5 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
9 Changes between 1.0.1b and 1.0.1c [10 May 2012]
11 *) Sanity check record length before skipping explicit IV in TLS
12 1.2, 1.1 and DTLS to fix DoS attack.
14 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
15 fuzzing as a service testing platform.
19 *) Initialise tkeylen properly when encrypting CMS messages.
20 Thanks to Solar Designer of Openwall for reporting this issue.
23 *) In FIPS mode don't try to use composite ciphers as they are not
27 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
29 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
30 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
31 mean any application compiled against OpenSSL 1.0.0 headers setting
32 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
33 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
34 0x10000000L Any application which was previously compiled against
35 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
36 will need to be recompiled as a result. Letting be results in
37 inability to disable specifically TLS 1.1 and in client context,
38 in unlike event, limit maximum offered version to TLS 1.0 [see below].
41 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
42 disable just protocol X, but all protocols above X *if* there are
43 protocols *below* X still enabled. In more practical terms it means
44 that if application wants to disable TLS1.0 in favor of TLS1.1 and
45 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
46 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
50 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
52 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
53 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
54 in CRYPTO_realloc_clean.
56 Thanks to Tavis Ormandy, Google Security Team, for discovering this
57 issue and to Adam Langley <agl@chromium.org> for fixing it.
59 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
61 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
64 *) Workarounds for some broken servers that "hang" if a client hello
65 record length exceeds 255 bytes.
67 1. Do not use record version number > TLS 1.0 in initial client
68 hello: some (but not all) hanging servers will now work.
69 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
70 the number of ciphers sent in the client hello. This should be
71 set to an even number, such as 50, for example by passing:
72 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
73 Most broken servers should now work.
74 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
75 TLS 1.2 client support entirely.
78 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
81 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
83 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
84 STRING form instead of a DigestInfo.
87 *) The format used for MDC2 RSA signatures is inconsistent between EVP
88 and the RSA_sign/RSA_verify functions. This was made more apparent when
89 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
90 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
91 the correct format in RSA_verify so both forms transparently work.
94 *) Some servers which support TLS 1.0 can choke if we initially indicate
95 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
96 encrypted premaster secret. As a workaround use the maximum pemitted
97 client version in client hello, this should keep such servers happy
98 and still work with previous versions of OpenSSL.
101 *) Add support for TLS/DTLS heartbeats.
102 [Robin Seggelmann <seggelmann@fh-muenster.de>]
104 *) Add support for SCTP.
105 [Robin Seggelmann <seggelmann@fh-muenster.de>]
107 *) Improved PRNG seeding for VOS.
108 [Paul Green <Paul.Green@stratus.com>]
110 *) Extensive assembler packs updates, most notably:
112 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
113 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
114 - x86_64: bit-sliced AES implementation;
115 - ARM: NEON support, contemporary platforms optimizations;
116 - s390x: z196 support;
117 - *: GHASH and GF(2^m) multiplication implementations;
121 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
122 (removal of unnecessary code)
123 [Peter Sylvester <peter.sylvester@edelweb.fr>]
125 *) Add TLS key material exporter from RFC 5705.
128 *) Add DTLS-SRTP negotiation from RFC 5764.
131 *) Add Next Protocol Negotiation,
132 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
133 disabled with a no-npn flag to config or Configure. Code donated
135 [Adam Langley <agl@google.com> and Ben Laurie]
137 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
138 NIST-P256, NIST-P521, with constant-time single point multiplication on
139 typical inputs. Compiler support for the nonstandard type __uint128_t is
140 required to use this (present in gcc 4.4 and later, for 64-bit builds).
141 Code made available under Apache License version 2.0.
143 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
144 line to include this in your build of OpenSSL, and run "make depend" (or
145 "make update"). This enables the following EC_METHODs:
147 EC_GFp_nistp224_method()
148 EC_GFp_nistp256_method()
149 EC_GFp_nistp521_method()
151 EC_GROUP_new_by_curve_name() will automatically use these (while
152 EC_GROUP_new_curve_GFp() currently prefers the more flexible
154 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
156 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
157 all platforms. Move ssize_t definition from e_os.h to the public
158 header file e_os2.h as it now appears in public header file cms.h
161 *) New -sigopt option to the ca, req and x509 utilities. Additional
162 signature parameters can be passed using this option and in
166 *) Add RSA PSS signing function. This will generate and set the
167 appropriate AlgorithmIdentifiers for PSS based on those in the
168 corresponding EVP_MD_CTX structure. No application support yet.
171 *) Support for companion algorithm specific ASN1 signing routines.
172 New function ASN1_item_sign_ctx() signs a pre-initialised
173 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
174 the appropriate parameters.
177 *) Add new algorithm specific ASN1 verification initialisation function
178 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
179 handling will be the same no matter what EVP_PKEY_METHOD is used.
180 Add a PSS handler to support verification of PSS signatures: checked
181 against a number of sample certificates.
184 *) Add signature printing for PSS. Add PSS OIDs.
185 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
187 *) Add algorithm specific signature printing. An individual ASN1 method
188 can now print out signatures instead of the standard hex dump.
190 More complex signatures (e.g. PSS) can print out more meaningful
191 information. Include DSA version that prints out the signature
195 *) Password based recipient info support for CMS library: implementing
199 *) Split password based encryption into PBES2 and PBKDF2 functions. This
200 neatly separates the code into cipher and PBE sections and is required
201 for some algorithms that split PBES2 into separate pieces (such as
205 *) Session-handling fixes:
206 - Fix handling of connections that are resuming with a session ID,
207 but also support Session Tickets.
208 - Fix a bug that suppressed issuing of a new ticket if the client
209 presented a ticket with an expired session.
210 - Try to set the ticket lifetime hint to something reasonable.
211 - Make tickets shorter by excluding irrelevant information.
212 - On the client side, don't ignore renewed tickets.
213 [Adam Langley, Bodo Moeller (Google)]
215 *) Fix PSK session representation.
218 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
220 This work was sponsored by Intel.
223 *) Add GCM support to TLS library. Some custom code is needed to split
224 the IV between the fixed (from PRF) and explicit (from TLS record)
225 portions. This adds all GCM ciphersuites supported by RFC5288 and
226 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
227 add a special AESGCM string for GCM only.
230 *) Expand range of ctrls for AES GCM. Permit setting invocation
231 field on decrypt and retrieval of invocation field only on encrypt.
234 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
235 As required by RFC5289 these ciphersuites cannot be used if for
236 versions of TLS earlier than 1.2.
239 *) For FIPS capable OpenSSL interpret a NULL default public key method
240 as unset and return the appopriate default but do *not* set the default.
241 This means we can return the appopriate method in applications that
242 swicth between FIPS and non-FIPS modes.
245 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
246 ENGINE is used then we cannot handle that in the FIPS module so we
247 keep original code iff non-FIPS operations are allowed.
250 *) Add -attime option to openssl utilities.
251 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
253 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
256 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
257 FIPS EC methods unconditionally for now.
260 *) New build option no-ec2m to disable characteristic 2 code.
263 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
264 all cases can be covered as some introduce binary incompatibilities.
267 *) Redirect RSA operations to FIPS module including keygen,
268 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
271 *) Add similar low level API blocking to ciphers.
274 *) Low level digest APIs are not approved in FIPS mode: any attempt
275 to use these will cause a fatal error. Applications that *really* want
276 to use them can use the private_* version instead.
279 *) Redirect cipher operations to FIPS module for FIPS builds.
282 *) Redirect digest operations to FIPS module for FIPS builds.
285 *) Update build system to add "fips" flag which will link in fipscanister.o
286 for static and shared library builds embedding a signature if needed.
289 *) Output TLS supported curves in preference order instead of numerical
290 order. This is currently hardcoded for the highest order curves first.
291 This should be configurable so applications can judge speed vs strength.
294 *) Add TLS v1.2 server support for client authentication.
297 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
301 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
302 FIPS modules versions.
305 *) Add TLS v1.2 client side support for client authentication. Keep cache
306 of handshake records longer as we don't know the hash algorithm to use
307 until after the certificate request message is received.
310 *) Initial TLS v1.2 client support. Add a default signature algorithms
311 extension including all the algorithms we support. Parse new signature
312 format in client key exchange. Relax some ECC signing restrictions for
313 TLS v1.2 as indicated in RFC5246.
316 *) Add server support for TLS v1.2 signature algorithms extension. Switch
317 to new signature format when needed using client digest preference.
318 All server ciphersuites should now work correctly in TLS v1.2. No client
319 support yet and no support for client certificates.
322 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
323 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
324 ciphersuites. At present only RSA key exchange ciphersuites work with
325 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
326 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
327 and version checking.
330 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
331 with this defined it will not be affected by any changes to ssl internal
332 structures. Add several utility functions to allow openssl application
333 to work with OPENSSL_NO_SSL_INTERN defined.
337 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
339 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
342 *) Permit abbreviated handshakes when renegotiating using the function
343 SSL_renegotiate_abbreviated().
344 [Robin Seggelmann <seggelmann@fh-muenster.de>]
346 *) Add call to ENGINE_register_all_complete() to
347 ENGINE_load_builtin_engines(), so some implementations get used
348 automatically instead of needing explicit application support.
351 *) Add support for TLS key exporter as described in RFC5705.
352 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
354 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
355 a few changes are required:
357 Add SSL_OP_NO_TLSv1_1 flag.
359 Update version checking logic to handle version 1.1.
360 Add explicit IV handling (ported from DTLS code).
361 Add command line options to s_client/s_server.
364 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
366 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
367 in CMS and PKCS7 code. When RSA decryption fails use a random key for
368 content decryption and always return the same error. Note: this attack
369 needs on average 2^20 messages so it only affects automated senders. The
370 old behaviour can be reenabled in the CMS code by setting the
371 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
372 an MMA defence is not necessary.
373 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
374 this issue. (CVE-2012-0884)
377 *) Fix CVE-2011-4619: make sure we really are receiving a
378 client hello before rejecting multiple SGC restarts. Thanks to
379 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
382 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
384 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
385 Thanks to Antonio Martin, Enterprise Secure Access Research and
386 Development, Cisco Systems, Inc. for discovering this bug and
387 preparing a fix. (CVE-2012-0050)
390 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
392 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
393 of the Vaudenay padding oracle attack on CBC mode encryption
394 which enables an efficient plaintext recovery attack against
395 the OpenSSL implementation of DTLS. Their attack exploits timing
396 differences arising during decryption processing. A research
397 paper describing this attack can be found at:
398 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
399 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
400 Security Group at Royal Holloway, University of London
401 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
402 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
403 for preparing the fix. (CVE-2011-4108)
404 [Robin Seggelmann, Michael Tuexen]
406 *) Clear bytes used for block padding of SSL 3.0 records.
408 [Adam Langley (Google)]
410 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
411 Kadianakis <desnacked@gmail.com> for discovering this issue and
412 Adam Langley for preparing the fix. (CVE-2011-4619)
413 [Adam Langley (Google)]
415 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
416 [Andrey Kulikov <amdeich@gmail.com>]
418 *) Prevent malformed RFC3779 data triggering an assertion failure.
419 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
420 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
421 [Rob Austein <sra@hactrn.net>]
423 *) Improved PRNG seeding for VOS.
424 [Paul Green <Paul.Green@stratus.com>]
426 *) Fix ssl_ciph.c set-up race.
427 [Adam Langley (Google)]
429 *) Fix spurious failures in ecdsatest.c.
430 [Emilia Käsper (Google)]
432 *) Fix the BIO_f_buffer() implementation (which was mixing different
433 interpretations of the '..._len' fields).
434 [Adam Langley (Google)]
436 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
437 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
438 threads won't reuse the same blinding coefficients.
440 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
441 lock to call BN_BLINDING_invert_ex, and avoids one use of
442 BN_BLINDING_update for each BN_BLINDING structure (previously,
443 the last update always remained unused).
444 [Emilia Käsper (Google)]
446 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
447 [Bob Buckholz (Google)]
449 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
451 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
452 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
453 [Kaspar Brand <ossl@velox.ch>]
455 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
456 for multi-threaded use of ECDH. (CVE-2011-3210)
457 [Adam Langley (Google)]
459 *) Fix x509_name_ex_d2i memory leak on bad inputs.
462 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
463 signature public key algorithm by using OID xref utilities instead.
464 Before this you could only use some ECC ciphersuites with SHA1 only.
467 *) Add protection against ECDSA timing attacks as mentioned in the paper
468 by Billy Bob Brumley and Nicola Tuveri, see:
470 http://eprint.iacr.org/2011/232.pdf
472 [Billy Bob Brumley and Nicola Tuveri]
474 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
476 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
477 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
479 *) Fix bug in string printing code: if *any* escaping is enabled we must
480 escape the escape character (backslash) or the resulting string is
484 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
486 *) Disable code workaround for ancient and obsolete Netscape browsers
487 and servers: an attacker can use it in a ciphersuite downgrade attack.
488 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
491 *) Fixed J-PAKE implementation error, originally discovered by
492 Sebastien Martini, further info and confirmation from Stefan
493 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
496 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
498 *) Fix extension code to avoid race conditions which can result in a buffer
499 overrun vulnerability: resumed sessions must not be modified as they can
500 be shared by multiple threads. CVE-2010-3864
503 *) Fix WIN32 build system to correctly link an ENGINE directory into
507 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
509 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
511 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
513 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
515 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
516 context. The operation can be customised via the ctrl mechanism in
517 case ENGINEs want to include additional functionality.
520 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
523 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
524 output hashes compatible with older versions of OpenSSL.
525 [Willy Weisz <weisz@vcpc.univie.ac.at>]
527 *) Fix compression algorithm handling: if resuming a session use the
528 compression algorithm of the resumed session instead of determining
529 it from client hello again. Don't allow server to change algorithm.
532 *) Add load_crls() function to apps tidying load_certs() too. Add option
533 to verify utility to allow additional CRLs to be included.
536 *) Update OCSP request code to permit adding custom headers to the request:
537 some responders need this.
540 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
542 [Julia Lawall <julia@diku.dk>]
544 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
545 needlessly dereferenced structures, used obsolete functions and
546 didn't handle all updated verify codes correctly.
549 *) Disable MD2 in the default configuration.
552 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
553 indicate the initial BIO being pushed or popped. This makes it possible
554 to determine whether the BIO is the one explicitly called or as a result
555 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
556 it handles reference counts correctly and doesn't zero out the I/O bio
557 when it is not being explicitly popped. WARNING: applications which
558 included workarounds for the old buggy behaviour will need to be modified
559 or they could free up already freed BIOs.
562 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
563 renaming to all platforms (within the 0.9.8 branch, this was
564 done conditionally on Netware platforms to avoid a name clash).
565 [Guenter <lists@gknw.net>]
567 *) Add ECDHE and PSK support to DTLS.
568 [Michael Tuexen <tuexen@fh-muenster.de>]
570 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
574 *) Add "missing" function EVP_MD_flags() (without this the only way to
575 retrieve a digest flags is by accessing the structure directly. Update
576 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
577 or cipher is registered as in the "from" argument. Print out all
578 registered digests in the dgst usage message instead of manually
579 attempting to work them out.
582 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
583 this allows the use of compression and extensions. Change default cipher
584 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
585 by default unless an application cipher string requests it.
588 *) Alter match criteria in PKCS12_parse(). It used to try to use local
589 key ids to find matching certificates and keys but some PKCS#12 files
590 don't follow the (somewhat unwritten) rules and this strategy fails.
591 Now just gather all certificates together and the first private key
592 then look for the first certificate that matches the key.
595 *) Support use of registered digest and cipher names for dgst and cipher
596 commands instead of having to add each one as a special case. So now
603 openssl dgst -sha256 foo
605 and this works for ENGINE based algorithms too.
609 *) Update Gost ENGINE to support parameter files.
610 [Victor B. Wagner <vitus@cryptocom.ru>]
612 *) Support GeneralizedTime in ca utility.
613 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
615 *) Enhance the hash format used for certificate directory links. The new
616 form uses the canonical encoding (meaning equivalent names will work
617 even if they aren't identical) and uses SHA1 instead of MD5. This form
618 is incompatible with the older format and as a result c_rehash should
619 be used to rebuild symbolic links.
622 *) Make PKCS#8 the default write format for private keys, replacing the
623 traditional format. This form is standardised, more secure and doesn't
624 include an implicit MD5 dependency.
627 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
628 committed to OpenSSL should pass this lot as a minimum.
631 *) Add session ticket override functionality for use by EAP-FAST.
632 [Jouni Malinen <j@w1.fi>]
634 *) Modify HMAC functions to return a value. Since these can be implemented
635 in an ENGINE errors can occur.
638 *) Type-checked OBJ_bsearch_ex.
641 *) Type-checked OBJ_bsearch. Also some constification necessitated
642 by type-checking. Still to come: TXT_DB, bsearch(?),
643 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
647 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
648 seconds to a tm structure directly, instead of going through OS
649 specific date routines. This avoids any issues with OS routines such
650 as the year 2038 bug. New *_adj() functions for ASN1 time structures
651 and X509_time_adj_ex() to cover the extended range. The existing
652 X509_time_adj() is still usable and will no longer have any date issues.
655 *) Delta CRL support. New use deltas option which will attempt to locate
656 and search any appropriate delta CRLs available.
658 This work was sponsored by Google.
661 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
662 code and add additional score elements. Validate alternate CRL paths
663 as part of the CRL checking and indicate a new error "CRL path validation
664 error" in this case. Applications wanting additional details can use
665 the verify callback and check the new "parent" field. If this is not
666 NULL CRL path validation is taking place. Existing applications wont
667 see this because it requires extended CRL support which is off by
670 This work was sponsored by Google.
673 *) Support for freshest CRL extension.
675 This work was sponsored by Google.
678 *) Initial indirect CRL support. Currently only supported in the CRLs
679 passed directly and not via lookup. Process certificate issuer
680 CRL entry extension and lookup CRL entries by bother issuer name
681 and serial number. Check and process CRL issuer entry in IDP extension.
683 This work was sponsored by Google.
686 *) Add support for distinct certificate and CRL paths. The CRL issuer
687 certificate is validated separately in this case. Only enabled if
688 an extended CRL support flag is set: this flag will enable additional
689 CRL functionality in future.
691 This work was sponsored by Google.
694 *) Add support for policy mappings extension.
696 This work was sponsored by Google.
699 *) Fixes to pathlength constraint, self issued certificate handling,
700 policy processing to align with RFC3280 and PKITS tests.
702 This work was sponsored by Google.
705 *) Support for name constraints certificate extension. DN, email, DNS
706 and URI types are currently supported.
708 This work was sponsored by Google.
711 *) To cater for systems that provide a pointer-based thread ID rather
712 than numeric, deprecate the current numeric thread ID mechanism and
713 replace it with a structure and associated callback type. This
714 mechanism allows a numeric "hash" to be extracted from a thread ID in
715 either case, and on platforms where pointers are larger than 'long',
716 mixing is done to help ensure the numeric 'hash' is usable even if it
717 can't be guaranteed unique. The default mechanism is to use "&errno"
718 as a pointer-based thread ID to distinguish between threads.
720 Applications that want to provide their own thread IDs should now use
721 CRYPTO_THREADID_set_callback() to register a callback that will call
722 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
724 Note that ERR_remove_state() is now deprecated, because it is tied
725 to the assumption that thread IDs are numeric. ERR_remove_state(0)
726 to free the current thread's error state should be replaced by
727 ERR_remove_thread_state(NULL).
729 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
730 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
731 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
732 application was previously providing a numeric thread callback that
733 was inappropriate for distinguishing threads, then uniqueness might
734 have been obtained with &errno that happened immediately in the
735 intermediate development versions of OpenSSL; this is no longer the
736 case, the numeric thread callback will now override the automatic use
738 [Geoff Thorpe, with help from Bodo Moeller]
740 *) Initial support for different CRL issuing certificates. This covers a
741 simple case where the self issued certificates in the chain exist and
742 the real CRL issuer is higher in the existing chain.
744 This work was sponsored by Google.
747 *) Removed effectively defunct crypto/store from the build.
750 *) Revamp of STACK to provide stronger type-checking. Still to come:
751 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
752 ASN1_STRING, CONF_VALUE.
755 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
756 RAM on SSL connections. This option can save about 34k per idle SSL.
759 *) Revamp of LHASH to provide stronger type-checking. Still to come:
760 STACK, TXT_DB, bsearch, qsort.
763 *) Initial support for Cryptographic Message Syntax (aka CMS) based
764 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
765 support for data, signedData, compressedData, digestedData and
766 encryptedData, envelopedData types included. Scripts to check against
767 RFC4134 examples draft and interop and consistency checks of many
768 content types and variants.
771 *) Add options to enc utility to support use of zlib compression BIO.
774 *) Extend mk1mf to support importing of options and assembly language
775 files from Configure script, currently only included in VC-WIN32.
776 The assembly language rules can now optionally generate the source
777 files from the associated perl scripts.
780 *) Implement remaining functionality needed to support GOST ciphersuites.
781 Interop testing has been performed using CryptoPro implementations.
782 [Victor B. Wagner <vitus@cryptocom.ru>]
784 *) s390x assembler pack.
787 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
791 *) Implement Opaque PRF Input TLS extension as specified in
792 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
793 official specification yet and no extension type assignment by
794 IANA exists, this extension (for now) will have to be explicitly
795 enabled when building OpenSSL by providing the extension number
796 to use. For example, specify an option
798 -DTLSEXT_TYPE_opaque_prf_input=0x9527
800 to the "config" or "Configure" script to enable the extension,
801 assuming extension number 0x9527 (which is a completely arbitrary
802 and unofficial assignment based on the MD5 hash of the Internet
803 Draft). Note that by doing so, you potentially lose
804 interoperability with other TLS implementations since these might
805 be using the same extension number for other purposes.
807 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
808 opaque PRF input value to use in the handshake. This will create
809 an interal copy of the length-'len' string at 'src', and will
810 return non-zero for success.
812 To get more control and flexibility, provide a callback function
815 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
816 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
820 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
823 Callback function 'cb' will be called in handshakes, and is
824 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
825 Argument 'arg' is for application purposes (the value as given to
826 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
827 be provided to the callback function). The callback function
828 has to return non-zero to report success: usually 1 to use opaque
829 PRF input just if possible, or 2 to enforce use of the opaque PRF
830 input. In the latter case, the library will abort the handshake
831 if opaque PRF input is not successfully negotiated.
833 Arguments 'peerinput' and 'len' given to the callback function
834 will always be NULL and 0 in the case of a client. A server will
835 see the client's opaque PRF input through these variables if
836 available (NULL and 0 otherwise). Note that if the server
837 provides an opaque PRF input, the length must be the same as the
838 length of the client's opaque PRF input.
840 Note that the callback function will only be called when creating
841 a new session (session resumption can resume whatever was
842 previously negotiated), and will not be called in SSL 2.0
843 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
844 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
845 for applications that need to enforce opaque PRF input.
849 *) Update ssl code to support digests other than SHA1+MD5 for handshake
852 [Victor B. Wagner <vitus@cryptocom.ru>]
854 *) Add RFC4507 support to OpenSSL. This includes the corrections in
855 RFC4507bis. The encrypted ticket format is an encrypted encoded
856 SSL_SESSION structure, that way new session features are automatically
859 If a client application caches session in an SSL_SESSION structure
860 support is transparent because tickets are now stored in the encoded
863 The SSL_CTX structure automatically generates keys for ticket
864 protection in servers so again support should be possible
865 with no application modification.
867 If a client or server wishes to disable RFC4507 support then the option
868 SSL_OP_NO_TICKET can be set.
870 Add a TLS extension debugging callback to allow the contents of any client
871 or server extensions to be examined.
873 This work was sponsored by Google.
876 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
877 OpenSSL should now compile cleanly on gcc 4.2
878 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
880 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
881 support including streaming MAC support: this is required for GOST
883 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
885 *) Add option -stream to use PKCS#7 streaming in smime utility. New
886 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
887 to output in BER and PEM format.
890 *) Experimental support for use of HMAC via EVP_PKEY interface. This
891 allows HMAC to be handled via the EVP_DigestSign*() interface. The
892 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
893 ENGINE support for HMAC keys which are unextractable. New -mac and
894 -macopt options to dgst utility.
897 *) New option -sigopt to dgst utility. Update dgst to use
898 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
899 alternative signing paramaters such as X9.31 or PSS in the dgst
903 *) Change ssl_cipher_apply_rule(), the internal function that does
904 the work each time a ciphersuite string requests enabling
905 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
906 removing ("!foo+bar") a class of ciphersuites: Now it maintains
907 the order of disabled ciphersuites such that those ciphersuites
908 that most recently went from enabled to disabled not only stay
909 in order with respect to each other, but also have higher priority
910 than other disabled ciphersuites the next time ciphersuites are
913 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
914 the same ciphersuites as with "HIGH" alone, but in a specific
915 order where the PSK ciphersuites come first (since they are the
916 most recently disabled ciphersuites when "HIGH" is parsed).
918 Also, change ssl_create_cipher_list() (using this new
919 funcionality) such that between otherwise identical
920 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
924 *) Change ssl_create_cipher_list() so that it automatically
925 arranges the ciphersuites in reasonable order before starting
926 to process the rule string. Thus, the definition for "DEFAULT"
927 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
928 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
929 This makes it much easier to arrive at a reasonable default order
930 in applications for which anonymous ciphers are OK (meaning
931 that you can't actually use DEFAULT).
932 [Bodo Moeller; suggested by Victor Duchovni]
934 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
935 processing) into multiple integers instead of setting
936 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
937 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
938 (These masks as well as the individual bit definitions are hidden
939 away into the non-exported interface ssl/ssl_locl.h, so this
940 change to the definition of the SSL_CIPHER structure shouldn't
941 affect applications.) This give us more bits for each of these
942 categories, so there is no longer a need to coagulate AES128 and
943 AES256 into a single algorithm bit, and to coagulate Camellia128
944 and Camellia256 into a single algorithm bit, which has led to all
947 Thus, among other things, the kludge introduced in 0.9.7m and
948 0.9.8e for masking out AES256 independently of AES128 or masking
949 out Camellia256 independently of AES256 is not needed here in 0.9.9.
951 With the change, we also introduce new ciphersuite aliases that
952 so far were missing: "AES128", "AES256", "CAMELLIA128", and
956 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
957 Use the leftmost N bytes of the signature input if the input is
958 larger than the prime q (with N being the size in bytes of q).
961 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
962 it yet and it is largely untested.
965 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
968 *) Initial incomplete changes to avoid need for function casts in OpenSSL
969 some compilers (gcc 4.2 and later) reject their use. Safestack is
970 reimplemented. Update ASN1 to avoid use of legacy functions.
973 *) Win32/64 targets are linked with Winsock2.
976 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
977 to external functions. This can be used to increase CRL handling
978 efficiency especially when CRLs are very large by (for example) storing
979 the CRL revoked certificates in a database.
982 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
983 new CRLs added to a directory can be used. New command line option
984 -verify_return_error to s_client and s_server. This causes real errors
985 to be returned by the verify callback instead of carrying on no matter
986 what. This reflects the way a "real world" verify callback would behave.
989 *) GOST engine, supporting several GOST algorithms and public key formats.
990 Kindly donated by Cryptocom.
993 *) Partial support for Issuing Distribution Point CRL extension. CRLs
994 partitioned by DP are handled but no indirect CRL or reason partitioning
995 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
996 selected via a scoring technique which handles IDP and AKID in CRLs.
999 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1000 will ultimately be used for all verify operations: this will remove the
1001 X509_STORE dependency on certificate verification and allow alternative
1002 lookup methods. X509_STORE based implementations of these two callbacks.
1005 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1006 Modify get_crl() to find a valid (unexpired) CRL if possible.
1009 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1010 this would be called X509_CRL_cmp() but that name is already used by
1011 a function that just compares CRL issuer names. Cache several CRL
1012 extensions in X509_CRL structure and cache CRLDP in X509.
1015 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1016 this maps equivalent X509_NAME structures into a consistent structure.
1017 Name comparison can then be performed rapidly using memcmp().
1020 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1024 *) Allow digests to supply their own micalg string for S/MIME type using
1025 the ctrl EVP_MD_CTRL_MICALG.
1028 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1029 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1030 ctrl. It can then customise the structure before and/or after signing
1034 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1035 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1036 to free up any added signature OIDs.
1039 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1040 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1041 digest and cipher tables. New options added to openssl utility:
1042 list-message-digest-algorithms and list-cipher-algorithms.
1045 *) Change the array representation of binary polynomials: the list
1046 of degrees of non-zero coefficients is now terminated with -1.
1047 Previously it was terminated with 0, which was also part of the
1048 value; thus, the array representation was not applicable to
1049 polynomials where t^0 has coefficient zero. This change makes
1050 the array representation useful in a more general context.
1053 *) Various modifications and fixes to SSL/TLS cipher string
1054 handling. For ECC, the code now distinguishes between fixed ECDH
1055 with RSA certificates on the one hand and with ECDSA certificates
1056 on the other hand, since these are separate ciphersuites. The
1057 unused code for Fortezza ciphersuites has been removed.
1059 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1060 (not "ECDHE"). For consistency with the code for DH
1061 certificates, use of ECDH certificates is now considered ECDH
1062 authentication, not RSA or ECDSA authentication (the latter is
1063 merely the CA's signing algorithm and not actively used in the
1066 The temporary ciphersuite alias "ECCdraft" is no longer
1067 available, and ECC ciphersuites are no longer excluded from "ALL"
1068 and "DEFAULT". The following aliases now exist for RFC 4492
1069 ciphersuites, most of these by analogy with the DH case:
1071 kECDHr - ECDH cert, signed with RSA
1072 kECDHe - ECDH cert, signed with ECDSA
1073 kECDH - ECDH cert (signed with either RSA or ECDSA)
1074 kEECDH - ephemeral ECDH
1075 ECDH - ECDH cert or ephemeral ECDH
1081 AECDH - anonymous ECDH
1082 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1086 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1087 Use correct micalg parameters depending on digest(s) in signed message.
1090 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1091 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1094 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1095 an engine to register a method. Add ENGINE lookups for methods and
1096 functional reference processing.
1099 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1100 EVP_{Sign,Verify}* which allow an application to customise the signature
1104 *) New -resign option to smime utility. This adds one or more signers
1105 to an existing PKCS#7 signedData structure. Also -md option to use an
1106 alternative message digest algorithm for signing.
1109 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1110 create PKCS7 structures containing multiple signers. Update smime
1111 application to support multiple signers.
1114 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1118 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1119 Reorganize PBE internals to lookup from a static table using NIDs,
1120 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1121 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1122 PRF which will be automatically used with PBES2.
1125 *) Replace the algorithm specific calls to generate keys in "req" with the
1129 *) Update PKCS#7 enveloped data routines to use new API. This is now
1130 supported by any public key method supporting the encrypt operation. A
1131 ctrl is added to allow the public key algorithm to examine or modify
1132 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1136 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1137 a default digest type to use. In most cases this will be SHA1 but some
1138 algorithms (such as GOST) need to specify an alternative digest. The
1139 return value indicates how strong the prefernce is 1 means optional and
1140 2 is mandatory (that is it is the only supported type). Modify
1141 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1142 use the default md. Update openssl utilities to use the default digest
1143 type for signing if it is not explicitly indicated.
1146 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1147 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1148 signing method from the key type. This effectively removes the link
1149 between digests and public key types.
1152 *) Add an OID cross reference table and utility functions. Its purpose is to
1153 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1154 rsaEncryption. This will allow some of the algorithm specific hackery
1155 needed to use the correct OID to be removed.
1158 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1159 structures for PKCS7_sign(). They are now set up by the relevant public
1163 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1166 *) Add support for key derivation (agreement) in the API, DH method and
1170 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1171 public and private key formats. As a side effect these add additional
1172 command line functionality not previously available: DSA signatures can be
1173 generated and verified using pkeyutl and DH key support and generation in
1178 [Oliver Tappe <zooey@hirschkaefer.de>]
1180 *) New make target "install_html_docs" installs HTML renditions of the
1182 [Oliver Tappe <zooey@hirschkaefer.de>]
1184 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1185 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1186 support key and parameter generation and add initial key generation
1187 functionality for RSA.
1190 *) Add functions for main EVP_PKEY_method operations. The undocumented
1191 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1192 EVP_PKEY_{encrypt,decrypt}_old.
1195 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1196 key API, doesn't do much yet.
1199 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1200 public key algorithms. New option to openssl utility:
1201 "list-public-key-algorithms" to print out info.
1204 *) Implement the Supported Elliptic Curves Extension for
1205 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1208 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1209 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1212 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1213 utilities such as rsa, dsa, dsaparam etc except they process any key
1217 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1218 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1219 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1223 *) Initial support for pluggable public key ASN1.
1224 De-spaghettify the public key ASN1 handling. Move public and private
1225 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1226 algorithm specific handling to a single module within the relevant
1227 algorithm directory. Add functions to allow (near) opaque processing
1228 of public and private key structures.
1231 *) Implement the Supported Point Formats Extension for
1232 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1235 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1236 for the psk identity [hint] and the psk callback functions to the
1237 SSL_SESSION, SSL and SSL_CTX structure.
1240 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1244 SSL_CTX_use_psk_identity_hint
1245 SSL_get_psk_identity_hint
1246 SSL_get_psk_identity
1247 SSL_use_psk_identity_hint
1249 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1251 *) Add RFC 3161 compliant time stamp request creation, response generation
1252 and response verification functionality.
1253 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1255 *) Add initial support for TLS extensions, specifically for the server_name
1256 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1257 have new members for a host name. The SSL data structure has an
1258 additional member SSL_CTX *initial_ctx so that new sessions can be
1259 stored in that context to allow for session resumption, even after the
1260 SSL has been switched to a new SSL_CTX in reaction to a client's
1261 server_name extension.
1263 New functions (subject to change):
1265 SSL_get_servername()
1266 SSL_get_servername_type()
1269 New CTRL codes and macros (subject to change):
1271 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1272 - SSL_CTX_set_tlsext_servername_callback()
1273 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1274 - SSL_CTX_set_tlsext_servername_arg()
1275 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1277 openssl s_client has a new '-servername ...' option.
1279 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1280 '-key2 ...', '-servername_fatal' (subject to change). This allows
1281 testing the HostName extension for a specific single host name ('-cert'
1282 and '-key' remain fallbacks for handshakes without HostName
1283 negotiation). If the unrecogninzed_name alert has to be sent, this by
1284 default is a warning; it becomes fatal with the '-servername_fatal'
1287 [Peter Sylvester, Remy Allais, Christophe Renou]
1289 *) Whirlpool hash implementation is added.
1292 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1293 bn(64,32). Because of instruction set limitations it doesn't have
1294 any negative impact on performance. This was done mostly in order
1295 to make it possible to share assembler modules, such as bn_mul_mont
1296 implementations, between 32- and 64-bit builds without hassle.
1299 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1300 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1304 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1305 dedicated Montgomery multiplication procedure, is introduced.
1306 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1307 "64-bit" performance on certain 32-bit targets.
1310 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1311 in SSL structures. New SSL ctrl to set maximum send fragment size.
1312 Save memory by seeting the I/O buffer sizes dynamically instead of
1313 using the maximum available value.
