, not in the OpenSSL 0.9.8 branch).
2277 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2278 opaque EVP_CIPHER_CTX handling.
2281 *) Fixes and enhancements to zlib compression code. We now only use
2282 "zlib1.dll" and use the default __cdecl calling convention on Win32
2283 to conform with the standards mentioned here:
2284 http://www.zlib.net/DLL_FAQ.txt
2285 Static zlib linking now works on Windows and the new --with-zlib-include
2286 --with-zlib-lib options to Configure can be used to supply the location
2287 of the headers and library. Gracefully handle case where zlib library
2291 *) Several fixes and enhancements to the OID generation code. The old code
2292 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2293 handle numbers larger than ULONG_MAX, truncated printing and had a
2294 non standard OBJ_obj2txt() behaviour.
2297 *) Add support for building of engines under engine/ as shared libraries
2298 under VC++ build system.
2301 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2302 Hopefully, we will not see any false combination of paths any more.
2305 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2307 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2308 (part of SSL_OP_ALL). This option used to disable the
2309 countermeasure against man-in-the-middle protocol-version
2310 rollback in the SSL 2.0 server implementation, which is a bad
2311 idea. (CVE-2005-2969)
2313 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2314 for Information Security, National Institute of Advanced Industrial
2315 Science and Technology [AIST], Japan)]
2317 *) Add two function to clear and return the verify parameter flags.
2320 *) Keep cipherlists sorted in the source instead of sorting them at
2321 runtime, thus removing the need for a lock.
2324 *) Avoid some small subgroup attacks in Diffie-Hellman.
2325 [Nick Mathewson and Ben Laurie]
2327 *) Add functions for well-known primes.
2330 *) Extended Windows CE support.
2331 [Satoshi Nakamura and Andy Polyakov]
2333 *) Initialize SSL_METHOD structures at compile time instead of during
2334 runtime, thus removing the need for a lock.
2337 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2338 attempting to decrypt each encrypted key in turn. Add support to
2342 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2344 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2347 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2350 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2351 key into the same file any more.
2354 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2357 *) Add -utf8 command line and config file option to 'ca'.
2358 [Stefan <stf@udoma.org]
2360 *) Removed the macro des_crypt(), as it seems to conflict with some
2361 libraries. Use DES_crypt().
2364 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2365 involves renaming the source and generated shared-libs for
2366 both. The engines will accept the corrected or legacy ids
2367 ('ncipher' and '4758_cca' respectively) when binding. NB,
2368 this only applies when building 'shared'.
2369 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2371 *) Add attribute functions to EVP_PKEY structure. Modify
2372 PKCS12_create() to recognize a CSP name attribute and
2373 use it. Make -CSP option work again in pkcs12 utility.
2376 *) Add new functionality to the bn blinding code:
2377 - automatic re-creation of the BN_BLINDING parameters after
2378 a fixed number of uses (currently 32)
2379 - add new function for parameter creation
2380 - introduce flags to control the update behaviour of the
2381 BN_BLINDING parameters
2382 - hide BN_BLINDING structure
2383 Add a second BN_BLINDING slot to the RSA structure to improve
2384 performance when a single RSA object is shared among several
2388 *) Add support for DTLS.
2389 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2391 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2392 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2395 *) Remove buggy and incompletet DH cert support from
2396 ssl/ssl_rsa.c and ssl/s3_both.c
2399 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2400 the apps/openssl applications.
2403 *) Compile clean with "-Wall -Wmissing-prototypes
2404 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2405 DEBUG_SAFESTACK must also be set.
2408 *) Change ./Configure so that certain algorithms can be disabled by default.
2409 The new counterpiece to "no-xxx" is "enable-xxx".
2411 The patented RC5 and MDC2 algorithms will now be disabled unless
2412 "enable-rc5" and "enable-mdc2", respectively, are specified.
2414 (IDEA remains enabled despite being patented. This is because IDEA
2415 is frequently required for interoperability, and there is no license
2416 fee for non-commercial use. As before, "no-idea" can be used to
2417 avoid this algorithm.)
2421 *) Add processing of proxy certificates (see RFC 3820). This work was
2422 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2423 EGEE (Enabling Grids for E-science in Europe).
2426 *) RC4 performance overhaul on modern architectures/implementations, such
2427 as Intel P4, IA-64 and AMD64.
2430 *) New utility extract-section.pl. This can be used specify an alternative
2431 section number in a pod file instead of having to treat each file as
2432 a separate case in Makefile. This can be done by adding two lines to the
2435 =for comment openssl_section:XXX
2437 The blank line is mandatory.
2441 *) New arguments -certform, -keyform and -pass for s_client and s_server
2442 to allow alternative format key and certificate files and passphrase
2446 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2447 update associated structures and add various utility functions.
2449 Add new policy related verify parameters, include policy checking in
2450 standard verify code. Enhance 'smime' application with extra parameters
2451 to support policy checking and print out.
2454 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2455 Nehemiah processors. These extensions support AES encryption in hardware
2456 as well as RNG (though RNG support is currently disabled).
2457 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2459 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2462 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2463 [Andy Polyakov and a number of other people]
2465 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2466 implementation contributed by IBM.
2467 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2469 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2470 exponent rather than 'unsigned long'. There is a corresponding change to
2471 the new 'rsa_keygen' element of the RSA_METHOD structure.
2472 [Jelte Jansen, Geoff Thorpe]
2474 *) Functionality for creating the initial serial number file is now
2475 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2477 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2478 number file to 1, which is bound to cause problems. To avoid
2479 the problems while respecting compatibility between different 0.9.7
2480 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2481 CA.pl for serial number initialization. With the new release 0.9.8,
2482 we can fix the problem directly in the 'ca' utility.)
2485 *) Reduced header interdepencies by declaring more opaque objects in
2486 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2487 give fewer recursive includes, which could break lazy source code - so
2488 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2489 developers should define this symbol when building and using openssl to
2490 ensure they track the recommended behaviour, interfaces, [etc], but
2491 backwards-compatible behaviour prevails when this isn't defined.
2494 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2497 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2498 This will generate a random key of the appropriate length based on the
2499 cipher context. The EVP_CIPHER can provide its own random key generation
2500 routine to support keys of a specific form. This is used in the des and
2501 3des routines to generate a key of the correct parity. Update S/MIME
2502 code to use new functions and hence generate correct parity DES keys.
2503 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2504 valid (weak or incorrect parity).
2507 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2508 as looking them up. This is useful when the verified structure may contain
2509 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2510 present unless the new PKCS7_NO_CRL flag is asserted.
2513 *) Extend ASN1 oid configuration module. It now additionally accepts the
2516 shortName = some long name, 1.2.3.4
2519 *) Reimplemented the BN_CTX implementation. There is now no more static
2520 limitation on the number of variables it can handle nor the depth of the
2521 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2522 information can now expand as required, and rather than having a single
2523 static array of bignums, BN_CTX now uses a linked-list of such arrays
2524 allowing it to expand on demand whilst maintaining the usefulness of
2525 BN_CTX's "bundling".
2528 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2529 to allow all RSA operations to function using a single BN_CTX.
2532 *) Preliminary support for certificate policy evaluation and checking. This
2533 is initially intended to pass the tests outlined in "Conformance Testing
2534 of Relying Party Client Certificate Path Processing Logic" v1.07.
2537 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2538 remained unused and not that useful. A variety of other little bignum
2539 tweaks and fixes have also been made continuing on from the audit (see
2543 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2544 associated ASN1, EVP and SSL functions and old ASN1 macros.
2547 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2548 and this should never fail. So the return value from the use of
2549 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2550 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2553 *) BN_CTX_get() should return zero-valued bignums, providing the same
2554 initialised value as BN_new().
2555 [Geoff Thorpe, suggested by Ulf Möller]
2557 *) Support for inhibitAnyPolicy certificate extension.
2560 *) An audit of the BIGNUM code is underway, for which debugging code is
2561 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2562 is considered valid when processing BIGNUMs, and causes execution to
2563 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2564 further steps are taken to deliberately pollute unused data in BIGNUM
2565 structures to try and expose faulty code further on. For now, openssl will
2566 (in its default mode of operation) continue to tolerate the inconsistent
2567 forms that it has tolerated in the past, but authors and packagers should
2568 consider trying openssl and their own applications when compiled with
2569 these debugging symbols defined. It will help highlight potential bugs in
2570 their own code, and will improve the test coverage for OpenSSL itself. At
2571 some point, these tighter rules will become openssl's default to improve
2572 maintainability, though the assert()s and other overheads will remain only
2573 in debugging configurations. See bn.h for more details.
2574 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2576 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2577 that can only be obtained through BN_CTX_new() (which implicitly
2578 initialises it). The presence of this function only made it possible
2579 to overwrite an existing structure (and cause memory leaks).
2582 *) Because of the callback-based approach for implementing LHASH as a
2583 template type, lh_insert() adds opaque objects to hash-tables and
2584 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2585 to clean up those corresponding objects before destroying the hash table
2586 (and losing the object pointers). So some over-zealous constifications in
2587 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2588 objects as "const" and the lh_doall[_arg] callback wrappers are not
2589 prototyped to have "const" restrictions on the object pointers they are
2590 given (and so aren't required to cast them away any more).
2593 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2594 (speed) prefers to use its own implementation. The two implementations
2595 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2596 its object type properly exposed (MS_TM) instead of casting to/from "char
2597 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2598 aren't necessarily the greatest nomenclatures - but this is what was used
2599 internally to the implementation so I've used that for now.
2602 *) Ensure that deprecated functions do not get compiled when
2603 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2604 the self-tests were still using deprecated key-generation functions so
2605 these have been updated also.
2608 *) Reorganise PKCS#7 code to separate the digest location functionality
2609 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2610 New function PKCS7_set_digest() to set the digest type for PKCS#7
2611 digestedData type. Add additional code to correctly generate the
2612 digestedData type and add support for this type in PKCS7 initialization
2616 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2617 structure of type "other".
2620 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2621 sure the loop does correctly stop and breaking ("division by zero")
2622 modulus operations are not performed. The (pre-generated) prime
2623 table crypto/bn/bn_prime.h was already correct, but it could not be
2624 re-generated on some platforms because of the "division by zero"
2625 situation in the script.
2626 [Ralf S. Engelschall]
2628 *) Update support for ECC-based TLS ciphersuites according to
2629 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2630 SHA-1 now is only used for "small" curves (where the
2631 representation of a field element takes up to 24 bytes); for
2632 larger curves, the field element resulting from ECDH is directly
2633 used as premaster secret.
2634 [Douglas Stebila (Sun Microsystems Laboratories)]
2636 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2637 curve secp160r1 to the tests.
2638 [Douglas Stebila (Sun Microsystems Laboratories)]
2640 *) Add the possibility to load symbols globally with DSO.
2641 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2643 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2644 control of the error stack.
2647 *) Add support for STORE in ENGINE.
2650 *) Add the STORE type. The intention is to provide a common interface
2651 to certificate and key stores, be they simple file-based stores, or
2652 HSM-type store, or LDAP stores, or...
2653 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2656 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2657 pass a list of arguments to any function as well as provide a way
2658 for a function to pass data back to the caller.
2661 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2662 works like BUF_strdup() but can be used to duplicate a portion of
2663 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2667 *) Add the function sk_find_ex() which works like sk_find(), but will
2668 return an index to an element even if an exact match couldn't be
2669 found. The index is guaranteed to point at the element where the
2670 searched-for key would be inserted to preserve sorting order.
2673 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2674 takes an extra flags argument for optional functionality. Currently,
2675 the following flags are defined:
2677 OBJ_BSEARCH_VALUE_ON_NOMATCH
2678 This one gets OBJ_bsearch_ex() to return a pointer to the first
2679 element where the comparing function returns a negative or zero
2682 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2683 This one gets OBJ_bsearch_ex() to return a pointer to the first
2684 element where the comparing function returns zero. This is useful
2685 if there are more than one element where the comparing function
2689 *) Make it possible to create self-signed certificates with 'openssl ca'
2690 in such a way that the self-signed certificate becomes part of the
2691 CA database and uses the same mechanisms for serial number generation
2692 as all other certificate signing. The new flag '-selfsign' enables
2693 this functionality. Adapt CA.sh and CA.pl.in.
2696 *) Add functionality to check the public key of a certificate request
2697 against a given private. This is useful to check that a certificate
2698 request can be signed by that key (self-signing).
2701 *) Make it possible to have multiple active certificates with the same
2702 subject in the CA index file. This is done only if the keyword
2703 'unique_subject' is set to 'no' in the main CA section (default
2704 if 'CA_default') of the configuration file. The value is saved
2705 with the database itself in a separate index attribute file,
2706 named like the index file with '.attr' appended to the name.
2709 *) Generate muti valued AVAs using '+' notation in config files for
2713 *) Support for nameConstraints certificate extension.
2716 *) Support for policyConstraints certificate extension.
2719 *) Support for policyMappings certificate extension.
2722 *) Make sure the default DSA_METHOD implementation only uses its
2723 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2724 and change its own handlers to be NULL so as to remove unnecessary
2725 indirection. This lets alternative implementations fallback to the
2726 default implementation more easily.
2729 *) Support for directoryName in GeneralName related extensions
2733 *) Make it possible to link applications using Makefile.shared.
2734 Make that possible even when linking against static libraries!
2737 *) Support for single pass processing for S/MIME signing. This now
2738 means that S/MIME signing can be done from a pipe, in addition
2739 cleartext signing (multipart/signed type) is effectively streaming
2740 and the signed data does not need to be all held in memory.
2742 This is done with a new flag PKCS7_STREAM. When this flag is set
2743 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2744 is done after the data is output (and digests calculated) in
2745 SMIME_write_PKCS7().
2748 *) Add full support for -rpath/-R, both in shared libraries and
2749 applications, at least on the platforms where it's known how
2753 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2754 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2755 will now compute a table of multiples of the generator that
2756 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2757 faster (notably in the case of a single point multiplication,
2758 scalar * generator).
2759 [Nils Larsch, Bodo Moeller]
2761 *) IPv6 support for certificate extensions. The various extensions
2762 which use the IP:a.b.c.d can now take IPv6 addresses using the
2763 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2767 *) Added an ENGINE that implements RSA by performing private key
2768 exponentiations with the GMP library. The conversions to and from
2769 GMP's mpz_t format aren't optimised nor are any montgomery forms
2770 cached, and on x86 it appears OpenSSL's own performance has caught up.
2771 However there are likely to be other architectures where GMP could
2772 provide a boost. This ENGINE is not built in by default, but it can be
2773 specified at Configure time and should be accompanied by the necessary
2774 linker additions, eg;
2775 ./config -DOPENSSL_USE_GMP -lgmp
2778 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2779 testing availability of engines with "-t" - the old behaviour is
2780 produced by increasing the feature's verbosity with "-tt".
2783 *) ECDSA routines: under certain error conditions uninitialized BN objects
2784 could be freed. Solution: make sure initialization is performed early
2785 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2789 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2790 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2791 software implementations. For DSA and DH, parameter generation can
2792 also be overriden by providing the appropriate method callbacks.
2795 *) Change the "progress" mechanism used in key-generation and
2796 primality testing to functions that take a new BN_GENCB pointer in
2797 place of callback/argument pairs. The new API functions have "_ex"
2798 postfixes and the older functions are reimplemented as wrappers for
2799 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2800 declarations of the old functions to help (graceful) attempts to
2801 migrate to the new functions. Also, the new key-generation API
2802 functions operate on a caller-supplied key-structure and return
2803 success/failure rather than returning a key or NULL - this is to
2804 help make "keygen" another member function of RSA_METHOD etc.
2806 Example for using the new callback interface:
2808 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2812 BN_GENCB_set(&my_cb, my_callback, my_arg);
2814 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2815 /* For the meaning of a, b in calls to my_callback(), see the
2816 * documentation of the function that calls the callback.
2817 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2818 * my_callback should return 1 if it wants BN_is_prime_ex()
2819 * to continue, or 0 to stop.
2824 *) Change the ZLIB compression method to be stateful, and make it
2825 available to TLS with the number defined in
2826 draft-ietf-tls-compression-04.txt.
2829 *) Add the ASN.1 structures and functions for CertificatePair, which
2830 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2832 CertificatePair ::= SEQUENCE {
2833 forward [0] Certificate OPTIONAL,
2834 reverse [1] Certificate OPTIONAL,
2835 -- at least one of the pair shall be present -- }
2837 Also implement the PEM functions to read and write certificate
2838 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2840 This needed to be defined, mostly for the sake of the LDAP
2841 attribute crossCertificatePair, but may prove useful elsewhere as
2845 *) Make it possible to inhibit symlinking of shared libraries in
2846 Makefile.shared, for Cygwin's sake.
2849 *) Extend the BIGNUM API by creating a function
2850 void BN_set_negative(BIGNUM *a, int neg);
2851 and a macro that behave like
2852 int BN_is_negative(const BIGNUM *a);
2854 to avoid the need to access 'a->neg' directly in applications.
2857 *) Implement fast modular reduction for pseudo-Mersenne primes
2858 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2859 EC_GROUP_new_curve_GFp() will now automatically use this
2861 [Nils Larsch <nla@trustcenter.de>]
2863 *) Add new lock type (CRYPTO_LOCK_BN).
2866 *) Change the ENGINE framework to automatically load engines
2867 dynamically from specific directories unless they could be
2868 found to already be built in or loaded. Move all the
2869 current engines except for the cryptodev one to a new
2871 The engines in engines/ are built as shared libraries if
2872 the "shared" options was given to ./Configure or ./config.
2873 Otherwise, they are inserted in libcrypto.a.
2874 /usr/local/ssl/engines is the default directory for dynamic
2875 engines, but that can be overriden at configure time through
2876 the usual use of --prefix and/or --openssldir, and at run
2877 time with the environment variable OPENSSL_ENGINES.
2878 [Geoff Thorpe and Richard Levitte]
2880 *) Add Makefile.shared, a helper makefile to build shared
2881 libraries. Addapt Makefile.org.
2884 *) Add version info to Win32 DLLs.
2885 [Peter 'Luna' Runestig" <peter@runestig.com>]
2887 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2888 can be added using this API to created arbitrary PKCS#12
2889 files while avoiding the low level API.
2891 New options to PKCS12_create(), key or cert can be NULL and
2892 will then be omitted from the output file. The encryption
2893 algorithm NIDs can be set to -1 for no encryption, the mac
2894 iteration count can be set to 0 to omit the mac.
2896 Enhance pkcs12 utility by making the -nokeys and -nocerts
2897 options work when creating a PKCS#12 file. New option -nomac
2898 to omit the mac, NONE can be set for an encryption algorithm.
2899 New code is modified to use the enhanced PKCS12_create()
2900 instead of the low level API.
2903 *) Extend ASN1 encoder to support indefinite length constructed
2904 encoding. This can output sequences tags and octet strings in
2905 this form. Modify pk7_asn1.c to support indefinite length
2906 encoding. This is experimental and needs additional code to
2907 be useful, such as an ASN1 bio and some enhanced streaming
2910 Extend template encode functionality so that tagging is passed
2911 down to the template encoder.
2914 *) Let 'openssl req' fail if an argument to '-newkey' is not
2915 recognized instead of using RSA as a default.
2918 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2919 As these are not official, they are not included in "ALL";
2920 the "ECCdraft" ciphersuite group alias can be used to select them.
2921 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2923 *) Add ECDH engine support.
2924 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2926 *) Add ECDH in new directory crypto/ecdh/.
2927 [Douglas Stebila (Sun Microsystems Laboratories)]
2929 *) Let BN_rand_range() abort with an error after 100 iterations
2930 without success (which indicates a broken PRNG).
2933 *) Change BN_mod_sqrt() so that it verifies that the input value
2934 is really the square of the return value. (Previously,
2935 BN_mod_sqrt would show GIGO behaviour.)
2938 *) Add named elliptic curves over binary fields from X9.62, SECG,
2939 and WAP/WTLS; add OIDs that were still missing.
2941 [Sheueling Chang Shantz and Douglas Stebila
2942 (Sun Microsystems Laboratories)]
2944 *) Extend the EC library for elliptic curves over binary fields
2945 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2948 EC_GF2m_simple_method
2952 EC_GROUP_new_curve_GF2m
2953 EC_GROUP_set_curve_GF2m
2954 EC_GROUP_get_curve_GF2m
2955 EC_POINT_set_affine_coordinates_GF2m
2956 EC_POINT_get_affine_coordinates_GF2m
2957 EC_POINT_set_compressed_coordinates_GF2m
2959 Point compression for binary fields is disabled by default for
2960 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2963 As binary polynomials are represented as BIGNUMs, various members
2964 of the EC_GROUP and EC_POINT data structures can be shared
2965 between the implementations for prime fields and binary fields;
2966 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2967 are essentially identical to their ..._GFp counterparts.
2968 (For simplicity, the '..._GFp' prefix has been dropped from
2969 various internal method names.)
2971 An internal 'field_div' method (similar to 'field_mul' and
2972 'field_sqr') has been added; this is used only for binary fields.
2974 [Sheueling Chang Shantz and Douglas Stebila
2975 (Sun Microsystems Laboratories)]
2977 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2978 through methods ('mul', 'precompute_mult').
2980 The generic implementations (now internally called 'ec_wNAF_mul'
2981 and 'ec_wNAF_precomputed_mult') remain the default if these
2982 methods are undefined.
2984 [Sheueling Chang Shantz and Douglas Stebila
2985 (Sun Microsystems Laboratories)]
2987 *) New function EC_GROUP_get_degree, which is defined through
2988 EC_METHOD. For curves over prime fields, this returns the bit
2989 length of the modulus.
2991 [Sheueling Chang Shantz and Douglas Stebila
2992 (Sun Microsystems Laboratories)]
2994 *) New functions EC_GROUP_dup, EC_POINT_dup.
2995 (These simply call ..._new and ..._copy).
2997 [Sheueling Chang Shantz and Douglas Stebila
2998 (Sun Microsystems Laboratories)]
3000 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3001 Polynomials are represented as BIGNUMs (where the sign bit is not
3002 used) in the following functions [macros]:
3005 BN_GF2m_sub [= BN_GF2m_add]
3006 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3007 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3008 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3010 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3011 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3012 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3013 BN_GF2m_cmp [= BN_ucmp]
3015 (Note that only the 'mod' functions are actually for fields GF(2^m).
3016 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3018 For some functions, an the irreducible polynomial defining a
3019 field can be given as an 'unsigned int[]' with strictly
3020 decreasing elements giving the indices of those bits that are set;
3021 i.e., p[] represents the polynomial
3022 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3024 p[0] > p[1] > ... > p[k] = 0.
3025 This applies to the following functions:
3030 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3031 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3033 BN_GF2m_mod_sqrt_arr
3034 BN_GF2m_mod_solve_quad_arr
3038 Conversion can be performed by the following functions:
3043 bntest.c has additional tests for binary polynomial arithmetic.
3045 Two implementations for BN_GF2m_mod_div() are available.
3046 The default algorithm simply uses BN_GF2m_mod_inv() and
3047 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3048 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3049 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3051 [Sheueling Chang Shantz and Douglas Stebila
3052 (Sun Microsystems Laboratories)]
3054 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3055 functionality is disabled at compile-time.
3056 [Douglas Stebila <douglas.stebila@sun.com>]
3058 *) Change default behaviour of 'openssl asn1parse' so that more
3059 information is visible when viewing, e.g., a certificate:
3061 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3062 mode the content of non-printable OCTET STRINGs is output in a
3063 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3064 avoid the appearance of a printable string.
3065 [Nils Larsch <nla@trustcenter.de>]
3067 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3069 EC_GROUP_set_asn1_flag()
3070 EC_GROUP_get_asn1_flag()
3071 EC_GROUP_set_point_conversion_form()
3072 EC_GROUP_get_point_conversion_form()
3073 These control ASN1 encoding details:
3074 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3075 has been set to OPENSSL_EC_NAMED_CURVE.
3076 - Points are encoded in uncompressed form by default; options for
3077 asn1_for are as for point2oct, namely
3078 POINT_CONVERSION_COMPRESSED
3079 POINT_CONVERSION_UNCOMPRESSED
3080 POINT_CONVERSION_HYBRID
3082 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3085 EC_GROUP_get0_seed()
3086 EC_GROUP_get_seed_len()
3087 This is used only for ASN1 purposes (so far).
3088 [Nils Larsch <nla@trustcenter.de>]
3090 *) Add 'field_type' member to EC_METHOD, which holds the NID
3091 of the appropriate field type OID. The new function
3092 EC_METHOD_get_field_type() returns this value.
3093 [Nils Larsch <nla@trustcenter.de>]
3098 EC_POINT_point2hex()
3099 EC_POINT_hex2point()
3100 providing useful interfaces to EC_POINT_point2oct() and
3101 EC_POINT_oct2point().
3102 [Nils Larsch <nla@trustcenter.de>]
3104 *) Change internals of the EC library so that the functions
3105 EC_GROUP_set_generator()
3106 EC_GROUP_get_generator()
3107 EC_GROUP_get_order()
3108 EC_GROUP_get_cofactor()
3109 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3110 to methods, which would lead to unnecessary code duplication when
3111 adding different types of curves.
3112 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3114 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3115 arithmetic, and such that modified wNAFs are generated
3116 (which avoid length expansion in many cases).
3119 *) Add a function EC_GROUP_check_discriminant() (defined via
3120 EC_METHOD) that verifies that the curve discriminant is non-zero.
3122 Add a function EC_GROUP_check() that makes some sanity tests
3123 on a EC_GROUP, its generator and order. This includes
3124 EC_GROUP_check_discriminant().
3125 [Nils Larsch <nla@trustcenter.de>]
3127 *) Add ECDSA in new directory crypto/ecdsa/.
3129 Add applications 'openssl ecparam' and 'openssl ecdsa'
3130 (these are based on 'openssl dsaparam' and 'openssl dsa').
3132 ECDSA support is also included in various other files across the
3133 library. Most notably,
3134 - 'openssl req' now has a '-newkey ecdsa:file' option;
3135 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3136 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3137 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3138 them suitable for ECDSA where domain parameters must be
3139 extracted before the specific public key;
3140 - ECDSA engine support has been added.
3141 [Nils Larsch <nla@trustcenter.de>]
3143 *) Include some named elliptic curves, and add OIDs from X9.62,
3144 SECG, and WAP/WTLS. Each curve can be obtained from the new
3146 EC_GROUP_new_by_curve_name(),
3147 and the list of available named curves can be obtained with
3148 EC_get_builtin_curves().
3149 Also add a 'curve_name' member to EC_GROUP objects, which can be
3151 EC_GROUP_set_curve_name()
3152 EC_GROUP_get_curve_name()
3153 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3155 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3156 was actually never needed) and in BN_mul(). The removal in BN_mul()
3157 required a small change in bn_mul_part_recursive() and the addition
3158 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3159 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3160 bn_sub_words() and bn_add_words() except they take arrays with
3164 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3166 *) Cleanse PEM buffers before freeing them since they may contain
3168 [Benjamin Bennett <ben@psc.edu>]
3170 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3171 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3172 authentication-only ciphersuites.
3175 *) Since AES128 and AES256 share a single mask bit in the logic of
3176 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3177 kludge to work properly if AES128 is available and AES256 isn't.
3180 *) Expand security boundary to match 1.1.1 module.
3183 *) Remove redundant features: hash file source, editing of test vectors
3184 modify fipsld to use external fips_premain.c signature.
