5 Changes between 1.0.1 and 1.0.2 [xx XXX xxxx]
7 *) Add new functions to allow customised supported signature algorithms
8 for SSL and SSL_CTX structures. Add options to s_client and s_server
12 *) New function SSL_certs_clear() to delete all references to certificates
13 from an SSL structure. Before this once a certificate had been added
14 it couldn't be removed.
17 *) Integrate hostname, email address and IP address checking with certificate
18 verification. New verify options supporting checking in opensl utility.
21 *) Fixes and wildcard matching support to hostname and email checking
22 functions. Add manual page.
23 [Florian Weimer (Red Hat Product Security Team)]
25 *) New functions to check a hostname email or IP address against a
26 certificate. Add options x509 utility to print results of checks against
31 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
33 *) Backport support for partial chain verification: if an intermediate
34 certificate is explicitly trusted (using -addtrust option to x509
35 utility for example) the verification is sucessful even if the chain
37 The OCSP checking fix depends on this backport.
38 [Steve Henson and Rob Stradling <rob.stradling@comodo.com>]
40 *) Add -trusted_first option which attempts to find certificates in the
41 trusted store even if an untrusted chain is also supplied.
44 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
45 platform support for Linux and Android.
48 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
49 the right response is stapled. Also change current certificate to
50 the certificate actually sent.
51 See http://rt.openssl.org/Ticket/Display.html?id=2836.
52 [Rob Stradling <rob.stradling@comodo.com>]
54 *) Support for linux-x32, ILP32 environment in x86_64 framework.
58 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
60 *) Experimental multi-implementation support for FIPS capable OpenSSL.
61 When in FIPS mode the approved implementations are used as normal,
62 when not in FIPS mode the internal unapproved versions are used instead.
63 This means that the FIPS capable OpenSSL isn't forced to use the
64 (often lower perfomance) FIPS implementations outside FIPS mode.
67 *) Transparently support X9.42 DH parameters when calling
68 PEM_read_bio_DHparameters. This means existing applications can handle
69 the new parameter format automatically.
72 *) Initial experimental support for X9.42 DH parameter format: mainly
73 to support use of 'q' parameter for RFC5114 parameters.
76 *) Add DH parameters from RFC5114 including test data to dhtest.
79 *) Support for automatic EC temporary key parameter selection. If enabled
80 the most preferred EC parameters are automatically used instead of
81 hardcoded fixed parameters. Now a server just has to call:
82 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
83 support ECDH and use the most appropriate parameters.
86 *) Enhance and tidy EC curve and point format TLS extension code. Use
87 static structures instead of allocation if default values are used.
88 New ctrls to set curves we wish to support and to retrieve shared curves.
89 Print out shared curves in s_server. New options to s_server and s_client
90 to set list of supported curves.
93 *) New ctrls to retrieve supported signature algorithms and
94 supported curve values as an array of NIDs. Extend openssl utility
95 to print out received values.
98 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
99 between NIDs and the more common NIST names such as "P-256". Enhance
100 ecparam utility and ECC method to recognise the NIST names for curves.
103 *) Enhance SSL/TLS certificate chain handling to support different
104 chains for each certificate instead of one chain in the parent SSL_CTX.
107 *) Support for fixed DH ciphersuite client authentication: where both
108 server and client use DH certificates with common parameters.
111 *) Support for fixed DH ciphersuites: those requiring DH server
115 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
117 *) Make openssl verify return errors.
118 [Chris Palmer <palmer@google.com> and Ben Laurie]
120 *) Fix possible deadlock when decoding public keys.
123 *) Don't use TLS 1.0 record version number in initial client hello
127 Changes between 1.0.1b and 1.0.1c [10 May 2012]
129 *) Sanity check record length before skipping explicit IV in TLS
130 1.2, 1.1 and DTLS to avoid DoS attack.
132 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
133 fuzzing as a service testing platform.
137 *) Initialise tkeylen properly when encrypting CMS messages.
138 Thanks to Solar Designer of Openwall for reporting this issue.
141 *) In FIPS mode don't try to use composite ciphers as they are not
145 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
147 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
148 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
149 mean any application compiled against OpenSSL 1.0.0 headers setting
150 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
151 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
152 0x10000000L Any application which was previously compiled against
153 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
154 will need to be recompiled as a result. Letting be results in
155 inability to disable specifically TLS 1.1 and in client context,
156 in unlike event, limit maximum offered version to TLS 1.0 [see below].
159 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
160 disable just protocol X, but all protocols above X *if* there are
161 protocols *below* X still enabled. In more practical terms it means
162 that if application wants to disable TLS1.0 in favor of TLS1.1 and
163 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
164 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
168 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
170 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
171 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
172 in CRYPTO_realloc_clean.
174 Thanks to Tavis Ormandy, Google Security Team, for discovering this
175 issue and to Adam Langley <agl@chromium.org> for fixing it.
177 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
179 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
182 *) Workarounds for some broken servers that "hang" if a client hello
183 record length exceeds 255 bytes:
185 1. Do not use record version number > TLS 1.0 in initial client
186 hello: some (but not all) hanging servers will now work.
187 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
188 the number of ciphers sent in the client hello. This should be
189 set to an even number, such as 50, for example by passing:
190 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
191 Most broken servers should now work.
192 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
193 TLS 1.2 client support entirely.
196 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
199 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
201 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
202 STRING form instead of a DigestInfo.
205 *) The format used for MDC2 RSA signatures is inconsistent between EVP
206 and the RSA_sign/RSA_verify functions. This was made more apparent when
207 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
208 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
209 the correct format in RSA_verify so both forms transparently work.
212 *) Some servers which support TLS 1.0 can choke if we initially indicate
213 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
214 encrypted premaster secret. As a workaround use the maximum pemitted
215 client version in client hello, this should keep such servers happy
216 and still work with previous versions of OpenSSL.
219 *) Add support for TLS/DTLS heartbeats.
220 [Robin Seggelmann <seggelmann@fh-muenster.de>]
222 *) Add support for SCTP.
223 [Robin Seggelmann <seggelmann@fh-muenster.de>]
225 *) Improved PRNG seeding for VOS.
226 [Paul Green <Paul.Green@stratus.com>]
228 *) Extensive assembler packs updates, most notably:
230 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
231 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
232 - x86_64: bit-sliced AES implementation;
233 - ARM: NEON support, contemporary platforms optimizations;
234 - s390x: z196 support;
235 - *: GHASH and GF(2^m) multiplication implementations;
239 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
240 (removal of unnecessary code)
241 [Peter Sylvester <peter.sylvester@edelweb.fr>]
243 *) Add TLS key material exporter from RFC 5705.
246 *) Add DTLS-SRTP negotiation from RFC 5764.
249 *) Add Next Protocol Negotiation,
250 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
251 disabled with a no-npn flag to config or Configure. Code donated
253 [Adam Langley <agl@google.com> and Ben Laurie]
255 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
256 NIST-P256, NIST-P521, with constant-time single point multiplication on
257 typical inputs. Compiler support for the nonstandard type __uint128_t is
258 required to use this (present in gcc 4.4 and later, for 64-bit builds).
259 Code made available under Apache License version 2.0.
261 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
262 line to include this in your build of OpenSSL, and run "make depend" (or
263 "make update"). This enables the following EC_METHODs:
265 EC_GFp_nistp224_method()
266 EC_GFp_nistp256_method()
267 EC_GFp_nistp521_method()
269 EC_GROUP_new_by_curve_name() will automatically use these (while
270 EC_GROUP_new_curve_GFp() currently prefers the more flexible
272 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
274 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
275 all platforms. Move ssize_t definition from e_os.h to the public
276 header file e_os2.h as it now appears in public header file cms.h
279 *) New -sigopt option to the ca, req and x509 utilities. Additional
280 signature parameters can be passed using this option and in
284 *) Add RSA PSS signing function. This will generate and set the
285 appropriate AlgorithmIdentifiers for PSS based on those in the
286 corresponding EVP_MD_CTX structure. No application support yet.
289 *) Support for companion algorithm specific ASN1 signing routines.
290 New function ASN1_item_sign_ctx() signs a pre-initialised
291 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
292 the appropriate parameters.
295 *) Add new algorithm specific ASN1 verification initialisation function
296 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
297 handling will be the same no matter what EVP_PKEY_METHOD is used.
298 Add a PSS handler to support verification of PSS signatures: checked
299 against a number of sample certificates.
302 *) Add signature printing for PSS. Add PSS OIDs.
303 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
305 *) Add algorithm specific signature printing. An individual ASN1 method
306 can now print out signatures instead of the standard hex dump.
308 More complex signatures (e.g. PSS) can print out more meaningful
309 information. Include DSA version that prints out the signature
313 *) Password based recipient info support for CMS library: implementing
317 *) Split password based encryption into PBES2 and PBKDF2 functions. This
318 neatly separates the code into cipher and PBE sections and is required
319 for some algorithms that split PBES2 into separate pieces (such as
323 *) Session-handling fixes:
324 - Fix handling of connections that are resuming with a session ID,
325 but also support Session Tickets.
326 - Fix a bug that suppressed issuing of a new ticket if the client
327 presented a ticket with an expired session.
328 - Try to set the ticket lifetime hint to something reasonable.
329 - Make tickets shorter by excluding irrelevant information.
330 - On the client side, don't ignore renewed tickets.
331 [Adam Langley, Bodo Moeller (Google)]
333 *) Fix PSK session representation.
336 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
338 This work was sponsored by Intel.
341 *) Add GCM support to TLS library. Some custom code is needed to split
342 the IV between the fixed (from PRF) and explicit (from TLS record)
343 portions. This adds all GCM ciphersuites supported by RFC5288 and
344 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
345 add a special AESGCM string for GCM only.
348 *) Expand range of ctrls for AES GCM. Permit setting invocation
349 field on decrypt and retrieval of invocation field only on encrypt.
352 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
353 As required by RFC5289 these ciphersuites cannot be used if for
354 versions of TLS earlier than 1.2.
357 *) For FIPS capable OpenSSL interpret a NULL default public key method
358 as unset and return the appopriate default but do *not* set the default.
359 This means we can return the appopriate method in applications that
360 swicth between FIPS and non-FIPS modes.
363 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
364 ENGINE is used then we cannot handle that in the FIPS module so we
365 keep original code iff non-FIPS operations are allowed.
368 *) Add -attime option to openssl utilities.
369 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
371 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
374 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
375 FIPS EC methods unconditionally for now.
378 *) New build option no-ec2m to disable characteristic 2 code.
381 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
382 all cases can be covered as some introduce binary incompatibilities.
385 *) Redirect RSA operations to FIPS module including keygen,
386 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
389 *) Add similar low level API blocking to ciphers.
392 *) Low level digest APIs are not approved in FIPS mode: any attempt
393 to use these will cause a fatal error. Applications that *really* want
394 to use them can use the private_* version instead.
397 *) Redirect cipher operations to FIPS module for FIPS builds.
400 *) Redirect digest operations to FIPS module for FIPS builds.
403 *) Update build system to add "fips" flag which will link in fipscanister.o
404 for static and shared library builds embedding a signature if needed.
407 *) Output TLS supported curves in preference order instead of numerical
408 order. This is currently hardcoded for the highest order curves first.
409 This should be configurable so applications can judge speed vs strength.
412 *) Add TLS v1.2 server support for client authentication.
415 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
419 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
420 FIPS modules versions.
423 *) Add TLS v1.2 client side support for client authentication. Keep cache
424 of handshake records longer as we don't know the hash algorithm to use
425 until after the certificate request message is received.
428 *) Initial TLS v1.2 client support. Add a default signature algorithms
429 extension including all the algorithms we support. Parse new signature
430 format in client key exchange. Relax some ECC signing restrictions for
431 TLS v1.2 as indicated in RFC5246.
434 *) Add server support for TLS v1.2 signature algorithms extension. Switch
435 to new signature format when needed using client digest preference.
436 All server ciphersuites should now work correctly in TLS v1.2. No client
437 support yet and no support for client certificates.
440 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
441 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
442 ciphersuites. At present only RSA key exchange ciphersuites work with
443 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
444 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
445 and version checking.
448 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
449 with this defined it will not be affected by any changes to ssl internal
450 structures. Add several utility functions to allow openssl application
451 to work with OPENSSL_NO_SSL_INTERN defined.
455 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
457 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
460 *) Permit abbreviated handshakes when renegotiating using the function
461 SSL_renegotiate_abbreviated().
462 [Robin Seggelmann <seggelmann@fh-muenster.de>]
464 *) Add call to ENGINE_register_all_complete() to
465 ENGINE_load_builtin_engines(), so some implementations get used
466 automatically instead of needing explicit application support.
469 *) Add support for TLS key exporter as described in RFC5705.
470 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
472 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
473 a few changes are required:
475 Add SSL_OP_NO_TLSv1_1 flag.
477 Update version checking logic to handle version 1.1.
478 Add explicit IV handling (ported from DTLS code).
479 Add command line options to s_client/s_server.
482 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
484 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
485 in CMS and PKCS7 code. When RSA decryption fails use a random key for
486 content decryption and always return the same error. Note: this attack
487 needs on average 2^20 messages so it only affects automated senders. The
488 old behaviour can be reenabled in the CMS code by setting the
489 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
490 an MMA defence is not necessary.
491 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
492 this issue. (CVE-2012-0884)
495 *) Fix CVE-2011-4619: make sure we really are receiving a
496 client hello before rejecting multiple SGC restarts. Thanks to
497 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
500 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
502 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
503 Thanks to Antonio Martin, Enterprise Secure Access Research and
504 Development, Cisco Systems, Inc. for discovering this bug and
505 preparing a fix. (CVE-2012-0050)
508 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
510 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
511 of the Vaudenay padding oracle attack on CBC mode encryption
512 which enables an efficient plaintext recovery attack against
513 the OpenSSL implementation of DTLS. Their attack exploits timing
514 differences arising during decryption processing. A research
515 paper describing this attack can be found at:
516 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
517 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
518 Security Group at Royal Holloway, University of London
519 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
520 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
521 for preparing the fix. (CVE-2011-4108)
522 [Robin Seggelmann, Michael Tuexen]
524 *) Clear bytes used for block padding of SSL 3.0 records.
526 [Adam Langley (Google)]
528 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
529 Kadianakis <desnacked@gmail.com> for discovering this issue and
530 Adam Langley for preparing the fix. (CVE-2011-4619)
531 [Adam Langley (Google)]
533 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
534 [Andrey Kulikov <amdeich@gmail.com>]
536 *) Prevent malformed RFC3779 data triggering an assertion failure.
537 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
538 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
539 [Rob Austein <sra@hactrn.net>]
541 *) Improved PRNG seeding for VOS.
542 [Paul Green <Paul.Green@stratus.com>]
544 *) Fix ssl_ciph.c set-up race.
545 [Adam Langley (Google)]
547 *) Fix spurious failures in ecdsatest.c.
548 [Emilia Käsper (Google)]
550 *) Fix the BIO_f_buffer() implementation (which was mixing different
551 interpretations of the '..._len' fields).
552 [Adam Langley (Google)]
554 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
555 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
556 threads won't reuse the same blinding coefficients.
558 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
559 lock to call BN_BLINDING_invert_ex, and avoids one use of
560 BN_BLINDING_update for each BN_BLINDING structure (previously,
561 the last update always remained unused).
562 [Emilia Käsper (Google)]
564 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
565 [Bob Buckholz (Google)]
567 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
569 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
570 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
571 [Kaspar Brand <ossl@velox.ch>]
573 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
574 for multi-threaded use of ECDH. (CVE-2011-3210)
575 [Adam Langley (Google)]
577 *) Fix x509_name_ex_d2i memory leak on bad inputs.
580 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
581 signature public key algorithm by using OID xref utilities instead.
582 Before this you could only use some ECC ciphersuites with SHA1 only.
585 *) Add protection against ECDSA timing attacks as mentioned in the paper
586 by Billy Bob Brumley and Nicola Tuveri, see:
588 http://eprint.iacr.org/2011/232.pdf
590 [Billy Bob Brumley and Nicola Tuveri]
592 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
594 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
595 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
597 *) Fix bug in string printing code: if *any* escaping is enabled we must
598 escape the escape character (backslash) or the resulting string is
602 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
604 *) Disable code workaround for ancient and obsolete Netscape browsers
605 and servers: an attacker can use it in a ciphersuite downgrade attack.
606 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
609 *) Fixed J-PAKE implementation error, originally discovered by
610 Sebastien Martini, further info and confirmation from Stefan
611 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
614 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
616 *) Fix extension code to avoid race conditions which can result in a buffer
617 overrun vulnerability: resumed sessions must not be modified as they can
618 be shared by multiple threads. CVE-2010-3864
621 *) Fix WIN32 build system to correctly link an ENGINE directory into
625 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
627 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
629 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
631 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
633 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
634 context. The operation can be customised via the ctrl mechanism in
635 case ENGINEs want to include additional functionality.
638 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
641 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
642 output hashes compatible with older versions of OpenSSL.
643 [Willy Weisz <weisz@vcpc.univie.ac.at>]
645 *) Fix compression algorithm handling: if resuming a session use the
646 compression algorithm of the resumed session instead of determining
647 it from client hello again. Don't allow server to change algorithm.
650 *) Add load_crls() function to apps tidying load_certs() too. Add option
651 to verify utility to allow additional CRLs to be included.
654 *) Update OCSP request code to permit adding custom headers to the request:
655 some responders need this.
658 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
660 [Julia Lawall <julia@diku.dk>]
662 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
663 needlessly dereferenced structures, used obsolete functions and
664 didn't handle all updated verify codes correctly.
667 *) Disable MD2 in the default configuration.
670 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
671 indicate the initial BIO being pushed or popped. This makes it possible
672 to determine whether the BIO is the one explicitly called or as a result
673 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
674 it handles reference counts correctly and doesn't zero out the I/O bio
675 when it is not being explicitly popped. WARNING: applications which
676 included workarounds for the old buggy behaviour will need to be modified
677 or they could free up already freed BIOs.
680 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
681 renaming to all platforms (within the 0.9.8 branch, this was
682 done conditionally on Netware platforms to avoid a name clash).
683 [Guenter <lists@gknw.net>]
685 *) Add ECDHE and PSK support to DTLS.
686 [Michael Tuexen <tuexen@fh-muenster.de>]
688 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
692 *) Add "missing" function EVP_MD_flags() (without this the only way to
693 retrieve a digest flags is by accessing the structure directly. Update
694 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
695 or cipher is registered as in the "from" argument. Print out all
696 registered digests in the dgst usage message instead of manually
697 attempting to work them out.
700 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
701 this allows the use of compression and extensions. Change default cipher
702 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
703 by default unless an application cipher string requests it.
706 *) Alter match criteria in PKCS12_parse(). It used to try to use local
707 key ids to find matching certificates and keys but some PKCS#12 files
708 don't follow the (somewhat unwritten) rules and this strategy fails.
709 Now just gather all certificates together and the first private key
710 then look for the first certificate that matches the key.
713 *) Support use of registered digest and cipher names for dgst and cipher
714 commands instead of having to add each one as a special case. So now
721 openssl dgst -sha256 foo
723 and this works for ENGINE based algorithms too.
727 *) Update Gost ENGINE to support parameter files.
728 [Victor B. Wagner <vitus@cryptocom.ru>]
730 *) Support GeneralizedTime in ca utility.
731 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
733 *) Enhance the hash format used for certificate directory links. The new
734 form uses the canonical encoding (meaning equivalent names will work
735 even if they aren't identical) and uses SHA1 instead of MD5. This form
736 is incompatible with the older format and as a result c_rehash should
737 be used to rebuild symbolic links.
740 *) Make PKCS#8 the default write format for private keys, replacing the
741 traditional format. This form is standardised, more secure and doesn't
742 include an implicit MD5 dependency.
745 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
746 committed to OpenSSL should pass this lot as a minimum.
749 *) Add session ticket override functionality for use by EAP-FAST.
750 [Jouni Malinen <j@w1.fi>]
752 *) Modify HMAC functions to return a value. Since these can be implemented
753 in an ENGINE errors can occur.
756 *) Type-checked OBJ_bsearch_ex.
759 *) Type-checked OBJ_bsearch. Also some constification necessitated
760 by type-checking. Still to come: TXT_DB, bsearch(?),
761 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
765 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
766 seconds to a tm structure directly, instead of going through OS
767 specific date routines. This avoids any issues with OS routines such
768 as the year 2038 bug. New *_adj() functions for ASN1 time structures
769 and X509_time_adj_ex() to cover the extended range. The existing
770 X509_time_adj() is still usable and will no longer have any date issues.
773 *) Delta CRL support. New use deltas option which will attempt to locate
774 and search any appropriate delta CRLs available.
776 This work was sponsored by Google.
779 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
780 code and add additional score elements. Validate alternate CRL paths
781 as part of the CRL checking and indicate a new error "CRL path validation
782 error" in this case. Applications wanting additional details can use
783 the verify callback and check the new "parent" field. If this is not
784 NULL CRL path validation is taking place. Existing applications wont
785 see this because it requires extended CRL support which is off by
788 This work was sponsored by Google.
791 *) Support for freshest CRL extension.
793 This work was sponsored by Google.
796 *) Initial indirect CRL support. Currently only supported in the CRLs
797 passed directly and not via lookup. Process certificate issuer
798 CRL entry extension and lookup CRL entries by bother issuer name
799 and serial number. Check and process CRL issuer entry in IDP extension.
801 This work was sponsored by Google.
804 *) Add support for distinct certificate and CRL paths. The CRL issuer
805 certificate is validated separately in this case. Only enabled if
806 an extended CRL support flag is set: this flag will enable additional
807 CRL functionality in future.
809 This work was sponsored by Google.
812 *) Add support for policy mappings extension.
814 This work was sponsored by Google.
817 *) Fixes to pathlength constraint, self issued certificate handling,
818 policy processing to align with RFC3280 and PKITS tests.
820 This work was sponsored by Google.
823 *) Support for name constraints certificate extension. DN, email, DNS
824 and URI types are currently supported.
826 This work was sponsored by Google.
829 *) To cater for systems that provide a pointer-based thread ID rather
830 than numeric, deprecate the current numeric thread ID mechanism and
831 replace it with a structure and associated callback type. This
832 mechanism allows a numeric "hash" to be extracted from a thread ID in
833 either case, and on platforms where pointers are larger than 'long',
834 mixing is done to help ensure the numeric 'hash' is usable even if it
835 can't be guaranteed unique. The default mechanism is to use "&errno"
836 as a pointer-based thread ID to distinguish between threads.
838 Applications that want to provide their own thread IDs should now use
839 CRYPTO_THREADID_set_callback() to register a callback that will call
840 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
842 Note that ERR_remove_state() is now deprecated, because it is tied
843 to the assumption that thread IDs are numeric. ERR_remove_state(0)
844 to free the current thread's error state should be replaced by
845 ERR_remove_thread_state(NULL).
847 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
848 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
849 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
850 application was previously providing a numeric thread callback that
851 was inappropriate for distinguishing threads, then uniqueness might
852 have been obtained with &errno that happened immediately in the
853 intermediate development versions of OpenSSL; this is no longer the
854 case, the numeric thread callback will now override the automatic use
856 [Geoff Thorpe, with help from Bodo Moeller]
858 *) Initial support for different CRL issuing certificates. This covers a
859 simple case where the self issued certificates in the chain exist and
860 the real CRL issuer is higher in the existing chain.
862 This work was sponsored by Google.
865 *) Removed effectively defunct crypto/store from the build.
868 *) Revamp of STACK to provide stronger type-checking. Still to come:
869 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
870 ASN1_STRING, CONF_VALUE.
873 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
874 RAM on SSL connections. This option can save about 34k per idle SSL.
877 *) Revamp of LHASH to provide stronger type-checking. Still to come:
878 STACK, TXT_DB, bsearch, qsort.
881 *) Initial support for Cryptographic Message Syntax (aka CMS) based
882 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
883 support for data, signedData, compressedData, digestedData and
884 encryptedData, envelopedData types included. Scripts to check against
885 RFC4134 examples draft and interop and consistency checks of many
886 content types and variants.
889 *) Add options to enc utility to support use of zlib compression BIO.
892 *) Extend mk1mf to support importing of options and assembly language
893 files from Configure script, currently only included in VC-WIN32.
894 The assembly language rules can now optionally generate the source
895 files from the associated perl scripts.
898 *) Implement remaining functionality needed to support GOST ciphersuites.
899 Interop testing has been performed using CryptoPro implementations.
900 [Victor B. Wagner <vitus@cryptocom.ru>]
902 *) s390x assembler pack.
905 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
909 *) Implement Opaque PRF Input TLS extension as specified in
910 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
911 official specification yet and no extension type assignment by
912 IANA exists, this extension (for now) will have to be explicitly
913 enabled when building OpenSSL by providing the extension number
914 to use. For example, specify an option
916 -DTLSEXT_TYPE_opaque_prf_input=0x9527
918 to the "config" or "Configure" script to enable the extension,
919 assuming extension number 0x9527 (which is a completely arbitrary
920 and unofficial assignment based on the MD5 hash of the Internet
921 Draft). Note that by doing so, you potentially lose
922 interoperability with other TLS implementations since these might
923 be using the same extension number for other purposes.
925 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
926 opaque PRF input value to use in the handshake. This will create
927 an interal copy of the length-'len' string at 'src', and will
928 return non-zero for success.
930 To get more control and flexibility, provide a callback function
933 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
934 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
938 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
941 Callback function 'cb' will be called in handshakes, and is
942 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
943 Argument 'arg' is for application purposes (the value as given to
944 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
945 be provided to the callback function). The callback function
946 has to return non-zero to report success: usually 1 to use opaque
947 PRF input just if possible, or 2 to enforce use of the opaque PRF
948 input. In the latter case, the library will abort the handshake
949 if opaque PRF input is not successfully negotiated.
951 Arguments 'peerinput' and 'len' given to the callback function
952 will always be NULL and 0 in the case of a client. A server will
953 see the client's opaque PRF input through these variables if
954 available (NULL and 0 otherwise). Note that if the server
955 provides an opaque PRF input, the length must be the same as the
956 length of the client's opaque PRF input.
958 Note that the callback function will only be called when creating
959 a new session (session resumption can resume whatever was
960 previously negotiated), and will not be called in SSL 2.0
961 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
962 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
963 for applications that need to enforce opaque PRF input.
967 *) Update ssl code to support digests other than SHA1+MD5 for handshake
970 [Victor B. Wagner <vitus@cryptocom.ru>]
972 *) Add RFC4507 support to OpenSSL. This includes the corrections in
973 RFC4507bis. The encrypted ticket format is an encrypted encoded
974 SSL_SESSION structure, that way new session features are automatically
977 If a client application caches session in an SSL_SESSION structure
978 support is transparent because tickets are now stored in the encoded
981 The SSL_CTX structure automatically generates keys for ticket
982 protection in servers so again support should be possible
983 with no application modification.
985 If a client or server wishes to disable RFC4507 support then the option
986 SSL_OP_NO_TICKET can be set.
988 Add a TLS extension debugging callback to allow the contents of any client
989 or server extensions to be examined.
991 This work was sponsored by Google.
994 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
995 OpenSSL should now compile cleanly on gcc 4.2
996 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
998 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
999 support including streaming MAC support: this is required for GOST
1000 ciphersuite support.
1001 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1003 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1004 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1005 to output in BER and PEM format.
1008 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1009 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1010 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1011 ENGINE support for HMAC keys which are unextractable. New -mac and
1012 -macopt options to dgst utility.
1015 *) New option -sigopt to dgst utility. Update dgst to use
1016 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1017 alternative signing paramaters such as X9.31 or PSS in the dgst
1021 *) Change ssl_cipher_apply_rule(), the internal function that does
1022 the work each time a ciphersuite string requests enabling
1023 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1024 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1025 the order of disabled ciphersuites such that those ciphersuites
1026 that most recently went from enabled to disabled not only stay
1027 in order with respect to each other, but also have higher priority
1028 than other disabled ciphersuites the next time ciphersuites are
1031 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1032 the same ciphersuites as with "HIGH" alone, but in a specific
1033 order where the PSK ciphersuites come first (since they are the
1034 most recently disabled ciphersuites when "HIGH" is parsed).
1036 Also, change ssl_create_cipher_list() (using this new
1037 funcionality) such that between otherwise identical
1038 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1042 *) Change ssl_create_cipher_list() so that it automatically
1043 arranges the ciphersuites in reasonable order before starting
1044 to process the rule string. Thus, the definition for "DEFAULT"
1045 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1046 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1047 This makes it much easier to arrive at a reasonable default order
1048 in applications for which anonymous ciphers are OK (meaning
1049 that you can't actually use DEFAULT).
1050 [Bodo Moeller; suggested by Victor Duchovni]
1052 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1053 processing) into multiple integers instead of setting
1054 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1055 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1056 (These masks as well as the individual bit definitions are hidden
1057 away into the non-exported interface ssl/ssl_locl.h, so this
1058 change to the definition of the SSL_CIPHER structure shouldn't
1059 affect applications.) This give us more bits for each of these
1060 categories, so there is no longer a need to coagulate AES128 and
1061 AES256 into a single algorithm bit, and to coagulate Camellia128
1062 and Camellia256 into a single algorithm bit, which has led to all
1065 Thus, among other things, the kludge introduced in 0.9.7m and
1066 0.9.8e for masking out AES256 independently of AES128 or masking
1067 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1069 With the change, we also introduce new ciphersuite aliases that
1070 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1074 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1075 Use the leftmost N bytes of the signature input if the input is
1076 larger than the prime q (with N being the size in bytes of q).
1079 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1080 it yet and it is largely untested.
1083 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1086 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1087 some compilers (gcc 4.2 and later) reject their use. Safestack is
1088 reimplemented. Update ASN1 to avoid use of legacy functions.
1091 *) Win32/64 targets are linked with Winsock2.
1094 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1095 to external functions. This can be used to increase CRL handling
1096 efficiency especially when CRLs are very large by (for example) storing
1097 the CRL revoked certificates in a database.
1100 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1101 new CRLs added to a directory can be used. New command line option
1102 -verify_return_error to s_client and s_server. This causes real errors
1103 to be returned by the verify callback instead of carrying on no matter
1104 what. This reflects the way a "real world" verify callback would behave.
1107 *) GOST engine, supporting several GOST algorithms and public key formats.
1108 Kindly donated by Cryptocom.
1111 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1112 partitioned by DP are handled but no indirect CRL or reason partitioning
1113 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1114 selected via a scoring technique which handles IDP and AKID in CRLs.
1117 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1118 will ultimately be used for all verify operations: this will remove the
1119 X509_STORE dependency on certificate verification and allow alternative
1120 lookup methods. X509_STORE based implementations of these two callbacks.
1123 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1124 Modify get_crl() to find a valid (unexpired) CRL if possible.
1127 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1128 this would be called X509_CRL_cmp() but that name is already used by
1129 a function that just compares CRL issuer names. Cache several CRL
1130 extensions in X509_CRL structure and cache CRLDP in X509.
1133 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1134 this maps equivalent X509_NAME structures into a consistent structure.
1135 Name comparison can then be performed rapidly using memcmp().
1138 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1142 *) Allow digests to supply their own micalg string for S/MIME type using
1143 the ctrl EVP_MD_CTRL_MICALG.
1146 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1147 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1148 ctrl. It can then customise the structure before and/or after signing
1152 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1153 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1154 to free up any added signature OIDs.