1316 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1317 in addition to the text details.
1320 *) Very, very preliminary EXPERIMENTAL support for printing of general
1321 ASN1 structures. This currently produces rather ugly output and doesn't
1322 handle several customised structures at all.
1325 *) Integrated support for PVK file format and some related formats such
1326 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1327 these in the 'rsa' and 'dsa' utilities.
1330 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1333 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1334 place for the (very old) "NETSCAPE" format certificates which are now
1335 handled using new ASN1 code equivalents.
1338 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1339 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1340 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1343 *) Modify CRL distribution points extension code to print out previously
1344 unsupported fields. Enhance extension setting code to allow setting of
1348 *) Add print and set support for Issuing Distribution Point CRL extension.
1351 *) Change 'Configure' script to enable Camellia by default.
1354 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1356 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1357 Thanks to Antonio Martin, Enterprise Secure Access Research and
1358 Development, Cisco Systems, Inc. for discovering this bug and
1359 preparing a fix. (CVE-2012-0050)
1362 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1364 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1365 of the Vaudenay padding oracle attack on CBC mode encryption
1366 which enables an efficient plaintext recovery attack against
1367 the OpenSSL implementation of DTLS. Their attack exploits timing
1368 differences arising during decryption processing. A research
1369 paper describing this attack can be found at:
1370 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1371 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1372 Security Group at Royal Holloway, University of London
1373 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1374 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1375 for preparing the fix. (CVE-2011-4108)
1376 [Robin Seggelmann, Michael Tuexen]
1378 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1379 [Ben Laurie, Kasper <ekasper@google.com>]
1381 *) Clear bytes used for block padding of SSL 3.0 records.
1383 [Adam Langley (Google)]
1385 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1386 Kadianakis <desnacked@gmail.com> for discovering this issue and
1387 Adam Langley for preparing the fix. (CVE-2011-4619)
1388 [Adam Langley (Google)]
1390 *) Prevent malformed RFC3779 data triggering an assertion failure.
1391 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1392 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1393 [Rob Austein <sra@hactrn.net>]
1395 *) Fix ssl_ciph.c set-up race.
1396 [Adam Langley (Google)]
1398 *) Fix spurious failures in ecdsatest.c.
1399 [Emilia Käsper (Google)]
1401 *) Fix the BIO_f_buffer() implementation (which was mixing different
1402 interpretations of the '..._len' fields).
1403 [Adam Langley (Google)]
1405 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1406 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1407 threads won't reuse the same blinding coefficients.
1409 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1410 lock to call BN_BLINDING_invert_ex, and avoids one use of
1411 BN_BLINDING_update for each BN_BLINDING structure (previously,
1412 the last update always remained unused).
1413 [Emilia Käsper (Google)]
1415 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1416 for multi-threaded use of ECDH.
1417 [Adam Langley (Google)]
1419 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1422 *) Add protection against ECDSA timing attacks as mentioned in the paper
1423 by Billy Bob Brumley and Nicola Tuveri, see:
1425 http://eprint.iacr.org/2011/232.pdf
1427 [Billy Bob Brumley and Nicola Tuveri]
1429 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1431 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1432 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1434 *) Fix bug in string printing code: if *any* escaping is enabled we must
1435 escape the escape character (backslash) or the resulting string is
1439 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1441 *) Disable code workaround for ancient and obsolete Netscape browsers
1442 and servers: an attacker can use it in a ciphersuite downgrade attack.
1443 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1446 *) Fixed J-PAKE implementation error, originally discovered by
1447 Sebastien Martini, further info and confirmation from Stefan
1448 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1451 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1453 *) Fix extension code to avoid race conditions which can result in a buffer
1454 overrun vulnerability: resumed sessions must not be modified as they can
1455 be shared by multiple threads. CVE-2010-3864
1458 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1461 *) Don't reencode certificate when calculating signature: cache and use
1462 the original encoding instead. This makes signature verification of
1463 some broken encodings work correctly.
1466 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1467 is also one of the inputs.
1468 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1470 *) Don't repeatedly append PBE algorithms to table if they already exist.
1471 Sort table on each new add. This effectively makes the table read only
1472 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1476 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1478 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1481 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1482 access or freeing data twice (CVE-2010-0742)
1483 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1485 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1486 common in certificates and some applications which only call
1487 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1491 Reduce copying into .apps and .test in makevms.com
1492 Don't try to use blank CA certificate in CA.com
1493 Allow use of C files from original directories in maketests.com
1494 [Steven M. Schweda" <sms@antinode.info>]
1496 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1498 *) When rejecting SSL/TLS records due to an incorrect version number, never
1499 update s->server with a new major version number. As of
1500 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1501 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1502 the previous behavior could result in a read attempt at NULL when
1503 receiving specific incorrect SSL/TLS records once record payload
1504 protection is active. (CVE-2010-0740)
1505 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1507 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1508 could be crashed if the relevant tables were not present (e.g. chrooted).
1509 [Tomas Hoger <thoger@redhat.com>]
1511 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1513 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1514 [Martin Olsson, Neel Mehta]
1516 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1517 accommodate for stack sorting, always a write lock!).
1520 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1521 excessive delays in the RAND_poll(): over a minute. As a workaround
1522 include a time check in the inner Heap32Next loop too.
1525 *) The code that handled flushing of data in SSL/TLS originally used the
1526 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1527 the problem outlined in PR#1949. The fix suggested there however can
1528 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1529 of Apache). So instead simplify the code to flush unconditionally.
1530 This should be fine since flushing with no data to flush is a no op.
1533 *) Handle TLS versions 2.0 and later properly and correctly use the
1534 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1535 off ancient servers have a habit of sticking around for a while...
1538 *) Modify compression code so it frees up structures without using the
1539 ex_data callbacks. This works around a problem where some applications
1540 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1541 restarting) then use compression (e.g. SSL with compression) later.
1542 This results in significant per-connection memory leaks and
1543 has caused some security issues including CVE-2008-1678 and
1547 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1548 change when encrypting or decrypting.
1551 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1552 connect and renegotiate with servers which do not support RI.
1553 Until RI is more widely deployed this option is enabled by default.
1556 *) Add "missing" ssl ctrls to clear options and mode.
1559 *) If client attempts to renegotiate and doesn't support RI respond with
1560 a no_renegotiation alert as required by RFC5746. Some renegotiating
1561 TLS clients will continue a connection gracefully when they receive
1562 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1563 waiting for a server hello which it will never receive. Now we treat a
1564 received no_renegotiation alert as a fatal error. This is because
1565 applications requesting a renegotiation might well expect it to succeed
1566 and would have no code in place to handle the server denying it so the
1567 only safe thing to do is to terminate the connection.
1570 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1571 peer supports secure renegotiation and 0 otherwise. Print out peer
1572 renegotiation support in s_client/s_server.
1575 *) Replace the highly broken and deprecated SPKAC certification method with
1576 the updated NID creation version. This should correctly handle UTF8.
1579 *) Implement RFC5746. Re-enable renegotiation but require the extension
1580 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1581 turns out to be a bad idea. It has been replaced by
1582 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1583 SSL_CTX_set_options(). This is really not recommended unless you
1584 know what you are doing.
1585 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1587 *) Fixes to stateless session resumption handling. Use initial_ctx when
1588 issuing and attempting to decrypt tickets in case it has changed during
1589 servername handling. Use a non-zero length session ID when attempting
1590 stateless session resumption: this makes it possible to determine if
1591 a resumption has occurred immediately after receiving server hello
1592 (several places in OpenSSL subtly assume this) instead of later in
1596 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1597 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1598 fixes for a few places where the return code is not checked
1600 [Julia Lawall <julia@diku.dk>]
1602 *) Add --strict-warnings option to Configure script to include devteam
1603 warnings in other configurations.
1606 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1607 makes it possible to install openssl libraries in locations which
1608 have names other than "lib", for example "/usr/lib64" which some
1610 [Steve Henson, based on patch from Jeremy Utley]
1612 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1613 X690 8.9.12 and can produce some misleading textual output of OIDs.
1614 [Steve Henson, reported by Dan Kaminsky]
1616 *) Delete MD2 from algorithm tables. This follows the recommendation in
1617 several standards that it is not used in new applications due to
1618 several cryptographic weaknesses. For binary compatibility reasons
1619 the MD2 API is still compiled in by default.
1622 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1626 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1627 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1629 [Guenter <lists@gknw.net>]
1631 *) Fix the server certificate chain building code to use X509_verify_cert(),
1632 it used to have an ad-hoc builder which was unable to cope with anything
1633 other than a simple chain.
1634 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1636 *) Don't check self signed certificate signatures in X509_verify_cert()
1637 by default (a flag can override this): it just wastes time without
1638 adding any security. As a useful side effect self signed root CAs
1639 with non-FIPS digests are now usable in FIPS mode.
1642 *) In dtls1_process_out_of_seq_message() the check if the current message
1643 is already buffered was missing. For every new message was memory
1644 allocated, allowing an attacker to perform an denial of service attack
1645 with sending out of seq handshake messages until there is no memory
1646 left. Additionally every future messege was buffered, even if the
1647 sequence number made no sense and would be part of another handshake.
1648 So only messages with sequence numbers less than 10 in advance will be
1649 buffered. (CVE-2009-1378)
1650 [Robin Seggelmann, discovered by Daniel Mentz]
1652 *) Records are buffered if they arrive with a future epoch to be
1653 processed after finishing the corresponding handshake. There is
1654 currently no limitation to this buffer allowing an attacker to perform
1655 a DOS attack with sending records with future epochs until there is no
1656 memory left. This patch adds the pqueue_size() function to detemine
1657 the size of a buffer and limits the record buffer to 100 entries.
1659 [Robin Seggelmann, discovered by Daniel Mentz]
1661 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1662 parent structure is freed. (CVE-2009-1379)
1665 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1666 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1669 [Ilya O. <vrghost@gmail.com>]
1671 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1673 *) Disable renegotiation completely - this fixes a severe security
1674 problem (CVE-2009-3555) at the cost of breaking all
1675 renegotiation. Renegotiation can be re-enabled by setting
1676 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1677 run-time. This is really not recommended unless you know what
1681 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1683 *) Don't set val to NULL when freeing up structures, it is freed up by
1684 underlying code. If sizeof(void *) > sizeof(long) this can result in
1685 zeroing past the valid field. (CVE-2009-0789)
1686 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1688 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1689 checked correctly. This would allow some invalid signed attributes to
1690 appear to verify correctly. (CVE-2009-0591)
1691 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1693 *) Reject UniversalString and BMPString types with invalid lengths. This
1694 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1695 a legal length. (CVE-2009-0590)
1698 *) Set S/MIME signing as the default purpose rather than setting it
1699 unconditionally. This allows applications to override it at the store
1703 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1704 to handle some structures.
1707 *) Improve efficiency of mem_gets: don't search whole buffer each time
1709 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1711 *) New -hex option for openssl rand.
1714 *) Print out UTF8String and NumericString when parsing ASN1.
1717 *) Support NumericString type for name components.
1720 *) Allow CC in the environment to override the automatically chosen
1721 compiler. Note that nothing is done to ensure flags work with the
1725 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1727 *) Properly check EVP_VerifyFinal() and similar return values
1729 [Ben Laurie, Bodo Moeller, Google Security Team]
1731 *) Enable TLS extensions by default.
1734 *) Allow the CHIL engine to be loaded, whether the application is
1735 multithreaded or not. (This does not release the developer from the
1736 obligation to set up the dynamic locking callbacks.)
1737 [Sander Temme <sander@temme.net>]
1739 *) Use correct exit code if there is an error in dgst command.
1740 [Steve Henson; problem pointed out by Roland Dirlewanger]
1742 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1743 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1746 *) Add experimental JPAKE support, including demo authentication in
1747 s_client and s_server.
1750 *) Set the comparison function in v3_addr_canonize().
1751 [Rob Austein <sra@hactrn.net>]
1753 *) Add support for XMPP STARTTLS in s_client.
1754 [Philip Paeps <philip@freebsd.org>]
1756 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1757 to ensure that even with this option, only ciphersuites in the
1758 server's preference list will be accepted. (Note that the option
1759 applies only when resuming a session, so the earlier behavior was
1760 just about the algorithm choice for symmetric cryptography.)
1763 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1765 *) Fix NULL pointer dereference if a DTLS server received
1766 ChangeCipherSpec as first record (CVE-2009-1386).
1769 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1770 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1773 *) The fix in 0.9.8c that supposedly got rid of unsafe
1774 double-checked locking was incomplete for RSA blinding,
1775 addressing just one layer of what turns out to have been
1776 doubly unsafe triple-checked locking.
1778 So now fix this for real by retiring the MONT_HELPER macro
1779 in crypto/rsa/rsa_eay.c.
1781 [Bodo Moeller; problem pointed out by Marius Schilder]
1783 *) Various precautionary measures:
1785 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1787 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1788 (NB: This would require knowledge of the secret session ticket key
1789 to exploit, in which case you'd be SOL either way.)
1791 - Change bn_nist.c so that it will properly handle input BIGNUMs
1792 outside the expected range.
1794 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1797 [Neel Mehta, Bodo Moeller]
1799 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1800 the load fails. Useful for distros.
1801 [Ben Laurie and the FreeBSD team]
1803 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1806 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1809 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1811 This work was sponsored by Logica.
1814 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1815 keystores. Support for SSL/TLS client authentication too.
1816 Not compiled unless enable-capieng specified to Configure.
1818 This work was sponsored by Logica.
1821 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1822 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1823 attribute creation routines such as certifcate requests and PKCS#12
1827 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1829 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1830 handshake which could lead to a cilent crash as found using the
1831 Codenomicon TLS test suite (CVE-2008-1672)
1832 [Steve Henson, Mark Cox]
1834 *) Fix double free in TLS server name extensions which could lead to
1835 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1838 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1840 Clear the error queue to ensure that error entries left from
1841 older function calls do not interfere with the correct operation.
1842 [Lutz Jaenicke, Erik de Castro Lopo]
1844 *) Remove root CA certificates of commercial CAs:
1846 The OpenSSL project does not recommend any specific CA and does not
1847 have any policy with respect to including or excluding any CA.
1848 Therefore it does not make any sense to ship an arbitrary selection
1849 of root CA certificates with the OpenSSL software.
1852 *) RSA OAEP patches to fix two separate invalid memory reads.
1853 The first one involves inputs when 'lzero' is greater than
1854 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1855 before the beginning of from). The second one involves inputs where
1856 the 'db' section contains nothing but zeroes (there is a one-byte
1857 invalid read after the end of 'db').
1858 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1860 *) Partial backport from 0.9.9-dev:
1862 Introduce bn_mul_mont (dedicated Montgomery multiplication
1863 procedure) as a candidate for BIGNUM assembler implementation.
1864 While 0.9.9-dev uses assembler for various architectures, only
1865 x86_64 is available by default here in the 0.9.8 branch, and
1866 32-bit x86 is available through a compile-time setting.
1868 To try the 32-bit x86 assembler implementation, use Configure
1869 option "enable-montasm" (which exists only for this backport).
1871 As "enable-montasm" for 32-bit x86 disclaims code stability
1872 anyway, in this constellation we activate additional code
1873 backported from 0.9.9-dev for further performance improvements,
1874 namely BN_from_montgomery_word. (To enable this otherwise,
1875 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1877 [Andy Polyakov (backport partially by Bodo Moeller)]
1879 *) Add TLS session ticket callback. This allows an application to set
1880 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1881 values. This is useful for key rollover for example where several key
1882 sets may exist with different names.
1885 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1886 This was broken until now in 0.9.8 releases, such that the only way
1887 a registered ENGINE could be used (assuming it initialises
1888 successfully on the host) was to explicitly set it as the default
1889 for the relevant algorithms. This is in contradiction with 0.9.7
1890 behaviour and the documentation. With this fix, when an ENGINE is
1891 registered into a given algorithm's table of implementations, the
1892 'uptodate' flag is reset so that auto-discovery will be used next
1893 time a new context for that algorithm attempts to select an
1895 [Ian Lister (tweaked by Geoff Thorpe)]
1897 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1898 implemention in the following ways:
1900 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1903 Lack of BER streaming support means one pass streaming processing is
1904 only supported if data is detached: setting the streaming flag is
1905 ignored for embedded content.
1907 CMS support is disabled by default and must be explicitly enabled
1908 with the enable-cms configuration option.
1911 *) Update the GMP engine glue to do direct copies between BIGNUM and
1912 mpz_t when openssl and GMP use the same limb size. Otherwise the
1913 existing "conversion via a text string export" trick is still used.
1914 [Paul Sheer <paulsheer@gmail.com>]
1916 *) Zlib compression BIO. This is a filter BIO which compressed and
1917 uncompresses any data passed through it.
1920 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1921 RFC3394 compatible AES key wrapping.
1924 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1925 sets string data without copying. X509_ALGOR_set0() and
1926 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1927 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1928 from an X509_ATTRIBUTE structure optionally checking it occurs only
1929 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1933 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1934 to get the expected BN_FLG_CONSTTIME behavior.
1935 [Bodo Moeller (Google)]
1939 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1940 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1941 - added some more tests to do_tests.pl
1942 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1943 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1944 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1945 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1946 - various changes to netware.pl to enable gcc-cross builds on Win32
1948 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1949 - various changes to fix missing prototype warnings
1950 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1951 - added AES, WHIRLPOOL and CPUID assembler code to build files
1952 - added missing AES assembler make rules to mk1mf.pl
1953 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1954 [Guenter Knauf <eflash@gmx.net>]
1956 *) Implement certificate status request TLS extension defined in RFC3546.
1957 A client can set the appropriate parameters and receive the encoded
1958 OCSP response via a callback. A server can query the supplied parameters
1959 and set the encoded OCSP response in the callback. Add simplified examples
1960 to s_client and s_server.
1963 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1965 *) Fix various bugs:
1966 + Binary incompatibility of ssl_ctx_st structure
1967 + DTLS interoperation with non-compliant servers
1968 + Don't call get_session_cb() without proposed session
1969 + Fix ia64 assembler code
1970 [Andy Polyakov, Steve Henson]
1972 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1974 *) DTLS Handshake overhaul. There were longstanding issues with
1975 OpenSSL DTLS implementation, which were making it impossible for
1976 RFC 4347 compliant client to communicate with OpenSSL server.
1977 Unfortunately just fixing these incompatibilities would "cut off"
1978 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1979 server keeps tolerating non RFC compliant syntax. The opposite is
1980 not true, 0.9.8f client can not communicate with earlier server.
1981 This update even addresses CVE-2007-4995.
1984 *) Changes to avoid need for function casts in OpenSSL: some compilers
1985 (gcc 4.2 and later) reject their use.
1986 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1989 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1990 RFC4507bis. The encrypted ticket format is an encrypted encoded
1991 SSL_SESSION structure, that way new session features are automatically
1994 If a client application caches session in an SSL_SESSION structure
1995 support is transparent because tickets are now stored in the encoded
1998 The SSL_CTX structure automatically generates keys for ticket
1999 protection in servers so again support should be possible
2000 with no application modification.
2002 If a client or server wishes to disable RFC4507 support then the option
2003 SSL_OP_NO_TICKET can be set.
2005 Add a TLS extension debugging callback to allow the contents of any client
2006 or server extensions to be examined.
2008 This work was sponsored by Google.
2011 *) Add initial support for TLS extensions, specifically for the server_name
2012 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2013 have new members for a host name. The SSL data structure has an
2014 additional member SSL_CTX *initial_ctx so that new sessions can be
2015 stored in that context to allow for session resumption, even after the
2016 SSL has been switched to a new SSL_CTX in reaction to a client's
2017 server_name extension.
2019 New functions (subject to change):
2021 SSL_get_servername()
2022 SSL_get_servername_type()
2025 New CTRL codes and macros (subject to change):
2027 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2028 - SSL_CTX_set_tlsext_servername_callback()
2029 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2030 - SSL_CTX_set_tlsext_servername_arg()
2031 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2033 openssl s_client has a new '-servername ...' option.
2035 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2036 '-key2 ...', '-servername_fatal' (subject to change). This allows
2037 testing the HostName extension for a specific single host name ('-cert'
2038 and '-key' remain fallbacks for handshakes without HostName
2039 negotiation). If the unrecogninzed_name alert has to be sent, this by
2040 default is a warning; it becomes fatal with the '-servername_fatal'
2043 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2045 *) Add AES and SSE2 assembly language support to VC++ build.
2048 *) Mitigate attack on final subtraction in Montgomery reduction.
2051 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2052 (which previously caused an internal error).
2055 *) Squeeze another 10% out of IGE mode when in != out.
2058 *) AES IGE mode speedup.
2059 [Dean Gaudet (Google)]
2061 *) Add the Korean symmetric 128-bit cipher SEED (see
2062 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2063 add SEED ciphersuites from RFC 4162:
2065 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2066 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2067 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2068 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2070 To minimize changes between patchlevels in the OpenSSL 0.9.8
2071 series, SEED remains excluded from compilation unless OpenSSL
2072 is configured with 'enable-seed'.
2073 [KISA, Bodo Moeller]
2075 *) Mitigate branch prediction attacks, which can be practical if a
2076 single processor is shared, allowing a spy process to extract
2077 information. For detailed background information, see
2078 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2079 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2080 and Necessary Software Countermeasures"). The core of the change
2081 are new versions BN_div_no_branch() and
2082 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2083 respectively, which are slower, but avoid the security-relevant
2084 conditional branches. These are automatically called by BN_div()
2085 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2086 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2087 remove a conditional branch.
2089 BN_FLG_CONSTTIME is the new name for the previous
2090 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2091 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2092 in the exponent causes BN_mod_exp_mont() to use the alternative
2093 implementation in BN_mod_exp_mont_consttime().) The old name
2094 remains as a deprecated alias.
2096 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2097 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2098 constant-time implementations for more than just exponentiation.
2099 Here too the old name is kept as a deprecated alias.
2101 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2102 the BN_BLINDING structure gets an independent copy of the
2103 modulus. This means that the previous "BIGNUM *m" argument to
2104 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2105 essentially becomes "const BIGNUM *m", although we can't actually
2106 change this in the header file before 0.9.9. It allows
2107 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2108 enable BN_FLG_CONSTTIME.
2110 [Matthew D Wood (Intel Corp)]
2112 *) In the SSL/TLS server implementation, be strict about session ID
2113 context matching (which matters if an application uses a single
2114 external cache for different purposes). Previously,
2115 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2116 set. This did ensure strict client verification, but meant that,
2117 with applications using a single external cache for quite
2118 different requirements, clients could circumvent ciphersuite
2119 restrictions for a given session ID context by starting a session
2120 in a different context.
2123 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2124 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2125 authentication-only ciphersuites.
2128 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2129 not complete and could lead to a possible single byte overflow
2130 (CVE-2007-5135) [Ben Laurie]
2132 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2134 *) Since AES128 and AES256 (and similarly Camellia128 and
2135 Camellia256) share a single mask bit in the logic of
2136 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2137 kludge to work properly if AES128 is available and AES256 isn't
2138 (or if Camellia128 is available and Camellia256 isn't).
2141 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2142 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2143 When a point or a seed is encoded in a BIT STRING, we need to
2144 prevent the removal of trailing zero bits to get the proper DER
2145 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2146 of a NamedBitList, for which trailing 0 bits need to be removed.)
2149 *) Have SSL/TLS server implementation tolerate "mismatched" record
2150 protocol version while receiving ClientHello even if the
2151 ClientHello is fragmented. (The server can't insist on the
2152 particular protocol version it has chosen before the ServerHello
2153 message has informed the client about his choice.)
2156 *) Add RFC 3779 support.
2157 [Rob Austein for ARIN, Ben Laurie]
2159 *) Load error codes if they are not already present instead of using a
2160 static variable. This allows them to be cleanly unloaded and reloaded.
2161 Improve header file function name parsing.
2164 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2165 or CAPABILITY handshake as required by RFCs.
2168 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2170 *) Introduce limits to prevent malicious keys being able to
2171 cause a denial of service. (CVE-2006-2940)
2172 [Steve Henson, Bodo Moeller]
2174 *) Fix ASN.1 parsing of certain invalid structures that can result
2175 in a denial of service. (CVE-2006-2937) [Steve Henson]
2177 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2178 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2180 *) Fix SSL client code which could crash if connecting to a
2181 malicious SSLv2 server. (CVE-2006-4343)
2182 [Tavis Ormandy and Will Drewry, Google Security Team]
2184 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2185 match only those. Before that, "AES256-SHA" would be interpreted
2186 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2187 the same strength classification in 0.9.7h) as we currently only
2188 have a single AES bit in the ciphersuite description bitmap.
2189 That change, however, also applied to ciphersuite strings such as
2190 "RC4-MD5" that intentionally matched multiple ciphersuites --
2191 namely, SSL 2.0 ciphersuites in addition to the more common ones
2192 from SSL 3.0/TLS 1.0.
2194 So we change the selection algorithm again: Naming an explicit
2195 ciphersuite selects this one ciphersuite, and any other similar
2196 ciphersuite (same bitmap) from *other* protocol versions.
2197 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2198 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2200 Since SSL 2.0 does not have any ciphersuites for which the
2201 128/256 bit distinction would be relevant, this works for now.
2202 The proper fix will be to use different bits for AES128 and
2203 AES256, which would have avoided the problems from the beginning;
2204 however, bits are scarce, so we can only do this in a new release
2205 (not just a patchlevel) when we can change the SSL_CIPHER
2206 definition to split the single 'unsigned long mask' bitmap into
2207 multiple values to extend the available space.
2211 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2213 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2214 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2216 *) Add AES IGE and biIGE modes.
2219 *) Change the Unix randomness entropy gathering to use poll() when
2220 possible instead of select(), since the latter has some
2221 undesirable limitations.
2222 [Darryl Miles via Richard Levitte and Bodo Moeller]
2224 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2225 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2226 cannot be implicitly activated as part of, e.g., the "AES" alias.
2227 However, please upgrade to OpenSSL 0.9.9[-dev] for
2228 non-experimental use of the ECC ciphersuites to get TLS extension
2229 support, which is required for curve and point format negotiation
2230 to avoid potential handshake problems.
2233 *) Disable rogue ciphersuites:
2235 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2236 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2237 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2239 The latter two were purportedly from
2240 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2243 Also deactivate the remaining ciphersuites from
2244 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2245 unofficial, and the ID has long expired.
2248 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2249 dual-core machines) and other potential thread-safety issues.
2252 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2253 versions), which is now available for royalty-free use
2254 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2255 Also, add Camellia TLS ciphersuites from RFC 4132.
2257 To minimize changes between patchlevels in the OpenSSL 0.9.8
2258 series, Camellia remains excluded from compilation unless OpenSSL
2259 is configured with 'enable-camellia'.
2262 *) Disable the padding bug check when compression is in use. The padding
2263 bug check assumes the first packet is of even length, this is not
2264 necessarily true if compresssion is enabled and can result in false
2265 positives causing handshake failure. The actual bug test is ancient
2266 code so it is hoped that implementations will either have fixed it by
2267 now or any which still have the bug do not support compression.
2270 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2272 *) When applying a cipher rule check to see if string match is an explicit
2273 cipher suite and only match that one cipher suite if it is.
2276 *) Link in manifests for VC++ if needed.
2277 [Austin Ziegler <halostatue@gmail.com>]
2279 *) Update support for ECC-based TLS ciphersuites according to
2280 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2281 TLS extensions, which are supported starting with the 0.9.9
2282 branch, not in the OpenSSL 0.9.8 branch).
2285 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2286 opaque EVP_CIPHER_CTX handling.
2289 *) Fixes and enhancements to zlib compression code. We now only use
2290 "zlib1.dll" and use the default __cdecl calling convention on Win32
2291 to conform with the standards mentioned here:
2292 http://www.zlib.net/DLL_FAQ.txt
2293 Static zlib linking now works on Windows and the new --with-zlib-include
2294 --with-zlib-lib options to Configure can be used to supply the location
2295 of the headers and library. Gracefully handle case where zlib library
2299 *) Several fixes and enhancements to the OID generation code. The old code
2300 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2301 handle numbers larger than ULONG_MAX, truncated printing and had a
2302 non standard OBJ_obj2txt() behaviour.
2305 *) Add support for building of engines under engine/ as shared libraries
2306 under VC++ build system.
2309 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2310 Hopefully, we will not see any false combination of paths any more.
2313 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2315 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2316 (part of SSL_OP_ALL). This option used to disable the
2317 countermeasure against man-in-the-middle protocol-version
2318 rollback in the SSL 2.0 server implementation, which is a bad
2319 idea. (CVE-2005-2969)
2321 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2322 for Information Security, National Institute of Advanced Industrial
2323 Science and Technology [AIST], Japan)]
2325 *) Add two function to clear and return the verify parameter flags.
2328 *) Keep cipherlists sorted in the source instead of sorting them at
2329 runtime, thus removing the need for a lock.
2332 *) Avoid some small subgroup attacks in Diffie-Hellman.
2333 [Nick Mathewson and Ben Laurie]
2335 *) Add functions for well-known primes.
2338 *) Extended Windows CE support.
2339 [Satoshi Nakamura and Andy Polyakov]
2341 *) Initialize SSL_METHOD structures at compile time instead of during
2342 runtime, thus removing the need for a lock.
2345 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2346 attempting to decrypt each encrypted key in turn. Add support to
2350 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2352 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2355 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2358 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2359 key into the same file any more.
2362 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2365 *) Add -utf8 command line and config file option to 'ca'.
2366 [Stefan <stf@udoma.org]
2368 *) Removed the macro des_crypt(), as it seems to conflict with some
2369 libraries. Use DES_crypt().
2372 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2373 involves renaming the source and generated shared-libs for
2374 both. The engines will accept the corrected or legacy ids
2375 ('ncipher' and '4758_cca' respectively) when binding. NB,
2376 this only applies when building 'shared'.
2377 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2379 *) Add attribute functions to EVP_PKEY structure. Modify
2380 PKCS12_create() to recognize a CSP name attribute and
2381 use it. Make -CSP option work again in pkcs12 utility.
2384 *) Add new functionality to the bn blinding code:
2385 - automatic re-creation of the BN_BLINDING parameters after
2386 a fixed number of uses (currently 32)
2387 - add new function for parameter creation
2388 - introduce flags to control the update behaviour of the
2389 BN_BLINDING parameters
2390 - hide BN_BLINDING structure
2391 Add a second BN_BLINDING slot to the RSA structure to improve
2392 performance when a single RSA object is shared among several
2396 *) Add support for DTLS.
2397 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2399 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2400 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2403 *) Remove buggy and incompletet DH cert support from
2404 ssl/ssl_rsa.c and ssl/s3_both.c
2407 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2408 the apps/openssl applications.
2411 *) Compile clean with "-Wall -Wmissing-prototypes
2412 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2413 DEBUG_SAFESTACK must also be set.
2416 *) Change ./Configure so that certain algorithms can be disabled by default.
2417 The new counterpiece to "no-xxx" is "enable-xxx".
2419 The patented RC5 and MDC2 algorithms will now be disabled unless
2420 "enable-rc5" and "enable-mdc2", respectively, are specified.
2422 (IDEA remains enabled despite being patented. This is because IDEA
2423 is frequently required for interoperability, and there is no license
2424 fee for non-commercial use. As before, "no-idea" can be used to
2425 avoid this algorithm.)
2429 *) Add processing of proxy certificates (see RFC 3820). This work was
2430 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2431 EGEE (Enabling Grids for E-science in Europe).
2434 *) RC4 performance overhaul on modern architectures/implementations, such
2435 as Intel P4, IA-64 and AMD64.
2438 *) New utility extract-section.pl. This can be used specify an alternative
2439 section number in a pod file instead of having to treat each file as
2440 a separate case in Makefile. This can be done by adding two lines to the
2443 =for comment openssl_section:XXX
2445 The blank line is mandatory.
2449 *) New arguments -certform, -keyform and -pass for s_client and s_server
2450 to allow alternative format key and certificate files and passphrase
2454 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2455 update associated structures and add various utility functions.
2457 Add new policy related verify parameters, include policy checking in
2458 standard verify code. Enhance 'smime' application with extra parameters
2459 to support policy checking and print out.
2462 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2463 Nehemiah processors. These extensions support AES encryption in hardware
2464 as well as RNG (though RNG support is currently disabled).
2465 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2467 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2470 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2471 [Andy Polyakov and a number of other people]
2473 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2474 implementation contributed by IBM.
2475 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2477 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2478 exponent rather than 'unsigned long'. There is a corresponding change to
2479 the new 'rsa_keygen' element of the RSA_METHOD structure.
2480 [Jelte Jansen, Geoff Thorpe]
2482 *) Functionality for creating the initial serial number file is now
2483 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2485 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2486 number file to 1, which is bound to cause problems. To avoid
2487 the problems while respecting compatibility between different 0.9.7
2488 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2489 CA.pl for serial number initialization. With the new release 0.9.8,
2490 we can fix the problem directly in the 'ca' utility.)
2493 *) Reduced header interdepencies by declaring more opaque objects in
2494 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2495 give fewer recursive includes, which could break lazy source code - so
2496 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2497 developers should define this symbol when building and using openssl to
2498 ensure they track the recommended behaviour, interfaces, [etc], but
2499 backwards-compatible behaviour prevails when this isn't defined.
2502 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2505 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2506 This will generate a random key of the appropriate length based on the
2507 cipher context. The EVP_CIPHER can provide its own random key generation
2508 routine to support keys of a specific form. This is used in the des and
2509 3des routines to generate a key of the correct parity. Update S/MIME
2510 code to use new functions and hence generate correct parity DES keys.
2511 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2512 valid (weak or incorrect parity).
2515 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2516 as looking them up. This is useful when the verified structure may contain
2517 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2518 present unless the new PKCS7_NO_CRL flag is asserted.
2521 *) Extend ASN1 oid configuration module. It now additionally accepts the
2524 shortName = some long name, 1.2.3.4
2527 *) Reimplemented the BN_CTX implementation. There is now no more static
2528 limitation on the number of variables it can handle nor the depth of the
2529 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2530 information can now expand as required, and rather than having a single
2531 static array of bignums, BN_CTX now uses a linked-list of such arrays
2532 allowing it to expand on demand whilst maintaining the usefulness of
2533 BN_CTX's "bundling".
2536 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2537 to allow all RSA operations to function using a single BN_CTX.
2540 *) Preliminary support for certificate policy evaluation and checking. This
2541 is initially intended to pass the tests outlined in "Conformance Testing
2542 of Relying Party Client Certificate Path Processing Logic" v1.07.
2545 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2546 remained unused and not that useful. A variety of other little bignum
2547 tweaks and fixes have also been made continuing on from the audit (see
2551 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2552 associated ASN1, EVP and SSL functions and old ASN1 macros.
2555 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2556 and this should never fail. So the return value from the use of
2557 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2558 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2561 *) BN_CTX_get() should return zero-valued bignums, providing the same
2562 initialised value as BN_new().
2563 [Geoff Thorpe, suggested by Ulf Möller]
2565 *) Support for inhibitAnyPolicy certificate extension.