3187 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3188 run algorithm test programs.
3191 *) Make algorithm test programs more tolerant of whitespace.
3194 *) Have SSL/TLS server implementation tolerate "mismatched" record
3195 protocol version while receiving ClientHello even if the
3196 ClientHello is fragmented. (The server can't insist on the
3197 particular protocol version it has chosen before the ServerHello
3198 message has informed the client about his choice.)
3201 *) Load error codes if they are not already present instead of using a
3202 static variable. This allows them to be cleanly unloaded and reloaded.
3205 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3207 *) Introduce limits to prevent malicious keys being able to
3208 cause a denial of service. (CVE-2006-2940)
3209 [Steve Henson, Bodo Moeller]
3211 *) Fix ASN.1 parsing of certain invalid structures that can result
3212 in a denial of service. (CVE-2006-2937) [Steve Henson]
3214 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3215 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3217 *) Fix SSL client code which could crash if connecting to a
3218 malicious SSLv2 server. (CVE-2006-4343)
3219 [Tavis Ormandy and Will Drewry, Google Security Team]
3221 *) Change ciphersuite string processing so that an explicit
3222 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3223 will no longer include "AES128-SHA"), and any other similar
3224 ciphersuite (same bitmap) from *other* protocol versions (so that
3225 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3226 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3227 changes from 0.9.8b and 0.9.8d.
3230 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3232 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3233 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3235 *) Change the Unix randomness entropy gathering to use poll() when
3236 possible instead of select(), since the latter has some
3237 undesirable limitations.
3238 [Darryl Miles via Richard Levitte and Bodo Moeller]
3240 *) Disable rogue ciphersuites:
3242 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3243 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3244 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3246 The latter two were purportedly from
3247 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3250 Also deactive the remaining ciphersuites from
3251 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3252 unofficial, and the ID has long expired.
3255 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3256 dual-core machines) and other potential thread-safety issues.
3259 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3261 *) Adapt fipsld and the build system to link against the validated FIPS
3262 module in FIPS mode.
3265 *) Fixes for VC++ 2005 build under Windows.
3268 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3269 from a Windows bash shell such as MSYS. It is autodetected from the
3270 "config" script when run from a VC++ environment. Modify standard VC++
3271 build to use fipscanister.o from the GNU make build.
3274 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3276 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3277 The value now differs depending on if you build for FIPS or not.
3278 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3279 safely run with a non-FIPSed libcrypto, as it may crash because of
3280 the difference induced by this change.
3283 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3285 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3286 (part of SSL_OP_ALL). This option used to disable the
3287 countermeasure against man-in-the-middle protocol-version
3288 rollback in the SSL 2.0 server implementation, which is a bad
3289 idea. (CVE-2005-2969)
3291 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3292 for Information Security, National Institute of Advanced Industrial
3293 Science and Technology [AIST], Japan)]
3295 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3296 mainly for FIPS compliance and not fully integrated at this stage.
3299 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3300 the exponentiation using a fixed-length exponent. (Otherwise,
3301 the information leaked through timing could expose the secret key
3302 after many signatures; cf. Bleichenbacher's attack on DSA with
3306 *) Make a new fixed-window mod_exp implementation the default for
3307 RSA, DSA, and DH private-key operations so that the sequence of
3308 squares and multiplies and the memory access pattern are
3309 independent of the particular secret key. This will mitigate
3310 cache-timing and potential related attacks.
3312 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3313 and this is automatically used by BN_mod_exp_mont() if the new flag
3314 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3315 will use this BN flag for private exponents unless the flag
3316 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3317 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3319 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3321 *) Change the client implementation for SSLv23_method() and
3322 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3323 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3324 (Previously, the SSL 2.0 backwards compatible Client Hello
3325 message format would be used even with SSL_OP_NO_SSLv2.)
3328 *) Add support for smime-type MIME parameter in S/MIME messages which some
3332 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3333 a threadsafe manner. Modify rsa code to use new function and add calls
3334 to dsa and dh code (which had race conditions before).
3337 *) Include the fixed error library code in the C error file definitions
3338 instead of fixing them up at runtime. This keeps the error code
3339 structures constant.
3342 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3344 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3347 *) Fixes for newer kerberos headers. NB: the casts are needed because
3348 the 'length' field is signed on one version and unsigned on another
3349 with no (?) obvious way to tell the difference, without these VC++
3350 complains. Also the "definition" of FAR (blank) is no longer included
3351 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3352 some needed definitions.
3355 *) Undo Cygwin change.
3358 *) Added support for proxy certificates according to RFC 3820.
3359 Because they may be a security thread to unaware applications,
3360 they must be explicitely allowed in run-time. See
3361 docs/HOWTO/proxy_certificates.txt for further information.
3364 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3366 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3367 server and client random values. Previously
3368 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3369 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3371 This change has negligible security impact because:
3373 1. Server and client random values still have 24 bytes of pseudo random
3376 2. Server and client random values are sent in the clear in the initial
3379 3. The master secret is derived using the premaster secret (48 bytes in
3380 size for static RSA ciphersuites) as well as client server and random
3383 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3386 [Stephen Henson, reported by UK NISCC]
3388 *) Use Windows randomness collection on Cygwin.
3391 *) Fix hang in EGD/PRNGD query when communication socket is closed
3392 prematurely by EGD/PRNGD.
3393 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3395 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3398 *) Back-port of selected performance improvements from development
3399 branch, as well as improved support for PowerPC platforms.
3402 *) Add lots of checks for memory allocation failure, error codes to indicate
3403 failure and freeing up memory if a failure occurs.
3404 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3406 *) Add new -passin argument to dgst.
3409 *) Perform some character comparisons of different types in X509_NAME_cmp:
3410 this is needed for some certificates that reencode DNs into UTF8Strings
3411 (in violation of RFC3280) and can't or wont issue name rollover
3415 *) Make an explicit check during certificate validation to see that
3416 the CA setting in each certificate on the chain is correct. As a
3417 side effect always do the following basic checks on extensions,
3418 not just when there's an associated purpose to the check:
3420 - if there is an unhandled critical extension (unless the user
3421 has chosen to ignore this fault)
3422 - if the path length has been exceeded (if one is set at all)
3423 - that certain extensions fit the associated purpose (if one has
3427 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3429 *) Avoid a race condition when CRLs are checked in a multi threaded
3430 environment. This would happen due to the reordering of the revoked
3431 entries during signature checking and serial number lookup. Now the
3432 encoding is cached and the serial number sort performed under a lock.
3433 Add new STACK function sk_is_sorted().
3436 *) Add Delta CRL to the extension code.
3439 *) Various fixes to s3_pkt.c so alerts are sent properly.
3440 [David Holmes <d.holmes@f5.com>]
3442 *) Reduce the chances of duplicate issuer name and serial numbers (in
3443 violation of RFC3280) using the OpenSSL certificate creation utilities.
3444 This is done by creating a random 64 bit value for the initial serial
3445 number when a serial number file is created or when a self signed
3446 certificate is created using 'openssl req -x509'. The initial serial
3447 number file is created using 'openssl x509 -next_serial' in CA.pl
3448 rather than being initialized to 1.
3451 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3453 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3454 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3455 [Joe Orton, Steve Henson]
3457 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3459 [Joe Orton, Steve Henson]
3461 *) Make it possible to have multiple active certificates with the same
3462 subject in the CA index file. This is done only if the keyword
3463 'unique_subject' is set to 'no' in the main CA section (default
3464 if 'CA_default') of the configuration file. The value is saved
3465 with the database itself in a separate index attribute file,
3466 named like the index file with '.attr' appended to the name.
3469 *) X509 verify fixes. Disable broken certificate workarounds when
3470 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3471 keyUsage extension present. Don't accept CRLs with unhandled critical
3472 extensions: since verify currently doesn't process CRL extensions this
3473 rejects a CRL with *any* critical extensions. Add new verify error codes
3477 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3478 A clarification of RFC2560 will require the use of OCTET STRINGs and
3479 some implementations cannot handle the current raw format. Since OpenSSL
3480 copies and compares OCSP nonces as opaque blobs without any attempt at
3481 parsing them this should not create any compatibility issues.
3484 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3485 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3486 this HMAC (and other) operations are several times slower than OpenSSL
3490 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3491 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3493 *) Use the correct content when signing type "other".
3496 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3498 *) Fix various bugs revealed by running the NISCC test suite:
3500 Stop out of bounds reads in the ASN1 code when presented with
3501 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3503 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3505 If verify callback ignores invalid public key errors don't try to check
3506 certificate signature with the NULL public key.
3510 *) New -ignore_err option in ocsp application to stop the server
3511 exiting on the first error in a request.
3514 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3515 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3519 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3520 extra data after the compression methods not only for TLS 1.0
3521 but also for SSL 3.0 (as required by the specification).
3522 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3524 *) Change X509_certificate_type() to mark the key as exported/exportable
3525 when it's 512 *bits* long, not 512 bytes.
3528 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3529 blocks during encryption.
3532 *) Various fixes to base64 BIO and non blocking I/O. On write
3533 flushes were not handled properly if the BIO retried. On read
3534 data was not being buffered properly and had various logic bugs.
3535 This also affects blocking I/O when the data being decoded is a
3539 *) Various S/MIME bugfixes and compatibility changes:
3540 output correct application/pkcs7 MIME type if
3541 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3542 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3543 of files as .eml work). Correctly handle very long lines in MIME
3547 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3549 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3550 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3551 a protocol version number mismatch like a decryption error
3552 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3555 *) Turn on RSA blinding by default in the default implementation
3556 to avoid a timing attack. Applications that don't want it can call
3557 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3558 They would be ill-advised to do so in most cases.
3559 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3561 *) Change RSA blinding code so that it works when the PRNG is not
3562 seeded (in this case, the secret RSA exponent is abused as
3563 an unpredictable seed -- if it is not unpredictable, there
3564 is no point in blinding anyway). Make RSA blinding thread-safe
3565 by remembering the creator's thread ID in rsa->blinding and
3566 having all other threads use local one-time blinding factors
3567 (this requires more computation than sharing rsa->blinding, but
3568 avoids excessive locking; and if an RSA object is not shared
3569 between threads, blinding will still be very fast).
3572 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3573 ENGINE as defaults for all supported algorithms irrespective of
3574 the 'flags' parameter. 'flags' is now honoured, so applications
3575 should make sure they are passing it correctly.
3578 *) Target "mingw" now allows native Windows code to be generated in
3579 the Cygwin environment as well as with the MinGW compiler.
3582 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3584 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3585 via timing by performing a MAC computation even if incorrrect
3586 block cipher padding has been found. This is a countermeasure
3587 against active attacks where the attacker has to distinguish
3588 between bad padding and a MAC verification error. (CVE-2003-0078)
3590 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3591 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3592 Martin Vuagnoux (EPFL, Ilion)]
3594 *) Make the no-err option work as intended. The intention with no-err
3595 is not to have the whole error stack handling routines removed from
3596 libcrypto, it's only intended to remove all the function name and
3597 reason texts, thereby removing some of the footprint that may not
3598 be interesting if those errors aren't displayed anyway.
3600 NOTE: it's still possible for any application or module to have it's
3601 own set of error texts inserted. The routines are there, just not
3602 used by default when no-err is given.
3605 *) Add support for FreeBSD on IA64.
3606 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3608 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3609 Kerberos function mit_des_cbc_cksum(). Before this change,
3610 the value returned by DES_cbc_cksum() was like the one from
3611 mit_des_cbc_cksum(), except the bytes were swapped.
3612 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3614 *) Allow an application to disable the automatic SSL chain building.
3615 Before this a rather primitive chain build was always performed in
3616 ssl3_output_cert_chain(): an application had no way to send the
3617 correct chain if the automatic operation produced an incorrect result.
3619 Now the chain builder is disabled if either:
3621 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3623 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3625 The reasoning behind this is that an application would not want the
3626 auto chain building to take place if extra chain certificates are
3627 present and it might also want a means of sending no additional
3628 certificates (for example the chain has two certificates and the
3632 *) Add the possibility to build without the ENGINE framework.
3633 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3635 *) Under Win32 gmtime() can return NULL: check return value in
3636 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3639 *) DSA routines: under certain error conditions uninitialized BN objects
3640 could be freed. Solution: make sure initialization is performed early
3641 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3642 Nils Larsch <nla@trustcenter.de> via PR#459)
3645 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3646 checked on reconnect on the client side, therefore session resumption
3647 could still fail with a "ssl session id is different" error. This
3648 behaviour is masked when SSL_OP_ALL is used due to
3649 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3650 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3651 followup to PR #377.
3654 *) IA-32 assembler support enhancements: unified ELF targets, support
3655 for SCO/Caldera platforms, fix for Cygwin shared build.
3658 *) Add support for FreeBSD on sparc64. As a consequence, support for
3659 FreeBSD on non-x86 processors is separate from x86 processors on
3660 the config script, much like the NetBSD support.
3661 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3663 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3665 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3668 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3669 code (06) was taken as the first octet of the session ID and the last
3670 octet was ignored consequently. As a result SSLv2 client side session
3671 caching could not have worked due to the session ID mismatch between
3673 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3677 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3678 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3682 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3683 seems that in spite of existing for more than a year, many application
3684 author have done nothing to provide the necessary callbacks, which
3685 means that this particular engine will not work properly anywhere.
3686 This is a very unfortunate situation which forces us, in the name
3687 of usability, to give the hw_ncipher.c a static lock, which is part
3689 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3690 appear in 0.9.8 or later. We EXPECT application authors to have
3691 dealt properly with this when 0.9.8 is released (unless we actually
3692 make such changes in the libcrypto locking code that changes will
3693 have to be made anyway).
3696 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3697 octets have been read, EOF or an error occurs. Without this change
3698 some truncated ASN1 structures will not produce an error.
3701 *) Disable Heimdal support, since it hasn't been fully implemented.
3702 Still give the possibility to force the use of Heimdal, but with
3703 warnings and a request that patches get sent to openssl-dev.
3706 *) Add the VC-CE target, introduce the WINCE sysname, and add
3707 INSTALL.WCE and appropriate conditionals to make it build.
3708 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3710 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3711 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3712 edit numbers of the version.
3713 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3715 *) Introduce safe string copy and catenation functions
3716 (BUF_strlcpy() and BUF_strlcat()).
3717 [Ben Laurie (CHATS) and Richard Levitte]
3719 *) Avoid using fixed-size buffers for one-line DNs.
3720 [Ben Laurie (CHATS)]
3722 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3723 resizing buffers containing secrets, and use where appropriate.
3724 [Ben Laurie (CHATS)]
3726 *) Avoid using fixed size buffers for configuration file location.
3727 [Ben Laurie (CHATS)]
3729 *) Avoid filename truncation for various CA files.
3730 [Ben Laurie (CHATS)]
3732 *) Use sizeof in preference to magic numbers.
3733 [Ben Laurie (CHATS)]
3735 *) Avoid filename truncation in cert requests.
3736 [Ben Laurie (CHATS)]
3738 *) Add assertions to check for (supposedly impossible) buffer
3740 [Ben Laurie (CHATS)]
3742 *) Don't cache truncated DNS entries in the local cache (this could
3743 potentially lead to a spoofing attack).
3744 [Ben Laurie (CHATS)]
3746 *) Fix various buffers to be large enough for hex/decimal
3747 representations in a platform independent manner.
3748 [Ben Laurie (CHATS)]
3750 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3751 resizing buffers containing secrets, and use where appropriate.
3752 [Ben Laurie (CHATS)]
3754 *) Add BIO_indent() to avoid much slightly worrying code to do
3756 [Ben Laurie (CHATS)]
3758 *) Convert sprintf()/BIO_puts() to BIO_printf().
3759 [Ben Laurie (CHATS)]
3761 *) buffer_gets() could terminate with the buffer only half
3763 [Ben Laurie (CHATS)]
3765 *) Add assertions to prevent user-supplied crypto functions from
3766 overflowing internal buffers by having large block sizes, etc.
3767 [Ben Laurie (CHATS)]
3769 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3771 [Ben Laurie (CHATS)]
3773 *) Eliminate unused copy of key in RC4.
3774 [Ben Laurie (CHATS)]
3776 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3777 [Ben Laurie (CHATS)]
3779 *) Fix off-by-one error in EGD path.
3780 [Ben Laurie (CHATS)]
3782 *) If RANDFILE path is too long, ignore instead of truncating.
3783 [Ben Laurie (CHATS)]
3785 *) Eliminate unused and incorrectly sized X.509 structure
3787 [Ben Laurie (CHATS)]
3789 *) Eliminate unused and dangerous function knumber().
3790 [Ben Laurie (CHATS)]
3792 *) Eliminate unused and dangerous structure, KSSL_ERR.
3793 [Ben Laurie (CHATS)]
3795 *) Protect against overlong session ID context length in an encoded
3796 session object. Since these are local, this does not appear to be
3798 [Ben Laurie (CHATS)]
3800 *) Change from security patch (see 0.9.6e below) that did not affect
3801 the 0.9.6 release series:
3803 Remote buffer overflow in SSL3 protocol - an attacker could
3804 supply an oversized master key in Kerberos-enabled versions.
3806 [Ben Laurie (CHATS)]
3808 *) Change the SSL kerb5 codes to match RFC 2712.
3811 *) Make -nameopt work fully for req and add -reqopt switch.
3812 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3814 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3815 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3817 *) Make sure tests can be performed even if the corresponding algorithms
3818 have been removed entirely. This was also the last step to make
3819 OpenSSL compilable with DJGPP under all reasonable conditions.
3820 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3822 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3823 to allow version independent disabling of normally unselected ciphers,
3824 which may be activated as a side-effect of selecting a single cipher.
3826 (E.g., cipher list string "RSA" enables ciphersuites that are left
3827 out of "ALL" because they do not provide symmetric encryption.
3828 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3829 [Lutz Jaenicke, Bodo Moeller]
3831 *) Add appropriate support for separate platform-dependent build
3832 directories. The recommended way to make a platform-dependent
3833 build directory is the following (tested on Linux), maybe with
3836 # Place yourself outside of the OpenSSL source tree. In
3837 # this example, the environment variable OPENSSL_SOURCE
3838 # is assumed to contain the absolute OpenSSL source directory.
3839 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3840 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3841 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3842 mkdir -p `dirname $F`
3843 ln -s $OPENSSL_SOURCE/$F $F
3846 To be absolutely sure not to disturb the source tree, a "make clean"
3847 is a good thing. If it isn't successfull, don't worry about it,
3848 it probably means the source directory is very clean.
3851 *) Make sure any ENGINE control commands make local copies of string
3852 pointers passed to them whenever necessary. Otherwise it is possible
3853 the caller may have overwritten (or deallocated) the original string
3854 data when a later ENGINE operation tries to use the stored values.
3855 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3857 *) Improve diagnostics in file reading and command-line digests.
3858 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3860 *) Add AES modes CFB and OFB to the object database. Correct an
3861 error in AES-CFB decryption.
3864 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3865 allows existing EVP_CIPHER_CTX structures to be reused after
3866 calling EVP_*Final(). This behaviour is used by encryption
3867 BIOs and some applications. This has the side effect that
3868 applications must explicitly clean up cipher contexts with
3869 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3872 *) Check the values of dna and dnb in bn_mul_recursive before calling
3873 bn_mul_comba (a non zero value means the a or b arrays do not contain
3874 n2 elements) and fallback to bn_mul_normal if either is not zero.
3877 *) Fix escaping of non-ASCII characters when using the -subj option
3878 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3881 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3882 form for "surname", serialNumber has no short form.
3883 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3884 therefore remove "mail" short name for "internet 7".
3885 The OID for unique identifiers in X509 certificates is
3886 x500UniqueIdentifier, not uniqueIdentifier.
3887 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3890 *) Add an "init" command to the ENGINE config module and auto initialize
3891 ENGINEs. Without any "init" command the ENGINE will be initialized
3892 after all ctrl commands have been executed on it. If init=1 the
3893 ENGINE is initailized at that point (ctrls before that point are run
3894 on the uninitialized ENGINE and after on the initialized one). If
3895 init=0 then the ENGINE will not be iniatialized at all.
3898 *) Fix the 'app_verify_callback' interface so that the user-defined
3899 argument is actually passed to the callback: In the
3900 SSL_CTX_set_cert_verify_callback() prototype, the callback
3901 declaration has been changed from
3904 int (*cb)(X509_STORE_CTX *,void *);
3905 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3906 i=s->ctx->app_verify_callback(&ctx)
3907 has been changed into
3908 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3910 To update applications using SSL_CTX_set_cert_verify_callback(),
3911 a dummy argument can be added to their callback functions.
3912 [D. K. Smetters <smetters@parc.xerox.com>]
3914 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3915 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3917 *) Add and OPENSSL_LOAD_CONF define which will cause
3918 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3919 This allows older applications to transparently support certain
3920 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3921 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3922 load the config file and OPENSSL_add_all_algorithms_conf() which will
3923 always load it have also been added.
3926 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3927 Adjust NIDs and EVP layer.
3928 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3930 *) Config modules support in openssl utility.
3932 Most commands now load modules from the config file,
3933 though in a few (such as version) this isn't done
3934 because it couldn't be used for anything.
3936 In the case of ca and req the config file used is
3937 the same as the utility itself: that is the -config
3938 command line option can be used to specify an
3942 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3943 use "openssl_conf" if filename is NULL use default openssl config file.
3946 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3947 config section name. Add a new flag to tolerate a missing config file
3948 and move code to CONF_modules_load_file().
3951 *) Support for crypto accelerator cards from Accelerated Encryption
3952 Processing, www.aep.ie. (Use engine 'aep')
3953 The support was copied from 0.9.6c [engine] and adapted/corrected
3954 to work with the new engine framework.
3955 [AEP Inc. and Richard Levitte]
3957 *) Support for SureWare crypto accelerator cards from Baltimore
3958 Technologies. (Use engine 'sureware')
3959 The support was copied from 0.9.6c [engine] and adapted
3960 to work with the new engine framework.
3963 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3964 make the newer ENGINE framework commands for the CHIL engine work.
3965 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3967 *) Make it possible to produce shared libraries on ReliantUNIX.
3968 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3970 *) Add the configuration target debug-linux-ppro.
3971 Make 'openssl rsa' use the general key loading routines
3972 implemented in apps.c, and make those routines able to
3973 handle the key format FORMAT_NETSCAPE and the variant
3975 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3977 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3978 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3980 *) Add -keyform to rsautl, and document -engine.
3981 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3983 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3984 BIO_R_NO_SUCH_FILE error code rather than the generic
3985 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3988 *) Add new functions
3990 ERR_peek_last_error_line
3991 ERR_peek_last_error_line_data.
3992 These are similar to
3995 ERR_peek_error_line_data,
3996 but report on the latest error recorded rather than the first one
3997 still in the error queue.
3998 [Ben Laurie, Bodo Moeller]
4000 *) default_algorithms option in ENGINE config module. This allows things
4002 default_algorithms = ALL
4003 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4006 *) Prelminary ENGINE config module.
4009 *) New experimental application configuration code.
4012 *) Change the AES code to follow the same name structure as all other
4013 symmetric ciphers, and behave the same way. Move everything to
4014 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4015 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4017 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4018 [Ben Laurie and Theo de Raadt]
4020 *) Add option to output public keys in req command.
4021 [Massimiliano Pala madwolf@openca.org]
4023 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4024 (up to about 10% better than before for P-192 and P-224).
4027 *) New functions/macros
4029 SSL_CTX_set_msg_callback(ctx, cb)
4030 SSL_CTX_set_msg_callback_arg(ctx, arg)
4031 SSL_set_msg_callback(ssl, cb)
4032 SSL_set_msg_callback_arg(ssl, arg)
4034 to request calling a callback function
4036 void cb(int write_p, int version, int content_type,
4037 const void *buf, size_t len, SSL *ssl, void *arg)
4039 whenever a protocol message has been completely received
4040 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4041 protocol version according to which the SSL library interprets
4042 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4043 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4044 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4045 specification (change_cipher_spec(20), alert(21), handshake(22)).
4046 'buf' and 'len' point to the actual message, 'ssl' to the
4047 SSL object, and 'arg' is the application-defined value set by
4048 SSL[_CTX]_set_msg_callback_arg().
4050 'openssl s_client' and 'openssl s_server' have new '-msg' options
4051 to enable a callback that displays all protocol messages.
4054 *) Change the shared library support so shared libraries are built as
4055 soon as the corresponding static library is finished, and thereby get
4056 openssl and the test programs linked against the shared library.
4057 This still only happens when the keyword "shard" has been given to
4058 the configuration scripts.
4060 NOTE: shared library support is still an experimental thing, and
4061 backward binary compatibility is still not guaranteed.
4062 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4064 *) Add support for Subject Information Access extension.
4065 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4067 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4068 additional bytes when new memory had to be allocated, not just
4069 when reusing an existing buffer.
4072 *) New command line and configuration option 'utf8' for the req command.
4073 This allows field values to be specified as UTF8 strings.
4076 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4077 runs for the former and machine-readable output for the latter.
4080 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4081 of the e-mail address in the DN (i.e., it will go into a certificate
4082 extension only). The new configuration file option 'email_in_dn = no'
4083 has the same effect.
4084 [Massimiliano Pala madwolf@openca.org]
4086 *) Change all functions with names starting with des_ to be starting
4087 with DES_ instead. Add wrappers that are compatible with libdes,
4088 but are named _ossl_old_des_*. Finally, add macros that map the
4089 des_* symbols to the corresponding _ossl_old_des_* if libdes
4090 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4091 desired, the des_* symbols will be mapped to DES_*, with one
4094 Since we provide two compatibility mappings, the user needs to
4095 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4096 compatibility is desired. The default (i.e., when that macro
4097 isn't defined) is OpenSSL 0.9.6c compatibility.
4099 There are also macros that enable and disable the support of old
4100 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4101 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4102 are defined, the default will apply: to support the old des routines.
4104 In either case, one must include openssl/des.h to get the correct
4105 definitions. Do not try to just include openssl/des_old.h, that
4108 NOTE: This is a major break of an old API into a new one. Software
4109 authors are encouraged to switch to the DES_ style functions. Some
4110 time in the future, des_old.h and the libdes compatibility functions
4111 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4112 default), and then completely removed.
4115 *) Test for certificates which contain unsupported critical extensions.
4116 If such a certificate is found during a verify operation it is
4117 rejected by default: this behaviour can be overridden by either
4118 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4119 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4120 X509_supported_extension() has also been added which returns 1 if a
4121 particular extension is supported.
4124 *) Modify the behaviour of EVP cipher functions in similar way to digests
4125 to retain compatibility with existing code.
4128 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4129 compatibility with existing code. In particular the 'ctx' parameter does
4130 not have to be to be initialized before the call to EVP_DigestInit() and
4131 it is tidied up after a call to EVP_DigestFinal(). New function
4132 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4133 EVP_MD_CTX_copy() changed to not require the destination to be
4134 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4135 requires the destination to be valid.
4137 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4138 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4141 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4142 so that complete 'Handshake' protocol structures are kept in memory
4143 instead of overwriting 'msg_type' and 'length' with 'body' data.