1157 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1158 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1159 digest and cipher tables. New options added to openssl utility:
1160 list-message-digest-algorithms and list-cipher-algorithms.
1163 *) Change the array representation of binary polynomials: the list
1164 of degrees of non-zero coefficients is now terminated with -1.
1165 Previously it was terminated with 0, which was also part of the
1166 value; thus, the array representation was not applicable to
1167 polynomials where t^0 has coefficient zero. This change makes
1168 the array representation useful in a more general context.
1171 *) Various modifications and fixes to SSL/TLS cipher string
1172 handling. For ECC, the code now distinguishes between fixed ECDH
1173 with RSA certificates on the one hand and with ECDSA certificates
1174 on the other hand, since these are separate ciphersuites. The
1175 unused code for Fortezza ciphersuites has been removed.
1177 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1178 (not "ECDHE"). For consistency with the code for DH
1179 certificates, use of ECDH certificates is now considered ECDH
1180 authentication, not RSA or ECDSA authentication (the latter is
1181 merely the CA's signing algorithm and not actively used in the
1184 The temporary ciphersuite alias "ECCdraft" is no longer
1185 available, and ECC ciphersuites are no longer excluded from "ALL"
1186 and "DEFAULT". The following aliases now exist for RFC 4492
1187 ciphersuites, most of these by analogy with the DH case:
1189 kECDHr - ECDH cert, signed with RSA
1190 kECDHe - ECDH cert, signed with ECDSA
1191 kECDH - ECDH cert (signed with either RSA or ECDSA)
1192 kEECDH - ephemeral ECDH
1193 ECDH - ECDH cert or ephemeral ECDH
1199 AECDH - anonymous ECDH
1200 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1204 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1205 Use correct micalg parameters depending on digest(s) in signed message.
1208 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1209 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1212 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1213 an engine to register a method. Add ENGINE lookups for methods and
1214 functional reference processing.
1217 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1218 EVP_{Sign,Verify}* which allow an application to customise the signature
1222 *) New -resign option to smime utility. This adds one or more signers
1223 to an existing PKCS#7 signedData structure. Also -md option to use an
1224 alternative message digest algorithm for signing.
1227 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1228 create PKCS7 structures containing multiple signers. Update smime
1229 application to support multiple signers.
1232 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1236 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1237 Reorganize PBE internals to lookup from a static table using NIDs,
1238 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1239 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1240 PRF which will be automatically used with PBES2.
1243 *) Replace the algorithm specific calls to generate keys in "req" with the
1247 *) Update PKCS#7 enveloped data routines to use new API. This is now
1248 supported by any public key method supporting the encrypt operation. A
1249 ctrl is added to allow the public key algorithm to examine or modify
1250 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1254 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1255 a default digest type to use. In most cases this will be SHA1 but some
1256 algorithms (such as GOST) need to specify an alternative digest. The
1257 return value indicates how strong the prefernce is 1 means optional and
1258 2 is mandatory (that is it is the only supported type). Modify
1259 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1260 use the default md. Update openssl utilities to use the default digest
1261 type for signing if it is not explicitly indicated.
1264 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1265 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1266 signing method from the key type. This effectively removes the link
1267 between digests and public key types.
1270 *) Add an OID cross reference table and utility functions. Its purpose is to
1271 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1272 rsaEncryption. This will allow some of the algorithm specific hackery
1273 needed to use the correct OID to be removed.
1276 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1277 structures for PKCS7_sign(). They are now set up by the relevant public
1281 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1284 *) Add support for key derivation (agreement) in the API, DH method and
1288 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1289 public and private key formats. As a side effect these add additional
1290 command line functionality not previously available: DSA signatures can be
1291 generated and verified using pkeyutl and DH key support and generation in
1296 [Oliver Tappe <zooey@hirschkaefer.de>]
1298 *) New make target "install_html_docs" installs HTML renditions of the
1300 [Oliver Tappe <zooey@hirschkaefer.de>]
1302 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1303 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1304 support key and parameter generation and add initial key generation
1305 functionality for RSA.
1308 *) Add functions for main EVP_PKEY_method operations. The undocumented
1309 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1310 EVP_PKEY_{encrypt,decrypt}_old.
1313 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1314 key API, doesn't do much yet.
1317 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1318 public key algorithms. New option to openssl utility:
1319 "list-public-key-algorithms" to print out info.
1322 *) Implement the Supported Elliptic Curves Extension for
1323 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1326 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1327 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1330 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1331 utilities such as rsa, dsa, dsaparam etc except they process any key
1335 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1336 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1337 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1341 *) Initial support for pluggable public key ASN1.
1342 De-spaghettify the public key ASN1 handling. Move public and private
1343 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1344 algorithm specific handling to a single module within the relevant
1345 algorithm directory. Add functions to allow (near) opaque processing
1346 of public and private key structures.
1349 *) Implement the Supported Point Formats Extension for
1350 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1353 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1354 for the psk identity [hint] and the psk callback functions to the
1355 SSL_SESSION, SSL and SSL_CTX structure.
1358 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1362 SSL_CTX_use_psk_identity_hint
1363 SSL_get_psk_identity_hint
1364 SSL_get_psk_identity
1365 SSL_use_psk_identity_hint
1367 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1369 *) Add RFC 3161 compliant time stamp request creation, response generation
1370 and response verification functionality.
1371 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1373 *) Add initial support for TLS extensions, specifically for the server_name
1374 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1375 have new members for a host name. The SSL data structure has an
1376 additional member SSL_CTX *initial_ctx so that new sessions can be
1377 stored in that context to allow for session resumption, even after the
1378 SSL has been switched to a new SSL_CTX in reaction to a client's
1379 server_name extension.
1381 New functions (subject to change):
1383 SSL_get_servername()
1384 SSL_get_servername_type()
1387 New CTRL codes and macros (subject to change):
1389 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1390 - SSL_CTX_set_tlsext_servername_callback()
1391 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1392 - SSL_CTX_set_tlsext_servername_arg()
1393 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1395 openssl s_client has a new '-servername ...' option.
1397 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1398 '-key2 ...', '-servername_fatal' (subject to change). This allows
1399 testing the HostName extension for a specific single host name ('-cert'
1400 and '-key' remain fallbacks for handshakes without HostName
1401 negotiation). If the unrecogninzed_name alert has to be sent, this by
1402 default is a warning; it becomes fatal with the '-servername_fatal'
1405 [Peter Sylvester, Remy Allais, Christophe Renou]
1407 *) Whirlpool hash implementation is added.
1410 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1411 bn(64,32). Because of instruction set limitations it doesn't have
1412 any negative impact on performance. This was done mostly in order
1413 to make it possible to share assembler modules, such as bn_mul_mont
1414 implementations, between 32- and 64-bit builds without hassle.
1417 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1418 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1422 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1423 dedicated Montgomery multiplication procedure, is introduced.
1424 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1425 "64-bit" performance on certain 32-bit targets.
1428 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1429 in SSL structures. New SSL ctrl to set maximum send fragment size.
1430 Save memory by seeting the I/O buffer sizes dynamically instead of
1431 using the maximum available value.
1434 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1435 in addition to the text details.
1438 *) Very, very preliminary EXPERIMENTAL support for printing of general
1439 ASN1 structures. This currently produces rather ugly output and doesn't
1440 handle several customised structures at all.
1443 *) Integrated support for PVK file format and some related formats such
1444 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1445 these in the 'rsa' and 'dsa' utilities.
1448 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1451 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1452 place for the (very old) "NETSCAPE" format certificates which are now
1453 handled using new ASN1 code equivalents.
1456 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1457 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1458 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1461 *) Modify CRL distribution points extension code to print out previously
1462 unsupported fields. Enhance extension setting code to allow setting of
1466 *) Add print and set support for Issuing Distribution Point CRL extension.
1469 *) Change 'Configure' script to enable Camellia by default.
1472 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1474 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1475 Thanks to Antonio Martin, Enterprise Secure Access Research and
1476 Development, Cisco Systems, Inc. for discovering this bug and
1477 preparing a fix. (CVE-2012-0050)
1480 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1482 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1483 of the Vaudenay padding oracle attack on CBC mode encryption
1484 which enables an efficient plaintext recovery attack against
1485 the OpenSSL implementation of DTLS. Their attack exploits timing
1486 differences arising during decryption processing. A research
1487 paper describing this attack can be found at:
1488 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1489 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1490 Security Group at Royal Holloway, University of London
1491 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1492 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1493 for preparing the fix. (CVE-2011-4108)
1494 [Robin Seggelmann, Michael Tuexen]
1496 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1497 [Ben Laurie, Kasper <ekasper@google.com>]
1499 *) Clear bytes used for block padding of SSL 3.0 records.
1501 [Adam Langley (Google)]
1503 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1504 Kadianakis <desnacked@gmail.com> for discovering this issue and
1505 Adam Langley for preparing the fix. (CVE-2011-4619)
1506 [Adam Langley (Google)]
1508 *) Prevent malformed RFC3779 data triggering an assertion failure.
1509 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1510 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1511 [Rob Austein <sra@hactrn.net>]
1513 *) Fix ssl_ciph.c set-up race.
1514 [Adam Langley (Google)]
1516 *) Fix spurious failures in ecdsatest.c.
1517 [Emilia Käsper (Google)]
1519 *) Fix the BIO_f_buffer() implementation (which was mixing different
1520 interpretations of the '..._len' fields).
1521 [Adam Langley (Google)]
1523 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1524 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1525 threads won't reuse the same blinding coefficients.
1527 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1528 lock to call BN_BLINDING_invert_ex, and avoids one use of
1529 BN_BLINDING_update for each BN_BLINDING structure (previously,
1530 the last update always remained unused).
1531 [Emilia Käsper (Google)]
1533 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1534 for multi-threaded use of ECDH.
1535 [Adam Langley (Google)]
1537 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1540 *) Add protection against ECDSA timing attacks as mentioned in the paper
1541 by Billy Bob Brumley and Nicola Tuveri, see:
1543 http://eprint.iacr.org/2011/232.pdf
1545 [Billy Bob Brumley and Nicola Tuveri]
1547 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1549 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1550 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1552 *) Fix bug in string printing code: if *any* escaping is enabled we must
1553 escape the escape character (backslash) or the resulting string is
1557 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1559 *) Disable code workaround for ancient and obsolete Netscape browsers
1560 and servers: an attacker can use it in a ciphersuite downgrade attack.
1561 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1564 *) Fixed J-PAKE implementation error, originally discovered by
1565 Sebastien Martini, further info and confirmation from Stefan
1566 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1569 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1571 *) Fix extension code to avoid race conditions which can result in a buffer
1572 overrun vulnerability: resumed sessions must not be modified as they can
1573 be shared by multiple threads. CVE-2010-3864
1576 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1579 *) Don't reencode certificate when calculating signature: cache and use
1580 the original encoding instead. This makes signature verification of
1581 some broken encodings work correctly.
1584 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1585 is also one of the inputs.
1586 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1588 *) Don't repeatedly append PBE algorithms to table if they already exist.
1589 Sort table on each new add. This effectively makes the table read only
1590 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1594 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1596 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1599 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1600 access or freeing data twice (CVE-2010-0742)
1601 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1603 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1604 common in certificates and some applications which only call
1605 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1609 Reduce copying into .apps and .test in makevms.com
1610 Don't try to use blank CA certificate in CA.com
1611 Allow use of C files from original directories in maketests.com
1612 [Steven M. Schweda" <sms@antinode.info>]
1614 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1616 *) When rejecting SSL/TLS records due to an incorrect version number, never
1617 update s->server with a new major version number. As of
1618 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1619 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1620 the previous behavior could result in a read attempt at NULL when
1621 receiving specific incorrect SSL/TLS records once record payload
1622 protection is active. (CVE-2010-0740)
1623 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1625 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1626 could be crashed if the relevant tables were not present (e.g. chrooted).
1627 [Tomas Hoger <thoger@redhat.com>]
1629 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1631 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1632 [Martin Olsson, Neel Mehta]
1634 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1635 accommodate for stack sorting, always a write lock!).
1638 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1639 excessive delays in the RAND_poll(): over a minute. As a workaround
1640 include a time check in the inner Heap32Next loop too.
1643 *) The code that handled flushing of data in SSL/TLS originally used the
1644 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1645 the problem outlined in PR#1949. The fix suggested there however can
1646 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1647 of Apache). So instead simplify the code to flush unconditionally.
1648 This should be fine since flushing with no data to flush is a no op.
1651 *) Handle TLS versions 2.0 and later properly and correctly use the
1652 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1653 off ancient servers have a habit of sticking around for a while...
1656 *) Modify compression code so it frees up structures without using the
1657 ex_data callbacks. This works around a problem where some applications
1658 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1659 restarting) then use compression (e.g. SSL with compression) later.
1660 This results in significant per-connection memory leaks and
1661 has caused some security issues including CVE-2008-1678 and
1665 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1666 change when encrypting or decrypting.
1669 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1670 connect and renegotiate with servers which do not support RI.
1671 Until RI is more widely deployed this option is enabled by default.
1674 *) Add "missing" ssl ctrls to clear options and mode.
1677 *) If client attempts to renegotiate and doesn't support RI respond with
1678 a no_renegotiation alert as required by RFC5746. Some renegotiating
1679 TLS clients will continue a connection gracefully when they receive
1680 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1681 waiting for a server hello which it will never receive. Now we treat a
1682 received no_renegotiation alert as a fatal error. This is because
1683 applications requesting a renegotiation might well expect it to succeed
1684 and would have no code in place to handle the server denying it so the
1685 only safe thing to do is to terminate the connection.
1688 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1689 peer supports secure renegotiation and 0 otherwise. Print out peer
1690 renegotiation support in s_client/s_server.
1693 *) Replace the highly broken and deprecated SPKAC certification method with
1694 the updated NID creation version. This should correctly handle UTF8.
1697 *) Implement RFC5746. Re-enable renegotiation but require the extension
1698 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1699 turns out to be a bad idea. It has been replaced by
1700 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1701 SSL_CTX_set_options(). This is really not recommended unless you
1702 know what you are doing.
1703 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1705 *) Fixes to stateless session resumption handling. Use initial_ctx when
1706 issuing and attempting to decrypt tickets in case it has changed during
1707 servername handling. Use a non-zero length session ID when attempting
1708 stateless session resumption: this makes it possible to determine if
1709 a resumption has occurred immediately after receiving server hello
1710 (several places in OpenSSL subtly assume this) instead of later in
1714 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1715 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1716 fixes for a few places where the return code is not checked
1718 [Julia Lawall <julia@diku.dk>]
1720 *) Add --strict-warnings option to Configure script to include devteam
1721 warnings in other configurations.
1724 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1725 makes it possible to install openssl libraries in locations which
1726 have names other than "lib", for example "/usr/lib64" which some
1728 [Steve Henson, based on patch from Jeremy Utley]
1730 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1731 X690 8.9.12 and can produce some misleading textual output of OIDs.
1732 [Steve Henson, reported by Dan Kaminsky]
1734 *) Delete MD2 from algorithm tables. This follows the recommendation in
1735 several standards that it is not used in new applications due to
1736 several cryptographic weaknesses. For binary compatibility reasons
1737 the MD2 API is still compiled in by default.
1740 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1744 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1745 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1747 [Guenter <lists@gknw.net>]
1749 *) Fix the server certificate chain building code to use X509_verify_cert(),
1750 it used to have an ad-hoc builder which was unable to cope with anything
1751 other than a simple chain.
1752 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1754 *) Don't check self signed certificate signatures in X509_verify_cert()
1755 by default (a flag can override this): it just wastes time without
1756 adding any security. As a useful side effect self signed root CAs
1757 with non-FIPS digests are now usable in FIPS mode.
1760 *) In dtls1_process_out_of_seq_message() the check if the current message
1761 is already buffered was missing. For every new message was memory
1762 allocated, allowing an attacker to perform an denial of service attack
1763 with sending out of seq handshake messages until there is no memory
1764 left. Additionally every future messege was buffered, even if the
1765 sequence number made no sense and would be part of another handshake.
1766 So only messages with sequence numbers less than 10 in advance will be
1767 buffered. (CVE-2009-1378)
1768 [Robin Seggelmann, discovered by Daniel Mentz]
1770 *) Records are buffered if they arrive with a future epoch to be
1771 processed after finishing the corresponding handshake. There is
1772 currently no limitation to this buffer allowing an attacker to perform
1773 a DOS attack with sending records with future epochs until there is no
1774 memory left. This patch adds the pqueue_size() function to detemine
1775 the size of a buffer and limits the record buffer to 100 entries.
1777 [Robin Seggelmann, discovered by Daniel Mentz]
1779 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1780 parent structure is freed. (CVE-2009-1379)
1783 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1784 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1787 [Ilya O. <vrghost@gmail.com>]
1789 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1791 *) Disable renegotiation completely - this fixes a severe security
1792 problem (CVE-2009-3555) at the cost of breaking all
1793 renegotiation. Renegotiation can be re-enabled by setting
1794 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1795 run-time. This is really not recommended unless you know what
1799 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1801 *) Don't set val to NULL when freeing up structures, it is freed up by
1802 underlying code. If sizeof(void *) > sizeof(long) this can result in
1803 zeroing past the valid field. (CVE-2009-0789)
1804 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1806 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1807 checked correctly. This would allow some invalid signed attributes to
1808 appear to verify correctly. (CVE-2009-0591)
1809 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1811 *) Reject UniversalString and BMPString types with invalid lengths. This
1812 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1813 a legal length. (CVE-2009-0590)
1816 *) Set S/MIME signing as the default purpose rather than setting it
1817 unconditionally. This allows applications to override it at the store
1821 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1822 to handle some structures.
1825 *) Improve efficiency of mem_gets: don't search whole buffer each time
1827 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1829 *) New -hex option for openssl rand.
1832 *) Print out UTF8String and NumericString when parsing ASN1.
1835 *) Support NumericString type for name components.
1838 *) Allow CC in the environment to override the automatically chosen
1839 compiler. Note that nothing is done to ensure flags work with the
1843 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1845 *) Properly check EVP_VerifyFinal() and similar return values
1847 [Ben Laurie, Bodo Moeller, Google Security Team]
1849 *) Enable TLS extensions by default.
1852 *) Allow the CHIL engine to be loaded, whether the application is
1853 multithreaded or not. (This does not release the developer from the
1854 obligation to set up the dynamic locking callbacks.)
1855 [Sander Temme <sander@temme.net>]
1857 *) Use correct exit code if there is an error in dgst command.
1858 [Steve Henson; problem pointed out by Roland Dirlewanger]
1860 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1861 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1864 *) Add experimental JPAKE support, including demo authentication in
1865 s_client and s_server.
1868 *) Set the comparison function in v3_addr_canonize().
1869 [Rob Austein <sra@hactrn.net>]
1871 *) Add support for XMPP STARTTLS in s_client.
1872 [Philip Paeps <philip@freebsd.org>]
1874 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1875 to ensure that even with this option, only ciphersuites in the
1876 server's preference list will be accepted. (Note that the option
1877 applies only when resuming a session, so the earlier behavior was
1878 just about the algorithm choice for symmetric cryptography.)
1881 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1883 *) Fix NULL pointer dereference if a DTLS server received
1884 ChangeCipherSpec as first record (CVE-2009-1386).
1887 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1888 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1891 *) The fix in 0.9.8c that supposedly got rid of unsafe
1892 double-checked locking was incomplete for RSA blinding,
1893 addressing just one layer of what turns out to have been
1894 doubly unsafe triple-checked locking.
1896 So now fix this for real by retiring the MONT_HELPER macro
1897 in crypto/rsa/rsa_eay.c.
1899 [Bodo Moeller; problem pointed out by Marius Schilder]
1901 *) Various precautionary measures:
1903 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1905 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1906 (NB: This would require knowledge of the secret session ticket key
1907 to exploit, in which case you'd be SOL either way.)
1909 - Change bn_nist.c so that it will properly handle input BIGNUMs
1910 outside the expected range.
1912 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1915 [Neel Mehta, Bodo Moeller]
1917 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1918 the load fails. Useful for distros.
1919 [Ben Laurie and the FreeBSD team]
1921 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1924 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1927 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1929 This work was sponsored by Logica.
1932 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1933 keystores. Support for SSL/TLS client authentication too.
1934 Not compiled unless enable-capieng specified to Configure.
1936 This work was sponsored by Logica.
1939 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1940 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1941 attribute creation routines such as certifcate requests and PKCS#12
1945 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1947 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1948 handshake which could lead to a cilent crash as found using the
1949 Codenomicon TLS test suite (CVE-2008-1672)
1950 [Steve Henson, Mark Cox]
1952 *) Fix double free in TLS server name extensions which could lead to
1953 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1956 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1958 Clear the error queue to ensure that error entries left from
1959 older function calls do not interfere with the correct operation.
1960 [Lutz Jaenicke, Erik de Castro Lopo]
1962 *) Remove root CA certificates of commercial CAs:
1964 The OpenSSL project does not recommend any specific CA and does not
1965 have any policy with respect to including or excluding any CA.
1966 Therefore it does not make any sense to ship an arbitrary selection
1967 of root CA certificates with the OpenSSL software.
1970 *) RSA OAEP patches to fix two separate invalid memory reads.
1971 The first one involves inputs when 'lzero' is greater than
1972 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1973 before the beginning of from). The second one involves inputs where
1974 the 'db' section contains nothing but zeroes (there is a one-byte
1975 invalid read after the end of 'db').
1976 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1978 *) Partial backport from 0.9.9-dev:
1980 Introduce bn_mul_mont (dedicated Montgomery multiplication
1981 procedure) as a candidate for BIGNUM assembler implementation.
1982 While 0.9.9-dev uses assembler for various architectures, only
1983 x86_64 is available by default here in the 0.9.8 branch, and
1984 32-bit x86 is available through a compile-time setting.
1986 To try the 32-bit x86 assembler implementation, use Configure
1987 option "enable-montasm" (which exists only for this backport).
1989 As "enable-montasm" for 32-bit x86 disclaims code stability
1990 anyway, in this constellation we activate additional code
1991 backported from 0.9.9-dev for further performance improvements,
1992 namely BN_from_montgomery_word. (To enable this otherwise,
1993 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1995 [Andy Polyakov (backport partially by Bodo Moeller)]
1997 *) Add TLS session ticket callback. This allows an application to set
1998 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1999 values. This is useful for key rollover for example where several key
2000 sets may exist with different names.
2003 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2004 This was broken until now in 0.9.8 releases, such that the only way
2005 a registered ENGINE could be used (assuming it initialises
2006 successfully on the host) was to explicitly set it as the default
2007 for the relevant algorithms. This is in contradiction with 0.9.7
2008 behaviour and the documentation. With this fix, when an ENGINE is
2009 registered into a given algorithm's table of implementations, the
2010 'uptodate' flag is reset so that auto-discovery will be used next
2011 time a new context for that algorithm attempts to select an
2013 [Ian Lister (tweaked by Geoff Thorpe)]
2015 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2016 implemention in the following ways:
2018 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2021 Lack of BER streaming support means one pass streaming processing is
2022 only supported if data is detached: setting the streaming flag is
2023 ignored for embedded content.
2025 CMS support is disabled by default and must be explicitly enabled
2026 with the enable-cms configuration option.
2029 *) Update the GMP engine glue to do direct copies between BIGNUM and
2030 mpz_t when openssl and GMP use the same limb size. Otherwise the
2031 existing "conversion via a text string export" trick is still used.
2032 [Paul Sheer <paulsheer@gmail.com>]
2034 *) Zlib compression BIO. This is a filter BIO which compressed and
2035 uncompresses any data passed through it.
2038 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2039 RFC3394 compatible AES key wrapping.
2042 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2043 sets string data without copying. X509_ALGOR_set0() and
2044 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2045 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2046 from an X509_ATTRIBUTE structure optionally checking it occurs only
2047 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2051 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2052 to get the expected BN_FLG_CONSTTIME behavior.
2053 [Bodo Moeller (Google)]
2057 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2058 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2059 - added some more tests to do_tests.pl
2060 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2061 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2062 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2063 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2064 - various changes to netware.pl to enable gcc-cross builds on Win32
2066 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2067 - various changes to fix missing prototype warnings
2068 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2069 - added AES, WHIRLPOOL and CPUID assembler code to build files
2070 - added missing AES assembler make rules to mk1mf.pl
2071 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2072 [Guenter Knauf <eflash@gmx.net>]
2074 *) Implement certificate status request TLS extension defined in RFC3546.
2075 A client can set the appropriate parameters and receive the encoded
2076 OCSP response via a callback. A server can query the supplied parameters
2077 and set the encoded OCSP response in the callback. Add simplified examples
2078 to s_client and s_server.
2081 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2083 *) Fix various bugs:
2084 + Binary incompatibility of ssl_ctx_st structure
2085 + DTLS interoperation with non-compliant servers
2086 + Don't call get_session_cb() without proposed session
2087 + Fix ia64 assembler code
2088 [Andy Polyakov, Steve Henson]
2090 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2092 *) DTLS Handshake overhaul. There were longstanding issues with
2093 OpenSSL DTLS implementation, which were making it impossible for
2094 RFC 4347 compliant client to communicate with OpenSSL server.
2095 Unfortunately just fixing these incompatibilities would "cut off"
2096 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2097 server keeps tolerating non RFC compliant syntax. The opposite is
2098 not true, 0.9.8f client can not communicate with earlier server.
2099 This update even addresses CVE-2007-4995.
2102 *) Changes to avoid need for function casts in OpenSSL: some compilers
2103 (gcc 4.2 and later) reject their use.
2104 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2107 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2108 RFC4507bis. The encrypted ticket format is an encrypted encoded
2109 SSL_SESSION structure, that way new session features are automatically
2112 If a client application caches session in an SSL_SESSION structure
2113 support is transparent because tickets are now stored in the encoded
2116 The SSL_CTX structure automatically generates keys for ticket
2117 protection in servers so again support should be possible
2118 with no application modification.
2120 If a client or server wishes to disable RFC4507 support then the option
2121 SSL_OP_NO_TICKET can be set.
2123 Add a TLS extension debugging callback to allow the contents of any client
2124 or server extensions to be examined.
2126 This work was sponsored by Google.
2129 *) Add initial support for TLS extensions, specifically for the server_name
2130 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2131 have new members for a host name. The SSL data structure has an
2132 additional member SSL_CTX *initial_ctx so that new sessions can be
2133 stored in that context to allow for session resumption, even after the
2134 SSL has been switched to a new SSL_CTX in reaction to a client's
2135 server_name extension.
2137 New functions (subject to change):
2139 SSL_get_servername()
2140 SSL_get_servername_type()
2143 New CTRL codes and macros (subject to change):
2145 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2146 - SSL_CTX_set_tlsext_servername_callback()
2147 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2148 - SSL_CTX_set_tlsext_servername_arg()
2149 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2151 openssl s_client has a new '-servername ...' option.
2153 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2154 '-key2 ...', '-servername_fatal' (subject to change). This allows
2155 testing the HostName extension for a specific single host name ('-cert'
2156 and '-key' remain fallbacks for handshakes without HostName
2157 negotiation). If the unrecogninzed_name alert has to be sent, this by
2158 default is a warning; it becomes fatal with the '-servername_fatal'
2161 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2163 *) Add AES and SSE2 assembly language support to VC++ build.
2166 *) Mitigate attack on final subtraction in Montgomery reduction.
2169 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2170 (which previously caused an internal error).
2173 *) Squeeze another 10% out of IGE mode when in != out.
2176 *) AES IGE mode speedup.
2177 [Dean Gaudet (Google)]
2179 *) Add the Korean symmetric 128-bit cipher SEED (see
2180 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2181 add SEED ciphersuites from RFC 4162:
2183 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2184 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2185 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2186 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2188 To minimize changes between patchlevels in the OpenSSL 0.9.8
2189 series, SEED remains excluded from compilation unless OpenSSL
2190 is configured with 'enable-seed'.
2191 [KISA, Bodo Moeller]
2193 *) Mitigate branch prediction attacks, which can be practical if a
2194 single processor is shared, allowing a spy process to extract
2195 information. For detailed background information, see
2196 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2197 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2198 and Necessary Software Countermeasures"). The core of the change
2199 are new versions BN_div_no_branch() and
2200 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2201 respectively, which are slower, but avoid the security-relevant
2202 conditional branches. These are automatically called by BN_div()
2203 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2204 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2205 remove a conditional branch.
2207 BN_FLG_CONSTTIME is the new name for the previous
2208 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2209 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2210 in the exponent causes BN_mod_exp_mont() to use the alternative
2211 implementation in BN_mod_exp_mont_consttime().) The old name
2212 remains as a deprecated alias.
2214 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2215 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2216 constant-time implementations for more than just exponentiation.
2217 Here too the old name is kept as a deprecated alias.
2219 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2220 the BN_BLINDING structure gets an independent copy of the
2221 modulus. This means that the previous "BIGNUM *m" argument to
2222 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2223 essentially becomes "const BIGNUM *m", although we can't actually
2224 change this in the header file before 0.9.9. It allows
2225 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2226 enable BN_FLG_CONSTTIME.
2228 [Matthew D Wood (Intel Corp)]
2230 *) In the SSL/TLS server implementation, be strict about session ID
2231 context matching (which matters if an application uses a single
2232 external cache for different purposes). Previously,
2233 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2234 set. This did ensure strict client verification, but meant that,
2235 with applications using a single external cache for quite
2236 different requirements, clients could circumvent ciphersuite
2237 restrictions for a given session ID context by starting a session
2238 in a different context.
2241 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2242 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2243 authentication-only ciphersuites.
2246 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2247 not complete and could lead to a possible single byte overflow
2248 (CVE-2007-5135) [Ben Laurie]
2250 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2252 *) Since AES128 and AES256 (and similarly Camellia128 and
2253 Camellia256) share a single mask bit in the logic of
2254 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2255 kludge to work properly if AES128 is available and AES256 isn't
2256 (or if Camellia128 is available and Camellia256 isn't).
2259 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2260 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2261 When a point or a seed is encoded in a BIT STRING, we need to
2262 prevent the removal of trailing zero bits to get the proper DER
2263 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2264 of a NamedBitList, for which trailing 0 bits need to be removed.)
2267 *) Have SSL/TLS server implementation tolerate "mismatched" record
2268 protocol version while receiving ClientHello even if the
2269 ClientHello is fragmented. (The server can't insist on the
2270 particular protocol version it has chosen before the ServerHello
2271 message has informed the client about his choice.)
2274 *) Add RFC 3779 support.
2275 [Rob Austein for ARIN, Ben Laurie]
2277 *) Load error codes if they are not already present instead of using a
2278 static variable. This allows them to be cleanly unloaded and reloaded.
2279 Improve header file function name parsing.
2282 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2283 or CAPABILITY handshake as required by RFCs.
2286 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2288 *) Introduce limits to prevent malicious keys being able to
2289 cause a denial of service. (CVE-2006-2940)
2290 [Steve Henson, Bodo Moeller]
2292 *) Fix ASN.1 parsing of certain invalid structures that can result
2293 in a denial of service. (CVE-2006-2937) [Steve Henson]
2295 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2296 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2298 *) Fix SSL client code which could crash if connecting to a
2299 malicious SSLv2 server. (CVE-2006-4343)
2300 [Tavis Ormandy and Will Drewry, Google Security Team]
2302 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2303 match only those. Before that, "AES256-SHA" would be interpreted
2304 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2305 the same strength classification in 0.9.7h) as we currently only
2306 have a single AES bit in the ciphersuite description bitmap.