2568 *) An audit of the BIGNUM code is underway, for which debugging code is
2569 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2570 is considered valid when processing BIGNUMs, and causes execution to
2571 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2572 further steps are taken to deliberately pollute unused data in BIGNUM
2573 structures to try and expose faulty code further on. For now, openssl will
2574 (in its default mode of operation) continue to tolerate the inconsistent
2575 forms that it has tolerated in the past, but authors and packagers should
2576 consider trying openssl and their own applications when compiled with
2577 these debugging symbols defined. It will help highlight potential bugs in
2578 their own code, and will improve the test coverage for OpenSSL itself. At
2579 some point, these tighter rules will become openssl's default to improve
2580 maintainability, though the assert()s and other overheads will remain only
2581 in debugging configurations. See bn.h for more details.
2582 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2584 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2585 that can only be obtained through BN_CTX_new() (which implicitly
2586 initialises it). The presence of this function only made it possible
2587 to overwrite an existing structure (and cause memory leaks).
2590 *) Because of the callback-based approach for implementing LHASH as a
2591 template type, lh_insert() adds opaque objects to hash-tables and
2592 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2593 to clean up those corresponding objects before destroying the hash table
2594 (and losing the object pointers). So some over-zealous constifications in
2595 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2596 objects as "const" and the lh_doall[_arg] callback wrappers are not
2597 prototyped to have "const" restrictions on the object pointers they are
2598 given (and so aren't required to cast them away any more).
2601 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2602 (speed) prefers to use its own implementation. The two implementations
2603 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2604 its object type properly exposed (MS_TM) instead of casting to/from "char
2605 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2606 aren't necessarily the greatest nomenclatures - but this is what was used
2607 internally to the implementation so I've used that for now.
2610 *) Ensure that deprecated functions do not get compiled when
2611 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2612 the self-tests were still using deprecated key-generation functions so
2613 these have been updated also.
2616 *) Reorganise PKCS#7 code to separate the digest location functionality
2617 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2618 New function PKCS7_set_digest() to set the digest type for PKCS#7
2619 digestedData type. Add additional code to correctly generate the
2620 digestedData type and add support for this type in PKCS7 initialization
2624 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2625 structure of type "other".
2628 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2629 sure the loop does correctly stop and breaking ("division by zero")
2630 modulus operations are not performed. The (pre-generated) prime
2631 table crypto/bn/bn_prime.h was already correct, but it could not be
2632 re-generated on some platforms because of the "division by zero"
2633 situation in the script.
2634 [Ralf S. Engelschall]
2636 *) Update support for ECC-based TLS ciphersuites according to
2637 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2638 SHA-1 now is only used for "small" curves (where the
2639 representation of a field element takes up to 24 bytes); for
2640 larger curves, the field element resulting from ECDH is directly
2641 used as premaster secret.
2642 [Douglas Stebila (Sun Microsystems Laboratories)]
2644 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2645 curve secp160r1 to the tests.
2646 [Douglas Stebila (Sun Microsystems Laboratories)]
2648 *) Add the possibility to load symbols globally with DSO.
2649 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2651 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2652 control of the error stack.
2655 *) Add support for STORE in ENGINE.
2658 *) Add the STORE type. The intention is to provide a common interface
2659 to certificate and key stores, be they simple file-based stores, or
2660 HSM-type store, or LDAP stores, or...
2661 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2664 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2665 pass a list of arguments to any function as well as provide a way
2666 for a function to pass data back to the caller.
2669 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2670 works like BUF_strdup() but can be used to duplicate a portion of
2671 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2675 *) Add the function sk_find_ex() which works like sk_find(), but will
2676 return an index to an element even if an exact match couldn't be
2677 found. The index is guaranteed to point at the element where the
2678 searched-for key would be inserted to preserve sorting order.
2681 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2682 takes an extra flags argument for optional functionality. Currently,
2683 the following flags are defined:
2685 OBJ_BSEARCH_VALUE_ON_NOMATCH
2686 This one gets OBJ_bsearch_ex() to return a pointer to the first
2687 element where the comparing function returns a negative or zero
2690 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2691 This one gets OBJ_bsearch_ex() to return a pointer to the first
2692 element where the comparing function returns zero. This is useful
2693 if there are more than one element where the comparing function
2697 *) Make it possible to create self-signed certificates with 'openssl ca'
2698 in such a way that the self-signed certificate becomes part of the
2699 CA database and uses the same mechanisms for serial number generation
2700 as all other certificate signing. The new flag '-selfsign' enables
2701 this functionality. Adapt CA.sh and CA.pl.in.
2704 *) Add functionality to check the public key of a certificate request
2705 against a given private. This is useful to check that a certificate
2706 request can be signed by that key (self-signing).
2709 *) Make it possible to have multiple active certificates with the same
2710 subject in the CA index file. This is done only if the keyword
2711 'unique_subject' is set to 'no' in the main CA section (default
2712 if 'CA_default') of the configuration file. The value is saved
2713 with the database itself in a separate index attribute file,
2714 named like the index file with '.attr' appended to the name.
2717 *) Generate muti valued AVAs using '+' notation in config files for
2721 *) Support for nameConstraints certificate extension.
2724 *) Support for policyConstraints certificate extension.
2727 *) Support for policyMappings certificate extension.
2730 *) Make sure the default DSA_METHOD implementation only uses its
2731 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2732 and change its own handlers to be NULL so as to remove unnecessary
2733 indirection. This lets alternative implementations fallback to the
2734 default implementation more easily.
2737 *) Support for directoryName in GeneralName related extensions
2741 *) Make it possible to link applications using Makefile.shared.
2742 Make that possible even when linking against static libraries!
2745 *) Support for single pass processing for S/MIME signing. This now
2746 means that S/MIME signing can be done from a pipe, in addition
2747 cleartext signing (multipart/signed type) is effectively streaming
2748 and the signed data does not need to be all held in memory.
2750 This is done with a new flag PKCS7_STREAM. When this flag is set
2751 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2752 is done after the data is output (and digests calculated) in
2753 SMIME_write_PKCS7().
2756 *) Add full support for -rpath/-R, both in shared libraries and
2757 applications, at least on the platforms where it's known how
2761 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2762 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2763 will now compute a table of multiples of the generator that
2764 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2765 faster (notably in the case of a single point multiplication,
2766 scalar * generator).
2767 [Nils Larsch, Bodo Moeller]
2769 *) IPv6 support for certificate extensions. The various extensions
2770 which use the IP:a.b.c.d can now take IPv6 addresses using the
2771 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2775 *) Added an ENGINE that implements RSA by performing private key
2776 exponentiations with the GMP library. The conversions to and from
2777 GMP's mpz_t format aren't optimised nor are any montgomery forms
2778 cached, and on x86 it appears OpenSSL's own performance has caught up.
2779 However there are likely to be other architectures where GMP could
2780 provide a boost. This ENGINE is not built in by default, but it can be
2781 specified at Configure time and should be accompanied by the necessary
2782 linker additions, eg;
2783 ./config -DOPENSSL_USE_GMP -lgmp
2786 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2787 testing availability of engines with "-t" - the old behaviour is
2788 produced by increasing the feature's verbosity with "-tt".
2791 *) ECDSA routines: under certain error conditions uninitialized BN objects
2792 could be freed. Solution: make sure initialization is performed early
2793 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2797 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2798 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2799 software implementations. For DSA and DH, parameter generation can
2800 also be overriden by providing the appropriate method callbacks.
2803 *) Change the "progress" mechanism used in key-generation and
2804 primality testing to functions that take a new BN_GENCB pointer in
2805 place of callback/argument pairs. The new API functions have "_ex"
2806 postfixes and the older functions are reimplemented as wrappers for
2807 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2808 declarations of the old functions to help (graceful) attempts to
2809 migrate to the new functions. Also, the new key-generation API
2810 functions operate on a caller-supplied key-structure and return
2811 success/failure rather than returning a key or NULL - this is to
2812 help make "keygen" another member function of RSA_METHOD etc.
2814 Example for using the new callback interface:
2816 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2820 BN_GENCB_set(&my_cb, my_callback, my_arg);
2822 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2823 /* For the meaning of a, b in calls to my_callback(), see the
2824 * documentation of the function that calls the callback.
2825 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2826 * my_callback should return 1 if it wants BN_is_prime_ex()
2827 * to continue, or 0 to stop.
2832 *) Change the ZLIB compression method to be stateful, and make it
2833 available to TLS with the number defined in
2834 draft-ietf-tls-compression-04.txt.
2837 *) Add the ASN.1 structures and functions for CertificatePair, which
2838 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2840 CertificatePair ::= SEQUENCE {
2841 forward [0] Certificate OPTIONAL,
2842 reverse [1] Certificate OPTIONAL,
2843 -- at least one of the pair shall be present -- }
2845 Also implement the PEM functions to read and write certificate
2846 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2848 This needed to be defined, mostly for the sake of the LDAP
2849 attribute crossCertificatePair, but may prove useful elsewhere as
2853 *) Make it possible to inhibit symlinking of shared libraries in
2854 Makefile.shared, for Cygwin's sake.
2857 *) Extend the BIGNUM API by creating a function
2858 void BN_set_negative(BIGNUM *a, int neg);
2859 and a macro that behave like
2860 int BN_is_negative(const BIGNUM *a);
2862 to avoid the need to access 'a->neg' directly in applications.
2865 *) Implement fast modular reduction for pseudo-Mersenne primes
2866 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2867 EC_GROUP_new_curve_GFp() will now automatically use this
2869 [Nils Larsch <nla@trustcenter.de>]
2871 *) Add new lock type (CRYPTO_LOCK_BN).
2874 *) Change the ENGINE framework to automatically load engines
2875 dynamically from specific directories unless they could be
2876 found to already be built in or loaded. Move all the
2877 current engines except for the cryptodev one to a new
2879 The engines in engines/ are built as shared libraries if
2880 the "shared" options was given to ./Configure or ./config.
2881 Otherwise, they are inserted in libcrypto.a.
2882 /usr/local/ssl/engines is the default directory for dynamic
2883 engines, but that can be overriden at configure time through
2884 the usual use of --prefix and/or --openssldir, and at run
2885 time with the environment variable OPENSSL_ENGINES.
2886 [Geoff Thorpe and Richard Levitte]
2888 *) Add Makefile.shared, a helper makefile to build shared
2889 libraries. Addapt Makefile.org.
2892 *) Add version info to Win32 DLLs.
2893 [Peter 'Luna' Runestig" <peter@runestig.com>]
2895 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2896 can be added using this API to created arbitrary PKCS#12
2897 files while avoiding the low level API.
2899 New options to PKCS12_create(), key or cert can be NULL and
2900 will then be omitted from the output file. The encryption
2901 algorithm NIDs can be set to -1 for no encryption, the mac
2902 iteration count can be set to 0 to omit the mac.
2904 Enhance pkcs12 utility by making the -nokeys and -nocerts
2905 options work when creating a PKCS#12 file. New option -nomac
2906 to omit the mac, NONE can be set for an encryption algorithm.
2907 New code is modified to use the enhanced PKCS12_create()
2908 instead of the low level API.
2911 *) Extend ASN1 encoder to support indefinite length constructed
2912 encoding. This can output sequences tags and octet strings in
2913 this form. Modify pk7_asn1.c to support indefinite length
2914 encoding. This is experimental and needs additional code to
2915 be useful, such as an ASN1 bio and some enhanced streaming
2918 Extend template encode functionality so that tagging is passed
2919 down to the template encoder.
2922 *) Let 'openssl req' fail if an argument to '-newkey' is not
2923 recognized instead of using RSA as a default.
2926 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2927 As these are not official, they are not included in "ALL";
2928 the "ECCdraft" ciphersuite group alias can be used to select them.
2929 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2931 *) Add ECDH engine support.
2932 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2934 *) Add ECDH in new directory crypto/ecdh/.
2935 [Douglas Stebila (Sun Microsystems Laboratories)]
2937 *) Let BN_rand_range() abort with an error after 100 iterations
2938 without success (which indicates a broken PRNG).
2941 *) Change BN_mod_sqrt() so that it verifies that the input value
2942 is really the square of the return value. (Previously,
2943 BN_mod_sqrt would show GIGO behaviour.)
2946 *) Add named elliptic curves over binary fields from X9.62, SECG,
2947 and WAP/WTLS; add OIDs that were still missing.
2949 [Sheueling Chang Shantz and Douglas Stebila
2950 (Sun Microsystems Laboratories)]
2952 *) Extend the EC library for elliptic curves over binary fields
2953 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2956 EC_GF2m_simple_method
2960 EC_GROUP_new_curve_GF2m
2961 EC_GROUP_set_curve_GF2m
2962 EC_GROUP_get_curve_GF2m
2963 EC_POINT_set_affine_coordinates_GF2m
2964 EC_POINT_get_affine_coordinates_GF2m
2965 EC_POINT_set_compressed_coordinates_GF2m
2967 Point compression for binary fields is disabled by default for
2968 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2971 As binary polynomials are represented as BIGNUMs, various members
2972 of the EC_GROUP and EC_POINT data structures can be shared
2973 between the implementations for prime fields and binary fields;
2974 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2975 are essentially identical to their ..._GFp counterparts.
2976 (For simplicity, the '..._GFp' prefix has been dropped from
2977 various internal method names.)
2979 An internal 'field_div' method (similar to 'field_mul' and
2980 'field_sqr') has been added; this is used only for binary fields.
2982 [Sheueling Chang Shantz and Douglas Stebila
2983 (Sun Microsystems Laboratories)]
2985 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2986 through methods ('mul', 'precompute_mult').
2988 The generic implementations (now internally called 'ec_wNAF_mul'
2989 and 'ec_wNAF_precomputed_mult') remain the default if these
2990 methods are undefined.
2992 [Sheueling Chang Shantz and Douglas Stebila
2993 (Sun Microsystems Laboratories)]
2995 *) New function EC_GROUP_get_degree, which is defined through
2996 EC_METHOD. For curves over prime fields, this returns the bit
2997 length of the modulus.
2999 [Sheueling Chang Shantz and Douglas Stebila
3000 (Sun Microsystems Laboratories)]
3002 *) New functions EC_GROUP_dup, EC_POINT_dup.
3003 (These simply call ..._new and ..._copy).
3005 [Sheueling Chang Shantz and Douglas Stebila
3006 (Sun Microsystems Laboratories)]
3008 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3009 Polynomials are represented as BIGNUMs (where the sign bit is not
3010 used) in the following functions [macros]:
3013 BN_GF2m_sub [= BN_GF2m_add]
3014 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3015 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3016 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3018 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3019 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3020 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3021 BN_GF2m_cmp [= BN_ucmp]
3023 (Note that only the 'mod' functions are actually for fields GF(2^m).
3024 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3026 For some functions, an the irreducible polynomial defining a
3027 field can be given as an 'unsigned int[]' with strictly
3028 decreasing elements giving the indices of those bits that are set;
3029 i.e., p[] represents the polynomial
3030 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3032 p[0] > p[1] > ... > p[k] = 0.
3033 This applies to the following functions:
3038 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3039 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3041 BN_GF2m_mod_sqrt_arr
3042 BN_GF2m_mod_solve_quad_arr
3046 Conversion can be performed by the following functions:
3051 bntest.c has additional tests for binary polynomial arithmetic.
3053 Two implementations for BN_GF2m_mod_div() are available.
3054 The default algorithm simply uses BN_GF2m_mod_inv() and
3055 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3056 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3057 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3059 [Sheueling Chang Shantz and Douglas Stebila
3060 (Sun Microsystems Laboratories)]
3062 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3063 functionality is disabled at compile-time.
3064 [Douglas Stebila <douglas.stebila@sun.com>]
3066 *) Change default behaviour of 'openssl asn1parse' so that more
3067 information is visible when viewing, e.g., a certificate:
3069 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3070 mode the content of non-printable OCTET STRINGs is output in a
3071 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3072 avoid the appearance of a printable string.
3073 [Nils Larsch <nla@trustcenter.de>]
3075 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3077 EC_GROUP_set_asn1_flag()
3078 EC_GROUP_get_asn1_flag()
3079 EC_GROUP_set_point_conversion_form()
3080 EC_GROUP_get_point_conversion_form()
3081 These control ASN1 encoding details:
3082 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3083 has been set to OPENSSL_EC_NAMED_CURVE.
3084 - Points are encoded in uncompressed form by default; options for
3085 asn1_for are as for point2oct, namely
3086 POINT_CONVERSION_COMPRESSED
3087 POINT_CONVERSION_UNCOMPRESSED
3088 POINT_CONVERSION_HYBRID
3090 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3093 EC_GROUP_get0_seed()
3094 EC_GROUP_get_seed_len()
3095 This is used only for ASN1 purposes (so far).
3096 [Nils Larsch <nla@trustcenter.de>]
3098 *) Add 'field_type' member to EC_METHOD, which holds the NID
3099 of the appropriate field type OID. The new function
3100 EC_METHOD_get_field_type() returns this value.
3101 [Nils Larsch <nla@trustcenter.de>]
3106 EC_POINT_point2hex()
3107 EC_POINT_hex2point()
3108 providing useful interfaces to EC_POINT_point2oct() and
3109 EC_POINT_oct2point().
3110 [Nils Larsch <nla@trustcenter.de>]
3112 *) Change internals of the EC library so that the functions
3113 EC_GROUP_set_generator()
3114 EC_GROUP_get_generator()
3115 EC_GROUP_get_order()
3116 EC_GROUP_get_cofactor()
3117 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3118 to methods, which would lead to unnecessary code duplication when
3119 adding different types of curves.
3120 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3122 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3123 arithmetic, and such that modified wNAFs are generated
3124 (which avoid length expansion in many cases).
3127 *) Add a function EC_GROUP_check_discriminant() (defined via
3128 EC_METHOD) that verifies that the curve discriminant is non-zero.
3130 Add a function EC_GROUP_check() that makes some sanity tests
3131 on a EC_GROUP, its generator and order. This includes
3132 EC_GROUP_check_discriminant().
3133 [Nils Larsch <nla@trustcenter.de>]
3135 *) Add ECDSA in new directory crypto/ecdsa/.
3137 Add applications 'openssl ecparam' and 'openssl ecdsa'
3138 (these are based on 'openssl dsaparam' and 'openssl dsa').
3140 ECDSA support is also included in various other files across the
3141 library. Most notably,
3142 - 'openssl req' now has a '-newkey ecdsa:file' option;
3143 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3144 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3145 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3146 them suitable for ECDSA where domain parameters must be
3147 extracted before the specific public key;
3148 - ECDSA engine support has been added.
3149 [Nils Larsch <nla@trustcenter.de>]
3151 *) Include some named elliptic curves, and add OIDs from X9.62,
3152 SECG, and WAP/WTLS. Each curve can be obtained from the new
3154 EC_GROUP_new_by_curve_name(),
3155 and the list of available named curves can be obtained with
3156 EC_get_builtin_curves().
3157 Also add a 'curve_name' member to EC_GROUP objects, which can be
3159 EC_GROUP_set_curve_name()
3160 EC_GROUP_get_curve_name()
3161 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3163 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3164 was actually never needed) and in BN_mul(). The removal in BN_mul()
3165 required a small change in bn_mul_part_recursive() and the addition
3166 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3167 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3168 bn_sub_words() and bn_add_words() except they take arrays with
3172 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3174 *) Cleanse PEM buffers before freeing them since they may contain
3176 [Benjamin Bennett <ben@psc.edu>]
3178 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3179 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3180 authentication-only ciphersuites.
3183 *) Since AES128 and AES256 share a single mask bit in the logic of
3184 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3185 kludge to work properly if AES128 is available and AES256 isn't.
3188 *) Expand security boundary to match 1.1.1 module.
3191 *) Remove redundant features: hash file source, editing of test vectors
3192 modify fipsld to use external fips_premain.c signature.
3195 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3196 run algorithm test programs.
3199 *) Make algorithm test programs more tolerant of whitespace.
3202 *) Have SSL/TLS server implementation tolerate "mismatched" record
3203 protocol version while receiving ClientHello even if the
3204 ClientHello is fragmented. (The server can't insist on the
3205 particular protocol version it has chosen before the ServerHello
3206 message has informed the client about his choice.)
3209 *) Load error codes if they are not already present instead of using a
3210 static variable. This allows them to be cleanly unloaded and reloaded.
3213 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3215 *) Introduce limits to prevent malicious keys being able to
3216 cause a denial of service. (CVE-2006-2940)
3217 [Steve Henson, Bodo Moeller]
3219 *) Fix ASN.1 parsing of certain invalid structures that can result
3220 in a denial of service. (CVE-2006-2937) [Steve Henson]
3222 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3223 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3225 *) Fix SSL client code which could crash if connecting to a
3226 malicious SSLv2 server. (CVE-2006-4343)
3227 [Tavis Ormandy and Will Drewry, Google Security Team]
3229 *) Change ciphersuite string processing so that an explicit
3230 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3231 will no longer include "AES128-SHA"), and any other similar
3232 ciphersuite (same bitmap) from *other* protocol versions (so that
3233 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3234 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3235 changes from 0.9.8b and 0.9.8d.
3238 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3240 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3241 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3243 *) Change the Unix randomness entropy gathering to use poll() when
3244 possible instead of select(), since the latter has some
3245 undesirable limitations.
3246 [Darryl Miles via Richard Levitte and Bodo Moeller]
3248 *) Disable rogue ciphersuites:
3250 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3251 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3252 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3254 The latter two were purportedly from
3255 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3258 Also deactive the remaining ciphersuites from
3259 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3260 unofficial, and the ID has long expired.
3263 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3264 dual-core machines) and other potential thread-safety issues.
3267 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3269 *) Adapt fipsld and the build system to link against the validated FIPS
3270 module in FIPS mode.
3273 *) Fixes for VC++ 2005 build under Windows.
3276 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3277 from a Windows bash shell such as MSYS. It is autodetected from the
3278 "config" script when run from a VC++ environment. Modify standard VC++
3279 build to use fipscanister.o from the GNU make build.
3282 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3284 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3285 The value now differs depending on if you build for FIPS or not.
3286 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3287 safely run with a non-FIPSed libcrypto, as it may crash because of
3288 the difference induced by this change.
3291 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3293 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3294 (part of SSL_OP_ALL). This option used to disable the
3295 countermeasure against man-in-the-middle protocol-version
3296 rollback in the SSL 2.0 server implementation, which is a bad
3297 idea. (CVE-2005-2969)
3299 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3300 for Information Security, National Institute of Advanced Industrial
3301 Science and Technology [AIST], Japan)]
3303 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3304 mainly for FIPS compliance and not fully integrated at this stage.
3307 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3308 the exponentiation using a fixed-length exponent. (Otherwise,
3309 the information leaked through timing could expose the secret key
3310 after many signatures; cf. Bleichenbacher's attack on DSA with
3314 *) Make a new fixed-window mod_exp implementation the default for
3315 RSA, DSA, and DH private-key operations so that the sequence of
3316 squares and multiplies and the memory access pattern are
3317 independent of the particular secret key. This will mitigate
3318 cache-timing and potential related attacks.
3320 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3321 and this is automatically used by BN_mod_exp_mont() if the new flag
3322 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3323 will use this BN flag for private exponents unless the flag
3324 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3325 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3327 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3329 *) Change the client implementation for SSLv23_method() and
3330 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3331 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3332 (Previously, the SSL 2.0 backwards compatible Client Hello
3333 message format would be used even with SSL_OP_NO_SSLv2.)
3336 *) Add support for smime-type MIME parameter in S/MIME messages which some
3340 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3341 a threadsafe manner. Modify rsa code to use new function and add calls
3342 to dsa and dh code (which had race conditions before).
3345 *) Include the fixed error library code in the C error file definitions
3346 instead of fixing them up at runtime. This keeps the error code
3347 structures constant.
3350 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3352 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3355 *) Fixes for newer kerberos headers. NB: the casts are needed because
3356 the 'length' field is signed on one version and unsigned on another
3357 with no (?) obvious way to tell the difference, without these VC++
3358 complains. Also the "definition" of FAR (blank) is no longer included
3359 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3360 some needed definitions.
3363 *) Undo Cygwin change.
3366 *) Added support for proxy certificates according to RFC 3820.
3367 Because they may be a security thread to unaware applications,
3368 they must be explicitely allowed in run-time. See
3369 docs/HOWTO/proxy_certificates.txt for further information.
3372 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3374 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3375 server and client random values. Previously
3376 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3377 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3379 This change has negligible security impact because:
3381 1. Server and client random values still have 24 bytes of pseudo random
3384 2. Server and client random values are sent in the clear in the initial
3387 3. The master secret is derived using the premaster secret (48 bytes in
3388 size for static RSA ciphersuites) as well as client server and random
3391 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3394 [Stephen Henson, reported by UK NISCC]
3396 *) Use Windows randomness collection on Cygwin.
3399 *) Fix hang in EGD/PRNGD query when communication socket is closed
3400 prematurely by EGD/PRNGD.
3401 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3403 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3406 *) Back-port of selected performance improvements from development
3407 branch, as well as improved support for PowerPC platforms.
3410 *) Add lots of checks for memory allocation failure, error codes to indicate
3411 failure and freeing up memory if a failure occurs.
3412 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3414 *) Add new -passin argument to dgst.
3417 *) Perform some character comparisons of different types in X509_NAME_cmp:
3418 this is needed for some certificates that reencode DNs into UTF8Strings
3419 (in violation of RFC3280) and can't or wont issue name rollover
3423 *) Make an explicit check during certificate validation to see that
3424 the CA setting in each certificate on the chain is correct. As a
3425 side effect always do the following basic checks on extensions,
3426 not just when there's an associated purpose to the check:
3428 - if there is an unhandled critical extension (unless the user
3429 has chosen to ignore this fault)
3430 - if the path length has been exceeded (if one is set at all)
3431 - that certain extensions fit the associated purpose (if one has
3435 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3437 *) Avoid a race condition when CRLs are checked in a multi threaded
3438 environment. This would happen due to the reordering of the revoked
3439 entries during signature checking and serial number lookup. Now the
3440 encoding is cached and the serial number sort performed under a lock.
3441 Add new STACK function sk_is_sorted().
3444 *) Add Delta CRL to the extension code.
3447 *) Various fixes to s3_pkt.c so alerts are sent properly.
3448 [David Holmes <d.holmes@f5.com>]
3450 *) Reduce the chances of duplicate issuer name and serial numbers (in
3451 violation of RFC3280) using the OpenSSL certificate creation utilities.
3452 This is done by creating a random 64 bit value for the initial serial
3453 number when a serial number file is created or when a self signed
3454 certificate is created using 'openssl req -x509'. The initial serial
3455 number file is created using 'openssl x509 -next_serial' in CA.pl
3456 rather than being initialized to 1.
3459 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3461 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3462 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3463 [Joe Orton, Steve Henson]
3465 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3467 [Joe Orton, Steve Henson]
3469 *) Make it possible to have multiple active certificates with the same
3470 subject in the CA index file. This is done only if the keyword
3471 'unique_subject' is set to 'no' in the main CA section (default
3472 if 'CA_default') of the configuration file. The value is saved
3473 with the database itself in a separate index attribute file,
3474 named like the index file with '.attr' appended to the name.
3477 *) X509 verify fixes. Disable broken certificate workarounds when
3478 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3479 keyUsage extension present. Don't accept CRLs with unhandled critical
3480 extensions: since verify currently doesn't process CRL extensions this
3481 rejects a CRL with *any* critical extensions. Add new verify error codes
3485 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3486 A clarification of RFC2560 will require the use of OCTET STRINGs and
3487 some implementations cannot handle the current raw format. Since OpenSSL
3488 copies and compares OCSP nonces as opaque blobs without any attempt at
3489 parsing them this should not create any compatibility issues.
3492 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3493 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3494 this HMAC (and other) operations are several times slower than OpenSSL
3498 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3499 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3501 *) Use the correct content when signing type "other".
3504 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3506 *) Fix various bugs revealed by running the NISCC test suite:
3508 Stop out of bounds reads in the ASN1 code when presented with
3509 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3511 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3513 If verify callback ignores invalid public key errors don't try to check
3514 certificate signature with the NULL public key.
3518 *) New -ignore_err option in ocsp application to stop the server
3519 exiting on the first error in a request.
3522 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3523 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3527 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3528 extra data after the compression methods not only for TLS 1.0
3529 but also for SSL 3.0 (as required by the specification).
3530 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3532 *) Change X509_certificate_type() to mark the key as exported/exportable
3533 when it's 512 *bits* long, not 512 bytes.
3536 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3537 blocks during encryption.
3540 *) Various fixes to base64 BIO and non blocking I/O. On write
3541 flushes were not handled properly if the BIO retried. On read
3542 data was not being buffered properly and had various logic bugs.
3543 This also affects blocking I/O when the data being decoded is a
3547 *) Various S/MIME bugfixes and compatibility changes:
3548 output correct application/pkcs7 MIME type if
3549 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3550 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3551 of files as .eml work). Correctly handle very long lines in MIME
3555 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3557 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3558 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3559 a protocol version number mismatch like a decryption error
3560 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3563 *) Turn on RSA blinding by default in the default implementation
3564 to avoid a timing attack. Applications that don't want it can call
3565 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3566 They would be ill-advised to do so in most cases.
3567 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3569 *) Change RSA blinding code so that it works when the PRNG is not
3570 seeded (in this case, the secret RSA exponent is abused as
3571 an unpredictable seed -- if it is not unpredictable, there
3572 is no point in blinding anyway). Make RSA blinding thread-safe
3573 by remembering the creator's thread ID in rsa->blinding and
3574 having all other threads use local one-time blinding factors
3575 (this requires more computation than sharing rsa->blinding, but
3576 avoids excessive locking; and if an RSA object is not shared
3577 between threads, blinding will still be very fast).
3580 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3581 ENGINE as defaults for all supported algorithms irrespective of
3582 the 'flags' parameter. 'flags' is now honoured, so applications
3583 should make sure they are passing it correctly.
3586 *) Target "mingw" now allows native Windows code to be generated in
3587 the Cygwin environment as well as with the MinGW compiler.
3590 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3592 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3593 via timing by performing a MAC computation even if incorrrect
3594 block cipher padding has been found. This is a countermeasure
3595 against active attacks where the attacker has to distinguish
3596 between bad padding and a MAC verification error. (CVE-2003-0078)
3598 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3599 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3600 Martin Vuagnoux (EPFL, Ilion)]
3602 *) Make the no-err option work as intended. The intention with no-err
3603 is not to have the whole error stack handling routines removed from
3604 libcrypto, it's only intended to remove all the function name and
3605 reason texts, thereby removing some of the footprint that may not
3606 be interesting if those errors aren't displayed anyway.
3608 NOTE: it's still possible for any application or module to have it's
3609 own set of error texts inserted. The routines are there, just not
3610 used by default when no-err is given.
3613 *) Add support for FreeBSD on IA64.
3614 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3616 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3617 Kerberos function mit_des_cbc_cksum(). Before this change,
3618 the value returned by DES_cbc_cksum() was like the one from
3619 mit_des_cbc_cksum(), except the bytes were swapped.
3620 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3622 *) Allow an application to disable the automatic SSL chain building.
3623 Before this a rather primitive chain build was always performed in
3624 ssl3_output_cert_chain(): an application had no way to send the
3625 correct chain if the automatic operation produced an incorrect result.
3627 Now the chain builder is disabled if either:
3629 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3631 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3633 The reasoning behind this is that an application would not want the
3634 auto chain building to take place if extra chain certificates are
3635 present and it might also want a means of sending no additional
3636 certificates (for example the chain has two certificates and the
3640 *) Add the possibility to build without the ENGINE framework.
3641 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3643 *) Under Win32 gmtime() can return NULL: check return value in
3644 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3647 *) DSA routines: under certain error conditions uninitialized BN objects
3648 could be freed. Solution: make sure initialization is performed early
3649 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3650 Nils Larsch <nla@trustcenter.de> via PR#459)
3653 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3654 checked on reconnect on the client side, therefore session resumption
3655 could still fail with a "ssl session id is different" error. This
3656 behaviour is masked when SSL_OP_ALL is used due to
3657 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3658 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3659 followup to PR #377.
3662 *) IA-32 assembler support enhancements: unified ELF targets, support
3663 for SCO/Caldera platforms, fix for Cygwin shared build.
3666 *) Add support for FreeBSD on sparc64. As a consequence, support for
3667 FreeBSD on non-x86 processors is separate from x86 processors on
3668 the config script, much like the NetBSD support.
3669 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3671 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3673 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3676 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3677 code (06) was taken as the first octet of the session ID and the last
3678 octet was ignored consequently. As a result SSLv2 client side session
3679 caching could not have worked due to the session ID mismatch between
3681 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3685 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3686 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3690 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3691 seems that in spite of existing for more than a year, many application
3692 author have done nothing to provide the necessary callbacks, which
3693 means that this particular engine will not work properly anywhere.
3694 This is a very unfortunate situation which forces us, in the name
3695 of usability, to give the hw_ncipher.c a static lock, which is part
3697 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3698 appear in 0.9.8 or later. We EXPECT application authors to have
3699 dealt properly with this when 0.9.8 is released (unless we actually
3700 make such changes in the libcrypto locking code that changes will
3701 have to be made anyway).
3704 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3705 octets have been read, EOF or an error occurs. Without this change
3706 some truncated ASN1 structures will not produce an error.
3709 *) Disable Heimdal support, since it hasn't been fully implemented.
3710 Still give the possibility to force the use of Heimdal, but with
3711 warnings and a request that patches get sent to openssl-dev.
3714 *) Add the VC-CE target, introduce the WINCE sysname, and add
3715 INSTALL.WCE and appropriate conditionals to make it build.
3716 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3718 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3719 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3720 edit numbers of the version.
3721 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3723 *) Introduce safe string copy and catenation functions
3724 (BUF_strlcpy() and BUF_strlcat()).
3725 [Ben Laurie (CHATS) and Richard Levitte]
3727 *) Avoid using fixed-size buffers for one-line DNs.
3728 [Ben Laurie (CHATS)]
3730 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3731 resizing buffers containing secrets, and use where appropriate.
3732 [Ben Laurie (CHATS)]
3734 *) Avoid using fixed size buffers for configuration file location.
3735 [Ben Laurie (CHATS)]
3737 *) Avoid filename truncation for various CA files.
3738 [Ben Laurie (CHATS)]
3740 *) Use sizeof in preference to magic numbers.
3741 [Ben Laurie (CHATS)]
3743 *) Avoid filename truncation in cert requests.
3744 [Ben Laurie (CHATS)]
3746 *) Add assertions to check for (supposedly impossible) buffer
3748 [Ben Laurie (CHATS)]
3750 *) Don't cache truncated DNS entries in the local cache (this could
3751 potentially lead to a spoofing attack).
3752 [Ben Laurie (CHATS)]
3754 *) Fix various buffers to be large enough for hex/decimal
3755 representations in a platform independent manner.
3756 [Ben Laurie (CHATS)]
3758 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3759 resizing buffers containing secrets, and use where appropriate.
3760 [Ben Laurie (CHATS)]
3762 *) Add BIO_indent() to avoid much slightly worrying code to do
3764 [Ben Laurie (CHATS)]
3766 *) Convert sprintf()/BIO_puts() to BIO_printf().