4146 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4147 [Massimo Santin via Richard Levitte]
4149 *) Major restructuring to the underlying ENGINE code. This includes
4150 reduction of linker bloat, separation of pure "ENGINE" manipulation
4151 (initialisation, etc) from functionality dealing with implementations
4152 of specific crypto iterfaces. This change also introduces integrated
4153 support for symmetric ciphers and digest implementations - so ENGINEs
4154 can now accelerate these by providing EVP_CIPHER and EVP_MD
4155 implementations of their own. This is detailed in crypto/engine/README
4156 as it couldn't be adequately described here. However, there are a few
4157 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4158 were changed in the original introduction of ENGINE code have now
4159 reverted back - the hooking from this code to ENGINE is now a good
4160 deal more passive and at run-time, operations deal directly with
4161 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4162 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4163 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4164 they were not being used by the framework as there is no concept of a
4165 BIGNUM_METHOD and they could not be generalised to the new
4166 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4167 ENGINE_cpy() has been removed as it cannot be consistently defined in
4171 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4174 *) Change mkdef.pl to sort symbols that get the same entry number,
4175 and make sure the automatically generated functions ERR_load_*
4176 become part of libeay.num as well.
4179 *) New function SSL_renegotiate_pending(). This returns true once
4180 renegotiation has been requested (either SSL_renegotiate() call
4181 or HelloRequest/ClientHello receveived from the peer) and becomes
4182 false once a handshake has been completed.
4183 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4184 sends a HelloRequest, but does not ensure that a handshake takes
4185 place. SSL_renegotiate_pending() is useful for checking if the
4186 client has followed the request.)
4189 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4190 By default, clients may request session resumption even during
4191 renegotiation (if session ID contexts permit); with this option,
4192 session resumption is possible only in the first handshake.
4194 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4195 more bits available for options that should not be part of
4196 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4199 *) Add some demos for certificate and certificate request creation.
4202 *) Make maximum certificate chain size accepted from the peer application
4203 settable (SSL*_get/set_max_cert_list()), as proposed by
4204 "Douglas E. Engert" <deengert@anl.gov>.
4207 *) Add support for shared libraries for Unixware-7
4208 (Boyd Lynn Gerber <gerberb@zenez.com>).
4211 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4212 be done prior to destruction. Use this to unload error strings from
4213 ENGINEs that load their own error strings. NB: This adds two new API
4214 functions to "get" and "set" this destroy handler in an ENGINE.
4217 *) Alter all existing ENGINE implementations (except "openssl" and
4218 "openbsd") to dynamically instantiate their own error strings. This
4219 makes them more flexible to be built both as statically-linked ENGINEs
4220 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4221 Also, add stub code to each that makes building them as self-contained
4222 shared-libraries easier (see README.ENGINE).
4225 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4226 implementations into applications that are completely implemented in
4227 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4228 commands that can be used to configure what shared-library to load and
4229 to control aspects of the way it is handled. Also, made an update to
4230 the README.ENGINE file that brings its information up-to-date and
4231 provides some information and instructions on the "dynamic" ENGINE
4232 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4235 *) Make it possible to unload ranges of ERR strings with a new
4236 "ERR_unload_strings" function.
4239 *) Add a copy() function to EVP_MD.
4242 *) Make EVP_MD routines take a context pointer instead of just the
4243 md_data void pointer.
4246 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4247 that the digest can only process a single chunk of data
4248 (typically because it is provided by a piece of
4249 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4250 is only going to provide a single chunk of data, and hence the
4251 framework needn't accumulate the data for oneshot drivers.
4254 *) As with "ERR", make it possible to replace the underlying "ex_data"
4255 functions. This change also alters the storage and management of global
4256 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4257 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4258 index counters. The API functions that use this state have been changed
4259 to take a "class_index" rather than pointers to the class's local STACK
4260 and counter, and there is now an API function to dynamically create new
4261 classes. This centralisation allows us to (a) plug a lot of the
4262 thread-safety problems that existed, and (b) makes it possible to clean
4263 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4264 such data would previously have always leaked in application code and
4265 workarounds were in place to make the memory debugging turn a blind eye
4266 to it. Application code that doesn't use this new function will still
4267 leak as before, but their memory debugging output will announce it now
4268 rather than letting it slide.
4270 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4271 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4272 has a return value to indicate success or failure.
4275 *) Make it possible to replace the underlying "ERR" functions such that the
4276 global state (2 LHASH tables and 2 locks) is only used by the "default"
4277 implementation. This change also adds two functions to "get" and "set"
4278 the implementation prior to it being automatically set the first time
4279 any other ERR function takes place. Ie. an application can call "get",
4280 pass the return value to a module it has just loaded, and that module
4281 can call its own "set" function using that value. This means the
4282 module's "ERR" operations will use (and modify) the error state in the
4283 application and not in its own statically linked copy of OpenSSL code.
4286 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4287 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4288 the operation, and provides a more encapsulated way for external code
4289 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4290 to use these functions rather than manually incrementing the counts.
4292 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4295 *) Add EVP test program.
4298 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4301 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4302 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4303 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4304 These allow a CRL to be built without having to access X509_CRL fields
4305 directly. Modify 'ca' application to use new functions.
4308 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4309 bug workarounds. Rollback attack detection is a security feature.
4310 The problem will only arise on OpenSSL servers when TLSv1 is not
4311 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4312 Software authors not wanting to support TLSv1 will have special reasons
4313 for their choice and can explicitly enable this option.
4314 [Bodo Moeller, Lutz Jaenicke]
4316 *) Rationalise EVP so it can be extended: don't include a union of
4317 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4318 (similar to those existing for EVP_CIPHER_CTX).
4323 EVP_MD_CTX_init(&md); /* new function call */
4324 EVP_DigestInit(&md, EVP_sha1());
4325 EVP_DigestUpdate(&md, in, len);
4326 EVP_DigestFinal(&md, out, NULL);
4327 EVP_MD_CTX_cleanup(&md); /* new function call */
4331 *) Make DES key schedule conform to the usual scheme, as well as
4332 correcting its structure. This means that calls to DES functions
4333 now have to pass a pointer to a des_key_schedule instead of a
4334 plain des_key_schedule (which was actually always a pointer
4337 des_key_schedule ks;
4339 des_set_key_checked(..., &ks);
4340 des_ncbc_encrypt(..., &ks, ...);
4342 (Note that a later change renames 'des_...' into 'DES_...'.)
4345 *) Initial reduction of linker bloat: the use of some functions, such as
4346 PEM causes large amounts of unused functions to be linked in due to
4347 poor organisation. For example pem_all.c contains every PEM function
4348 which has a knock on effect of linking in large amounts of (unused)
4349 ASN1 code. Grouping together similar functions and splitting unrelated
4350 functions prevents this.
4353 *) Cleanup of EVP macros.
4356 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4357 correct _ecb suffix.
4360 *) Add initial OCSP responder support to ocsp application. The
4361 revocation information is handled using the text based index
4362 use by the ca application. The responder can either handle
4363 requests generated internally, supplied in files (for example
4364 via a CGI script) or using an internal minimal server.
4367 *) Add configuration choices to get zlib compression for TLS.
4370 *) Changes to Kerberos SSL for RFC 2712 compliance:
4371 1. Implemented real KerberosWrapper, instead of just using
4372 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4373 2. Implemented optional authenticator field of KerberosWrapper.
4375 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4376 and authenticator structs; see crypto/krb5/.
4378 Generalized Kerberos calls to support multiple Kerberos libraries.
4379 [Vern Staats <staatsvr@asc.hpc.mil>,
4380 Jeffrey Altman <jaltman@columbia.edu>
4381 via Richard Levitte]
4383 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4384 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4385 values for each of the key sizes rather than having just
4386 parameters (and 'speed' generating keys each time).
4389 *) Speed up EVP routines.
4392 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4393 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4394 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4395 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4397 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4398 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4399 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4402 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4404 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4407 *) Added the OS2-EMX target.
4408 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4410 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4411 to support NCONF routines in extension code. New function CONF_set_nconf()
4412 to allow functions which take an NCONF to also handle the old LHASH
4413 structure: this means that the old CONF compatible routines can be
4414 retained (in particular wrt extensions) without having to duplicate the
4415 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4418 *) Enhance the general user interface with mechanisms for inner control
4419 and with possibilities to have yes/no kind of prompts.
4422 *) Change all calls to low level digest routines in the library and
4423 applications to use EVP. Add missing calls to HMAC_cleanup() and
4424 don't assume HMAC_CTX can be copied using memcpy().
4425 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4427 *) Add the possibility to control engines through control names but with
4428 arbitrary arguments instead of just a string.
4429 Change the key loaders to take a UI_METHOD instead of a callback
4430 function pointer. NOTE: this breaks binary compatibility with earlier
4431 versions of OpenSSL [engine].
4432 Adapt the nCipher code for these new conditions and add a card insertion
4436 *) Enhance the general user interface with mechanisms to better support
4437 dialog box interfaces, application-defined prompts, the possibility
4438 to use defaults (for example default passwords from somewhere else)
4439 and interrupts/cancellations.
4442 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4443 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4446 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4447 tidy up some unnecessarily weird code in 'sk_new()').
4448 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4450 *) Change the key loading routines for ENGINEs to use the same kind
4451 callback (pem_password_cb) as all other routines that need this
4455 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4456 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4457 than this minimum value is recommended.
4460 *) New random seeder for OpenVMS, using the system process statistics
4461 that are easily reachable.
4464 *) Windows apparently can't transparently handle global
4465 variables defined in DLLs. Initialisations such as:
4467 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4469 wont compile. This is used by the any applications that need to
4470 declare their own ASN1 modules. This was fixed by adding the option
4471 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4472 needed for static libraries under Win32.
4475 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4476 setting of purpose and trust fields. New X509_STORE trust and
4477 purpose functions and tidy up setting in other SSL functions.
4480 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4481 structure. These are inherited by X509_STORE_CTX when it is
4482 initialised. This allows various defaults to be set in the
4483 X509_STORE structure (such as flags for CRL checking and custom
4484 purpose or trust settings) for functions which only use X509_STORE_CTX
4485 internally such as S/MIME.
4487 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4488 trust settings if they are not set in X509_STORE. This allows X509_STORE
4489 purposes and trust (in S/MIME for example) to override any set by default.
4491 Add command line options for CRL checking to smime, s_client and s_server
4495 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4496 are set then the CRL is looked up in the X509_STORE structure and
4497 its validity and signature checked, then if the certificate is found
4498 in the CRL the verify fails with a revoked error.
4500 Various new CRL related callbacks added to X509_STORE_CTX structure.
4502 Command line options added to 'verify' application to support this.
4504 This needs some additional work, such as being able to handle multiple
4505 CRLs with different times, extension based lookup (rather than just
4506 by subject name) and ultimately more complete V2 CRL extension
4510 *) Add a general user interface API (crypto/ui/). This is designed
4511 to replace things like des_read_password and friends (backward
4512 compatibility functions using this new API are provided).
4513 The purpose is to remove prompting functions from the DES code
4514 section as well as provide for prompting through dialog boxes in
4515 a window system and the like.
4518 *) Add "ex_data" support to ENGINE so implementations can add state at a
4519 per-structure level rather than having to store it globally.
4522 *) Make it possible for ENGINE structures to be copied when retrieved by
4523 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4524 This causes the "original" ENGINE structure to act like a template,
4525 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4526 operational state can be localised to each ENGINE structure, despite the
4527 fact they all share the same "methods". New ENGINE structures returned in
4528 this case have no functional references and the return value is the single
4529 structural reference. This matches the single structural reference returned
4530 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4534 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4535 needs to match any other type at all we need to manually clear the
4539 *) Changes to the "openssl engine" utility to include;
4540 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4541 about an ENGINE's available control commands.
4542 - executing control commands from command line arguments using the
4543 '-pre' and '-post' switches. '-post' is only used if '-t' is
4544 specified and the ENGINE is successfully initialised. The syntax for
4545 the individual commands are colon-separated, for example;
4546 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4549 *) New dynamic control command support for ENGINEs. ENGINEs can now
4550 declare their own commands (numbers), names (strings), descriptions,
4551 and input types for run-time discovery by calling applications. A
4552 subset of these commands are implicitly classed as "executable"
4553 depending on their input type, and only these can be invoked through
4554 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4555 can be based on user input, config files, etc). The distinction is
4556 that "executable" commands cannot return anything other than a boolean
4557 result and can only support numeric or string input, whereas some
4558 discoverable commands may only be for direct use through
4559 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4560 pointers, or other custom uses. The "executable" commands are to
4561 support parameterisations of ENGINE behaviour that can be
4562 unambiguously defined by ENGINEs and used consistently across any
4563 OpenSSL-based application. Commands have been added to all the
4564 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4565 control over shared-library paths without source code alterations.
4568 *) Changed all ENGINE implementations to dynamically allocate their
4569 ENGINEs rather than declaring them statically. Apart from this being
4570 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4571 this also allows the implementations to compile without using the
4572 internal engine_int.h header.
4575 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4576 'const' value. Any code that should be able to modify a RAND_METHOD
4577 should already have non-const pointers to it (ie. they should only
4578 modify their own ones).
4581 *) Made a variety of little tweaks to the ENGINE code.
4582 - "atalla" and "ubsec" string definitions were moved from header files
4583 to C code. "nuron" string definitions were placed in variables
4584 rather than hard-coded - allowing parameterisation of these values
4585 later on via ctrl() commands.
4586 - Removed unused "#if 0"'d code.
4587 - Fixed engine list iteration code so it uses ENGINE_free() to release
4588 structural references.
4589 - Constified the RAND_METHOD element of ENGINE structures.
4590 - Constified various get/set functions as appropriate and added
4591 missing functions (including a catch-all ENGINE_cpy that duplicates
4592 all ENGINE values onto a new ENGINE except reference counts/state).
4593 - Removed NULL parameter checks in get/set functions. Setting a method
4594 or function to NULL is a way of cancelling out a previously set
4595 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4596 and doesn't justify the extra error symbols and code.
4597 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4598 flags from engine_int.h to engine.h.
4599 - Changed prototypes for ENGINE handler functions (init(), finish(),
4600 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4603 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4604 to the algorithm using long division. The binary algorithm can be
4605 used only if the modulus is odd. On 32-bit systems, it is faster
4606 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4607 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4608 up to 450 bits. In 64-bit environments, the binary algorithm
4609 appears to be advantageous for much longer moduli; here we use it
4610 for moduli up to 2048 bits.
4613 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4614 could not support the combine flag in choice fields.
4617 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4618 extensions from a certificate request to the certificate.
4621 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4622 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4623 file: this allows the display of the certificate about to be
4624 signed to be customised, to allow certain fields to be included
4625 or excluded and extension details. The old system didn't display
4626 multicharacter strings properly, omitted fields not in the policy
4627 and couldn't display additional details such as extensions.
4630 *) Function EC_POINTs_mul for multiple scalar multiplication
4631 of an arbitrary number of elliptic curve points
4632 \sum scalars[i]*points[i],
4633 optionally including the generator defined for the EC_GROUP:
4634 scalar*generator + \sum scalars[i]*points[i].
4636 EC_POINT_mul is a simple wrapper function for the typical case
4637 that the point list has just one item (besides the optional
4641 *) First EC_METHODs for curves over GF(p):
4643 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4644 operations and provides various method functions that can also
4645 operate with faster implementations of modular arithmetic.
4647 EC_GFp_mont_method() reuses most functions that are part of
4648 EC_GFp_simple_method, but uses Montgomery arithmetic.
4650 [Bodo Moeller; point addition and point doubling
4651 implementation directly derived from source code provided by
4652 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4654 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4655 crypto/ec/ec_lib.c):
4657 Curves are EC_GROUP objects (with an optional group generator)
4658 based on EC_METHODs that are built into the library.
4660 Points are EC_POINT objects based on EC_GROUP objects.
4662 Most of the framework would be able to handle curves over arbitrary
4663 finite fields, but as there are no obvious types for fields other
4664 than GF(p), some functions are limited to that for now.
4667 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4668 that the file contains a complete HTTP response.
4671 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4672 change the def and num file printf format specifier from "%-40sXXX"
4673 to "%-39s XXX". The latter will always guarantee a space after the
4674 field while the former will cause them to run together if the field
4675 is 40 of more characters long.
4678 *) Constify the cipher and digest 'method' functions and structures
4679 and modify related functions to take constant EVP_MD and EVP_CIPHER
4683 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4684 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4687 *) Modify EVP_Digest*() routines so they now return values. Although the
4688 internal software routines can never fail additional hardware versions
4692 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4694 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4695 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4698 ERR_R_NESTED_ASN1_ERROR
4700 ERR_R_MISSING_ASN1_EOS
4701 were 4 .. 9, conflicting with
4702 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4704 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4705 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4707 Add new error code 'ERR_R_INTERNAL_ERROR'.
4710 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4714 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4715 sets the subject name for a new request or supersedes the
4716 subject name in a given request. Formats that can be parsed are
4717 'CN=Some Name, OU=myOU, C=IT'
4719 'CN=Some Name/OU=myOU/C=IT'.
4721 Add options '-batch' and '-verbose' to 'openssl req'.
4722 [Massimiliano Pala <madwolf@hackmasters.net>]
4724 *) Introduce the possibility to access global variables through
4725 functions on platform were that's the best way to handle exporting
4726 global variables in shared libraries. To enable this functionality,
4727 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4728 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4729 is normally done by Configure or something similar).
4731 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4732 in the source file (foo.c) like this:
4734 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4735 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4737 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4738 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4740 OPENSSL_DECLARE_GLOBAL(int,foo);
4741 #define foo OPENSSL_GLOBAL_REF(foo)
4742 OPENSSL_DECLARE_GLOBAL(double,bar);
4743 #define bar OPENSSL_GLOBAL_REF(bar)
4745 The #defines are very important, and therefore so is including the
4746 header file everywhere where the defined globals are used.
4748 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4749 of ASN.1 items, but that structure is a bit different.
4751 The largest change is in util/mkdef.pl which has been enhanced with
4752 better and easier to understand logic to choose which symbols should
4753 go into the Windows .def files as well as a number of fixes and code
4754 cleanup (among others, algorithm keywords are now sorted
4755 lexicographically to avoid constant rewrites).
4758 *) In BN_div() keep a copy of the sign of 'num' before writing the
4759 result to 'rm' because if rm==num the value will be overwritten
4760 and produce the wrong result if 'num' is negative: this caused
4761 problems with BN_mod() and BN_nnmod().
4764 *) Function OCSP_request_verify(). This checks the signature on an
4765 OCSP request and verifies the signer certificate. The signer
4766 certificate is just checked for a generic purpose and OCSP request
4770 *) Add OCSP_check_validity() function to check the validity of OCSP
4771 responses. OCSP responses are prepared in real time and may only
4772 be a few seconds old. Simply checking that the current time lies
4773 between thisUpdate and nextUpdate max reject otherwise valid responses
4774 caused by either OCSP responder or client clock inaccuracy. Instead
4775 we allow thisUpdate and nextUpdate to fall within a certain period of
4776 the current time. The age of the response can also optionally be
4777 checked. Two new options -validity_period and -status_age added to
4781 *) If signature or public key algorithm is unrecognized print out its
4782 OID rather that just UNKNOWN.
4785 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4786 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4787 ID to be generated from the issuer certificate alone which can then be
4788 passed to OCSP_id_issuer_cmp().
4791 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4792 ASN1 modules to export functions returning ASN1_ITEM pointers
4793 instead of the ASN1_ITEM structures themselves. This adds several
4794 new macros which allow the underlying ASN1 function/structure to
4795 be accessed transparently. As a result code should not use ASN1_ITEM
4796 references directly (such as &X509_it) but instead use the relevant
4797 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4798 use of the new ASN1 code on platforms where exporting structures
4799 is problematical (for example in shared libraries) but exporting
4800 functions returning pointers to structures is not.
4803 *) Add support for overriding the generation of SSL/TLS session IDs.
4804 These callbacks can be registered either in an SSL_CTX or per SSL.
4805 The purpose of this is to allow applications to control, if they wish,
4806 the arbitrary values chosen for use as session IDs, particularly as it
4807 can be useful for session caching in multiple-server environments. A
4808 command-line switch for testing this (and any client code that wishes
4809 to use such a feature) has been added to "s_server".
4810 [Geoff Thorpe, Lutz Jaenicke]
4812 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4813 of the form '#if defined(...) || defined(...) || ...' and
4814 '#if !defined(...) && !defined(...) && ...'. This also avoids
4815 the growing number of special cases it was previously handling.
4818 *) Make all configuration macros available for application by making
4819 sure they are available in opensslconf.h, by giving them names starting
4820 with "OPENSSL_" to avoid conflicts with other packages and by making
4821 sure e_os2.h will cover all platform-specific cases together with
4823 Additionally, it is now possible to define configuration/platform-
4824 specific names (called "system identities"). In the C code, these
4825 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4826 macro with the name beginning with "OPENSSL_SYS_", which is determined
4827 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4831 *) New option -set_serial to 'req' and 'x509' this allows the serial
4832 number to use to be specified on the command line. Previously self
4833 signed certificates were hard coded with serial number 0 and the
4834 CA options of 'x509' had to use a serial number in a file which was
4838 *) New options to 'ca' utility to support V2 CRL entry extensions.
4839 Currently CRL reason, invalidity date and hold instruction are
4840 supported. Add new CRL extensions to V3 code and some new objects.
4843 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4844 disable standard block padding (aka PKCS#5 padding) in the EVP
4845 API, which was previously mandatory. This means that the data is
4846 not padded in any way and so the total length much be a multiple
4847 of the block size, otherwise an error occurs.
4850 *) Initial (incomplete) OCSP SSL support.
4853 *) New function OCSP_parse_url(). This splits up a URL into its host,
4854 port and path components: primarily to parse OCSP URLs. New -url
4855 option to ocsp utility.
4858 *) New nonce behavior. The return value of OCSP_check_nonce() now
4859 reflects the various checks performed. Applications can decide
4860 whether to tolerate certain situations such as an absent nonce
4861 in a response when one was present in a request: the ocsp application
4862 just prints out a warning. New function OCSP_add1_basic_nonce()
4863 this is to allow responders to include a nonce in a response even if
4864 the request is nonce-less.
4867 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4868 skipped when using openssl x509 multiple times on a single input file,
4869 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4872 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4873 set string type: to handle setting ASN1_TIME structures. Fix ca
4874 utility to correctly initialize revocation date of CRLs.
4877 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4878 the clients preferred ciphersuites and rather use its own preferences.
4879 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4880 Internet Explorer by ensuring unchanged hash method during stepup.
4881 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4884 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4885 to aes and add a new 'exist' option to print out symbols that don't
4889 *) Additional options to ocsp utility to allow flags to be set and
4890 additional certificates supplied.
4893 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4894 OCSP client a number of certificate to only verify the response
4898 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4899 handle the new API. Currently only ECB, CBC modes supported. Add new
4902 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4903 Encryption Standard (AES) Ciphersuites for Transport Layer
4904 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4905 not enabled by default and were not part of the "ALL" ciphersuite
4906 alias because they were not yet official; they could be
4907 explicitly requested by specifying the "AESdraft" ciphersuite
4908 group alias. In the final release of OpenSSL 0.9.7, the group
4909 alias is called "AES" and is part of "ALL".)
4910 [Ben Laurie, Steve Henson, Bodo Moeller]
4912 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4913 request to response.
4916 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4917 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4918 extract information from a certificate request. OCSP_response_create()
4919 creates a response and optionally adds a basic response structure.
4920 OCSP_basic_add1_status() adds a complete single response to a basic
4921 response and returns the OCSP_SINGLERESP structure just added (to allow
4922 extensions to be included for example). OCSP_basic_add1_cert() adds a
4923 certificate to a basic response and OCSP_basic_sign() signs a basic
4924 response with various flags. New helper functions ASN1_TIME_check()
4925 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4926 (converts ASN1_TIME to GeneralizedTime).
4929 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4930 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4931 structure from a certificate. X509_pubkey_digest() digests the public_key
4932 contents: this is used in various key identifiers.
4935 *) Make sk_sort() tolerate a NULL argument.
4936 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4938 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4939 passed by the function are trusted implicitly. If any of them signed the
4940 response then it is assumed to be valid and is not verified.
4943 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4944 to data. This was previously part of the PKCS7 ASN1 code. This
4945 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4946 [Steve Henson, reported by Kenneth R. Robinette
4947 <support@securenetterm.com>]
4949 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4950 routines: without these tracing memory leaks is very painful.
4951 Fix leaks in PKCS12 and PKCS7 routines.
4954 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4955 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4956 effectively meant GeneralizedTime would never be used. Now it
4957 is initialised to -1 but X509_time_adj() now has to check the value
4958 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4959 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4960 [Steve Henson, reported by Kenneth R. Robinette
4961 <support@securenetterm.com>]
4963 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4964 result in a zero length in the ASN1_INTEGER structure which was
4965 not consistent with the structure when d2i_ASN1_INTEGER() was used
4966 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4967 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4968 where it did not print out a minus for negative ASN1_INTEGER.
4971 *) Add summary printout to ocsp utility. The various functions which
4972 convert status values to strings have been renamed to:
4973 OCSP_response_status_str(), OCSP_cert_status_str() and
4974 OCSP_crl_reason_str() and are no longer static. New options
4975 to verify nonce values and to disable verification. OCSP response
4976 printout format cleaned up.
4979 *) Add additional OCSP certificate checks. These are those specified
4980 in RFC2560. This consists of two separate checks: the CA of the
4981 certificate being checked must either be the OCSP signer certificate
4982 or the issuer of the OCSP signer certificate. In the latter case the
4983 OCSP signer certificate must contain the OCSP signing extended key
4984 usage. This check is performed by attempting to match the OCSP
4985 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4986 in the OCSP_CERTID structures of the response.
4989 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4990 and related routines. This uses the standard OpenSSL certificate
4991 verify routines to perform initial checks (just CA validity) and
4992 to obtain the certificate chain. Then additional checks will be
4993 performed on the chain. Currently the root CA is checked to see
4994 if it is explicitly trusted for OCSP signing. This is used to set
4995 a root CA as a global signing root: that is any certificate that
4996 chains to that CA is an acceptable OCSP signing certificate.
4999 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5000 extensions from a separate configuration file.
5001 As when reading extensions from the main configuration file,
5002 the '-extensions ...' option may be used for specifying the
5004 [Massimiliano Pala <madwolf@comune.modena.it>]
5006 *) New OCSP utility. Allows OCSP requests to be generated or
5007 read. The request can be sent to a responder and the output
5008 parsed, outputed or printed in text form. Not complete yet:
5009 still needs to check the OCSP response validity.
5012 *) New subcommands for 'openssl ca':
5013 'openssl ca -status <serial>' prints the status of the cert with
5014 the given serial number (according to the index file).
5015 'openssl ca -updatedb' updates the expiry status of certificates
5017 [Massimiliano Pala <madwolf@comune.modena.it>]
5019 *) New '-newreq-nodes' command option to CA.pl. This is like
5020 '-newreq', but calls 'openssl req' with the '-nodes' option
5021 so that the resulting key is not encrypted.
5022 [Damien Miller <djm@mindrot.org>]
5024 *) New configuration for the GNU Hurd.
5025 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5027 *) Initial code to implement OCSP basic response verify. This
5028 is currently incomplete. Currently just finds the signer's
5029 certificate and verifies the signature on the response.
5032 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5033 value of OPENSSLDIR. This is available via the new '-d' option
5034 to 'openssl version', and is also included in 'openssl version -a'.