2307 That change, however, also applied to ciphersuite strings such as
2308 "RC4-MD5" that intentionally matched multiple ciphersuites --
2309 namely, SSL 2.0 ciphersuites in addition to the more common ones
2310 from SSL 3.0/TLS 1.0.
2312 So we change the selection algorithm again: Naming an explicit
2313 ciphersuite selects this one ciphersuite, and any other similar
2314 ciphersuite (same bitmap) from *other* protocol versions.
2315 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2316 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2318 Since SSL 2.0 does not have any ciphersuites for which the
2319 128/256 bit distinction would be relevant, this works for now.
2320 The proper fix will be to use different bits for AES128 and
2321 AES256, which would have avoided the problems from the beginning;
2322 however, bits are scarce, so we can only do this in a new release
2323 (not just a patchlevel) when we can change the SSL_CIPHER
2324 definition to split the single 'unsigned long mask' bitmap into
2325 multiple values to extend the available space.
2329 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2331 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2332 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2334 *) Add AES IGE and biIGE modes.
2337 *) Change the Unix randomness entropy gathering to use poll() when
2338 possible instead of select(), since the latter has some
2339 undesirable limitations.
2340 [Darryl Miles via Richard Levitte and Bodo Moeller]
2342 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2343 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2344 cannot be implicitly activated as part of, e.g., the "AES" alias.
2345 However, please upgrade to OpenSSL 0.9.9[-dev] for
2346 non-experimental use of the ECC ciphersuites to get TLS extension
2347 support, which is required for curve and point format negotiation
2348 to avoid potential handshake problems.
2351 *) Disable rogue ciphersuites:
2353 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2354 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2355 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2357 The latter two were purportedly from
2358 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2361 Also deactivate the remaining ciphersuites from
2362 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2363 unofficial, and the ID has long expired.
2366 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2367 dual-core machines) and other potential thread-safety issues.
2370 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2371 versions), which is now available for royalty-free use
2372 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2373 Also, add Camellia TLS ciphersuites from RFC 4132.
2375 To minimize changes between patchlevels in the OpenSSL 0.9.8
2376 series, Camellia remains excluded from compilation unless OpenSSL
2377 is configured with 'enable-camellia'.
2380 *) Disable the padding bug check when compression is in use. The padding
2381 bug check assumes the first packet is of even length, this is not
2382 necessarily true if compresssion is enabled and can result in false
2383 positives causing handshake failure. The actual bug test is ancient
2384 code so it is hoped that implementations will either have fixed it by
2385 now or any which still have the bug do not support compression.
2388 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2390 *) When applying a cipher rule check to see if string match is an explicit
2391 cipher suite and only match that one cipher suite if it is.
2394 *) Link in manifests for VC++ if needed.
2395 [Austin Ziegler <halostatue@gmail.com>]
2397 *) Update support for ECC-based TLS ciphersuites according to
2398 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2399 TLS extensions, which are supported starting with the 0.9.9
2400 branch, not in the OpenSSL 0.9.8 branch).
2403 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2404 opaque EVP_CIPHER_CTX handling.
2407 *) Fixes and enhancements to zlib compression code. We now only use
2408 "zlib1.dll" and use the default __cdecl calling convention on Win32
2409 to conform with the standards mentioned here:
2410 http://www.zlib.net/DLL_FAQ.txt
2411 Static zlib linking now works on Windows and the new --with-zlib-include
2412 --with-zlib-lib options to Configure can be used to supply the location
2413 of the headers and library. Gracefully handle case where zlib library
2417 *) Several fixes and enhancements to the OID generation code. The old code
2418 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2419 handle numbers larger than ULONG_MAX, truncated printing and had a
2420 non standard OBJ_obj2txt() behaviour.
2423 *) Add support for building of engines under engine/ as shared libraries
2424 under VC++ build system.
2427 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2428 Hopefully, we will not see any false combination of paths any more.
2431 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2433 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2434 (part of SSL_OP_ALL). This option used to disable the
2435 countermeasure against man-in-the-middle protocol-version
2436 rollback in the SSL 2.0 server implementation, which is a bad
2437 idea. (CVE-2005-2969)
2439 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2440 for Information Security, National Institute of Advanced Industrial
2441 Science and Technology [AIST], Japan)]
2443 *) Add two function to clear and return the verify parameter flags.
2446 *) Keep cipherlists sorted in the source instead of sorting them at
2447 runtime, thus removing the need for a lock.
2450 *) Avoid some small subgroup attacks in Diffie-Hellman.
2451 [Nick Mathewson and Ben Laurie]
2453 *) Add functions for well-known primes.
2456 *) Extended Windows CE support.
2457 [Satoshi Nakamura and Andy Polyakov]
2459 *) Initialize SSL_METHOD structures at compile time instead of during
2460 runtime, thus removing the need for a lock.
2463 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2464 attempting to decrypt each encrypted key in turn. Add support to
2468 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2470 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2473 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2476 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2477 key into the same file any more.
2480 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2483 *) Add -utf8 command line and config file option to 'ca'.
2484 [Stefan <stf@udoma.org]
2486 *) Removed the macro des_crypt(), as it seems to conflict with some
2487 libraries. Use DES_crypt().
2490 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2491 involves renaming the source and generated shared-libs for
2492 both. The engines will accept the corrected or legacy ids
2493 ('ncipher' and '4758_cca' respectively) when binding. NB,
2494 this only applies when building 'shared'.
2495 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2497 *) Add attribute functions to EVP_PKEY structure. Modify
2498 PKCS12_create() to recognize a CSP name attribute and
2499 use it. Make -CSP option work again in pkcs12 utility.
2502 *) Add new functionality to the bn blinding code:
2503 - automatic re-creation of the BN_BLINDING parameters after
2504 a fixed number of uses (currently 32)
2505 - add new function for parameter creation
2506 - introduce flags to control the update behaviour of the
2507 BN_BLINDING parameters
2508 - hide BN_BLINDING structure
2509 Add a second BN_BLINDING slot to the RSA structure to improve
2510 performance when a single RSA object is shared among several
2514 *) Add support for DTLS.
2515 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2517 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2518 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2521 *) Remove buggy and incompletet DH cert support from
2522 ssl/ssl_rsa.c and ssl/s3_both.c
2525 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2526 the apps/openssl applications.
2529 *) Compile clean with "-Wall -Wmissing-prototypes
2530 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2531 DEBUG_SAFESTACK must also be set.
2534 *) Change ./Configure so that certain algorithms can be disabled by default.
2535 The new counterpiece to "no-xxx" is "enable-xxx".
2537 The patented RC5 and MDC2 algorithms will now be disabled unless
2538 "enable-rc5" and "enable-mdc2", respectively, are specified.
2540 (IDEA remains enabled despite being patented. This is because IDEA
2541 is frequently required for interoperability, and there is no license
2542 fee for non-commercial use. As before, "no-idea" can be used to
2543 avoid this algorithm.)
2547 *) Add processing of proxy certificates (see RFC 3820). This work was
2548 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2549 EGEE (Enabling Grids for E-science in Europe).
2552 *) RC4 performance overhaul on modern architectures/implementations, such
2553 as Intel P4, IA-64 and AMD64.
2556 *) New utility extract-section.pl. This can be used specify an alternative
2557 section number in a pod file instead of having to treat each file as
2558 a separate case in Makefile. This can be done by adding two lines to the
2561 =for comment openssl_section:XXX
2563 The blank line is mandatory.
2567 *) New arguments -certform, -keyform and -pass for s_client and s_server
2568 to allow alternative format key and certificate files and passphrase
2572 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2573 update associated structures and add various utility functions.
2575 Add new policy related verify parameters, include policy checking in
2576 standard verify code. Enhance 'smime' application with extra parameters
2577 to support policy checking and print out.
2580 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2581 Nehemiah processors. These extensions support AES encryption in hardware
2582 as well as RNG (though RNG support is currently disabled).
2583 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2585 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2588 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2589 [Andy Polyakov and a number of other people]
2591 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2592 implementation contributed by IBM.
2593 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2595 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2596 exponent rather than 'unsigned long'. There is a corresponding change to
2597 the new 'rsa_keygen' element of the RSA_METHOD structure.
2598 [Jelte Jansen, Geoff Thorpe]
2600 *) Functionality for creating the initial serial number file is now
2601 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2603 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2604 number file to 1, which is bound to cause problems. To avoid
2605 the problems while respecting compatibility between different 0.9.7
2606 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2607 CA.pl for serial number initialization. With the new release 0.9.8,
2608 we can fix the problem directly in the 'ca' utility.)
2611 *) Reduced header interdepencies by declaring more opaque objects in
2612 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2613 give fewer recursive includes, which could break lazy source code - so
2614 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2615 developers should define this symbol when building and using openssl to
2616 ensure they track the recommended behaviour, interfaces, [etc], but
2617 backwards-compatible behaviour prevails when this isn't defined.
2620 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2623 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2624 This will generate a random key of the appropriate length based on the
2625 cipher context. The EVP_CIPHER can provide its own random key generation
2626 routine to support keys of a specific form. This is used in the des and
2627 3des routines to generate a key of the correct parity. Update S/MIME
2628 code to use new functions and hence generate correct parity DES keys.
2629 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2630 valid (weak or incorrect parity).
2633 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2634 as looking them up. This is useful when the verified structure may contain
2635 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2636 present unless the new PKCS7_NO_CRL flag is asserted.
2639 *) Extend ASN1 oid configuration module. It now additionally accepts the
2642 shortName = some long name, 1.2.3.4
2645 *) Reimplemented the BN_CTX implementation. There is now no more static
2646 limitation on the number of variables it can handle nor the depth of the
2647 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2648 information can now expand as required, and rather than having a single
2649 static array of bignums, BN_CTX now uses a linked-list of such arrays
2650 allowing it to expand on demand whilst maintaining the usefulness of
2651 BN_CTX's "bundling".
2654 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2655 to allow all RSA operations to function using a single BN_CTX.
2658 *) Preliminary support for certificate policy evaluation and checking. This
2659 is initially intended to pass the tests outlined in "Conformance Testing
2660 of Relying Party Client Certificate Path Processing Logic" v1.07.
2663 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2664 remained unused and not that useful. A variety of other little bignum
2665 tweaks and fixes have also been made continuing on from the audit (see
2669 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2670 associated ASN1, EVP and SSL functions and old ASN1 macros.
2673 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2674 and this should never fail. So the return value from the use of
2675 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2676 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2679 *) BN_CTX_get() should return zero-valued bignums, providing the same
2680 initialised value as BN_new().
2681 [Geoff Thorpe, suggested by Ulf Möller]
2683 *) Support for inhibitAnyPolicy certificate extension.
2686 *) An audit of the BIGNUM code is underway, for which debugging code is
2687 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2688 is considered valid when processing BIGNUMs, and causes execution to
2689 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2690 further steps are taken to deliberately pollute unused data in BIGNUM
2691 structures to try and expose faulty code further on. For now, openssl will
2692 (in its default mode of operation) continue to tolerate the inconsistent
2693 forms that it has tolerated in the past, but authors and packagers should
2694 consider trying openssl and their own applications when compiled with
2695 these debugging symbols defined. It will help highlight potential bugs in
2696 their own code, and will improve the test coverage for OpenSSL itself. At
2697 some point, these tighter rules will become openssl's default to improve
2698 maintainability, though the assert()s and other overheads will remain only
2699 in debugging configurations. See bn.h for more details.
2700 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2702 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2703 that can only be obtained through BN_CTX_new() (which implicitly
2704 initialises it). The presence of this function only made it possible
2705 to overwrite an existing structure (and cause memory leaks).
2708 *) Because of the callback-based approach for implementing LHASH as a
2709 template type, lh_insert() adds opaque objects to hash-tables and
2710 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2711 to clean up those corresponding objects before destroying the hash table
2712 (and losing the object pointers). So some over-zealous constifications in
2713 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2714 objects as "const" and the lh_doall[_arg] callback wrappers are not
2715 prototyped to have "const" restrictions on the object pointers they are
2716 given (and so aren't required to cast them away any more).
2719 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2720 (speed) prefers to use its own implementation. The two implementations
2721 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2722 its object type properly exposed (MS_TM) instead of casting to/from "char
2723 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2724 aren't necessarily the greatest nomenclatures - but this is what was used
2725 internally to the implementation so I've used that for now.
2728 *) Ensure that deprecated functions do not get compiled when
2729 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2730 the self-tests were still using deprecated key-generation functions so
2731 these have been updated also.
2734 *) Reorganise PKCS#7 code to separate the digest location functionality
2735 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2736 New function PKCS7_set_digest() to set the digest type for PKCS#7
2737 digestedData type. Add additional code to correctly generate the
2738 digestedData type and add support for this type in PKCS7 initialization
2742 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2743 structure of type "other".
2746 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2747 sure the loop does correctly stop and breaking ("division by zero")
2748 modulus operations are not performed. The (pre-generated) prime
2749 table crypto/bn/bn_prime.h was already correct, but it could not be
2750 re-generated on some platforms because of the "division by zero"
2751 situation in the script.
2752 [Ralf S. Engelschall]
2754 *) Update support for ECC-based TLS ciphersuites according to
2755 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2756 SHA-1 now is only used for "small" curves (where the
2757 representation of a field element takes up to 24 bytes); for
2758 larger curves, the field element resulting from ECDH is directly
2759 used as premaster secret.
2760 [Douglas Stebila (Sun Microsystems Laboratories)]
2762 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2763 curve secp160r1 to the tests.
2764 [Douglas Stebila (Sun Microsystems Laboratories)]
2766 *) Add the possibility to load symbols globally with DSO.
2767 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2769 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2770 control of the error stack.
2773 *) Add support for STORE in ENGINE.
2776 *) Add the STORE type. The intention is to provide a common interface
2777 to certificate and key stores, be they simple file-based stores, or
2778 HSM-type store, or LDAP stores, or...
2779 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2782 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2783 pass a list of arguments to any function as well as provide a way
2784 for a function to pass data back to the caller.
2787 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2788 works like BUF_strdup() but can be used to duplicate a portion of
2789 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2793 *) Add the function sk_find_ex() which works like sk_find(), but will
2794 return an index to an element even if an exact match couldn't be
2795 found. The index is guaranteed to point at the element where the
2796 searched-for key would be inserted to preserve sorting order.
2799 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2800 takes an extra flags argument for optional functionality. Currently,
2801 the following flags are defined:
2803 OBJ_BSEARCH_VALUE_ON_NOMATCH
2804 This one gets OBJ_bsearch_ex() to return a pointer to the first
2805 element where the comparing function returns a negative or zero
2808 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2809 This one gets OBJ_bsearch_ex() to return a pointer to the first
2810 element where the comparing function returns zero. This is useful
2811 if there are more than one element where the comparing function
2815 *) Make it possible to create self-signed certificates with 'openssl ca'
2816 in such a way that the self-signed certificate becomes part of the
2817 CA database and uses the same mechanisms for serial number generation
2818 as all other certificate signing. The new flag '-selfsign' enables
2819 this functionality. Adapt CA.sh and CA.pl.in.
2822 *) Add functionality to check the public key of a certificate request
2823 against a given private. This is useful to check that a certificate
2824 request can be signed by that key (self-signing).
2827 *) Make it possible to have multiple active certificates with the same
2828 subject in the CA index file. This is done only if the keyword
2829 'unique_subject' is set to 'no' in the main CA section (default
2830 if 'CA_default') of the configuration file. The value is saved
2831 with the database itself in a separate index attribute file,
2832 named like the index file with '.attr' appended to the name.
2835 *) Generate muti valued AVAs using '+' notation in config files for
2839 *) Support for nameConstraints certificate extension.
2842 *) Support for policyConstraints certificate extension.
2845 *) Support for policyMappings certificate extension.
2848 *) Make sure the default DSA_METHOD implementation only uses its
2849 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2850 and change its own handlers to be NULL so as to remove unnecessary
2851 indirection. This lets alternative implementations fallback to the
2852 default implementation more easily.
2855 *) Support for directoryName in GeneralName related extensions
2859 *) Make it possible to link applications using Makefile.shared.
2860 Make that possible even when linking against static libraries!
2863 *) Support for single pass processing for S/MIME signing. This now
2864 means that S/MIME signing can be done from a pipe, in addition
2865 cleartext signing (multipart/signed type) is effectively streaming
2866 and the signed data does not need to be all held in memory.
2868 This is done with a new flag PKCS7_STREAM. When this flag is set
2869 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2870 is done after the data is output (and digests calculated) in
2871 SMIME_write_PKCS7().
2874 *) Add full support for -rpath/-R, both in shared libraries and
2875 applications, at least on the platforms where it's known how
2879 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2880 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2881 will now compute a table of multiples of the generator that
2882 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2883 faster (notably in the case of a single point multiplication,
2884 scalar * generator).
2885 [Nils Larsch, Bodo Moeller]
2887 *) IPv6 support for certificate extensions. The various extensions
2888 which use the IP:a.b.c.d can now take IPv6 addresses using the
2889 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2893 *) Added an ENGINE that implements RSA by performing private key
2894 exponentiations with the GMP library. The conversions to and from
2895 GMP's mpz_t format aren't optimised nor are any montgomery forms
2896 cached, and on x86 it appears OpenSSL's own performance has caught up.
2897 However there are likely to be other architectures where GMP could
2898 provide a boost. This ENGINE is not built in by default, but it can be
2899 specified at Configure time and should be accompanied by the necessary
2900 linker additions, eg;
2901 ./config -DOPENSSL_USE_GMP -lgmp
2904 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2905 testing availability of engines with "-t" - the old behaviour is
2906 produced by increasing the feature's verbosity with "-tt".
2909 *) ECDSA routines: under certain error conditions uninitialized BN objects
2910 could be freed. Solution: make sure initialization is performed early
2911 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2915 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2916 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2917 software implementations. For DSA and DH, parameter generation can
2918 also be overriden by providing the appropriate method callbacks.
2921 *) Change the "progress" mechanism used in key-generation and
2922 primality testing to functions that take a new BN_GENCB pointer in
2923 place of callback/argument pairs. The new API functions have "_ex"
2924 postfixes and the older functions are reimplemented as wrappers for
2925 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2926 declarations of the old functions to help (graceful) attempts to
2927 migrate to the new functions. Also, the new key-generation API
2928 functions operate on a caller-supplied key-structure and return
2929 success/failure rather than returning a key or NULL - this is to
2930 help make "keygen" another member function of RSA_METHOD etc.
2932 Example for using the new callback interface:
2934 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2938 BN_GENCB_set(&my_cb, my_callback, my_arg);
2940 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2941 /* For the meaning of a, b in calls to my_callback(), see the
2942 * documentation of the function that calls the callback.
2943 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2944 * my_callback should return 1 if it wants BN_is_prime_ex()
2945 * to continue, or 0 to stop.
2950 *) Change the ZLIB compression method to be stateful, and make it
2951 available to TLS with the number defined in
2952 draft-ietf-tls-compression-04.txt.
2955 *) Add the ASN.1 structures and functions for CertificatePair, which
2956 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2958 CertificatePair ::= SEQUENCE {
2959 forward [0] Certificate OPTIONAL,
2960 reverse [1] Certificate OPTIONAL,
2961 -- at least one of the pair shall be present -- }
2963 Also implement the PEM functions to read and write certificate
2964 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2966 This needed to be defined, mostly for the sake of the LDAP
2967 attribute crossCertificatePair, but may prove useful elsewhere as
2971 *) Make it possible to inhibit symlinking of shared libraries in
2972 Makefile.shared, for Cygwin's sake.
2975 *) Extend the BIGNUM API by creating a function
2976 void BN_set_negative(BIGNUM *a, int neg);
2977 and a macro that behave like
2978 int BN_is_negative(const BIGNUM *a);
2980 to avoid the need to access 'a->neg' directly in applications.
2983 *) Implement fast modular reduction for pseudo-Mersenne primes
2984 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2985 EC_GROUP_new_curve_GFp() will now automatically use this
2987 [Nils Larsch <nla@trustcenter.de>]
2989 *) Add new lock type (CRYPTO_LOCK_BN).
2992 *) Change the ENGINE framework to automatically load engines
2993 dynamically from specific directories unless they could be
2994 found to already be built in or loaded. Move all the
2995 current engines except for the cryptodev one to a new
2997 The engines in engines/ are built as shared libraries if
2998 the "shared" options was given to ./Configure or ./config.
2999 Otherwise, they are inserted in libcrypto.a.
3000 /usr/local/ssl/engines is the default directory for dynamic
3001 engines, but that can be overriden at configure time through
3002 the usual use of --prefix and/or --openssldir, and at run
3003 time with the environment variable OPENSSL_ENGINES.
3004 [Geoff Thorpe and Richard Levitte]
3006 *) Add Makefile.shared, a helper makefile to build shared
3007 libraries. Addapt Makefile.org.
3010 *) Add version info to Win32 DLLs.
3011 [Peter 'Luna' Runestig" <peter@runestig.com>]
3013 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3014 can be added using this API to created arbitrary PKCS#12
3015 files while avoiding the low level API.
3017 New options to PKCS12_create(), key or cert can be NULL and
3018 will then be omitted from the output file. The encryption
3019 algorithm NIDs can be set to -1 for no encryption, the mac
3020 iteration count can be set to 0 to omit the mac.
3022 Enhance pkcs12 utility by making the -nokeys and -nocerts
3023 options work when creating a PKCS#12 file. New option -nomac
3024 to omit the mac, NONE can be set for an encryption algorithm.
3025 New code is modified to use the enhanced PKCS12_create()
3026 instead of the low level API.
3029 *) Extend ASN1 encoder to support indefinite length constructed
3030 encoding. This can output sequences tags and octet strings in
3031 this form. Modify pk7_asn1.c to support indefinite length
3032 encoding. This is experimental and needs additional code to
3033 be useful, such as an ASN1 bio and some enhanced streaming
3036 Extend template encode functionality so that tagging is passed
3037 down to the template encoder.
3040 *) Let 'openssl req' fail if an argument to '-newkey' is not
3041 recognized instead of using RSA as a default.
3044 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3045 As these are not official, they are not included in "ALL";
3046 the "ECCdraft" ciphersuite group alias can be used to select them.
3047 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3049 *) Add ECDH engine support.
3050 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3052 *) Add ECDH in new directory crypto/ecdh/.
3053 [Douglas Stebila (Sun Microsystems Laboratories)]
3055 *) Let BN_rand_range() abort with an error after 100 iterations
3056 without success (which indicates a broken PRNG).
3059 *) Change BN_mod_sqrt() so that it verifies that the input value
3060 is really the square of the return value. (Previously,
3061 BN_mod_sqrt would show GIGO behaviour.)
3064 *) Add named elliptic curves over binary fields from X9.62, SECG,
3065 and WAP/WTLS; add OIDs that were still missing.
3067 [Sheueling Chang Shantz and Douglas Stebila
3068 (Sun Microsystems Laboratories)]
3070 *) Extend the EC library for elliptic curves over binary fields
3071 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3074 EC_GF2m_simple_method
3078 EC_GROUP_new_curve_GF2m
3079 EC_GROUP_set_curve_GF2m
3080 EC_GROUP_get_curve_GF2m
3081 EC_POINT_set_affine_coordinates_GF2m
3082 EC_POINT_get_affine_coordinates_GF2m
3083 EC_POINT_set_compressed_coordinates_GF2m
3085 Point compression for binary fields is disabled by default for
3086 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3089 As binary polynomials are represented as BIGNUMs, various members
3090 of the EC_GROUP and EC_POINT data structures can be shared
3091 between the implementations for prime fields and binary fields;
3092 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3093 are essentially identical to their ..._GFp counterparts.
3094 (For simplicity, the '..._GFp' prefix has been dropped from
3095 various internal method names.)
3097 An internal 'field_div' method (similar to 'field_mul' and
3098 'field_sqr') has been added; this is used only for binary fields.
3100 [Sheueling Chang Shantz and Douglas Stebila
3101 (Sun Microsystems Laboratories)]
3103 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3104 through methods ('mul', 'precompute_mult').
3106 The generic implementations (now internally called 'ec_wNAF_mul'
3107 and 'ec_wNAF_precomputed_mult') remain the default if these
3108 methods are undefined.
3110 [Sheueling Chang Shantz and Douglas Stebila
3111 (Sun Microsystems Laboratories)]
3113 *) New function EC_GROUP_get_degree, which is defined through
3114 EC_METHOD. For curves over prime fields, this returns the bit
3115 length of the modulus.
3117 [Sheueling Chang Shantz and Douglas Stebila
3118 (Sun Microsystems Laboratories)]
3120 *) New functions EC_GROUP_dup, EC_POINT_dup.
3121 (These simply call ..._new and ..._copy).
3123 [Sheueling Chang Shantz and Douglas Stebila
3124 (Sun Microsystems Laboratories)]
3126 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3127 Polynomials are represented as BIGNUMs (where the sign bit is not
3128 used) in the following functions [macros]:
3131 BN_GF2m_sub [= BN_GF2m_add]
3132 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3133 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3134 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3136 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3137 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3138 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3139 BN_GF2m_cmp [= BN_ucmp]
3141 (Note that only the 'mod' functions are actually for fields GF(2^m).
3142 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3144 For some functions, an the irreducible polynomial defining a
3145 field can be given as an 'unsigned int[]' with strictly
3146 decreasing elements giving the indices of those bits that are set;
3147 i.e., p[] represents the polynomial
3148 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3150 p[0] > p[1] > ... > p[k] = 0.
3151 This applies to the following functions:
3156 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3157 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3159 BN_GF2m_mod_sqrt_arr
3160 BN_GF2m_mod_solve_quad_arr
3164 Conversion can be performed by the following functions:
3169 bntest.c has additional tests for binary polynomial arithmetic.
3171 Two implementations for BN_GF2m_mod_div() are available.
3172 The default algorithm simply uses BN_GF2m_mod_inv() and
3173 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3174 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3175 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3177 [Sheueling Chang Shantz and Douglas Stebila
3178 (Sun Microsystems Laboratories)]
3180 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3181 functionality is disabled at compile-time.
3182 [Douglas Stebila <douglas.stebila@sun.com>]
3184 *) Change default behaviour of 'openssl asn1parse' so that more
3185 information is visible when viewing, e.g., a certificate:
3187 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3188 mode the content of non-printable OCTET STRINGs is output in a
3189 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3190 avoid the appearance of a printable string.
3191 [Nils Larsch <nla@trustcenter.de>]
3193 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3195 EC_GROUP_set_asn1_flag()
3196 EC_GROUP_get_asn1_flag()
3197 EC_GROUP_set_point_conversion_form()
3198 EC_GROUP_get_point_conversion_form()
3199 These control ASN1 encoding details:
3200 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3201 has been set to OPENSSL_EC_NAMED_CURVE.
3202 - Points are encoded in uncompressed form by default; options for
3203 asn1_for are as for point2oct, namely
3204 POINT_CONVERSION_COMPRESSED
3205 POINT_CONVERSION_UNCOMPRESSED
3206 POINT_CONVERSION_HYBRID
3208 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3211 EC_GROUP_get0_seed()
3212 EC_GROUP_get_seed_len()
3213 This is used only for ASN1 purposes (so far).
3214 [Nils Larsch <nla@trustcenter.de>]
3216 *) Add 'field_type' member to EC_METHOD, which holds the NID
3217 of the appropriate field type OID. The new function
3218 EC_METHOD_get_field_type() returns this value.
3219 [Nils Larsch <nla@trustcenter.de>]
3224 EC_POINT_point2hex()
3225 EC_POINT_hex2point()
3226 providing useful interfaces to EC_POINT_point2oct() and
3227 EC_POINT_oct2point().
3228 [Nils Larsch <nla@trustcenter.de>]
3230 *) Change internals of the EC library so that the functions
3231 EC_GROUP_set_generator()
3232 EC_GROUP_get_generator()
3233 EC_GROUP_get_order()
3234 EC_GROUP_get_cofactor()
3235 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3236 to methods, which would lead to unnecessary code duplication when
3237 adding different types of curves.
3238 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3240 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3241 arithmetic, and such that modified wNAFs are generated
3242 (which avoid length expansion in many cases).
3245 *) Add a function EC_GROUP_check_discriminant() (defined via
3246 EC_METHOD) that verifies that the curve discriminant is non-zero.
3248 Add a function EC_GROUP_check() that makes some sanity tests
3249 on a EC_GROUP, its generator and order. This includes
3250 EC_GROUP_check_discriminant().
3251 [Nils Larsch <nla@trustcenter.de>]
3253 *) Add ECDSA in new directory crypto/ecdsa/.
3255 Add applications 'openssl ecparam' and 'openssl ecdsa'
3256 (these are based on 'openssl dsaparam' and 'openssl dsa').
3258 ECDSA support is also included in various other files across the
3259 library. Most notably,
3260 - 'openssl req' now has a '-newkey ecdsa:file' option;
3261 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3262 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3263 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3264 them suitable for ECDSA where domain parameters must be
3265 extracted before the specific public key;
3266 - ECDSA engine support has been added.
3267 [Nils Larsch <nla@trustcenter.de>]
3269 *) Include some named elliptic curves, and add OIDs from X9.62,
3270 SECG, and WAP/WTLS. Each curve can be obtained from the new
3272 EC_GROUP_new_by_curve_name(),
3273 and the list of available named curves can be obtained with
3274 EC_get_builtin_curves().
3275 Also add a 'curve_name' member to EC_GROUP objects, which can be
3277 EC_GROUP_set_curve_name()
3278 EC_GROUP_get_curve_name()
3279 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3281 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3282 was actually never needed) and in BN_mul(). The removal in BN_mul()
3283 required a small change in bn_mul_part_recursive() and the addition
3284 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3285 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3286 bn_sub_words() and bn_add_words() except they take arrays with
3290 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3292 *) Cleanse PEM buffers before freeing them since they may contain
3294 [Benjamin Bennett <ben@psc.edu>]
3296 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3297 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3298 authentication-only ciphersuites.
3301 *) Since AES128 and AES256 share a single mask bit in the logic of
3302 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3303 kludge to work properly if AES128 is available and AES256 isn't.
3306 *) Expand security boundary to match 1.1.1 module.
3309 *) Remove redundant features: hash file source, editing of test vectors
3310 modify fipsld to use external fips_premain.c signature.
3313 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3314 run algorithm test programs.
3317 *) Make algorithm test programs more tolerant of whitespace.
3320 *) Have SSL/TLS server implementation tolerate "mismatched" record
3321 protocol version while receiving ClientHello even if the
3322 ClientHello is fragmented. (The server can't insist on the
3323 particular protocol version it has chosen before the ServerHello
3324 message has informed the client about his choice.)
3327 *) Load error codes if they are not already present instead of using a
3328 static variable. This allows them to be cleanly unloaded and reloaded.
3331 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3333 *) Introduce limits to prevent malicious keys being able to
3334 cause a denial of service. (CVE-2006-2940)
3335 [Steve Henson, Bodo Moeller]
3337 *) Fix ASN.1 parsing of certain invalid structures that can result
3338 in a denial of service. (CVE-2006-2937) [Steve Henson]
3340 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3341 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3343 *) Fix SSL client code which could crash if connecting to a
3344 malicious SSLv2 server. (CVE-2006-4343)
3345 [Tavis Ormandy and Will Drewry, Google Security Team]
3347 *) Change ciphersuite string processing so that an explicit
3348 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3349 will no longer include "AES128-SHA"), and any other similar
3350 ciphersuite (same bitmap) from *other* protocol versions (so that
3351 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3352 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3353 changes from 0.9.8b and 0.9.8d.