3767 [Ben Laurie (CHATS)]
3769 *) buffer_gets() could terminate with the buffer only half
3771 [Ben Laurie (CHATS)]
3773 *) Add assertions to prevent user-supplied crypto functions from
3774 overflowing internal buffers by having large block sizes, etc.
3775 [Ben Laurie (CHATS)]
3777 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3779 [Ben Laurie (CHATS)]
3781 *) Eliminate unused copy of key in RC4.
3782 [Ben Laurie (CHATS)]
3784 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3785 [Ben Laurie (CHATS)]
3787 *) Fix off-by-one error in EGD path.
3788 [Ben Laurie (CHATS)]
3790 *) If RANDFILE path is too long, ignore instead of truncating.
3791 [Ben Laurie (CHATS)]
3793 *) Eliminate unused and incorrectly sized X.509 structure
3795 [Ben Laurie (CHATS)]
3797 *) Eliminate unused and dangerous function knumber().
3798 [Ben Laurie (CHATS)]
3800 *) Eliminate unused and dangerous structure, KSSL_ERR.
3801 [Ben Laurie (CHATS)]
3803 *) Protect against overlong session ID context length in an encoded
3804 session object. Since these are local, this does not appear to be
3806 [Ben Laurie (CHATS)]
3808 *) Change from security patch (see 0.9.6e below) that did not affect
3809 the 0.9.6 release series:
3811 Remote buffer overflow in SSL3 protocol - an attacker could
3812 supply an oversized master key in Kerberos-enabled versions.
3814 [Ben Laurie (CHATS)]
3816 *) Change the SSL kerb5 codes to match RFC 2712.
3819 *) Make -nameopt work fully for req and add -reqopt switch.
3820 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3822 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3823 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3825 *) Make sure tests can be performed even if the corresponding algorithms
3826 have been removed entirely. This was also the last step to make
3827 OpenSSL compilable with DJGPP under all reasonable conditions.
3828 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3830 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3831 to allow version independent disabling of normally unselected ciphers,
3832 which may be activated as a side-effect of selecting a single cipher.
3834 (E.g., cipher list string "RSA" enables ciphersuites that are left
3835 out of "ALL" because they do not provide symmetric encryption.
3836 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3837 [Lutz Jaenicke, Bodo Moeller]
3839 *) Add appropriate support for separate platform-dependent build
3840 directories. The recommended way to make a platform-dependent
3841 build directory is the following (tested on Linux), maybe with
3844 # Place yourself outside of the OpenSSL source tree. In
3845 # this example, the environment variable OPENSSL_SOURCE
3846 # is assumed to contain the absolute OpenSSL source directory.
3847 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3848 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3849 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3850 mkdir -p `dirname $F`
3851 ln -s $OPENSSL_SOURCE/$F $F
3854 To be absolutely sure not to disturb the source tree, a "make clean"
3855 is a good thing. If it isn't successfull, don't worry about it,
3856 it probably means the source directory is very clean.
3859 *) Make sure any ENGINE control commands make local copies of string
3860 pointers passed to them whenever necessary. Otherwise it is possible
3861 the caller may have overwritten (or deallocated) the original string
3862 data when a later ENGINE operation tries to use the stored values.
3863 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3865 *) Improve diagnostics in file reading and command-line digests.
3866 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3868 *) Add AES modes CFB and OFB to the object database. Correct an
3869 error in AES-CFB decryption.
3872 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3873 allows existing EVP_CIPHER_CTX structures to be reused after
3874 calling EVP_*Final(). This behaviour is used by encryption
3875 BIOs and some applications. This has the side effect that
3876 applications must explicitly clean up cipher contexts with
3877 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3880 *) Check the values of dna and dnb in bn_mul_recursive before calling
3881 bn_mul_comba (a non zero value means the a or b arrays do not contain
3882 n2 elements) and fallback to bn_mul_normal if either is not zero.
3885 *) Fix escaping of non-ASCII characters when using the -subj option
3886 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3889 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3890 form for "surname", serialNumber has no short form.
3891 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3892 therefore remove "mail" short name for "internet 7".
3893 The OID for unique identifiers in X509 certificates is
3894 x500UniqueIdentifier, not uniqueIdentifier.
3895 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3898 *) Add an "init" command to the ENGINE config module and auto initialize
3899 ENGINEs. Without any "init" command the ENGINE will be initialized
3900 after all ctrl commands have been executed on it. If init=1 the
3901 ENGINE is initailized at that point (ctrls before that point are run
3902 on the uninitialized ENGINE and after on the initialized one). If
3903 init=0 then the ENGINE will not be iniatialized at all.
3906 *) Fix the 'app_verify_callback' interface so that the user-defined
3907 argument is actually passed to the callback: In the
3908 SSL_CTX_set_cert_verify_callback() prototype, the callback
3909 declaration has been changed from
3912 int (*cb)(X509_STORE_CTX *,void *);
3913 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3914 i=s->ctx->app_verify_callback(&ctx)
3915 has been changed into
3916 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3918 To update applications using SSL_CTX_set_cert_verify_callback(),
3919 a dummy argument can be added to their callback functions.
3920 [D. K. Smetters <smetters@parc.xerox.com>]
3922 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3923 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3925 *) Add and OPENSSL_LOAD_CONF define which will cause
3926 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3927 This allows older applications to transparently support certain
3928 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3929 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3930 load the config file and OPENSSL_add_all_algorithms_conf() which will
3931 always load it have also been added.
3934 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3935 Adjust NIDs and EVP layer.
3936 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3938 *) Config modules support in openssl utility.
3940 Most commands now load modules from the config file,
3941 though in a few (such as version) this isn't done
3942 because it couldn't be used for anything.
3944 In the case of ca and req the config file used is
3945 the same as the utility itself: that is the -config
3946 command line option can be used to specify an
3950 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3951 use "openssl_conf" if filename is NULL use default openssl config file.
3954 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3955 config section name. Add a new flag to tolerate a missing config file
3956 and move code to CONF_modules_load_file().
3959 *) Support for crypto accelerator cards from Accelerated Encryption
3960 Processing, www.aep.ie. (Use engine 'aep')
3961 The support was copied from 0.9.6c [engine] and adapted/corrected
3962 to work with the new engine framework.
3963 [AEP Inc. and Richard Levitte]
3965 *) Support for SureWare crypto accelerator cards from Baltimore
3966 Technologies. (Use engine 'sureware')
3967 The support was copied from 0.9.6c [engine] and adapted
3968 to work with the new engine framework.
3971 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3972 make the newer ENGINE framework commands for the CHIL engine work.
3973 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3975 *) Make it possible to produce shared libraries on ReliantUNIX.
3976 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3978 *) Add the configuration target debug-linux-ppro.
3979 Make 'openssl rsa' use the general key loading routines
3980 implemented in apps.c, and make those routines able to
3981 handle the key format FORMAT_NETSCAPE and the variant
3983 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3985 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3986 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3988 *) Add -keyform to rsautl, and document -engine.
3989 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3991 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3992 BIO_R_NO_SUCH_FILE error code rather than the generic
3993 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3996 *) Add new functions
3998 ERR_peek_last_error_line
3999 ERR_peek_last_error_line_data.
4000 These are similar to
4003 ERR_peek_error_line_data,
4004 but report on the latest error recorded rather than the first one
4005 still in the error queue.
4006 [Ben Laurie, Bodo Moeller]
4008 *) default_algorithms option in ENGINE config module. This allows things
4010 default_algorithms = ALL
4011 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4014 *) Prelminary ENGINE config module.
4017 *) New experimental application configuration code.
4020 *) Change the AES code to follow the same name structure as all other
4021 symmetric ciphers, and behave the same way. Move everything to
4022 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4023 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4025 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4026 [Ben Laurie and Theo de Raadt]
4028 *) Add option to output public keys in req command.
4029 [Massimiliano Pala madwolf@openca.org]
4031 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4032 (up to about 10% better than before for P-192 and P-224).
4035 *) New functions/macros
4037 SSL_CTX_set_msg_callback(ctx, cb)
4038 SSL_CTX_set_msg_callback_arg(ctx, arg)
4039 SSL_set_msg_callback(ssl, cb)
4040 SSL_set_msg_callback_arg(ssl, arg)
4042 to request calling a callback function
4044 void cb(int write_p, int version, int content_type,
4045 const void *buf, size_t len, SSL *ssl, void *arg)
4047 whenever a protocol message has been completely received
4048 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4049 protocol version according to which the SSL library interprets
4050 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4051 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4052 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4053 specification (change_cipher_spec(20), alert(21), handshake(22)).
4054 'buf' and 'len' point to the actual message, 'ssl' to the
4055 SSL object, and 'arg' is the application-defined value set by
4056 SSL[_CTX]_set_msg_callback_arg().
4058 'openssl s_client' and 'openssl s_server' have new '-msg' options
4059 to enable a callback that displays all protocol messages.
4062 *) Change the shared library support so shared libraries are built as
4063 soon as the corresponding static library is finished, and thereby get
4064 openssl and the test programs linked against the shared library.
4065 This still only happens when the keyword "shard" has been given to
4066 the configuration scripts.
4068 NOTE: shared library support is still an experimental thing, and
4069 backward binary compatibility is still not guaranteed.
4070 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4072 *) Add support for Subject Information Access extension.
4073 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4075 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4076 additional bytes when new memory had to be allocated, not just
4077 when reusing an existing buffer.
4080 *) New command line and configuration option 'utf8' for the req command.
4081 This allows field values to be specified as UTF8 strings.
4084 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4085 runs for the former and machine-readable output for the latter.
4088 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4089 of the e-mail address in the DN (i.e., it will go into a certificate
4090 extension only). The new configuration file option 'email_in_dn = no'
4091 has the same effect.
4092 [Massimiliano Pala madwolf@openca.org]
4094 *) Change all functions with names starting with des_ to be starting
4095 with DES_ instead. Add wrappers that are compatible with libdes,
4096 but are named _ossl_old_des_*. Finally, add macros that map the
4097 des_* symbols to the corresponding _ossl_old_des_* if libdes
4098 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4099 desired, the des_* symbols will be mapped to DES_*, with one
4102 Since we provide two compatibility mappings, the user needs to
4103 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4104 compatibility is desired. The default (i.e., when that macro
4105 isn't defined) is OpenSSL 0.9.6c compatibility.
4107 There are also macros that enable and disable the support of old
4108 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4109 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4110 are defined, the default will apply: to support the old des routines.
4112 In either case, one must include openssl/des.h to get the correct
4113 definitions. Do not try to just include openssl/des_old.h, that
4116 NOTE: This is a major break of an old API into a new one. Software
4117 authors are encouraged to switch to the DES_ style functions. Some
4118 time in the future, des_old.h and the libdes compatibility functions
4119 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4120 default), and then completely removed.
4123 *) Test for certificates which contain unsupported critical extensions.
4124 If such a certificate is found during a verify operation it is
4125 rejected by default: this behaviour can be overridden by either
4126 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4127 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4128 X509_supported_extension() has also been added which returns 1 if a
4129 particular extension is supported.
4132 *) Modify the behaviour of EVP cipher functions in similar way to digests
4133 to retain compatibility with existing code.
4136 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4137 compatibility with existing code. In particular the 'ctx' parameter does
4138 not have to be to be initialized before the call to EVP_DigestInit() and
4139 it is tidied up after a call to EVP_DigestFinal(). New function
4140 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4141 EVP_MD_CTX_copy() changed to not require the destination to be
4142 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4143 requires the destination to be valid.
4145 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4146 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4149 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4150 so that complete 'Handshake' protocol structures are kept in memory
4151 instead of overwriting 'msg_type' and 'length' with 'body' data.
4154 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4155 [Massimo Santin via Richard Levitte]
4157 *) Major restructuring to the underlying ENGINE code. This includes
4158 reduction of linker bloat, separation of pure "ENGINE" manipulation
4159 (initialisation, etc) from functionality dealing with implementations
4160 of specific crypto iterfaces. This change also introduces integrated
4161 support for symmetric ciphers and digest implementations - so ENGINEs
4162 can now accelerate these by providing EVP_CIPHER and EVP_MD
4163 implementations of their own. This is detailed in crypto/engine/README
4164 as it couldn't be adequately described here. However, there are a few
4165 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4166 were changed in the original introduction of ENGINE code have now
4167 reverted back - the hooking from this code to ENGINE is now a good
4168 deal more passive and at run-time, operations deal directly with
4169 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4170 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4171 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4172 they were not being used by the framework as there is no concept of a
4173 BIGNUM_METHOD and they could not be generalised to the new
4174 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4175 ENGINE_cpy() has been removed as it cannot be consistently defined in
4179 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4182 *) Change mkdef.pl to sort symbols that get the same entry number,
4183 and make sure the automatically generated functions ERR_load_*
4184 become part of libeay.num as well.
4187 *) New function SSL_renegotiate_pending(). This returns true once
4188 renegotiation has been requested (either SSL_renegotiate() call
4189 or HelloRequest/ClientHello receveived from the peer) and becomes
4190 false once a handshake has been completed.
4191 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4192 sends a HelloRequest, but does not ensure that a handshake takes
4193 place. SSL_renegotiate_pending() is useful for checking if the
4194 client has followed the request.)
4197 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4198 By default, clients may request session resumption even during
4199 renegotiation (if session ID contexts permit); with this option,
4200 session resumption is possible only in the first handshake.
4202 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4203 more bits available for options that should not be part of
4204 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4207 *) Add some demos for certificate and certificate request creation.
4210 *) Make maximum certificate chain size accepted from the peer application
4211 settable (SSL*_get/set_max_cert_list()), as proposed by
4212 "Douglas E. Engert" <deengert@anl.gov>.
4215 *) Add support for shared libraries for Unixware-7
4216 (Boyd Lynn Gerber <gerberb@zenez.com>).
4219 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4220 be done prior to destruction. Use this to unload error strings from
4221 ENGINEs that load their own error strings. NB: This adds two new API
4222 functions to "get" and "set" this destroy handler in an ENGINE.
4225 *) Alter all existing ENGINE implementations (except "openssl" and
4226 "openbsd") to dynamically instantiate their own error strings. This
4227 makes them more flexible to be built both as statically-linked ENGINEs
4228 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4229 Also, add stub code to each that makes building them as self-contained
4230 shared-libraries easier (see README.ENGINE).
4233 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4234 implementations into applications that are completely implemented in
4235 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4236 commands that can be used to configure what shared-library to load and
4237 to control aspects of the way it is handled. Also, made an update to
4238 the README.ENGINE file that brings its information up-to-date and
4239 provides some information and instructions on the "dynamic" ENGINE
4240 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4243 *) Make it possible to unload ranges of ERR strings with a new
4244 "ERR_unload_strings" function.
4247 *) Add a copy() function to EVP_MD.
4250 *) Make EVP_MD routines take a context pointer instead of just the
4251 md_data void pointer.
4254 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4255 that the digest can only process a single chunk of data
4256 (typically because it is provided by a piece of
4257 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4258 is only going to provide a single chunk of data, and hence the
4259 framework needn't accumulate the data for oneshot drivers.
4262 *) As with "ERR", make it possible to replace the underlying "ex_data"
4263 functions. This change also alters the storage and management of global
4264 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4265 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4266 index counters. The API functions that use this state have been changed
4267 to take a "class_index" rather than pointers to the class's local STACK
4268 and counter, and there is now an API function to dynamically create new
4269 classes. This centralisation allows us to (a) plug a lot of the
4270 thread-safety problems that existed, and (b) makes it possible to clean
4271 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4272 such data would previously have always leaked in application code and
4273 workarounds were in place to make the memory debugging turn a blind eye
4274 to it. Application code that doesn't use this new function will still
4275 leak as before, but their memory debugging output will announce it now
4276 rather than letting it slide.
4278 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4279 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4280 has a return value to indicate success or failure.
4283 *) Make it possible to replace the underlying "ERR" functions such that the
4284 global state (2 LHASH tables and 2 locks) is only used by the "default"
4285 implementation. This change also adds two functions to "get" and "set"
4286 the implementation prior to it being automatically set the first time
4287 any other ERR function takes place. Ie. an application can call "get",
4288 pass the return value to a module it has just loaded, and that module
4289 can call its own "set" function using that value. This means the
4290 module's "ERR" operations will use (and modify) the error state in the
4291 application and not in its own statically linked copy of OpenSSL code.
4294 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4295 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4296 the operation, and provides a more encapsulated way for external code
4297 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4298 to use these functions rather than manually incrementing the counts.
4300 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4303 *) Add EVP test program.
4306 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4309 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4310 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4311 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4312 These allow a CRL to be built without having to access X509_CRL fields
4313 directly. Modify 'ca' application to use new functions.
4316 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4317 bug workarounds. Rollback attack detection is a security feature.
4318 The problem will only arise on OpenSSL servers when TLSv1 is not
4319 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4320 Software authors not wanting to support TLSv1 will have special reasons
4321 for their choice and can explicitly enable this option.
4322 [Bodo Moeller, Lutz Jaenicke]
4324 *) Rationalise EVP so it can be extended: don't include a union of
4325 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4326 (similar to those existing for EVP_CIPHER_CTX).
4331 EVP_MD_CTX_init(&md); /* new function call */
4332 EVP_DigestInit(&md, EVP_sha1());
4333 EVP_DigestUpdate(&md, in, len);
4334 EVP_DigestFinal(&md, out, NULL);
4335 EVP_MD_CTX_cleanup(&md); /* new function call */
4339 *) Make DES key schedule conform to the usual scheme, as well as
4340 correcting its structure. This means that calls to DES functions
4341 now have to pass a pointer to a des_key_schedule instead of a
4342 plain des_key_schedule (which was actually always a pointer
4345 des_key_schedule ks;
4347 des_set_key_checked(..., &ks);
4348 des_ncbc_encrypt(..., &ks, ...);
4350 (Note that a later change renames 'des_...' into 'DES_...'.)
4353 *) Initial reduction of linker bloat: the use of some functions, such as
4354 PEM causes large amounts of unused functions to be linked in due to
4355 poor organisation. For example pem_all.c contains every PEM function
4356 which has a knock on effect of linking in large amounts of (unused)
4357 ASN1 code. Grouping together similar functions and splitting unrelated
4358 functions prevents this.
4361 *) Cleanup of EVP macros.
4364 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4365 correct _ecb suffix.
4368 *) Add initial OCSP responder support to ocsp application. The
4369 revocation information is handled using the text based index
4370 use by the ca application. The responder can either handle
4371 requests generated internally, supplied in files (for example
4372 via a CGI script) or using an internal minimal server.
4375 *) Add configuration choices to get zlib compression for TLS.
4378 *) Changes to Kerberos SSL for RFC 2712 compliance:
4379 1. Implemented real KerberosWrapper, instead of just using
4380 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4381 2. Implemented optional authenticator field of KerberosWrapper.
4383 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4384 and authenticator structs; see crypto/krb5/.
4386 Generalized Kerberos calls to support multiple Kerberos libraries.
4387 [Vern Staats <staatsvr@asc.hpc.mil>,
4388 Jeffrey Altman <jaltman@columbia.edu>
4389 via Richard Levitte]
4391 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4392 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4393 values for each of the key sizes rather than having just
4394 parameters (and 'speed' generating keys each time).
4397 *) Speed up EVP routines.
4400 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4401 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4402 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4403 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4405 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4406 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4407 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4410 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4412 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4415 *) Added the OS2-EMX target.
4416 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4418 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4419 to support NCONF routines in extension code. New function CONF_set_nconf()
4420 to allow functions which take an NCONF to also handle the old LHASH
4421 structure: this means that the old CONF compatible routines can be
4422 retained (in particular wrt extensions) without having to duplicate the
4423 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4426 *) Enhance the general user interface with mechanisms for inner control
4427 and with possibilities to have yes/no kind of prompts.
4430 *) Change all calls to low level digest routines in the library and
4431 applications to use EVP. Add missing calls to HMAC_cleanup() and
4432 don't assume HMAC_CTX can be copied using memcpy().
4433 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4435 *) Add the possibility to control engines through control names but with
4436 arbitrary arguments instead of just a string.
4437 Change the key loaders to take a UI_METHOD instead of a callback
4438 function pointer. NOTE: this breaks binary compatibility with earlier
4439 versions of OpenSSL [engine].
4440 Adapt the nCipher code for these new conditions and add a card insertion
4444 *) Enhance the general user interface with mechanisms to better support
4445 dialog box interfaces, application-defined prompts, the possibility
4446 to use defaults (for example default passwords from somewhere else)
4447 and interrupts/cancellations.
4450 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4451 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4454 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4455 tidy up some unnecessarily weird code in 'sk_new()').
4456 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4458 *) Change the key loading routines for ENGINEs to use the same kind
4459 callback (pem_password_cb) as all other routines that need this
4463 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4464 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4465 than this minimum value is recommended.
4468 *) New random seeder for OpenVMS, using the system process statistics
4469 that are easily reachable.
4472 *) Windows apparently can't transparently handle global
4473 variables defined in DLLs. Initialisations such as:
4475 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4477 wont compile. This is used by the any applications that need to
4478 declare their own ASN1 modules. This was fixed by adding the option
4479 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4480 needed for static libraries under Win32.
4483 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4484 setting of purpose and trust fields. New X509_STORE trust and
4485 purpose functions and tidy up setting in other SSL functions.
4488 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4489 structure. These are inherited by X509_STORE_CTX when it is
4490 initialised. This allows various defaults to be set in the
4491 X509_STORE structure (such as flags for CRL checking and custom
4492 purpose or trust settings) for functions which only use X509_STORE_CTX
4493 internally such as S/MIME.
4495 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4496 trust settings if they are not set in X509_STORE. This allows X509_STORE
4497 purposes and trust (in S/MIME for example) to override any set by default.
4499 Add command line options for CRL checking to smime, s_client and s_server
4503 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4504 are set then the CRL is looked up in the X509_STORE structure and
4505 its validity and signature checked, then if the certificate is found
4506 in the CRL the verify fails with a revoked error.
4508 Various new CRL related callbacks added to X509_STORE_CTX structure.
4510 Command line options added to 'verify' application to support this.
4512 This needs some additional work, such as being able to handle multiple
4513 CRLs with different times, extension based lookup (rather than just
4514 by subject name) and ultimately more complete V2 CRL extension
4518 *) Add a general user interface API (crypto/ui/). This is designed
4519 to replace things like des_read_password and friends (backward
4520 compatibility functions using this new API are provided).
4521 The purpose is to remove prompting functions from the DES code
4522 section as well as provide for prompting through dialog boxes in
4523 a window system and the like.
4526 *) Add "ex_data" support to ENGINE so implementations can add state at a
4527 per-structure level rather than having to store it globally.
4530 *) Make it possible for ENGINE structures to be copied when retrieved by
4531 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4532 This causes the "original" ENGINE structure to act like a template,
4533 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4534 operational state can be localised to each ENGINE structure, despite the
4535 fact they all share the same "methods". New ENGINE structures returned in
4536 this case have no functional references and the return value is the single
4537 structural reference. This matches the single structural reference returned
4538 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4542 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4543 needs to match any other type at all we need to manually clear the
4547 *) Changes to the "openssl engine" utility to include;
4548 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4549 about an ENGINE's available control commands.
4550 - executing control commands from command line arguments using the
4551 '-pre' and '-post' switches. '-post' is only used if '-t' is
4552 specified and the ENGINE is successfully initialised. The syntax for
4553 the individual commands are colon-separated, for example;
4554 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4557 *) New dynamic control command support for ENGINEs. ENGINEs can now
4558 declare their own commands (numbers), names (strings), descriptions,
4559 and input types for run-time discovery by calling applications. A
4560 subset of these commands are implicitly classed as "executable"
4561 depending on their input type, and only these can be invoked through
4562 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4563 can be based on user input, config files, etc). The distinction is
4564 that "executable" commands cannot return anything other than a boolean
4565 result and can only support numeric or string input, whereas some
4566 discoverable commands may only be for direct use through
4567 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4568 pointers, or other custom uses. The "executable" commands are to
4569 support parameterisations of ENGINE behaviour that can be
4570 unambiguously defined by ENGINEs and used consistently across any
4571 OpenSSL-based application. Commands have been added to all the
4572 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4573 control over shared-library paths without source code alterations.
4576 *) Changed all ENGINE implementations to dynamically allocate their
4577 ENGINEs rather than declaring them statically. Apart from this being
4578 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4579 this also allows the implementations to compile without using the
4580 internal engine_int.h header.
4583 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4584 'const' value. Any code that should be able to modify a RAND_METHOD
4585 should already have non-const pointers to it (ie. they should only
4586 modify their own ones).
4589 *) Made a variety of little tweaks to the ENGINE code.
4590 - "atalla" and "ubsec" string definitions were moved from header files
4591 to C code. "nuron" string definitions were placed in variables
4592 rather than hard-coded - allowing parameterisation of these values
4593 later on via ctrl() commands.
4594 - Removed unused "#if 0"'d code.
4595 - Fixed engine list iteration code so it uses ENGINE_free() to release
4596 structural references.
4597 - Constified the RAND_METHOD element of ENGINE structures.
4598 - Constified various get/set functions as appropriate and added
4599 missing functions (including a catch-all ENGINE_cpy that duplicates
4600 all ENGINE values onto a new ENGINE except reference counts/state).
4601 - Removed NULL parameter checks in get/set functions. Setting a method
4602 or function to NULL is a way of cancelling out a previously set
4603 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4604 and doesn't justify the extra error symbols and code.
4605 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4606 flags from engine_int.h to engine.h.
4607 - Changed prototypes for ENGINE handler functions (init(), finish(),
4608 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4611 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4612 to the algorithm using long division. The binary algorithm can be
4613 used only if the modulus is odd. On 32-bit systems, it is faster
4614 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4615 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4616 up to 450 bits. In 64-bit environments, the binary algorithm
4617 appears to be advantageous for much longer moduli; here we use it
4618 for moduli up to 2048 bits.
4621 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4622 could not support the combine flag in choice fields.
4625 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4626 extensions from a certificate request to the certificate.
4629 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4630 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4631 file: this allows the display of the certificate about to be
4632 signed to be customised, to allow certain fields to be included
4633 or excluded and extension details. The old system didn't display
4634 multicharacter strings properly, omitted fields not in the policy
4635 and couldn't display additional details such as extensions.
4638 *) Function EC_POINTs_mul for multiple scalar multiplication
4639 of an arbitrary number of elliptic curve points
4640 \sum scalars[i]*points[i],
4641 optionally including the generator defined for the EC_GROUP:
4642 scalar*generator + \sum scalars[i]*points[i].
4644 EC_POINT_mul is a simple wrapper function for the typical case
4645 that the point list has just one item (besides the optional
4649 *) First EC_METHODs for curves over GF(p):
4651 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4652 operations and provides various method functions that can also
4653 operate with faster implementations of modular arithmetic.
4655 EC_GFp_mont_method() reuses most functions that are part of
4656 EC_GFp_simple_method, but uses Montgomery arithmetic.
4658 [Bodo Moeller; point addition and point doubling
4659 implementation directly derived from source code provided by
4660 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4662 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4663 crypto/ec/ec_lib.c):
4665 Curves are EC_GROUP objects (with an optional group generator)
4666 based on EC_METHODs that are built into the library.
4668 Points are EC_POINT objects based on EC_GROUP objects.
4670 Most of the framework would be able to handle curves over arbitrary
4671 finite fields, but as there are no obvious types for fields other
4672 than GF(p), some functions are limited to that for now.
4675 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4676 that the file contains a complete HTTP response.
4679 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4680 change the def and num file printf format specifier from "%-40sXXX"
4681 to "%-39s XXX". The latter will always guarantee a space after the
4682 field while the former will cause them to run together if the field
4683 is 40 of more characters long.
4686 *) Constify the cipher and digest 'method' functions and structures
4687 and modify related functions to take constant EVP_MD and EVP_CIPHER
4691 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4692 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4695 *) Modify EVP_Digest*() routines so they now return values. Although the
4696 internal software routines can never fail additional hardware versions
4700 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4702 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4703 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4706 ERR_R_NESTED_ASN1_ERROR
4708 ERR_R_MISSING_ASN1_EOS
4709 were 4 .. 9, conflicting with
4710 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4712 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4713 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4715 Add new error code 'ERR_R_INTERNAL_ERROR'.
4718 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4722 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4723 sets the subject name for a new request or supersedes the
4724 subject name in a given request. Formats that can be parsed are
4725 'CN=Some Name, OU=myOU, C=IT'
4727 'CN=Some Name/OU=myOU/C=IT'.
4729 Add options '-batch' and '-verbose' to 'openssl req'.
4730 [Massimiliano Pala <madwolf@hackmasters.net>]
4732 *) Introduce the possibility to access global variables through
4733 functions on platform were that's the best way to handle exporting
4734 global variables in shared libraries. To enable this functionality,
4735 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4736 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4737 is normally done by Configure or something similar).
4739 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4740 in the source file (foo.c) like this:
4742 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4743 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4745 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4746 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4748 OPENSSL_DECLARE_GLOBAL(int,foo);
4749 #define foo OPENSSL_GLOBAL_REF(foo)
4750 OPENSSL_DECLARE_GLOBAL(double,bar);
4751 #define bar OPENSSL_GLOBAL_REF(bar)
4753 The #defines are very important, and therefore so is including the
4754 header file everywhere where the defined globals are used.
4756 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4757 of ASN.1 items, but that structure is a bit different.
4759 The largest change is in util/mkdef.pl which has been enhanced with
4760 better and easier to understand logic to choose which symbols should
4761 go into the Windows .def files as well as a number of fixes and code
4762 cleanup (among others, algorithm keywords are now sorted
4763 lexicographically to avoid constant rewrites).
4766 *) In BN_div() keep a copy of the sign of 'num' before writing the
4767 result to 'rm' because if rm==num the value will be overwritten
4768 and produce the wrong result if 'num' is negative: this caused
4769 problems with BN_mod() and BN_nnmod().
4772 *) Function OCSP_request_verify(). This checks the signature on an
4773 OCSP request and verifies the signer certificate. The signer
4774 certificate is just checked for a generic purpose and OCSP request
4778 *) Add OCSP_check_validity() function to check the validity of OCSP
4779 responses. OCSP responses are prepared in real time and may only
4780 be a few seconds old. Simply checking that the current time lies
4781 between thisUpdate and nextUpdate max reject otherwise valid responses
4782 caused by either OCSP responder or client clock inaccuracy. Instead
4783 we allow thisUpdate and nextUpdate to fall within a certain period of
4784 the current time. The age of the response can also optionally be
4785 checked. Two new options -validity_period and -status_age added to
4789 *) If signature or public key algorithm is unrecognized print out its
4790 OID rather that just UNKNOWN.
4793 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4794 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4795 ID to be generated from the issuer certificate alone which can then be
4796 passed to OCSP_id_issuer_cmp().
4799 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4800 ASN1 modules to export functions returning ASN1_ITEM pointers
4801 instead of the ASN1_ITEM structures themselves. This adds several
4802 new macros which allow the underlying ASN1 function/structure to
4803 be accessed transparently. As a result code should not use ASN1_ITEM
4804 references directly (such as &X509_it) but instead use the relevant
4805 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4806 use of the new ASN1 code on platforms where exporting structures
4807 is problematical (for example in shared libraries) but exporting
4808 functions returning pointers to structures is not.
4811 *) Add support for overriding the generation of SSL/TLS session IDs.
4812 These callbacks can be registered either in an SSL_CTX or per SSL.
4813 The purpose of this is to allow applications to control, if they wish,
4814 the arbitrary values chosen for use as session IDs, particularly as it
4815 can be useful for session caching in multiple-server environments. A
4816 command-line switch for testing this (and any client code that wishes
4817 to use such a feature) has been added to "s_server".
4818 [Geoff Thorpe, Lutz Jaenicke]
4820 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4821 of the form '#if defined(...) || defined(...) || ...' and
4822 '#if !defined(...) && !defined(...) && ...'. This also avoids
4823 the growing number of special cases it was previously handling.
4826 *) Make all configuration macros available for application by making
4827 sure they are available in opensslconf.h, by giving them names starting
4828 with "OPENSSL_" to avoid conflicts with other packages and by making
4829 sure e_os2.h will cover all platform-specific cases together with
4831 Additionally, it is now possible to define configuration/platform-
4832 specific names (called "system identities"). In the C code, these
4833 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4834 macro with the name beginning with "OPENSSL_SYS_", which is determined
4835 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4839 *) New option -set_serial to 'req' and 'x509' this allows the serial
4840 number to use to be specified on the command line. Previously self
4841 signed certificates were hard coded with serial number 0 and the
4842 CA options of 'x509' had to use a serial number in a file which was
4846 *) New options to 'ca' utility to support V2 CRL entry extensions.
4847 Currently CRL reason, invalidity date and hold instruction are
4848 supported. Add new CRL extensions to V3 code and some new objects.
4851 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4852 disable standard block padding (aka PKCS#5 padding) in the EVP
4853 API, which was previously mandatory. This means that the data is
4854 not padded in any way and so the total length much be a multiple
4855 of the block size, otherwise an error occurs.
4858 *) Initial (incomplete) OCSP SSL support.
4861 *) New function OCSP_parse_url(). This splits up a URL into its host,
4862 port and path components: primarily to parse OCSP URLs. New -url
4863 option to ocsp utility.
4866 *) New nonce behavior. The return value of OCSP_check_nonce() now
4867 reflects the various checks performed. Applications can decide
4868 whether to tolerate certain situations such as an absent nonce
4869 in a response when one was present in a request: the ocsp application
4870 just prints out a warning. New function OCSP_add1_basic_nonce()
4871 this is to allow responders to include a nonce in a response even if
4872 the request is nonce-less.
4875 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4876 skipped when using openssl x509 multiple times on a single input file,
4877 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4880 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4881 set string type: to handle setting ASN1_TIME structures. Fix ca
4882 utility to correctly initialize revocation date of CRLs.
4885 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4886 the clients preferred ciphersuites and rather use its own preferences.
4887 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4888 Internet Explorer by ensuring unchanged hash method during stepup.
4889 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4892 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4893 to aes and add a new 'exist' option to print out symbols that don't
4897 *) Additional options to ocsp utility to allow flags to be set and
4898 additional certificates supplied.
4901 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4902 OCSP client a number of certificate to only verify the response
4906 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4907 handle the new API. Currently only ECB, CBC modes supported. Add new
4910 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4911 Encryption Standard (AES) Ciphersuites for Transport Layer
4912 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4913 not enabled by default and were not part of the "ALL" ciphersuite
4914 alias because they were not yet official; they could be
4915 explicitly requested by specifying the "AESdraft" ciphersuite
4916 group alias. In the final release of OpenSSL 0.9.7, the group
4917 alias is called "AES" and is part of "ALL".)
4918 [Ben Laurie, Steve Henson, Bodo Moeller]
4920 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4921 request to response.
4924 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4925 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4926 extract information from a certificate request. OCSP_response_create()
4927 creates a response and optionally adds a basic response structure.
4928 OCSP_basic_add1_status() adds a complete single response to a basic
4929 response and returns the OCSP_SINGLERESP structure just added (to allow
4930 extensions to be included for example). OCSP_basic_add1_cert() adds a
4931 certificate to a basic response and OCSP_basic_sign() signs a basic
4932 response with various flags. New helper functions ASN1_TIME_check()
4933 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4934 (converts ASN1_TIME to GeneralizedTime).