5037 *) Allowing defining memory allocation callbacks that will be given
5038 file name and line number information in additional arguments
5039 (a const char* and an int). The basic functionality remains, as
5040 well as the original possibility to just replace malloc(),
5041 realloc() and free() by functions that do not know about these
5042 additional arguments. To register and find out the current
5043 settings for extended allocation functions, the following
5044 functions are provided:
5046 CRYPTO_set_mem_ex_functions
5047 CRYPTO_set_locked_mem_ex_functions
5048 CRYPTO_get_mem_ex_functions
5049 CRYPTO_get_locked_mem_ex_functions
5051 These work the same way as CRYPTO_set_mem_functions and friends.
5052 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5053 extended allocation function is enabled.
5054 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5055 a conventional allocation function is enabled.
5056 [Richard Levitte, Bodo Moeller]
5058 *) Finish off removing the remaining LHASH function pointer casts.
5059 There should no longer be any prototype-casting required when using
5060 the LHASH abstraction, and any casts that remain are "bugs". See
5061 the callback types and macros at the head of lhash.h for details
5062 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5065 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5066 If /dev/[u]random devices are not available or do not return enough
5067 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5069 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5070 /etc/entropy will be queried once each in this sequence, quering stops
5071 when enough entropy was collected without querying more sockets.
5074 *) Change the Unix RAND_poll() variant to be able to poll several
5075 random devices, as specified by DEVRANDOM, until a sufficient amount
5076 of data has been collected. We spend at most 10 ms on each file
5077 (select timeout) and read in non-blocking mode. DEVRANDOM now
5078 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5079 (previously it was just the string "/dev/urandom"), so on typical
5080 platforms the 10 ms delay will never occur.
5081 Also separate out the Unix variant to its own file, rand_unix.c.
5082 For VMS, there's a currently-empty rand_vms.c.
5085 *) Move OCSP client related routines to ocsp_cl.c. These
5086 provide utility functions which an application needing
5087 to issue a request to an OCSP responder and analyse the
5088 response will typically need: as opposed to those which an
5089 OCSP responder itself would need which will be added later.
5091 OCSP_request_sign() signs an OCSP request with an API similar
5092 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5093 response. OCSP_response_get1_basic() extracts basic response
5094 from response. OCSP_resp_find_status(): finds and extracts status
5095 information from an OCSP_CERTID structure (which will be created
5096 when the request structure is built). These are built from lower
5097 level functions which work on OCSP_SINGLERESP structures but
5098 wont normally be used unless the application wishes to examine
5099 extensions in the OCSP response for example.
5101 Replace nonce routines with a pair of functions.
5102 OCSP_request_add1_nonce() adds a nonce value and optionally
5103 generates a random value. OCSP_check_nonce() checks the
5104 validity of the nonce in an OCSP response.
5107 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5108 This doesn't copy the supplied OCSP_CERTID and avoids the
5109 need to free up the newly created id. Change return type
5110 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5111 This can then be used to add extensions to the request.
5112 Deleted OCSP_request_new(), since most of its functionality
5113 is now in OCSP_REQUEST_new() (and the case insensitive name
5114 clash) apart from the ability to set the request name which
5115 will be added elsewhere.
5118 *) Update OCSP API. Remove obsolete extensions argument from
5119 various functions. Extensions are now handled using the new
5120 OCSP extension code. New simple OCSP HTTP function which
5121 can be used to send requests and parse the response.
5124 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5125 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5126 uses the special reorder version of SET OF to sort the attributes
5127 and reorder them to match the encoded order. This resolves a long
5128 standing problem: a verify on a PKCS7 structure just after signing
5129 it used to fail because the attribute order did not match the
5130 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5131 it uses the received order. This is necessary to tolerate some broken
5132 software that does not order SET OF. This is handled by encoding
5133 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5134 to produce the required SET OF.
5137 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5138 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5139 files to get correct declarations of the ASN.1 item variables.
5142 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5143 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5144 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5145 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5146 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5147 ASN1_ITEM and no wrapper functions.
5150 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5151 replace the old function pointer based I/O routines. Change most of
5152 the *_d2i_bio() and *_d2i_fp() functions to use these.
5155 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5156 lines, recognice more "algorithms" that can be deselected, and make
5157 it complain about algorithm deselection that isn't recognised.
5160 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5161 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5162 to use new functions. Add NO_ASN1_OLD which can be set to remove
5163 some old style ASN1 functions: this can be used to determine if old
5164 code will still work when these eventually go away.
5167 *) New extension functions for OCSP structures, these follow the
5168 same conventions as certificates and CRLs.
5171 *) New function X509V3_add1_i2d(). This automatically encodes and
5172 adds an extension. Its behaviour can be customised with various
5173 flags to append, replace or delete. Various wrappers added for
5174 certifcates and CRLs.
5177 *) Fix to avoid calling the underlying ASN1 print routine when
5178 an extension cannot be parsed. Correct a typo in the
5179 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5182 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5183 entries for variables.
5186 *) Add functionality to apps/openssl.c for detecting locking
5187 problems: As the program is single-threaded, all we have
5188 to do is register a locking callback using an array for
5189 storing which locks are currently held by the program.
5192 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5193 SSL_get_ex_data_X509_STORE_idx(), which is used in
5194 ssl_verify_cert_chain() and thus can be called at any time
5195 during TLS/SSL handshakes so that thread-safety is essential.
5196 Unfortunately, the ex_data design is not at all suited
5197 for multi-threaded use, so it probably should be abolished.
5200 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5201 [Broadcom, tweaked and integrated by Geoff Thorpe]
5203 *) Move common extension printing code to new function
5204 X509V3_print_extensions(). Reorganise OCSP print routines and
5205 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5208 *) New function X509_signature_print() to remove duplication in some
5212 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5213 set (this was treated exactly the same as SET OF previously). This
5214 is used to reorder the STACK representing the structure to match the
5215 encoding. This will be used to get round a problem where a PKCS7
5216 structure which was signed could not be verified because the STACK
5217 order did not reflect the encoded order.
5220 *) Reimplement the OCSP ASN1 module using the new code.
5223 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5224 for its ASN1 operations. The old style function pointers still exist
5225 for now but they will eventually go away.
5228 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5229 completely replaces the old ASN1 functionality with a table driven
5230 encoder and decoder which interprets an ASN1_ITEM structure describing
5231 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5232 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5233 has also been converted to the new form.
5236 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5237 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5238 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5239 for negative moduli.
5242 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5243 of not touching the result's sign bit.
5246 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5250 *) Changed the LHASH code to use prototypes for callbacks, and created
5251 macros to declare and implement thin (optionally static) functions
5252 that provide type-safety and avoid function pointer casting for the
5253 type-specific callbacks.
5256 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5258 [Veers Staats <staatsvr@asc.hpc.mil>,
5259 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5261 *) Reformat the FAQ so the different questions and answers can be divided
5262 in sections depending on the subject.
5265 *) Have the zlib compression code load ZLIB.DLL dynamically under
5269 *) New function BN_mod_sqrt for computing square roots modulo a prime
5270 (using the probabilistic Tonelli-Shanks algorithm unless
5271 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5272 be handled deterministically).
5273 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5275 *) Make BN_mod_inverse faster by explicitly handling small quotients
5276 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5277 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5280 *) New function BN_kronecker.
5283 *) Fix BN_gcd so that it works on negative inputs; the result is
5284 positive unless both parameters are zero.
5285 Previously something reasonably close to an infinite loop was
5286 possible because numbers could be growing instead of shrinking
5287 in the implementation of Euclid's algorithm.
5290 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5291 sign of the number in question.
5293 Fix BN_is_word(a,w) to work correctly for w == 0.
5295 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5296 because its test if the absolute value of 'a' equals 'w'.
5297 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5298 it exists mostly for use in the implementations of BN_is_zero(),
5299 BN_is_one(), and BN_is_word().
5302 *) New function BN_swap.
5305 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5306 the exponentiation functions are more likely to produce reasonable
5307 results on negative inputs.
5310 *) Change BN_mod_mul so that the result is always non-negative.
5311 Previously, it could be negative if one of the factors was negative;
5312 I don't think anyone really wanted that behaviour.
5315 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5316 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5317 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5318 and add new functions:
5327 BN_mod_lshift1_quick
5331 These functions always generate non-negative results.
5333 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5334 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5336 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5337 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5338 be reduced modulo m.
5339 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5342 The following entry accidentily appeared in the CHANGES file
5343 distributed with OpenSSL 0.9.7. The modifications described in
5344 it do *not* apply to OpenSSL 0.9.7.
5346 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5347 was actually never needed) and in BN_mul(). The removal in BN_mul()
5348 required a small change in bn_mul_part_recursive() and the addition
5349 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5350 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5351 bn_sub_words() and bn_add_words() except they take arrays with
5356 *) In 'openssl passwd', verify passwords read from the terminal
5357 unless the '-salt' option is used (which usually means that
5358 verification would just waste user's time since the resulting
5359 hash is going to be compared with some given password hash)
5360 or the new '-noverify' option is used.
5362 This is an incompatible change, but it does not affect
5363 non-interactive use of 'openssl passwd' (passwords on the command
5364 line, '-stdin' option, '-in ...' option) and thus should not
5368 *) Remove all references to RSAref, since there's no more need for it.
5371 *) Make DSO load along a path given through an environment variable
5372 (SHLIB_PATH) with shl_load().
5375 *) Constify the ENGINE code as a result of BIGNUM constification.
5376 Also constify the RSA code and most things related to it. In a
5377 few places, most notable in the depth of the ASN.1 code, ugly
5378 casts back to non-const were required (to be solved at a later
5382 *) Make it so the openssl application has all engines loaded by default.
5385 *) Constify the BIGNUM routines a little more.
5388 *) Add the following functions:
5390 ENGINE_load_cswift()
5392 ENGINE_load_atalla()
5394 ENGINE_load_builtin_engines()
5396 That way, an application can itself choose if external engines that
5397 are built-in in OpenSSL shall ever be used or not. The benefit is
5398 that applications won't have to be linked with libdl or other dso
5399 libraries unless it's really needed.
5401 Changed 'openssl engine' to load all engines on demand.
5402 Changed the engine header files to avoid the duplication of some
5403 declarations (they differed!).
5406 *) 'openssl engine' can now list capabilities.
5409 *) Better error reporting in 'openssl engine'.
5412 *) Never call load_dh_param(NULL) in s_server.
5415 *) Add engine application. It can currently list engines by name and
5416 identity, and test if they are actually available.
5419 *) Improve RPM specification file by forcing symbolic linking and making
5420 sure the installed documentation is also owned by root.root.
5421 [Damien Miller <djm@mindrot.org>]
5423 *) Give the OpenSSL applications more possibilities to make use of
5424 keys (public as well as private) handled by engines.
5427 *) Add OCSP code that comes from CertCo.
5430 *) Add VMS support for the Rijndael code.
5433 *) Added untested support for Nuron crypto accelerator.
5436 *) Add support for external cryptographic devices. This code was
5437 previously distributed separately as the "engine" branch.
5438 [Geoff Thorpe, Richard Levitte]
5440 *) Rework the filename-translation in the DSO code. It is now possible to
5441 have far greater control over how a "name" is turned into a filename
5442 depending on the operating environment and any oddities about the
5443 different shared library filenames on each system.
5446 *) Support threads on FreeBSD-elf in Configure.
5449 *) Fix for SHA1 assembly problem with MASM: it produces
5450 warnings about corrupt line number information when assembling
5451 with debugging information. This is caused by the overlapping
5453 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5456 NCONF_get_number() has no error checking at all. As a replacement,
5457 NCONF_get_number_e() is defined (_e for "error checking") and is
5458 promoted strongly. The old NCONF_get_number is kept around for
5459 binary backward compatibility.
5460 Make it possible for methods to load from something other than a BIO,
5461 by providing a function pointer that is given a name instead of a BIO.
5462 For example, this could be used to load configuration data from an
5466 *) Fix for non blocking accept BIOs. Added new I/O special reason
5467 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5468 with non blocking I/O was not possible because no retry code was
5469 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5473 *) Added the beginnings of Rijndael support.
5476 *) Fix for bug in DirectoryString mask setting. Add support for
5477 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5478 to allow certificate printing to more controllable, additional
5479 'certopt' option to 'x509' to allow new printing options to be
5483 *) Clean old EAY MD5 hack from e_os.h.
5486 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5488 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5489 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5490 [Joe Orton, Steve Henson]
5492 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5494 *) Fix additional bug revealed by the NISCC test suite:
5496 Stop bug triggering large recursion when presented with
5497 certain ASN.1 tags (CVE-2003-0851)
5500 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5502 *) Fix various bugs revealed by running the NISCC test suite:
5504 Stop out of bounds reads in the ASN1 code when presented with
5505 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5507 If verify callback ignores invalid public key errors don't try to check
5508 certificate signature with the NULL public key.
5512 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5513 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5517 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5518 extra data after the compression methods not only for TLS 1.0
5519 but also for SSL 3.0 (as required by the specification).
5520 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5522 *) Change X509_certificate_type() to mark the key as exported/exportable
5523 when it's 512 *bits* long, not 512 bytes.
5526 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5528 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5529 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5530 a protocol version number mismatch like a decryption error
5531 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5534 *) Turn on RSA blinding by default in the default implementation
5535 to avoid a timing attack. Applications that don't want it can call
5536 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5537 They would be ill-advised to do so in most cases.
5538 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5540 *) Change RSA blinding code so that it works when the PRNG is not
5541 seeded (in this case, the secret RSA exponent is abused as
5542 an unpredictable seed -- if it is not unpredictable, there
5543 is no point in blinding anyway). Make RSA blinding thread-safe
5544 by remembering the creator's thread ID in rsa->blinding and
5545 having all other threads use local one-time blinding factors
5546 (this requires more computation than sharing rsa->blinding, but
5547 avoids excessive locking; and if an RSA object is not shared
5548 between threads, blinding will still be very fast).
5551 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5553 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5554 via timing by performing a MAC computation even if incorrrect
5555 block cipher padding has been found. This is a countermeasure
5556 against active attacks where the attacker has to distinguish
5557 between bad padding and a MAC verification error. (CVE-2003-0078)
5559 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5560 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5561 Martin Vuagnoux (EPFL, Ilion)]
5563 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5565 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5566 memory from it's contents. This is done with a counter that will
5567 place alternating values in each byte. This can be used to solve
5568 two issues: 1) the removal of calls to memset() by highly optimizing
5569 compilers, and 2) cleansing with other values than 0, since those can
5570 be read through on certain media, for example a swap space on disk.
5573 *) Bugfix: client side session caching did not work with external caching,
5574 because the session->cipher setting was not restored when reloading
5575 from the external cache. This problem was masked, when
5576 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5577 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5580 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5581 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5582 [Zeev Lieber <zeev-l@yahoo.com>]
5584 *) Undo an undocumented change introduced in 0.9.6e which caused
5585 repeated calls to OpenSSL_add_all_ciphers() and
5586 OpenSSL_add_all_digests() to be ignored, even after calling
5590 *) Change the default configuration reader to deal with last line not
5591 being properly terminated.
5594 *) Change X509_NAME_cmp() so it applies the special rules on handling
5595 DN values that are of type PrintableString, as well as RDNs of type
5596 emailAddress where the value has the type ia5String.
5597 [stefank@valicert.com via Richard Levitte]
5599 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5600 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5601 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5602 the bitwise-OR of the two for use by the majority of applications
5603 wanting this behaviour, and update the docs. The documented
5604 behaviour and actual behaviour were inconsistent and had been
5605 changing anyway, so this is more a bug-fix than a behavioural
5607 [Geoff Thorpe, diagnosed by Nadav Har'El]
5609 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5610 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5613 *) Fix initialization code race conditions in
5614 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5615 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5616 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5617 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5618 ssl2_get_cipher_by_char(),
5619 ssl3_get_cipher_by_char().
5620 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5622 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5623 the cached sessions are flushed, as the remove_cb() might use ex_data
5624 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5625 (see [openssl.org #212]).
5626 [Geoff Thorpe, Lutz Jaenicke]
5628 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5629 length, instead of the encoding length to d2i_ASN1_OBJECT.
5632 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5634 *) [In 0.9.6g-engine release:]
5635 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5636 [Lynn Gazis <lgazis@rainbow.com>]
5638 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5640 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5641 and get fix the header length calculation.
5642 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5643 Alon Kantor <alonk@checkpoint.com> (and others),
5646 *) Use proper error handling instead of 'assertions' in buffer
5647 overflow checks added in 0.9.6e. This prevents DoS (the
5648 assertions could call abort()).
5649 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5651 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5653 *) Add various sanity checks to asn1_get_length() to reject
5654 the ASN1 length bytes if they exceed sizeof(long), will appear
5655 negative or the content length exceeds the length of the
5657 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5659 *) Fix cipher selection routines: ciphers without encryption had no flags
5660 for the cipher strength set and where therefore not handled correctly
5661 by the selection routines (PR #130).
5664 *) Fix EVP_dsa_sha macro.
5668 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5669 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5670 that was added in OpenSSL 0.9.6d.
5672 As the countermeasure turned out to be incompatible with some
5673 broken SSL implementations, the new option is part of SSL_OP_ALL.
5674 SSL_OP_ALL is usually employed when compatibility with weird SSL
5675 implementations is desired (e.g. '-bugs' option to 's_client' and
5676 's_server'), so the new option is automatically set in many
5680 *) Changes in security patch:
5682 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5683 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5684 Air Force Materiel Command, USAF, under agreement number
5687 *) Add various sanity checks to asn1_get_length() to reject
5688 the ASN1 length bytes if they exceed sizeof(long), will appear
5689 negative or the content length exceeds the length of the
5690 supplied buffer. (CVE-2002-0659)
5691 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5693 *) Assertions for various potential buffer overflows, not known to
5695 [Ben Laurie (CHATS)]
5697 *) Various temporary buffers to hold ASCII versions of integers were
5698 too small for 64 bit platforms. (CVE-2002-0655)
5699 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5701 *) Remote buffer overflow in SSL3 protocol - an attacker could
5702 supply an oversized session ID to a client. (CVE-2002-0656)
5703 [Ben Laurie (CHATS)]
5705 *) Remote buffer overflow in SSL2 protocol - an attacker could
5706 supply an oversized client master key. (CVE-2002-0656)
5707 [Ben Laurie (CHATS)]
5709 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5711 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5712 encoded as NULL) with id-dsa-with-sha1.
5713 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5715 *) Check various X509_...() return values in apps/req.c.
5716 [Nils Larsch <nla@trustcenter.de>]
5718 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5719 an end-of-file condition would erronously be flagged, when the CRLF
5720 was just at the end of a processed block. The bug was discovered when
5721 processing data through a buffering memory BIO handing the data to a
5722 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5723 <ptsekov@syntrex.com> and Nedelcho Stanev.
5726 *) Implement a countermeasure against a vulnerability recently found
5727 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5728 before application data chunks to avoid the use of known IVs
5729 with data potentially chosen by the attacker.
5732 *) Fix length checks in ssl3_get_client_hello().
5735 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5736 to prevent ssl3_read_internal() from incorrectly assuming that
5737 ssl3_read_bytes() found application data while handshake
5738 processing was enabled when in fact s->s3->in_read_app_data was
5739 merely automatically cleared during the initial handshake.
5740 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5742 *) Fix object definitions for Private and Enterprise: they were not
5743 recognized in their shortname (=lowercase) representation. Extend
5744 obj_dat.pl to issue an error when using undefined keywords instead
5745 of silently ignoring the problem (Svenning Sorensen
5746 <sss@sss.dnsalias.net>).
5749 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5750 generators, i.e. generators other than 2 and 5. (Previously, the
5751 code did not properly initialise the 'add' and 'rem' values to
5752 BN_generate_prime().)
5754 In the new general case, we do not insist that 'generator' is
5755 actually a primitive root: This requirement is rather pointless;
5756 a generator of the order-q subgroup is just as good, if not
5760 *) Map new X509 verification errors to alerts. Discovered and submitted by
5761 Tom Wu <tom@arcot.com>.
5764 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5765 returning non-zero before the data has been completely received
5766 when using non-blocking I/O.
5767 [Bodo Moeller; problem pointed out by John Hughes]
5769 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5770 [Ben Laurie, Lutz Jaenicke]
5772 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5773 Yoram Zahavi <YoramZ@gilian.com>).
5776 *) Add information about CygWin 1.3 and on, and preserve proper
5777 configuration for the versions before that.
5778 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5780 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5781 check whether we deal with a copy of a session and do not delete from
5782 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5783 <izhar@checkpoint.com>.
5786 *) Do not store session data into the internal session cache, if it
5787 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5788 flag is set). Proposed by Aslam <aslam@funk.com>.
5791 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5795 *) [In 0.9.6d-engine release:]
5796 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5797 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5799 *) Add the configuration target linux-s390x.
5800 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5802 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5803 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5804 variable as an indication that a ClientHello message has been
5805 received. As the flag value will be lost between multiple
5806 invocations of ssl3_accept when using non-blocking I/O, the
5807 function may not be aware that a handshake has actually taken
5808 place, thus preventing a new session from being added to the
5811 To avoid this problem, we now set s->new_session to 2 instead of
5812 using a local variable.
5813 [Lutz Jaenicke, Bodo Moeller]
5815 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5816 if the SSL_R_LENGTH_MISMATCH error is detected.
5817 [Geoff Thorpe, Bodo Moeller]
5819 *) New 'shared_ldflag' column in Configure platform table.
5822 *) Fix EVP_CIPHER_mode macro.
5823 ["Dan S. Camper" <dan@bti.net>]
5825 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5826 type, we must throw them away by setting rr->length to 0.
5827 [D P Chang <dpc@qualys.com>]
5829 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5831 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5832 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5833 worked incorrectly for those cases where range = 10..._2 and
5834 3*range is two bits longer than range.)
5837 *) Only add signing time to PKCS7 structures if it is not already
5841 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5842 OBJ_ld_ce should be OBJ_id_ce.
5843 Also some ip-pda OIDs in crypto/objects/objects.txt were
5844 incorrect (cf. RFC 3039).
5845 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5847 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5848 returns early because it has nothing to do.
5849 [Andy Schneider <andy.schneider@bjss.co.uk>]
5851 *) [In 0.9.6c-engine release:]
5852 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5853 [Andy Schneider <andy.schneider@bjss.co.uk>]
5855 *) [In 0.9.6c-engine release:]
5856 Add support for Cryptographic Appliance's keyserver technology.
5857 (Use engine 'keyclient')
5858 [Cryptographic Appliances and Geoff Thorpe]
5860 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5861 is called via tools/c89.sh because arguments have to be
5862 rearranged (all '-L' options must appear before the first object
5864 [Richard Shapiro <rshapiro@abinitio.com>]
5866 *) [In 0.9.6c-engine release:]
5867 Add support for Broadcom crypto accelerator cards, backported
5869 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5871 *) [In 0.9.6c-engine release:]
5872 Add support for SureWare crypto accelerator cards from
5873 Baltimore Technologies. (Use engine 'sureware')
5874 [Baltimore Technologies and Mark Cox]
5876 *) [In 0.9.6c-engine release:]
5877 Add support for crypto accelerator cards from Accelerated
5878 Encryption Processing, www.aep.ie. (Use engine 'aep')
5879 [AEP Inc. and Mark Cox]
5881 *) Add a configuration entry for gcc on UnixWare.
5882 [Gary Benson <gbenson@redhat.com>]
5884 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5885 messages are stored in a single piece (fixed-length part and
5886 variable-length part combined) and fix various bugs found on the way.
5889 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5890 instead. BIO_gethostbyname() does not know what timeouts are
5891 appropriate, so entries would stay in cache even when they have
5893 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5895 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5896 faced with a pathologically small ClientHello fragment that does
5897 not contain client_version: Instead of aborting with an error,
5898 simply choose the highest available protocol version (i.e.,
5899 TLS 1.0 unless it is disabled). In practice, ClientHello
5900 messages are never sent like this, but this change gives us
5901 strictly correct behaviour at least for TLS.
5904 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5905 never resets s->method to s->ctx->method when called from within
5906 one of the SSL handshake functions.
5907 [Bodo Moeller; problem pointed out by Niko Baric]
5909 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5910 (sent using the client's version number) if client_version is
5911 smaller than the protocol version in use. Also change
5912 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5913 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5914 the client will at least see that alert.
5917 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5921 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5922 client receives HelloRequest while in a handshake.
5923 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5925 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5926 should end in 'break', not 'goto end' which circuments various
5927 cleanups done in state SSL_ST_OK. But session related stuff
5928 must be disabled for SSL_ST_OK in the case that we just sent a
5931 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5932 before just sending a HelloRequest.
5933 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5935 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5936 reveal whether illegal block cipher padding was found or a MAC
5937 verification error occured. (Neither SSLerr() codes nor alerts
5938 are directly visible to potential attackers, but the information
5939 may leak via logfiles.)
5941 Similar changes are not required for the SSL 2.0 implementation
5942 because the number of padding bytes is sent in clear for SSL 2.0,
5943 and the extra bytes are just ignored. However ssl/s2_pkt.c
5944 failed to verify that the purported number of padding bytes is in
5948 *) Add OpenUNIX-8 support including shared libraries
5949 (Boyd Lynn Gerber <gerberb@zenez.com>).
5952 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5953 'wristwatch attack' using huge encoding parameters (cf.
5954 James H. Manger's CRYPTO 2001 paper). Note that the
5955 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5956 encoding parameters and hence was not vulnerable.
5959 *) BN_sqr() bug fix.
5960 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5962 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5963 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5964 followed by modular reduction.
5965 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5967 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5968 equivalent based on BN_pseudo_rand() instead of BN_rand().
5971 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5972 This function was broken, as the check for a new client hello message
5973 to handle SGC did not allow these large messages.
5974 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5977 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5980 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5981 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5984 *) Rework the configuration and shared library support for Tru64 Unix.
5985 The configuration part makes use of modern compiler features and
5986 still retains old compiler behavior for those that run older versions
5987 of the OS. The shared library support part includes a variant that
5988 uses the RPATH feature, and is available through the special
5989 configuration target "alpha-cc-rpath", which will never be selected
5991 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5993 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5994 with the same message size as in ssl3_get_certificate_request().
5995 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5996 messages might inadvertently be reject as too long.
5997 [Petr Lampa <lampa@fee.vutbr.cz>]
5999 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6002 *) Modified SSL library such that the verify_callback that has been set
6003 specificly for an SSL object with SSL_set_verify() is actually being
6004 used. Before the change, a verify_callback set with this function was
6005 ignored and the verify_callback() set in the SSL_CTX at the time of
6006 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6007 to allow the necessary settings.
6010 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6011 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6012 done automatically (in contradiction to the requirements of the C
6013 standard). This made problems when used from OpenSSH.
6016 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6017 dh->length and always used
6019 BN_rand_range(priv_key, dh->p).
6021 BN_rand_range() is not necessary for Diffie-Hellman, and this
6022 specific range makes Diffie-Hellman unnecessarily inefficient if
6023 dh->length (recommended exponent length) is much smaller than the
6024 length of dh->p. We could use BN_rand_range() if the order of
6025 the subgroup was stored in the DH structure, but we only have
6030 BN_rand(priv_key, l, ...)
6032 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6038 RSA_eay_public_encrypt
6039 RSA_eay_private_decrypt
6040 RSA_eay_private_encrypt (signing)
6041 RSA_eay_public_decrypt (signature verification)
6043 (default implementations for RSA_public_encrypt,
6044 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6045 always reject numbers >= n.