3356 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3358 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3359 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3361 *) Change the Unix randomness entropy gathering to use poll() when
3362 possible instead of select(), since the latter has some
3363 undesirable limitations.
3364 [Darryl Miles via Richard Levitte and Bodo Moeller]
3366 *) Disable rogue ciphersuites:
3368 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3369 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3370 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3372 The latter two were purportedly from
3373 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3376 Also deactive the remaining ciphersuites from
3377 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3378 unofficial, and the ID has long expired.
3381 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3382 dual-core machines) and other potential thread-safety issues.
3385 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3387 *) Adapt fipsld and the build system to link against the validated FIPS
3388 module in FIPS mode.
3391 *) Fixes for VC++ 2005 build under Windows.
3394 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3395 from a Windows bash shell such as MSYS. It is autodetected from the
3396 "config" script when run from a VC++ environment. Modify standard VC++
3397 build to use fipscanister.o from the GNU make build.
3400 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3402 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3403 The value now differs depending on if you build for FIPS or not.
3404 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3405 safely run with a non-FIPSed libcrypto, as it may crash because of
3406 the difference induced by this change.
3409 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3411 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3412 (part of SSL_OP_ALL). This option used to disable the
3413 countermeasure against man-in-the-middle protocol-version
3414 rollback in the SSL 2.0 server implementation, which is a bad
3415 idea. (CVE-2005-2969)
3417 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3418 for Information Security, National Institute of Advanced Industrial
3419 Science and Technology [AIST], Japan)]
3421 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3422 mainly for FIPS compliance and not fully integrated at this stage.
3425 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3426 the exponentiation using a fixed-length exponent. (Otherwise,
3427 the information leaked through timing could expose the secret key
3428 after many signatures; cf. Bleichenbacher's attack on DSA with
3432 *) Make a new fixed-window mod_exp implementation the default for
3433 RSA, DSA, and DH private-key operations so that the sequence of
3434 squares and multiplies and the memory access pattern are
3435 independent of the particular secret key. This will mitigate
3436 cache-timing and potential related attacks.
3438 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3439 and this is automatically used by BN_mod_exp_mont() if the new flag
3440 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3441 will use this BN flag for private exponents unless the flag
3442 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3443 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3445 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3447 *) Change the client implementation for SSLv23_method() and
3448 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3449 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3450 (Previously, the SSL 2.0 backwards compatible Client Hello
3451 message format would be used even with SSL_OP_NO_SSLv2.)
3454 *) Add support for smime-type MIME parameter in S/MIME messages which some
3458 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3459 a threadsafe manner. Modify rsa code to use new function and add calls
3460 to dsa and dh code (which had race conditions before).
3463 *) Include the fixed error library code in the C error file definitions
3464 instead of fixing them up at runtime. This keeps the error code
3465 structures constant.
3468 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3470 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3473 *) Fixes for newer kerberos headers. NB: the casts are needed because
3474 the 'length' field is signed on one version and unsigned on another
3475 with no (?) obvious way to tell the difference, without these VC++
3476 complains. Also the "definition" of FAR (blank) is no longer included
3477 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3478 some needed definitions.
3481 *) Undo Cygwin change.
3484 *) Added support for proxy certificates according to RFC 3820.
3485 Because they may be a security thread to unaware applications,
3486 they must be explicitely allowed in run-time. See
3487 docs/HOWTO/proxy_certificates.txt for further information.
3490 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3492 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3493 server and client random values. Previously
3494 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3495 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3497 This change has negligible security impact because:
3499 1. Server and client random values still have 24 bytes of pseudo random
3502 2. Server and client random values are sent in the clear in the initial
3505 3. The master secret is derived using the premaster secret (48 bytes in
3506 size for static RSA ciphersuites) as well as client server and random
3509 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3512 [Stephen Henson, reported by UK NISCC]
3514 *) Use Windows randomness collection on Cygwin.
3517 *) Fix hang in EGD/PRNGD query when communication socket is closed
3518 prematurely by EGD/PRNGD.
3519 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3521 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3524 *) Back-port of selected performance improvements from development
3525 branch, as well as improved support for PowerPC platforms.
3528 *) Add lots of checks for memory allocation failure, error codes to indicate
3529 failure and freeing up memory if a failure occurs.
3530 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3532 *) Add new -passin argument to dgst.
3535 *) Perform some character comparisons of different types in X509_NAME_cmp:
3536 this is needed for some certificates that reencode DNs into UTF8Strings
3537 (in violation of RFC3280) and can't or wont issue name rollover
3541 *) Make an explicit check during certificate validation to see that
3542 the CA setting in each certificate on the chain is correct. As a
3543 side effect always do the following basic checks on extensions,
3544 not just when there's an associated purpose to the check:
3546 - if there is an unhandled critical extension (unless the user
3547 has chosen to ignore this fault)
3548 - if the path length has been exceeded (if one is set at all)
3549 - that certain extensions fit the associated purpose (if one has
3553 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3555 *) Avoid a race condition when CRLs are checked in a multi threaded
3556 environment. This would happen due to the reordering of the revoked
3557 entries during signature checking and serial number lookup. Now the
3558 encoding is cached and the serial number sort performed under a lock.
3559 Add new STACK function sk_is_sorted().
3562 *) Add Delta CRL to the extension code.
3565 *) Various fixes to s3_pkt.c so alerts are sent properly.
3566 [David Holmes <d.holmes@f5.com>]
3568 *) Reduce the chances of duplicate issuer name and serial numbers (in
3569 violation of RFC3280) using the OpenSSL certificate creation utilities.
3570 This is done by creating a random 64 bit value for the initial serial
3571 number when a serial number file is created or when a self signed
3572 certificate is created using 'openssl req -x509'. The initial serial
3573 number file is created using 'openssl x509 -next_serial' in CA.pl
3574 rather than being initialized to 1.
3577 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3579 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3580 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3581 [Joe Orton, Steve Henson]
3583 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3585 [Joe Orton, Steve Henson]
3587 *) Make it possible to have multiple active certificates with the same
3588 subject in the CA index file. This is done only if the keyword
3589 'unique_subject' is set to 'no' in the main CA section (default
3590 if 'CA_default') of the configuration file. The value is saved
3591 with the database itself in a separate index attribute file,
3592 named like the index file with '.attr' appended to the name.
3595 *) X509 verify fixes. Disable broken certificate workarounds when
3596 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3597 keyUsage extension present. Don't accept CRLs with unhandled critical
3598 extensions: since verify currently doesn't process CRL extensions this
3599 rejects a CRL with *any* critical extensions. Add new verify error codes
3603 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3604 A clarification of RFC2560 will require the use of OCTET STRINGs and
3605 some implementations cannot handle the current raw format. Since OpenSSL
3606 copies and compares OCSP nonces as opaque blobs without any attempt at
3607 parsing them this should not create any compatibility issues.
3610 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3611 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3612 this HMAC (and other) operations are several times slower than OpenSSL
3616 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3617 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3619 *) Use the correct content when signing type "other".
3622 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3624 *) Fix various bugs revealed by running the NISCC test suite:
3626 Stop out of bounds reads in the ASN1 code when presented with
3627 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3629 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3631 If verify callback ignores invalid public key errors don't try to check
3632 certificate signature with the NULL public key.
3636 *) New -ignore_err option in ocsp application to stop the server
3637 exiting on the first error in a request.
3640 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3641 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3645 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3646 extra data after the compression methods not only for TLS 1.0
3647 but also for SSL 3.0 (as required by the specification).
3648 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3650 *) Change X509_certificate_type() to mark the key as exported/exportable
3651 when it's 512 *bits* long, not 512 bytes.
3654 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3655 blocks during encryption.
3658 *) Various fixes to base64 BIO and non blocking I/O. On write
3659 flushes were not handled properly if the BIO retried. On read
3660 data was not being buffered properly and had various logic bugs.
3661 This also affects blocking I/O when the data being decoded is a
3665 *) Various S/MIME bugfixes and compatibility changes:
3666 output correct application/pkcs7 MIME type if
3667 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3668 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3669 of files as .eml work). Correctly handle very long lines in MIME
3673 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3675 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3676 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3677 a protocol version number mismatch like a decryption error
3678 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3681 *) Turn on RSA blinding by default in the default implementation
3682 to avoid a timing attack. Applications that don't want it can call
3683 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3684 They would be ill-advised to do so in most cases.
3685 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3687 *) Change RSA blinding code so that it works when the PRNG is not
3688 seeded (in this case, the secret RSA exponent is abused as
3689 an unpredictable seed -- if it is not unpredictable, there
3690 is no point in blinding anyway). Make RSA blinding thread-safe
3691 by remembering the creator's thread ID in rsa->blinding and
3692 having all other threads use local one-time blinding factors
3693 (this requires more computation than sharing rsa->blinding, but
3694 avoids excessive locking; and if an RSA object is not shared
3695 between threads, blinding will still be very fast).
3698 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3699 ENGINE as defaults for all supported algorithms irrespective of
3700 the 'flags' parameter. 'flags' is now honoured, so applications
3701 should make sure they are passing it correctly.
3704 *) Target "mingw" now allows native Windows code to be generated in
3705 the Cygwin environment as well as with the MinGW compiler.
3708 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3710 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3711 via timing by performing a MAC computation even if incorrrect
3712 block cipher padding has been found. This is a countermeasure
3713 against active attacks where the attacker has to distinguish
3714 between bad padding and a MAC verification error. (CVE-2003-0078)
3716 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3717 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3718 Martin Vuagnoux (EPFL, Ilion)]
3720 *) Make the no-err option work as intended. The intention with no-err
3721 is not to have the whole error stack handling routines removed from
3722 libcrypto, it's only intended to remove all the function name and
3723 reason texts, thereby removing some of the footprint that may not
3724 be interesting if those errors aren't displayed anyway.
3726 NOTE: it's still possible for any application or module to have it's
3727 own set of error texts inserted. The routines are there, just not
3728 used by default when no-err is given.
3731 *) Add support for FreeBSD on IA64.
3732 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3734 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3735 Kerberos function mit_des_cbc_cksum(). Before this change,
3736 the value returned by DES_cbc_cksum() was like the one from
3737 mit_des_cbc_cksum(), except the bytes were swapped.
3738 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3740 *) Allow an application to disable the automatic SSL chain building.
3741 Before this a rather primitive chain build was always performed in
3742 ssl3_output_cert_chain(): an application had no way to send the
3743 correct chain if the automatic operation produced an incorrect result.
3745 Now the chain builder is disabled if either:
3747 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3749 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3751 The reasoning behind this is that an application would not want the
3752 auto chain building to take place if extra chain certificates are
3753 present and it might also want a means of sending no additional
3754 certificates (for example the chain has two certificates and the
3758 *) Add the possibility to build without the ENGINE framework.
3759 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3761 *) Under Win32 gmtime() can return NULL: check return value in
3762 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3765 *) DSA routines: under certain error conditions uninitialized BN objects
3766 could be freed. Solution: make sure initialization is performed early
3767 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3768 Nils Larsch <nla@trustcenter.de> via PR#459)
3771 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3772 checked on reconnect on the client side, therefore session resumption
3773 could still fail with a "ssl session id is different" error. This
3774 behaviour is masked when SSL_OP_ALL is used due to
3775 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3776 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3777 followup to PR #377.
3780 *) IA-32 assembler support enhancements: unified ELF targets, support
3781 for SCO/Caldera platforms, fix for Cygwin shared build.
3784 *) Add support for FreeBSD on sparc64. As a consequence, support for
3785 FreeBSD on non-x86 processors is separate from x86 processors on
3786 the config script, much like the NetBSD support.
3787 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3789 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3791 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3794 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3795 code (06) was taken as the first octet of the session ID and the last
3796 octet was ignored consequently. As a result SSLv2 client side session
3797 caching could not have worked due to the session ID mismatch between
3799 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3803 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3804 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3808 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3809 seems that in spite of existing for more than a year, many application
3810 author have done nothing to provide the necessary callbacks, which
3811 means that this particular engine will not work properly anywhere.
3812 This is a very unfortunate situation which forces us, in the name
3813 of usability, to give the hw_ncipher.c a static lock, which is part
3815 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3816 appear in 0.9.8 or later. We EXPECT application authors to have
3817 dealt properly with this when 0.9.8 is released (unless we actually
3818 make such changes in the libcrypto locking code that changes will
3819 have to be made anyway).
3822 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3823 octets have been read, EOF or an error occurs. Without this change
3824 some truncated ASN1 structures will not produce an error.
3827 *) Disable Heimdal support, since it hasn't been fully implemented.
3828 Still give the possibility to force the use of Heimdal, but with
3829 warnings and a request that patches get sent to openssl-dev.
3832 *) Add the VC-CE target, introduce the WINCE sysname, and add
3833 INSTALL.WCE and appropriate conditionals to make it build.
3834 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3836 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3837 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3838 edit numbers of the version.
3839 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3841 *) Introduce safe string copy and catenation functions
3842 (BUF_strlcpy() and BUF_strlcat()).
3843 [Ben Laurie (CHATS) and Richard Levitte]
3845 *) Avoid using fixed-size buffers for one-line DNs.
3846 [Ben Laurie (CHATS)]
3848 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3849 resizing buffers containing secrets, and use where appropriate.
3850 [Ben Laurie (CHATS)]
3852 *) Avoid using fixed size buffers for configuration file location.
3853 [Ben Laurie (CHATS)]
3855 *) Avoid filename truncation for various CA files.
3856 [Ben Laurie (CHATS)]
3858 *) Use sizeof in preference to magic numbers.
3859 [Ben Laurie (CHATS)]
3861 *) Avoid filename truncation in cert requests.
3862 [Ben Laurie (CHATS)]
3864 *) Add assertions to check for (supposedly impossible) buffer
3866 [Ben Laurie (CHATS)]
3868 *) Don't cache truncated DNS entries in the local cache (this could
3869 potentially lead to a spoofing attack).
3870 [Ben Laurie (CHATS)]
3872 *) Fix various buffers to be large enough for hex/decimal
3873 representations in a platform independent manner.
3874 [Ben Laurie (CHATS)]
3876 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3877 resizing buffers containing secrets, and use where appropriate.
3878 [Ben Laurie (CHATS)]
3880 *) Add BIO_indent() to avoid much slightly worrying code to do
3882 [Ben Laurie (CHATS)]
3884 *) Convert sprintf()/BIO_puts() to BIO_printf().
3885 [Ben Laurie (CHATS)]
3887 *) buffer_gets() could terminate with the buffer only half
3889 [Ben Laurie (CHATS)]
3891 *) Add assertions to prevent user-supplied crypto functions from
3892 overflowing internal buffers by having large block sizes, etc.
3893 [Ben Laurie (CHATS)]
3895 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3897 [Ben Laurie (CHATS)]
3899 *) Eliminate unused copy of key in RC4.
3900 [Ben Laurie (CHATS)]
3902 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3903 [Ben Laurie (CHATS)]
3905 *) Fix off-by-one error in EGD path.
3906 [Ben Laurie (CHATS)]
3908 *) If RANDFILE path is too long, ignore instead of truncating.
3909 [Ben Laurie (CHATS)]
3911 *) Eliminate unused and incorrectly sized X.509 structure
3913 [Ben Laurie (CHATS)]
3915 *) Eliminate unused and dangerous function knumber().
3916 [Ben Laurie (CHATS)]
3918 *) Eliminate unused and dangerous structure, KSSL_ERR.
3919 [Ben Laurie (CHATS)]
3921 *) Protect against overlong session ID context length in an encoded
3922 session object. Since these are local, this does not appear to be
3924 [Ben Laurie (CHATS)]
3926 *) Change from security patch (see 0.9.6e below) that did not affect
3927 the 0.9.6 release series:
3929 Remote buffer overflow in SSL3 protocol - an attacker could
3930 supply an oversized master key in Kerberos-enabled versions.
3932 [Ben Laurie (CHATS)]
3934 *) Change the SSL kerb5 codes to match RFC 2712.
3937 *) Make -nameopt work fully for req and add -reqopt switch.
3938 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3940 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3941 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3943 *) Make sure tests can be performed even if the corresponding algorithms
3944 have been removed entirely. This was also the last step to make
3945 OpenSSL compilable with DJGPP under all reasonable conditions.
3946 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3948 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3949 to allow version independent disabling of normally unselected ciphers,
3950 which may be activated as a side-effect of selecting a single cipher.
3952 (E.g., cipher list string "RSA" enables ciphersuites that are left
3953 out of "ALL" because they do not provide symmetric encryption.
3954 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3955 [Lutz Jaenicke, Bodo Moeller]
3957 *) Add appropriate support for separate platform-dependent build
3958 directories. The recommended way to make a platform-dependent
3959 build directory is the following (tested on Linux), maybe with
3962 # Place yourself outside of the OpenSSL source tree. In
3963 # this example, the environment variable OPENSSL_SOURCE
3964 # is assumed to contain the absolute OpenSSL source directory.
3965 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3966 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3967 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3968 mkdir -p `dirname $F`
3969 ln -s $OPENSSL_SOURCE/$F $F
3972 To be absolutely sure not to disturb the source tree, a "make clean"
3973 is a good thing. If it isn't successfull, don't worry about it,
3974 it probably means the source directory is very clean.
3977 *) Make sure any ENGINE control commands make local copies of string
3978 pointers passed to them whenever necessary. Otherwise it is possible
3979 the caller may have overwritten (or deallocated) the original string
3980 data when a later ENGINE operation tries to use the stored values.
3981 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3983 *) Improve diagnostics in file reading and command-line digests.
3984 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3986 *) Add AES modes CFB and OFB to the object database. Correct an
3987 error in AES-CFB decryption.
3990 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3991 allows existing EVP_CIPHER_CTX structures to be reused after
3992 calling EVP_*Final(). This behaviour is used by encryption
3993 BIOs and some applications. This has the side effect that
3994 applications must explicitly clean up cipher contexts with
3995 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3998 *) Check the values of dna and dnb in bn_mul_recursive before calling
3999 bn_mul_comba (a non zero value means the a or b arrays do not contain
4000 n2 elements) and fallback to bn_mul_normal if either is not zero.
4003 *) Fix escaping of non-ASCII characters when using the -subj option
4004 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4007 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4008 form for "surname", serialNumber has no short form.
4009 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4010 therefore remove "mail" short name for "internet 7".
4011 The OID for unique identifiers in X509 certificates is
4012 x500UniqueIdentifier, not uniqueIdentifier.
4013 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4016 *) Add an "init" command to the ENGINE config module and auto initialize
4017 ENGINEs. Without any "init" command the ENGINE will be initialized
4018 after all ctrl commands have been executed on it. If init=1 the
4019 ENGINE is initailized at that point (ctrls before that point are run
4020 on the uninitialized ENGINE and after on the initialized one). If
4021 init=0 then the ENGINE will not be iniatialized at all.
4024 *) Fix the 'app_verify_callback' interface so that the user-defined
4025 argument is actually passed to the callback: In the
4026 SSL_CTX_set_cert_verify_callback() prototype, the callback
4027 declaration has been changed from
4030 int (*cb)(X509_STORE_CTX *,void *);
4031 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4032 i=s->ctx->app_verify_callback(&ctx)
4033 has been changed into
4034 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4036 To update applications using SSL_CTX_set_cert_verify_callback(),
4037 a dummy argument can be added to their callback functions.
4038 [D. K. Smetters <smetters@parc.xerox.com>]
4040 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4041 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4043 *) Add and OPENSSL_LOAD_CONF define which will cause
4044 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4045 This allows older applications to transparently support certain
4046 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4047 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4048 load the config file and OPENSSL_add_all_algorithms_conf() which will
4049 always load it have also been added.
4052 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4053 Adjust NIDs and EVP layer.
4054 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4056 *) Config modules support in openssl utility.
4058 Most commands now load modules from the config file,
4059 though in a few (such as version) this isn't done
4060 because it couldn't be used for anything.
4062 In the case of ca and req the config file used is
4063 the same as the utility itself: that is the -config
4064 command line option can be used to specify an
4068 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4069 use "openssl_conf" if filename is NULL use default openssl config file.
4072 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4073 config section name. Add a new flag to tolerate a missing config file
4074 and move code to CONF_modules_load_file().
4077 *) Support for crypto accelerator cards from Accelerated Encryption
4078 Processing, www.aep.ie. (Use engine 'aep')
4079 The support was copied from 0.9.6c [engine] and adapted/corrected
4080 to work with the new engine framework.
4081 [AEP Inc. and Richard Levitte]
4083 *) Support for SureWare crypto accelerator cards from Baltimore
4084 Technologies. (Use engine 'sureware')
4085 The support was copied from 0.9.6c [engine] and adapted
4086 to work with the new engine framework.
4089 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4090 make the newer ENGINE framework commands for the CHIL engine work.
4091 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4093 *) Make it possible to produce shared libraries on ReliantUNIX.
4094 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4096 *) Add the configuration target debug-linux-ppro.
4097 Make 'openssl rsa' use the general key loading routines
4098 implemented in apps.c, and make those routines able to
4099 handle the key format FORMAT_NETSCAPE and the variant
4101 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4103 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4104 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4106 *) Add -keyform to rsautl, and document -engine.
4107 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4109 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4110 BIO_R_NO_SUCH_FILE error code rather than the generic
4111 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4114 *) Add new functions
4116 ERR_peek_last_error_line
4117 ERR_peek_last_error_line_data.
4118 These are similar to
4121 ERR_peek_error_line_data,
4122 but report on the latest error recorded rather than the first one
4123 still in the error queue.
4124 [Ben Laurie, Bodo Moeller]
4126 *) default_algorithms option in ENGINE config module. This allows things
4128 default_algorithms = ALL
4129 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4132 *) Prelminary ENGINE config module.
4135 *) New experimental application configuration code.
4138 *) Change the AES code to follow the same name structure as all other
4139 symmetric ciphers, and behave the same way. Move everything to
4140 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4141 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4143 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4144 [Ben Laurie and Theo de Raadt]
4146 *) Add option to output public keys in req command.
4147 [Massimiliano Pala madwolf@openca.org]
4149 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4150 (up to about 10% better than before for P-192 and P-224).
4153 *) New functions/macros
4155 SSL_CTX_set_msg_callback(ctx, cb)
4156 SSL_CTX_set_msg_callback_arg(ctx, arg)
4157 SSL_set_msg_callback(ssl, cb)
4158 SSL_set_msg_callback_arg(ssl, arg)
4160 to request calling a callback function
4162 void cb(int write_p, int version, int content_type,
4163 const void *buf, size_t len, SSL *ssl, void *arg)
4165 whenever a protocol message has been completely received
4166 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4167 protocol version according to which the SSL library interprets
4168 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4169 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4170 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4171 specification (change_cipher_spec(20), alert(21), handshake(22)).
4172 'buf' and 'len' point to the actual message, 'ssl' to the
4173 SSL object, and 'arg' is the application-defined value set by
4174 SSL[_CTX]_set_msg_callback_arg().
4176 'openssl s_client' and 'openssl s_server' have new '-msg' options
4177 to enable a callback that displays all protocol messages.
4180 *) Change the shared library support so shared libraries are built as
4181 soon as the corresponding static library is finished, and thereby get
4182 openssl and the test programs linked against the shared library.
4183 This still only happens when the keyword "shard" has been given to
4184 the configuration scripts.
4186 NOTE: shared library support is still an experimental thing, and
4187 backward binary compatibility is still not guaranteed.
4188 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4190 *) Add support for Subject Information Access extension.
4191 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4193 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4194 additional bytes when new memory had to be allocated, not just
4195 when reusing an existing buffer.
4198 *) New command line and configuration option 'utf8' for the req command.
4199 This allows field values to be specified as UTF8 strings.
4202 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4203 runs for the former and machine-readable output for the latter.
4206 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4207 of the e-mail address in the DN (i.e., it will go into a certificate
4208 extension only). The new configuration file option 'email_in_dn = no'
4209 has the same effect.
4210 [Massimiliano Pala madwolf@openca.org]
4212 *) Change all functions with names starting with des_ to be starting
4213 with DES_ instead. Add wrappers that are compatible with libdes,
4214 but are named _ossl_old_des_*. Finally, add macros that map the
4215 des_* symbols to the corresponding _ossl_old_des_* if libdes
4216 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4217 desired, the des_* symbols will be mapped to DES_*, with one
4220 Since we provide two compatibility mappings, the user needs to
4221 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4222 compatibility is desired. The default (i.e., when that macro
4223 isn't defined) is OpenSSL 0.9.6c compatibility.
4225 There are also macros that enable and disable the support of old
4226 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4227 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4228 are defined, the default will apply: to support the old des routines.
4230 In either case, one must include openssl/des.h to get the correct
4231 definitions. Do not try to just include openssl/des_old.h, that
4234 NOTE: This is a major break of an old API into a new one. Software
4235 authors are encouraged to switch to the DES_ style functions. Some
4236 time in the future, des_old.h and the libdes compatibility functions
4237 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4238 default), and then completely removed.
4241 *) Test for certificates which contain unsupported critical extensions.
4242 If such a certificate is found during a verify operation it is
4243 rejected by default: this behaviour can be overridden by either
4244 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4245 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4246 X509_supported_extension() has also been added which returns 1 if a
4247 particular extension is supported.
4250 *) Modify the behaviour of EVP cipher functions in similar way to digests
4251 to retain compatibility with existing code.
4254 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4255 compatibility with existing code. In particular the 'ctx' parameter does
4256 not have to be to be initialized before the call to EVP_DigestInit() and
4257 it is tidied up after a call to EVP_DigestFinal(). New function
4258 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4259 EVP_MD_CTX_copy() changed to not require the destination to be
4260 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4261 requires the destination to be valid.
4263 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4264 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4267 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4268 so that complete 'Handshake' protocol structures are kept in memory
4269 instead of overwriting 'msg_type' and 'length' with 'body' data.
4272 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4273 [Massimo Santin via Richard Levitte]
4275 *) Major restructuring to the underlying ENGINE code. This includes
4276 reduction of linker bloat, separation of pure "ENGINE" manipulation
4277 (initialisation, etc) from functionality dealing with implementations
4278 of specific crypto iterfaces. This change also introduces integrated
4279 support for symmetric ciphers and digest implementations - so ENGINEs
4280 can now accelerate these by providing EVP_CIPHER and EVP_MD
4281 implementations of their own. This is detailed in crypto/engine/README
4282 as it couldn't be adequately described here. However, there are a few
4283 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4284 were changed in the original introduction of ENGINE code have now
4285 reverted back - the hooking from this code to ENGINE is now a good
4286 deal more passive and at run-time, operations deal directly with
4287 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4288 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4289 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4290 they were not being used by the framework as there is no concept of a
4291 BIGNUM_METHOD and they could not be generalised to the new
4292 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4293 ENGINE_cpy() has been removed as it cannot be consistently defined in
4297 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4300 *) Change mkdef.pl to sort symbols that get the same entry number,
4301 and make sure the automatically generated functions ERR_load_*
4302 become part of libeay.num as well.
4305 *) New function SSL_renegotiate_pending(). This returns true once
4306 renegotiation has been requested (either SSL_renegotiate() call
4307 or HelloRequest/ClientHello receveived from the peer) and becomes
4308 false once a handshake has been completed.
4309 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4310 sends a HelloRequest, but does not ensure that a handshake takes
4311 place. SSL_renegotiate_pending() is useful for checking if the
4312 client has followed the request.)
4315 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4316 By default, clients may request session resumption even during
4317 renegotiation (if session ID contexts permit); with this option,
4318 session resumption is possible only in the first handshake.
4320 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4321 more bits available for options that should not be part of
4322 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4325 *) Add some demos for certificate and certificate request creation.
4328 *) Make maximum certificate chain size accepted from the peer application
4329 settable (SSL*_get/set_max_cert_list()), as proposed by
4330 "Douglas E. Engert" <deengert@anl.gov>.
4333 *) Add support for shared libraries for Unixware-7
4334 (Boyd Lynn Gerber <gerberb@zenez.com>).
4337 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4338 be done prior to destruction. Use this to unload error strings from
4339 ENGINEs that load their own error strings. NB: This adds two new API
4340 functions to "get" and "set" this destroy handler in an ENGINE.
4343 *) Alter all existing ENGINE implementations (except "openssl" and
4344 "openbsd") to dynamically instantiate their own error strings. This
4345 makes them more flexible to be built both as statically-linked ENGINEs
4346 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4347 Also, add stub code to each that makes building them as self-contained
4348 shared-libraries easier (see README.ENGINE).
4351 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4352 implementations into applications that are completely implemented in
4353 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4354 commands that can be used to configure what shared-library to load and
4355 to control aspects of the way it is handled. Also, made an update to
4356 the README.ENGINE file that brings its information up-to-date and
4357 provides some information and instructions on the "dynamic" ENGINE
4358 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4361 *) Make it possible to unload ranges of ERR strings with a new
4362 "ERR_unload_strings" function.
4365 *) Add a copy() function to EVP_MD.
4368 *) Make EVP_MD routines take a context pointer instead of just the
4369 md_data void pointer.
4372 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4373 that the digest can only process a single chunk of data
4374 (typically because it is provided by a piece of
4375 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4376 is only going to provide a single chunk of data, and hence the
4377 framework needn't accumulate the data for oneshot drivers.
4380 *) As with "ERR", make it possible to replace the underlying "ex_data"
4381 functions. This change also alters the storage and management of global
4382 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4383 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4384 index counters. The API functions that use this state have been changed
4385 to take a "class_index" rather than pointers to the class's local STACK
4386 and counter, and there is now an API function to dynamically create new
4387 classes. This centralisation allows us to (a) plug a lot of the
4388 thread-safety problems that existed, and (b) makes it possible to clean
4389 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4390 such data would previously have always leaked in application code and
4391 workarounds were in place to make the memory debugging turn a blind eye
4392 to it. Application code that doesn't use this new function will still
4393 leak as before, but their memory debugging output will announce it now
4394 rather than letting it slide.
4396 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4397 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4398 has a return value to indicate success or failure.
4401 *) Make it possible to replace the underlying "ERR" functions such that the
4402 global state (2 LHASH tables and 2 locks) is only used by the "default"
4403 implementation. This change also adds two functions to "get" and "set"
4404 the implementation prior to it being automatically set the first time
4405 any other ERR function takes place. Ie. an application can call "get",
4406 pass the return value to a module it has just loaded, and that module
4407 can call its own "set" function using that value. This means the
4408 module's "ERR" operations will use (and modify) the error state in the
4409 application and not in its own statically linked copy of OpenSSL code.
4412 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4413 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4414 the operation, and provides a more encapsulated way for external code
4415 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4416 to use these functions rather than manually incrementing the counts.
4418 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4421 *) Add EVP test program.
4424 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4427 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4428 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4429 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4430 These allow a CRL to be built without having to access X509_CRL fields
4431 directly. Modify 'ca' application to use new functions.
4434 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4435 bug workarounds. Rollback attack detection is a security feature.
4436 The problem will only arise on OpenSSL servers when TLSv1 is not
4437 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4438 Software authors not wanting to support TLSv1 will have special reasons
4439 for their choice and can explicitly enable this option.