4937 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4938 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4939 structure from a certificate. X509_pubkey_digest() digests the public_key
4940 contents: this is used in various key identifiers.
4943 *) Make sk_sort() tolerate a NULL argument.
4944 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4946 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4947 passed by the function are trusted implicitly. If any of them signed the
4948 response then it is assumed to be valid and is not verified.
4951 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4952 to data. This was previously part of the PKCS7 ASN1 code. This
4953 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4954 [Steve Henson, reported by Kenneth R. Robinette
4955 <support@securenetterm.com>]
4957 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4958 routines: without these tracing memory leaks is very painful.
4959 Fix leaks in PKCS12 and PKCS7 routines.
4962 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4963 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4964 effectively meant GeneralizedTime would never be used. Now it
4965 is initialised to -1 but X509_time_adj() now has to check the value
4966 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4967 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4968 [Steve Henson, reported by Kenneth R. Robinette
4969 <support@securenetterm.com>]
4971 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4972 result in a zero length in the ASN1_INTEGER structure which was
4973 not consistent with the structure when d2i_ASN1_INTEGER() was used
4974 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4975 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4976 where it did not print out a minus for negative ASN1_INTEGER.
4979 *) Add summary printout to ocsp utility. The various functions which
4980 convert status values to strings have been renamed to:
4981 OCSP_response_status_str(), OCSP_cert_status_str() and
4982 OCSP_crl_reason_str() and are no longer static. New options
4983 to verify nonce values and to disable verification. OCSP response
4984 printout format cleaned up.
4987 *) Add additional OCSP certificate checks. These are those specified
4988 in RFC2560. This consists of two separate checks: the CA of the
4989 certificate being checked must either be the OCSP signer certificate
4990 or the issuer of the OCSP signer certificate. In the latter case the
4991 OCSP signer certificate must contain the OCSP signing extended key
4992 usage. This check is performed by attempting to match the OCSP
4993 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4994 in the OCSP_CERTID structures of the response.
4997 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4998 and related routines. This uses the standard OpenSSL certificate
4999 verify routines to perform initial checks (just CA validity) and
5000 to obtain the certificate chain. Then additional checks will be
5001 performed on the chain. Currently the root CA is checked to see
5002 if it is explicitly trusted for OCSP signing. This is used to set
5003 a root CA as a global signing root: that is any certificate that
5004 chains to that CA is an acceptable OCSP signing certificate.
5007 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5008 extensions from a separate configuration file.
5009 As when reading extensions from the main configuration file,
5010 the '-extensions ...' option may be used for specifying the
5012 [Massimiliano Pala <madwolf@comune.modena.it>]
5014 *) New OCSP utility. Allows OCSP requests to be generated or
5015 read. The request can be sent to a responder and the output
5016 parsed, outputed or printed in text form. Not complete yet:
5017 still needs to check the OCSP response validity.
5020 *) New subcommands for 'openssl ca':
5021 'openssl ca -status <serial>' prints the status of the cert with
5022 the given serial number (according to the index file).
5023 'openssl ca -updatedb' updates the expiry status of certificates
5025 [Massimiliano Pala <madwolf@comune.modena.it>]
5027 *) New '-newreq-nodes' command option to CA.pl. This is like
5028 '-newreq', but calls 'openssl req' with the '-nodes' option
5029 so that the resulting key is not encrypted.
5030 [Damien Miller <djm@mindrot.org>]
5032 *) New configuration for the GNU Hurd.
5033 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5035 *) Initial code to implement OCSP basic response verify. This
5036 is currently incomplete. Currently just finds the signer's
5037 certificate and verifies the signature on the response.
5040 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5041 value of OPENSSLDIR. This is available via the new '-d' option
5042 to 'openssl version', and is also included in 'openssl version -a'.
5045 *) Allowing defining memory allocation callbacks that will be given
5046 file name and line number information in additional arguments
5047 (a const char* and an int). The basic functionality remains, as
5048 well as the original possibility to just replace malloc(),
5049 realloc() and free() by functions that do not know about these
5050 additional arguments. To register and find out the current
5051 settings for extended allocation functions, the following
5052 functions are provided:
5054 CRYPTO_set_mem_ex_functions
5055 CRYPTO_set_locked_mem_ex_functions
5056 CRYPTO_get_mem_ex_functions
5057 CRYPTO_get_locked_mem_ex_functions
5059 These work the same way as CRYPTO_set_mem_functions and friends.
5060 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5061 extended allocation function is enabled.
5062 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5063 a conventional allocation function is enabled.
5064 [Richard Levitte, Bodo Moeller]
5066 *) Finish off removing the remaining LHASH function pointer casts.
5067 There should no longer be any prototype-casting required when using
5068 the LHASH abstraction, and any casts that remain are "bugs". See
5069 the callback types and macros at the head of lhash.h for details
5070 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5073 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5074 If /dev/[u]random devices are not available or do not return enough
5075 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5077 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5078 /etc/entropy will be queried once each in this sequence, quering stops
5079 when enough entropy was collected without querying more sockets.
5082 *) Change the Unix RAND_poll() variant to be able to poll several
5083 random devices, as specified by DEVRANDOM, until a sufficient amount
5084 of data has been collected. We spend at most 10 ms on each file
5085 (select timeout) and read in non-blocking mode. DEVRANDOM now
5086 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5087 (previously it was just the string "/dev/urandom"), so on typical
5088 platforms the 10 ms delay will never occur.
5089 Also separate out the Unix variant to its own file, rand_unix.c.
5090 For VMS, there's a currently-empty rand_vms.c.
5093 *) Move OCSP client related routines to ocsp_cl.c. These
5094 provide utility functions which an application needing
5095 to issue a request to an OCSP responder and analyse the
5096 response will typically need: as opposed to those which an
5097 OCSP responder itself would need which will be added later.
5099 OCSP_request_sign() signs an OCSP request with an API similar
5100 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5101 response. OCSP_response_get1_basic() extracts basic response
5102 from response. OCSP_resp_find_status(): finds and extracts status
5103 information from an OCSP_CERTID structure (which will be created
5104 when the request structure is built). These are built from lower
5105 level functions which work on OCSP_SINGLERESP structures but
5106 wont normally be used unless the application wishes to examine
5107 extensions in the OCSP response for example.
5109 Replace nonce routines with a pair of functions.
5110 OCSP_request_add1_nonce() adds a nonce value and optionally
5111 generates a random value. OCSP_check_nonce() checks the
5112 validity of the nonce in an OCSP response.
5115 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5116 This doesn't copy the supplied OCSP_CERTID and avoids the
5117 need to free up the newly created id. Change return type
5118 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5119 This can then be used to add extensions to the request.
5120 Deleted OCSP_request_new(), since most of its functionality
5121 is now in OCSP_REQUEST_new() (and the case insensitive name
5122 clash) apart from the ability to set the request name which
5123 will be added elsewhere.
5126 *) Update OCSP API. Remove obsolete extensions argument from
5127 various functions. Extensions are now handled using the new
5128 OCSP extension code. New simple OCSP HTTP function which
5129 can be used to send requests and parse the response.
5132 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5133 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5134 uses the special reorder version of SET OF to sort the attributes
5135 and reorder them to match the encoded order. This resolves a long
5136 standing problem: a verify on a PKCS7 structure just after signing
5137 it used to fail because the attribute order did not match the
5138 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5139 it uses the received order. This is necessary to tolerate some broken
5140 software that does not order SET OF. This is handled by encoding
5141 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5142 to produce the required SET OF.
5145 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5146 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5147 files to get correct declarations of the ASN.1 item variables.
5150 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5151 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5152 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5153 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5154 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5155 ASN1_ITEM and no wrapper functions.
5158 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5159 replace the old function pointer based I/O routines. Change most of
5160 the *_d2i_bio() and *_d2i_fp() functions to use these.
5163 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5164 lines, recognice more "algorithms" that can be deselected, and make
5165 it complain about algorithm deselection that isn't recognised.
5168 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5169 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5170 to use new functions. Add NO_ASN1_OLD which can be set to remove
5171 some old style ASN1 functions: this can be used to determine if old
5172 code will still work when these eventually go away.
5175 *) New extension functions for OCSP structures, these follow the
5176 same conventions as certificates and CRLs.
5179 *) New function X509V3_add1_i2d(). This automatically encodes and
5180 adds an extension. Its behaviour can be customised with various
5181 flags to append, replace or delete. Various wrappers added for
5182 certifcates and CRLs.
5185 *) Fix to avoid calling the underlying ASN1 print routine when
5186 an extension cannot be parsed. Correct a typo in the
5187 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5190 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5191 entries for variables.
5194 *) Add functionality to apps/openssl.c for detecting locking
5195 problems: As the program is single-threaded, all we have
5196 to do is register a locking callback using an array for
5197 storing which locks are currently held by the program.
5200 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5201 SSL_get_ex_data_X509_STORE_idx(), which is used in
5202 ssl_verify_cert_chain() and thus can be called at any time
5203 during TLS/SSL handshakes so that thread-safety is essential.
5204 Unfortunately, the ex_data design is not at all suited
5205 for multi-threaded use, so it probably should be abolished.
5208 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5209 [Broadcom, tweaked and integrated by Geoff Thorpe]
5211 *) Move common extension printing code to new function
5212 X509V3_print_extensions(). Reorganise OCSP print routines and
5213 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5216 *) New function X509_signature_print() to remove duplication in some
5220 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5221 set (this was treated exactly the same as SET OF previously). This
5222 is used to reorder the STACK representing the structure to match the
5223 encoding. This will be used to get round a problem where a PKCS7
5224 structure which was signed could not be verified because the STACK
5225 order did not reflect the encoded order.
5228 *) Reimplement the OCSP ASN1 module using the new code.
5231 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5232 for its ASN1 operations. The old style function pointers still exist
5233 for now but they will eventually go away.
5236 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5237 completely replaces the old ASN1 functionality with a table driven
5238 encoder and decoder which interprets an ASN1_ITEM structure describing
5239 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5240 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5241 has also been converted to the new form.
5244 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5245 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5246 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5247 for negative moduli.
5250 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5251 of not touching the result's sign bit.
5254 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5258 *) Changed the LHASH code to use prototypes for callbacks, and created
5259 macros to declare and implement thin (optionally static) functions
5260 that provide type-safety and avoid function pointer casting for the
5261 type-specific callbacks.
5264 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5266 [Veers Staats <staatsvr@asc.hpc.mil>,
5267 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5269 *) Reformat the FAQ so the different questions and answers can be divided
5270 in sections depending on the subject.
5273 *) Have the zlib compression code load ZLIB.DLL dynamically under
5277 *) New function BN_mod_sqrt for computing square roots modulo a prime
5278 (using the probabilistic Tonelli-Shanks algorithm unless
5279 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5280 be handled deterministically).
5281 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5283 *) Make BN_mod_inverse faster by explicitly handling small quotients
5284 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5285 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5288 *) New function BN_kronecker.
5291 *) Fix BN_gcd so that it works on negative inputs; the result is
5292 positive unless both parameters are zero.
5293 Previously something reasonably close to an infinite loop was
5294 possible because numbers could be growing instead of shrinking
5295 in the implementation of Euclid's algorithm.
5298 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5299 sign of the number in question.
5301 Fix BN_is_word(a,w) to work correctly for w == 0.
5303 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5304 because its test if the absolute value of 'a' equals 'w'.
5305 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5306 it exists mostly for use in the implementations of BN_is_zero(),
5307 BN_is_one(), and BN_is_word().
5310 *) New function BN_swap.
5313 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5314 the exponentiation functions are more likely to produce reasonable
5315 results on negative inputs.
5318 *) Change BN_mod_mul so that the result is always non-negative.
5319 Previously, it could be negative if one of the factors was negative;
5320 I don't think anyone really wanted that behaviour.
5323 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5324 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5325 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5326 and add new functions:
5335 BN_mod_lshift1_quick
5339 These functions always generate non-negative results.
5341 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5342 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5344 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5345 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5346 be reduced modulo m.
5347 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5350 The following entry accidentily appeared in the CHANGES file
5351 distributed with OpenSSL 0.9.7. The modifications described in
5352 it do *not* apply to OpenSSL 0.9.7.
5354 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5355 was actually never needed) and in BN_mul(). The removal in BN_mul()
5356 required a small change in bn_mul_part_recursive() and the addition
5357 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5358 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5359 bn_sub_words() and bn_add_words() except they take arrays with
5364 *) In 'openssl passwd', verify passwords read from the terminal
5365 unless the '-salt' option is used (which usually means that
5366 verification would just waste user's time since the resulting
5367 hash is going to be compared with some given password hash)
5368 or the new '-noverify' option is used.
5370 This is an incompatible change, but it does not affect
5371 non-interactive use of 'openssl passwd' (passwords on the command
5372 line, '-stdin' option, '-in ...' option) and thus should not
5376 *) Remove all references to RSAref, since there's no more need for it.
5379 *) Make DSO load along a path given through an environment variable
5380 (SHLIB_PATH) with shl_load().
5383 *) Constify the ENGINE code as a result of BIGNUM constification.
5384 Also constify the RSA code and most things related to it. In a
5385 few places, most notable in the depth of the ASN.1 code, ugly
5386 casts back to non-const were required (to be solved at a later
5390 *) Make it so the openssl application has all engines loaded by default.
5393 *) Constify the BIGNUM routines a little more.
5396 *) Add the following functions:
5398 ENGINE_load_cswift()
5400 ENGINE_load_atalla()
5402 ENGINE_load_builtin_engines()
5404 That way, an application can itself choose if external engines that
5405 are built-in in OpenSSL shall ever be used or not. The benefit is
5406 that applications won't have to be linked with libdl or other dso
5407 libraries unless it's really needed.
5409 Changed 'openssl engine' to load all engines on demand.
5410 Changed the engine header files to avoid the duplication of some
5411 declarations (they differed!).
5414 *) 'openssl engine' can now list capabilities.
5417 *) Better error reporting in 'openssl engine'.
5420 *) Never call load_dh_param(NULL) in s_server.
5423 *) Add engine application. It can currently list engines by name and
5424 identity, and test if they are actually available.
5427 *) Improve RPM specification file by forcing symbolic linking and making
5428 sure the installed documentation is also owned by root.root.
5429 [Damien Miller <djm@mindrot.org>]
5431 *) Give the OpenSSL applications more possibilities to make use of
5432 keys (public as well as private) handled by engines.
5435 *) Add OCSP code that comes from CertCo.
5438 *) Add VMS support for the Rijndael code.
5441 *) Added untested support for Nuron crypto accelerator.
5444 *) Add support for external cryptographic devices. This code was
5445 previously distributed separately as the "engine" branch.
5446 [Geoff Thorpe, Richard Levitte]
5448 *) Rework the filename-translation in the DSO code. It is now possible to
5449 have far greater control over how a "name" is turned into a filename
5450 depending on the operating environment and any oddities about the
5451 different shared library filenames on each system.
5454 *) Support threads on FreeBSD-elf in Configure.
5457 *) Fix for SHA1 assembly problem with MASM: it produces
5458 warnings about corrupt line number information when assembling
5459 with debugging information. This is caused by the overlapping
5461 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5464 NCONF_get_number() has no error checking at all. As a replacement,
5465 NCONF_get_number_e() is defined (_e for "error checking") and is
5466 promoted strongly. The old NCONF_get_number is kept around for
5467 binary backward compatibility.
5468 Make it possible for methods to load from something other than a BIO,
5469 by providing a function pointer that is given a name instead of a BIO.
5470 For example, this could be used to load configuration data from an
5474 *) Fix for non blocking accept BIOs. Added new I/O special reason
5475 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5476 with non blocking I/O was not possible because no retry code was
5477 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5481 *) Added the beginnings of Rijndael support.
5484 *) Fix for bug in DirectoryString mask setting. Add support for
5485 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5486 to allow certificate printing to more controllable, additional
5487 'certopt' option to 'x509' to allow new printing options to be
5491 *) Clean old EAY MD5 hack from e_os.h.
5494 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5496 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5497 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5498 [Joe Orton, Steve Henson]
5500 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5502 *) Fix additional bug revealed by the NISCC test suite:
5504 Stop bug triggering large recursion when presented with
5505 certain ASN.1 tags (CVE-2003-0851)
5508 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5510 *) Fix various bugs revealed by running the NISCC test suite:
5512 Stop out of bounds reads in the ASN1 code when presented with
5513 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5515 If verify callback ignores invalid public key errors don't try to check
5516 certificate signature with the NULL public key.
5520 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5521 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5525 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5526 extra data after the compression methods not only for TLS 1.0
5527 but also for SSL 3.0 (as required by the specification).
5528 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5530 *) Change X509_certificate_type() to mark the key as exported/exportable
5531 when it's 512 *bits* long, not 512 bytes.
5534 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5536 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5537 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5538 a protocol version number mismatch like a decryption error
5539 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5542 *) Turn on RSA blinding by default in the default implementation
5543 to avoid a timing attack. Applications that don't want it can call
5544 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5545 They would be ill-advised to do so in most cases.
5546 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5548 *) Change RSA blinding code so that it works when the PRNG is not
5549 seeded (in this case, the secret RSA exponent is abused as
5550 an unpredictable seed -- if it is not unpredictable, there
5551 is no point in blinding anyway). Make RSA blinding thread-safe
5552 by remembering the creator's thread ID in rsa->blinding and
5553 having all other threads use local one-time blinding factors
5554 (this requires more computation than sharing rsa->blinding, but
5555 avoids excessive locking; and if an RSA object is not shared
5556 between threads, blinding will still be very fast).
5559 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5561 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5562 via timing by performing a MAC computation even if incorrrect
5563 block cipher padding has been found. This is a countermeasure
5564 against active attacks where the attacker has to distinguish
5565 between bad padding and a MAC verification error. (CVE-2003-0078)
5567 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5568 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5569 Martin Vuagnoux (EPFL, Ilion)]
5571 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5573 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5574 memory from it's contents. This is done with a counter that will
5575 place alternating values in each byte. This can be used to solve
5576 two issues: 1) the removal of calls to memset() by highly optimizing
5577 compilers, and 2) cleansing with other values than 0, since those can
5578 be read through on certain media, for example a swap space on disk.
5581 *) Bugfix: client side session caching did not work with external caching,
5582 because the session->cipher setting was not restored when reloading
5583 from the external cache. This problem was masked, when
5584 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5585 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5588 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5589 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5590 [Zeev Lieber <zeev-l@yahoo.com>]
5592 *) Undo an undocumented change introduced in 0.9.6e which caused
5593 repeated calls to OpenSSL_add_all_ciphers() and
5594 OpenSSL_add_all_digests() to be ignored, even after calling
5598 *) Change the default configuration reader to deal with last line not
5599 being properly terminated.
5602 *) Change X509_NAME_cmp() so it applies the special rules on handling
5603 DN values that are of type PrintableString, as well as RDNs of type
5604 emailAddress where the value has the type ia5String.
5605 [stefank@valicert.com via Richard Levitte]
5607 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5608 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5609 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5610 the bitwise-OR of the two for use by the majority of applications
5611 wanting this behaviour, and update the docs. The documented
5612 behaviour and actual behaviour were inconsistent and had been
5613 changing anyway, so this is more a bug-fix than a behavioural
5615 [Geoff Thorpe, diagnosed by Nadav Har'El]
5617 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5618 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5621 *) Fix initialization code race conditions in
5622 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5623 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5624 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5625 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5626 ssl2_get_cipher_by_char(),
5627 ssl3_get_cipher_by_char().
5628 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5630 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5631 the cached sessions are flushed, as the remove_cb() might use ex_data
5632 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5633 (see [openssl.org #212]).
5634 [Geoff Thorpe, Lutz Jaenicke]
5636 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5637 length, instead of the encoding length to d2i_ASN1_OBJECT.
5640 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5642 *) [In 0.9.6g-engine release:]
5643 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5644 [Lynn Gazis <lgazis@rainbow.com>]
5646 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5648 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5649 and get fix the header length calculation.
5650 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5651 Alon Kantor <alonk@checkpoint.com> (and others),
5654 *) Use proper error handling instead of 'assertions' in buffer
5655 overflow checks added in 0.9.6e. This prevents DoS (the
5656 assertions could call abort()).
5657 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5659 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5661 *) Add various sanity checks to asn1_get_length() to reject
5662 the ASN1 length bytes if they exceed sizeof(long), will appear
5663 negative or the content length exceeds the length of the
5665 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5667 *) Fix cipher selection routines: ciphers without encryption had no flags
5668 for the cipher strength set and where therefore not handled correctly
5669 by the selection routines (PR #130).
5672 *) Fix EVP_dsa_sha macro.
5676 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5677 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5678 that was added in OpenSSL 0.9.6d.
5680 As the countermeasure turned out to be incompatible with some
5681 broken SSL implementations, the new option is part of SSL_OP_ALL.
5682 SSL_OP_ALL is usually employed when compatibility with weird SSL
5683 implementations is desired (e.g. '-bugs' option to 's_client' and
5684 's_server'), so the new option is automatically set in many
5688 *) Changes in security patch:
5690 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5691 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5692 Air Force Materiel Command, USAF, under agreement number
5695 *) Add various sanity checks to asn1_get_length() to reject
5696 the ASN1 length bytes if they exceed sizeof(long), will appear
5697 negative or the content length exceeds the length of the
5698 supplied buffer. (CVE-2002-0659)
5699 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5701 *) Assertions for various potential buffer overflows, not known to
5703 [Ben Laurie (CHATS)]
5705 *) Various temporary buffers to hold ASCII versions of integers were
5706 too small for 64 bit platforms. (CVE-2002-0655)
5707 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5709 *) Remote buffer overflow in SSL3 protocol - an attacker could
5710 supply an oversized session ID to a client. (CVE-2002-0656)
5711 [Ben Laurie (CHATS)]
5713 *) Remote buffer overflow in SSL2 protocol - an attacker could
5714 supply an oversized client master key. (CVE-2002-0656)
5715 [Ben Laurie (CHATS)]
5717 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5719 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5720 encoded as NULL) with id-dsa-with-sha1.
5721 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5723 *) Check various X509_...() return values in apps/req.c.
5724 [Nils Larsch <nla@trustcenter.de>]
5726 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5727 an end-of-file condition would erronously be flagged, when the CRLF
5728 was just at the end of a processed block. The bug was discovered when
5729 processing data through a buffering memory BIO handing the data to a
5730 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5731 <ptsekov@syntrex.com> and Nedelcho Stanev.
5734 *) Implement a countermeasure against a vulnerability recently found
5735 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5736 before application data chunks to avoid the use of known IVs
5737 with data potentially chosen by the attacker.
5740 *) Fix length checks in ssl3_get_client_hello().
5743 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5744 to prevent ssl3_read_internal() from incorrectly assuming that
5745 ssl3_read_bytes() found application data while handshake
5746 processing was enabled when in fact s->s3->in_read_app_data was
5747 merely automatically cleared during the initial handshake.
5748 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5750 *) Fix object definitions for Private and Enterprise: they were not
5751 recognized in their shortname (=lowercase) representation. Extend
5752 obj_dat.pl to issue an error when using undefined keywords instead
5753 of silently ignoring the problem (Svenning Sorensen
5754 <sss@sss.dnsalias.net>).
5757 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5758 generators, i.e. generators other than 2 and 5. (Previously, the
5759 code did not properly initialise the 'add' and 'rem' values to
5760 BN_generate_prime().)
5762 In the new general case, we do not insist that 'generator' is
5763 actually a primitive root: This requirement is rather pointless;
5764 a generator of the order-q subgroup is just as good, if not
5768 *) Map new X509 verification errors to alerts. Discovered and submitted by
5769 Tom Wu <tom@arcot.com>.
5772 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5773 returning non-zero before the data has been completely received
5774 when using non-blocking I/O.
5775 [Bodo Moeller; problem pointed out by John Hughes]
5777 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5778 [Ben Laurie, Lutz Jaenicke]
5780 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5781 Yoram Zahavi <YoramZ@gilian.com>).
5784 *) Add information about CygWin 1.3 and on, and preserve proper
5785 configuration for the versions before that.
5786 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5788 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5789 check whether we deal with a copy of a session and do not delete from
5790 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5791 <izhar@checkpoint.com>.
5794 *) Do not store session data into the internal session cache, if it
5795 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5796 flag is set). Proposed by Aslam <aslam@funk.com>.
5799 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5803 *) [In 0.9.6d-engine release:]
5804 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5805 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5807 *) Add the configuration target linux-s390x.
5808 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5810 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5811 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5812 variable as an indication that a ClientHello message has been
5813 received. As the flag value will be lost between multiple
5814 invocations of ssl3_accept when using non-blocking I/O, the
5815 function may not be aware that a handshake has actually taken
5816 place, thus preventing a new session from being added to the
5819 To avoid this problem, we now set s->new_session to 2 instead of
5820 using a local variable.
5821 [Lutz Jaenicke, Bodo Moeller]
5823 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5824 if the SSL_R_LENGTH_MISMATCH error is detected.
5825 [Geoff Thorpe, Bodo Moeller]
5827 *) New 'shared_ldflag' column in Configure platform table.
5830 *) Fix EVP_CIPHER_mode macro.
5831 ["Dan S. Camper" <dan@bti.net>]
5833 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5834 type, we must throw them away by setting rr->length to 0.
5835 [D P Chang <dpc@qualys.com>]
5837 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5839 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5840 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5841 worked incorrectly for those cases where range = 10..._2 and
5842 3*range is two bits longer than range.)
5845 *) Only add signing time to PKCS7 structures if it is not already
5849 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5850 OBJ_ld_ce should be OBJ_id_ce.
5851 Also some ip-pda OIDs in crypto/objects/objects.txt were
5852 incorrect (cf. RFC 3039).
5853 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5855 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5856 returns early because it has nothing to do.
5857 [Andy Schneider <andy.schneider@bjss.co.uk>]
5859 *) [In 0.9.6c-engine release:]
5860 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5861 [Andy Schneider <andy.schneider@bjss.co.uk>]
5863 *) [In 0.9.6c-engine release:]
5864 Add support for Cryptographic Appliance's keyserver technology.
5865 (Use engine 'keyclient')
5866 [Cryptographic Appliances and Geoff Thorpe]
5868 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5869 is called via tools/c89.sh because arguments have to be
5870 rearranged (all '-L' options must appear before the first object
5872 [Richard Shapiro <rshapiro@abinitio.com>]
5874 *) [In 0.9.6c-engine release:]
5875 Add support for Broadcom crypto accelerator cards, backported
5877 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5879 *) [In 0.9.6c-engine release:]
5880 Add support for SureWare crypto accelerator cards from
5881 Baltimore Technologies. (Use engine 'sureware')
5882 [Baltimore Technologies and Mark Cox]
5884 *) [In 0.9.6c-engine release:]
5885 Add support for crypto accelerator cards from Accelerated
5886 Encryption Processing, www.aep.ie. (Use engine 'aep')
5887 [AEP Inc. and Mark Cox]
5889 *) Add a configuration entry for gcc on UnixWare.
5890 [Gary Benson <gbenson@redhat.com>]
5892 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5893 messages are stored in a single piece (fixed-length part and
5894 variable-length part combined) and fix various bugs found on the way.
5897 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5898 instead. BIO_gethostbyname() does not know what timeouts are
5899 appropriate, so entries would stay in cache even when they have
5901 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5903 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5904 faced with a pathologically small ClientHello fragment that does
5905 not contain client_version: Instead of aborting with an error,
5906 simply choose the highest available protocol version (i.e.,
5907 TLS 1.0 unless it is disabled). In practice, ClientHello
5908 messages are never sent like this, but this change gives us
5909 strictly correct behaviour at least for TLS.
5912 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5913 never resets s->method to s->ctx->method when called from within
5914 one of the SSL handshake functions.
5915 [Bodo Moeller; problem pointed out by Niko Baric]
5917 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5918 (sent using the client's version number) if client_version is
5919 smaller than the protocol version in use. Also change
5920 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5921 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5922 the client will at least see that alert.
5925 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5929 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5930 client receives HelloRequest while in a handshake.
5931 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5933 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5934 should end in 'break', not 'goto end' which circuments various
5935 cleanups done in state SSL_ST_OK. But session related stuff
5936 must be disabled for SSL_ST_OK in the case that we just sent a
5939 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5940 before just sending a HelloRequest.
5941 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5943 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5944 reveal whether illegal block cipher padding was found or a MAC
5945 verification error occured. (Neither SSLerr() codes nor alerts
5946 are directly visible to potential attackers, but the information
5947 may leak via logfiles.)
5949 Similar changes are not required for the SSL 2.0 implementation
5950 because the number of padding bytes is sent in clear for SSL 2.0,
5951 and the extra bytes are just ignored. However ssl/s2_pkt.c
5952 failed to verify that the purported number of padding bytes is in
5956 *) Add OpenUNIX-8 support including shared libraries
5957 (Boyd Lynn Gerber <gerberb@zenez.com>).
5960 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5961 'wristwatch attack' using huge encoding parameters (cf.
5962 James H. Manger's CRYPTO 2001 paper). Note that the
5963 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5964 encoding parameters and hence was not vulnerable.
5967 *) BN_sqr() bug fix.
5968 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5970 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5971 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5972 followed by modular reduction.
5973 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5975 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5976 equivalent based on BN_pseudo_rand() instead of BN_rand().
5979 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5980 This function was broken, as the check for a new client hello message
5981 to handle SGC did not allow these large messages.
5982 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5985 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5988 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5989 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5992 *) Rework the configuration and shared library support for Tru64 Unix.
5993 The configuration part makes use of modern compiler features and
5994 still retains old compiler behavior for those that run older versions
5995 of the OS. The shared library support part includes a variant that
5996 uses the RPATH feature, and is available through the special
5997 configuration target "alpha-cc-rpath", which will never be selected
5999 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6001 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6002 with the same message size as in ssl3_get_certificate_request().
6003 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6004 messages might inadvertently be reject as too long.
6005 [Petr Lampa <lampa@fee.vutbr.cz>]
6007 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6010 *) Modified SSL library such that the verify_callback that has been set
6011 specificly for an SSL object with SSL_set_verify() is actually being
6012 used. Before the change, a verify_callback set with this function was
6013 ignored and the verify_callback() set in the SSL_CTX at the time of
6014 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6015 to allow the necessary settings.
6018 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6019 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6020 done automatically (in contradiction to the requirements of the C
6021 standard). This made problems when used from OpenSSH.
6024 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6025 dh->length and always used
6027 BN_rand_range(priv_key, dh->p).
6029 BN_rand_range() is not necessary for Diffie-Hellman, and this
6030 specific range makes Diffie-Hellman unnecessarily inefficient if
6031 dh->length (recommended exponent length) is much smaller than the
6032 length of dh->p. We could use BN_rand_range() if the order of
6033 the subgroup was stored in the DH structure, but we only have
6038 BN_rand(priv_key, l, ...)
6040 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6046 RSA_eay_public_encrypt
6047 RSA_eay_private_decrypt
6048 RSA_eay_private_encrypt (signing)
6049 RSA_eay_public_decrypt (signature verification)
6051 (default implementations for RSA_public_encrypt,
6052 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6053 always reject numbers >= n.
6056 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6057 to synchronize access to 'locking_thread'. This is necessary on
6058 systems where access to 'locking_thread' (an 'unsigned long'
6059 variable) is not atomic.
6062 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6063 *before* setting the 'crypto_lock_rand' flag. The previous code had
6064 a race condition if 0 is a valid thread ID.
6065 [Travis Vitek <vitek@roguewave.com>]
6067 *) Add support for shared libraries under Irix.
6068 [Albert Chin-A-Young <china@thewrittenword.com>]
6070 *) Add configuration option to build on Linux on both big-endian and
6072 [Ralf Baechle <ralf@uni-koblenz.de>]
6074 *) Add the possibility to create shared libraries on HP-UX.
6077 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6079 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6080 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6081 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6082 PRNG state recovery was possible based on the output of
6083 one PRNG request appropriately sized to gain knowledge on
6084 'md' followed by enough consecutive 1-byte PRNG requests
6085 to traverse all of 'state'.
6087 1. When updating 'md_local' (the current thread's copy of 'md')
6088 during PRNG output generation, hash all of the previous
6089 'md_local' value, not just the half used for PRNG output.
6091 2. Make the number of bytes from 'state' included into the hash
6092 independent from the number of PRNG bytes requested.
6094 The first measure alone would be sufficient to avoid
6095 Markku-Juhani's attack. (Actually it had never occurred
6096 to me that the half of 'md_local' used for chaining was the
6097 half from which PRNG output bytes were taken -- I had always
6098 assumed that the secret half would be used.) The second
6099 measure makes sure that additional data from 'state' is never
6100 mixed into 'md_local' in small portions; this heuristically
6101 further strengthens the PRNG.
6104 *) Fix crypto/bn/asm/mips3.s.
6107 *) When only the key is given to "enc", the IV is undefined. Print out
6108 an error message in this case.
6111 *) Handle special case when X509_NAME is empty in X509 printing routines.
6114 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6115 positive and less than q.
6118 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6119 used: it isn't thread safe and the add_lock_callback should handle
6121 [Paul Rose <Paul.Rose@bridge.com>]
6123 *) Verify that incoming data obeys the block size in
6124 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6128 [Ulf Möller, Bodo Möller]
6130 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6131 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6132 when fixing the server behaviour for backwards-compatible 'client
6133 hello' messages. (Note that the attack is impractical against
6134 SSL 3.0 and TLS 1.0 anyway because length and version checking
6135 means that the probability of guessing a valid ciphertext is
6136 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6139 Before 0.9.5, the countermeasure (hide the error by generating a
6140 random 'decryption result') did not work properly because
6141 ERR_clear_error() was missing, meaning that SSL_get_error() would
6142 detect the supposedly ignored error.
6144 Both problems are now fixed.
6147 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6148 (previously it was 1024).
6151 *) Fix for compatibility mode trust settings: ignore trust settings
6152 unless some valid trust or reject settings are present.
6155 *) Fix for blowfish EVP: its a variable length cipher.
6158 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6159 parameters in DSA public key structures and return an error in the
6160 DSA routines if parameters are absent.
6163 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6164 in the current directory if neither $RANDFILE nor $HOME was set.
6165 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6166 caused some confusion to Windows users who haven't defined $HOME.
6167 Thus RAND_file_name() is changed again: e_os.h can define a
6168 DEFAULT_HOME, which will be used if $HOME is not set.
6169 For Windows, we use "C:"; on other platforms, we still require
6170 environment variables.
6172 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6173 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6174 having multiple threads call RAND_poll() concurrently.
6177 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6178 combination of a flag and a thread ID variable.
6179 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6180 flag), *other* threads can enter ssleay_add_bytes without obeying
6181 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6182 that they do not hold after the first thread unsets add_do_not_lock).
6185 *) Change bctest again: '-x' expressions are not available in all
6189 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6191 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6192 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6194 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6195 the default extension for executables, if any. Also, make the perl
6196 scripts that use symlink() to test if it really exists and use "cp"
6197 if it doesn't. All this made OpenSSL compilable and installable in
6201 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6202 If SEQUENCE is length is indefinite just set c->slen to the total
6203 amount of data available.