6048 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6049 to synchronize access to 'locking_thread'. This is necessary on
6050 systems where access to 'locking_thread' (an 'unsigned long'
6051 variable) is not atomic.
6054 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6055 *before* setting the 'crypto_lock_rand' flag. The previous code had
6056 a race condition if 0 is a valid thread ID.
6057 [Travis Vitek <vitek@roguewave.com>]
6059 *) Add support for shared libraries under Irix.
6060 [Albert Chin-A-Young <china@thewrittenword.com>]
6062 *) Add configuration option to build on Linux on both big-endian and
6064 [Ralf Baechle <ralf@uni-koblenz.de>]
6066 *) Add the possibility to create shared libraries on HP-UX.
6069 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6071 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6072 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6073 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6074 PRNG state recovery was possible based on the output of
6075 one PRNG request appropriately sized to gain knowledge on
6076 'md' followed by enough consecutive 1-byte PRNG requests
6077 to traverse all of 'state'.
6079 1. When updating 'md_local' (the current thread's copy of 'md')
6080 during PRNG output generation, hash all of the previous
6081 'md_local' value, not just the half used for PRNG output.
6083 2. Make the number of bytes from 'state' included into the hash
6084 independent from the number of PRNG bytes requested.
6086 The first measure alone would be sufficient to avoid
6087 Markku-Juhani's attack. (Actually it had never occurred
6088 to me that the half of 'md_local' used for chaining was the
6089 half from which PRNG output bytes were taken -- I had always
6090 assumed that the secret half would be used.) The second
6091 measure makes sure that additional data from 'state' is never
6092 mixed into 'md_local' in small portions; this heuristically
6093 further strengthens the PRNG.
6096 *) Fix crypto/bn/asm/mips3.s.
6099 *) When only the key is given to "enc", the IV is undefined. Print out
6100 an error message in this case.
6103 *) Handle special case when X509_NAME is empty in X509 printing routines.
6106 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6107 positive and less than q.
6110 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6111 used: it isn't thread safe and the add_lock_callback should handle
6113 [Paul Rose <Paul.Rose@bridge.com>]
6115 *) Verify that incoming data obeys the block size in
6116 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6120 [Ulf Möller, Bodo Möller]
6122 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6123 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6124 when fixing the server behaviour for backwards-compatible 'client
6125 hello' messages. (Note that the attack is impractical against
6126 SSL 3.0 and TLS 1.0 anyway because length and version checking
6127 means that the probability of guessing a valid ciphertext is
6128 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6131 Before 0.9.5, the countermeasure (hide the error by generating a
6132 random 'decryption result') did not work properly because
6133 ERR_clear_error() was missing, meaning that SSL_get_error() would
6134 detect the supposedly ignored error.
6136 Both problems are now fixed.
6139 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6140 (previously it was 1024).
6143 *) Fix for compatibility mode trust settings: ignore trust settings
6144 unless some valid trust or reject settings are present.
6147 *) Fix for blowfish EVP: its a variable length cipher.
6150 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6151 parameters in DSA public key structures and return an error in the
6152 DSA routines if parameters are absent.
6155 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6156 in the current directory if neither $RANDFILE nor $HOME was set.
6157 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6158 caused some confusion to Windows users who haven't defined $HOME.
6159 Thus RAND_file_name() is changed again: e_os.h can define a
6160 DEFAULT_HOME, which will be used if $HOME is not set.
6161 For Windows, we use "C:"; on other platforms, we still require
6162 environment variables.
6164 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6165 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6166 having multiple threads call RAND_poll() concurrently.
6169 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6170 combination of a flag and a thread ID variable.
6171 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6172 flag), *other* threads can enter ssleay_add_bytes without obeying
6173 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6174 that they do not hold after the first thread unsets add_do_not_lock).
6177 *) Change bctest again: '-x' expressions are not available in all
6181 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6183 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6184 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6186 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6187 the default extension for executables, if any. Also, make the perl
6188 scripts that use symlink() to test if it really exists and use "cp"
6189 if it doesn't. All this made OpenSSL compilable and installable in
6193 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6194 If SEQUENCE is length is indefinite just set c->slen to the total
6195 amount of data available.
6196 [Steve Henson, reported by shige@FreeBSD.org]
6197 [This change does not apply to 0.9.7.]
6199 *) Change bctest to avoid here-documents inside command substitution
6200 (workaround for FreeBSD /bin/sh bug).
6201 For compatibility with Ultrix, avoid shell functions (introduced
6202 in the bctest version that searches along $PATH).
6205 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6206 with des_encrypt() defined on some operating systems, like Solaris
6210 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6211 On the Importance of Eliminating Errors in Cryptographic
6212 Computations, J. Cryptology 14 (2001) 2, 101-119,
6213 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6216 *) MIPS assembler BIGNUM division bug fix.
6219 *) Disabled incorrect Alpha assembler code.
6222 *) Fix PKCS#7 decode routines so they correctly update the length
6223 after reading an EOC for the EXPLICIT tag.
6225 [This change does not apply to 0.9.7.]
6227 *) Fix bug in PKCS#12 key generation routines. This was triggered
6228 if a 3DES key was generated with a 0 initial byte. Include
6229 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6230 (but broken) behaviour.
6233 *) Enhance bctest to search for a working bc along $PATH and print
6235 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6237 *) Fix memory leaks in err.c: free err_data string if necessary;
6238 don't write to the wrong index in ERR_set_error_data.
6241 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6245 *) Replace rdtsc with _emit statements for VC++ version 5.
6246 [Jeremy Cooper <jeremy@baymoo.org>]
6248 *) Make it possible to reuse SSLv2 sessions.
6251 *) In copy_email() check for >= 0 as a return value for
6252 X509_NAME_get_index_by_NID() since 0 is a valid index.
6253 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6255 *) Avoid coredump with unsupported or invalid public keys by checking if
6256 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6257 PKCS7_verify() fails with non detached data.
6260 *) Don't use getenv in library functions when run as setuid/setgid.
6261 New function OPENSSL_issetugid().
6264 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6265 due to incorrect handling of multi-threading:
6267 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6269 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6271 3. Count how many times MemCheck_off() has been called so that
6272 nested use can be treated correctly. This also avoids
6273 inband-signalling in the previous code (which relied on the
6274 assumption that thread ID 0 is impossible).
6277 *) Add "-rand" option also to s_client and s_server.
6280 *) Fix CPU detection on Irix 6.x.
6281 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6282 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6284 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6287 [This change does not apply to 0.9.7.]
6289 *) Use the cached encoding of an X509_NAME structure rather than
6290 copying it. This is apparently the reason for the libsafe "errors"
6291 but the code is actually correct.
6294 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6295 Bleichenbacher's DSA attack.
6296 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6297 to be set and top=0 forces the highest bit to be set; top=-1 is new
6298 and leaves the highest bit random.
6299 [Ulf Moeller, Bodo Moeller]
6301 *) In the NCONF_...-based implementations for CONF_... queries
6302 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6303 a temporary CONF structure with the data component set to NULL
6304 (which gives segmentation faults in lh_retrieve).
6305 Instead, use NULL for the CONF pointer in CONF_get_string and
6306 CONF_get_number (which may use environment variables) and directly
6307 return NULL from CONF_get_section.
6310 *) Fix potential buffer overrun for EBCDIC.
6313 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6314 keyUsage if basicConstraints absent for a CA.
6317 *) Make SMIME_write_PKCS7() write mail header values with a format that
6318 is more generally accepted (no spaces before the semicolon), since
6319 some programs can't parse those values properly otherwise. Also make
6320 sure BIO's that break lines after each write do not create invalid
6324 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6325 macros previously used would not encode an empty SEQUENCE OF
6326 and break the signature.
6328 [This change does not apply to 0.9.7.]
6330 *) Zero the premaster secret after deriving the master secret in
6334 *) Add some EVP_add_digest_alias registrations (as found in
6335 OpenSSL_add_all_digests()) to SSL_library_init()
6336 aka OpenSSL_add_ssl_algorithms(). This provides improved
6337 compatibility with peers using X.509 certificates
6338 with unconventional AlgorithmIdentifier OIDs.
6341 *) Fix for Irix with NO_ASM.
6342 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6344 *) ./config script fixes.
6345 [Ulf Moeller, Richard Levitte]
6347 *) Fix 'openssl passwd -1'.
6350 *) Change PKCS12_key_gen_asc() so it can cope with non null
6351 terminated strings whose length is passed in the passlen
6352 parameter, for example from PEM callbacks. This was done
6353 by adding an extra length parameter to asc2uni().
6354 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6356 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6357 call failed, free the DSA structure.
6360 *) Fix to uni2asc() to cope with zero length Unicode strings.
6361 These are present in some PKCS#12 files.
6364 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6365 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6366 when writing a 32767 byte record.
6367 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6369 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6370 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6372 (RSA objects have a reference count access to which is protected
6373 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6374 so they are meant to be shared between threads.)
6375 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6376 "Reddie, Steven" <Steven.Reddie@ca.com>]
6378 *) Fix a deadlock in CRYPTO_mem_leaks().
6381 *) Use better test patterns in bntest.
6384 *) rand_win.c fix for Borland C.
6387 *) BN_rshift bugfix for n == 0.
6390 *) Add a 'bctest' script that checks for some known 'bc' bugs
6391 so that 'make test' does not abort just because 'bc' is broken.
6394 *) Store verify_result within SSL_SESSION also for client side to
6395 avoid potential security hole. (Re-used sessions on the client side
6396 always resulted in verify_result==X509_V_OK, not using the original
6397 result of the server certificate verification.)
6400 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6401 SSL3_RT_APPLICATION_DATA, return 0.
6402 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6406 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6407 releases, have been re-implemented by renaming the previous
6408 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6409 and ssl3_read_internal, respectively, and adding 'peek' parameters
6410 to them. The new ssl[23]_{read,peek} functions are calls to
6411 ssl[23]_read_internal with the 'peek' flag set appropriately.
6412 A 'peek' parameter has also been added to ssl3_read_bytes, which
6413 does the actual work for ssl3_read_internal.
6416 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6417 the method-specific "init()" handler. Also clean up ex_data after
6418 calling the method-specific "finish()" handler. Previously, this was
6419 happening the other way round.
6422 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6423 The previous value, 12, was not always sufficient for BN_mod_exp().
6426 *) Make sure that shared libraries get the internal name engine with
6427 the full version number and not just 0. This should mark the
6428 shared libraries as not backward compatible. Of course, this should
6429 be changed again when we can guarantee backward binary compatibility.
6432 *) Fix typo in get_cert_by_subject() in by_dir.c
6433 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6435 *) Rework the system to generate shared libraries:
6437 - Make note of the expected extension for the shared libraries and
6438 if there is a need for symbolic links from for example libcrypto.so.0
6439 to libcrypto.so.0.9.7. There is extended info in Configure for
6442 - Make as few rebuilds of the shared libraries as possible.
6444 - Still avoid linking the OpenSSL programs with the shared libraries.
6446 - When installing, install the shared libraries separately from the
6450 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6452 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6453 and not in SSL_clear because the latter is also used by the
6454 accept/connect functions; previously, the settings made by
6455 SSL_set_read_ahead would be lost during the handshake.
6456 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6458 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6459 Previously, it would create entries for disableed algorithms no
6463 *) Added several new manual pages for SSL_* function.
6466 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6468 *) In ssl23_get_client_hello, generate an error message when faced
6469 with an initial SSL 3.0/TLS record that is too small to contain the
6470 first two bytes of the ClientHello message, i.e. client_version.
6471 (Note that this is a pathologic case that probably has never happened
6472 in real life.) The previous approach was to use the version number
6473 from the record header as a substitute; but our protocol choice
6474 should not depend on that one because it is not authenticated
6475 by the Finished messages.
6478 *) More robust randomness gathering functions for Windows.
6479 [Jeffrey Altman <jaltman@columbia.edu>]
6481 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6482 not set then we don't setup the error code for issuer check errors
6483 to avoid possibly overwriting other errors which the callback does
6484 handle. If an application does set the flag then we assume it knows
6485 what it is doing and can handle the new informational codes
6489 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6490 a general "ANY" type, as such it should be able to decode anything
6491 including tagged types. However it didn't check the class so it would
6492 wrongly interpret tagged types in the same way as their universal
6493 counterpart and unknown types were just rejected. Changed so that the
6494 tagged and unknown types are handled in the same way as a SEQUENCE:
6495 that is the encoding is stored intact. There is also a new type
6496 "V_ASN1_OTHER" which is used when the class is not universal, in this
6497 case we have no idea what the actual type is so we just lump them all
6501 *) On VMS, stdout may very well lead to a file that is written to
6502 in a record-oriented fashion. That means that every write() will
6503 write a separate record, which will be read separately by the
6504 programs trying to read from it. This can be very confusing.
6506 The solution is to put a BIO filter in the way that will buffer
6507 text until a linefeed is reached, and then write everything a
6508 line at a time, so every record written will be an actual line,
6509 not chunks of lines and not (usually doesn't happen, but I've
6510 seen it once) several lines in one record. BIO_f_linebuffer() is
6513 Currently, it's a VMS-only method, because that's where it has
6514 been tested well enough.
6517 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6518 it can return incorrect results.
6519 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6520 but it was in 0.9.6-beta[12].)
6523 *) Disable the check for content being present when verifying detached
6524 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6525 include zero length content when signing messages.
6528 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6529 BIO_ctrl (for BIO pairs).
6532 *) Add DSO method for VMS.
6535 *) Bug fix: Montgomery multiplication could produce results with the
6539 *) Add RPM specification openssl.spec and modify it to build three
6540 packages. The default package contains applications, application
6541 documentation and run-time libraries. The devel package contains
6542 include files, static libraries and function documentation. The
6543 doc package contains the contents of the doc directory. The original
6544 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6547 *) Add a large number of documentation files for many SSL routines.
6548 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6550 *) Add a configuration entry for Sony News 4.
6551 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6553 *) Don't set the two most significant bits to one when generating a
6554 random number < q in the DSA library.
6557 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6558 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6559 the underlying transport is blocking) if a handshake took place.
6560 (The default behaviour is needed by applications such as s_client
6561 and s_server that use select() to determine when to use SSL_read;
6562 but for applications that know in advance when to expect data, it
6563 just makes things more complicated.)
6566 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6570 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6571 work better on such systems.
6572 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6574 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6575 Update PKCS12_parse() so it copies the friendlyName and the
6576 keyid to the certificates aux info.
6579 *) Fix bug in PKCS7_verify() which caused an infinite loop
6580 if there was more than one signature.
6581 [Sven Uszpelkat <su@celocom.de>]
6583 *) Major change in util/mkdef.pl to include extra information
6584 about each symbol, as well as presentig variables as well
6585 as functions. This change means that there's n more need
6586 to rebuild the .num files when some algorithms are excluded.
6589 *) Allow the verify time to be set by an application,
6590 rather than always using the current time.
6593 *) Phase 2 verify code reorganisation. The certificate
6594 verify code now looks up an issuer certificate by a
6595 number of criteria: subject name, authority key id
6596 and key usage. It also verifies self signed certificates
6597 by the same criteria. The main comparison function is
6598 X509_check_issued() which performs these checks.
6600 Lot of changes were necessary in order to support this
6601 without completely rewriting the lookup code.
6603 Authority and subject key identifier are now cached.
6605 The LHASH 'certs' is X509_STORE has now been replaced
6606 by a STACK_OF(X509_OBJECT). This is mainly because an
6607 LHASH can't store or retrieve multiple objects with
6608 the same hash value.
6610 As a result various functions (which were all internal
6611 use only) have changed to handle the new X509_STORE
6612 structure. This will break anything that messed round
6613 with X509_STORE internally.
6615 The functions X509_STORE_add_cert() now checks for an
6616 exact match, rather than just subject name.
6618 The X509_STORE API doesn't directly support the retrieval
6619 of multiple certificates matching a given criteria, however
6620 this can be worked round by performing a lookup first
6621 (which will fill the cache with candidate certificates)
6622 and then examining the cache for matches. This is probably
6623 the best we can do without throwing out X509_LOOKUP
6624 entirely (maybe later...).
6626 The X509_VERIFY_CTX structure has been enhanced considerably.
6628 All certificate lookup operations now go via a get_issuer()
6629 callback. Although this currently uses an X509_STORE it
6630 can be replaced by custom lookups. This is a simple way
6631 to bypass the X509_STORE hackery necessary to make this
6632 work and makes it possible to use more efficient techniques
6633 in future. A very simple version which uses a simple
6634 STACK for its trusted certificate store is also provided
6635 using X509_STORE_CTX_trusted_stack().
6637 The verify_cb() and verify() callbacks now have equivalents
6638 in the X509_STORE_CTX structure.
6640 X509_STORE_CTX also has a 'flags' field which can be used
6641 to customise the verify behaviour.
6644 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6645 excludes S/MIME capabilities.
6648 *) When a certificate request is read in keep a copy of the
6649 original encoding of the signed data and use it when outputing
6650 again. Signatures then use the original encoding rather than
6651 a decoded, encoded version which may cause problems if the
6652 request is improperly encoded.
6655 *) For consistency with other BIO_puts implementations, call
6656 buffer_write(b, ...) directly in buffer_puts instead of calling
6659 In BIO_puts, increment b->num_write as in BIO_write.
6660 [Peter.Sylvester@EdelWeb.fr]
6662 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6663 BN_zero, we may not return a BIGNUM with an array consisting of
6667 *) Avoid calling abort() from within the library when problems are
6668 detected, except if preprocessor symbols have been defined
6669 (such as REF_CHECK, BN_DEBUG etc.).
6672 *) New openssl application 'rsautl'. This utility can be
6673 used for low level RSA operations. DER public key
6674 BIO/fp routines also added.
6677 *) New Configure entry and patches for compiling on QNX 4.
6678 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6680 *) A demo state-machine implementation was sponsored by
6681 Nuron (http://www.nuron.com/) and is now available in
6682 demos/state_machine.
6685 *) New options added to the 'dgst' utility for signature
6686 generation and verification.
6689 *) Unrecognized PKCS#7 content types are now handled via a
6690 catch all ASN1_TYPE structure. This allows unsupported
6691 types to be stored as a "blob" and an application can
6692 encode and decode it manually.
6695 *) Fix various signed/unsigned issues to make a_strex.c
6697 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6699 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6700 length if passed a buffer. ASN1_INTEGER_to_BN failed
6701 if passed a NULL BN and its argument was negative.
6702 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6704 *) Modification to PKCS#7 encoding routines to output definite
6705 length encoding. Since currently the whole structures are in
6706 memory there's not real point in using indefinite length
6707 constructed encoding. However if OpenSSL is compiled with
6708 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6711 *) Added BIO_vprintf() and BIO_vsnprintf().
6714 *) Added more prefixes to parse for in the the strings written
6715 through a logging bio, to cover all the levels that are available
6716 through syslog. The prefixes are now:
6718 PANIC, EMERG, EMR => LOG_EMERG
6719 ALERT, ALR => LOG_ALERT
6720 CRIT, CRI => LOG_CRIT
6721 ERROR, ERR => LOG_ERR
6722 WARNING, WARN, WAR => LOG_WARNING
6723 NOTICE, NOTE, NOT => LOG_NOTICE
6724 INFO, INF => LOG_INFO
6725 DEBUG, DBG => LOG_DEBUG
6727 and as before, if none of those prefixes are present at the
6728 beginning of the string, LOG_ERR is chosen.
6730 On Win32, the LOG_* levels are mapped according to this:
6732 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6733 LOG_WARNING => EVENTLOG_WARNING_TYPE
6734 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6738 *) Made it possible to reconfigure with just the configuration
6739 argument "reconf" or "reconfigure". The command line arguments
6740 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6741 and are retrieved from there when reconfiguring.
6745 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6747 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6750 *) The obj_dat.pl script was messing up the sorting of object
6751 names. The reason was that it compared the quoted version
6752 of strings as a result "OCSP" > "OCSP Signing" because
6753 " > SPACE. Changed script to store unquoted versions of
6754 names and add quotes on output. It was also omitting some
6755 names from the lookup table if they were given a default
6756 value (that is if SN is missing it is given the same
6757 value as LN and vice versa), these are now added on the
6758 grounds that if an object has a name we should be able to
6759 look it up. Finally added warning output when duplicate
6760 short or long names are found.
6763 *) Changes needed for Tandem NSK.
6764 [Scott Uroff <scott@xypro.com>]
6766 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6767 RSA_padding_check_SSLv23(), special padding was never detected
6768 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6769 version rollback attacks was not effective.
6771 In s23_clnt.c, don't use special rollback-attack detection padding
6772 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6773 client; similarly, in s23_srvr.c, don't do the rollback check if
6774 SSL 2.0 is the only protocol enabled in the server.
6777 *) Make it possible to get hexdumps of unprintable data with 'openssl
6778 asn1parse'. By implication, the functions ASN1_parse_dump() and
6779 BIO_dump_indent() are added.
6782 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6783 these print out strings and name structures based on various
6784 flags including RFC2253 support and proper handling of
6785 multibyte characters. Added options to the 'x509' utility
6786 to allow the various flags to be set.
6789 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6790 Also change the functions X509_cmp_current_time() and
6791 X509_gmtime_adj() work with an ASN1_TIME structure,
6792 this will enable certificates using GeneralizedTime in validity
6793 dates to be checked.
6796 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6797 negative public key encodings) on by default,
6798 NO_NEG_PUBKEY_BUG can be set to disable it.
6801 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6802 content octets. An i2c_ASN1_OBJECT is unnecessary because
6803 the encoding can be trivially obtained from the structure.
6806 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6807 not read locks (CRYPTO_r_[un]lock).
6810 *) A first attempt at creating official support for shared
6811 libraries through configuration. I've kept it so the
6812 default is static libraries only, and the OpenSSL programs
6813 are always statically linked for now, but there are
6814 preparations for dynamic linking in place.
6815 This has been tested on Linux and Tru64.
6818 *) Randomness polling function for Win9x, as described in:
6819 Peter Gutmann, Software Generation of Practically Strong
6823 *) Fix so PRNG is seeded in req if using an already existing
6827 *) New options to smime application. -inform and -outform
6828 allow alternative formats for the S/MIME message including
6829 PEM and DER. The -content option allows the content to be
6830 specified separately. This should allow things like Netscape
6831 form signing output easier to verify.
6834 *) Fix the ASN1 encoding of tags using the 'long form'.
6837 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6838 STRING types. These convert content octets to and from the
6839 underlying type. The actual tag and length octets are
6840 already assumed to have been read in and checked. These
6841 are needed because all other string types have virtually
6842 identical handling apart from the tag. By having versions
6843 of the ASN1 functions that just operate on content octets
6844 IMPLICIT tagging can be handled properly. It also allows
6845 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6846 and ASN1_INTEGER are identical apart from the tag.
6849 *) Change the handling of OID objects as follows:
6851 - New object identifiers are inserted in objects.txt, following
6852 the syntax given in objects.README.
6853 - objects.pl is used to process obj_mac.num and create a new
6855 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6858 This is currently kind of a hack, and the perl code in objects.pl
6859 isn't very elegant, but it works as I intended. The simplest way
6860 to check that it worked correctly is to look in obj_dat.h and
6861 check the array nid_objs and make sure the objects haven't moved
6862 around (this is important!). Additions are OK, as well as
6863 consistent name changes.
6866 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6869 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6870 The given file adds to whatever has already been seeded into the
6871 random pool through the RANDFILE configuration file option or
6872 environment variable, or the default random state file.
6875 *) mkstack.pl now sorts each macro group into lexical order.
6876 Previously the output order depended on the order the files
6877 appeared in the directory, resulting in needless rewriting
6881 *) Patches to make OpenSSL compile under Win32 again. Mostly
6882 work arounds for the VC++ problem that it treats func() as
6883 func(void). Also stripped out the parts of mkdef.pl that
6884 added extra typesafe functions: these no longer exist.
6887 *) Reorganisation of the stack code. The macros are now all
6888 collected in safestack.h . Each macro is defined in terms of
6889 a "stack macro" of the form SKM_<name>(type, a, b). The
6890 DEBUG_SAFESTACK is now handled in terms of function casts,
6891 this has the advantage of retaining type safety without the
6892 use of additional functions. If DEBUG_SAFESTACK is not defined
6893 then the non typesafe macros are used instead. Also modified the
6894 mkstack.pl script to handle the new form. Needs testing to see
6895 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6896 the default if no major problems. Similar behaviour for ASN1_SET_OF
6897 and PKCS12_STACK_OF.
6900 *) When some versions of IIS use the 'NET' form of private key the
6901 key derivation algorithm is different. Normally MD5(password) is
6902 used as a 128 bit RC4 key. In the modified case
6903 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6904 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6905 as the old Netscape_RSA functions except they have an additional
6906 'sgckey' parameter which uses the modified algorithm. Also added
6907 an -sgckey command line option to the rsa utility. Thanks to
6908 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6909 algorithm to openssl-dev.
6912 *) The evp_local.h macros were using 'c.##kname' which resulted in
6913 invalid expansion on some systems (SCO 5.0.5 for example).
6914 Corrected to 'c.kname'.
6915 [Phillip Porch <root@theporch.com>]
6917 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6918 a STACK of email addresses from a certificate or request, these look
6919 in the subject name and the subject alternative name extensions and
6920 omit any duplicate addresses.
6923 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6924 This makes DSA verification about 2 % faster.
6927 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6928 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6929 plus overhead for 1024 bit moduli).
6930 This makes exponentiations about 0.5 % faster for 1024 bit
6931 exponents (as measured by "openssl speed rsa2048").
6934 *) Rename memory handling macros to avoid conflicts with other
6936 Malloc => OPENSSL_malloc
6937 Malloc_locked => OPENSSL_malloc_locked
6938 Realloc => OPENSSL_realloc
6939 Free => OPENSSL_free
6942 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6943 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6946 *) CygWin32 support.
6947 [John Jarvie <jjarvie@newsguy.com>]
6949 *) The type-safe stack code has been rejigged. It is now only compiled
6950 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6951 by default all type-specific stack functions are "#define"d back to
6952 standard stack functions. This results in more streamlined output
6953 but retains the type-safety checking possibilities of the original
6957 *) The STACK code has been cleaned up, and certain type declarations
6958 that didn't make a lot of sense have been brought in line. This has
6959 also involved a cleanup of sorts in safestack.h to more correctly
6960 map type-safe stack functions onto their plain stack counterparts.
6961 This work has also resulted in a variety of "const"ifications of
6962 lots of the code, especially "_cmp" operations which should normally
6963 be prototyped with "const" parameters anyway.
6966 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6967 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6968 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6969 where all of 'md' is used each time the PRNG is used, but 'state'
6970 is used only indexed by a cyclic counter. As entropy may not be
6971 well distributed from the beginning, 'md' is important as a
6972 chaining variable. However, the output function chains only half
6973 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6974 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6975 in all of 'state' being rewritten, with the new values depending
6976 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6979 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6980 the handshake is continued after ssl_verify_cert_chain();
6981 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6982 can lead to 'unexplainable' connection aborts later.
6983 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6985 *) Major EVP API cipher revision.
6986 Add hooks for extra EVP features. This allows various cipher
6987 parameters to be set in the EVP interface. Support added for variable
6988 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6989 setting of RC2 and RC5 parameters.
6991 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6994 Remove lots of duplicated code from the EVP library. For example *every*
6995 cipher init() function handles the 'iv' in the same way according to the
6996 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6997 for CFB and OFB modes they zero ctx->num.