4440 [Bodo Moeller, Lutz Jaenicke]
4442 *) Rationalise EVP so it can be extended: don't include a union of
4443 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4444 (similar to those existing for EVP_CIPHER_CTX).
4449 EVP_MD_CTX_init(&md); /* new function call */
4450 EVP_DigestInit(&md, EVP_sha1());
4451 EVP_DigestUpdate(&md, in, len);
4452 EVP_DigestFinal(&md, out, NULL);
4453 EVP_MD_CTX_cleanup(&md); /* new function call */
4457 *) Make DES key schedule conform to the usual scheme, as well as
4458 correcting its structure. This means that calls to DES functions
4459 now have to pass a pointer to a des_key_schedule instead of a
4460 plain des_key_schedule (which was actually always a pointer
4463 des_key_schedule ks;
4465 des_set_key_checked(..., &ks);
4466 des_ncbc_encrypt(..., &ks, ...);
4468 (Note that a later change renames 'des_...' into 'DES_...'.)
4471 *) Initial reduction of linker bloat: the use of some functions, such as
4472 PEM causes large amounts of unused functions to be linked in due to
4473 poor organisation. For example pem_all.c contains every PEM function
4474 which has a knock on effect of linking in large amounts of (unused)
4475 ASN1 code. Grouping together similar functions and splitting unrelated
4476 functions prevents this.
4479 *) Cleanup of EVP macros.
4482 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4483 correct _ecb suffix.
4486 *) Add initial OCSP responder support to ocsp application. The
4487 revocation information is handled using the text based index
4488 use by the ca application. The responder can either handle
4489 requests generated internally, supplied in files (for example
4490 via a CGI script) or using an internal minimal server.
4493 *) Add configuration choices to get zlib compression for TLS.
4496 *) Changes to Kerberos SSL for RFC 2712 compliance:
4497 1. Implemented real KerberosWrapper, instead of just using
4498 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4499 2. Implemented optional authenticator field of KerberosWrapper.
4501 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4502 and authenticator structs; see crypto/krb5/.
4504 Generalized Kerberos calls to support multiple Kerberos libraries.
4505 [Vern Staats <staatsvr@asc.hpc.mil>,
4506 Jeffrey Altman <jaltman@columbia.edu>
4507 via Richard Levitte]
4509 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4510 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4511 values for each of the key sizes rather than having just
4512 parameters (and 'speed' generating keys each time).
4515 *) Speed up EVP routines.
4518 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4519 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4520 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4521 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4523 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4524 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4525 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4528 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4530 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4533 *) Added the OS2-EMX target.
4534 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4536 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4537 to support NCONF routines in extension code. New function CONF_set_nconf()
4538 to allow functions which take an NCONF to also handle the old LHASH
4539 structure: this means that the old CONF compatible routines can be
4540 retained (in particular wrt extensions) without having to duplicate the
4541 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4544 *) Enhance the general user interface with mechanisms for inner control
4545 and with possibilities to have yes/no kind of prompts.
4548 *) Change all calls to low level digest routines in the library and
4549 applications to use EVP. Add missing calls to HMAC_cleanup() and
4550 don't assume HMAC_CTX can be copied using memcpy().
4551 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4553 *) Add the possibility to control engines through control names but with
4554 arbitrary arguments instead of just a string.
4555 Change the key loaders to take a UI_METHOD instead of a callback
4556 function pointer. NOTE: this breaks binary compatibility with earlier
4557 versions of OpenSSL [engine].
4558 Adapt the nCipher code for these new conditions and add a card insertion
4562 *) Enhance the general user interface with mechanisms to better support
4563 dialog box interfaces, application-defined prompts, the possibility
4564 to use defaults (for example default passwords from somewhere else)
4565 and interrupts/cancellations.
4568 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4569 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4572 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4573 tidy up some unnecessarily weird code in 'sk_new()').
4574 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4576 *) Change the key loading routines for ENGINEs to use the same kind
4577 callback (pem_password_cb) as all other routines that need this
4581 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4582 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4583 than this minimum value is recommended.
4586 *) New random seeder for OpenVMS, using the system process statistics
4587 that are easily reachable.
4590 *) Windows apparently can't transparently handle global
4591 variables defined in DLLs. Initialisations such as:
4593 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4595 wont compile. This is used by the any applications that need to
4596 declare their own ASN1 modules. This was fixed by adding the option
4597 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4598 needed for static libraries under Win32.
4601 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4602 setting of purpose and trust fields. New X509_STORE trust and
4603 purpose functions and tidy up setting in other SSL functions.
4606 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4607 structure. These are inherited by X509_STORE_CTX when it is
4608 initialised. This allows various defaults to be set in the
4609 X509_STORE structure (such as flags for CRL checking and custom
4610 purpose or trust settings) for functions which only use X509_STORE_CTX
4611 internally such as S/MIME.
4613 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4614 trust settings if they are not set in X509_STORE. This allows X509_STORE
4615 purposes and trust (in S/MIME for example) to override any set by default.
4617 Add command line options for CRL checking to smime, s_client and s_server
4621 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4622 are set then the CRL is looked up in the X509_STORE structure and
4623 its validity and signature checked, then if the certificate is found
4624 in the CRL the verify fails with a revoked error.
4626 Various new CRL related callbacks added to X509_STORE_CTX structure.
4628 Command line options added to 'verify' application to support this.
4630 This needs some additional work, such as being able to handle multiple
4631 CRLs with different times, extension based lookup (rather than just
4632 by subject name) and ultimately more complete V2 CRL extension
4636 *) Add a general user interface API (crypto/ui/). This is designed
4637 to replace things like des_read_password and friends (backward
4638 compatibility functions using this new API are provided).
4639 The purpose is to remove prompting functions from the DES code
4640 section as well as provide for prompting through dialog boxes in
4641 a window system and the like.
4644 *) Add "ex_data" support to ENGINE so implementations can add state at a
4645 per-structure level rather than having to store it globally.
4648 *) Make it possible for ENGINE structures to be copied when retrieved by
4649 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4650 This causes the "original" ENGINE structure to act like a template,
4651 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4652 operational state can be localised to each ENGINE structure, despite the
4653 fact they all share the same "methods". New ENGINE structures returned in
4654 this case have no functional references and the return value is the single
4655 structural reference. This matches the single structural reference returned
4656 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4660 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4661 needs to match any other type at all we need to manually clear the
4665 *) Changes to the "openssl engine" utility to include;
4666 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4667 about an ENGINE's available control commands.
4668 - executing control commands from command line arguments using the
4669 '-pre' and '-post' switches. '-post' is only used if '-t' is
4670 specified and the ENGINE is successfully initialised. The syntax for
4671 the individual commands are colon-separated, for example;
4672 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4675 *) New dynamic control command support for ENGINEs. ENGINEs can now
4676 declare their own commands (numbers), names (strings), descriptions,
4677 and input types for run-time discovery by calling applications. A
4678 subset of these commands are implicitly classed as "executable"
4679 depending on their input type, and only these can be invoked through
4680 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4681 can be based on user input, config files, etc). The distinction is
4682 that "executable" commands cannot return anything other than a boolean
4683 result and can only support numeric or string input, whereas some
4684 discoverable commands may only be for direct use through
4685 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4686 pointers, or other custom uses. The "executable" commands are to
4687 support parameterisations of ENGINE behaviour that can be
4688 unambiguously defined by ENGINEs and used consistently across any
4689 OpenSSL-based application. Commands have been added to all the
4690 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4691 control over shared-library paths without source code alterations.
4694 *) Changed all ENGINE implementations to dynamically allocate their
4695 ENGINEs rather than declaring them statically. Apart from this being
4696 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4697 this also allows the implementations to compile without using the
4698 internal engine_int.h header.
4701 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4702 'const' value. Any code that should be able to modify a RAND_METHOD
4703 should already have non-const pointers to it (ie. they should only
4704 modify their own ones).
4707 *) Made a variety of little tweaks to the ENGINE code.
4708 - "atalla" and "ubsec" string definitions were moved from header files
4709 to C code. "nuron" string definitions were placed in variables
4710 rather than hard-coded - allowing parameterisation of these values
4711 later on via ctrl() commands.
4712 - Removed unused "#if 0"'d code.
4713 - Fixed engine list iteration code so it uses ENGINE_free() to release
4714 structural references.
4715 - Constified the RAND_METHOD element of ENGINE structures.
4716 - Constified various get/set functions as appropriate and added
4717 missing functions (including a catch-all ENGINE_cpy that duplicates
4718 all ENGINE values onto a new ENGINE except reference counts/state).
4719 - Removed NULL parameter checks in get/set functions. Setting a method
4720 or function to NULL is a way of cancelling out a previously set
4721 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4722 and doesn't justify the extra error symbols and code.
4723 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4724 flags from engine_int.h to engine.h.
4725 - Changed prototypes for ENGINE handler functions (init(), finish(),
4726 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4729 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4730 to the algorithm using long division. The binary algorithm can be
4731 used only if the modulus is odd. On 32-bit systems, it is faster
4732 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4733 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4734 up to 450 bits. In 64-bit environments, the binary algorithm
4735 appears to be advantageous for much longer moduli; here we use it
4736 for moduli up to 2048 bits.
4739 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4740 could not support the combine flag in choice fields.
4743 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4744 extensions from a certificate request to the certificate.
4747 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4748 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4749 file: this allows the display of the certificate about to be
4750 signed to be customised, to allow certain fields to be included
4751 or excluded and extension details. The old system didn't display
4752 multicharacter strings properly, omitted fields not in the policy
4753 and couldn't display additional details such as extensions.
4756 *) Function EC_POINTs_mul for multiple scalar multiplication
4757 of an arbitrary number of elliptic curve points
4758 \sum scalars[i]*points[i],
4759 optionally including the generator defined for the EC_GROUP:
4760 scalar*generator + \sum scalars[i]*points[i].
4762 EC_POINT_mul is a simple wrapper function for the typical case
4763 that the point list has just one item (besides the optional
4767 *) First EC_METHODs for curves over GF(p):
4769 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4770 operations and provides various method functions that can also
4771 operate with faster implementations of modular arithmetic.
4773 EC_GFp_mont_method() reuses most functions that are part of
4774 EC_GFp_simple_method, but uses Montgomery arithmetic.
4776 [Bodo Moeller; point addition and point doubling
4777 implementation directly derived from source code provided by
4778 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4780 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4781 crypto/ec/ec_lib.c):
4783 Curves are EC_GROUP objects (with an optional group generator)
4784 based on EC_METHODs that are built into the library.
4786 Points are EC_POINT objects based on EC_GROUP objects.
4788 Most of the framework would be able to handle curves over arbitrary
4789 finite fields, but as there are no obvious types for fields other
4790 than GF(p), some functions are limited to that for now.
4793 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4794 that the file contains a complete HTTP response.
4797 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4798 change the def and num file printf format specifier from "%-40sXXX"
4799 to "%-39s XXX". The latter will always guarantee a space after the
4800 field while the former will cause them to run together if the field
4801 is 40 of more characters long.
4804 *) Constify the cipher and digest 'method' functions and structures
4805 and modify related functions to take constant EVP_MD and EVP_CIPHER
4809 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4810 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4813 *) Modify EVP_Digest*() routines so they now return values. Although the
4814 internal software routines can never fail additional hardware versions
4818 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4820 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4821 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4824 ERR_R_NESTED_ASN1_ERROR
4826 ERR_R_MISSING_ASN1_EOS
4827 were 4 .. 9, conflicting with
4828 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4830 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4831 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4833 Add new error code 'ERR_R_INTERNAL_ERROR'.
4836 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4840 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4841 sets the subject name for a new request or supersedes the
4842 subject name in a given request. Formats that can be parsed are
4843 'CN=Some Name, OU=myOU, C=IT'
4845 'CN=Some Name/OU=myOU/C=IT'.
4847 Add options '-batch' and '-verbose' to 'openssl req'.
4848 [Massimiliano Pala <madwolf@hackmasters.net>]
4850 *) Introduce the possibility to access global variables through
4851 functions on platform were that's the best way to handle exporting
4852 global variables in shared libraries. To enable this functionality,
4853 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4854 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4855 is normally done by Configure or something similar).
4857 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4858 in the source file (foo.c) like this:
4860 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4861 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4863 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4864 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4866 OPENSSL_DECLARE_GLOBAL(int,foo);
4867 #define foo OPENSSL_GLOBAL_REF(foo)
4868 OPENSSL_DECLARE_GLOBAL(double,bar);
4869 #define bar OPENSSL_GLOBAL_REF(bar)
4871 The #defines are very important, and therefore so is including the
4872 header file everywhere where the defined globals are used.
4874 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4875 of ASN.1 items, but that structure is a bit different.
4877 The largest change is in util/mkdef.pl which has been enhanced with
4878 better and easier to understand logic to choose which symbols should
4879 go into the Windows .def files as well as a number of fixes and code
4880 cleanup (among others, algorithm keywords are now sorted
4881 lexicographically to avoid constant rewrites).
4884 *) In BN_div() keep a copy of the sign of 'num' before writing the
4885 result to 'rm' because if rm==num the value will be overwritten
4886 and produce the wrong result if 'num' is negative: this caused
4887 problems with BN_mod() and BN_nnmod().
4890 *) Function OCSP_request_verify(). This checks the signature on an
4891 OCSP request and verifies the signer certificate. The signer
4892 certificate is just checked for a generic purpose and OCSP request
4896 *) Add OCSP_check_validity() function to check the validity of OCSP
4897 responses. OCSP responses are prepared in real time and may only
4898 be a few seconds old. Simply checking that the current time lies
4899 between thisUpdate and nextUpdate max reject otherwise valid responses
4900 caused by either OCSP responder or client clock inaccuracy. Instead
4901 we allow thisUpdate and nextUpdate to fall within a certain period of
4902 the current time. The age of the response can also optionally be
4903 checked. Two new options -validity_period and -status_age added to
4907 *) If signature or public key algorithm is unrecognized print out its
4908 OID rather that just UNKNOWN.
4911 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4912 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4913 ID to be generated from the issuer certificate alone which can then be
4914 passed to OCSP_id_issuer_cmp().
4917 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4918 ASN1 modules to export functions returning ASN1_ITEM pointers
4919 instead of the ASN1_ITEM structures themselves. This adds several
4920 new macros which allow the underlying ASN1 function/structure to
4921 be accessed transparently. As a result code should not use ASN1_ITEM
4922 references directly (such as &X509_it) but instead use the relevant
4923 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4924 use of the new ASN1 code on platforms where exporting structures
4925 is problematical (for example in shared libraries) but exporting
4926 functions returning pointers to structures is not.
4929 *) Add support for overriding the generation of SSL/TLS session IDs.
4930 These callbacks can be registered either in an SSL_CTX or per SSL.
4931 The purpose of this is to allow applications to control, if they wish,
4932 the arbitrary values chosen for use as session IDs, particularly as it
4933 can be useful for session caching in multiple-server environments. A
4934 command-line switch for testing this (and any client code that wishes
4935 to use such a feature) has been added to "s_server".
4936 [Geoff Thorpe, Lutz Jaenicke]
4938 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4939 of the form '#if defined(...) || defined(...) || ...' and
4940 '#if !defined(...) && !defined(...) && ...'. This also avoids
4941 the growing number of special cases it was previously handling.
4944 *) Make all configuration macros available for application by making
4945 sure they are available in opensslconf.h, by giving them names starting
4946 with "OPENSSL_" to avoid conflicts with other packages and by making
4947 sure e_os2.h will cover all platform-specific cases together with
4949 Additionally, it is now possible to define configuration/platform-
4950 specific names (called "system identities"). In the C code, these
4951 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4952 macro with the name beginning with "OPENSSL_SYS_", which is determined
4953 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4957 *) New option -set_serial to 'req' and 'x509' this allows the serial
4958 number to use to be specified on the command line. Previously self
4959 signed certificates were hard coded with serial number 0 and the
4960 CA options of 'x509' had to use a serial number in a file which was
4964 *) New options to 'ca' utility to support V2 CRL entry extensions.
4965 Currently CRL reason, invalidity date and hold instruction are
4966 supported. Add new CRL extensions to V3 code and some new objects.
4969 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4970 disable standard block padding (aka PKCS#5 padding) in the EVP
4971 API, which was previously mandatory. This means that the data is
4972 not padded in any way and so the total length much be a multiple
4973 of the block size, otherwise an error occurs.
4976 *) Initial (incomplete) OCSP SSL support.
4979 *) New function OCSP_parse_url(). This splits up a URL into its host,
4980 port and path components: primarily to parse OCSP URLs. New -url
4981 option to ocsp utility.
4984 *) New nonce behavior. The return value of OCSP_check_nonce() now
4985 reflects the various checks performed. Applications can decide
4986 whether to tolerate certain situations such as an absent nonce
4987 in a response when one was present in a request: the ocsp application
4988 just prints out a warning. New function OCSP_add1_basic_nonce()
4989 this is to allow responders to include a nonce in a response even if
4990 the request is nonce-less.
4993 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4994 skipped when using openssl x509 multiple times on a single input file,
4995 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4998 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4999 set string type: to handle setting ASN1_TIME structures. Fix ca
5000 utility to correctly initialize revocation date of CRLs.
5003 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5004 the clients preferred ciphersuites and rather use its own preferences.
5005 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5006 Internet Explorer by ensuring unchanged hash method during stepup.
5007 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5010 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5011 to aes and add a new 'exist' option to print out symbols that don't
5015 *) Additional options to ocsp utility to allow flags to be set and
5016 additional certificates supplied.
5019 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5020 OCSP client a number of certificate to only verify the response
5024 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5025 handle the new API. Currently only ECB, CBC modes supported. Add new
5028 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5029 Encryption Standard (AES) Ciphersuites for Transport Layer
5030 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5031 not enabled by default and were not part of the "ALL" ciphersuite
5032 alias because they were not yet official; they could be
5033 explicitly requested by specifying the "AESdraft" ciphersuite
5034 group alias. In the final release of OpenSSL 0.9.7, the group
5035 alias is called "AES" and is part of "ALL".)
5036 [Ben Laurie, Steve Henson, Bodo Moeller]
5038 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5039 request to response.
5042 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5043 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5044 extract information from a certificate request. OCSP_response_create()
5045 creates a response and optionally adds a basic response structure.
5046 OCSP_basic_add1_status() adds a complete single response to a basic
5047 response and returns the OCSP_SINGLERESP structure just added (to allow
5048 extensions to be included for example). OCSP_basic_add1_cert() adds a
5049 certificate to a basic response and OCSP_basic_sign() signs a basic
5050 response with various flags. New helper functions ASN1_TIME_check()
5051 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5052 (converts ASN1_TIME to GeneralizedTime).
5055 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5056 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5057 structure from a certificate. X509_pubkey_digest() digests the public_key
5058 contents: this is used in various key identifiers.
5061 *) Make sk_sort() tolerate a NULL argument.
5062 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5064 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5065 passed by the function are trusted implicitly. If any of them signed the
5066 response then it is assumed to be valid and is not verified.
5069 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5070 to data. This was previously part of the PKCS7 ASN1 code. This
5071 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5072 [Steve Henson, reported by Kenneth R. Robinette
5073 <support@securenetterm.com>]
5075 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5076 routines: without these tracing memory leaks is very painful.
5077 Fix leaks in PKCS12 and PKCS7 routines.
5080 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5081 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5082 effectively meant GeneralizedTime would never be used. Now it
5083 is initialised to -1 but X509_time_adj() now has to check the value
5084 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5085 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5086 [Steve Henson, reported by Kenneth R. Robinette
5087 <support@securenetterm.com>]
5089 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5090 result in a zero length in the ASN1_INTEGER structure which was
5091 not consistent with the structure when d2i_ASN1_INTEGER() was used
5092 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5093 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5094 where it did not print out a minus for negative ASN1_INTEGER.
5097 *) Add summary printout to ocsp utility. The various functions which
5098 convert status values to strings have been renamed to:
5099 OCSP_response_status_str(), OCSP_cert_status_str() and
5100 OCSP_crl_reason_str() and are no longer static. New options
5101 to verify nonce values and to disable verification. OCSP response
5102 printout format cleaned up.
5105 *) Add additional OCSP certificate checks. These are those specified
5106 in RFC2560. This consists of two separate checks: the CA of the
5107 certificate being checked must either be the OCSP signer certificate
5108 or the issuer of the OCSP signer certificate. In the latter case the
5109 OCSP signer certificate must contain the OCSP signing extended key
5110 usage. This check is performed by attempting to match the OCSP
5111 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5112 in the OCSP_CERTID structures of the response.
5115 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5116 and related routines. This uses the standard OpenSSL certificate
5117 verify routines to perform initial checks (just CA validity) and
5118 to obtain the certificate chain. Then additional checks will be
5119 performed on the chain. Currently the root CA is checked to see
5120 if it is explicitly trusted for OCSP signing. This is used to set
5121 a root CA as a global signing root: that is any certificate that
5122 chains to that CA is an acceptable OCSP signing certificate.
5125 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5126 extensions from a separate configuration file.
5127 As when reading extensions from the main configuration file,
5128 the '-extensions ...' option may be used for specifying the
5130 [Massimiliano Pala <madwolf@comune.modena.it>]
5132 *) New OCSP utility. Allows OCSP requests to be generated or
5133 read. The request can be sent to a responder and the output
5134 parsed, outputed or printed in text form. Not complete yet:
5135 still needs to check the OCSP response validity.
5138 *) New subcommands for 'openssl ca':
5139 'openssl ca -status <serial>' prints the status of the cert with
5140 the given serial number (according to the index file).
5141 'openssl ca -updatedb' updates the expiry status of certificates
5143 [Massimiliano Pala <madwolf@comune.modena.it>]
5145 *) New '-newreq-nodes' command option to CA.pl. This is like
5146 '-newreq', but calls 'openssl req' with the '-nodes' option
5147 so that the resulting key is not encrypted.
5148 [Damien Miller <djm@mindrot.org>]
5150 *) New configuration for the GNU Hurd.
5151 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5153 *) Initial code to implement OCSP basic response verify. This
5154 is currently incomplete. Currently just finds the signer's
5155 certificate and verifies the signature on the response.
5158 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5159 value of OPENSSLDIR. This is available via the new '-d' option
5160 to 'openssl version', and is also included in 'openssl version -a'.
5163 *) Allowing defining memory allocation callbacks that will be given
5164 file name and line number information in additional arguments
5165 (a const char* and an int). The basic functionality remains, as
5166 well as the original possibility to just replace malloc(),
5167 realloc() and free() by functions that do not know about these
5168 additional arguments. To register and find out the current
5169 settings for extended allocation functions, the following
5170 functions are provided:
5172 CRYPTO_set_mem_ex_functions
5173 CRYPTO_set_locked_mem_ex_functions
5174 CRYPTO_get_mem_ex_functions
5175 CRYPTO_get_locked_mem_ex_functions
5177 These work the same way as CRYPTO_set_mem_functions and friends.
5178 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5179 extended allocation function is enabled.
5180 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5181 a conventional allocation function is enabled.
5182 [Richard Levitte, Bodo Moeller]
5184 *) Finish off removing the remaining LHASH function pointer casts.
5185 There should no longer be any prototype-casting required when using
5186 the LHASH abstraction, and any casts that remain are "bugs". See
5187 the callback types and macros at the head of lhash.h for details
5188 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5191 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5192 If /dev/[u]random devices are not available or do not return enough
5193 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5195 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5196 /etc/entropy will be queried once each in this sequence, quering stops
5197 when enough entropy was collected without querying more sockets.
5200 *) Change the Unix RAND_poll() variant to be able to poll several
5201 random devices, as specified by DEVRANDOM, until a sufficient amount
5202 of data has been collected. We spend at most 10 ms on each file
5203 (select timeout) and read in non-blocking mode. DEVRANDOM now
5204 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5205 (previously it was just the string "/dev/urandom"), so on typical
5206 platforms the 10 ms delay will never occur.
5207 Also separate out the Unix variant to its own file, rand_unix.c.
5208 For VMS, there's a currently-empty rand_vms.c.
5211 *) Move OCSP client related routines to ocsp_cl.c. These
5212 provide utility functions which an application needing
5213 to issue a request to an OCSP responder and analyse the
5214 response will typically need: as opposed to those which an
5215 OCSP responder itself would need which will be added later.
5217 OCSP_request_sign() signs an OCSP request with an API similar
5218 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5219 response. OCSP_response_get1_basic() extracts basic response
5220 from response. OCSP_resp_find_status(): finds and extracts status
5221 information from an OCSP_CERTID structure (which will be created
5222 when the request structure is built). These are built from lower
5223 level functions which work on OCSP_SINGLERESP structures but
5224 wont normally be used unless the application wishes to examine
5225 extensions in the OCSP response for example.
5227 Replace nonce routines with a pair of functions.
5228 OCSP_request_add1_nonce() adds a nonce value and optionally
5229 generates a random value. OCSP_check_nonce() checks the
5230 validity of the nonce in an OCSP response.
5233 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5234 This doesn't copy the supplied OCSP_CERTID and avoids the
5235 need to free up the newly created id. Change return type
5236 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5237 This can then be used to add extensions to the request.
5238 Deleted OCSP_request_new(), since most of its functionality
5239 is now in OCSP_REQUEST_new() (and the case insensitive name
5240 clash) apart from the ability to set the request name which
5241 will be added elsewhere.
5244 *) Update OCSP API. Remove obsolete extensions argument from
5245 various functions. Extensions are now handled using the new
5246 OCSP extension code. New simple OCSP HTTP function which
5247 can be used to send requests and parse the response.
5250 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5251 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5252 uses the special reorder version of SET OF to sort the attributes
5253 and reorder them to match the encoded order. This resolves a long
5254 standing problem: a verify on a PKCS7 structure just after signing
5255 it used to fail because the attribute order did not match the
5256 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5257 it uses the received order. This is necessary to tolerate some broken
5258 software that does not order SET OF. This is handled by encoding
5259 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5260 to produce the required SET OF.
5263 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5264 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5265 files to get correct declarations of the ASN.1 item variables.
5268 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5269 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5270 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5271 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5272 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5273 ASN1_ITEM and no wrapper functions.
5276 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5277 replace the old function pointer based I/O routines. Change most of
5278 the *_d2i_bio() and *_d2i_fp() functions to use these.
5281 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5282 lines, recognice more "algorithms" that can be deselected, and make
5283 it complain about algorithm deselection that isn't recognised.
5286 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5287 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5288 to use new functions. Add NO_ASN1_OLD which can be set to remove
5289 some old style ASN1 functions: this can be used to determine if old
5290 code will still work when these eventually go away.
5293 *) New extension functions for OCSP structures, these follow the
5294 same conventions as certificates and CRLs.
5297 *) New function X509V3_add1_i2d(). This automatically encodes and
5298 adds an extension. Its behaviour can be customised with various
5299 flags to append, replace or delete. Various wrappers added for
5300 certifcates and CRLs.
5303 *) Fix to avoid calling the underlying ASN1 print routine when
5304 an extension cannot be parsed. Correct a typo in the
5305 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5308 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5309 entries for variables.
5312 *) Add functionality to apps/openssl.c for detecting locking
5313 problems: As the program is single-threaded, all we have
5314 to do is register a locking callback using an array for
5315 storing which locks are currently held by the program.
5318 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5319 SSL_get_ex_data_X509_STORE_idx(), which is used in
5320 ssl_verify_cert_chain() and thus can be called at any time
5321 during TLS/SSL handshakes so that thread-safety is essential.
5322 Unfortunately, the ex_data design is not at all suited
5323 for multi-threaded use, so it probably should be abolished.
5326 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5327 [Broadcom, tweaked and integrated by Geoff Thorpe]
5329 *) Move common extension printing code to new function
5330 X509V3_print_extensions(). Reorganise OCSP print routines and
5331 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5334 *) New function X509_signature_print() to remove duplication in some
5338 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5339 set (this was treated exactly the same as SET OF previously). This
5340 is used to reorder the STACK representing the structure to match the
5341 encoding. This will be used to get round a problem where a PKCS7
5342 structure which was signed could not be verified because the STACK
5343 order did not reflect the encoded order.
5346 *) Reimplement the OCSP ASN1 module using the new code.
5349 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5350 for its ASN1 operations. The old style function pointers still exist
5351 for now but they will eventually go away.
5354 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5355 completely replaces the old ASN1 functionality with a table driven
5356 encoder and decoder which interprets an ASN1_ITEM structure describing
5357 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5358 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5359 has also been converted to the new form.
5362 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5363 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5364 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5365 for negative moduli.
5368 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5369 of not touching the result's sign bit.
5372 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5376 *) Changed the LHASH code to use prototypes for callbacks, and created
5377 macros to declare and implement thin (optionally static) functions
5378 that provide type-safety and avoid function pointer casting for the
5379 type-specific callbacks.
5382 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5384 [Veers Staats <staatsvr@asc.hpc.mil>,
5385 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5387 *) Reformat the FAQ so the different questions and answers can be divided
5388 in sections depending on the subject.
5391 *) Have the zlib compression code load ZLIB.DLL dynamically under
5395 *) New function BN_mod_sqrt for computing square roots modulo a prime
5396 (using the probabilistic Tonelli-Shanks algorithm unless
5397 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5398 be handled deterministically).
5399 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5401 *) Make BN_mod_inverse faster by explicitly handling small quotients
5402 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5403 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5406 *) New function BN_kronecker.
5409 *) Fix BN_gcd so that it works on negative inputs; the result is
5410 positive unless both parameters are zero.
5411 Previously something reasonably close to an infinite loop was
5412 possible because numbers could be growing instead of shrinking
5413 in the implementation of Euclid's algorithm.
5416 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5417 sign of the number in question.
5419 Fix BN_is_word(a,w) to work correctly for w == 0.
5421 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5422 because its test if the absolute value of 'a' equals 'w'.
5423 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5424 it exists mostly for use in the implementations of BN_is_zero(),
5425 BN_is_one(), and BN_is_word().
5428 *) New function BN_swap.
5431 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5432 the exponentiation functions are more likely to produce reasonable
5433 results on negative inputs.
5436 *) Change BN_mod_mul so that the result is always non-negative.
5437 Previously, it could be negative if one of the factors was negative;
5438 I don't think anyone really wanted that behaviour.
5441 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5442 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5443 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5444 and add new functions:
5453 BN_mod_lshift1_quick
5457 These functions always generate non-negative results.
5459 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5460 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5462 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5463 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5464 be reduced modulo m.
5465 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5468 The following entry accidentily appeared in the CHANGES file
5469 distributed with OpenSSL 0.9.7. The modifications described in
5470 it do *not* apply to OpenSSL 0.9.7.
5472 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5473 was actually never needed) and in BN_mul(). The removal in BN_mul()
5474 required a small change in bn_mul_part_recursive() and the addition
5475 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5476 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5477 bn_sub_words() and bn_add_words() except they take arrays with
5482 *) In 'openssl passwd', verify passwords read from the terminal
5483 unless the '-salt' option is used (which usually means that
5484 verification would just waste user's time since the resulting
5485 hash is going to be compared with some given password hash)
5486 or the new '-noverify' option is used.
5488 This is an incompatible change, but it does not affect
5489 non-interactive use of 'openssl passwd' (passwords on the command
5490 line, '-stdin' option, '-in ...' option) and thus should not
5494 *) Remove all references to RSAref, since there's no more need for it.
5497 *) Make DSO load along a path given through an environment variable
5498 (SHLIB_PATH) with shl_load().