6204 [Steve Henson, reported by shige@FreeBSD.org]
6205 [This change does not apply to 0.9.7.]
6207 *) Change bctest to avoid here-documents inside command substitution
6208 (workaround for FreeBSD /bin/sh bug).
6209 For compatibility with Ultrix, avoid shell functions (introduced
6210 in the bctest version that searches along $PATH).
6213 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6214 with des_encrypt() defined on some operating systems, like Solaris
6218 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6219 On the Importance of Eliminating Errors in Cryptographic
6220 Computations, J. Cryptology 14 (2001) 2, 101-119,
6221 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6224 *) MIPS assembler BIGNUM division bug fix.
6227 *) Disabled incorrect Alpha assembler code.
6230 *) Fix PKCS#7 decode routines so they correctly update the length
6231 after reading an EOC for the EXPLICIT tag.
6233 [This change does not apply to 0.9.7.]
6235 *) Fix bug in PKCS#12 key generation routines. This was triggered
6236 if a 3DES key was generated with a 0 initial byte. Include
6237 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6238 (but broken) behaviour.
6241 *) Enhance bctest to search for a working bc along $PATH and print
6243 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6245 *) Fix memory leaks in err.c: free err_data string if necessary;
6246 don't write to the wrong index in ERR_set_error_data.
6249 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6253 *) Replace rdtsc with _emit statements for VC++ version 5.
6254 [Jeremy Cooper <jeremy@baymoo.org>]
6256 *) Make it possible to reuse SSLv2 sessions.
6259 *) In copy_email() check for >= 0 as a return value for
6260 X509_NAME_get_index_by_NID() since 0 is a valid index.
6261 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6263 *) Avoid coredump with unsupported or invalid public keys by checking if
6264 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6265 PKCS7_verify() fails with non detached data.
6268 *) Don't use getenv in library functions when run as setuid/setgid.
6269 New function OPENSSL_issetugid().
6272 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6273 due to incorrect handling of multi-threading:
6275 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6277 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6279 3. Count how many times MemCheck_off() has been called so that
6280 nested use can be treated correctly. This also avoids
6281 inband-signalling in the previous code (which relied on the
6282 assumption that thread ID 0 is impossible).
6285 *) Add "-rand" option also to s_client and s_server.
6288 *) Fix CPU detection on Irix 6.x.
6289 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6290 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6292 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6295 [This change does not apply to 0.9.7.]
6297 *) Use the cached encoding of an X509_NAME structure rather than
6298 copying it. This is apparently the reason for the libsafe "errors"
6299 but the code is actually correct.
6302 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6303 Bleichenbacher's DSA attack.
6304 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6305 to be set and top=0 forces the highest bit to be set; top=-1 is new
6306 and leaves the highest bit random.
6307 [Ulf Moeller, Bodo Moeller]
6309 *) In the NCONF_...-based implementations for CONF_... queries
6310 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6311 a temporary CONF structure with the data component set to NULL
6312 (which gives segmentation faults in lh_retrieve).
6313 Instead, use NULL for the CONF pointer in CONF_get_string and
6314 CONF_get_number (which may use environment variables) and directly
6315 return NULL from CONF_get_section.
6318 *) Fix potential buffer overrun for EBCDIC.
6321 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6322 keyUsage if basicConstraints absent for a CA.
6325 *) Make SMIME_write_PKCS7() write mail header values with a format that
6326 is more generally accepted (no spaces before the semicolon), since
6327 some programs can't parse those values properly otherwise. Also make
6328 sure BIO's that break lines after each write do not create invalid
6332 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6333 macros previously used would not encode an empty SEQUENCE OF
6334 and break the signature.
6336 [This change does not apply to 0.9.7.]
6338 *) Zero the premaster secret after deriving the master secret in
6342 *) Add some EVP_add_digest_alias registrations (as found in
6343 OpenSSL_add_all_digests()) to SSL_library_init()
6344 aka OpenSSL_add_ssl_algorithms(). This provides improved
6345 compatibility with peers using X.509 certificates
6346 with unconventional AlgorithmIdentifier OIDs.
6349 *) Fix for Irix with NO_ASM.
6350 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6352 *) ./config script fixes.
6353 [Ulf Moeller, Richard Levitte]
6355 *) Fix 'openssl passwd -1'.
6358 *) Change PKCS12_key_gen_asc() so it can cope with non null
6359 terminated strings whose length is passed in the passlen
6360 parameter, for example from PEM callbacks. This was done
6361 by adding an extra length parameter to asc2uni().
6362 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6364 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6365 call failed, free the DSA structure.
6368 *) Fix to uni2asc() to cope with zero length Unicode strings.
6369 These are present in some PKCS#12 files.
6372 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6373 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6374 when writing a 32767 byte record.
6375 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6377 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6378 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6380 (RSA objects have a reference count access to which is protected
6381 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6382 so they are meant to be shared between threads.)
6383 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6384 "Reddie, Steven" <Steven.Reddie@ca.com>]
6386 *) Fix a deadlock in CRYPTO_mem_leaks().
6389 *) Use better test patterns in bntest.
6392 *) rand_win.c fix for Borland C.
6395 *) BN_rshift bugfix for n == 0.
6398 *) Add a 'bctest' script that checks for some known 'bc' bugs
6399 so that 'make test' does not abort just because 'bc' is broken.
6402 *) Store verify_result within SSL_SESSION also for client side to
6403 avoid potential security hole. (Re-used sessions on the client side
6404 always resulted in verify_result==X509_V_OK, not using the original
6405 result of the server certificate verification.)
6408 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6409 SSL3_RT_APPLICATION_DATA, return 0.
6410 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6414 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6415 releases, have been re-implemented by renaming the previous
6416 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6417 and ssl3_read_internal, respectively, and adding 'peek' parameters
6418 to them. The new ssl[23]_{read,peek} functions are calls to
6419 ssl[23]_read_internal with the 'peek' flag set appropriately.
6420 A 'peek' parameter has also been added to ssl3_read_bytes, which
6421 does the actual work for ssl3_read_internal.
6424 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6425 the method-specific "init()" handler. Also clean up ex_data after
6426 calling the method-specific "finish()" handler. Previously, this was
6427 happening the other way round.
6430 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6431 The previous value, 12, was not always sufficient for BN_mod_exp().
6434 *) Make sure that shared libraries get the internal name engine with
6435 the full version number and not just 0. This should mark the
6436 shared libraries as not backward compatible. Of course, this should
6437 be changed again when we can guarantee backward binary compatibility.
6440 *) Fix typo in get_cert_by_subject() in by_dir.c
6441 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6443 *) Rework the system to generate shared libraries:
6445 - Make note of the expected extension for the shared libraries and
6446 if there is a need for symbolic links from for example libcrypto.so.0
6447 to libcrypto.so.0.9.7. There is extended info in Configure for
6450 - Make as few rebuilds of the shared libraries as possible.
6452 - Still avoid linking the OpenSSL programs with the shared libraries.
6454 - When installing, install the shared libraries separately from the
6458 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6460 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6461 and not in SSL_clear because the latter is also used by the
6462 accept/connect functions; previously, the settings made by
6463 SSL_set_read_ahead would be lost during the handshake.
6464 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6466 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6467 Previously, it would create entries for disableed algorithms no
6471 *) Added several new manual pages for SSL_* function.
6474 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6476 *) In ssl23_get_client_hello, generate an error message when faced
6477 with an initial SSL 3.0/TLS record that is too small to contain the
6478 first two bytes of the ClientHello message, i.e. client_version.
6479 (Note that this is a pathologic case that probably has never happened
6480 in real life.) The previous approach was to use the version number
6481 from the record header as a substitute; but our protocol choice
6482 should not depend on that one because it is not authenticated
6483 by the Finished messages.
6486 *) More robust randomness gathering functions for Windows.
6487 [Jeffrey Altman <jaltman@columbia.edu>]
6489 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6490 not set then we don't setup the error code for issuer check errors
6491 to avoid possibly overwriting other errors which the callback does
6492 handle. If an application does set the flag then we assume it knows
6493 what it is doing and can handle the new informational codes
6497 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6498 a general "ANY" type, as such it should be able to decode anything
6499 including tagged types. However it didn't check the class so it would
6500 wrongly interpret tagged types in the same way as their universal
6501 counterpart and unknown types were just rejected. Changed so that the
6502 tagged and unknown types are handled in the same way as a SEQUENCE:
6503 that is the encoding is stored intact. There is also a new type
6504 "V_ASN1_OTHER" which is used when the class is not universal, in this
6505 case we have no idea what the actual type is so we just lump them all
6509 *) On VMS, stdout may very well lead to a file that is written to
6510 in a record-oriented fashion. That means that every write() will
6511 write a separate record, which will be read separately by the
6512 programs trying to read from it. This can be very confusing.
6514 The solution is to put a BIO filter in the way that will buffer
6515 text until a linefeed is reached, and then write everything a
6516 line at a time, so every record written will be an actual line,
6517 not chunks of lines and not (usually doesn't happen, but I've
6518 seen it once) several lines in one record. BIO_f_linebuffer() is
6521 Currently, it's a VMS-only method, because that's where it has
6522 been tested well enough.
6525 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6526 it can return incorrect results.
6527 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6528 but it was in 0.9.6-beta[12].)
6531 *) Disable the check for content being present when verifying detached
6532 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6533 include zero length content when signing messages.
6536 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6537 BIO_ctrl (for BIO pairs).
6540 *) Add DSO method for VMS.
6543 *) Bug fix: Montgomery multiplication could produce results with the
6547 *) Add RPM specification openssl.spec and modify it to build three
6548 packages. The default package contains applications, application
6549 documentation and run-time libraries. The devel package contains
6550 include files, static libraries and function documentation. The
6551 doc package contains the contents of the doc directory. The original
6552 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6555 *) Add a large number of documentation files for many SSL routines.
6556 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6558 *) Add a configuration entry for Sony News 4.
6559 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6561 *) Don't set the two most significant bits to one when generating a
6562 random number < q in the DSA library.
6565 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6566 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6567 the underlying transport is blocking) if a handshake took place.
6568 (The default behaviour is needed by applications such as s_client
6569 and s_server that use select() to determine when to use SSL_read;
6570 but for applications that know in advance when to expect data, it
6571 just makes things more complicated.)
6574 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6578 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6579 work better on such systems.
6580 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6582 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6583 Update PKCS12_parse() so it copies the friendlyName and the
6584 keyid to the certificates aux info.
6587 *) Fix bug in PKCS7_verify() which caused an infinite loop
6588 if there was more than one signature.
6589 [Sven Uszpelkat <su@celocom.de>]
6591 *) Major change in util/mkdef.pl to include extra information
6592 about each symbol, as well as presentig variables as well
6593 as functions. This change means that there's n more need
6594 to rebuild the .num files when some algorithms are excluded.
6597 *) Allow the verify time to be set by an application,
6598 rather than always using the current time.
6601 *) Phase 2 verify code reorganisation. The certificate
6602 verify code now looks up an issuer certificate by a
6603 number of criteria: subject name, authority key id
6604 and key usage. It also verifies self signed certificates
6605 by the same criteria. The main comparison function is
6606 X509_check_issued() which performs these checks.
6608 Lot of changes were necessary in order to support this
6609 without completely rewriting the lookup code.
6611 Authority and subject key identifier are now cached.
6613 The LHASH 'certs' is X509_STORE has now been replaced
6614 by a STACK_OF(X509_OBJECT). This is mainly because an
6615 LHASH can't store or retrieve multiple objects with
6616 the same hash value.
6618 As a result various functions (which were all internal
6619 use only) have changed to handle the new X509_STORE
6620 structure. This will break anything that messed round
6621 with X509_STORE internally.
6623 The functions X509_STORE_add_cert() now checks for an
6624 exact match, rather than just subject name.
6626 The X509_STORE API doesn't directly support the retrieval
6627 of multiple certificates matching a given criteria, however
6628 this can be worked round by performing a lookup first
6629 (which will fill the cache with candidate certificates)
6630 and then examining the cache for matches. This is probably
6631 the best we can do without throwing out X509_LOOKUP
6632 entirely (maybe later...).
6634 The X509_VERIFY_CTX structure has been enhanced considerably.
6636 All certificate lookup operations now go via a get_issuer()
6637 callback. Although this currently uses an X509_STORE it
6638 can be replaced by custom lookups. This is a simple way
6639 to bypass the X509_STORE hackery necessary to make this
6640 work and makes it possible to use more efficient techniques
6641 in future. A very simple version which uses a simple
6642 STACK for its trusted certificate store is also provided
6643 using X509_STORE_CTX_trusted_stack().
6645 The verify_cb() and verify() callbacks now have equivalents
6646 in the X509_STORE_CTX structure.
6648 X509_STORE_CTX also has a 'flags' field which can be used
6649 to customise the verify behaviour.
6652 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6653 excludes S/MIME capabilities.
6656 *) When a certificate request is read in keep a copy of the
6657 original encoding of the signed data and use it when outputing
6658 again. Signatures then use the original encoding rather than
6659 a decoded, encoded version which may cause problems if the
6660 request is improperly encoded.
6663 *) For consistency with other BIO_puts implementations, call
6664 buffer_write(b, ...) directly in buffer_puts instead of calling
6667 In BIO_puts, increment b->num_write as in BIO_write.
6668 [Peter.Sylvester@EdelWeb.fr]
6670 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6671 BN_zero, we may not return a BIGNUM with an array consisting of
6675 *) Avoid calling abort() from within the library when problems are
6676 detected, except if preprocessor symbols have been defined
6677 (such as REF_CHECK, BN_DEBUG etc.).
6680 *) New openssl application 'rsautl'. This utility can be
6681 used for low level RSA operations. DER public key
6682 BIO/fp routines also added.
6685 *) New Configure entry and patches for compiling on QNX 4.
6686 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6688 *) A demo state-machine implementation was sponsored by
6689 Nuron (http://www.nuron.com/) and is now available in
6690 demos/state_machine.
6693 *) New options added to the 'dgst' utility for signature
6694 generation and verification.
6697 *) Unrecognized PKCS#7 content types are now handled via a
6698 catch all ASN1_TYPE structure. This allows unsupported
6699 types to be stored as a "blob" and an application can
6700 encode and decode it manually.
6703 *) Fix various signed/unsigned issues to make a_strex.c
6705 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6707 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6708 length if passed a buffer. ASN1_INTEGER_to_BN failed
6709 if passed a NULL BN and its argument was negative.
6710 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6712 *) Modification to PKCS#7 encoding routines to output definite
6713 length encoding. Since currently the whole structures are in
6714 memory there's not real point in using indefinite length
6715 constructed encoding. However if OpenSSL is compiled with
6716 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6719 *) Added BIO_vprintf() and BIO_vsnprintf().
6722 *) Added more prefixes to parse for in the the strings written
6723 through a logging bio, to cover all the levels that are available
6724 through syslog. The prefixes are now:
6726 PANIC, EMERG, EMR => LOG_EMERG
6727 ALERT, ALR => LOG_ALERT
6728 CRIT, CRI => LOG_CRIT
6729 ERROR, ERR => LOG_ERR
6730 WARNING, WARN, WAR => LOG_WARNING
6731 NOTICE, NOTE, NOT => LOG_NOTICE
6732 INFO, INF => LOG_INFO
6733 DEBUG, DBG => LOG_DEBUG
6735 and as before, if none of those prefixes are present at the
6736 beginning of the string, LOG_ERR is chosen.
6738 On Win32, the LOG_* levels are mapped according to this:
6740 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6741 LOG_WARNING => EVENTLOG_WARNING_TYPE
6742 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6746 *) Made it possible to reconfigure with just the configuration
6747 argument "reconf" or "reconfigure". The command line arguments
6748 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6749 and are retrieved from there when reconfiguring.
6753 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6755 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6758 *) The obj_dat.pl script was messing up the sorting of object
6759 names. The reason was that it compared the quoted version
6760 of strings as a result "OCSP" > "OCSP Signing" because
6761 " > SPACE. Changed script to store unquoted versions of
6762 names and add quotes on output. It was also omitting some
6763 names from the lookup table if they were given a default
6764 value (that is if SN is missing it is given the same
6765 value as LN and vice versa), these are now added on the
6766 grounds that if an object has a name we should be able to
6767 look it up. Finally added warning output when duplicate
6768 short or long names are found.
6771 *) Changes needed for Tandem NSK.
6772 [Scott Uroff <scott@xypro.com>]
6774 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6775 RSA_padding_check_SSLv23(), special padding was never detected
6776 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6777 version rollback attacks was not effective.
6779 In s23_clnt.c, don't use special rollback-attack detection padding
6780 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6781 client; similarly, in s23_srvr.c, don't do the rollback check if
6782 SSL 2.0 is the only protocol enabled in the server.
6785 *) Make it possible to get hexdumps of unprintable data with 'openssl
6786 asn1parse'. By implication, the functions ASN1_parse_dump() and
6787 BIO_dump_indent() are added.
6790 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6791 these print out strings and name structures based on various
6792 flags including RFC2253 support and proper handling of
6793 multibyte characters. Added options to the 'x509' utility
6794 to allow the various flags to be set.
6797 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6798 Also change the functions X509_cmp_current_time() and
6799 X509_gmtime_adj() work with an ASN1_TIME structure,
6800 this will enable certificates using GeneralizedTime in validity
6801 dates to be checked.
6804 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6805 negative public key encodings) on by default,
6806 NO_NEG_PUBKEY_BUG can be set to disable it.
6809 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6810 content octets. An i2c_ASN1_OBJECT is unnecessary because
6811 the encoding can be trivially obtained from the structure.
6814 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6815 not read locks (CRYPTO_r_[un]lock).
6818 *) A first attempt at creating official support for shared
6819 libraries through configuration. I've kept it so the
6820 default is static libraries only, and the OpenSSL programs
6821 are always statically linked for now, but there are
6822 preparations for dynamic linking in place.
6823 This has been tested on Linux and Tru64.
6826 *) Randomness polling function for Win9x, as described in:
6827 Peter Gutmann, Software Generation of Practically Strong
6831 *) Fix so PRNG is seeded in req if using an already existing
6835 *) New options to smime application. -inform and -outform
6836 allow alternative formats for the S/MIME message including
6837 PEM and DER. The -content option allows the content to be
6838 specified separately. This should allow things like Netscape
6839 form signing output easier to verify.
6842 *) Fix the ASN1 encoding of tags using the 'long form'.
6845 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6846 STRING types. These convert content octets to and from the
6847 underlying type. The actual tag and length octets are
6848 already assumed to have been read in and checked. These
6849 are needed because all other string types have virtually
6850 identical handling apart from the tag. By having versions
6851 of the ASN1 functions that just operate on content octets
6852 IMPLICIT tagging can be handled properly. It also allows
6853 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6854 and ASN1_INTEGER are identical apart from the tag.
6857 *) Change the handling of OID objects as follows:
6859 - New object identifiers are inserted in objects.txt, following
6860 the syntax given in objects.README.
6861 - objects.pl is used to process obj_mac.num and create a new
6863 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6866 This is currently kind of a hack, and the perl code in objects.pl
6867 isn't very elegant, but it works as I intended. The simplest way
6868 to check that it worked correctly is to look in obj_dat.h and
6869 check the array nid_objs and make sure the objects haven't moved
6870 around (this is important!). Additions are OK, as well as
6871 consistent name changes.
6874 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6877 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6878 The given file adds to whatever has already been seeded into the
6879 random pool through the RANDFILE configuration file option or
6880 environment variable, or the default random state file.
6883 *) mkstack.pl now sorts each macro group into lexical order.
6884 Previously the output order depended on the order the files
6885 appeared in the directory, resulting in needless rewriting
6889 *) Patches to make OpenSSL compile under Win32 again. Mostly
6890 work arounds for the VC++ problem that it treats func() as
6891 func(void). Also stripped out the parts of mkdef.pl that
6892 added extra typesafe functions: these no longer exist.
6895 *) Reorganisation of the stack code. The macros are now all
6896 collected in safestack.h . Each macro is defined in terms of
6897 a "stack macro" of the form SKM_<name>(type, a, b). The
6898 DEBUG_SAFESTACK is now handled in terms of function casts,
6899 this has the advantage of retaining type safety without the
6900 use of additional functions. If DEBUG_SAFESTACK is not defined
6901 then the non typesafe macros are used instead. Also modified the
6902 mkstack.pl script to handle the new form. Needs testing to see
6903 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6904 the default if no major problems. Similar behaviour for ASN1_SET_OF
6905 and PKCS12_STACK_OF.
6908 *) When some versions of IIS use the 'NET' form of private key the
6909 key derivation algorithm is different. Normally MD5(password) is
6910 used as a 128 bit RC4 key. In the modified case
6911 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6912 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6913 as the old Netscape_RSA functions except they have an additional
6914 'sgckey' parameter which uses the modified algorithm. Also added
6915 an -sgckey command line option to the rsa utility. Thanks to
6916 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6917 algorithm to openssl-dev.
6920 *) The evp_local.h macros were using 'c.##kname' which resulted in
6921 invalid expansion on some systems (SCO 5.0.5 for example).
6922 Corrected to 'c.kname'.
6923 [Phillip Porch <root@theporch.com>]
6925 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6926 a STACK of email addresses from a certificate or request, these look
6927 in the subject name and the subject alternative name extensions and
6928 omit any duplicate addresses.
6931 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6932 This makes DSA verification about 2 % faster.
6935 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6936 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6937 plus overhead for 1024 bit moduli).
6938 This makes exponentiations about 0.5 % faster for 1024 bit
6939 exponents (as measured by "openssl speed rsa2048").
6942 *) Rename memory handling macros to avoid conflicts with other
6944 Malloc => OPENSSL_malloc
6945 Malloc_locked => OPENSSL_malloc_locked
6946 Realloc => OPENSSL_realloc
6947 Free => OPENSSL_free
6950 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6951 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6954 *) CygWin32 support.
6955 [John Jarvie <jjarvie@newsguy.com>]
6957 *) The type-safe stack code has been rejigged. It is now only compiled
6958 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6959 by default all type-specific stack functions are "#define"d back to
6960 standard stack functions. This results in more streamlined output
6961 but retains the type-safety checking possibilities of the original
6965 *) The STACK code has been cleaned up, and certain type declarations
6966 that didn't make a lot of sense have been brought in line. This has
6967 also involved a cleanup of sorts in safestack.h to more correctly
6968 map type-safe stack functions onto their plain stack counterparts.
6969 This work has also resulted in a variety of "const"ifications of
6970 lots of the code, especially "_cmp" operations which should normally
6971 be prototyped with "const" parameters anyway.
6974 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6975 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6976 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6977 where all of 'md' is used each time the PRNG is used, but 'state'
6978 is used only indexed by a cyclic counter. As entropy may not be
6979 well distributed from the beginning, 'md' is important as a
6980 chaining variable. However, the output function chains only half
6981 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6982 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6983 in all of 'state' being rewritten, with the new values depending
6984 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6987 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6988 the handshake is continued after ssl_verify_cert_chain();
6989 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6990 can lead to 'unexplainable' connection aborts later.
6991 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6993 *) Major EVP API cipher revision.
6994 Add hooks for extra EVP features. This allows various cipher
6995 parameters to be set in the EVP interface. Support added for variable
6996 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6997 setting of RC2 and RC5 parameters.
6999 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7002 Remove lots of duplicated code from the EVP library. For example *every*
7003 cipher init() function handles the 'iv' in the same way according to the
7004 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7005 for CFB and OFB modes they zero ctx->num.
7007 New functionality allows removal of S/MIME code RC2 hack.
7009 Most of the routines have the same form and so can be declared in terms
7012 By shifting this to the top level EVP_CipherInit() it can be removed from
7013 all individual ciphers. If the cipher wants to handle IVs or keys
7014 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7017 Change lots of functions like EVP_EncryptUpdate() to now return a
7018 value: although software versions of the algorithms cannot fail
7019 any installed hardware versions can.
7022 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7023 this option is set, tolerate broken clients that send the negotiated
7024 protocol version number instead of the requested protocol version
7028 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7029 i.e. non-zero for export ciphersuites, zero otherwise.
7030 Previous versions had this flag inverted, inconsistent with
7031 rsa_tmp_cb (..._TMP_RSA_CB).
7032 [Bodo Moeller; problem reported by Amit Chopra]
7034 *) Add missing DSA library text string. Work around for some IIS
7035 key files with invalid SEQUENCE encoding.
7038 *) Add a document (doc/standards.txt) that list all kinds of standards
7039 and so on that are implemented in OpenSSL.
7042 *) Enhance c_rehash script. Old version would mishandle certificates
7043 with the same subject name hash and wouldn't handle CRLs at all.
7044 Added -fingerprint option to crl utility, to support new c_rehash
7048 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7051 *) Fix for SSL server purpose checking. Server checking was
7052 rejecting certificates which had extended key usage present
7053 but no ssl client purpose.
7054 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7056 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7057 is a little unclear about how a blank password is handled.
7058 Since the password in encoded as a BMPString with terminating
7059 double NULL a zero length password would end up as just the
7060 double NULL. However no password at all is different and is
7061 handled differently in the PKCS#12 key generation code. NS
7062 treats a blank password as zero length. MSIE treats it as no
7063 password on export: but it will try both on import. We now do
7064 the same: PKCS12_parse() tries zero length and no password if
7065 the password is set to "" or NULL (NULL is now a valid password:
7066 it wasn't before) as does the pkcs12 application.
7069 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7070 perror when PEM_read_bio_X509_REQ fails, the error message must
7071 be obtained from the error queue.
7074 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7075 it in ERR_remove_state if appropriate, and change ERR_get_state
7076 accordingly to avoid race conditions (this is necessary because
7077 thread_hash is no longer constant once set).
7080 *) Bugfix for linux-elf makefile.one.
7083 *) RSA_get_default_method() will now cause a default
7084 RSA_METHOD to be chosen if one doesn't exist already.
7085 Previously this was only set during a call to RSA_new()
7086 or RSA_new_method(NULL) meaning it was possible for
7087 RSA_get_default_method() to return NULL.
7090 *) Added native name translation to the existing DSO code
7091 that will convert (if the flag to do so is set) filenames
7092 that are sufficiently small and have no path information
7093 into a canonical native form. Eg. "blah" converted to
7094 "libblah.so" or "blah.dll" etc.
7097 *) New function ERR_error_string_n(e, buf, len) which is like
7098 ERR_error_string(e, buf), but writes at most 'len' bytes
7099 including the 0 terminator. For ERR_error_string_n, 'buf'
7101 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7103 *) CONF library reworked to become more general. A new CONF
7104 configuration file reader "class" is implemented as well as a
7105 new functions (NCONF_*, for "New CONF") to handle it. The now
7106 old CONF_* functions are still there, but are reimplemented to
7107 work in terms of the new functions. Also, a set of functions
7108 to handle the internal storage of the configuration data is
7109 provided to make it easier to write new configuration file
7110 reader "classes" (I can definitely see something reading a
7111 configuration file in XML format, for example), called _CONF_*,
7112 or "the configuration storage API"...
7114 The new configuration file reading functions are:
7116 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7117 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7119 NCONF_default, NCONF_WIN32
7121 NCONF_dump_fp, NCONF_dump_bio
7123 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7124 NCONF_new creates a new CONF object. This works in the same way
7125 as other interfaces in OpenSSL, like the BIO interface.
7126 NCONF_dump_* dump the internal storage of the configuration file,
7127 which is useful for debugging. All other functions take the same
7128 arguments as the old CONF_* functions wth the exception of the
7129 first that must be a `CONF *' instead of a `LHASH *'.
7131 To make it easer to use the new classes with the old CONF_* functions,
7132 the function CONF_set_default_method is provided.
7135 *) Add '-tls1' option to 'openssl ciphers', which was already
7136 mentioned in the documentation but had not been implemented.
7137 (This option is not yet really useful because even the additional
7138 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7141 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7142 OpenSSL-based applications) load shared libraries and bind to
7143 them in a portable way.
7144 [Geoff Thorpe, with contributions from Richard Levitte]
7146 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7148 *) Make sure _lrotl and _lrotr are only used with MSVC.
7150 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7151 (the default implementation of RAND_status).
7153 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7154 to '-clrext' (= clear extensions), as intended and documented.
7155 [Bodo Moeller; inconsistency pointed out by Michael Attili
7158 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7159 was larger than the MD block size.
7160 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7162 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7163 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7164 using the passed key: if the passed key was a private key the result
7165 of X509_print(), for example, would be to print out all the private key
7169 *) des_quad_cksum() byte order bug fix.
7170 [Ulf Möller, using the problem description in krb4-0.9.7, where
7171 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7173 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7175 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7177 *) For easily testing in shell scripts whether some command
7178 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7179 returns with exit code 0 iff no command of the given name is available.
7180 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7181 the output goes to stdout and nothing is printed to stderr.
7182 Additional arguments are always ignored.
7184 Since for each cipher there is a command of the same name,
7185 the 'no-cipher' compilation switches can be tested this way.
7187 ('openssl no-XXX' is not able to detect pseudo-commands such
7188 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7191 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7194 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7195 is set; it will be thrown away anyway because each handshake creates
7197 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7198 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7199 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7200 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7203 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7204 'Q' and 'R' lose their special meanings (quit/renegotiate).
7205 This is part of what -quiet does; unlike -quiet, -ign_eof
7206 does not suppress any output.
7209 *) Add compatibility options to the purpose and trust code. The
7210 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7211 accepts a certificate or CA, this was the previous behaviour,
7212 with all the associated security issues.
7214 X509_TRUST_COMPAT is the old trust behaviour: only and
7215 automatically trust self signed roots in certificate store. A
7216 new trust setting X509_TRUST_DEFAULT is used to specify that
7217 a purpose has no associated trust setting and it should instead
7218 use the value in the default purpose.
7221 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7222 and fix a memory leak.
7225 *) In util/mkerr.pl (which implements 'make errors'), preserve
7226 reason strings from the previous version of the .c file, as
7227 the default to have only downcase letters (and digits) in
7228 automatically generated reasons codes is not always appropriate.
7231 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7232 using strerror. Previously, ERR_reason_error_string() returned
7233 library names as reason strings for SYSerr; but SYSerr is a special
7234 case where small numbers are errno values, not library numbers.
7237 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7238 converts DSA parameters into DH parameters. (When creating parameters,
7239 DSA_generate_parameters is used.)
7242 *) Include 'length' (recommended exponent length) in C code generated
7243 by 'openssl dhparam -C'.
7246 *) The second argument to set_label in perlasm was already being used
7247 so couldn't be used as a "file scope" flag. Moved to third argument
7251 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7252 instead of RAND_bytes for encryption IVs and salts.
7255 *) Include RAND_status() into RAND_METHOD instead of implementing
7256 it only for md_rand.c Otherwise replacing the PRNG by calling
7257 RAND_set_rand_method would be impossible.
7260 *) Don't let DSA_generate_key() enter an infinite loop if the random
7261 number generation fails.
7264 *) New 'rand' application for creating pseudo-random output.
7267 *) Added configuration support for Linux/IA64
7268 [Rolf Haberrecker <rolf@suse.de>]
7270 *) Assembler module support for Mingw32.
7273 *) Shared library support for HPUX (in shlib/).
7274 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7276 *) Shared library support for Solaris gcc.
7277 [Lutz Behnke <behnke@trustcenter.de>]
7279 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7281 *) PKCS7_encrypt() was adding text MIME headers twice because they
7282 were added manually and by SMIME_crlf_copy().
7285 *) In bntest.c don't call BN_rand with zero bits argument.
7286 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7288 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7289 case was implemented. This caused BN_div_recp() to fail occasionally.
7292 *) Add an optional second argument to the set_label() in the perl
7293 assembly language builder. If this argument exists and is set
7294 to 1 it signals that the assembler should use a symbol whose
7295 scope is the entire file, not just the current function. This
7296 is needed with MASM which uses the format label:: for this scope.
7297 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7299 *) Change the ASN1 types so they are typedefs by default. Before
7300 almost all types were #define'd to ASN1_STRING which was causing
7301 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7305 *) Change names of new functions to the new get1/get0 naming
7306 convention: After 'get1', the caller owns a reference count
7307 and has to call ..._free; 'get0' returns a pointer to some
7308 data structure without incrementing reference counters.
7309 (Some of the existing 'get' functions increment a reference
7310 counter, some don't.)
7311 Similarly, 'set1' and 'add1' functions increase reference
7312 counters or duplicate objects.
7315 *) Allow for the possibility of temp RSA key generation failure:
7316 the code used to assume it always worked and crashed on failure.
7319 *) Fix potential buffer overrun problem in BIO_printf().
7320 [Ulf Möller, using public domain code by Patrick Powell; problem
7321 pointed out by David Sacerdote <das33@cornell.edu>]
7323 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7324 RAND_egd() and RAND_status(). In the command line application,
7325 the EGD socket can be specified like a seed file using RANDFILE
7329 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7330 Some CAs (e.g. Verisign) distribute certificates in this form.
7333 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7334 list to exclude them. This means that no special compilation option
7335 is needed to use anonymous DH: it just needs to be included in the
7339 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7340 EVP_MD_type. The old functionality is available in a new macro called
7341 EVP_MD_md(). Change code that uses it and update docs.
7344 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7345 where the 'void *' argument is replaced by a function pointer argument.
7346 Previously 'void *' was abused to point to functions, which works on
7347 many platforms, but is not correct. As these functions are usually
7348 called by macros defined in OpenSSL header files, most source code
7349 should work without changes.
7352 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7353 sections with information on -D... compiler switches used for
7354 compiling the library so that applications can see them. To enable
7355 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7356 must be defined. E.g.,
7357 #define OPENSSL_ALGORITHM_DEFINES
7358 #include <openssl/opensslconf.h>
7359 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7360 [Richard Levitte, Ulf and Bodo Möller]
7362 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7366 *) Change the 'other' type in certificate aux info to a STACK_OF
7367 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7368 the required ASN1 format: arbitrary types determined by an OID.
7371 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7372 argument to 'req'. This is not because the function is newer or
7373 better than others it just uses the work 'NEW' in the certificate
7374 request header lines. Some software needs this.
7377 *) Reorganise password command line arguments: now passwords can be
7378 obtained from various sources. Delete the PEM_cb function and make
7379 it the default behaviour: i.e. if the callback is NULL and the
7380 usrdata argument is not NULL interpret it as a null terminated pass
7381 phrase. If usrdata and the callback are NULL then the pass phrase
7382 is prompted for as usual.
7385 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7386 the support is automatically enabled. The resulting binaries will
7387 autodetect the card and use it if present.
7388 [Ben Laurie and Compaq Inc.]
7390 *) Work around for Netscape hang bug. This sends certificate request
7391 and server done in one record. Since this is perfectly legal in the
7392 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7393 the bugs/SSLv3 entry for more info.
7396 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7399 *) Add -rand argument to smime and pkcs12 applications and read/write
7403 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7406 *) Add command line password options to the remaining applications.
7409 *) Bug fix for BN_div_recp() for numerators with an even number of
7413 *) More tests in bntest.c, and changed test_bn output.