6999 New functionality allows removal of S/MIME code RC2 hack.
7001 Most of the routines have the same form and so can be declared in terms
7004 By shifting this to the top level EVP_CipherInit() it can be removed from
7005 all individual ciphers. If the cipher wants to handle IVs or keys
7006 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7009 Change lots of functions like EVP_EncryptUpdate() to now return a
7010 value: although software versions of the algorithms cannot fail
7011 any installed hardware versions can.
7014 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7015 this option is set, tolerate broken clients that send the negotiated
7016 protocol version number instead of the requested protocol version
7020 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7021 i.e. non-zero for export ciphersuites, zero otherwise.
7022 Previous versions had this flag inverted, inconsistent with
7023 rsa_tmp_cb (..._TMP_RSA_CB).
7024 [Bodo Moeller; problem reported by Amit Chopra]
7026 *) Add missing DSA library text string. Work around for some IIS
7027 key files with invalid SEQUENCE encoding.
7030 *) Add a document (doc/standards.txt) that list all kinds of standards
7031 and so on that are implemented in OpenSSL.
7034 *) Enhance c_rehash script. Old version would mishandle certificates
7035 with the same subject name hash and wouldn't handle CRLs at all.
7036 Added -fingerprint option to crl utility, to support new c_rehash
7040 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7043 *) Fix for SSL server purpose checking. Server checking was
7044 rejecting certificates which had extended key usage present
7045 but no ssl client purpose.
7046 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7048 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7049 is a little unclear about how a blank password is handled.
7050 Since the password in encoded as a BMPString with terminating
7051 double NULL a zero length password would end up as just the
7052 double NULL. However no password at all is different and is
7053 handled differently in the PKCS#12 key generation code. NS
7054 treats a blank password as zero length. MSIE treats it as no
7055 password on export: but it will try both on import. We now do
7056 the same: PKCS12_parse() tries zero length and no password if
7057 the password is set to "" or NULL (NULL is now a valid password:
7058 it wasn't before) as does the pkcs12 application.
7061 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7062 perror when PEM_read_bio_X509_REQ fails, the error message must
7063 be obtained from the error queue.
7066 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7067 it in ERR_remove_state if appropriate, and change ERR_get_state
7068 accordingly to avoid race conditions (this is necessary because
7069 thread_hash is no longer constant once set).
7072 *) Bugfix for linux-elf makefile.one.
7075 *) RSA_get_default_method() will now cause a default
7076 RSA_METHOD to be chosen if one doesn't exist already.
7077 Previously this was only set during a call to RSA_new()
7078 or RSA_new_method(NULL) meaning it was possible for
7079 RSA_get_default_method() to return NULL.
7082 *) Added native name translation to the existing DSO code
7083 that will convert (if the flag to do so is set) filenames
7084 that are sufficiently small and have no path information
7085 into a canonical native form. Eg. "blah" converted to
7086 "libblah.so" or "blah.dll" etc.
7089 *) New function ERR_error_string_n(e, buf, len) which is like
7090 ERR_error_string(e, buf), but writes at most 'len' bytes
7091 including the 0 terminator. For ERR_error_string_n, 'buf'
7093 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7095 *) CONF library reworked to become more general. A new CONF
7096 configuration file reader "class" is implemented as well as a
7097 new functions (NCONF_*, for "New CONF") to handle it. The now
7098 old CONF_* functions are still there, but are reimplemented to
7099 work in terms of the new functions. Also, a set of functions
7100 to handle the internal storage of the configuration data is
7101 provided to make it easier to write new configuration file
7102 reader "classes" (I can definitely see something reading a
7103 configuration file in XML format, for example), called _CONF_*,
7104 or "the configuration storage API"...
7106 The new configuration file reading functions are:
7108 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7109 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7111 NCONF_default, NCONF_WIN32
7113 NCONF_dump_fp, NCONF_dump_bio
7115 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7116 NCONF_new creates a new CONF object. This works in the same way
7117 as other interfaces in OpenSSL, like the BIO interface.
7118 NCONF_dump_* dump the internal storage of the configuration file,
7119 which is useful for debugging. All other functions take the same
7120 arguments as the old CONF_* functions wth the exception of the
7121 first that must be a `CONF *' instead of a `LHASH *'.
7123 To make it easer to use the new classes with the old CONF_* functions,
7124 the function CONF_set_default_method is provided.
7127 *) Add '-tls1' option to 'openssl ciphers', which was already
7128 mentioned in the documentation but had not been implemented.
7129 (This option is not yet really useful because even the additional
7130 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7133 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7134 OpenSSL-based applications) load shared libraries and bind to
7135 them in a portable way.
7136 [Geoff Thorpe, with contributions from Richard Levitte]
7138 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7140 *) Make sure _lrotl and _lrotr are only used with MSVC.
7142 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7143 (the default implementation of RAND_status).
7145 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7146 to '-clrext' (= clear extensions), as intended and documented.
7147 [Bodo Moeller; inconsistency pointed out by Michael Attili
7150 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7151 was larger than the MD block size.
7152 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7154 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7155 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7156 using the passed key: if the passed key was a private key the result
7157 of X509_print(), for example, would be to print out all the private key
7161 *) des_quad_cksum() byte order bug fix.
7162 [Ulf Möller, using the problem description in krb4-0.9.7, where
7163 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7165 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7167 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7169 *) For easily testing in shell scripts whether some command
7170 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7171 returns with exit code 0 iff no command of the given name is available.
7172 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7173 the output goes to stdout and nothing is printed to stderr.
7174 Additional arguments are always ignored.
7176 Since for each cipher there is a command of the same name,
7177 the 'no-cipher' compilation switches can be tested this way.
7179 ('openssl no-XXX' is not able to detect pseudo-commands such
7180 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7183 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7186 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7187 is set; it will be thrown away anyway because each handshake creates
7189 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7190 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7191 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7192 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7195 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7196 'Q' and 'R' lose their special meanings (quit/renegotiate).
7197 This is part of what -quiet does; unlike -quiet, -ign_eof
7198 does not suppress any output.
7201 *) Add compatibility options to the purpose and trust code. The
7202 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7203 accepts a certificate or CA, this was the previous behaviour,
7204 with all the associated security issues.
7206 X509_TRUST_COMPAT is the old trust behaviour: only and
7207 automatically trust self signed roots in certificate store. A
7208 new trust setting X509_TRUST_DEFAULT is used to specify that
7209 a purpose has no associated trust setting and it should instead
7210 use the value in the default purpose.
7213 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7214 and fix a memory leak.
7217 *) In util/mkerr.pl (which implements 'make errors'), preserve
7218 reason strings from the previous version of the .c file, as
7219 the default to have only downcase letters (and digits) in
7220 automatically generated reasons codes is not always appropriate.
7223 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7224 using strerror. Previously, ERR_reason_error_string() returned
7225 library names as reason strings for SYSerr; but SYSerr is a special
7226 case where small numbers are errno values, not library numbers.
7229 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7230 converts DSA parameters into DH parameters. (When creating parameters,
7231 DSA_generate_parameters is used.)
7234 *) Include 'length' (recommended exponent length) in C code generated
7235 by 'openssl dhparam -C'.
7238 *) The second argument to set_label in perlasm was already being used
7239 so couldn't be used as a "file scope" flag. Moved to third argument
7243 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7244 instead of RAND_bytes for encryption IVs and salts.
7247 *) Include RAND_status() into RAND_METHOD instead of implementing
7248 it only for md_rand.c Otherwise replacing the PRNG by calling
7249 RAND_set_rand_method would be impossible.
7252 *) Don't let DSA_generate_key() enter an infinite loop if the random
7253 number generation fails.
7256 *) New 'rand' application for creating pseudo-random output.
7259 *) Added configuration support for Linux/IA64
7260 [Rolf Haberrecker <rolf@suse.de>]
7262 *) Assembler module support for Mingw32.
7265 *) Shared library support for HPUX (in shlib/).
7266 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7268 *) Shared library support for Solaris gcc.
7269 [Lutz Behnke <behnke@trustcenter.de>]
7271 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7273 *) PKCS7_encrypt() was adding text MIME headers twice because they
7274 were added manually and by SMIME_crlf_copy().
7277 *) In bntest.c don't call BN_rand with zero bits argument.
7278 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7280 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7281 case was implemented. This caused BN_div_recp() to fail occasionally.
7284 *) Add an optional second argument to the set_label() in the perl
7285 assembly language builder. If this argument exists and is set
7286 to 1 it signals that the assembler should use a symbol whose
7287 scope is the entire file, not just the current function. This
7288 is needed with MASM which uses the format label:: for this scope.
7289 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7291 *) Change the ASN1 types so they are typedefs by default. Before
7292 almost all types were #define'd to ASN1_STRING which was causing
7293 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7297 *) Change names of new functions to the new get1/get0 naming
7298 convention: After 'get1', the caller owns a reference count
7299 and has to call ..._free; 'get0' returns a pointer to some
7300 data structure without incrementing reference counters.
7301 (Some of the existing 'get' functions increment a reference
7302 counter, some don't.)
7303 Similarly, 'set1' and 'add1' functions increase reference
7304 counters or duplicate objects.
7307 *) Allow for the possibility of temp RSA key generation failure:
7308 the code used to assume it always worked and crashed on failure.
7311 *) Fix potential buffer overrun problem in BIO_printf().
7312 [Ulf Möller, using public domain code by Patrick Powell; problem
7313 pointed out by David Sacerdote <das33@cornell.edu>]
7315 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7316 RAND_egd() and RAND_status(). In the command line application,
7317 the EGD socket can be specified like a seed file using RANDFILE
7321 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7322 Some CAs (e.g. Verisign) distribute certificates in this form.
7325 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7326 list to exclude them. This means that no special compilation option
7327 is needed to use anonymous DH: it just needs to be included in the
7331 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7332 EVP_MD_type. The old functionality is available in a new macro called
7333 EVP_MD_md(). Change code that uses it and update docs.
7336 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7337 where the 'void *' argument is replaced by a function pointer argument.
7338 Previously 'void *' was abused to point to functions, which works on
7339 many platforms, but is not correct. As these functions are usually
7340 called by macros defined in OpenSSL header files, most source code
7341 should work without changes.
7344 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7345 sections with information on -D... compiler switches used for
7346 compiling the library so that applications can see them. To enable
7347 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7348 must be defined. E.g.,
7349 #define OPENSSL_ALGORITHM_DEFINES
7350 #include <openssl/opensslconf.h>
7351 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7352 [Richard Levitte, Ulf and Bodo Möller]
7354 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7358 *) Change the 'other' type in certificate aux info to a STACK_OF
7359 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7360 the required ASN1 format: arbitrary types determined by an OID.
7363 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7364 argument to 'req'. This is not because the function is newer or
7365 better than others it just uses the work 'NEW' in the certificate
7366 request header lines. Some software needs this.
7369 *) Reorganise password command line arguments: now passwords can be
7370 obtained from various sources. Delete the PEM_cb function and make
7371 it the default behaviour: i.e. if the callback is NULL and the
7372 usrdata argument is not NULL interpret it as a null terminated pass
7373 phrase. If usrdata and the callback are NULL then the pass phrase
7374 is prompted for as usual.
7377 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7378 the support is automatically enabled. The resulting binaries will
7379 autodetect the card and use it if present.
7380 [Ben Laurie and Compaq Inc.]
7382 *) Work around for Netscape hang bug. This sends certificate request
7383 and server done in one record. Since this is perfectly legal in the
7384 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7385 the bugs/SSLv3 entry for more info.
7388 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7391 *) Add -rand argument to smime and pkcs12 applications and read/write
7395 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7398 *) Add command line password options to the remaining applications.
7401 *) Bug fix for BN_div_recp() for numerators with an even number of
7405 *) More tests in bntest.c, and changed test_bn output.
7408 *) ./config recognizes MacOS X now.
7411 *) Bug fix for BN_div() when the first words of num and divsor are
7412 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7415 *) Add support for various broken PKCS#8 formats, and command line
7416 options to produce them.
7419 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7420 get temporary BIGNUMs from a BN_CTX.
7423 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7427 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7428 include a #define from the old name to the new. The original intent
7429 was that statically linked binaries could for example just call
7430 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7431 link with digests. This never worked becayse SSLeay_add_all_digests()
7432 and SSLeay_add_all_ciphers() were in the same source file so calling
7433 one would link with the other. They are now in separate source files.
7436 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7439 *) Use a less unusual form of the Miller-Rabin primality test (it used
7440 a binary algorithm for exponentiation integrated into the Miller-Rabin
7441 loop, our standard modexp algorithms are faster).
7444 *) Support for the EBCDIC character set completed.
7445 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7447 *) Source code cleanups: use const where appropriate, eliminate casts,
7448 use void * instead of char * in lhash.
7451 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7452 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7453 this the server could overwrite ephemeral keys that the client
7457 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7458 using 50 iterations of the Rabin-Miller test.
7460 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7461 iterations of the Rabin-Miller test as required by the appendix
7462 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7463 As BN_is_prime_fasttest includes trial division, DSA parameter
7464 generation becomes much faster.
7466 This implies a change for the callback functions in DSA_is_prime
7467 and DSA_generate_parameters: The callback function is called once
7468 for each positive witness in the Rabin-Miller test, not just
7469 occasionally in the inner loop; and the parameters to the
7470 callback function now provide an iteration count for the outer
7471 loop rather than for the current invocation of the inner loop.
7472 DSA_generate_parameters additionally can call the callback
7473 function with an 'iteration count' of -1, meaning that a
7474 candidate has passed the trial division test (when q is generated
7475 from an application-provided seed, trial division is skipped).
7478 *) New function BN_is_prime_fasttest that optionally does trial
7479 division before starting the Rabin-Miller test and has
7480 an additional BN_CTX * argument (whereas BN_is_prime always
7481 has to allocate at least one BN_CTX).
7482 'callback(1, -1, cb_arg)' is called when a number has passed the
7483 trial division stage.
7486 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7490 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7493 *) New function BN_pseudo_rand().
7496 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7497 bignum version of BN_from_montgomery() with the working code from
7498 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7502 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7503 made it impossible to use the same SSL_SESSION data structure in
7504 SSL2 clients in multiple threads.
7507 *) The return value of RAND_load_file() no longer counts bytes obtained
7508 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7509 to seed the PRNG (previously an explicit byte count was required).
7510 [Ulf Möller, Bodo Möller]
7512 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7513 used (char *) instead of (void *) and had casts all over the place.
7516 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7519 *) Retain source code compatibility for BN_prime_checks macro:
7520 BN_is_prime(..., BN_prime_checks, ...) now uses
7521 BN_prime_checks_for_size to determine the appropriate number of
7522 Rabin-Miller iterations.
7525 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7526 DH_CHECK_P_NOT_SAFE_PRIME.
7527 (Check if this is true? OpenPGP calls them "strong".)
7530 *) Merge the functionality of "dh" and "gendh" programs into a new program
7531 "dhparam". The old programs are retained for now but will handle DH keys
7532 (instead of parameters) in future.
7535 *) Make the ciphers, s_server and s_client programs check the return values
7536 when a new cipher list is set.
7539 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7540 ciphers. Before when the 56bit ciphers were enabled the sorting was
7543 The syntax for the cipher sorting has been extended to support sorting by
7544 cipher-strength (using the strength_bits hard coded in the tables).
7545 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7547 Fix a bug in the cipher-command parser: when supplying a cipher command
7548 string with an "undefined" symbol (neither command nor alphanumeric
7549 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7550 an error is flagged.
7552 Due to the strength-sorting extension, the code of the
7553 ssl_create_cipher_list() function was completely rearranged. I hope that
7554 the readability was also increased :-)
7555 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7557 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7558 for the first serial number and places 2 in the serial number file. This
7559 avoids problems when the root CA is created with serial number zero and
7560 the first user certificate has the same issuer name and serial number
7564 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7565 the new code. Add documentation for this stuff.
7568 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7569 X509_*() to X509at_*() on the grounds that they don't handle X509
7570 structures and behave in an analagous way to the X509v3 functions:
7571 they shouldn't be called directly but wrapper functions should be used
7574 So we also now have some wrapper functions that call the X509at functions
7575 when passed certificate requests. (TO DO: similar things can be done with
7576 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7577 things. Some of these need some d2i or i2d and print functionality
7578 because they handle more complex structures.)
7581 *) Add missing #ifndefs that caused missing symbols when building libssl
7582 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7583 NO_RSA in ssl/s2*.c.
7584 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7586 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7587 has a return value which indicates the quality of the random data
7588 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7589 error queue. New function RAND_pseudo_bytes() generates output that is
7590 guaranteed to be unique but not unpredictable. RAND_add is like
7591 RAND_seed, but takes an extra argument for an entropy estimate
7592 (RAND_seed always assumes full entropy).
7595 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7596 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7597 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7598 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7599 false-positive rate of at most 2^-80 for random input.
7602 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7605 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7606 in the 0.9.5 release), this returns the chain
7607 from an X509_CTX structure with a dup of the stack and all
7608 the X509 reference counts upped: so the stack will exist
7609 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7612 Also make SSL_SESSION_print() print out the verify return
7616 *) Add manpage for the pkcs12 command. Also change the default
7617 behaviour so MAC iteration counts are used unless the new
7618 -nomaciter option is used. This improves file security and
7619 only older versions of MSIE (4.0 for example) need it.
7622 *) Honor the no-xxx Configure options when creating .DEF files.
7625 *) Add PKCS#10 attributes to field table: challengePassword,
7626 unstructuredName and unstructuredAddress. These are taken from
7627 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7628 international characters are used.
7630 More changes to X509_ATTRIBUTE code: allow the setting of types
7631 based on strings. Remove the 'loc' parameter when adding
7632 attributes because these will be a SET OF encoding which is sorted
7636 *) Initial changes to the 'req' utility to allow request generation
7637 automation. This will allow an application to just generate a template
7638 file containing all the field values and have req construct the
7641 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7642 used all over the place including certificate requests and PKCS#7
7643 structures. They are currently handled manually where necessary with
7644 some primitive wrappers for PKCS#7. The new functions behave in a
7645 manner analogous to the X509 extension functions: they allow
7646 attributes to be looked up by NID and added.
7648 Later something similar to the X509V3 code would be desirable to
7649 automatically handle the encoding, decoding and printing of the
7650 more complex types. The string types like challengePassword can
7651 be handled by the string table functions.
7653 Also modified the multi byte string table handling. Now there is
7654 a 'global mask' which masks out certain types. The table itself
7655 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7656 is useful when for example there is only one permissible type
7657 (as in countryName) and using the mask might result in no valid
7661 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7662 SSL_get_peer_finished to allow applications to obtain the latest
7663 Finished messages sent to the peer or expected from the peer,
7664 respectively. (SSL_get_peer_finished is usually the Finished message
7665 actually received from the peer, otherwise the protocol will be aborted.)
7667 As the Finished message are message digests of the complete handshake
7668 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7669 be used for external authentication procedures when the authentication
7670 provided by SSL/TLS is not desired or is not enough.
7673 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7674 the host supports BWX extension and if Compaq C is present on the
7675 $PATH. Just exploiting of the BWX extension results in 20-30%
7676 performance kick for some algorithms, e.g. DES and RC4 to mention
7677 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7681 *) Add support for MS "fast SGC". This is arguably a violation of the
7682 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7683 weak crypto and after checking the certificate is SGC a second one
7684 with strong crypto. MS SGC stops the first handshake after receiving
7685 the server certificate message and sends a second client hello. Since
7686 a server will typically do all the time consuming operations before
7687 expecting any further messages from the client (server key exchange
7688 is the most expensive) there is little difference between the two.
7690 To get OpenSSL to support MS SGC we have to permit a second client
7691 hello message after we have sent server done. In addition we have to
7692 reset the MAC if we do get this second client hello.
7695 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7696 if a DER encoded private key is RSA or DSA traditional format. Changed
7697 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7698 format DER encoded private key. Newer code should use PKCS#8 format which
7699 has the key type encoded in the ASN1 structure. Added DER private key
7700 support to pkcs8 application.
7703 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7704 ciphersuites has been selected (as required by the SSL 3/TLS 1
7705 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7706 is set, we interpret this as a request to violate the specification
7707 (the worst that can happen is a handshake failure, and 'correct'
7708 behaviour would result in a handshake failure anyway).
7711 *) In SSL_CTX_add_session, take into account that there might be multiple
7712 SSL_SESSION structures with the same session ID (e.g. when two threads
7713 concurrently obtain them from an external cache).
7714 The internal cache can handle only one SSL_SESSION with a given ID,
7715 so if there's a conflict, we now throw out the old one to achieve
7719 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7720 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7721 some routines that use cipher OIDs: some ciphers do not have OIDs
7722 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7726 *) Simplify the trust setting structure and code. Now we just have
7727 two sequences of OIDs for trusted and rejected settings. These will
7728 typically have values the same as the extended key usage extension
7729 and any application specific purposes.
7731 The trust checking code now has a default behaviour: it will just
7732 check for an object with the same NID as the passed id. Functions can
7733 be provided to override either the default behaviour or the behaviour
7734 for a given id. SSL client, server and email already have functions
7735 in place for compatibility: they check the NID and also return "trusted"
7736 if the certificate is self signed.
7739 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7740 traditional format into an EVP_PKEY structure.
7743 *) Add a password callback function PEM_cb() which either prompts for
7744 a password if usr_data is NULL or otherwise assumes it is a null
7745 terminated password. Allow passwords to be passed on command line
7746 environment or config files in a few more utilities.
7749 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7750 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7751 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7752 Update documentation.
7755 *) Support for ASN1 "NULL" type. This could be handled before by using
7756 ASN1_TYPE but there wasn't any function that would try to read a NULL
7757 and produce an error if it couldn't. For compatibility we also have
7758 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7759 don't allocate anything because they don't need to.
7762 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7764 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7766 *) Rebuild of the memory allocation routines used by OpenSSL code and
7767 possibly others as well. The purpose is to make an interface that
7768 provide hooks so anyone can build a separate set of allocation and
7769 deallocation routines to be used by OpenSSL, for example memory
7770 pool implementations, or something else, which was previously hard
7771 since Malloc(), Realloc() and Free() were defined as macros having
7772 the values malloc, realloc and free, respectively (except for Win32
7773 compilations). The same is provided for memory debugging code.
7774 OpenSSL already comes with functionality to find memory leaks, but
7775 this gives people a chance to debug other memory problems.
7777 With these changes, a new set of functions and macros have appeared:
7779 CRYPTO_set_mem_debug_functions() [F]
7780 CRYPTO_get_mem_debug_functions() [F]
7781 CRYPTO_dbg_set_options() [F]
7782 CRYPTO_dbg_get_options() [F]
7783 CRYPTO_malloc_debug_init() [M]
7785 The memory debug functions are NULL by default, unless the library
7786 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7787 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7788 gives the standard debugging functions that come with OpenSSL) or
7789 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7790 provided by the library user) must be used. When the standard
7791 debugging functions are used, CRYPTO_dbg_set_options can be used to
7792 request additional information:
7793 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7794 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7796 Also, things like CRYPTO_set_mem_functions will always give the
7797 expected result (the new set of functions is used for allocation
7798 and deallocation) at all times, regardless of platform and compiler
7801 To finish it up, some functions that were never use in any other
7802 way than through macros have a new API and new semantic:
7805 CRYPTO_dbg_realloc()
7808 All macros of value have retained their old syntax.
7809 [Richard Levitte and Bodo Moeller]
7811 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7812 ordering of SMIMECapabilities wasn't in "strength order" and there
7813 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7817 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7818 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7819 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7821 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7822 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7823 functionality to handle multipart/signed properly) and a utility
7824 called 'smime' to call all this stuff. This is based on code I
7825 originally wrote for Celo who have kindly allowed it to be
7826 included in OpenSSL.
7829 *) Add variants des_set_key_checked and des_set_key_unchecked of
7830 des_set_key (aka des_key_sched). Global variable des_check_key
7831 decides which of these is called by des_set_key; this way
7832 des_check_key behaves as it always did, but applications and
7833 the library itself, which was buggy for des_check_key == 1,
7834 have a cleaner way to pick the version they need.
7837 *) New function PKCS12_newpass() which changes the password of a
7841 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7842 dynamic mix. In both cases the ids can be used as an index into the
7843 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7844 functions so they accept a list of the field values and the
7845 application doesn't need to directly manipulate the X509_TRUST
7849 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7853 *) Modify the way the V3 extension code looks up extensions. This now
7854 works in a similar way to the object code: we have some "standard"
7855 extensions in a static table which is searched with OBJ_bsearch()
7856 and the application can add dynamic ones if needed. The file
7857 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7858 updated whenever a new extension is added to the core code and kept
7859 in ext_nid order. There is a simple program 'tabtest.c' which checks
7860 this. New extensions are not added too often so this file can readily
7861 be maintained manually.
7863 There are two big advantages in doing things this way. The extensions
7864 can be looked up immediately and no longer need to be "added" using
7865 X509V3_add_standard_extensions(): this function now does nothing.
7866 [Side note: I get *lots* of email saying the extension code doesn't
7867 work because people forget to call this function]
7868 Also no dynamic allocation is done unless new extensions are added:
7869 so if we don't add custom extensions there is no need to call
7870 X509V3_EXT_cleanup().
7873 *) Modify enc utility's salting as follows: make salting the default. Add a
7874 magic header, so unsalted files fail gracefully instead of just decrypting
7875 to garbage. This is because not salting is a big security hole, so people
7876 should be discouraged from doing it.
7879 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7880 digest to be passed on the command line but it only used this
7881 parameter when signing a certificate. Modified so all relevant
7882 operations are affected by the digest parameter including the
7883 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7884 DSA key was used because it didn't fix the digest.
7887 *) Initial certificate chain verify code. Currently tests the untrusted
7888 certificates for consistency with the verify purpose (which is set
7889 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7891 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7892 this is because it will reject chains with invalid extensions whereas
7893 every previous version of OpenSSL and SSLeay made no checks at all.
7895 Trust code: checks the root CA for the relevant trust settings. Trust
7896 settings have an initial value consistent with the verify purpose: e.g.
7897 if the verify purpose is for SSL client use it expects the CA to be
7898 trusted for SSL client use. However the default value can be changed to
7899 permit custom trust settings: one example of this would be to only trust
7900 certificates from a specific "secure" set of CAs.
7902 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7903 which should be used for version portability: especially since the
7904 verify structure is likely to change more often now.
7906 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7907 to set them. If not set then assume SSL clients will verify SSL servers
7910 Two new options to the verify program: -untrusted allows a set of
7911 untrusted certificates to be passed in and -purpose which sets the
7912 intended purpose of the certificate. If a purpose is set then the
7913 new chain verify code is used to check extension consistency.
7916 *) Support for the authority information access extension.
7919 *) Modify RSA and DSA PEM read routines to transparently handle
7920 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7921 public keys in a format compatible with certificate
7922 SubjectPublicKeyInfo structures. Unfortunately there were already
7923 functions called *_PublicKey_* which used various odd formats so
7924 these are retained for compatibility: however the DSA variants were
7925 never in a public release so they have been deleted. Changed dsa/rsa
7926 utilities to handle the new format: note no releases ever handled public
7927 keys so we should be OK.