5501 *) Constify the ENGINE code as a result of BIGNUM constification.
5502 Also constify the RSA code and most things related to it. In a
5503 few places, most notable in the depth of the ASN.1 code, ugly
5504 casts back to non-const were required (to be solved at a later
5508 *) Make it so the openssl application has all engines loaded by default.
5511 *) Constify the BIGNUM routines a little more.
5514 *) Add the following functions:
5516 ENGINE_load_cswift()
5518 ENGINE_load_atalla()
5520 ENGINE_load_builtin_engines()
5522 That way, an application can itself choose if external engines that
5523 are built-in in OpenSSL shall ever be used or not. The benefit is
5524 that applications won't have to be linked with libdl or other dso
5525 libraries unless it's really needed.
5527 Changed 'openssl engine' to load all engines on demand.
5528 Changed the engine header files to avoid the duplication of some
5529 declarations (they differed!).
5532 *) 'openssl engine' can now list capabilities.
5535 *) Better error reporting in 'openssl engine'.
5538 *) Never call load_dh_param(NULL) in s_server.
5541 *) Add engine application. It can currently list engines by name and
5542 identity, and test if they are actually available.
5545 *) Improve RPM specification file by forcing symbolic linking and making
5546 sure the installed documentation is also owned by root.root.
5547 [Damien Miller <djm@mindrot.org>]
5549 *) Give the OpenSSL applications more possibilities to make use of
5550 keys (public as well as private) handled by engines.
5553 *) Add OCSP code that comes from CertCo.
5556 *) Add VMS support for the Rijndael code.
5559 *) Added untested support for Nuron crypto accelerator.
5562 *) Add support for external cryptographic devices. This code was
5563 previously distributed separately as the "engine" branch.
5564 [Geoff Thorpe, Richard Levitte]
5566 *) Rework the filename-translation in the DSO code. It is now possible to
5567 have far greater control over how a "name" is turned into a filename
5568 depending on the operating environment and any oddities about the
5569 different shared library filenames on each system.
5572 *) Support threads on FreeBSD-elf in Configure.
5575 *) Fix for SHA1 assembly problem with MASM: it produces
5576 warnings about corrupt line number information when assembling
5577 with debugging information. This is caused by the overlapping
5579 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5582 NCONF_get_number() has no error checking at all. As a replacement,
5583 NCONF_get_number_e() is defined (_e for "error checking") and is
5584 promoted strongly. The old NCONF_get_number is kept around for
5585 binary backward compatibility.
5586 Make it possible for methods to load from something other than a BIO,
5587 by providing a function pointer that is given a name instead of a BIO.
5588 For example, this could be used to load configuration data from an
5592 *) Fix for non blocking accept BIOs. Added new I/O special reason
5593 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5594 with non blocking I/O was not possible because no retry code was
5595 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5599 *) Added the beginnings of Rijndael support.
5602 *) Fix for bug in DirectoryString mask setting. Add support for
5603 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5604 to allow certificate printing to more controllable, additional
5605 'certopt' option to 'x509' to allow new printing options to be
5609 *) Clean old EAY MD5 hack from e_os.h.
5612 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5614 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5615 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5616 [Joe Orton, Steve Henson]
5618 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5620 *) Fix additional bug revealed by the NISCC test suite:
5622 Stop bug triggering large recursion when presented with
5623 certain ASN.1 tags (CVE-2003-0851)
5626 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5628 *) Fix various bugs revealed by running the NISCC test suite:
5630 Stop out of bounds reads in the ASN1 code when presented with
5631 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5633 If verify callback ignores invalid public key errors don't try to check
5634 certificate signature with the NULL public key.
5638 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5639 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5643 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5644 extra data after the compression methods not only for TLS 1.0
5645 but also for SSL 3.0 (as required by the specification).
5646 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5648 *) Change X509_certificate_type() to mark the key as exported/exportable
5649 when it's 512 *bits* long, not 512 bytes.
5652 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5654 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5655 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5656 a protocol version number mismatch like a decryption error
5657 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5660 *) Turn on RSA blinding by default in the default implementation
5661 to avoid a timing attack. Applications that don't want it can call
5662 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5663 They would be ill-advised to do so in most cases.
5664 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5666 *) Change RSA blinding code so that it works when the PRNG is not
5667 seeded (in this case, the secret RSA exponent is abused as
5668 an unpredictable seed -- if it is not unpredictable, there
5669 is no point in blinding anyway). Make RSA blinding thread-safe
5670 by remembering the creator's thread ID in rsa->blinding and
5671 having all other threads use local one-time blinding factors
5672 (this requires more computation than sharing rsa->blinding, but
5673 avoids excessive locking; and if an RSA object is not shared
5674 between threads, blinding will still be very fast).
5677 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5679 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5680 via timing by performing a MAC computation even if incorrrect
5681 block cipher padding has been found. This is a countermeasure
5682 against active attacks where the attacker has to distinguish
5683 between bad padding and a MAC verification error. (CVE-2003-0078)
5685 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5686 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5687 Martin Vuagnoux (EPFL, Ilion)]
5689 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5691 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5692 memory from it's contents. This is done with a counter that will
5693 place alternating values in each byte. This can be used to solve
5694 two issues: 1) the removal of calls to memset() by highly optimizing
5695 compilers, and 2) cleansing with other values than 0, since those can
5696 be read through on certain media, for example a swap space on disk.
5699 *) Bugfix: client side session caching did not work with external caching,
5700 because the session->cipher setting was not restored when reloading
5701 from the external cache. This problem was masked, when
5702 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5703 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5706 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5707 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5708 [Zeev Lieber <zeev-l@yahoo.com>]
5710 *) Undo an undocumented change introduced in 0.9.6e which caused
5711 repeated calls to OpenSSL_add_all_ciphers() and
5712 OpenSSL_add_all_digests() to be ignored, even after calling
5716 *) Change the default configuration reader to deal with last line not
5717 being properly terminated.
5720 *) Change X509_NAME_cmp() so it applies the special rules on handling
5721 DN values that are of type PrintableString, as well as RDNs of type
5722 emailAddress where the value has the type ia5String.
5723 [stefank@valicert.com via Richard Levitte]
5725 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5726 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5727 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5728 the bitwise-OR of the two for use by the majority of applications
5729 wanting this behaviour, and update the docs. The documented
5730 behaviour and actual behaviour were inconsistent and had been
5731 changing anyway, so this is more a bug-fix than a behavioural
5733 [Geoff Thorpe, diagnosed by Nadav Har'El]
5735 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5736 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5739 *) Fix initialization code race conditions in
5740 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5741 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5742 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5743 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5744 ssl2_get_cipher_by_char(),
5745 ssl3_get_cipher_by_char().
5746 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5748 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5749 the cached sessions are flushed, as the remove_cb() might use ex_data
5750 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5751 (see [openssl.org #212]).
5752 [Geoff Thorpe, Lutz Jaenicke]
5754 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5755 length, instead of the encoding length to d2i_ASN1_OBJECT.
5758 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5760 *) [In 0.9.6g-engine release:]
5761 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5762 [Lynn Gazis <lgazis@rainbow.com>]
5764 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5766 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5767 and get fix the header length calculation.
5768 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5769 Alon Kantor <alonk@checkpoint.com> (and others),
5772 *) Use proper error handling instead of 'assertions' in buffer
5773 overflow checks added in 0.9.6e. This prevents DoS (the
5774 assertions could call abort()).
5775 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5777 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5779 *) Add various sanity checks to asn1_get_length() to reject
5780 the ASN1 length bytes if they exceed sizeof(long), will appear
5781 negative or the content length exceeds the length of the
5783 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5785 *) Fix cipher selection routines: ciphers without encryption had no flags
5786 for the cipher strength set and where therefore not handled correctly
5787 by the selection routines (PR #130).
5790 *) Fix EVP_dsa_sha macro.
5794 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5795 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5796 that was added in OpenSSL 0.9.6d.
5798 As the countermeasure turned out to be incompatible with some
5799 broken SSL implementations, the new option is part of SSL_OP_ALL.
5800 SSL_OP_ALL is usually employed when compatibility with weird SSL
5801 implementations is desired (e.g. '-bugs' option to 's_client' and
5802 's_server'), so the new option is automatically set in many
5806 *) Changes in security patch:
5808 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5809 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5810 Air Force Materiel Command, USAF, under agreement number
5813 *) Add various sanity checks to asn1_get_length() to reject
5814 the ASN1 length bytes if they exceed sizeof(long), will appear
5815 negative or the content length exceeds the length of the
5816 supplied buffer. (CVE-2002-0659)
5817 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5819 *) Assertions for various potential buffer overflows, not known to
5821 [Ben Laurie (CHATS)]
5823 *) Various temporary buffers to hold ASCII versions of integers were
5824 too small for 64 bit platforms. (CVE-2002-0655)
5825 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5827 *) Remote buffer overflow in SSL3 protocol - an attacker could
5828 supply an oversized session ID to a client. (CVE-2002-0656)
5829 [Ben Laurie (CHATS)]
5831 *) Remote buffer overflow in SSL2 protocol - an attacker could
5832 supply an oversized client master key. (CVE-2002-0656)
5833 [Ben Laurie (CHATS)]
5835 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5837 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5838 encoded as NULL) with id-dsa-with-sha1.
5839 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5841 *) Check various X509_...() return values in apps/req.c.
5842 [Nils Larsch <nla@trustcenter.de>]
5844 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5845 an end-of-file condition would erronously be flagged, when the CRLF
5846 was just at the end of a processed block. The bug was discovered when
5847 processing data through a buffering memory BIO handing the data to a
5848 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5849 <ptsekov@syntrex.com> and Nedelcho Stanev.
5852 *) Implement a countermeasure against a vulnerability recently found
5853 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5854 before application data chunks to avoid the use of known IVs
5855 with data potentially chosen by the attacker.
5858 *) Fix length checks in ssl3_get_client_hello().
5861 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5862 to prevent ssl3_read_internal() from incorrectly assuming that
5863 ssl3_read_bytes() found application data while handshake
5864 processing was enabled when in fact s->s3->in_read_app_data was
5865 merely automatically cleared during the initial handshake.
5866 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5868 *) Fix object definitions for Private and Enterprise: they were not
5869 recognized in their shortname (=lowercase) representation. Extend
5870 obj_dat.pl to issue an error when using undefined keywords instead
5871 of silently ignoring the problem (Svenning Sorensen
5872 <sss@sss.dnsalias.net>).
5875 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5876 generators, i.e. generators other than 2 and 5. (Previously, the
5877 code did not properly initialise the 'add' and 'rem' values to
5878 BN_generate_prime().)
5880 In the new general case, we do not insist that 'generator' is
5881 actually a primitive root: This requirement is rather pointless;
5882 a generator of the order-q subgroup is just as good, if not
5886 *) Map new X509 verification errors to alerts. Discovered and submitted by
5887 Tom Wu <tom@arcot.com>.
5890 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5891 returning non-zero before the data has been completely received
5892 when using non-blocking I/O.
5893 [Bodo Moeller; problem pointed out by John Hughes]
5895 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5896 [Ben Laurie, Lutz Jaenicke]
5898 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5899 Yoram Zahavi <YoramZ@gilian.com>).
5902 *) Add information about CygWin 1.3 and on, and preserve proper
5903 configuration for the versions before that.
5904 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5906 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5907 check whether we deal with a copy of a session and do not delete from
5908 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5909 <izhar@checkpoint.com>.
5912 *) Do not store session data into the internal session cache, if it
5913 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5914 flag is set). Proposed by Aslam <aslam@funk.com>.
5917 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5921 *) [In 0.9.6d-engine release:]
5922 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5923 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5925 *) Add the configuration target linux-s390x.
5926 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5928 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5929 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5930 variable as an indication that a ClientHello message has been
5931 received. As the flag value will be lost between multiple
5932 invocations of ssl3_accept when using non-blocking I/O, the
5933 function may not be aware that a handshake has actually taken
5934 place, thus preventing a new session from being added to the
5937 To avoid this problem, we now set s->new_session to 2 instead of
5938 using a local variable.
5939 [Lutz Jaenicke, Bodo Moeller]
5941 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5942 if the SSL_R_LENGTH_MISMATCH error is detected.
5943 [Geoff Thorpe, Bodo Moeller]
5945 *) New 'shared_ldflag' column in Configure platform table.
5948 *) Fix EVP_CIPHER_mode macro.
5949 ["Dan S. Camper" <dan@bti.net>]
5951 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5952 type, we must throw them away by setting rr->length to 0.
5953 [D P Chang <dpc@qualys.com>]
5955 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5957 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5958 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5959 worked incorrectly for those cases where range = 10..._2 and
5960 3*range is two bits longer than range.)
5963 *) Only add signing time to PKCS7 structures if it is not already
5967 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5968 OBJ_ld_ce should be OBJ_id_ce.
5969 Also some ip-pda OIDs in crypto/objects/objects.txt were
5970 incorrect (cf. RFC 3039).
5971 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5973 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5974 returns early because it has nothing to do.
5975 [Andy Schneider <andy.schneider@bjss.co.uk>]
5977 *) [In 0.9.6c-engine release:]
5978 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5979 [Andy Schneider <andy.schneider@bjss.co.uk>]
5981 *) [In 0.9.6c-engine release:]
5982 Add support for Cryptographic Appliance's keyserver technology.
5983 (Use engine 'keyclient')
5984 [Cryptographic Appliances and Geoff Thorpe]
5986 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5987 is called via tools/c89.sh because arguments have to be
5988 rearranged (all '-L' options must appear before the first object
5990 [Richard Shapiro <rshapiro@abinitio.com>]
5992 *) [In 0.9.6c-engine release:]
5993 Add support for Broadcom crypto accelerator cards, backported
5995 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5997 *) [In 0.9.6c-engine release:]
5998 Add support for SureWare crypto accelerator cards from
5999 Baltimore Technologies. (Use engine 'sureware')
6000 [Baltimore Technologies and Mark Cox]
6002 *) [In 0.9.6c-engine release:]
6003 Add support for crypto accelerator cards from Accelerated
6004 Encryption Processing, www.aep.ie. (Use engine 'aep')
6005 [AEP Inc. and Mark Cox]
6007 *) Add a configuration entry for gcc on UnixWare.
6008 [Gary Benson <gbenson@redhat.com>]
6010 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6011 messages are stored in a single piece (fixed-length part and
6012 variable-length part combined) and fix various bugs found on the way.
6015 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6016 instead. BIO_gethostbyname() does not know what timeouts are
6017 appropriate, so entries would stay in cache even when they have
6019 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6021 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6022 faced with a pathologically small ClientHello fragment that does
6023 not contain client_version: Instead of aborting with an error,
6024 simply choose the highest available protocol version (i.e.,
6025 TLS 1.0 unless it is disabled). In practice, ClientHello
6026 messages are never sent like this, but this change gives us
6027 strictly correct behaviour at least for TLS.
6030 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6031 never resets s->method to s->ctx->method when called from within
6032 one of the SSL handshake functions.
6033 [Bodo Moeller; problem pointed out by Niko Baric]
6035 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6036 (sent using the client's version number) if client_version is
6037 smaller than the protocol version in use. Also change
6038 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6039 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6040 the client will at least see that alert.
6043 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6047 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6048 client receives HelloRequest while in a handshake.
6049 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6051 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6052 should end in 'break', not 'goto end' which circuments various
6053 cleanups done in state SSL_ST_OK. But session related stuff
6054 must be disabled for SSL_ST_OK in the case that we just sent a
6057 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6058 before just sending a HelloRequest.
6059 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6061 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6062 reveal whether illegal block cipher padding was found or a MAC
6063 verification error occured. (Neither SSLerr() codes nor alerts
6064 are directly visible to potential attackers, but the information
6065 may leak via logfiles.)
6067 Similar changes are not required for the SSL 2.0 implementation
6068 because the number of padding bytes is sent in clear for SSL 2.0,
6069 and the extra bytes are just ignored. However ssl/s2_pkt.c
6070 failed to verify that the purported number of padding bytes is in
6074 *) Add OpenUNIX-8 support including shared libraries
6075 (Boyd Lynn Gerber <gerberb@zenez.com>).
6078 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6079 'wristwatch attack' using huge encoding parameters (cf.
6080 James H. Manger's CRYPTO 2001 paper). Note that the
6081 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6082 encoding parameters and hence was not vulnerable.
6085 *) BN_sqr() bug fix.
6086 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6088 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6089 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6090 followed by modular reduction.
6091 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6093 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6094 equivalent based on BN_pseudo_rand() instead of BN_rand().
6097 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6098 This function was broken, as the check for a new client hello message
6099 to handle SGC did not allow these large messages.
6100 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6103 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6106 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6107 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6110 *) Rework the configuration and shared library support for Tru64 Unix.
6111 The configuration part makes use of modern compiler features and
6112 still retains old compiler behavior for those that run older versions
6113 of the OS. The shared library support part includes a variant that
6114 uses the RPATH feature, and is available through the special
6115 configuration target "alpha-cc-rpath", which will never be selected
6117 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6119 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6120 with the same message size as in ssl3_get_certificate_request().
6121 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6122 messages might inadvertently be reject as too long.
6123 [Petr Lampa <lampa@fee.vutbr.cz>]
6125 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6128 *) Modified SSL library such that the verify_callback that has been set
6129 specificly for an SSL object with SSL_set_verify() is actually being
6130 used. Before the change, a verify_callback set with this function was
6131 ignored and the verify_callback() set in the SSL_CTX at the time of
6132 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6133 to allow the necessary settings.
6136 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6137 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6138 done automatically (in contradiction to the requirements of the C
6139 standard). This made problems when used from OpenSSH.
6142 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6143 dh->length and always used
6145 BN_rand_range(priv_key, dh->p).
6147 BN_rand_range() is not necessary for Diffie-Hellman, and this
6148 specific range makes Diffie-Hellman unnecessarily inefficient if
6149 dh->length (recommended exponent length) is much smaller than the
6150 length of dh->p. We could use BN_rand_range() if the order of
6151 the subgroup was stored in the DH structure, but we only have
6156 BN_rand(priv_key, l, ...)
6158 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6164 RSA_eay_public_encrypt
6165 RSA_eay_private_decrypt
6166 RSA_eay_private_encrypt (signing)
6167 RSA_eay_public_decrypt (signature verification)
6169 (default implementations for RSA_public_encrypt,
6170 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6171 always reject numbers >= n.
6174 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6175 to synchronize access to 'locking_thread'. This is necessary on
6176 systems where access to 'locking_thread' (an 'unsigned long'
6177 variable) is not atomic.
6180 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6181 *before* setting the 'crypto_lock_rand' flag. The previous code had
6182 a race condition if 0 is a valid thread ID.
6183 [Travis Vitek <vitek@roguewave.com>]
6185 *) Add support for shared libraries under Irix.
6186 [Albert Chin-A-Young <china@thewrittenword.com>]
6188 *) Add configuration option to build on Linux on both big-endian and
6190 [Ralf Baechle <ralf@uni-koblenz.de>]
6192 *) Add the possibility to create shared libraries on HP-UX.
6195 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6197 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6198 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6199 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6200 PRNG state recovery was possible based on the output of
6201 one PRNG request appropriately sized to gain knowledge on
6202 'md' followed by enough consecutive 1-byte PRNG requests
6203 to traverse all of 'state'.
6205 1. When updating 'md_local' (the current thread's copy of 'md')
6206 during PRNG output generation, hash all of the previous
6207 'md_local' value, not just the half used for PRNG output.
6209 2. Make the number of bytes from 'state' included into the hash
6210 independent from the number of PRNG bytes requested.
6212 The first measure alone would be sufficient to avoid
6213 Markku-Juhani's attack. (Actually it had never occurred
6214 to me that the half of 'md_local' used for chaining was the
6215 half from which PRNG output bytes were taken -- I had always
6216 assumed that the secret half would be used.) The second
6217 measure makes sure that additional data from 'state' is never
6218 mixed into 'md_local' in small portions; this heuristically
6219 further strengthens the PRNG.
6222 *) Fix crypto/bn/asm/mips3.s.
6225 *) When only the key is given to "enc", the IV is undefined. Print out
6226 an error message in this case.
6229 *) Handle special case when X509_NAME is empty in X509 printing routines.
6232 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6233 positive and less than q.
6236 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6237 used: it isn't thread safe and the add_lock_callback should handle
6239 [Paul Rose <Paul.Rose@bridge.com>]
6241 *) Verify that incoming data obeys the block size in
6242 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6246 [Ulf Möller, Bodo Möller]
6248 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6249 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6250 when fixing the server behaviour for backwards-compatible 'client
6251 hello' messages. (Note that the attack is impractical against
6252 SSL 3.0 and TLS 1.0 anyway because length and version checking
6253 means that the probability of guessing a valid ciphertext is
6254 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6257 Before 0.9.5, the countermeasure (hide the error by generating a
6258 random 'decryption result') did not work properly because
6259 ERR_clear_error() was missing, meaning that SSL_get_error() would
6260 detect the supposedly ignored error.
6262 Both problems are now fixed.
6265 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6266 (previously it was 1024).
6269 *) Fix for compatibility mode trust settings: ignore trust settings
6270 unless some valid trust or reject settings are present.
6273 *) Fix for blowfish EVP: its a variable length cipher.
6276 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6277 parameters in DSA public key structures and return an error in the
6278 DSA routines if parameters are absent.
6281 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6282 in the current directory if neither $RANDFILE nor $HOME was set.
6283 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6284 caused some confusion to Windows users who haven't defined $HOME.
6285 Thus RAND_file_name() is changed again: e_os.h can define a
6286 DEFAULT_HOME, which will be used if $HOME is not set.
6287 For Windows, we use "C:"; on other platforms, we still require
6288 environment variables.
6290 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6291 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6292 having multiple threads call RAND_poll() concurrently.
6295 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6296 combination of a flag and a thread ID variable.
6297 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6298 flag), *other* threads can enter ssleay_add_bytes without obeying
6299 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6300 that they do not hold after the first thread unsets add_do_not_lock).
6303 *) Change bctest again: '-x' expressions are not available in all
6307 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6309 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6310 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6312 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6313 the default extension for executables, if any. Also, make the perl
6314 scripts that use symlink() to test if it really exists and use "cp"
6315 if it doesn't. All this made OpenSSL compilable and installable in
6319 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6320 If SEQUENCE is length is indefinite just set c->slen to the total
6321 amount of data available.
6322 [Steve Henson, reported by shige@FreeBSD.org]
6323 [This change does not apply to 0.9.7.]
6325 *) Change bctest to avoid here-documents inside command substitution
6326 (workaround for FreeBSD /bin/sh bug).
6327 For compatibility with Ultrix, avoid shell functions (introduced
6328 in the bctest version that searches along $PATH).
6331 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6332 with des_encrypt() defined on some operating systems, like Solaris
6336 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6337 On the Importance of Eliminating Errors in Cryptographic
6338 Computations, J. Cryptology 14 (2001) 2, 101-119,
6339 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6342 *) MIPS assembler BIGNUM division bug fix.
6345 *) Disabled incorrect Alpha assembler code.
6348 *) Fix PKCS#7 decode routines so they correctly update the length
6349 after reading an EOC for the EXPLICIT tag.
6351 [This change does not apply to 0.9.7.]
6353 *) Fix bug in PKCS#12 key generation routines. This was triggered
6354 if a 3DES key was generated with a 0 initial byte. Include
6355 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6356 (but broken) behaviour.
6359 *) Enhance bctest to search for a working bc along $PATH and print
6361 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6363 *) Fix memory leaks in err.c: free err_data string if necessary;
6364 don't write to the wrong index in ERR_set_error_data.
6367 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6371 *) Replace rdtsc with _emit statements for VC++ version 5.
6372 [Jeremy Cooper <jeremy@baymoo.org>]
6374 *) Make it possible to reuse SSLv2 sessions.
6377 *) In copy_email() check for >= 0 as a return value for
6378 X509_NAME_get_index_by_NID() since 0 is a valid index.
6379 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6381 *) Avoid coredump with unsupported or invalid public keys by checking if
6382 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6383 PKCS7_verify() fails with non detached data.
6386 *) Don't use getenv in library functions when run as setuid/setgid.
6387 New function OPENSSL_issetugid().
6390 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6391 due to incorrect handling of multi-threading:
6393 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6395 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6397 3. Count how many times MemCheck_off() has been called so that
6398 nested use can be treated correctly. This also avoids
6399 inband-signalling in the previous code (which relied on the
6400 assumption that thread ID 0 is impossible).
6403 *) Add "-rand" option also to s_client and s_server.
6406 *) Fix CPU detection on Irix 6.x.
6407 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6408 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6410 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6413 [This change does not apply to 0.9.7.]
6415 *) Use the cached encoding of an X509_NAME structure rather than
6416 copying it. This is apparently the reason for the libsafe "errors"
6417 but the code is actually correct.
6420 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6421 Bleichenbacher's DSA attack.
6422 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6423 to be set and top=0 forces the highest bit to be set; top=-1 is new
6424 and leaves the highest bit random.
6425 [Ulf Moeller, Bodo Moeller]
6427 *) In the NCONF_...-based implementations for CONF_... queries
6428 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6429 a temporary CONF structure with the data component set to NULL
6430 (which gives segmentation faults in lh_retrieve).
6431 Instead, use NULL for the CONF pointer in CONF_get_string and
6432 CONF_get_number (which may use environment variables) and directly
6433 return NULL from CONF_get_section.
6436 *) Fix potential buffer overrun for EBCDIC.
6439 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6440 keyUsage if basicConstraints absent for a CA.
6443 *) Make SMIME_write_PKCS7() write mail header values with a format that
6444 is more generally accepted (no spaces before the semicolon), since
6445 some programs can't parse those values properly otherwise. Also make
6446 sure BIO's that break lines after each write do not create invalid
6450 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6451 macros previously used would not encode an empty SEQUENCE OF
6452 and break the signature.
6454 [This change does not apply to 0.9.7.]
6456 *) Zero the premaster secret after deriving the master secret in
6460 *) Add some EVP_add_digest_alias registrations (as found in
6461 OpenSSL_add_all_digests()) to SSL_library_init()
6462 aka OpenSSL_add_ssl_algorithms(). This provides improved
6463 compatibility with peers using X.509 certificates
6464 with unconventional AlgorithmIdentifier OIDs.
6467 *) Fix for Irix with NO_ASM.
6468 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6470 *) ./config script fixes.
6471 [Ulf Moeller, Richard Levitte]
6473 *) Fix 'openssl passwd -1'.
6476 *) Change PKCS12_key_gen_asc() so it can cope with non null
6477 terminated strings whose length is passed in the passlen
6478 parameter, for example from PEM callbacks. This was done
6479 by adding an extra length parameter to asc2uni().
6480 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6482 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6483 call failed, free the DSA structure.
6486 *) Fix to uni2asc() to cope with zero length Unicode strings.
6487 These are present in some PKCS#12 files.
6490 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6491 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6492 when writing a 32767 byte record.
6493 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6495 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6496 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6498 (RSA objects have a reference count access to which is protected
6499 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6500 so they are meant to be shared between threads.)
6501 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6502 "Reddie, Steven" <Steven.Reddie@ca.com>]
6504 *) Fix a deadlock in CRYPTO_mem_leaks().
6507 *) Use better test patterns in bntest.
6510 *) rand_win.c fix for Borland C.
6513 *) BN_rshift bugfix for n == 0.
6516 *) Add a 'bctest' script that checks for some known 'bc' bugs
6517 so that 'make test' does not abort just because 'bc' is broken.
6520 *) Store verify_result within SSL_SESSION also for client side to
6521 avoid potential security hole. (Re-used sessions on the client side
6522 always resulted in verify_result==X509_V_OK, not using the original
6523 result of the server certificate verification.)
6526 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6527 SSL3_RT_APPLICATION_DATA, return 0.
6528 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6532 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6533 releases, have been re-implemented by renaming the previous
6534 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6535 and ssl3_read_internal, respectively, and adding 'peek' parameters
6536 to them. The new ssl[23]_{read,peek} functions are calls to
6537 ssl[23]_read_internal with the 'peek' flag set appropriately.
6538 A 'peek' parameter has also been added to ssl3_read_bytes, which
6539 does the actual work for ssl3_read_internal.
6542 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6543 the method-specific "init()" handler. Also clean up ex_data after
6544 calling the method-specific "finish()" handler. Previously, this was
6545 happening the other way round.
6548 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6549 The previous value, 12, was not always sufficient for BN_mod_exp().
6552 *) Make sure that shared libraries get the internal name engine with
6553 the full version number and not just 0. This should mark the
6554 shared libraries as not backward compatible. Of course, this should
6555 be changed again when we can guarantee backward binary compatibility.
6558 *) Fix typo in get_cert_by_subject() in by_dir.c
6559 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6561 *) Rework the system to generate shared libraries:
6563 - Make note of the expected extension for the shared libraries and
6564 if there is a need for symbolic links from for example libcrypto.so.0
6565 to libcrypto.so.0.9.7. There is extended info in Configure for
6568 - Make as few rebuilds of the shared libraries as possible.
6570 - Still avoid linking the OpenSSL programs with the shared libraries.
6572 - When installing, install the shared libraries separately from the
6576 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6578 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6579 and not in SSL_clear because the latter is also used by the
6580 accept/connect functions; previously, the settings made by
6581 SSL_set_read_ahead would be lost during the handshake.
6582 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6584 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6585 Previously, it would create entries for disableed algorithms no
6589 *) Added several new manual pages for SSL_* function.
6592 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6594 *) In ssl23_get_client_hello, generate an error message when faced
6595 with an initial SSL 3.0/TLS record that is too small to contain the
6596 first two bytes of the ClientHello message, i.e. client_version.
6597 (Note that this is a pathologic case that probably has never happened
6598 in real life.) The previous approach was to use the version number
6599 from the record header as a substitute; but our protocol choice
6600 should not depend on that one because it is not authenticated
6601 by the Finished messages.
6604 *) More robust randomness gathering functions for Windows.
6605 [Jeffrey Altman <jaltman@columbia.edu>]
6607 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6608 not set then we don't setup the error code for issuer check errors
6609 to avoid possibly overwriting other errors which the callback does
6610 handle. If an application does set the flag then we assume it knows
6611 what it is doing and can handle the new informational codes
6615 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6616 a general "ANY" type, as such it should be able to decode anything
6617 including tagged types. However it didn't check the class so it would
6618 wrongly interpret tagged types in the same way as their universal
6619 counterpart and unknown types were just rejected. Changed so that the
6620 tagged and unknown types are handled in the same way as a SEQUENCE:
6621 that is the encoding is stored intact. There is also a new type
6622 "V_ASN1_OTHER" which is used when the class is not universal, in this
6623 case we have no idea what the actual type is so we just lump them all
6627 *) On VMS, stdout may very well lead to a file that is written to
6628 in a record-oriented fashion. That means that every write() will
6629 write a separate record, which will be read separately by the
6630 programs trying to read from it. This can be very confusing.
6632 The solution is to put a BIO filter in the way that will buffer
6633 text until a linefeed is reached, and then write everything a
6634 line at a time, so every record written will be an actual line,
6635 not chunks of lines and not (usually doesn't happen, but I've
6636 seen it once) several lines in one record. BIO_f_linebuffer() is
6639 Currently, it's a VMS-only method, because that's where it has
6640 been tested well enough.
6643 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6644 it can return incorrect results.