7416 *) ./config recognizes MacOS X now.
7419 *) Bug fix for BN_div() when the first words of num and divsor are
7420 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7423 *) Add support for various broken PKCS#8 formats, and command line
7424 options to produce them.
7427 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7428 get temporary BIGNUMs from a BN_CTX.
7431 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7435 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7436 include a #define from the old name to the new. The original intent
7437 was that statically linked binaries could for example just call
7438 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7439 link with digests. This never worked becayse SSLeay_add_all_digests()
7440 and SSLeay_add_all_ciphers() were in the same source file so calling
7441 one would link with the other. They are now in separate source files.
7444 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7447 *) Use a less unusual form of the Miller-Rabin primality test (it used
7448 a binary algorithm for exponentiation integrated into the Miller-Rabin
7449 loop, our standard modexp algorithms are faster).
7452 *) Support for the EBCDIC character set completed.
7453 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7455 *) Source code cleanups: use const where appropriate, eliminate casts,
7456 use void * instead of char * in lhash.
7459 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7460 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7461 this the server could overwrite ephemeral keys that the client
7465 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7466 using 50 iterations of the Rabin-Miller test.
7468 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7469 iterations of the Rabin-Miller test as required by the appendix
7470 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7471 As BN_is_prime_fasttest includes trial division, DSA parameter
7472 generation becomes much faster.
7474 This implies a change for the callback functions in DSA_is_prime
7475 and DSA_generate_parameters: The callback function is called once
7476 for each positive witness in the Rabin-Miller test, not just
7477 occasionally in the inner loop; and the parameters to the
7478 callback function now provide an iteration count for the outer
7479 loop rather than for the current invocation of the inner loop.
7480 DSA_generate_parameters additionally can call the callback
7481 function with an 'iteration count' of -1, meaning that a
7482 candidate has passed the trial division test (when q is generated
7483 from an application-provided seed, trial division is skipped).
7486 *) New function BN_is_prime_fasttest that optionally does trial
7487 division before starting the Rabin-Miller test and has
7488 an additional BN_CTX * argument (whereas BN_is_prime always
7489 has to allocate at least one BN_CTX).
7490 'callback(1, -1, cb_arg)' is called when a number has passed the
7491 trial division stage.
7494 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7498 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7501 *) New function BN_pseudo_rand().
7504 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7505 bignum version of BN_from_montgomery() with the working code from
7506 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7510 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7511 made it impossible to use the same SSL_SESSION data structure in
7512 SSL2 clients in multiple threads.
7515 *) The return value of RAND_load_file() no longer counts bytes obtained
7516 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7517 to seed the PRNG (previously an explicit byte count was required).
7518 [Ulf Möller, Bodo Möller]
7520 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7521 used (char *) instead of (void *) and had casts all over the place.
7524 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7527 *) Retain source code compatibility for BN_prime_checks macro:
7528 BN_is_prime(..., BN_prime_checks, ...) now uses
7529 BN_prime_checks_for_size to determine the appropriate number of
7530 Rabin-Miller iterations.
7533 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7534 DH_CHECK_P_NOT_SAFE_PRIME.
7535 (Check if this is true? OpenPGP calls them "strong".)
7538 *) Merge the functionality of "dh" and "gendh" programs into a new program
7539 "dhparam". The old programs are retained for now but will handle DH keys
7540 (instead of parameters) in future.
7543 *) Make the ciphers, s_server and s_client programs check the return values
7544 when a new cipher list is set.
7547 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7548 ciphers. Before when the 56bit ciphers were enabled the sorting was
7551 The syntax for the cipher sorting has been extended to support sorting by
7552 cipher-strength (using the strength_bits hard coded in the tables).
7553 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7555 Fix a bug in the cipher-command parser: when supplying a cipher command
7556 string with an "undefined" symbol (neither command nor alphanumeric
7557 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7558 an error is flagged.
7560 Due to the strength-sorting extension, the code of the
7561 ssl_create_cipher_list() function was completely rearranged. I hope that
7562 the readability was also increased :-)
7563 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7565 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7566 for the first serial number and places 2 in the serial number file. This
7567 avoids problems when the root CA is created with serial number zero and
7568 the first user certificate has the same issuer name and serial number
7572 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7573 the new code. Add documentation for this stuff.
7576 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7577 X509_*() to X509at_*() on the grounds that they don't handle X509
7578 structures and behave in an analagous way to the X509v3 functions:
7579 they shouldn't be called directly but wrapper functions should be used
7582 So we also now have some wrapper functions that call the X509at functions
7583 when passed certificate requests. (TO DO: similar things can be done with
7584 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7585 things. Some of these need some d2i or i2d and print functionality
7586 because they handle more complex structures.)
7589 *) Add missing #ifndefs that caused missing symbols when building libssl
7590 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7591 NO_RSA in ssl/s2*.c.
7592 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7594 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7595 has a return value which indicates the quality of the random data
7596 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7597 error queue. New function RAND_pseudo_bytes() generates output that is
7598 guaranteed to be unique but not unpredictable. RAND_add is like
7599 RAND_seed, but takes an extra argument for an entropy estimate
7600 (RAND_seed always assumes full entropy).
7603 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7604 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7605 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7606 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7607 false-positive rate of at most 2^-80 for random input.
7610 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7613 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7614 in the 0.9.5 release), this returns the chain
7615 from an X509_CTX structure with a dup of the stack and all
7616 the X509 reference counts upped: so the stack will exist
7617 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7620 Also make SSL_SESSION_print() print out the verify return
7624 *) Add manpage for the pkcs12 command. Also change the default
7625 behaviour so MAC iteration counts are used unless the new
7626 -nomaciter option is used. This improves file security and
7627 only older versions of MSIE (4.0 for example) need it.
7630 *) Honor the no-xxx Configure options when creating .DEF files.
7633 *) Add PKCS#10 attributes to field table: challengePassword,
7634 unstructuredName and unstructuredAddress. These are taken from
7635 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7636 international characters are used.
7638 More changes to X509_ATTRIBUTE code: allow the setting of types
7639 based on strings. Remove the 'loc' parameter when adding
7640 attributes because these will be a SET OF encoding which is sorted
7644 *) Initial changes to the 'req' utility to allow request generation
7645 automation. This will allow an application to just generate a template
7646 file containing all the field values and have req construct the
7649 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7650 used all over the place including certificate requests and PKCS#7
7651 structures. They are currently handled manually where necessary with
7652 some primitive wrappers for PKCS#7. The new functions behave in a
7653 manner analogous to the X509 extension functions: they allow
7654 attributes to be looked up by NID and added.
7656 Later something similar to the X509V3 code would be desirable to
7657 automatically handle the encoding, decoding and printing of the
7658 more complex types. The string types like challengePassword can
7659 be handled by the string table functions.
7661 Also modified the multi byte string table handling. Now there is
7662 a 'global mask' which masks out certain types. The table itself
7663 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7664 is useful when for example there is only one permissible type
7665 (as in countryName) and using the mask might result in no valid
7669 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7670 SSL_get_peer_finished to allow applications to obtain the latest
7671 Finished messages sent to the peer or expected from the peer,
7672 respectively. (SSL_get_peer_finished is usually the Finished message
7673 actually received from the peer, otherwise the protocol will be aborted.)
7675 As the Finished message are message digests of the complete handshake
7676 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7677 be used for external authentication procedures when the authentication
7678 provided by SSL/TLS is not desired or is not enough.
7681 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7682 the host supports BWX extension and if Compaq C is present on the
7683 $PATH. Just exploiting of the BWX extension results in 20-30%
7684 performance kick for some algorithms, e.g. DES and RC4 to mention
7685 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7689 *) Add support for MS "fast SGC". This is arguably a violation of the
7690 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7691 weak crypto and after checking the certificate is SGC a second one
7692 with strong crypto. MS SGC stops the first handshake after receiving
7693 the server certificate message and sends a second client hello. Since
7694 a server will typically do all the time consuming operations before
7695 expecting any further messages from the client (server key exchange
7696 is the most expensive) there is little difference between the two.
7698 To get OpenSSL to support MS SGC we have to permit a second client
7699 hello message after we have sent server done. In addition we have to
7700 reset the MAC if we do get this second client hello.
7703 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7704 if a DER encoded private key is RSA or DSA traditional format. Changed
7705 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7706 format DER encoded private key. Newer code should use PKCS#8 format which
7707 has the key type encoded in the ASN1 structure. Added DER private key
7708 support to pkcs8 application.
7711 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7712 ciphersuites has been selected (as required by the SSL 3/TLS 1
7713 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7714 is set, we interpret this as a request to violate the specification
7715 (the worst that can happen is a handshake failure, and 'correct'
7716 behaviour would result in a handshake failure anyway).
7719 *) In SSL_CTX_add_session, take into account that there might be multiple
7720 SSL_SESSION structures with the same session ID (e.g. when two threads
7721 concurrently obtain them from an external cache).
7722 The internal cache can handle only one SSL_SESSION with a given ID,
7723 so if there's a conflict, we now throw out the old one to achieve
7727 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7728 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7729 some routines that use cipher OIDs: some ciphers do not have OIDs
7730 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7734 *) Simplify the trust setting structure and code. Now we just have
7735 two sequences of OIDs for trusted and rejected settings. These will
7736 typically have values the same as the extended key usage extension
7737 and any application specific purposes.
7739 The trust checking code now has a default behaviour: it will just
7740 check for an object with the same NID as the passed id. Functions can
7741 be provided to override either the default behaviour or the behaviour
7742 for a given id. SSL client, server and email already have functions
7743 in place for compatibility: they check the NID and also return "trusted"
7744 if the certificate is self signed.
7747 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7748 traditional format into an EVP_PKEY structure.
7751 *) Add a password callback function PEM_cb() which either prompts for
7752 a password if usr_data is NULL or otherwise assumes it is a null
7753 terminated password. Allow passwords to be passed on command line
7754 environment or config files in a few more utilities.
7757 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7758 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7759 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7760 Update documentation.
7763 *) Support for ASN1 "NULL" type. This could be handled before by using
7764 ASN1_TYPE but there wasn't any function that would try to read a NULL
7765 and produce an error if it couldn't. For compatibility we also have
7766 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7767 don't allocate anything because they don't need to.
7770 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7772 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7774 *) Rebuild of the memory allocation routines used by OpenSSL code and
7775 possibly others as well. The purpose is to make an interface that
7776 provide hooks so anyone can build a separate set of allocation and
7777 deallocation routines to be used by OpenSSL, for example memory
7778 pool implementations, or something else, which was previously hard
7779 since Malloc(), Realloc() and Free() were defined as macros having
7780 the values malloc, realloc and free, respectively (except for Win32
7781 compilations). The same is provided for memory debugging code.
7782 OpenSSL already comes with functionality to find memory leaks, but
7783 this gives people a chance to debug other memory problems.
7785 With these changes, a new set of functions and macros have appeared:
7787 CRYPTO_set_mem_debug_functions() [F]
7788 CRYPTO_get_mem_debug_functions() [F]
7789 CRYPTO_dbg_set_options() [F]
7790 CRYPTO_dbg_get_options() [F]
7791 CRYPTO_malloc_debug_init() [M]
7793 The memory debug functions are NULL by default, unless the library
7794 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7795 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7796 gives the standard debugging functions that come with OpenSSL) or
7797 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7798 provided by the library user) must be used. When the standard
7799 debugging functions are used, CRYPTO_dbg_set_options can be used to
7800 request additional information:
7801 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7802 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7804 Also, things like CRYPTO_set_mem_functions will always give the
7805 expected result (the new set of functions is used for allocation
7806 and deallocation) at all times, regardless of platform and compiler
7809 To finish it up, some functions that were never use in any other
7810 way than through macros have a new API and new semantic:
7813 CRYPTO_dbg_realloc()
7816 All macros of value have retained their old syntax.
7817 [Richard Levitte and Bodo Moeller]
7819 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7820 ordering of SMIMECapabilities wasn't in "strength order" and there
7821 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7825 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7826 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7827 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7829 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7830 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7831 functionality to handle multipart/signed properly) and a utility
7832 called 'smime' to call all this stuff. This is based on code I
7833 originally wrote for Celo who have kindly allowed it to be
7834 included in OpenSSL.
7837 *) Add variants des_set_key_checked and des_set_key_unchecked of
7838 des_set_key (aka des_key_sched). Global variable des_check_key
7839 decides which of these is called by des_set_key; this way
7840 des_check_key behaves as it always did, but applications and
7841 the library itself, which was buggy for des_check_key == 1,
7842 have a cleaner way to pick the version they need.
7845 *) New function PKCS12_newpass() which changes the password of a
7849 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7850 dynamic mix. In both cases the ids can be used as an index into the
7851 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7852 functions so they accept a list of the field values and the
7853 application doesn't need to directly manipulate the X509_TRUST
7857 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7861 *) Modify the way the V3 extension code looks up extensions. This now
7862 works in a similar way to the object code: we have some "standard"
7863 extensions in a static table which is searched with OBJ_bsearch()
7864 and the application can add dynamic ones if needed. The file
7865 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7866 updated whenever a new extension is added to the core code and kept
7867 in ext_nid order. There is a simple program 'tabtest.c' which checks
7868 this. New extensions are not added too often so this file can readily
7869 be maintained manually.
7871 There are two big advantages in doing things this way. The extensions
7872 can be looked up immediately and no longer need to be "added" using
7873 X509V3_add_standard_extensions(): this function now does nothing.
7874 [Side note: I get *lots* of email saying the extension code doesn't
7875 work because people forget to call this function]
7876 Also no dynamic allocation is done unless new extensions are added:
7877 so if we don't add custom extensions there is no need to call
7878 X509V3_EXT_cleanup().
7881 *) Modify enc utility's salting as follows: make salting the default. Add a
7882 magic header, so unsalted files fail gracefully instead of just decrypting
7883 to garbage. This is because not salting is a big security hole, so people
7884 should be discouraged from doing it.
7887 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7888 digest to be passed on the command line but it only used this
7889 parameter when signing a certificate. Modified so all relevant
7890 operations are affected by the digest parameter including the
7891 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7892 DSA key was used because it didn't fix the digest.
7895 *) Initial certificate chain verify code. Currently tests the untrusted
7896 certificates for consistency with the verify purpose (which is set
7897 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7899 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7900 this is because it will reject chains with invalid extensions whereas
7901 every previous version of OpenSSL and SSLeay made no checks at all.
7903 Trust code: checks the root CA for the relevant trust settings. Trust
7904 settings have an initial value consistent with the verify purpose: e.g.
7905 if the verify purpose is for SSL client use it expects the CA to be
7906 trusted for SSL client use. However the default value can be changed to
7907 permit custom trust settings: one example of this would be to only trust
7908 certificates from a specific "secure" set of CAs.
7910 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7911 which should be used for version portability: especially since the
7912 verify structure is likely to change more often now.
7914 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7915 to set them. If not set then assume SSL clients will verify SSL servers
7918 Two new options to the verify program: -untrusted allows a set of
7919 untrusted certificates to be passed in and -purpose which sets the
7920 intended purpose of the certificate. If a purpose is set then the
7921 new chain verify code is used to check extension consistency.
7924 *) Support for the authority information access extension.
7927 *) Modify RSA and DSA PEM read routines to transparently handle
7928 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7929 public keys in a format compatible with certificate
7930 SubjectPublicKeyInfo structures. Unfortunately there were already
7931 functions called *_PublicKey_* which used various odd formats so
7932 these are retained for compatibility: however the DSA variants were
7933 never in a public release so they have been deleted. Changed dsa/rsa
7934 utilities to handle the new format: note no releases ever handled public
7935 keys so we should be OK.
7937 The primary motivation for this change is to avoid the same fiasco
7938 that dogs private keys: there are several incompatible private key
7939 formats some of which are standard and some OpenSSL specific and
7940 require various evil hacks to allow partial transparent handling and
7941 even then it doesn't work with DER formats. Given the option anything
7942 other than PKCS#8 should be dumped: but the other formats have to
7943 stay in the name of compatibility.
7945 With public keys and the benefit of hindsight one standard format
7946 is used which works with EVP_PKEY, RSA or DSA structures: though
7947 it clearly returns an error if you try to read the wrong kind of key.
7949 Added a -pubkey option to the 'x509' utility to output the public key.
7950 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7951 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7952 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7953 that do the same as the EVP_PKEY_assign_*() except they up the
7954 reference count of the added key (they don't "swallow" the
7958 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7959 CRLs would fail if the file contained no certificates or no CRLs:
7960 added a new function to read in both types and return the number
7961 read: this means that if none are read it will be an error. The
7962 DER versions of the certificate and CRL reader would always fail
7963 because it isn't possible to mix certificates and CRLs in DER format
7964 without choking one or the other routine. Changed this to just read
7965 a certificate: this is the best we can do. Also modified the code
7966 in apps/verify.c to take notice of return codes: it was previously
7967 attempting to read in certificates from NULL pointers and ignoring
7968 any errors: this is one reason why the cert and CRL reader seemed
7969 to work. It doesn't check return codes from the default certificate
7970 routines: these may well fail if the certificates aren't installed.
7973 *) Code to support otherName option in GeneralName.
7976 *) First update to verify code. Change the verify utility
7977 so it warns if it is passed a self signed certificate:
7978 for consistency with the normal behaviour. X509_verify
7979 has been modified to it will now verify a self signed
7980 certificate if *exactly* the same certificate appears
7981 in the store: it was previously impossible to trust a
7982 single self signed certificate. This means that:
7983 openssl verify ss.pem
7984 now gives a warning about a self signed certificate but
7985 openssl verify -CAfile ss.pem ss.pem
7989 *) For servers, store verify_result in SSL_SESSION data structure
7990 (and add it to external session representation).
7991 This is needed when client certificate verifications fails,
7992 but an application-provided verification callback (set by
7993 SSL_CTX_set_cert_verify_callback) allows accepting the session
7994 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7995 but returns 1): When the session is reused, we have to set
7996 ssl->verify_result to the appropriate error code to avoid
7998 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8000 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8001 case in PKCS7_dataInit() where the signed PKCS7 structure
8002 didn't contain any existing data because it was being created.
8003 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8005 *) Add a salt to the key derivation routines in enc.c. This
8006 forms the first 8 bytes of the encrypted file. Also add a
8007 -S option to allow a salt to be input on the command line.
8010 *) New function X509_cmp(). Oddly enough there wasn't a function
8011 to compare two certificates. We do this by working out the SHA1
8012 hash and comparing that. X509_cmp() will be needed by the trust
8016 *) SSL_get1_session() is like SSL_get_session(), but increments
8017 the reference count in the SSL_SESSION returned.
8018 [Geoff Thorpe <geoff@eu.c2.net>]
8020 *) Fix for 'req': it was adding a null to request attributes.
8021 Also change the X509_LOOKUP and X509_INFO code to handle
8022 certificate auxiliary information.
8025 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8029 *) Add the possibility to add extra information to the memory leak
8030 detecting output, to form tracebacks, showing from where each
8031 allocation was originated: CRYPTO_push_info("constant string") adds
8032 the string plus current file name and line number to a per-thread
8033 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8034 is like calling CYRPTO_pop_info() until the stack is empty.
8035 Also updated memory leak detection code to be multi-thread-safe.
8038 *) Add options -text and -noout to pkcs7 utility and delete the
8039 encryption options which never did anything. Update docs.
8042 *) Add options to some of the utilities to allow the pass phrase
8043 to be included on either the command line (not recommended on
8044 OSes like Unix) or read from the environment. Update the
8045 manpages and fix a few bugs.
8048 *) Add a few manpages for some of the openssl commands.
8051 *) Fix the -revoke option in ca. It was freeing up memory twice,
8052 leaking and not finding already revoked certificates.
8055 *) Extensive changes to support certificate auxiliary information.
8056 This involves the use of X509_CERT_AUX structure and X509_AUX
8057 functions. An X509_AUX function such as PEM_read_X509_AUX()
8058 can still read in a certificate file in the usual way but it
8059 will also read in any additional "auxiliary information". By
8060 doing things this way a fair degree of compatibility can be
8061 retained: existing certificates can have this information added
8062 using the new 'x509' options.
8064 Current auxiliary information includes an "alias" and some trust
8065 settings. The trust settings will ultimately be used in enhanced
8066 certificate chain verification routines: currently a certificate
8067 can only be trusted if it is self signed and then it is trusted
8071 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8072 The problem was that one of the replacement routines had not been working
8073 since SSLeay releases. For now the offending routine has been replaced
8074 with non-optimised assembler. Even so, this now gives around 95%
8075 performance improvement for 1024 bit RSA signs.
8078 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8079 handling. Most clients have the effective key size in bits equal to
8080 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8081 A few however don't do this and instead use the size of the decrypted key
8082 to determine the RC2 key length and the AlgorithmIdentifier to determine
8083 the effective key length. In this case the effective key length can still
8084 be 40 bits but the key length can be 168 bits for example. This is fixed
8085 by manually forcing an RC2 key into the EVP_PKEY structure because the
8086 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8087 the key length and effective key length are equal.
8090 *) Add a bunch of functions that should simplify the creation of
8091 X509_NAME structures. Now you should be able to do:
8092 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8093 and have it automatically work out the correct field type and fill in
8094 the structures. The more adventurous can try:
8095 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8096 and it will (hopefully) work out the correct multibyte encoding.
8099 *) Change the 'req' utility to use the new field handling and multibyte
8100 copy routines. Before the DN field creation was handled in an ad hoc
8101 way in req, ca, and x509 which was rather broken and didn't support
8102 BMPStrings or UTF8Strings. Since some software doesn't implement
8103 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8104 using the dirstring_type option. See the new comment in the default
8105 openssl.cnf for more info.
8108 *) Make crypto/rand/md_rand.c more robust:
8109 - Assure unique random numbers after fork().
8110 - Make sure that concurrent threads access the global counter and
8111 md serializably so that we never lose entropy in them
8112 or use exactly the same state in multiple threads.
8113 Access to the large state is not always serializable because
8114 the additional locking could be a performance killer, and
8115 md should be large enough anyway.
8118 *) New file apps/app_rand.c with commonly needed functionality
8119 for handling the random seed file.
8121 Use the random seed file in some applications that previously did not:
8123 dsaparam -genkey (which also ignored its '-rand' option),
8126 x509 (when signing).
8127 Except on systems with /dev/urandom, it is crucial to have a random
8128 seed file at least for key creation, DSA signing, and for DH exchanges;
8129 for RSA signatures we could do without one.
8131 gendh and gendsa (unlike genrsa) used to read only the first byte
8132 of each file listed in the '-rand' option. The function as previously
8133 found in genrsa is now in app_rand.c and is used by all programs
8134 that support '-rand'.
8137 *) In RAND_write_file, use mode 0600 for creating files;
8138 don't just chmod when it may be too late.
8141 *) Report an error from X509_STORE_load_locations
8142 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8145 *) New function ASN1_mbstring_copy() this copies a string in either
8146 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8147 into an ASN1_STRING type. A mask of permissible types is passed
8148 and it chooses the "minimal" type to use or an error if not type
8152 *) Add function equivalents to the various macros in asn1.h. The old
8153 macros are retained with an M_ prefix. Code inside the library can
8154 use the M_ macros. External code (including the openssl utility)
8155 should *NOT* in order to be "shared library friendly".
8158 *) Add various functions that can check a certificate's extensions
8159 to see if it usable for various purposes such as SSL client,
8160 server or S/MIME and CAs of these types. This is currently
8161 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8162 verification. Also added a -purpose flag to x509 utility to
8163 print out all the purposes.
8166 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8170 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8171 for, obtain and decode and extension and obtain its critical flag.
8172 This allows all the necessary extension code to be handled in a
8173 single function call.
8176 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8177 platforms. See crypto/rc4/rc4_enc.c for further details.
8180 *) New -noout option to asn1parse. This causes no output to be produced
8181 its main use is when combined with -strparse and -out to extract data
8182 from a file (which may not be in ASN.1 format).
8185 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8186 when producing the local key id.
8187 [Richard Levitte <levitte@stacken.kth.se>]
8189 *) New option -dhparam in s_server. This allows a DH parameter file to be
8190 stated explicitly. If it is not stated then it tries the first server
8191 certificate file. The previous behaviour hard coded the filename
8195 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8196 a public key to be input or output. For example:
8197 openssl rsa -in key.pem -pubout -out pubkey.pem
8198 Also added necessary DSA public key functions to handle this.
8201 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8202 in the message. This was handled by allowing
8203 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8204 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8206 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8207 to the end of the strings whereas this didn't. This would cause problems
8208 if strings read with d2i_ASN1_bytes() were later modified.
8209 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8211 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8212 data and it contains EOF it will end up returning an error. This is
8213 caused by input 46 bytes long. The cause is due to the way base64
8214 BIOs find the start of base64 encoded data. They do this by trying a
8215 trial decode on each line until they find one that works. When they
8216 do a flag is set and it starts again knowing it can pass all the
8217 data directly through the decoder. Unfortunately it doesn't reset
8218 the context it uses. This means that if EOF is reached an attempt
8219 is made to pass two EOFs through the context and this causes the
8220 resulting error. This can also cause other problems as well. As is
8221 usual with these problems it takes *ages* to find and the fix is
8222 trivial: move one line.
8223 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8225 *) Ugly workaround to get s_client and s_server working under Windows. The
8226 old code wouldn't work because it needed to select() on sockets and the
8227 tty (for keypresses and to see if data could be written). Win32 only
8228 supports select() on sockets so we select() with a 1s timeout on the
8229 sockets and then see if any characters are waiting to be read, if none
8230 are present then we retry, we also assume we can always write data to
8231 the tty. This isn't nice because the code then blocks until we've
8232 received a complete line of data and it is effectively polling the
8233 keyboard at 1s intervals: however it's quite a bit better than not
8234 working at all :-) A dedicated Windows application might handle this
8235 with an event loop for example.
8238 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8239 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8240 will be called when RSA_sign() and RSA_verify() are used. This is useful
8241 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8242 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8243 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8244 This necessitated the support of an extra signature type NID_md5_sha1
8245 for SSL signatures and modifications to the SSL library to use it instead
8246 of calling RSA_public_decrypt() and RSA_private_encrypt().
8249 *) Add new -verify -CAfile and -CApath options to the crl program, these
8250 will lookup a CRL issuers certificate and verify the signature in a
8251 similar way to the verify program. Tidy up the crl program so it
8252 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8253 less strict. It will now permit CRL extensions even if it is not
8254 a V2 CRL: this will allow it to tolerate some broken CRLs.
8257 *) Initialize all non-automatic variables each time one of the openssl
8258 sub-programs is started (this is necessary as they may be started
8259 multiple times from the "OpenSSL>" prompt).
8260 [Lennart Bang, Bodo Moeller]
8262 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8263 removing all other RSA functionality (this is what NO_RSA does). This
8264 is so (for example) those in the US can disable those operations covered
8265 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8269 *) Non-copying interface to BIO pairs.
8270 (still largely untested)
8273 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8274 ASCII string. This was handled independently in various places before.
8277 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8278 UTF8 strings a character at a time.
8281 *) Use client_version from client hello to select the protocol
8282 (s23_srvr.c) and for RSA client key exchange verification
8283 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8286 *) Add various utility functions to handle SPKACs, these were previously
8287 handled by poking round in the structure internals. Added new function
8288 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8289 print, verify and generate SPKACs. Based on an original idea from
8290 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8293 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8296 *) Allow the config file extension section to be overwritten on the
8297 command line. Based on an original idea from Massimiliano Pala
8298 <madwolf@comune.modena.it>. The new option is called -extensions
8299 and can be applied to ca, req and x509. Also -reqexts to override
8300 the request extensions in req and -crlexts to override the crl extensions
8304 *) Add new feature to the SPKAC handling in ca. Now you can include
8305 the same field multiple times by preceding it by "XXXX." for example:
8308 this is the same syntax as used in the req config file.
8311 *) Allow certificate extensions to be added to certificate requests. These
8312 are specified in a 'req_extensions' option of the req section of the
8313 config file. They can be printed out with the -text option to req but
8314 are otherwise ignored at present.
8317 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8318 data read consists of only the final block it would not decrypted because
8319 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8320 A misplaced 'break' also meant the decrypted final block might not be
8321 copied until the next read.
8324 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8325 a few extra parameters to the DH structure: these will be useful if
8326 for example we want the value of 'q' or implement X9.42 DH.
8329 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8330 provides hooks that allow the default DSA functions or functions on a
8331 "per key" basis to be replaced. This allows hardware acceleration and
8332 hardware key storage to be handled without major modification to the
8333 library. Also added low level modexp hooks and CRYPTO_EX structure and
8334 associated functions.
8337 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8338 as "read only": it can't be written to and the buffer it points to will
8339 not be freed. Reading from a read only BIO is much more efficient than
8340 a normal memory BIO. This was added because there are several times when
8341 an area of memory needs to be read from a BIO. The previous method was
8342 to create a memory BIO and write the data to it, this results in two
8343 copies of the data and an O(n^2) reading algorithm. There is a new
8344 function BIO_new_mem_buf() which creates a read only memory BIO from
8345 an area of memory. Also modified the PKCS#7 routines to use read only
8349 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8350 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8351 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8352 but a retry condition occured while trying to read the rest.
8355 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8356 NID_pkcs7_encrypted by default: this was wrong since this should almost
8357 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8358 the encrypted data type: this is a more sensible place to put it and it
8359 allows the PKCS#12 code to be tidied up that duplicated this
8363 *) Changed obj_dat.pl script so it takes its input and output files on
8364 the command line. This should avoid shell escape redirection problems
8368 *) Initial support for certificate extension requests, these are included
8369 in things like Xenroll certificate requests. Included functions to allow
8370 extensions to be obtained and added.
8373 *) -crlf option to s_client and s_server for sending newlines as
8374 CRLF (as required by many protocols).
8377 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8379 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8380 [Ralf S. Engelschall]
8382 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8383 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8385 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8389 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8390 DH parameters/keys (q is lost during that conversion, but the resulting
8391 DH parameters contain its length).
8393 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8394 much faster than DH_generate_parameters (which creates parameters
8395 where p = 2*q + 1), and also the smaller q makes DH computations
8396 much more efficient (160-bit exponentiation instead of 1024-bit
8397 exponentiation); so this provides a convenient way to support DHE
8398 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8399 utter importance to use
8400 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8402 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8403 when such DH parameters are used, because otherwise small subgroup
8404 attacks may become possible!
8407 *) Avoid memory leak in i2d_DHparams.
8410 *) Allow the -k option to be used more than once in the enc program:
8411 this allows the same encrypted message to be read by multiple recipients.
8414 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8415 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8416 it will always use the numerical form of the OID, even if it has a short
8420 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8421 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8422 otherwise bn_mod_exp was called. In the case of hardware keys for example
8423 no private key components need be present and it might store extra data
8424 in the RSA structure, which cannot be accessed from bn_mod_exp.
8425 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8426 private key operations.
8429 *) Added support for SPARC Linux.
8432 *) pem_password_cb function type incompatibly changed from
8433 typedef int pem_password_cb(char *buf, int size, int rwflag);
8435 ....(char *buf, int size, int rwflag, void *userdata);
8436 so that applications can pass data to their callbacks:
8437 The PEM[_ASN1]_{read,write}... functions and macros now take an
8438 additional void * argument, which is just handed through whenever
8439 the password callback is called.
8440 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8442 New function SSL_CTX_set_default_passwd_cb_userdata.
8444 Compatibility note: As many C implementations push function arguments
8445 onto the stack in reverse order, the new library version is likely to
8446 interoperate with programs that have been compiled with the old
8447 pem_password_cb definition (PEM_whatever takes some data that
8448 happens to be on the stack as its last argument, and the callback
8449 just ignores this garbage); but there is no guarantee whatsoever that
8452 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8453 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8454 problems not only on Windows, but also on some Unix platforms.
8455 To avoid problematic command lines, these definitions are now in an
8456 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8457 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8460 *) MIPS III/IV assembler module is reimplemented.
8463 *) More DES library cleanups: remove references to srand/rand and
8464 delete an unused file.
8467 *) Add support for the the free Netwide assembler (NASM) under Win32,
8468 since not many people have MASM (ml) and it can be hard to obtain.
8469 This is currently experimental but it seems to work OK and pass all
8470 the tests. Check out INSTALL.W32 for info.
8473 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8474 without temporary keys kept an extra copy of the server key,
8475 and connections with temporary keys did not free everything in case
8479 *) New function RSA_check_key and new openssl rsa option -check
8480 for verifying the consistency of RSA keys.
8481 [Ulf Moeller, Bodo Moeller]
8483 *) Various changes to make Win32 compile work:
8484 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8485 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8486 comparison" warnings.
8487 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8490 *) Add a debugging option to PKCS#5 v2 key generation function: when
8491 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8492 derived keys are printed to stderr.
8495 *) Copy the flags in ASN1_STRING_dup().
8496 [Roman E. Pavlov <pre@mo.msk.ru>]
8498 *) The x509 application mishandled signing requests containing DSA
8499 keys when the signing key was also DSA and the parameters didn't match.
8501 It was supposed to omit the parameters when they matched the signing key:
8502 the verifying software was then supposed to automatically use the CA's
8503 parameters if they were absent from the end user certificate.
8505 Omitting parameters is no longer recommended. The test was also
8506 the wrong way round! This was probably due to unusual behaviour in
8507 EVP_cmp_parameters() which returns 1 if the parameters match.
8508 This meant that parameters were omitted when they *didn't* match and
8509 the certificate was useless. Certificates signed with 'ca' didn't have
8511 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8513 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8514 The interface is as follows:
8515 Applications can use
8516 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8517 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8518 "off" is now the default.
8519 The library internally uses
8520 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8521 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8522 to disable memory-checking temporarily.
8524 Some inconsistent states that previously were possible (and were
8525 even the default) are now avoided.
8527 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8528 with each memory chunk allocated; this is occasionally more helpful
8529 than just having a counter.
8531 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8533 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8537 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8538 which largely parallels "options", but is for changing API behaviour,
8539 whereas "options" are about protocol behaviour.
8540 Initial "mode" flags are:
8542 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8543 a single record has been written.
8544 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8545 retries use the same buffer location.
8546 (But all of the contents must be
8550 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8553 *) Fix problems with no-hmac etc.
8554 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8556 *) New functions RSA_get_default_method(), RSA_set_method() and
8557 RSA_get_method(). These allows replacement of RSA_METHODs without having
8558 to mess around with the internals of an RSA structure.
8561 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8562 Also really enable memory leak checks in openssl.c and in some
8564 [Chad C. Mulligan, Bodo Moeller]
8566 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8567 up the length of negative integers. This has now been simplified to just
8568 store the length when it is first determined and use it later, rather
8569 than trying to keep track of where data is copied and updating it to
8571 [Steve Henson, reported by Brien Wheeler
8572 <bwheeler@authentica-security.com>]
8574 *) Add a new function PKCS7_signatureVerify. This allows the verification
8575 of a PKCS#7 signature but with the signing certificate passed to the
8576 function itself. This contrasts with PKCS7_dataVerify which assumes the
8577 certificate is present in the PKCS#7 structure. This isn't always the
8578 case: certificates can be omitted from a PKCS#7 structure and be
8579 distributed by "out of band" means (such as a certificate database).