7929 The primary motivation for this change is to avoid the same fiasco
7930 that dogs private keys: there are several incompatible private key
7931 formats some of which are standard and some OpenSSL specific and
7932 require various evil hacks to allow partial transparent handling and
7933 even then it doesn't work with DER formats. Given the option anything
7934 other than PKCS#8 should be dumped: but the other formats have to
7935 stay in the name of compatibility.
7937 With public keys and the benefit of hindsight one standard format
7938 is used which works with EVP_PKEY, RSA or DSA structures: though
7939 it clearly returns an error if you try to read the wrong kind of key.
7941 Added a -pubkey option to the 'x509' utility to output the public key.
7942 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7943 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7944 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7945 that do the same as the EVP_PKEY_assign_*() except they up the
7946 reference count of the added key (they don't "swallow" the
7950 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7951 CRLs would fail if the file contained no certificates or no CRLs:
7952 added a new function to read in both types and return the number
7953 read: this means that if none are read it will be an error. The
7954 DER versions of the certificate and CRL reader would always fail
7955 because it isn't possible to mix certificates and CRLs in DER format
7956 without choking one or the other routine. Changed this to just read
7957 a certificate: this is the best we can do. Also modified the code
7958 in apps/verify.c to take notice of return codes: it was previously
7959 attempting to read in certificates from NULL pointers and ignoring
7960 any errors: this is one reason why the cert and CRL reader seemed
7961 to work. It doesn't check return codes from the default certificate
7962 routines: these may well fail if the certificates aren't installed.
7965 *) Code to support otherName option in GeneralName.
7968 *) First update to verify code. Change the verify utility
7969 so it warns if it is passed a self signed certificate:
7970 for consistency with the normal behaviour. X509_verify
7971 has been modified to it will now verify a self signed
7972 certificate if *exactly* the same certificate appears
7973 in the store: it was previously impossible to trust a
7974 single self signed certificate. This means that:
7975 openssl verify ss.pem
7976 now gives a warning about a self signed certificate but
7977 openssl verify -CAfile ss.pem ss.pem
7981 *) For servers, store verify_result in SSL_SESSION data structure
7982 (and add it to external session representation).
7983 This is needed when client certificate verifications fails,
7984 but an application-provided verification callback (set by
7985 SSL_CTX_set_cert_verify_callback) allows accepting the session
7986 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7987 but returns 1): When the session is reused, we have to set
7988 ssl->verify_result to the appropriate error code to avoid
7990 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7992 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7993 case in PKCS7_dataInit() where the signed PKCS7 structure
7994 didn't contain any existing data because it was being created.
7995 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7997 *) Add a salt to the key derivation routines in enc.c. This
7998 forms the first 8 bytes of the encrypted file. Also add a
7999 -S option to allow a salt to be input on the command line.
8002 *) New function X509_cmp(). Oddly enough there wasn't a function
8003 to compare two certificates. We do this by working out the SHA1
8004 hash and comparing that. X509_cmp() will be needed by the trust
8008 *) SSL_get1_session() is like SSL_get_session(), but increments
8009 the reference count in the SSL_SESSION returned.
8010 [Geoff Thorpe <geoff@eu.c2.net>]
8012 *) Fix for 'req': it was adding a null to request attributes.
8013 Also change the X509_LOOKUP and X509_INFO code to handle
8014 certificate auxiliary information.
8017 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8021 *) Add the possibility to add extra information to the memory leak
8022 detecting output, to form tracebacks, showing from where each
8023 allocation was originated: CRYPTO_push_info("constant string") adds
8024 the string plus current file name and line number to a per-thread
8025 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8026 is like calling CYRPTO_pop_info() until the stack is empty.
8027 Also updated memory leak detection code to be multi-thread-safe.
8030 *) Add options -text and -noout to pkcs7 utility and delete the
8031 encryption options which never did anything. Update docs.
8034 *) Add options to some of the utilities to allow the pass phrase
8035 to be included on either the command line (not recommended on
8036 OSes like Unix) or read from the environment. Update the
8037 manpages and fix a few bugs.
8040 *) Add a few manpages for some of the openssl commands.
8043 *) Fix the -revoke option in ca. It was freeing up memory twice,
8044 leaking and not finding already revoked certificates.
8047 *) Extensive changes to support certificate auxiliary information.
8048 This involves the use of X509_CERT_AUX structure and X509_AUX
8049 functions. An X509_AUX function such as PEM_read_X509_AUX()
8050 can still read in a certificate file in the usual way but it
8051 will also read in any additional "auxiliary information". By
8052 doing things this way a fair degree of compatibility can be
8053 retained: existing certificates can have this information added
8054 using the new 'x509' options.
8056 Current auxiliary information includes an "alias" and some trust
8057 settings. The trust settings will ultimately be used in enhanced
8058 certificate chain verification routines: currently a certificate
8059 can only be trusted if it is self signed and then it is trusted
8063 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8064 The problem was that one of the replacement routines had not been working
8065 since SSLeay releases. For now the offending routine has been replaced
8066 with non-optimised assembler. Even so, this now gives around 95%
8067 performance improvement for 1024 bit RSA signs.
8070 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8071 handling. Most clients have the effective key size in bits equal to
8072 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8073 A few however don't do this and instead use the size of the decrypted key
8074 to determine the RC2 key length and the AlgorithmIdentifier to determine
8075 the effective key length. In this case the effective key length can still
8076 be 40 bits but the key length can be 168 bits for example. This is fixed
8077 by manually forcing an RC2 key into the EVP_PKEY structure because the
8078 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8079 the key length and effective key length are equal.
8082 *) Add a bunch of functions that should simplify the creation of
8083 X509_NAME structures. Now you should be able to do:
8084 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8085 and have it automatically work out the correct field type and fill in
8086 the structures. The more adventurous can try:
8087 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8088 and it will (hopefully) work out the correct multibyte encoding.
8091 *) Change the 'req' utility to use the new field handling and multibyte
8092 copy routines. Before the DN field creation was handled in an ad hoc
8093 way in req, ca, and x509 which was rather broken and didn't support
8094 BMPStrings or UTF8Strings. Since some software doesn't implement
8095 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8096 using the dirstring_type option. See the new comment in the default
8097 openssl.cnf for more info.
8100 *) Make crypto/rand/md_rand.c more robust:
8101 - Assure unique random numbers after fork().
8102 - Make sure that concurrent threads access the global counter and
8103 md serializably so that we never lose entropy in them
8104 or use exactly the same state in multiple threads.
8105 Access to the large state is not always serializable because
8106 the additional locking could be a performance killer, and
8107 md should be large enough anyway.
8110 *) New file apps/app_rand.c with commonly needed functionality
8111 for handling the random seed file.
8113 Use the random seed file in some applications that previously did not:
8115 dsaparam -genkey (which also ignored its '-rand' option),
8118 x509 (when signing).
8119 Except on systems with /dev/urandom, it is crucial to have a random
8120 seed file at least for key creation, DSA signing, and for DH exchanges;
8121 for RSA signatures we could do without one.
8123 gendh and gendsa (unlike genrsa) used to read only the first byte
8124 of each file listed in the '-rand' option. The function as previously
8125 found in genrsa is now in app_rand.c and is used by all programs
8126 that support '-rand'.
8129 *) In RAND_write_file, use mode 0600 for creating files;
8130 don't just chmod when it may be too late.
8133 *) Report an error from X509_STORE_load_locations
8134 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8137 *) New function ASN1_mbstring_copy() this copies a string in either
8138 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8139 into an ASN1_STRING type. A mask of permissible types is passed
8140 and it chooses the "minimal" type to use or an error if not type
8144 *) Add function equivalents to the various macros in asn1.h. The old
8145 macros are retained with an M_ prefix. Code inside the library can
8146 use the M_ macros. External code (including the openssl utility)
8147 should *NOT* in order to be "shared library friendly".
8150 *) Add various functions that can check a certificate's extensions
8151 to see if it usable for various purposes such as SSL client,
8152 server or S/MIME and CAs of these types. This is currently
8153 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8154 verification. Also added a -purpose flag to x509 utility to
8155 print out all the purposes.
8158 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8162 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8163 for, obtain and decode and extension and obtain its critical flag.
8164 This allows all the necessary extension code to be handled in a
8165 single function call.
8168 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8169 platforms. See crypto/rc4/rc4_enc.c for further details.
8172 *) New -noout option to asn1parse. This causes no output to be produced
8173 its main use is when combined with -strparse and -out to extract data
8174 from a file (which may not be in ASN.1 format).
8177 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8178 when producing the local key id.
8179 [Richard Levitte <levitte@stacken.kth.se>]
8181 *) New option -dhparam in s_server. This allows a DH parameter file to be
8182 stated explicitly. If it is not stated then it tries the first server
8183 certificate file. The previous behaviour hard coded the filename
8187 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8188 a public key to be input or output. For example:
8189 openssl rsa -in key.pem -pubout -out pubkey.pem
8190 Also added necessary DSA public key functions to handle this.
8193 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8194 in the message. This was handled by allowing
8195 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8196 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8198 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8199 to the end of the strings whereas this didn't. This would cause problems
8200 if strings read with d2i_ASN1_bytes() were later modified.
8201 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8203 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8204 data and it contains EOF it will end up returning an error. This is
8205 caused by input 46 bytes long. The cause is due to the way base64
8206 BIOs find the start of base64 encoded data. They do this by trying a
8207 trial decode on each line until they find one that works. When they
8208 do a flag is set and it starts again knowing it can pass all the
8209 data directly through the decoder. Unfortunately it doesn't reset
8210 the context it uses. This means that if EOF is reached an attempt
8211 is made to pass two EOFs through the context and this causes the
8212 resulting error. This can also cause other problems as well. As is
8213 usual with these problems it takes *ages* to find and the fix is
8214 trivial: move one line.
8215 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8217 *) Ugly workaround to get s_client and s_server working under Windows. The
8218 old code wouldn't work because it needed to select() on sockets and the
8219 tty (for keypresses and to see if data could be written). Win32 only
8220 supports select() on sockets so we select() with a 1s timeout on the
8221 sockets and then see if any characters are waiting to be read, if none
8222 are present then we retry, we also assume we can always write data to
8223 the tty. This isn't nice because the code then blocks until we've
8224 received a complete line of data and it is effectively polling the
8225 keyboard at 1s intervals: however it's quite a bit better than not
8226 working at all :-) A dedicated Windows application might handle this
8227 with an event loop for example.
8230 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8231 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8232 will be called when RSA_sign() and RSA_verify() are used. This is useful
8233 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8234 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8235 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8236 This necessitated the support of an extra signature type NID_md5_sha1
8237 for SSL signatures and modifications to the SSL library to use it instead
8238 of calling RSA_public_decrypt() and RSA_private_encrypt().
8241 *) Add new -verify -CAfile and -CApath options to the crl program, these
8242 will lookup a CRL issuers certificate and verify the signature in a
8243 similar way to the verify program. Tidy up the crl program so it
8244 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8245 less strict. It will now permit CRL extensions even if it is not
8246 a V2 CRL: this will allow it to tolerate some broken CRLs.
8249 *) Initialize all non-automatic variables each time one of the openssl
8250 sub-programs is started (this is necessary as they may be started
8251 multiple times from the "OpenSSL>" prompt).
8252 [Lennart Bang, Bodo Moeller]
8254 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8255 removing all other RSA functionality (this is what NO_RSA does). This
8256 is so (for example) those in the US can disable those operations covered
8257 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8261 *) Non-copying interface to BIO pairs.
8262 (still largely untested)
8265 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8266 ASCII string. This was handled independently in various places before.
8269 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8270 UTF8 strings a character at a time.
8273 *) Use client_version from client hello to select the protocol
8274 (s23_srvr.c) and for RSA client key exchange verification
8275 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8278 *) Add various utility functions to handle SPKACs, these were previously
8279 handled by poking round in the structure internals. Added new function
8280 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8281 print, verify and generate SPKACs. Based on an original idea from
8282 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8285 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8288 *) Allow the config file extension section to be overwritten on the
8289 command line. Based on an original idea from Massimiliano Pala
8290 <madwolf@comune.modena.it>. The new option is called -extensions
8291 and can be applied to ca, req and x509. Also -reqexts to override
8292 the request extensions in req and -crlexts to override the crl extensions
8296 *) Add new feature to the SPKAC handling in ca. Now you can include
8297 the same field multiple times by preceding it by "XXXX." for example:
8300 this is the same syntax as used in the req config file.
8303 *) Allow certificate extensions to be added to certificate requests. These
8304 are specified in a 'req_extensions' option of the req section of the
8305 config file. They can be printed out with the -text option to req but
8306 are otherwise ignored at present.
8309 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8310 data read consists of only the final block it would not decrypted because
8311 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8312 A misplaced 'break' also meant the decrypted final block might not be
8313 copied until the next read.
8316 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8317 a few extra parameters to the DH structure: these will be useful if
8318 for example we want the value of 'q' or implement X9.42 DH.
8321 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8322 provides hooks that allow the default DSA functions or functions on a
8323 "per key" basis to be replaced. This allows hardware acceleration and
8324 hardware key storage to be handled without major modification to the
8325 library. Also added low level modexp hooks and CRYPTO_EX structure and
8326 associated functions.
8329 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8330 as "read only": it can't be written to and the buffer it points to will
8331 not be freed. Reading from a read only BIO is much more efficient than
8332 a normal memory BIO. This was added because there are several times when
8333 an area of memory needs to be read from a BIO. The previous method was
8334 to create a memory BIO and write the data to it, this results in two
8335 copies of the data and an O(n^2) reading algorithm. There is a new
8336 function BIO_new_mem_buf() which creates a read only memory BIO from
8337 an area of memory. Also modified the PKCS#7 routines to use read only
8341 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8342 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8343 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8344 but a retry condition occured while trying to read the rest.
8347 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8348 NID_pkcs7_encrypted by default: this was wrong since this should almost
8349 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8350 the encrypted data type: this is a more sensible place to put it and it
8351 allows the PKCS#12 code to be tidied up that duplicated this
8355 *) Changed obj_dat.pl script so it takes its input and output files on
8356 the command line. This should avoid shell escape redirection problems
8360 *) Initial support for certificate extension requests, these are included
8361 in things like Xenroll certificate requests. Included functions to allow
8362 extensions to be obtained and added.
8365 *) -crlf option to s_client and s_server for sending newlines as
8366 CRLF (as required by many protocols).
8369 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8371 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8372 [Ralf S. Engelschall]
8374 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8375 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8377 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8381 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8382 DH parameters/keys (q is lost during that conversion, but the resulting
8383 DH parameters contain its length).
8385 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8386 much faster than DH_generate_parameters (which creates parameters
8387 where p = 2*q + 1), and also the smaller q makes DH computations
8388 much more efficient (160-bit exponentiation instead of 1024-bit
8389 exponentiation); so this provides a convenient way to support DHE
8390 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8391 utter importance to use
8392 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8394 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8395 when such DH parameters are used, because otherwise small subgroup
8396 attacks may become possible!
8399 *) Avoid memory leak in i2d_DHparams.
8402 *) Allow the -k option to be used more than once in the enc program:
8403 this allows the same encrypted message to be read by multiple recipients.
8406 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8407 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8408 it will always use the numerical form of the OID, even if it has a short
8412 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8413 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8414 otherwise bn_mod_exp was called. In the case of hardware keys for example
8415 no private key components need be present and it might store extra data
8416 in the RSA structure, which cannot be accessed from bn_mod_exp.
8417 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8418 private key operations.
8421 *) Added support for SPARC Linux.
8424 *) pem_password_cb function type incompatibly changed from
8425 typedef int pem_password_cb(char *buf, int size, int rwflag);
8427 ....(char *buf, int size, int rwflag, void *userdata);
8428 so that applications can pass data to their callbacks:
8429 The PEM[_ASN1]_{read,write}... functions and macros now take an
8430 additional void * argument, which is just handed through whenever
8431 the password callback is called.
8432 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8434 New function SSL_CTX_set_default_passwd_cb_userdata.
8436 Compatibility note: As many C implementations push function arguments
8437 onto the stack in reverse order, the new library version is likely to
8438 interoperate with programs that have been compiled with the old
8439 pem_password_cb definition (PEM_whatever takes some data that
8440 happens to be on the stack as its last argument, and the callback
8441 just ignores this garbage); but there is no guarantee whatsoever that
8444 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8445 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8446 problems not only on Windows, but also on some Unix platforms.
8447 To avoid problematic command lines, these definitions are now in an
8448 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8449 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8452 *) MIPS III/IV assembler module is reimplemented.
8455 *) More DES library cleanups: remove references to srand/rand and
8456 delete an unused file.
8459 *) Add support for the the free Netwide assembler (NASM) under Win32,
8460 since not many people have MASM (ml) and it can be hard to obtain.
8461 This is currently experimental but it seems to work OK and pass all
8462 the tests. Check out INSTALL.W32 for info.
8465 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8466 without temporary keys kept an extra copy of the server key,
8467 and connections with temporary keys did not free everything in case
8471 *) New function RSA_check_key and new openssl rsa option -check
8472 for verifying the consistency of RSA keys.
8473 [Ulf Moeller, Bodo Moeller]
8475 *) Various changes to make Win32 compile work:
8476 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8477 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8478 comparison" warnings.
8479 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8482 *) Add a debugging option to PKCS#5 v2 key generation function: when
8483 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8484 derived keys are printed to stderr.
8487 *) Copy the flags in ASN1_STRING_dup().
8488 [Roman E. Pavlov <pre@mo.msk.ru>]
8490 *) The x509 application mishandled signing requests containing DSA
8491 keys when the signing key was also DSA and the parameters didn't match.
8493 It was supposed to omit the parameters when they matched the signing key:
8494 the verifying software was then supposed to automatically use the CA's
8495 parameters if they were absent from the end user certificate.
8497 Omitting parameters is no longer recommended. The test was also
8498 the wrong way round! This was probably due to unusual behaviour in
8499 EVP_cmp_parameters() which returns 1 if the parameters match.
8500 This meant that parameters were omitted when they *didn't* match and
8501 the certificate was useless. Certificates signed with 'ca' didn't have
8503 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8505 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8506 The interface is as follows:
8507 Applications can use
8508 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8509 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8510 "off" is now the default.
8511 The library internally uses
8512 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8513 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8514 to disable memory-checking temporarily.
8516 Some inconsistent states that previously were possible (and were
8517 even the default) are now avoided.
8519 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8520 with each memory chunk allocated; this is occasionally more helpful
8521 than just having a counter.
8523 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8525 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8529 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8530 which largely parallels "options", but is for changing API behaviour,
8531 whereas "options" are about protocol behaviour.
8532 Initial "mode" flags are:
8534 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8535 a single record has been written.
8536 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8537 retries use the same buffer location.
8538 (But all of the contents must be
8542 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8545 *) Fix problems with no-hmac etc.
8546 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8548 *) New functions RSA_get_default_method(), RSA_set_method() and
8549 RSA_get_method(). These allows replacement of RSA_METHODs without having
8550 to mess around with the internals of an RSA structure.
8553 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8554 Also really enable memory leak checks in openssl.c and in some
8556 [Chad C. Mulligan, Bodo Moeller]
8558 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8559 up the length of negative integers. This has now been simplified to just
8560 store the length when it is first determined and use it later, rather
8561 than trying to keep track of where data is copied and updating it to
8563 [Steve Henson, reported by Brien Wheeler
8564 <bwheeler@authentica-security.com>]
8566 *) Add a new function PKCS7_signatureVerify. This allows the verification
8567 of a PKCS#7 signature but with the signing certificate passed to the
8568 function itself. This contrasts with PKCS7_dataVerify which assumes the
8569 certificate is present in the PKCS#7 structure. This isn't always the
8570 case: certificates can be omitted from a PKCS#7 structure and be
8571 distributed by "out of band" means (such as a certificate database).
8574 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8575 function prototypes in pem.h, also change util/mkdef.pl to add the
8576 necessary function names.
8579 *) mk1mf.pl (used by Windows builds) did not properly read the
8580 options set by Configure in the top level Makefile, and Configure
8581 was not even able to write more than one option correctly.
8582 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8585 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8586 file to be loaded from a BIO or FILE pointer. The BIO version will
8587 for example allow memory BIOs to contain config info.
8590 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8591 Whoever hopes to achieve shared-library compatibility across versions
8592 must use this, not the compile-time macro.
8593 (Exercise 0.9.4: Which is the minimum library version required by
8595 Note: All this applies only to multi-threaded programs, others don't
8599 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8600 through a BIO pair triggered the default case, i.e.
8601 SSLerr(...,SSL_R_UNKNOWN_STATE).
8604 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8605 can use the SSL library even if none of the specific BIOs is
8609 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8610 for the encoded length.
8611 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8613 *) Add initial documentation of the X509V3 functions.
8616 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8617 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8618 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8619 secure PKCS#8 private key format with a high iteration count.
8622 *) Fix determination of Perl interpreter: A perl or perl5
8623 _directory_ in $PATH was also accepted as the interpreter.
8624 [Ralf S. Engelschall]
8626 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8627 wrong with it but it was very old and did things like calling
8628 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8632 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8633 to use the new extension code.
8636 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8637 with macros. This should make it easier to change their form, add extra
8638 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8642 *) Add to configuration table a new entry that can specify an alternative
8643 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8644 according to Mark Crispin <MRC@Panda.COM>.
8648 *) DES CBC did not update the IV. Weird.
8651 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8652 Changing the behaviour of the former might break existing programs --
8653 where IV updating is needed, des_ncbc_encrypt can be used.
8656 *) When bntest is run from "make test" it drives bc to check its
8657 calculations, as well as internally checking them. If an internal check
8658 fails, it needs to cause bc to give a non-zero result or make test carries
8659 on without noticing the failure. Fixed.
8662 *) DES library cleanups.
8665 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8666 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8667 ciphers. NOTE: although the key derivation function has been verified
8668 against some published test vectors it has not been extensively tested
8669 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8673 *) Instead of "mkdir -p", which is not fully portable, use new
8674 Perl script "util/mkdir-p.pl".
8677 *) Rewrite the way password based encryption (PBE) is handled. It used to
8678 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8679 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8680 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8681 the 'parameter' field of the AlgorithmIdentifier is passed to the
8682 underlying key generation function so it must do its own ASN1 parsing.
8683 This has also changed the EVP_PBE_CipherInit() function which now has a
8684 'parameter' argument instead of literal salt and iteration count values
8685 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8688 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8689 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8690 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8691 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8692 value was just used as a "magic string" and not used directly its
8693 value doesn't matter.
8696 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8700 *) "linux-sparc64" configuration (ultrapenguin).
8701 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8702 "linux-sparc" configuration.
8703 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8705 *) config now generates no-xxx options for missing ciphers.
8708 *) Support the EBCDIC character set (work in progress).
8709 File ebcdic.c not yet included because it has a different license.
8710 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8712 *) Support BS2000/OSD-POSIX.
8713 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8715 *) Make callbacks for key generation use void * instead of char *.
8718 *) Make S/MIME samples compile (not yet tested).
8721 *) Additional typesafe stacks.
8724 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8728 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8730 *) New configuration variant "sco5-gcc".
8732 *) Updated some demos.
8733 [Sean O Riordain, Wade Scholine]
8735 *) Add missing BIO_free at exit of pkcs12 application.
8738 *) Fix memory leak in conf.c.
8741 *) Updates for Win32 to assembler version of MD5.
8744 *) Set #! path to perl in apps/der_chop to where we found it
8745 instead of using a fixed path.
8748 *) SHA library changes for irix64-mips4-cc.
8751 *) Improvements for VMS support.
8755 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8757 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8758 This also avoids the problems with SC4.2 and unpatched SC5.
8759 [Andy Polyakov <appro@fy.chalmers.se>]
8761 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8762 These are required because of the typesafe stack would otherwise break
8763 existing code. If old code used a structure member which used to be STACK
8764 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8765 sk_num or sk_value it would produce an error because the num, data members
8766 are not present in STACK_OF. Now it just produces a warning. sk_set
8767 replaces the old method of assigning a value to sk_value
8768 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8769 that does this will no longer work (and should use sk_set instead) but
8770 this could be regarded as a "questionable" behaviour anyway.
8773 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8774 correctly handle encrypted S/MIME data.
8777 *) Change type of various DES function arguments from des_cblock
8778 (which means, in function argument declarations, pointer to char)
8779 to des_cblock * (meaning pointer to array with 8 char elements),
8780 which allows the compiler to do more typechecking; it was like
8781 that back in SSLeay, but with lots of ugly casts.
8783 Introduce new type const_des_cblock.
8786 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8787 problems: find RecipientInfo structure that matches recipient certificate
8788 and initialise the ASN1 structures properly based on passed cipher.
8791 *) Belatedly make the BN tests actually check the results.
8794 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8795 to and from BNs: it was completely broken. New compilation option
8796 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8797 key elements as negative integers.
8800 *) Reorganize and speed up MD5.
8801 [Andy Polyakov <appro@fy.chalmers.se>]
8804 [Richard Levitte <richard@levitte.org>]
8806 *) New option -out to asn1parse to allow the parsed structure to be
8807 output to a file. This is most useful when combined with the -strparse
8808 option to examine the output of things like OCTET STRINGS.
8811 *) Make SSL library a little more fool-proof by not requiring any longer
8812 that SSL_set_{accept,connect}_state be called before
8813 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8814 in many applications because usually everything *appeared* to work as
8815 intended anyway -- now it really works as intended).
8818 *) Move openssl.cnf out of lib/.
8821 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8822 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8823 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8824 [Ralf S. Engelschall]
8826 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8827 handle PKCS#7 enveloped data properly.
8828 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8830 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8831 copying pointers. The cert_st handling is changed by this in
8832 various ways (and thus what used to be known as ctx->default_cert
8833 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8834 any longer when s->cert does not give us what we need).
8835 ssl_cert_instantiate becomes obsolete by this change.
8836 As soon as we've got the new code right (possibly it already is?),
8837 we have solved a couple of bugs of the earlier code where s->cert
8838 was used as if it could not have been shared with other SSL structures.
8840 Note that using the SSL API in certain dirty ways now will result
8841 in different behaviour than observed with earlier library versions:
8842 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8843 does not influence s as it used to.
8845 In order to clean up things more thoroughly, inside SSL_SESSION
8846 we don't use CERT any longer, but a new structure SESS_CERT
8847 that holds per-session data (if available); currently, this is
8848 the peer's certificate chain and, for clients, the server's certificate
8849 and temporary key. CERT holds only those values that can have
8850 meaningful defaults in an SSL_CTX.
8853 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8854 from the internal representation. Various PKCS#7 fixes: remove some
8855 evil casts and set the enc_dig_alg field properly based on the signing
8859 *) Allow PKCS#12 password to be set from the command line or the
8860 environment. Let 'ca' get its config file name from the environment
8861 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8865 *) Allow certificate policies extension to use an IA5STRING for the
8866 organization field. This is contrary to the PKIX definition but
8867 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8871 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8872 without disallowing inline assembler and the like for non-pedantic builds.
8875 *) Support Borland C++ builder.
8876 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8881 *) SHA-1 cleanups and performance enhancements.
8882 [Andy Polyakov <appro@fy.chalmers.se>]
8884 *) Sparc v8plus assembler for the bignum library.
8885 [Andy Polyakov <appro@fy.chalmers.se>]
8887 *) Accept any -xxx and +xxx compiler options in Configure.
8890 *) Update HPUX configuration.