6645 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6646 but it was in 0.9.6-beta[12].)
6649 *) Disable the check for content being present when verifying detached
6650 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6651 include zero length content when signing messages.
6654 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6655 BIO_ctrl (for BIO pairs).
6658 *) Add DSO method for VMS.
6661 *) Bug fix: Montgomery multiplication could produce results with the
6665 *) Add RPM specification openssl.spec and modify it to build three
6666 packages. The default package contains applications, application
6667 documentation and run-time libraries. The devel package contains
6668 include files, static libraries and function documentation. The
6669 doc package contains the contents of the doc directory. The original
6670 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6673 *) Add a large number of documentation files for many SSL routines.
6674 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6676 *) Add a configuration entry for Sony News 4.
6677 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6679 *) Don't set the two most significant bits to one when generating a
6680 random number < q in the DSA library.
6683 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6684 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6685 the underlying transport is blocking) if a handshake took place.
6686 (The default behaviour is needed by applications such as s_client
6687 and s_server that use select() to determine when to use SSL_read;
6688 but for applications that know in advance when to expect data, it
6689 just makes things more complicated.)
6692 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6696 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6697 work better on such systems.
6698 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6700 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6701 Update PKCS12_parse() so it copies the friendlyName and the
6702 keyid to the certificates aux info.
6705 *) Fix bug in PKCS7_verify() which caused an infinite loop
6706 if there was more than one signature.
6707 [Sven Uszpelkat <su@celocom.de>]
6709 *) Major change in util/mkdef.pl to include extra information
6710 about each symbol, as well as presentig variables as well
6711 as functions. This change means that there's n more need
6712 to rebuild the .num files when some algorithms are excluded.
6715 *) Allow the verify time to be set by an application,
6716 rather than always using the current time.
6719 *) Phase 2 verify code reorganisation. The certificate
6720 verify code now looks up an issuer certificate by a
6721 number of criteria: subject name, authority key id
6722 and key usage. It also verifies self signed certificates
6723 by the same criteria. The main comparison function is
6724 X509_check_issued() which performs these checks.
6726 Lot of changes were necessary in order to support this
6727 without completely rewriting the lookup code.
6729 Authority and subject key identifier are now cached.
6731 The LHASH 'certs' is X509_STORE has now been replaced
6732 by a STACK_OF(X509_OBJECT). This is mainly because an
6733 LHASH can't store or retrieve multiple objects with
6734 the same hash value.
6736 As a result various functions (which were all internal
6737 use only) have changed to handle the new X509_STORE
6738 structure. This will break anything that messed round
6739 with X509_STORE internally.
6741 The functions X509_STORE_add_cert() now checks for an
6742 exact match, rather than just subject name.
6744 The X509_STORE API doesn't directly support the retrieval
6745 of multiple certificates matching a given criteria, however
6746 this can be worked round by performing a lookup first
6747 (which will fill the cache with candidate certificates)
6748 and then examining the cache for matches. This is probably
6749 the best we can do without throwing out X509_LOOKUP
6750 entirely (maybe later...).
6752 The X509_VERIFY_CTX structure has been enhanced considerably.
6754 All certificate lookup operations now go via a get_issuer()
6755 callback. Although this currently uses an X509_STORE it
6756 can be replaced by custom lookups. This is a simple way
6757 to bypass the X509_STORE hackery necessary to make this
6758 work and makes it possible to use more efficient techniques
6759 in future. A very simple version which uses a simple
6760 STACK for its trusted certificate store is also provided
6761 using X509_STORE_CTX_trusted_stack().
6763 The verify_cb() and verify() callbacks now have equivalents
6764 in the X509_STORE_CTX structure.
6766 X509_STORE_CTX also has a 'flags' field which can be used
6767 to customise the verify behaviour.
6770 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6771 excludes S/MIME capabilities.
6774 *) When a certificate request is read in keep a copy of the
6775 original encoding of the signed data and use it when outputing
6776 again. Signatures then use the original encoding rather than
6777 a decoded, encoded version which may cause problems if the
6778 request is improperly encoded.
6781 *) For consistency with other BIO_puts implementations, call
6782 buffer_write(b, ...) directly in buffer_puts instead of calling
6785 In BIO_puts, increment b->num_write as in BIO_write.
6786 [Peter.Sylvester@EdelWeb.fr]
6788 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6789 BN_zero, we may not return a BIGNUM with an array consisting of
6793 *) Avoid calling abort() from within the library when problems are
6794 detected, except if preprocessor symbols have been defined
6795 (such as REF_CHECK, BN_DEBUG etc.).
6798 *) New openssl application 'rsautl'. This utility can be
6799 used for low level RSA operations. DER public key
6800 BIO/fp routines also added.
6803 *) New Configure entry and patches for compiling on QNX 4.
6804 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6806 *) A demo state-machine implementation was sponsored by
6807 Nuron (http://www.nuron.com/) and is now available in
6808 demos/state_machine.
6811 *) New options added to the 'dgst' utility for signature
6812 generation and verification.
6815 *) Unrecognized PKCS#7 content types are now handled via a
6816 catch all ASN1_TYPE structure. This allows unsupported
6817 types to be stored as a "blob" and an application can
6818 encode and decode it manually.
6821 *) Fix various signed/unsigned issues to make a_strex.c
6823 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6825 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6826 length if passed a buffer. ASN1_INTEGER_to_BN failed
6827 if passed a NULL BN and its argument was negative.
6828 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6830 *) Modification to PKCS#7 encoding routines to output definite
6831 length encoding. Since currently the whole structures are in
6832 memory there's not real point in using indefinite length
6833 constructed encoding. However if OpenSSL is compiled with
6834 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6837 *) Added BIO_vprintf() and BIO_vsnprintf().
6840 *) Added more prefixes to parse for in the the strings written
6841 through a logging bio, to cover all the levels that are available
6842 through syslog. The prefixes are now:
6844 PANIC, EMERG, EMR => LOG_EMERG
6845 ALERT, ALR => LOG_ALERT
6846 CRIT, CRI => LOG_CRIT
6847 ERROR, ERR => LOG_ERR
6848 WARNING, WARN, WAR => LOG_WARNING
6849 NOTICE, NOTE, NOT => LOG_NOTICE
6850 INFO, INF => LOG_INFO
6851 DEBUG, DBG => LOG_DEBUG
6853 and as before, if none of those prefixes are present at the
6854 beginning of the string, LOG_ERR is chosen.
6856 On Win32, the LOG_* levels are mapped according to this:
6858 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6859 LOG_WARNING => EVENTLOG_WARNING_TYPE
6860 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6864 *) Made it possible to reconfigure with just the configuration
6865 argument "reconf" or "reconfigure". The command line arguments
6866 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6867 and are retrieved from there when reconfiguring.
6871 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6873 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6876 *) The obj_dat.pl script was messing up the sorting of object
6877 names. The reason was that it compared the quoted version
6878 of strings as a result "OCSP" > "OCSP Signing" because
6879 " > SPACE. Changed script to store unquoted versions of
6880 names and add quotes on output. It was also omitting some
6881 names from the lookup table if they were given a default
6882 value (that is if SN is missing it is given the same
6883 value as LN and vice versa), these are now added on the
6884 grounds that if an object has a name we should be able to
6885 look it up. Finally added warning output when duplicate
6886 short or long names are found.
6889 *) Changes needed for Tandem NSK.
6890 [Scott Uroff <scott@xypro.com>]
6892 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6893 RSA_padding_check_SSLv23(), special padding was never detected
6894 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6895 version rollback attacks was not effective.
6897 In s23_clnt.c, don't use special rollback-attack detection padding
6898 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6899 client; similarly, in s23_srvr.c, don't do the rollback check if
6900 SSL 2.0 is the only protocol enabled in the server.
6903 *) Make it possible to get hexdumps of unprintable data with 'openssl
6904 asn1parse'. By implication, the functions ASN1_parse_dump() and
6905 BIO_dump_indent() are added.
6908 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6909 these print out strings and name structures based on various
6910 flags including RFC2253 support and proper handling of
6911 multibyte characters. Added options to the 'x509' utility
6912 to allow the various flags to be set.
6915 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6916 Also change the functions X509_cmp_current_time() and
6917 X509_gmtime_adj() work with an ASN1_TIME structure,
6918 this will enable certificates using GeneralizedTime in validity
6919 dates to be checked.
6922 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6923 negative public key encodings) on by default,
6924 NO_NEG_PUBKEY_BUG can be set to disable it.
6927 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6928 content octets. An i2c_ASN1_OBJECT is unnecessary because
6929 the encoding can be trivially obtained from the structure.
6932 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6933 not read locks (CRYPTO_r_[un]lock).
6936 *) A first attempt at creating official support for shared
6937 libraries through configuration. I've kept it so the
6938 default is static libraries only, and the OpenSSL programs
6939 are always statically linked for now, but there are
6940 preparations for dynamic linking in place.
6941 This has been tested on Linux and Tru64.
6944 *) Randomness polling function for Win9x, as described in:
6945 Peter Gutmann, Software Generation of Practically Strong
6949 *) Fix so PRNG is seeded in req if using an already existing
6953 *) New options to smime application. -inform and -outform
6954 allow alternative formats for the S/MIME message including
6955 PEM and DER. The -content option allows the content to be
6956 specified separately. This should allow things like Netscape
6957 form signing output easier to verify.
6960 *) Fix the ASN1 encoding of tags using the 'long form'.
6963 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6964 STRING types. These convert content octets to and from the
6965 underlying type. The actual tag and length octets are
6966 already assumed to have been read in and checked. These
6967 are needed because all other string types have virtually
6968 identical handling apart from the tag. By having versions
6969 of the ASN1 functions that just operate on content octets
6970 IMPLICIT tagging can be handled properly. It also allows
6971 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6972 and ASN1_INTEGER are identical apart from the tag.
6975 *) Change the handling of OID objects as follows:
6977 - New object identifiers are inserted in objects.txt, following
6978 the syntax given in objects.README.
6979 - objects.pl is used to process obj_mac.num and create a new
6981 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6984 This is currently kind of a hack, and the perl code in objects.pl
6985 isn't very elegant, but it works as I intended. The simplest way
6986 to check that it worked correctly is to look in obj_dat.h and
6987 check the array nid_objs and make sure the objects haven't moved
6988 around (this is important!). Additions are OK, as well as
6989 consistent name changes.
6992 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6995 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6996 The given file adds to whatever has already been seeded into the
6997 random pool through the RANDFILE configuration file option or
6998 environment variable, or the default random state file.
7001 *) mkstack.pl now sorts each macro group into lexical order.
7002 Previously the output order depended on the order the files
7003 appeared in the directory, resulting in needless rewriting
7007 *) Patches to make OpenSSL compile under Win32 again. Mostly
7008 work arounds for the VC++ problem that it treats func() as
7009 func(void). Also stripped out the parts of mkdef.pl that
7010 added extra typesafe functions: these no longer exist.
7013 *) Reorganisation of the stack code. The macros are now all
7014 collected in safestack.h . Each macro is defined in terms of
7015 a "stack macro" of the form SKM_<name>(type, a, b). The
7016 DEBUG_SAFESTACK is now handled in terms of function casts,
7017 this has the advantage of retaining type safety without the
7018 use of additional functions. If DEBUG_SAFESTACK is not defined
7019 then the non typesafe macros are used instead. Also modified the
7020 mkstack.pl script to handle the new form. Needs testing to see
7021 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7022 the default if no major problems. Similar behaviour for ASN1_SET_OF
7023 and PKCS12_STACK_OF.
7026 *) When some versions of IIS use the 'NET' form of private key the
7027 key derivation algorithm is different. Normally MD5(password) is
7028 used as a 128 bit RC4 key. In the modified case
7029 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7030 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7031 as the old Netscape_RSA functions except they have an additional
7032 'sgckey' parameter which uses the modified algorithm. Also added
7033 an -sgckey command line option to the rsa utility. Thanks to
7034 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7035 algorithm to openssl-dev.
7038 *) The evp_local.h macros were using 'c.##kname' which resulted in
7039 invalid expansion on some systems (SCO 5.0.5 for example).
7040 Corrected to 'c.kname'.
7041 [Phillip Porch <root@theporch.com>]
7043 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7044 a STACK of email addresses from a certificate or request, these look
7045 in the subject name and the subject alternative name extensions and
7046 omit any duplicate addresses.
7049 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7050 This makes DSA verification about 2 % faster.
7053 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7054 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7055 plus overhead for 1024 bit moduli).
7056 This makes exponentiations about 0.5 % faster for 1024 bit
7057 exponents (as measured by "openssl speed rsa2048").
7060 *) Rename memory handling macros to avoid conflicts with other
7062 Malloc => OPENSSL_malloc
7063 Malloc_locked => OPENSSL_malloc_locked
7064 Realloc => OPENSSL_realloc
7065 Free => OPENSSL_free
7068 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7069 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7072 *) CygWin32 support.
7073 [John Jarvie <jjarvie@newsguy.com>]
7075 *) The type-safe stack code has been rejigged. It is now only compiled
7076 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7077 by default all type-specific stack functions are "#define"d back to
7078 standard stack functions. This results in more streamlined output
7079 but retains the type-safety checking possibilities of the original
7083 *) The STACK code has been cleaned up, and certain type declarations
7084 that didn't make a lot of sense have been brought in line. This has
7085 also involved a cleanup of sorts in safestack.h to more correctly
7086 map type-safe stack functions onto their plain stack counterparts.
7087 This work has also resulted in a variety of "const"ifications of
7088 lots of the code, especially "_cmp" operations which should normally
7089 be prototyped with "const" parameters anyway.
7092 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7093 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7094 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7095 where all of 'md' is used each time the PRNG is used, but 'state'
7096 is used only indexed by a cyclic counter. As entropy may not be
7097 well distributed from the beginning, 'md' is important as a
7098 chaining variable. However, the output function chains only half
7099 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7100 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7101 in all of 'state' being rewritten, with the new values depending
7102 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7105 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7106 the handshake is continued after ssl_verify_cert_chain();
7107 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7108 can lead to 'unexplainable' connection aborts later.
7109 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7111 *) Major EVP API cipher revision.
7112 Add hooks for extra EVP features. This allows various cipher
7113 parameters to be set in the EVP interface. Support added for variable
7114 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7115 setting of RC2 and RC5 parameters.
7117 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7120 Remove lots of duplicated code from the EVP library. For example *every*
7121 cipher init() function handles the 'iv' in the same way according to the
7122 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7123 for CFB and OFB modes they zero ctx->num.
7125 New functionality allows removal of S/MIME code RC2 hack.
7127 Most of the routines have the same form and so can be declared in terms
7130 By shifting this to the top level EVP_CipherInit() it can be removed from
7131 all individual ciphers. If the cipher wants to handle IVs or keys
7132 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7135 Change lots of functions like EVP_EncryptUpdate() to now return a
7136 value: although software versions of the algorithms cannot fail
7137 any installed hardware versions can.
7140 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7141 this option is set, tolerate broken clients that send the negotiated
7142 protocol version number instead of the requested protocol version
7146 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7147 i.e. non-zero for export ciphersuites, zero otherwise.
7148 Previous versions had this flag inverted, inconsistent with
7149 rsa_tmp_cb (..._TMP_RSA_CB).
7150 [Bodo Moeller; problem reported by Amit Chopra]
7152 *) Add missing DSA library text string. Work around for some IIS
7153 key files with invalid SEQUENCE encoding.
7156 *) Add a document (doc/standards.txt) that list all kinds of standards
7157 and so on that are implemented in OpenSSL.
7160 *) Enhance c_rehash script. Old version would mishandle certificates
7161 with the same subject name hash and wouldn't handle CRLs at all.
7162 Added -fingerprint option to crl utility, to support new c_rehash
7166 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7169 *) Fix for SSL server purpose checking. Server checking was
7170 rejecting certificates which had extended key usage present
7171 but no ssl client purpose.
7172 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7174 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7175 is a little unclear about how a blank password is handled.
7176 Since the password in encoded as a BMPString with terminating
7177 double NULL a zero length password would end up as just the
7178 double NULL. However no password at all is different and is
7179 handled differently in the PKCS#12 key generation code. NS
7180 treats a blank password as zero length. MSIE treats it as no
7181 password on export: but it will try both on import. We now do
7182 the same: PKCS12_parse() tries zero length and no password if
7183 the password is set to "" or NULL (NULL is now a valid password:
7184 it wasn't before) as does the pkcs12 application.
7187 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7188 perror when PEM_read_bio_X509_REQ fails, the error message must
7189 be obtained from the error queue.
7192 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7193 it in ERR_remove_state if appropriate, and change ERR_get_state
7194 accordingly to avoid race conditions (this is necessary because
7195 thread_hash is no longer constant once set).
7198 *) Bugfix for linux-elf makefile.one.
7201 *) RSA_get_default_method() will now cause a default
7202 RSA_METHOD to be chosen if one doesn't exist already.
7203 Previously this was only set during a call to RSA_new()
7204 or RSA_new_method(NULL) meaning it was possible for
7205 RSA_get_default_method() to return NULL.
7208 *) Added native name translation to the existing DSO code
7209 that will convert (if the flag to do so is set) filenames
7210 that are sufficiently small and have no path information
7211 into a canonical native form. Eg. "blah" converted to
7212 "libblah.so" or "blah.dll" etc.
7215 *) New function ERR_error_string_n(e, buf, len) which is like
7216 ERR_error_string(e, buf), but writes at most 'len' bytes
7217 including the 0 terminator. For ERR_error_string_n, 'buf'
7219 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7221 *) CONF library reworked to become more general. A new CONF
7222 configuration file reader "class" is implemented as well as a
7223 new functions (NCONF_*, for "New CONF") to handle it. The now
7224 old CONF_* functions are still there, but are reimplemented to
7225 work in terms of the new functions. Also, a set of functions
7226 to handle the internal storage of the configuration data is
7227 provided to make it easier to write new configuration file
7228 reader "classes" (I can definitely see something reading a
7229 configuration file in XML format, for example), called _CONF_*,
7230 or "the configuration storage API"...
7232 The new configuration file reading functions are:
7234 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7235 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7237 NCONF_default, NCONF_WIN32
7239 NCONF_dump_fp, NCONF_dump_bio
7241 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7242 NCONF_new creates a new CONF object. This works in the same way
7243 as other interfaces in OpenSSL, like the BIO interface.
7244 NCONF_dump_* dump the internal storage of the configuration file,
7245 which is useful for debugging. All other functions take the same
7246 arguments as the old CONF_* functions wth the exception of the
7247 first that must be a `CONF *' instead of a `LHASH *'.
7249 To make it easer to use the new classes with the old CONF_* functions,
7250 the function CONF_set_default_method is provided.
7253 *) Add '-tls1' option to 'openssl ciphers', which was already
7254 mentioned in the documentation but had not been implemented.
7255 (This option is not yet really useful because even the additional
7256 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7259 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7260 OpenSSL-based applications) load shared libraries and bind to
7261 them in a portable way.
7262 [Geoff Thorpe, with contributions from Richard Levitte]
7264 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7266 *) Make sure _lrotl and _lrotr are only used with MSVC.
7268 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7269 (the default implementation of RAND_status).
7271 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7272 to '-clrext' (= clear extensions), as intended and documented.
7273 [Bodo Moeller; inconsistency pointed out by Michael Attili
7276 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7277 was larger than the MD block size.
7278 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7280 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7281 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7282 using the passed key: if the passed key was a private key the result
7283 of X509_print(), for example, would be to print out all the private key
7287 *) des_quad_cksum() byte order bug fix.
7288 [Ulf Möller, using the problem description in krb4-0.9.7, where
7289 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7291 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7293 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7295 *) For easily testing in shell scripts whether some command
7296 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7297 returns with exit code 0 iff no command of the given name is available.
7298 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7299 the output goes to stdout and nothing is printed to stderr.
7300 Additional arguments are always ignored.
7302 Since for each cipher there is a command of the same name,
7303 the 'no-cipher' compilation switches can be tested this way.
7305 ('openssl no-XXX' is not able to detect pseudo-commands such
7306 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7309 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7312 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7313 is set; it will be thrown away anyway because each handshake creates
7315 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7316 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7317 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7318 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7321 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7322 'Q' and 'R' lose their special meanings (quit/renegotiate).
7323 This is part of what -quiet does; unlike -quiet, -ign_eof
7324 does not suppress any output.
7327 *) Add compatibility options to the purpose and trust code. The
7328 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7329 accepts a certificate or CA, this was the previous behaviour,
7330 with all the associated security issues.
7332 X509_TRUST_COMPAT is the old trust behaviour: only and
7333 automatically trust self signed roots in certificate store. A
7334 new trust setting X509_TRUST_DEFAULT is used to specify that
7335 a purpose has no associated trust setting and it should instead
7336 use the value in the default purpose.
7339 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7340 and fix a memory leak.
7343 *) In util/mkerr.pl (which implements 'make errors'), preserve
7344 reason strings from the previous version of the .c file, as
7345 the default to have only downcase letters (and digits) in
7346 automatically generated reasons codes is not always appropriate.
7349 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7350 using strerror. Previously, ERR_reason_error_string() returned
7351 library names as reason strings for SYSerr; but SYSerr is a special
7352 case where small numbers are errno values, not library numbers.
7355 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7356 converts DSA parameters into DH parameters. (When creating parameters,
7357 DSA_generate_parameters is used.)
7360 *) Include 'length' (recommended exponent length) in C code generated
7361 by 'openssl dhparam -C'.
7364 *) The second argument to set_label in perlasm was already being used
7365 so couldn't be used as a "file scope" flag. Moved to third argument
7369 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7370 instead of RAND_bytes for encryption IVs and salts.
7373 *) Include RAND_status() into RAND_METHOD instead of implementing
7374 it only for md_rand.c Otherwise replacing the PRNG by calling
7375 RAND_set_rand_method would be impossible.
7378 *) Don't let DSA_generate_key() enter an infinite loop if the random
7379 number generation fails.
7382 *) New 'rand' application for creating pseudo-random output.
7385 *) Added configuration support for Linux/IA64
7386 [Rolf Haberrecker <rolf@suse.de>]
7388 *) Assembler module support for Mingw32.
7391 *) Shared library support for HPUX (in shlib/).
7392 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7394 *) Shared library support for Solaris gcc.
7395 [Lutz Behnke <behnke@trustcenter.de>]
7397 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7399 *) PKCS7_encrypt() was adding text MIME headers twice because they
7400 were added manually and by SMIME_crlf_copy().
7403 *) In bntest.c don't call BN_rand with zero bits argument.
7404 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7406 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7407 case was implemented. This caused BN_div_recp() to fail occasionally.
7410 *) Add an optional second argument to the set_label() in the perl
7411 assembly language builder. If this argument exists and is set
7412 to 1 it signals that the assembler should use a symbol whose
7413 scope is the entire file, not just the current function. This
7414 is needed with MASM which uses the format label:: for this scope.
7415 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7417 *) Change the ASN1 types so they are typedefs by default. Before
7418 almost all types were #define'd to ASN1_STRING which was causing
7419 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7423 *) Change names of new functions to the new get1/get0 naming
7424 convention: After 'get1', the caller owns a reference count
7425 and has to call ..._free; 'get0' returns a pointer to some
7426 data structure without incrementing reference counters.
7427 (Some of the existing 'get' functions increment a reference
7428 counter, some don't.)
7429 Similarly, 'set1' and 'add1' functions increase reference
7430 counters or duplicate objects.
7433 *) Allow for the possibility of temp RSA key generation failure:
7434 the code used to assume it always worked and crashed on failure.
7437 *) Fix potential buffer overrun problem in BIO_printf().
7438 [Ulf Möller, using public domain code by Patrick Powell; problem
7439 pointed out by David Sacerdote <das33@cornell.edu>]
7441 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7442 RAND_egd() and RAND_status(). In the command line application,
7443 the EGD socket can be specified like a seed file using RANDFILE
7447 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7448 Some CAs (e.g. Verisign) distribute certificates in this form.
7451 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7452 list to exclude them. This means that no special compilation option
7453 is needed to use anonymous DH: it just needs to be included in the
7457 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7458 EVP_MD_type. The old functionality is available in a new macro called
7459 EVP_MD_md(). Change code that uses it and update docs.
7462 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7463 where the 'void *' argument is replaced by a function pointer argument.
7464 Previously 'void *' was abused to point to functions, which works on
7465 many platforms, but is not correct. As these functions are usually
7466 called by macros defined in OpenSSL header files, most source code
7467 should work without changes.
7470 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7471 sections with information on -D... compiler switches used for
7472 compiling the library so that applications can see them. To enable
7473 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7474 must be defined. E.g.,
7475 #define OPENSSL_ALGORITHM_DEFINES
7476 #include <openssl/opensslconf.h>
7477 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7478 [Richard Levitte, Ulf and Bodo Möller]
7480 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7484 *) Change the 'other' type in certificate aux info to a STACK_OF
7485 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7486 the required ASN1 format: arbitrary types determined by an OID.
7489 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7490 argument to 'req'. This is not because the function is newer or
7491 better than others it just uses the work 'NEW' in the certificate
7492 request header lines. Some software needs this.
7495 *) Reorganise password command line arguments: now passwords can be
7496 obtained from various sources. Delete the PEM_cb function and make
7497 it the default behaviour: i.e. if the callback is NULL and the
7498 usrdata argument is not NULL interpret it as a null terminated pass
7499 phrase. If usrdata and the callback are NULL then the pass phrase
7500 is prompted for as usual.
7503 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7504 the support is automatically enabled. The resulting binaries will
7505 autodetect the card and use it if present.
7506 [Ben Laurie and Compaq Inc.]
7508 *) Work around for Netscape hang bug. This sends certificate request
7509 and server done in one record. Since this is perfectly legal in the
7510 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7511 the bugs/SSLv3 entry for more info.
7514 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7517 *) Add -rand argument to smime and pkcs12 applications and read/write
7521 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7524 *) Add command line password options to the remaining applications.
7527 *) Bug fix for BN_div_recp() for numerators with an even number of
7531 *) More tests in bntest.c, and changed test_bn output.
7534 *) ./config recognizes MacOS X now.
7537 *) Bug fix for BN_div() when the first words of num and divsor are
7538 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7541 *) Add support for various broken PKCS#8 formats, and command line
7542 options to produce them.
7545 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7546 get temporary BIGNUMs from a BN_CTX.
7549 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7553 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7554 include a #define from the old name to the new. The original intent
7555 was that statically linked binaries could for example just call
7556 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7557 link with digests. This never worked becayse SSLeay_add_all_digests()
7558 and SSLeay_add_all_ciphers() were in the same source file so calling
7559 one would link with the other. They are now in separate source files.
7562 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7565 *) Use a less unusual form of the Miller-Rabin primality test (it used
7566 a binary algorithm for exponentiation integrated into the Miller-Rabin
7567 loop, our standard modexp algorithms are faster).
7570 *) Support for the EBCDIC character set completed.
7571 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7573 *) Source code cleanups: use const where appropriate, eliminate casts,
7574 use void * instead of char * in lhash.
7577 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7578 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7579 this the server could overwrite ephemeral keys that the client
7583 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7584 using 50 iterations of the Rabin-Miller test.
7586 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7587 iterations of the Rabin-Miller test as required by the appendix
7588 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7589 As BN_is_prime_fasttest includes trial division, DSA parameter
7590 generation becomes much faster.
7592 This implies a change for the callback functions in DSA_is_prime
7593 and DSA_generate_parameters: The callback function is called once
7594 for each positive witness in the Rabin-Miller test, not just
7595 occasionally in the inner loop; and the parameters to the
7596 callback function now provide an iteration count for the outer
7597 loop rather than for the current invocation of the inner loop.
7598 DSA_generate_parameters additionally can call the callback
7599 function with an 'iteration count' of -1, meaning that a
7600 candidate has passed the trial division test (when q is generated
7601 from an application-provided seed, trial division is skipped).
7604 *) New function BN_is_prime_fasttest that optionally does trial
7605 division before starting the Rabin-Miller test and has
7606 an additional BN_CTX * argument (whereas BN_is_prime always
7607 has to allocate at least one BN_CTX).
7608 'callback(1, -1, cb_arg)' is called when a number has passed the
7609 trial division stage.
7612 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7616 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7619 *) New function BN_pseudo_rand().
7622 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7623 bignum version of BN_from_montgomery() with the working code from
7624 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7628 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7629 made it impossible to use the same SSL_SESSION data structure in
7630 SSL2 clients in multiple threads.
7633 *) The return value of RAND_load_file() no longer counts bytes obtained
7634 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7635 to seed the PRNG (previously an explicit byte count was required).
7636 [Ulf Möller, Bodo Möller]
7638 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7639 used (char *) instead of (void *) and had casts all over the place.
7642 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7645 *) Retain source code compatibility for BN_prime_checks macro:
7646 BN_is_prime(..., BN_prime_checks, ...) now uses
7647 BN_prime_checks_for_size to determine the appropriate number of
7648 Rabin-Miller iterations.
7651 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7652 DH_CHECK_P_NOT_SAFE_PRIME.
7653 (Check if this is true? OpenPGP calls them "strong".)
7656 *) Merge the functionality of "dh" and "gendh" programs into a new program
7657 "dhparam". The old programs are retained for now but will handle DH keys
7658 (instead of parameters) in future.
7661 *) Make the ciphers, s_server and s_client programs check the return values
7662 when a new cipher list is set.
7665 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7666 ciphers. Before when the 56bit ciphers were enabled the sorting was
7669 The syntax for the cipher sorting has been extended to support sorting by
7670 cipher-strength (using the strength_bits hard coded in the tables).
7671 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7673 Fix a bug in the cipher-command parser: when supplying a cipher command
7674 string with an "undefined" symbol (neither command nor alphanumeric
7675 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7676 an error is flagged.
7678 Due to the strength-sorting extension, the code of the
7679 ssl_create_cipher_list() function was completely rearranged. I hope that
7680 the readability was also increased :-)
7681 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7683 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7684 for the first serial number and places 2 in the serial number file. This
7685 avoids problems when the root CA is created with serial number zero and
7686 the first user certificate has the same issuer name and serial number
7690 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7691 the new code. Add documentation for this stuff.
7694 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7695 X509_*() to X509at_*() on the grounds that they don't handle X509
7696 structures and behave in an analagous way to the X509v3 functions:
7697 they shouldn't be called directly but wrapper functions should be used
7700 So we also now have some wrapper functions that call the X509at functions
7701 when passed certificate requests. (TO DO: similar things can be done with
7702 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7703 things. Some of these need some d2i or i2d and print functionality
7704 because they handle more complex structures.)
7707 *) Add missing #ifndefs that caused missing symbols when building libssl
7708 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7709 NO_RSA in ssl/s2*.c.
7710 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7712 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7713 has a return value which indicates the quality of the random data
7714 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7715 error queue. New function RAND_pseudo_bytes() generates output that is
7716 guaranteed to be unique but not unpredictable. RAND_add is like
7717 RAND_seed, but takes an extra argument for an entropy estimate
7718 (RAND_seed always assumes full entropy).
7721 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7722 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7723 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7724 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7725 false-positive rate of at most 2^-80 for random input.
7728 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7731 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7732 in the 0.9.5 release), this returns the chain
7733 from an X509_CTX structure with a dup of the stack and all
7734 the X509 reference counts upped: so the stack will exist
7735 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7738 Also make SSL_SESSION_print() print out the verify return
7742 *) Add manpage for the pkcs12 command. Also change the default
7743 behaviour so MAC iteration counts are used unless the new
7744 -nomaciter option is used. This improves file security and
7745 only older versions of MSIE (4.0 for example) need it.