8582 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8583 function prototypes in pem.h, also change util/mkdef.pl to add the
8584 necessary function names.
8587 *) mk1mf.pl (used by Windows builds) did not properly read the
8588 options set by Configure in the top level Makefile, and Configure
8589 was not even able to write more than one option correctly.
8590 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8593 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8594 file to be loaded from a BIO or FILE pointer. The BIO version will
8595 for example allow memory BIOs to contain config info.
8598 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8599 Whoever hopes to achieve shared-library compatibility across versions
8600 must use this, not the compile-time macro.
8601 (Exercise 0.9.4: Which is the minimum library version required by
8603 Note: All this applies only to multi-threaded programs, others don't
8607 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8608 through a BIO pair triggered the default case, i.e.
8609 SSLerr(...,SSL_R_UNKNOWN_STATE).
8612 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8613 can use the SSL library even if none of the specific BIOs is
8617 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8618 for the encoded length.
8619 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8621 *) Add initial documentation of the X509V3 functions.
8624 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8625 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8626 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8627 secure PKCS#8 private key format with a high iteration count.
8630 *) Fix determination of Perl interpreter: A perl or perl5
8631 _directory_ in $PATH was also accepted as the interpreter.
8632 [Ralf S. Engelschall]
8634 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8635 wrong with it but it was very old and did things like calling
8636 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8640 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8641 to use the new extension code.
8644 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8645 with macros. This should make it easier to change their form, add extra
8646 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8650 *) Add to configuration table a new entry that can specify an alternative
8651 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8652 according to Mark Crispin <MRC@Panda.COM>.
8656 *) DES CBC did not update the IV. Weird.
8659 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8660 Changing the behaviour of the former might break existing programs --
8661 where IV updating is needed, des_ncbc_encrypt can be used.
8664 *) When bntest is run from "make test" it drives bc to check its
8665 calculations, as well as internally checking them. If an internal check
8666 fails, it needs to cause bc to give a non-zero result or make test carries
8667 on without noticing the failure. Fixed.
8670 *) DES library cleanups.
8673 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8674 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8675 ciphers. NOTE: although the key derivation function has been verified
8676 against some published test vectors it has not been extensively tested
8677 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8681 *) Instead of "mkdir -p", which is not fully portable, use new
8682 Perl script "util/mkdir-p.pl".
8685 *) Rewrite the way password based encryption (PBE) is handled. It used to
8686 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8687 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8688 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8689 the 'parameter' field of the AlgorithmIdentifier is passed to the
8690 underlying key generation function so it must do its own ASN1 parsing.
8691 This has also changed the EVP_PBE_CipherInit() function which now has a
8692 'parameter' argument instead of literal salt and iteration count values
8693 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8696 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8697 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8698 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8699 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8700 value was just used as a "magic string" and not used directly its
8701 value doesn't matter.
8704 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8708 *) "linux-sparc64" configuration (ultrapenguin).
8709 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8710 "linux-sparc" configuration.
8711 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8713 *) config now generates no-xxx options for missing ciphers.
8716 *) Support the EBCDIC character set (work in progress).
8717 File ebcdic.c not yet included because it has a different license.
8718 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8720 *) Support BS2000/OSD-POSIX.
8721 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8723 *) Make callbacks for key generation use void * instead of char *.
8726 *) Make S/MIME samples compile (not yet tested).
8729 *) Additional typesafe stacks.
8732 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8736 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8738 *) New configuration variant "sco5-gcc".
8740 *) Updated some demos.
8741 [Sean O Riordain, Wade Scholine]
8743 *) Add missing BIO_free at exit of pkcs12 application.
8746 *) Fix memory leak in conf.c.
8749 *) Updates for Win32 to assembler version of MD5.
8752 *) Set #! path to perl in apps/der_chop to where we found it
8753 instead of using a fixed path.
8756 *) SHA library changes for irix64-mips4-cc.
8759 *) Improvements for VMS support.
8763 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8765 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8766 This also avoids the problems with SC4.2 and unpatched SC5.
8767 [Andy Polyakov <appro@fy.chalmers.se>]
8769 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8770 These are required because of the typesafe stack would otherwise break
8771 existing code. If old code used a structure member which used to be STACK
8772 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8773 sk_num or sk_value it would produce an error because the num, data members
8774 are not present in STACK_OF. Now it just produces a warning. sk_set
8775 replaces the old method of assigning a value to sk_value
8776 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8777 that does this will no longer work (and should use sk_set instead) but
8778 this could be regarded as a "questionable" behaviour anyway.
8781 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8782 correctly handle encrypted S/MIME data.
8785 *) Change type of various DES function arguments from des_cblock
8786 (which means, in function argument declarations, pointer to char)
8787 to des_cblock * (meaning pointer to array with 8 char elements),
8788 which allows the compiler to do more typechecking; it was like
8789 that back in SSLeay, but with lots of ugly casts.
8791 Introduce new type const_des_cblock.
8794 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8795 problems: find RecipientInfo structure that matches recipient certificate
8796 and initialise the ASN1 structures properly based on passed cipher.
8799 *) Belatedly make the BN tests actually check the results.
8802 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8803 to and from BNs: it was completely broken. New compilation option
8804 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8805 key elements as negative integers.
8808 *) Reorganize and speed up MD5.
8809 [Andy Polyakov <appro@fy.chalmers.se>]
8812 [Richard Levitte <richard@levitte.org>]
8814 *) New option -out to asn1parse to allow the parsed structure to be
8815 output to a file. This is most useful when combined with the -strparse
8816 option to examine the output of things like OCTET STRINGS.
8819 *) Make SSL library a little more fool-proof by not requiring any longer
8820 that SSL_set_{accept,connect}_state be called before
8821 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8822 in many applications because usually everything *appeared* to work as
8823 intended anyway -- now it really works as intended).
8826 *) Move openssl.cnf out of lib/.
8829 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8830 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8831 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8832 [Ralf S. Engelschall]
8834 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8835 handle PKCS#7 enveloped data properly.
8836 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8838 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8839 copying pointers. The cert_st handling is changed by this in
8840 various ways (and thus what used to be known as ctx->default_cert
8841 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8842 any longer when s->cert does not give us what we need).
8843 ssl_cert_instantiate becomes obsolete by this change.
8844 As soon as we've got the new code right (possibly it already is?),
8845 we have solved a couple of bugs of the earlier code where s->cert
8846 was used as if it could not have been shared with other SSL structures.
8848 Note that using the SSL API in certain dirty ways now will result
8849 in different behaviour than observed with earlier library versions:
8850 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8851 does not influence s as it used to.
8853 In order to clean up things more thoroughly, inside SSL_SESSION
8854 we don't use CERT any longer, but a new structure SESS_CERT
8855 that holds per-session data (if available); currently, this is
8856 the peer's certificate chain and, for clients, the server's certificate
8857 and temporary key. CERT holds only those values that can have
8858 meaningful defaults in an SSL_CTX.
8861 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8862 from the internal representation. Various PKCS#7 fixes: remove some
8863 evil casts and set the enc_dig_alg field properly based on the signing
8867 *) Allow PKCS#12 password to be set from the command line or the
8868 environment. Let 'ca' get its config file name from the environment
8869 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8873 *) Allow certificate policies extension to use an IA5STRING for the
8874 organization field. This is contrary to the PKIX definition but
8875 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8879 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8880 without disallowing inline assembler and the like for non-pedantic builds.
8883 *) Support Borland C++ builder.
8884 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8889 *) SHA-1 cleanups and performance enhancements.
8890 [Andy Polyakov <appro@fy.chalmers.se>]
8892 *) Sparc v8plus assembler for the bignum library.
8893 [Andy Polyakov <appro@fy.chalmers.se>]
8895 *) Accept any -xxx and +xxx compiler options in Configure.
8898 *) Update HPUX configuration.
8901 *) Add missing sk_<type>_unshift() function to safestack.h
8902 [Ralf S. Engelschall]
8904 *) New function SSL_CTX_use_certificate_chain_file that sets the
8905 "extra_cert"s in addition to the certificate. (This makes sense
8906 only for "PEM" format files, as chains as a whole are not
8910 *) Support verify_depth from the SSL API.
8911 x509_vfy.c had what can be considered an off-by-one-error:
8912 Its depth (which was not part of the external interface)
8913 was actually counting the number of certificates in a chain;
8914 now it really counts the depth.
8917 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8918 instead of X509err, which often resulted in confusing error
8919 messages since the error codes are not globally unique
8920 (e.g. an alleged error in ssl3_accept when a certificate
8921 didn't match the private key).
8923 *) New function SSL_CTX_set_session_id_context that allows to set a default
8924 value (so that you don't need SSL_set_session_id_context for each
8925 connection using the SSL_CTX).
8928 *) OAEP decoding bug fix.
8931 *) Support INSTALL_PREFIX for package builders, as proposed by
8935 *) New Configure options "threads" and "no-threads". For systems
8936 where the proper compiler options are known (currently Solaris
8937 and Linux), "threads" is the default.
8940 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8943 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8944 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8945 such as /usr/local/bin.
8948 *) "make linux-shared" to build shared libraries.
8949 [Niels Poppe <niels@netbox.org>]
8951 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8954 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8955 extension adding in x509 utility.
8958 *) Remove NOPROTO sections and error code comments.
8961 *) Partial rewrite of the DEF file generator to now parse the ANSI
8965 *) New Configure options --prefix=DIR and --openssldir=DIR.
8968 *) Complete rewrite of the error code script(s). It is all now handled
8969 by one script at the top level which handles error code gathering,
8970 header rewriting and C source file generation. It should be much better
8971 than the old method: it now uses a modified version of Ulf's parser to
8972 read the ANSI prototypes in all header files (thus the old K&R definitions
8973 aren't needed for error creation any more) and do a better job of
8974 translating function codes into names. The old 'ASN1 error code imbedded
8975 in a comment' is no longer necessary and it doesn't use .err files which
8976 have now been deleted. Also the error code call doesn't have to appear all
8977 on one line (which resulted in some large lines...).
8980 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8983 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8984 0 (which usually indicates a closed connection), but continue reading.
8987 *) Fix some race conditions.
8990 *) Add support for CRL distribution points extension. Add Certificate
8991 Policies and CRL distribution points documentation.
8994 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8997 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8998 8 of keying material. Merlin has also confirmed interop with this fix
8999 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9000 [Merlin Hughes <merlin@baltimore.ie>]
9002 *) Fix lots of warnings.
9003 [Richard Levitte <levitte@stacken.kth.se>]
9005 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9006 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9007 [Richard Levitte <levitte@stacken.kth.se>]
9009 *) Fix problems with sizeof(long) == 8.
9010 [Andy Polyakov <appro@fy.chalmers.se>]
9012 *) Change functions to ANSI C.
9015 *) Fix typos in error codes.
9016 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9018 *) Remove defunct assembler files from Configure.
9021 *) SPARC v8 assembler BIGNUM implementation.
9022 [Andy Polyakov <appro@fy.chalmers.se>]
9024 *) Support for Certificate Policies extension: both print and set.
9025 Various additions to support the r2i method this uses.
9028 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9029 return a const string when you are expecting an allocated buffer.
9032 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9033 types DirectoryString and DisplayText.
9036 *) Add code to allow r2i extensions to access the configuration database,
9037 add an LHASH database driver and add several ctx helper functions.
9040 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9041 fail when they extended the size of a BIGNUM.
9044 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9045 support typesafe stack.
9048 *) Fix typo in SSL_[gs]et_options().
9049 [Nils Frostberg <nils@medcom.se>]
9051 *) Delete various functions and files that belonged to the (now obsolete)
9052 old X509V3 handling code.
9055 *) New Configure option "rsaref".
9058 *) Don't auto-generate pem.h.
9061 *) Introduce type-safe ASN.1 SETs.
9064 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9065 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9067 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9068 that links with OpenSSL (well at least cause lots of warnings), but fear
9069 not: the conversion is trivial, and it eliminates loads of evil casts. A
9070 few STACKed things have been converted already. Feel free to convert more.
9071 In the fullness of time, I'll do away with the STACK type altogether.
9074 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9075 specified in <certfile> by updating the entry in the index.txt file.
9076 This way one no longer has to edit the index.txt file manually for
9077 revoking a certificate. The -revoke option does the gory details now.
9078 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9080 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9081 `-text' option at all and this way the `-noout -text' combination was
9082 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9083 [Ralf S. Engelschall]
9085 *) Make sure a corresponding plain text error message exists for the
9086 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9087 verify callback function determined that a certificate was revoked.
9088 [Ralf S. Engelschall]
9090 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9091 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9092 all available cipers including rc5, which was forgotten until now.
9093 In order to let the testing shell script know which algorithms
9094 are available, a new (up to now undocumented) command
9095 "openssl list-cipher-commands" is used.
9098 *) Bugfix: s_client occasionally would sleep in select() when
9099 it should have checked SSL_pending() first.
9102 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9103 the raw DSA values prior to ASN.1 encoding.
9106 *) Tweaks to Configure
9107 [Niels Poppe <niels@netbox.org>]
9109 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9113 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9116 *) New config option to avoid instructions that are illegal on the 80386.
9117 The default code is faster, but requires at least a 486.
9120 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9121 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9122 same as SSL2_VERSION anyway.
9125 *) New "-showcerts" option for s_client.
9128 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9129 application. Various cleanups and fixes.
9132 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9133 modify error routines to work internally. Add error codes and PBE init
9134 to library startup routines.
9137 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9138 packing functions to asn1 and evp. Changed function names and error
9139 codes along the way.
9142 *) PKCS12 integration: and so it begins... First of several patches to
9143 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9144 objects to objects.h
9147 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9148 and display support for Thawte strong extranet extension.
9151 *) Add LinuxPPC support.
9152 [Jeff Dubrule <igor@pobox.org>]
9154 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9155 bn_div_words in alpha.s.
9156 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9158 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9159 OAEP isn't supported when OpenSSL is built with RSAref.
9160 [Ulf Moeller <ulf@fitug.de>]
9162 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9163 so they no longer are missing under -DNOPROTO.
9164 [Soren S. Jorvang <soren@t.dk>]
9167 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9169 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9170 doesn't work when the session is reused. Coming soon!
9173 *) Fix a security hole, that allows sessions to be reused in the wrong
9174 context thus bypassing client cert protection! All software that uses
9175 client certs and session caches in multiple contexts NEEDS PATCHING to
9176 allow session reuse! A fuller solution is in the works.
9177 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9179 *) Some more source tree cleanups (removed obsolete files
9180 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9181 permission on "config" script to be executable) and a fix for the INSTALL
9183 [Ulf Moeller <ulf@fitug.de>]
9185 *) Remove some legacy and erroneous uses of malloc, free instead of
9187 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9189 *) Make rsa_oaep_test return non-zero on error.
9190 [Ulf Moeller <ulf@fitug.de>]
9192 *) Add support for native Solaris shared libraries. Configure
9193 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9194 if someone would make that last step automatic.
9195 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9197 *) ctx_size was not built with the right compiler during "make links". Fixed.
9200 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9201 except NULL ciphers". This means the default cipher list will no longer
9202 enable NULL ciphers. They need to be specifically enabled e.g. with
9203 the string "DEFAULT:eNULL".
9206 *) Fix to RSA private encryption routines: if p < q then it would
9207 occasionally produce an invalid result. This will only happen with
9208 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9211 *) Be less restrictive and allow also `perl util/perlpath.pl
9212 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9213 because this way one can also use an interpreter named `perl5' (which is
9214 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9215 installed as `perl').
9216 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9218 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9219 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9221 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9222 advapi32.lib to Win32 build and change the pem test comparision
9223 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9224 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9225 and crypto/des/ede_cbcm_enc.c.
9228 *) DES quad checksum was broken on big-endian architectures. Fixed.
9231 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9232 Win32 test batch file so it (might) work again. The Win32 test batch file
9233 is horrible: I feel ill....
9236 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9237 in e_os.h. Audit of header files to check ANSI and non ANSI
9238 sections: 10 functions were absent from non ANSI section and not exported
9239 from Windows DLLs. Fixed up libeay.num for new functions.
9242 *) Make `openssl version' output lines consistent.
9243 [Ralf S. Engelschall]
9245 *) Fix Win32 symbol export lists for BIO functions: Added
9246 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9247 to ms/libeay{16,32}.def.
9248 [Ralf S. Engelschall]
9250 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9251 fine under Unix and passes some trivial tests I've now added. But the
9252 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9253 added to make sure no one expects that this stuff really works in the
9254 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9255 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9257 [Ralf S. Engelschall]
9259 *) Fix the generation of two part addresses in perl.
9260 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9262 *) Add config entry for Linux on MIPS.
9263 [John Tobey <jtobey@channel1.com>]
9265 *) Make links whenever Configure is run, unless we are on Windoze.
9268 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9269 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9273 *) Add a useful kludge to allow package maintainers to specify compiler and
9274 other platforms details on the command line without having to patch the
9275 Configure script everytime: One now can use ``perl Configure
9276 <id>:<details>'', i.e. platform ids are allowed to have details appended
9277 to them (seperated by colons). This is treated as there would be a static
9278 pre-configured entry in Configure's %table under key <id> with value
9279 <details> and ``perl Configure <id>'' is called. So, when you want to
9280 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9281 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9282 now, which overrides the FreeBSD-elf entry on-the-fly.
9283 [Ralf S. Engelschall]
9285 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9288 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9289 on the `perl Configure ...' command line. This way one can compile
9290 OpenSSL libraries with Position Independent Code (PIC) which is needed
9291 for linking it into DSOs.
9292 [Ralf S. Engelschall]
9294 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9298 *) Cleaned up the LICENSE document: The official contact for any license
9299 questions now is the OpenSSL core team under openssl-core@openssl.org.
9300 And add a paragraph about the dual-license situation to make sure people
9301 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9302 to the OpenSSL toolkit.
9303 [Ralf S. Engelschall]
9305 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9306 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9307 Additonally cleaned up the `make links' target: Remove unnecessary
9308 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9309 to speed processing and no longer clutter the display with confusing
9310 stuff. Instead only the actually done links are displayed.
9311 [Ralf S. Engelschall]
9313 *) Permit null encryption ciphersuites, used for authentication only. It used
9314 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9315 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9319 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9320 signed attributes when verifying signatures (this would break them),
9321 the detached data encoding was wrong and public keys obtained using
9322 X509_get_pubkey() weren't freed.
9325 *) Add text documentation for the BUFFER functions. Also added a work around
9326 to a Win95 console bug. This was triggered by the password read stuff: the
9327 last character typed gets carried over to the next fread(). If you were
9328 generating a new cert request using 'req' for example then the last
9329 character of the passphrase would be CR which would then enter the first
9333 *) Added the new `Includes OpenSSL Cryptography Software' button as
9334 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9335 button and can be used by applications based on OpenSSL to show the
9336 relationship to the OpenSSL project.
9337 [Ralf S. Engelschall]
9339 *) Remove confusing variables in function signatures in files
9340 ssl/ssl_lib.c and ssl/ssl.h.
9341 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9343 *) Don't install bss_file.c under PREFIX/include/
9344 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9346 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9347 functions that return function pointers and has support for NT specific
9348 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9349 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9350 unsigned to signed types: this was killing the Win32 compile.
9353 *) Add new certificate file to stack functions,
9354 SSL_add_dir_cert_subjects_to_stack() and
9355 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9356 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9357 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9358 This means that Apache-SSL and similar packages don't have to mess around
9359 to add as many CAs as they want to the preferred list.
9362 *) Experiment with doxygen documentation. Currently only partially applied to
9364 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9365 openssl.doxy as the configuration file.
9368 *) Get rid of remaining C++-style comments which strict C compilers hate.
9369 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9371 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9372 compiled in by default: it has problems with large keys.
9375 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9376 DH private keys and/or callback functions which directly correspond to
9377 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9378 is needed for applications which have to configure certificates on a
9379 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9381 For the RSA certificate situation is makes no difference, but
9382 for the DSA certificate situation this fixes the "no shared cipher"
9383 problem where the OpenSSL cipher selection procedure failed because the
9384 temporary keys were not overtaken from the context and the API provided
9385 no way to reconfigure them.
9386 The new functions now let applications reconfigure the stuff and they
9387 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9388 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9389 non-public-API function ssl_cert_instantiate() is used as a helper
9390 function and also to reduce code redundancy inside ssl_rsa.c.
9391 [Ralf S. Engelschall]
9393 *) Move s_server -dcert and -dkey options out of the undocumented feature
9394 area because they are useful for the DSA situation and should be
9395 recognized by the users.
9396 [Ralf S. Engelschall]
9398 *) Fix the cipher decision scheme for export ciphers: the export bits are
9399 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9400 SSL_EXP_MASK. So, the original variable has to be used instead of the
9401 already masked variable.
9402 [Richard Levitte <levitte@stacken.kth.se>]
9404 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9405 [Richard Levitte <levitte@stacken.kth.se>]
9407 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9408 from `int' to `unsigned int' because it's a length and initialized by
9409 EVP_DigestFinal() which expects an `unsigned int *'.
9410 [Richard Levitte <levitte@stacken.kth.se>]
9412 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9413 script. Instead use the usual Shell->Perl transition trick.
9414 [Ralf S. Engelschall]
9416 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9417 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9418 -noout -modulus' as it's already the case for `openssl rsa -noout
9419 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9420 currently the public key is printed (a decision which was already done by
9421 `openssl dsa -modulus' in the past) which serves a similar purpose.
9422 Additionally the NO_RSA no longer completely removes the whole -modulus
9423 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9425 [Ralf S. Engelschall]
9427 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9428 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9429 [Arne Ansper <arne@ats.cyber.ee>]
9431 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9432 to be added. Now both 'req' and 'ca' can use new objects defined in the
9436 *) Add cool BIO that does syslog (or event log on NT).
9437 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9439 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9440 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9441 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9442 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9445 *) Add preliminary config info for new extension code.
9448 *) Make RSA_NO_PADDING really use no padding.
9449 [Ulf Moeller <ulf@fitug.de>]
9451 *) Generate errors when private/public key check is done.
9454 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9455 for some CRL extensions and new objects added.
9458 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9459 key usage extension and fuller support for authority key id.
9462 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9463 padding method for RSA, which is recommended for new applications in PKCS
9464 #1 v2.0 (RFC 2437, October 1998).
9465 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9466 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9467 against Bleichbacher's attack on RSA.
9468 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9471 *) Updates to the new SSL compression code
9472 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9474 *) Fix so that the version number in the master secret, when passed
9475 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9476 (because the server will not accept higher), that the version number
9477 is 0x03,0x01, not 0x03,0x00
9478 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9480 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9481 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9482 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9485 *) Support for RAW extensions where an arbitrary extension can be
9486 created by including its DER encoding. See apps/openssl.cnf for
9490 *) Make sure latest Perl versions don't interpret some generated C array
9491 code as Perl array code in the crypto/err/err_genc.pl script.
9492 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9494 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9495 not many people have the assembler. Various Win32 compilation fixes and
9496 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9500 *) Modify configure script 'Configure' to automatically create crypto/date.h
9501 file under Win32 and also build pem.h from pem.org. New script
9502 util/mkfiles.pl to create the MINFO file on environments that can't do a
9503 'make files': perl util/mkfiles.pl >MINFO should work.
9506 *) Major rework of DES function declarations, in the pursuit of correctness
9507 and purity. As a result, many evil casts evaporated, and some weirdness,
9508 too. You may find this causes warnings in your code. Zapping your evil
9509 casts will probably fix them. Mostly.
9512 *) Fix for a typo in asn1.h. Bug fix to object creation script
9513 obj_dat.pl. It considered a zero in an object definition to mean
9514 "end of object": none of the objects in objects.h have any zeros
9515 so it wasn't spotted.
9516 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9518 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9519 Masking (CBCM). In the absence of test vectors, the best I have been able
9520 to do is check that the decrypt undoes the encrypt, so far. Send me test
9521 vectors if you have them.
9524 *) Correct calculation of key length for export ciphers (too much space was
9525 allocated for null ciphers). This has not been tested!
9528 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9529 message is now correct (it understands "crypto" and "ssl" on its
9530 command line). There is also now an "update" option. This will update
9531 the util/ssleay.num and util/libeay.num files with any new functions.
9533 perl util/mkdef.pl crypto ssl update
9534 it will update them.
9537 *) Overhauled the Perl interface (perl/*):
9538 - ported BN stuff to OpenSSL's different BN library
9539 - made the perl/ source tree CVS-aware
9540 - renamed the package from SSLeay to OpenSSL (the files still contain
9541 their history because I've copied them in the repository)
9542 - removed obsolete files (the test scripts will be replaced
9543 by better Test::Harness variants in the future)
9544 [Ralf S. Engelschall]
9546 *) First cut for a very conservative source tree cleanup:
9547 1. merge various obsolete readme texts into doc/ssleay.txt
9548 where we collect the old documents and readme texts.
9549 2. remove the first part of files where I'm already sure that we no
9550 longer need them because of three reasons: either they are just temporary
9551 files which were left by Eric or they are preserved original files where
9552 I've verified that the diff is also available in the CVS via "cvs diff
9553 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9554 the crypto/md/ stuff).
9555 [Ralf S. Engelschall]
9557 *) More extension code. Incomplete support for subject and issuer alt
9558 name, issuer and authority key id. Change the i2v function parameters
9559 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9560 what that's for :-) Fix to ASN1 macro which messed up
9561 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9564 *) Preliminary support for ENUMERATED type. This is largely copied from the
9568 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9569 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9571 *) Make sure `make rehash' target really finds the `openssl' program.
9572 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9574 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9575 like to hear about it if this slows down other processors.
9578 *) Add CygWin32 platform information to Configure script.
9579 [Alan Batie <batie@aahz.jf.intel.com>]
9581 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9582 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9584 *) New program nseq to manipulate netscape certificate sequences
9587 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9591 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9592 but the BN code had some problems that would cause failures when
9593 doing certificate verification and some other functions.
9594 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9596 *) Add ASN1 and PEM code to support netscape certificate sequences.
9599 *) Add ASN1 and PEM code to support netscape certificate sequences.
9602 *) Add several PKIX and private extended key usage OIDs.
9605 *) Modify the 'ca' program to handle the new extension code. Modify
9606 openssl.cnf for new extension format, add comments.
9609 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9610 and add a sample to openssl.cnf so req -x509 now adds appropriate
9614 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9615 error code, add initial support to X509_print() and x509 application.
9618 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9619 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9620 stuff is currently isolated and isn't even compiled yet.
9623 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9624 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9625 Removed the versions check from X509 routines when loading extensions:
9626 this allows certain broken certificates that don't set the version
9627 properly to be processed.
9630 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9631 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9632 can still be regenerated with "make depend".
9635 *) Spelling mistake in C version of CAST-128.
9636 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9638 *) Changes to the error generation code. The perl script err-code.pl
9639 now reads in the old error codes and retains the old numbers, only
9640 adding new ones if necessary. It also only changes the .err files if new
9641 codes are added. The makefiles have been modified to only insert errors
9642 when needed (to avoid needlessly modifying header files). This is done
9643 by only inserting errors if the .err file is newer than the auto generated
9644 C file. To rebuild all the error codes from scratch (the old behaviour)
9645 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9646 or delete all the .err files.
9649 *) CAST-128 was incorrectly implemented for short keys. The C version has
9650 been fixed, but is untested. The assembler versions are also fixed, but
9651 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9652 to regenerate it if needed.
9653 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9654 Hagino <itojun@kame.net>]
9656 *) File was opened incorrectly in randfile.c.
9657 [Ulf Möller <ulf@fitug.de>]
9659 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9660 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9661 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9662 al: it's just almost always a UTCTime. Note this patch adds new error
9663 codes so do a "make errors" if there are problems.
9666 *) Correct Linux 1 recognition in config.
9667 [Ulf Möller <ulf@fitug.de>]
9669 *) Remove pointless MD5 hash when using DSA keys in ca.
9670 [Anonymous <nobody@replay.com>]
9672 *) Generate an error if given an empty string as a cert directory. Also
9673 generate an error if handed NULL (previously returned 0 to indicate an
9674 error, but didn't set one).
9675 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9677 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9680 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9681 parameters. This was causing a warning which killed off the Win32 compile.
9684 *) Remove C++ style comments from crypto/bn/bn_local.h.
9685 [Neil Costigan <neil.costigan@celocom.com>]
9687 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9688 based on a text string, looking up short and long names and finally
9689 "dot" format. The "dot" format stuff didn't work. Added new function
9690 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9691 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9692 OID is not part of the table.
9695 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9696 X509_LOOKUP_by_alias().
9699 *) Sort openssl functions by name.
9702 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9703 encryption from sample DSA keys (in case anyone is interested the password
9707 *) Make _all_ *_free functions accept a NULL pointer.
9708 [Frans Heymans <fheymans@isaserver.be>]
9710 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9712 [Anonymous <nobody@replay.com>]
9714 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9715 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9717 *) Don't blow it for numeric -newkey arguments to apps/req.
9718 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9720 *) Temp key "for export" tests were wrong in s3_srvr.c.
9721 [Anonymous <nobody@replay.com>]
9723 *) Add prototype for temp key callback functions
9724 SSL_CTX_set_tmp_{rsa,dh}_callback().
9727 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9728 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9731 *) X509_name_add_entry() freed the wrong thing after an error.
9732 [Arne Ansper <arne@ats.cyber.ee>]
9734 *) rsa_eay.c would attempt to free a NULL context.
9735 [Arne Ansper <arne@ats.cyber.ee>]
9737 *) BIO_s_socket() had a broken should_retry() on Windoze.
9738 [Arne Ansper <arne@ats.cyber.ee>]
9740 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9741 [Arne Ansper <arne@ats.cyber.ee>]
9743 *) Make sure the already existing X509_STORE->depth variable is initialized
9744 in X509_STORE_new(), but document the fact that this variable is still
9745 unused in the certificate verification process.
9746 [Ralf S. Engelschall]
9748 *) Fix the various library and apps files to free up pkeys obtained from
9749 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9752 *) Fix reference counting in X509_PUBKEY_get(). This makes
9753 demos/maurice/example2.c work, amongst others, probably.
9754 [Steve Henson and Ben Laurie]
9756 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9757 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9758 are no longer created. This way we have a single and consistent command
9759 line interface `openssl <command>', similar to `cvs <command>'.
9760 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9762 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9763 BIT STRING wrapper always have zero unused bits.
9766 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9769 *) Make the top-level INSTALL documentation easier to understand.
9772 *) Makefiles updated to exit if an error occurs in a sub-directory
9773 make (including if user presses ^C) [Paul Sutton]
9775 *) Make Montgomery context stuff explicit in RSA data structure.
9778 *) Fix build order of pem and err to allow for generated pem.h.
9781 *) Fix renumbering bug in X509_NAME_delete_entry().
9784 *) Enhanced the err-ins.pl script so it makes the error library number
9785 global and can add a library name. This is needed for external ASN1 and
9786 other error libraries.
9789 *) Fixed sk_insert which never worked properly.
9792 *) Fix ASN1 macros so they can handle indefinite length construted
9793 EXPLICIT tags. Some non standard certificates use these: they can now
9797 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9798 into a single doc/ssleay.txt bundle. This way the information is still
9799 preserved but no longer messes up this directory. Now it's new room for
9800 the new set of documenation files.
9801 [Ralf S. Engelschall]
9803 *) SETs were incorrectly DER encoded. This was a major pain, because they
9804 shared code with SEQUENCEs, which aren't coded the same. This means that
9805 almost everything to do with SETs or SEQUENCEs has either changed name or
9806 number of arguments.
9807 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9809 *) Fix test data to work with the above.
9812 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9813 was already fixed by Eric for 0.9.1 it seems.
9814 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9816 *) Autodetect FreeBSD3.
9819 *) Fix various bugs in Configure. This affects the following platforms:
9823 unixware-2.0-pentium
9827 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9828 before they are needed.
9831 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9835 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9837 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9838 changed SSLeay to OpenSSL in version strings.
9839 [Ralf S. Engelschall]
9841 *) Some fixups to the top-level documents.
9844 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9845 because the symlink to include/ was missing.
9846 [Ralf S. Engelschall]
9848 *) Incorporated the popular no-RSA/DSA-only patches
9849 which allow to compile a RSA-free SSLeay.
9850 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9852 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9853 when "ssleay" is still not found.
9854 [Ralf S. Engelschall]
9856 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9857 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9859 *) Updated the README file.
9860 [Ralf S. Engelschall]
9862 *) Added various .cvsignore files in the CVS repository subdirs
9863 to make a "cvs update" really silent.
9864 [Ralf S. Engelschall]
9866 *) Recompiled the error-definition header files and added
9867 missing symbols to the Win32 linker tables.
9868 [Ralf S. Engelschall]
9870 *) Cleaned up the top-level documents;
9871 o new files: CHANGES and LICENSE
9872 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9873 o merged COPYRIGHT into LICENSE
9874 o removed obsolete TODO file
9875 o renamed MICROSOFT to INSTALL.W32
9876 [Ralf S. Engelschall]
9878 *) Removed dummy files from the 0.9.1b source tree:
9879 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9880 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9881 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9882 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9883 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9884 [Ralf S. Engelschall]
9886 *) Added various platform portability fixes.
9889 *) The Genesis of the OpenSSL rpject:
9890 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9891 Young and Tim J. Hudson created while they were working for C2Net until
9893 [The OpenSSL Project]
9896 Changes between 0.9.0b and 0.9.1b [not released]
9898 *) Updated a few CA certificates under certs/
9901 *) Changed some BIGNUM api stuff.
9904 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9905 DGUX x86, Linux Alpha, etc.
9908 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9909 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9913 *) Add -strparse option to asn1pars program which parses nested
9915 [Dr Stephen Henson <shenson@bigfoot.com>]
9917 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9920 *) DSA fix for "ca" program.
9923 *) Added "-genkey" option to "dsaparam" program.
9926 *) Added RIPE MD160 (rmd160) message digest.
9929 *) Added -a (all) option to "ssleay version" command.
9932 *) Added PLATFORM define which is the id given to Configure.
9935 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9938 *) Extended the ASN.1 parser routines.
9941 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9944 *) Added a BN_CTX to the BN library.
9947 *) Fixed the weak key values in DES library
9950 *) Changed API in EVP library for cipher aliases.
9953 *) Added support for RC2/64bit cipher.
9956 *) Converted the lhash library to the crypto/mem.c functions.
9959 *) Added more recognized ASN.1 object ids.
9962 *) Added more RSA padding checks for SSL/TLS.
9965 *) Added BIO proxy/filter functionality.
9968 *) Added extra_certs to SSL_CTX which can be used
9969 send extra CA certificates to the client in the CA cert chain sending
9970 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9973 *) Now Fortezza is denied in the authentication phase because
9974 this is key exchange mechanism is not supported by SSLeay at all.
9977 *) Additional PKCS1 checks.
9980 *) Support the string "TLSv1" for all TLS v1 ciphers.
9983 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9984 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9987 *) Fixed a few memory leaks.
9990 *) Fixed various code and comment typos.
9993 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9994 bytes sent in the client random.
9995 [Edward Bishop <ebishop@spyglass.com>]