8893 *) Add missing sk_<type>_unshift() function to safestack.h
8894 [Ralf S. Engelschall]
8896 *) New function SSL_CTX_use_certificate_chain_file that sets the
8897 "extra_cert"s in addition to the certificate. (This makes sense
8898 only for "PEM" format files, as chains as a whole are not
8902 *) Support verify_depth from the SSL API.
8903 x509_vfy.c had what can be considered an off-by-one-error:
8904 Its depth (which was not part of the external interface)
8905 was actually counting the number of certificates in a chain;
8906 now it really counts the depth.
8909 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8910 instead of X509err, which often resulted in confusing error
8911 messages since the error codes are not globally unique
8912 (e.g. an alleged error in ssl3_accept when a certificate
8913 didn't match the private key).
8915 *) New function SSL_CTX_set_session_id_context that allows to set a default
8916 value (so that you don't need SSL_set_session_id_context for each
8917 connection using the SSL_CTX).
8920 *) OAEP decoding bug fix.
8923 *) Support INSTALL_PREFIX for package builders, as proposed by
8927 *) New Configure options "threads" and "no-threads". For systems
8928 where the proper compiler options are known (currently Solaris
8929 and Linux), "threads" is the default.
8932 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8935 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8936 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8937 such as /usr/local/bin.
8940 *) "make linux-shared" to build shared libraries.
8941 [Niels Poppe <niels@netbox.org>]
8943 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8946 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8947 extension adding in x509 utility.
8950 *) Remove NOPROTO sections and error code comments.
8953 *) Partial rewrite of the DEF file generator to now parse the ANSI
8957 *) New Configure options --prefix=DIR and --openssldir=DIR.
8960 *) Complete rewrite of the error code script(s). It is all now handled
8961 by one script at the top level which handles error code gathering,
8962 header rewriting and C source file generation. It should be much better
8963 than the old method: it now uses a modified version of Ulf's parser to
8964 read the ANSI prototypes in all header files (thus the old K&R definitions
8965 aren't needed for error creation any more) and do a better job of
8966 translating function codes into names. The old 'ASN1 error code imbedded
8967 in a comment' is no longer necessary and it doesn't use .err files which
8968 have now been deleted. Also the error code call doesn't have to appear all
8969 on one line (which resulted in some large lines...).
8972 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8975 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8976 0 (which usually indicates a closed connection), but continue reading.
8979 *) Fix some race conditions.
8982 *) Add support for CRL distribution points extension. Add Certificate
8983 Policies and CRL distribution points documentation.
8986 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8989 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8990 8 of keying material. Merlin has also confirmed interop with this fix
8991 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8992 [Merlin Hughes <merlin@baltimore.ie>]
8994 *) Fix lots of warnings.
8995 [Richard Levitte <levitte@stacken.kth.se>]
8997 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8998 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8999 [Richard Levitte <levitte@stacken.kth.se>]
9001 *) Fix problems with sizeof(long) == 8.
9002 [Andy Polyakov <appro@fy.chalmers.se>]
9004 *) Change functions to ANSI C.
9007 *) Fix typos in error codes.
9008 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9010 *) Remove defunct assembler files from Configure.
9013 *) SPARC v8 assembler BIGNUM implementation.
9014 [Andy Polyakov <appro@fy.chalmers.se>]
9016 *) Support for Certificate Policies extension: both print and set.
9017 Various additions to support the r2i method this uses.
9020 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9021 return a const string when you are expecting an allocated buffer.
9024 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9025 types DirectoryString and DisplayText.
9028 *) Add code to allow r2i extensions to access the configuration database,
9029 add an LHASH database driver and add several ctx helper functions.
9032 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9033 fail when they extended the size of a BIGNUM.
9036 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9037 support typesafe stack.
9040 *) Fix typo in SSL_[gs]et_options().
9041 [Nils Frostberg <nils@medcom.se>]
9043 *) Delete various functions and files that belonged to the (now obsolete)
9044 old X509V3 handling code.
9047 *) New Configure option "rsaref".
9050 *) Don't auto-generate pem.h.
9053 *) Introduce type-safe ASN.1 SETs.
9056 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9057 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9059 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9060 that links with OpenSSL (well at least cause lots of warnings), but fear
9061 not: the conversion is trivial, and it eliminates loads of evil casts. A
9062 few STACKed things have been converted already. Feel free to convert more.
9063 In the fullness of time, I'll do away with the STACK type altogether.
9066 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9067 specified in <certfile> by updating the entry in the index.txt file.
9068 This way one no longer has to edit the index.txt file manually for
9069 revoking a certificate. The -revoke option does the gory details now.
9070 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9072 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9073 `-text' option at all and this way the `-noout -text' combination was
9074 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9075 [Ralf S. Engelschall]
9077 *) Make sure a corresponding plain text error message exists for the
9078 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9079 verify callback function determined that a certificate was revoked.
9080 [Ralf S. Engelschall]
9082 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9083 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9084 all available cipers including rc5, which was forgotten until now.
9085 In order to let the testing shell script know which algorithms
9086 are available, a new (up to now undocumented) command
9087 "openssl list-cipher-commands" is used.
9090 *) Bugfix: s_client occasionally would sleep in select() when
9091 it should have checked SSL_pending() first.
9094 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9095 the raw DSA values prior to ASN.1 encoding.
9098 *) Tweaks to Configure
9099 [Niels Poppe <niels@netbox.org>]
9101 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9105 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9108 *) New config option to avoid instructions that are illegal on the 80386.
9109 The default code is faster, but requires at least a 486.
9112 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9113 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9114 same as SSL2_VERSION anyway.
9117 *) New "-showcerts" option for s_client.
9120 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9121 application. Various cleanups and fixes.
9124 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9125 modify error routines to work internally. Add error codes and PBE init
9126 to library startup routines.
9129 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9130 packing functions to asn1 and evp. Changed function names and error
9131 codes along the way.
9134 *) PKCS12 integration: and so it begins... First of several patches to
9135 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9136 objects to objects.h
9139 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9140 and display support for Thawte strong extranet extension.
9143 *) Add LinuxPPC support.
9144 [Jeff Dubrule <igor@pobox.org>]
9146 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9147 bn_div_words in alpha.s.
9148 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9150 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9151 OAEP isn't supported when OpenSSL is built with RSAref.
9152 [Ulf Moeller <ulf@fitug.de>]
9154 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9155 so they no longer are missing under -DNOPROTO.
9156 [Soren S. Jorvang <soren@t.dk>]
9159 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9161 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9162 doesn't work when the session is reused. Coming soon!
9165 *) Fix a security hole, that allows sessions to be reused in the wrong
9166 context thus bypassing client cert protection! All software that uses
9167 client certs and session caches in multiple contexts NEEDS PATCHING to
9168 allow session reuse! A fuller solution is in the works.
9169 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9171 *) Some more source tree cleanups (removed obsolete files
9172 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9173 permission on "config" script to be executable) and a fix for the INSTALL
9175 [Ulf Moeller <ulf@fitug.de>]
9177 *) Remove some legacy and erroneous uses of malloc, free instead of
9179 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9181 *) Make rsa_oaep_test return non-zero on error.
9182 [Ulf Moeller <ulf@fitug.de>]
9184 *) Add support for native Solaris shared libraries. Configure
9185 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9186 if someone would make that last step automatic.
9187 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9189 *) ctx_size was not built with the right compiler during "make links". Fixed.
9192 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9193 except NULL ciphers". This means the default cipher list will no longer
9194 enable NULL ciphers. They need to be specifically enabled e.g. with
9195 the string "DEFAULT:eNULL".
9198 *) Fix to RSA private encryption routines: if p < q then it would
9199 occasionally produce an invalid result. This will only happen with
9200 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9203 *) Be less restrictive and allow also `perl util/perlpath.pl
9204 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9205 because this way one can also use an interpreter named `perl5' (which is
9206 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9207 installed as `perl').
9208 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9210 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9211 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9213 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9214 advapi32.lib to Win32 build and change the pem test comparision
9215 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9216 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9217 and crypto/des/ede_cbcm_enc.c.
9220 *) DES quad checksum was broken on big-endian architectures. Fixed.
9223 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9224 Win32 test batch file so it (might) work again. The Win32 test batch file
9225 is horrible: I feel ill....
9228 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9229 in e_os.h. Audit of header files to check ANSI and non ANSI
9230 sections: 10 functions were absent from non ANSI section and not exported
9231 from Windows DLLs. Fixed up libeay.num for new functions.
9234 *) Make `openssl version' output lines consistent.
9235 [Ralf S. Engelschall]
9237 *) Fix Win32 symbol export lists for BIO functions: Added
9238 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9239 to ms/libeay{16,32}.def.
9240 [Ralf S. Engelschall]
9242 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9243 fine under Unix and passes some trivial tests I've now added. But the
9244 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9245 added to make sure no one expects that this stuff really works in the
9246 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9247 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9249 [Ralf S. Engelschall]
9251 *) Fix the generation of two part addresses in perl.
9252 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9254 *) Add config entry for Linux on MIPS.
9255 [John Tobey <jtobey@channel1.com>]
9257 *) Make links whenever Configure is run, unless we are on Windoze.
9260 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9261 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9265 *) Add a useful kludge to allow package maintainers to specify compiler and
9266 other platforms details on the command line without having to patch the
9267 Configure script everytime: One now can use ``perl Configure
9268 <id>:<details>'', i.e. platform ids are allowed to have details appended
9269 to them (seperated by colons). This is treated as there would be a static
9270 pre-configured entry in Configure's %table under key <id> with value
9271 <details> and ``perl Configure <id>'' is called. So, when you want to
9272 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9273 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9274 now, which overrides the FreeBSD-elf entry on-the-fly.
9275 [Ralf S. Engelschall]
9277 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9280 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9281 on the `perl Configure ...' command line. This way one can compile
9282 OpenSSL libraries with Position Independent Code (PIC) which is needed
9283 for linking it into DSOs.
9284 [Ralf S. Engelschall]
9286 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9290 *) Cleaned up the LICENSE document: The official contact for any license
9291 questions now is the OpenSSL core team under openssl-core@openssl.org.
9292 And add a paragraph about the dual-license situation to make sure people
9293 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9294 to the OpenSSL toolkit.
9295 [Ralf S. Engelschall]
9297 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9298 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9299 Additonally cleaned up the `make links' target: Remove unnecessary
9300 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9301 to speed processing and no longer clutter the display with confusing
9302 stuff. Instead only the actually done links are displayed.
9303 [Ralf S. Engelschall]
9305 *) Permit null encryption ciphersuites, used for authentication only. It used
9306 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9307 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9311 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9312 signed attributes when verifying signatures (this would break them),
9313 the detached data encoding was wrong and public keys obtained using
9314 X509_get_pubkey() weren't freed.
9317 *) Add text documentation for the BUFFER functions. Also added a work around
9318 to a Win95 console bug. This was triggered by the password read stuff: the
9319 last character typed gets carried over to the next fread(). If you were
9320 generating a new cert request using 'req' for example then the last
9321 character of the passphrase would be CR which would then enter the first
9325 *) Added the new `Includes OpenSSL Cryptography Software' button as
9326 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9327 button and can be used by applications based on OpenSSL to show the
9328 relationship to the OpenSSL project.
9329 [Ralf S. Engelschall]
9331 *) Remove confusing variables in function signatures in files
9332 ssl/ssl_lib.c and ssl/ssl.h.
9333 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9335 *) Don't install bss_file.c under PREFIX/include/
9336 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9338 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9339 functions that return function pointers and has support for NT specific
9340 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9341 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9342 unsigned to signed types: this was killing the Win32 compile.
9345 *) Add new certificate file to stack functions,
9346 SSL_add_dir_cert_subjects_to_stack() and
9347 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9348 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9349 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9350 This means that Apache-SSL and similar packages don't have to mess around
9351 to add as many CAs as they want to the preferred list.
9354 *) Experiment with doxygen documentation. Currently only partially applied to
9356 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9357 openssl.doxy as the configuration file.
9360 *) Get rid of remaining C++-style comments which strict C compilers hate.
9361 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9363 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9364 compiled in by default: it has problems with large keys.
9367 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9368 DH private keys and/or callback functions which directly correspond to
9369 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9370 is needed for applications which have to configure certificates on a
9371 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9373 For the RSA certificate situation is makes no difference, but
9374 for the DSA certificate situation this fixes the "no shared cipher"
9375 problem where the OpenSSL cipher selection procedure failed because the
9376 temporary keys were not overtaken from the context and the API provided
9377 no way to reconfigure them.
9378 The new functions now let applications reconfigure the stuff and they
9379 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9380 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9381 non-public-API function ssl_cert_instantiate() is used as a helper
9382 function and also to reduce code redundancy inside ssl_rsa.c.
9383 [Ralf S. Engelschall]
9385 *) Move s_server -dcert and -dkey options out of the undocumented feature
9386 area because they are useful for the DSA situation and should be
9387 recognized by the users.
9388 [Ralf S. Engelschall]
9390 *) Fix the cipher decision scheme for export ciphers: the export bits are
9391 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9392 SSL_EXP_MASK. So, the original variable has to be used instead of the
9393 already masked variable.
9394 [Richard Levitte <levitte@stacken.kth.se>]
9396 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9397 [Richard Levitte <levitte@stacken.kth.se>]
9399 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9400 from `int' to `unsigned int' because it's a length and initialized by
9401 EVP_DigestFinal() which expects an `unsigned int *'.
9402 [Richard Levitte <levitte@stacken.kth.se>]
9404 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9405 script. Instead use the usual Shell->Perl transition trick.
9406 [Ralf S. Engelschall]
9408 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9409 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9410 -noout -modulus' as it's already the case for `openssl rsa -noout
9411 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9412 currently the public key is printed (a decision which was already done by
9413 `openssl dsa -modulus' in the past) which serves a similar purpose.
9414 Additionally the NO_RSA no longer completely removes the whole -modulus
9415 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9417 [Ralf S. Engelschall]
9419 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9420 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9421 [Arne Ansper <arne@ats.cyber.ee>]
9423 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9424 to be added. Now both 'req' and 'ca' can use new objects defined in the
9428 *) Add cool BIO that does syslog (or event log on NT).
9429 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9431 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9432 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9433 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9434 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9437 *) Add preliminary config info for new extension code.
9440 *) Make RSA_NO_PADDING really use no padding.
9441 [Ulf Moeller <ulf@fitug.de>]
9443 *) Generate errors when private/public key check is done.
9446 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9447 for some CRL extensions and new objects added.
9450 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9451 key usage extension and fuller support for authority key id.
9454 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9455 padding method for RSA, which is recommended for new applications in PKCS
9456 #1 v2.0 (RFC 2437, October 1998).
9457 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9458 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9459 against Bleichbacher's attack on RSA.
9460 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9463 *) Updates to the new SSL compression code
9464 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9466 *) Fix so that the version number in the master secret, when passed
9467 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9468 (because the server will not accept higher), that the version number
9469 is 0x03,0x01, not 0x03,0x00
9470 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9472 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9473 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9474 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9477 *) Support for RAW extensions where an arbitrary extension can be
9478 created by including its DER encoding. See apps/openssl.cnf for
9482 *) Make sure latest Perl versions don't interpret some generated C array
9483 code as Perl array code in the crypto/err/err_genc.pl script.
9484 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9486 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9487 not many people have the assembler. Various Win32 compilation fixes and
9488 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9492 *) Modify configure script 'Configure' to automatically create crypto/date.h
9493 file under Win32 and also build pem.h from pem.org. New script
9494 util/mkfiles.pl to create the MINFO file on environments that can't do a
9495 'make files': perl util/mkfiles.pl >MINFO should work.
9498 *) Major rework of DES function declarations, in the pursuit of correctness
9499 and purity. As a result, many evil casts evaporated, and some weirdness,
9500 too. You may find this causes warnings in your code. Zapping your evil
9501 casts will probably fix them. Mostly.
9504 *) Fix for a typo in asn1.h. Bug fix to object creation script
9505 obj_dat.pl. It considered a zero in an object definition to mean
9506 "end of object": none of the objects in objects.h have any zeros
9507 so it wasn't spotted.
9508 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9510 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9511 Masking (CBCM). In the absence of test vectors, the best I have been able
9512 to do is check that the decrypt undoes the encrypt, so far. Send me test
9513 vectors if you have them.
9516 *) Correct calculation of key length for export ciphers (too much space was
9517 allocated for null ciphers). This has not been tested!
9520 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9521 message is now correct (it understands "crypto" and "ssl" on its
9522 command line). There is also now an "update" option. This will update
9523 the util/ssleay.num and util/libeay.num files with any new functions.
9525 perl util/mkdef.pl crypto ssl update
9526 it will update them.
9529 *) Overhauled the Perl interface (perl/*):
9530 - ported BN stuff to OpenSSL's different BN library
9531 - made the perl/ source tree CVS-aware
9532 - renamed the package from SSLeay to OpenSSL (the files still contain
9533 their history because I've copied them in the repository)
9534 - removed obsolete files (the test scripts will be replaced
9535 by better Test::Harness variants in the future)
9536 [Ralf S. Engelschall]
9538 *) First cut for a very conservative source tree cleanup:
9539 1. merge various obsolete readme texts into doc/ssleay.txt
9540 where we collect the old documents and readme texts.
9541 2. remove the first part of files where I'm already sure that we no
9542 longer need them because of three reasons: either they are just temporary
9543 files which were left by Eric or they are preserved original files where
9544 I've verified that the diff is also available in the CVS via "cvs diff
9545 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9546 the crypto/md/ stuff).
9547 [Ralf S. Engelschall]
9549 *) More extension code. Incomplete support for subject and issuer alt
9550 name, issuer and authority key id. Change the i2v function parameters
9551 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9552 what that's for :-) Fix to ASN1 macro which messed up
9553 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9556 *) Preliminary support for ENUMERATED type. This is largely copied from the
9560 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9561 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9563 *) Make sure `make rehash' target really finds the `openssl' program.
9564 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9566 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9567 like to hear about it if this slows down other processors.
9570 *) Add CygWin32 platform information to Configure script.
9571 [Alan Batie <batie@aahz.jf.intel.com>]
9573 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9574 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9576 *) New program nseq to manipulate netscape certificate sequences
9579 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9583 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9584 but the BN code had some problems that would cause failures when
9585 doing certificate verification and some other functions.
9586 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9588 *) Add ASN1 and PEM code to support netscape certificate sequences.
9591 *) Add ASN1 and PEM code to support netscape certificate sequences.
9594 *) Add several PKIX and private extended key usage OIDs.
9597 *) Modify the 'ca' program to handle the new extension code. Modify
9598 openssl.cnf for new extension format, add comments.
9601 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9602 and add a sample to openssl.cnf so req -x509 now adds appropriate
9606 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9607 error code, add initial support to X509_print() and x509 application.
9610 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9611 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9612 stuff is currently isolated and isn't even compiled yet.
9615 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9616 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9617 Removed the versions check from X509 routines when loading extensions:
9618 this allows certain broken certificates that don't set the version
9619 properly to be processed.
9622 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9623 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9624 can still be regenerated with "make depend".
9627 *) Spelling mistake in C version of CAST-128.
9628 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9630 *) Changes to the error generation code. The perl script err-code.pl
9631 now reads in the old error codes and retains the old numbers, only
9632 adding new ones if necessary. It also only changes the .err files if new
9633 codes are added. The makefiles have been modified to only insert errors
9634 when needed (to avoid needlessly modifying header files). This is done
9635 by only inserting errors if the .err file is newer than the auto generated
9636 C file. To rebuild all the error codes from scratch (the old behaviour)
9637 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9638 or delete all the .err files.
9641 *) CAST-128 was incorrectly implemented for short keys. The C version has
9642 been fixed, but is untested. The assembler versions are also fixed, but
9643 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9644 to regenerate it if needed.
9645 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9646 Hagino <itojun@kame.net>]
9648 *) File was opened incorrectly in randfile.c.
9649 [Ulf Möller <ulf@fitug.de>]
9651 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9652 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9653 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9654 al: it's just almost always a UTCTime. Note this patch adds new error
9655 codes so do a "make errors" if there are problems.
9658 *) Correct Linux 1 recognition in config.
9659 [Ulf Möller <ulf@fitug.de>]
9661 *) Remove pointless MD5 hash when using DSA keys in ca.
9662 [Anonymous <nobody@replay.com>]
9664 *) Generate an error if given an empty string as a cert directory. Also
9665 generate an error if handed NULL (previously returned 0 to indicate an
9666 error, but didn't set one).
9667 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9669 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9672 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9673 parameters. This was causing a warning which killed off the Win32 compile.
9676 *) Remove C++ style comments from crypto/bn/bn_local.h.
9677 [Neil Costigan <neil.costigan@celocom.com>]
9679 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9680 based on a text string, looking up short and long names and finally
9681 "dot" format. The "dot" format stuff didn't work. Added new function
9682 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9683 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9684 OID is not part of the table.
9687 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9688 X509_LOOKUP_by_alias().
9691 *) Sort openssl functions by name.
9694 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9695 encryption from sample DSA keys (in case anyone is interested the password
9699 *) Make _all_ *_free functions accept a NULL pointer.
9700 [Frans Heymans <fheymans@isaserver.be>]
9702 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9704 [Anonymous <nobody@replay.com>]
9706 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9707 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9709 *) Don't blow it for numeric -newkey arguments to apps/req.
9710 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9712 *) Temp key "for export" tests were wrong in s3_srvr.c.
9713 [Anonymous <nobody@replay.com>]
9715 *) Add prototype for temp key callback functions
9716 SSL_CTX_set_tmp_{rsa,dh}_callback().
9719 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9720 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9723 *) X509_name_add_entry() freed the wrong thing after an error.
9724 [Arne Ansper <arne@ats.cyber.ee>]
9726 *) rsa_eay.c would attempt to free a NULL context.
9727 [Arne Ansper <arne@ats.cyber.ee>]
9729 *) BIO_s_socket() had a broken should_retry() on Windoze.
9730 [Arne Ansper <arne@ats.cyber.ee>]
9732 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9733 [Arne Ansper <arne@ats.cyber.ee>]
9735 *) Make sure the already existing X509_STORE->depth variable is initialized
9736 in X509_STORE_new(), but document the fact that this variable is still
9737 unused in the certificate verification process.
9738 [Ralf S. Engelschall]
9740 *) Fix the various library and apps files to free up pkeys obtained from
9741 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9744 *) Fix reference counting in X509_PUBKEY_get(). This makes
9745 demos/maurice/example2.c work, amongst others, probably.
9746 [Steve Henson and Ben Laurie]
9748 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9749 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9750 are no longer created. This way we have a single and consistent command
9751 line interface `openssl <command>', similar to `cvs <command>'.
9752 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9754 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9755 BIT STRING wrapper always have zero unused bits.
9758 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9761 *) Make the top-level INSTALL documentation easier to understand.
9764 *) Makefiles updated to exit if an error occurs in a sub-directory
9765 make (including if user presses ^C) [Paul Sutton]
9767 *) Make Montgomery context stuff explicit in RSA data structure.
9770 *) Fix build order of pem and err to allow for generated pem.h.
9773 *) Fix renumbering bug in X509_NAME_delete_entry().
9776 *) Enhanced the err-ins.pl script so it makes the error library number
9777 global and can add a library name. This is needed for external ASN1 and
9778 other error libraries.
9781 *) Fixed sk_insert which never worked properly.
9784 *) Fix ASN1 macros so they can handle indefinite length construted
9785 EXPLICIT tags. Some non standard certificates use these: they can now
9789 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9790 into a single doc/ssleay.txt bundle. This way the information is still
9791 preserved but no longer messes up this directory. Now it's new room for
9792 the new set of documenation files.
9793 [Ralf S. Engelschall]
9795 *) SETs were incorrectly DER encoded. This was a major pain, because they
9796 shared code with SEQUENCEs, which aren't coded the same. This means that
9797 almost everything to do with SETs or SEQUENCEs has either changed name or
9798 number of arguments.
9799 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9801 *) Fix test data to work with the above.
9804 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9805 was already fixed by Eric for 0.9.1 it seems.
9806 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9808 *) Autodetect FreeBSD3.
9811 *) Fix various bugs in Configure. This affects the following platforms:
9815 unixware-2.0-pentium
9819 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9820 before they are needed.
9823 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9827 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9829 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9830 changed SSLeay to OpenSSL in version strings.
9831 [Ralf S. Engelschall]
9833 *) Some fixups to the top-level documents.
9836 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9837 because the symlink to include/ was missing.
9838 [Ralf S. Engelschall]
9840 *) Incorporated the popular no-RSA/DSA-only patches
9841 which allow to compile a RSA-free SSLeay.
9842 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9844 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9845 when "ssleay" is still not found.
9846 [Ralf S. Engelschall]
9848 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9849 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9851 *) Updated the README file.
9852 [Ralf S. Engelschall]
9854 *) Added various .cvsignore files in the CVS repository subdirs
9855 to make a "cvs update" really silent.
9856 [Ralf S. Engelschall]
9858 *) Recompiled the error-definition header files and added
9859 missing symbols to the Win32 linker tables.
9860 [Ralf S. Engelschall]
9862 *) Cleaned up the top-level documents;
9863 o new files: CHANGES and LICENSE
9864 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9865 o merged COPYRIGHT into LICENSE
9866 o removed obsolete TODO file
9867 o renamed MICROSOFT to INSTALL.W32
9868 [Ralf S. Engelschall]
9870 *) Removed dummy files from the 0.9.1b source tree:
9871 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9872 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9873 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9874 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9875 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9876 [Ralf S. Engelschall]
9878 *) Added various platform portability fixes.
9881 *) The Genesis of the OpenSSL rpject:
9882 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9883 Young and Tim J. Hudson created while they were working for C2Net until
9885 [The OpenSSL Project]
9888 Changes between 0.9.0b and 0.9.1b [not released]
9890 *) Updated a few CA certificates under certs/
9893 *) Changed some BIGNUM api stuff.
9896 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9897 DGUX x86, Linux Alpha, etc.
9900 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9901 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9905 *) Add -strparse option to asn1pars program which parses nested
9907 [Dr Stephen Henson <shenson@bigfoot.com>]
9909 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9912 *) DSA fix for "ca" program.
9915 *) Added "-genkey" option to "dsaparam" program.
9918 *) Added RIPE MD160 (rmd160) message digest.
9921 *) Added -a (all) option to "ssleay version" command.
9924 *) Added PLATFORM define which is the id given to Configure.
9927 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9930 *) Extended the ASN.1 parser routines.
9933 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9936 *) Added a BN_CTX to the BN library.
9939 *) Fixed the weak key values in DES library
9942 *) Changed API in EVP library for cipher aliases.
9945 *) Added support for RC2/64bit cipher.
9948 *) Converted the lhash library to the crypto/mem.c functions.
9951 *) Added more recognized ASN.1 object ids.
9954 *) Added more RSA padding checks for SSL/TLS.
9957 *) Added BIO proxy/filter functionality.
9960 *) Added extra_certs to SSL_CTX which can be used
9961 send extra CA certificates to the client in the CA cert chain sending
9962 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9965 *) Now Fortezza is denied in the authentication phase because
9966 this is key exchange mechanism is not supported by SSLeay at all.
9969 *) Additional PKCS1 checks.
9972 *) Support the string "TLSv1" for all TLS v1 ciphers.
9975 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9976 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9979 *) Fixed a few memory leaks.
9982 *) Fixed various code and comment typos.
9985 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9986 bytes sent in the client random.
9987 [Edward Bishop <ebishop@spyglass.com>]