7748 *) Honor the no-xxx Configure options when creating .DEF files.
7751 *) Add PKCS#10 attributes to field table: challengePassword,
7752 unstructuredName and unstructuredAddress. These are taken from
7753 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7754 international characters are used.
7756 More changes to X509_ATTRIBUTE code: allow the setting of types
7757 based on strings. Remove the 'loc' parameter when adding
7758 attributes because these will be a SET OF encoding which is sorted
7762 *) Initial changes to the 'req' utility to allow request generation
7763 automation. This will allow an application to just generate a template
7764 file containing all the field values and have req construct the
7767 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7768 used all over the place including certificate requests and PKCS#7
7769 structures. They are currently handled manually where necessary with
7770 some primitive wrappers for PKCS#7. The new functions behave in a
7771 manner analogous to the X509 extension functions: they allow
7772 attributes to be looked up by NID and added.
7774 Later something similar to the X509V3 code would be desirable to
7775 automatically handle the encoding, decoding and printing of the
7776 more complex types. The string types like challengePassword can
7777 be handled by the string table functions.
7779 Also modified the multi byte string table handling. Now there is
7780 a 'global mask' which masks out certain types. The table itself
7781 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7782 is useful when for example there is only one permissible type
7783 (as in countryName) and using the mask might result in no valid
7787 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7788 SSL_get_peer_finished to allow applications to obtain the latest
7789 Finished messages sent to the peer or expected from the peer,
7790 respectively. (SSL_get_peer_finished is usually the Finished message
7791 actually received from the peer, otherwise the protocol will be aborted.)
7793 As the Finished message are message digests of the complete handshake
7794 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7795 be used for external authentication procedures when the authentication
7796 provided by SSL/TLS is not desired or is not enough.
7799 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7800 the host supports BWX extension and if Compaq C is present on the
7801 $PATH. Just exploiting of the BWX extension results in 20-30%
7802 performance kick for some algorithms, e.g. DES and RC4 to mention
7803 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7807 *) Add support for MS "fast SGC". This is arguably a violation of the
7808 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7809 weak crypto and after checking the certificate is SGC a second one
7810 with strong crypto. MS SGC stops the first handshake after receiving
7811 the server certificate message and sends a second client hello. Since
7812 a server will typically do all the time consuming operations before
7813 expecting any further messages from the client (server key exchange
7814 is the most expensive) there is little difference between the two.
7816 To get OpenSSL to support MS SGC we have to permit a second client
7817 hello message after we have sent server done. In addition we have to
7818 reset the MAC if we do get this second client hello.
7821 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7822 if a DER encoded private key is RSA or DSA traditional format. Changed
7823 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7824 format DER encoded private key. Newer code should use PKCS#8 format which
7825 has the key type encoded in the ASN1 structure. Added DER private key
7826 support to pkcs8 application.
7829 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7830 ciphersuites has been selected (as required by the SSL 3/TLS 1
7831 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7832 is set, we interpret this as a request to violate the specification
7833 (the worst that can happen is a handshake failure, and 'correct'
7834 behaviour would result in a handshake failure anyway).
7837 *) In SSL_CTX_add_session, take into account that there might be multiple
7838 SSL_SESSION structures with the same session ID (e.g. when two threads
7839 concurrently obtain them from an external cache).
7840 The internal cache can handle only one SSL_SESSION with a given ID,
7841 so if there's a conflict, we now throw out the old one to achieve
7845 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7846 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7847 some routines that use cipher OIDs: some ciphers do not have OIDs
7848 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7852 *) Simplify the trust setting structure and code. Now we just have
7853 two sequences of OIDs for trusted and rejected settings. These will
7854 typically have values the same as the extended key usage extension
7855 and any application specific purposes.
7857 The trust checking code now has a default behaviour: it will just
7858 check for an object with the same NID as the passed id. Functions can
7859 be provided to override either the default behaviour or the behaviour
7860 for a given id. SSL client, server and email already have functions
7861 in place for compatibility: they check the NID and also return "trusted"
7862 if the certificate is self signed.
7865 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7866 traditional format into an EVP_PKEY structure.
7869 *) Add a password callback function PEM_cb() which either prompts for
7870 a password if usr_data is NULL or otherwise assumes it is a null
7871 terminated password. Allow passwords to be passed on command line
7872 environment or config files in a few more utilities.
7875 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7876 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7877 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7878 Update documentation.
7881 *) Support for ASN1 "NULL" type. This could be handled before by using
7882 ASN1_TYPE but there wasn't any function that would try to read a NULL
7883 and produce an error if it couldn't. For compatibility we also have
7884 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7885 don't allocate anything because they don't need to.
7888 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7890 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7892 *) Rebuild of the memory allocation routines used by OpenSSL code and
7893 possibly others as well. The purpose is to make an interface that
7894 provide hooks so anyone can build a separate set of allocation and
7895 deallocation routines to be used by OpenSSL, for example memory
7896 pool implementations, or something else, which was previously hard
7897 since Malloc(), Realloc() and Free() were defined as macros having
7898 the values malloc, realloc and free, respectively (except for Win32
7899 compilations). The same is provided for memory debugging code.
7900 OpenSSL already comes with functionality to find memory leaks, but
7901 this gives people a chance to debug other memory problems.
7903 With these changes, a new set of functions and macros have appeared:
7905 CRYPTO_set_mem_debug_functions() [F]
7906 CRYPTO_get_mem_debug_functions() [F]
7907 CRYPTO_dbg_set_options() [F]
7908 CRYPTO_dbg_get_options() [F]
7909 CRYPTO_malloc_debug_init() [M]
7911 The memory debug functions are NULL by default, unless the library
7912 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7913 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7914 gives the standard debugging functions that come with OpenSSL) or
7915 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7916 provided by the library user) must be used. When the standard
7917 debugging functions are used, CRYPTO_dbg_set_options can be used to
7918 request additional information:
7919 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7920 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7922 Also, things like CRYPTO_set_mem_functions will always give the
7923 expected result (the new set of functions is used for allocation
7924 and deallocation) at all times, regardless of platform and compiler
7927 To finish it up, some functions that were never use in any other
7928 way than through macros have a new API and new semantic:
7931 CRYPTO_dbg_realloc()
7934 All macros of value have retained their old syntax.
7935 [Richard Levitte and Bodo Moeller]
7937 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7938 ordering of SMIMECapabilities wasn't in "strength order" and there
7939 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7943 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7944 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7945 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7947 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7948 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7949 functionality to handle multipart/signed properly) and a utility
7950 called 'smime' to call all this stuff. This is based on code I
7951 originally wrote for Celo who have kindly allowed it to be
7952 included in OpenSSL.
7955 *) Add variants des_set_key_checked and des_set_key_unchecked of
7956 des_set_key (aka des_key_sched). Global variable des_check_key
7957 decides which of these is called by des_set_key; this way
7958 des_check_key behaves as it always did, but applications and
7959 the library itself, which was buggy for des_check_key == 1,
7960 have a cleaner way to pick the version they need.
7963 *) New function PKCS12_newpass() which changes the password of a
7967 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7968 dynamic mix. In both cases the ids can be used as an index into the
7969 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7970 functions so they accept a list of the field values and the
7971 application doesn't need to directly manipulate the X509_TRUST
7975 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7979 *) Modify the way the V3 extension code looks up extensions. This now
7980 works in a similar way to the object code: we have some "standard"
7981 extensions in a static table which is searched with OBJ_bsearch()
7982 and the application can add dynamic ones if needed. The file
7983 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7984 updated whenever a new extension is added to the core code and kept
7985 in ext_nid order. There is a simple program 'tabtest.c' which checks
7986 this. New extensions are not added too often so this file can readily
7987 be maintained manually.
7989 There are two big advantages in doing things this way. The extensions
7990 can be looked up immediately and no longer need to be "added" using
7991 X509V3_add_standard_extensions(): this function now does nothing.
7992 [Side note: I get *lots* of email saying the extension code doesn't
7993 work because people forget to call this function]
7994 Also no dynamic allocation is done unless new extensions are added:
7995 so if we don't add custom extensions there is no need to call
7996 X509V3_EXT_cleanup().
7999 *) Modify enc utility's salting as follows: make salting the default. Add a
8000 magic header, so unsalted files fail gracefully instead of just decrypting
8001 to garbage. This is because not salting is a big security hole, so people
8002 should be discouraged from doing it.
8005 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8006 digest to be passed on the command line but it only used this
8007 parameter when signing a certificate. Modified so all relevant
8008 operations are affected by the digest parameter including the
8009 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8010 DSA key was used because it didn't fix the digest.
8013 *) Initial certificate chain verify code. Currently tests the untrusted
8014 certificates for consistency with the verify purpose (which is set
8015 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8017 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8018 this is because it will reject chains with invalid extensions whereas
8019 every previous version of OpenSSL and SSLeay made no checks at all.
8021 Trust code: checks the root CA for the relevant trust settings. Trust
8022 settings have an initial value consistent with the verify purpose: e.g.
8023 if the verify purpose is for SSL client use it expects the CA to be
8024 trusted for SSL client use. However the default value can be changed to
8025 permit custom trust settings: one example of this would be to only trust
8026 certificates from a specific "secure" set of CAs.
8028 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8029 which should be used for version portability: especially since the
8030 verify structure is likely to change more often now.
8032 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8033 to set them. If not set then assume SSL clients will verify SSL servers
8036 Two new options to the verify program: -untrusted allows a set of
8037 untrusted certificates to be passed in and -purpose which sets the
8038 intended purpose of the certificate. If a purpose is set then the
8039 new chain verify code is used to check extension consistency.
8042 *) Support for the authority information access extension.
8045 *) Modify RSA and DSA PEM read routines to transparently handle
8046 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8047 public keys in a format compatible with certificate
8048 SubjectPublicKeyInfo structures. Unfortunately there were already
8049 functions called *_PublicKey_* which used various odd formats so
8050 these are retained for compatibility: however the DSA variants were
8051 never in a public release so they have been deleted. Changed dsa/rsa
8052 utilities to handle the new format: note no releases ever handled public
8053 keys so we should be OK.
8055 The primary motivation for this change is to avoid the same fiasco
8056 that dogs private keys: there are several incompatible private key
8057 formats some of which are standard and some OpenSSL specific and
8058 require various evil hacks to allow partial transparent handling and
8059 even then it doesn't work with DER formats. Given the option anything
8060 other than PKCS#8 should be dumped: but the other formats have to
8061 stay in the name of compatibility.
8063 With public keys and the benefit of hindsight one standard format
8064 is used which works with EVP_PKEY, RSA or DSA structures: though
8065 it clearly returns an error if you try to read the wrong kind of key.
8067 Added a -pubkey option to the 'x509' utility to output the public key.
8068 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8069 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8070 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8071 that do the same as the EVP_PKEY_assign_*() except they up the
8072 reference count of the added key (they don't "swallow" the
8076 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8077 CRLs would fail if the file contained no certificates or no CRLs:
8078 added a new function to read in both types and return the number
8079 read: this means that if none are read it will be an error. The
8080 DER versions of the certificate and CRL reader would always fail
8081 because it isn't possible to mix certificates and CRLs in DER format
8082 without choking one or the other routine. Changed this to just read
8083 a certificate: this is the best we can do. Also modified the code
8084 in apps/verify.c to take notice of return codes: it was previously
8085 attempting to read in certificates from NULL pointers and ignoring
8086 any errors: this is one reason why the cert and CRL reader seemed
8087 to work. It doesn't check return codes from the default certificate
8088 routines: these may well fail if the certificates aren't installed.
8091 *) Code to support otherName option in GeneralName.
8094 *) First update to verify code. Change the verify utility
8095 so it warns if it is passed a self signed certificate:
8096 for consistency with the normal behaviour. X509_verify
8097 has been modified to it will now verify a self signed
8098 certificate if *exactly* the same certificate appears
8099 in the store: it was previously impossible to trust a
8100 single self signed certificate. This means that:
8101 openssl verify ss.pem
8102 now gives a warning about a self signed certificate but
8103 openssl verify -CAfile ss.pem ss.pem
8107 *) For servers, store verify_result in SSL_SESSION data structure
8108 (and add it to external session representation).
8109 This is needed when client certificate verifications fails,
8110 but an application-provided verification callback (set by
8111 SSL_CTX_set_cert_verify_callback) allows accepting the session
8112 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8113 but returns 1): When the session is reused, we have to set
8114 ssl->verify_result to the appropriate error code to avoid
8116 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8118 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8119 case in PKCS7_dataInit() where the signed PKCS7 structure
8120 didn't contain any existing data because it was being created.
8121 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8123 *) Add a salt to the key derivation routines in enc.c. This
8124 forms the first 8 bytes of the encrypted file. Also add a
8125 -S option to allow a salt to be input on the command line.
8128 *) New function X509_cmp(). Oddly enough there wasn't a function
8129 to compare two certificates. We do this by working out the SHA1
8130 hash and comparing that. X509_cmp() will be needed by the trust
8134 *) SSL_get1_session() is like SSL_get_session(), but increments
8135 the reference count in the SSL_SESSION returned.
8136 [Geoff Thorpe <geoff@eu.c2.net>]
8138 *) Fix for 'req': it was adding a null to request attributes.
8139 Also change the X509_LOOKUP and X509_INFO code to handle
8140 certificate auxiliary information.
8143 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8147 *) Add the possibility to add extra information to the memory leak
8148 detecting output, to form tracebacks, showing from where each
8149 allocation was originated: CRYPTO_push_info("constant string") adds
8150 the string plus current file name and line number to a per-thread
8151 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8152 is like calling CYRPTO_pop_info() until the stack is empty.
8153 Also updated memory leak detection code to be multi-thread-safe.
8156 *) Add options -text and -noout to pkcs7 utility and delete the
8157 encryption options which never did anything. Update docs.
8160 *) Add options to some of the utilities to allow the pass phrase
8161 to be included on either the command line (not recommended on
8162 OSes like Unix) or read from the environment. Update the
8163 manpages and fix a few bugs.
8166 *) Add a few manpages for some of the openssl commands.
8169 *) Fix the -revoke option in ca. It was freeing up memory twice,
8170 leaking and not finding already revoked certificates.
8173 *) Extensive changes to support certificate auxiliary information.
8174 This involves the use of X509_CERT_AUX structure and X509_AUX
8175 functions. An X509_AUX function such as PEM_read_X509_AUX()
8176 can still read in a certificate file in the usual way but it
8177 will also read in any additional "auxiliary information". By
8178 doing things this way a fair degree of compatibility can be
8179 retained: existing certificates can have this information added
8180 using the new 'x509' options.
8182 Current auxiliary information includes an "alias" and some trust
8183 settings. The trust settings will ultimately be used in enhanced
8184 certificate chain verification routines: currently a certificate
8185 can only be trusted if it is self signed and then it is trusted
8189 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8190 The problem was that one of the replacement routines had not been working
8191 since SSLeay releases. For now the offending routine has been replaced
8192 with non-optimised assembler. Even so, this now gives around 95%
8193 performance improvement for 1024 bit RSA signs.
8196 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8197 handling. Most clients have the effective key size in bits equal to
8198 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8199 A few however don't do this and instead use the size of the decrypted key
8200 to determine the RC2 key length and the AlgorithmIdentifier to determine
8201 the effective key length. In this case the effective key length can still
8202 be 40 bits but the key length can be 168 bits for example. This is fixed
8203 by manually forcing an RC2 key into the EVP_PKEY structure because the
8204 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8205 the key length and effective key length are equal.
8208 *) Add a bunch of functions that should simplify the creation of
8209 X509_NAME structures. Now you should be able to do:
8210 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8211 and have it automatically work out the correct field type and fill in
8212 the structures. The more adventurous can try:
8213 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8214 and it will (hopefully) work out the correct multibyte encoding.
8217 *) Change the 'req' utility to use the new field handling and multibyte
8218 copy routines. Before the DN field creation was handled in an ad hoc
8219 way in req, ca, and x509 which was rather broken and didn't support
8220 BMPStrings or UTF8Strings. Since some software doesn't implement
8221 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8222 using the dirstring_type option. See the new comment in the default
8223 openssl.cnf for more info.
8226 *) Make crypto/rand/md_rand.c more robust:
8227 - Assure unique random numbers after fork().
8228 - Make sure that concurrent threads access the global counter and
8229 md serializably so that we never lose entropy in them
8230 or use exactly the same state in multiple threads.
8231 Access to the large state is not always serializable because
8232 the additional locking could be a performance killer, and
8233 md should be large enough anyway.
8236 *) New file apps/app_rand.c with commonly needed functionality
8237 for handling the random seed file.
8239 Use the random seed file in some applications that previously did not:
8241 dsaparam -genkey (which also ignored its '-rand' option),
8244 x509 (when signing).
8245 Except on systems with /dev/urandom, it is crucial to have a random
8246 seed file at least for key creation, DSA signing, and for DH exchanges;
8247 for RSA signatures we could do without one.
8249 gendh and gendsa (unlike genrsa) used to read only the first byte
8250 of each file listed in the '-rand' option. The function as previously
8251 found in genrsa is now in app_rand.c and is used by all programs
8252 that support '-rand'.
8255 *) In RAND_write_file, use mode 0600 for creating files;
8256 don't just chmod when it may be too late.
8259 *) Report an error from X509_STORE_load_locations
8260 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8263 *) New function ASN1_mbstring_copy() this copies a string in either
8264 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8265 into an ASN1_STRING type. A mask of permissible types is passed
8266 and it chooses the "minimal" type to use or an error if not type
8270 *) Add function equivalents to the various macros in asn1.h. The old
8271 macros are retained with an M_ prefix. Code inside the library can
8272 use the M_ macros. External code (including the openssl utility)
8273 should *NOT* in order to be "shared library friendly".
8276 *) Add various functions that can check a certificate's extensions
8277 to see if it usable for various purposes such as SSL client,
8278 server or S/MIME and CAs of these types. This is currently
8279 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8280 verification. Also added a -purpose flag to x509 utility to
8281 print out all the purposes.
8284 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8288 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8289 for, obtain and decode and extension and obtain its critical flag.
8290 This allows all the necessary extension code to be handled in a
8291 single function call.
8294 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8295 platforms. See crypto/rc4/rc4_enc.c for further details.
8298 *) New -noout option to asn1parse. This causes no output to be produced
8299 its main use is when combined with -strparse and -out to extract data
8300 from a file (which may not be in ASN.1 format).
8303 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8304 when producing the local key id.
8305 [Richard Levitte <levitte@stacken.kth.se>]
8307 *) New option -dhparam in s_server. This allows a DH parameter file to be
8308 stated explicitly. If it is not stated then it tries the first server
8309 certificate file. The previous behaviour hard coded the filename
8313 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8314 a public key to be input or output. For example:
8315 openssl rsa -in key.pem -pubout -out pubkey.pem
8316 Also added necessary DSA public key functions to handle this.
8319 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8320 in the message. This was handled by allowing
8321 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8322 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8324 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8325 to the end of the strings whereas this didn't. This would cause problems
8326 if strings read with d2i_ASN1_bytes() were later modified.
8327 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8329 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8330 data and it contains EOF it will end up returning an error. This is
8331 caused by input 46 bytes long. The cause is due to the way base64
8332 BIOs find the start of base64 encoded data. They do this by trying a
8333 trial decode on each line until they find one that works. When they
8334 do a flag is set and it starts again knowing it can pass all the
8335 data directly through the decoder. Unfortunately it doesn't reset
8336 the context it uses. This means that if EOF is reached an attempt
8337 is made to pass two EOFs through the context and this causes the
8338 resulting error. This can also cause other problems as well. As is
8339 usual with these problems it takes *ages* to find and the fix is
8340 trivial: move one line.
8341 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8343 *) Ugly workaround to get s_client and s_server working under Windows. The
8344 old code wouldn't work because it needed to select() on sockets and the
8345 tty (for keypresses and to see if data could be written). Win32 only
8346 supports select() on sockets so we select() with a 1s timeout on the
8347 sockets and then see if any characters are waiting to be read, if none
8348 are present then we retry, we also assume we can always write data to
8349 the tty. This isn't nice because the code then blocks until we've
8350 received a complete line of data and it is effectively polling the
8351 keyboard at 1s intervals: however it's quite a bit better than not
8352 working at all :-) A dedicated Windows application might handle this
8353 with an event loop for example.
8356 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8357 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8358 will be called when RSA_sign() and RSA_verify() are used. This is useful
8359 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8360 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8361 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8362 This necessitated the support of an extra signature type NID_md5_sha1
8363 for SSL signatures and modifications to the SSL library to use it instead
8364 of calling RSA_public_decrypt() and RSA_private_encrypt().
8367 *) Add new -verify -CAfile and -CApath options to the crl program, these
8368 will lookup a CRL issuers certificate and verify the signature in a
8369 similar way to the verify program. Tidy up the crl program so it
8370 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8371 less strict. It will now permit CRL extensions even if it is not
8372 a V2 CRL: this will allow it to tolerate some broken CRLs.
8375 *) Initialize all non-automatic variables each time one of the openssl
8376 sub-programs is started (this is necessary as they may be started
8377 multiple times from the "OpenSSL>" prompt).
8378 [Lennart Bang, Bodo Moeller]
8380 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8381 removing all other RSA functionality (this is what NO_RSA does). This
8382 is so (for example) those in the US can disable those operations covered
8383 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8387 *) Non-copying interface to BIO pairs.
8388 (still largely untested)
8391 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8392 ASCII string. This was handled independently in various places before.
8395 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8396 UTF8 strings a character at a time.
8399 *) Use client_version from client hello to select the protocol
8400 (s23_srvr.c) and for RSA client key exchange verification
8401 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8404 *) Add various utility functions to handle SPKACs, these were previously
8405 handled by poking round in the structure internals. Added new function
8406 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8407 print, verify and generate SPKACs. Based on an original idea from
8408 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8411 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8414 *) Allow the config file extension section to be overwritten on the
8415 command line. Based on an original idea from Massimiliano Pala
8416 <madwolf@comune.modena.it>. The new option is called -extensions
8417 and can be applied to ca, req and x509. Also -reqexts to override
8418 the request extensions in req and -crlexts to override the crl extensions
8422 *) Add new feature to the SPKAC handling in ca. Now you can include
8423 the same field multiple times by preceding it by "XXXX." for example:
8426 this is the same syntax as used in the req config file.
8429 *) Allow certificate extensions to be added to certificate requests. These
8430 are specified in a 'req_extensions' option of the req section of the
8431 config file. They can be printed out with the -text option to req but
8432 are otherwise ignored at present.
8435 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8436 data read consists of only the final block it would not decrypted because
8437 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8438 A misplaced 'break' also meant the decrypted final block might not be
8439 copied until the next read.
8442 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8443 a few extra parameters to the DH structure: these will be useful if
8444 for example we want the value of 'q' or implement X9.42 DH.
8447 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8448 provides hooks that allow the default DSA functions or functions on a
8449 "per key" basis to be replaced. This allows hardware acceleration and
8450 hardware key storage to be handled without major modification to the
8451 library. Also added low level modexp hooks and CRYPTO_EX structure and
8452 associated functions.
8455 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8456 as "read only": it can't be written to and the buffer it points to will
8457 not be freed. Reading from a read only BIO is much more efficient than
8458 a normal memory BIO. This was added because there are several times when
8459 an area of memory needs to be read from a BIO. The previous method was
8460 to create a memory BIO and write the data to it, this results in two
8461 copies of the data and an O(n^2) reading algorithm. There is a new
8462 function BIO_new_mem_buf() which creates a read only memory BIO from
8463 an area of memory. Also modified the PKCS#7 routines to use read only
8467 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8468 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8469 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8470 but a retry condition occured while trying to read the rest.
8473 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8474 NID_pkcs7_encrypted by default: this was wrong since this should almost
8475 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8476 the encrypted data type: this is a more sensible place to put it and it
8477 allows the PKCS#12 code to be tidied up that duplicated this
8481 *) Changed obj_dat.pl script so it takes its input and output files on
8482 the command line. This should avoid shell escape redirection problems
8486 *) Initial support for certificate extension requests, these are included
8487 in things like Xenroll certificate requests. Included functions to allow
8488 extensions to be obtained and added.
8491 *) -crlf option to s_client and s_server for sending newlines as
8492 CRLF (as required by many protocols).
8495 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8497 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8498 [Ralf S. Engelschall]
8500 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8501 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8503 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8507 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8508 DH parameters/keys (q is lost during that conversion, but the resulting
8509 DH parameters contain its length).
8511 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8512 much faster than DH_generate_parameters (which creates parameters
8513 where p = 2*q + 1), and also the smaller q makes DH computations
8514 much more efficient (160-bit exponentiation instead of 1024-bit
8515 exponentiation); so this provides a convenient way to support DHE
8516 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8517 utter importance to use
8518 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8520 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8521 when such DH parameters are used, because otherwise small subgroup
8522 attacks may become possible!
8525 *) Avoid memory leak in i2d_DHparams.
8528 *) Allow the -k option to be used more than once in the enc program:
8529 this allows the same encrypted message to be read by multiple recipients.
8532 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8533 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8534 it will always use the numerical form of the OID, even if it has a short
8538 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8539 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8540 otherwise bn_mod_exp was called. In the case of hardware keys for example
8541 no private key components need be present and it might store extra data
8542 in the RSA structure, which cannot be accessed from bn_mod_exp.
8543 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8544 private key operations.
8547 *) Added support for SPARC Linux.
8550 *) pem_password_cb function type incompatibly changed from
8551 typedef int pem_password_cb(char *buf, int size, int rwflag);
8553 ....(char *buf, int size, int rwflag, void *userdata);
8554 so that applications can pass data to their callbacks:
8555 The PEM[_ASN1]_{read,write}... functions and macros now take an
8556 additional void * argument, which is just handed through whenever
8557 the password callback is called.
8558 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8560 New function SSL_CTX_set_default_passwd_cb_userdata.
8562 Compatibility note: As many C implementations push function arguments
8563 onto the stack in reverse order, the new library version is likely to
8564 interoperate with programs that have been compiled with the old
8565 pem_password_cb definition (PEM_whatever takes some data that
8566 happens to be on the stack as its last argument, and the callback
8567 just ignores this garbage); but there is no guarantee whatsoever that
8570 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8571 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8572 problems not only on Windows, but also on some Unix platforms.
8573 To avoid problematic command lines, these definitions are now in an
8574 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8575 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8578 *) MIPS III/IV assembler module is reimplemented.
8581 *) More DES library cleanups: remove references to srand/rand and
8582 delete an unused file.
8585 *) Add support for the the free Netwide assembler (NASM) under Win32,
8586 since not many people have MASM (ml) and it can be hard to obtain.
8587 This is currently experimental but it seems to work OK and pass all
8588 the tests. Check out INSTALL.W32 for info.
8591 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8592 without temporary keys kept an extra copy of the server key,
8593 and connections with temporary keys did not free everything in case
8597 *) New function RSA_check_key and new openssl rsa option -check
8598 for verifying the consistency of RSA keys.
8599 [Ulf Moeller, Bodo Moeller]
8601 *) Various changes to make Win32 compile work:
8602 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8603 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8604 comparison" warnings.
8605 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8608 *) Add a debugging option to PKCS#5 v2 key generation function: when
8609 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8610 derived keys are printed to stderr.
8613 *) Copy the flags in ASN1_STRING_dup().
8614 [Roman E. Pavlov <pre@mo.msk.ru>]
8616 *) The x509 application mishandled signing requests containing DSA
8617 keys when the signing key was also DSA and the parameters didn't match.
8619 It was supposed to omit the parameters when they matched the signing key:
8620 the verifying software was then supposed to automatically use the CA's
8621 parameters if they were absent from the end user certificate.
8623 Omitting parameters is no longer recommended. The test was also
8624 the wrong way round! This was probably due to unusual behaviour in
8625 EVP_cmp_parameters() which returns 1 if the parameters match.
8626 This meant that parameters were omitted when they *didn't* match and
8627 the certificate was useless. Certificates signed with 'ca' didn't have
8629 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8631 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8632 The interface is as follows:
8633 Applications can use
8634 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8635 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8636 "off" is now the default.
8637 The library internally uses
8638 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8639 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8640 to disable memory-checking temporarily.
8642 Some inconsistent states that previously were possible (and were
8643 even the default) are now avoided.
8645 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8646 with each memory chunk allocated; this is occasionally more helpful
8647 than just having a counter.
8649 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8651 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8655 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8656 which largely parallels "options", but is for changing API behaviour,
8657 whereas "options" are about protocol behaviour.
8658 Initial "mode" flags are:
8660 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8661 a single record has been written.
8662 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8663 retries use the same buffer location.
8664 (But all of the contents must be
8668 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8671 *) Fix problems with no-hmac etc.
8672 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8674 *) New functions RSA_get_default_method(), RSA_set_method() and
8675 RSA_get_method(). These allows replacement of RSA_METHODs without having
8676 to mess around with the internals of an RSA structure.
8679 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8680 Also really enable memory leak checks in openssl.c and in some
8682 [Chad C. Mulligan, Bodo Moeller]
8684 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8685 up the length of negative integers. This has now been simplified to just
8686 store the length when it is first determined and use it later, rather
8687 than trying to keep track of where data is copied and updating it to
8689 [Steve Henson, reported by Brien Wheeler
8690 <bwheeler@authentica-security.com>]
8692 *) Add a new function PKCS7_signatureVerify. This allows the verification
8693 of a PKCS#7 signature but with the signing certificate passed to the
8694 function itself. This contrasts with PKCS7_dataVerify which assumes the
8695 certificate is present in the PKCS#7 structure. This isn't always the
8696 case: certificates can be omitted from a PKCS#7 structure and be
8697 distributed by "out of band" means (such as a certificate database).
8700 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8701 function prototypes in pem.h, also change util/mkdef.pl to add the
8702 necessary function names.
8705 *) mk1mf.pl (used by Windows builds) did not properly read the
8706 options set by Configure in the top level Makefile, and Configure
8707 was not even able to write more than one option correctly.
8708 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8711 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8712 file to be loaded from a BIO or FILE pointer. The BIO version will
8713 for example allow memory BIOs to contain config info.
8716 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8717 Whoever hopes to achieve shared-library compatibility across versions
8718 must use this, not the compile-time macro.
8719 (Exercise 0.9.4: Which is the minimum library version required by
8721 Note: All this applies only to multi-threaded programs, others don't
8725 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8726 through a BIO pair triggered the default case, i.e.
8727 SSLerr(...,SSL_R_UNKNOWN_STATE).
8730 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8731 can use the SSL library even if none of the specific BIOs is
8735 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8736 for the encoded length.
8737 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8739 *) Add initial documentation of the X509V3 functions.
8742 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8743 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8744 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8745 secure PKCS#8 private key format with a high iteration count.
8748 *) Fix determination of Perl interpreter: A perl or perl5
8749 _directory_ in $PATH was also accepted as the interpreter.
8750 [Ralf S. Engelschall]
8752 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8753 wrong with it but it was very old and did things like calling
8754 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8758 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8759 to use the new extension code.
8762 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8763 with macros. This should make it easier to change their form, add extra
8764 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8768 *) Add to configuration table a new entry that can specify an alternative
8769 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8770 according to Mark Crispin <MRC@Panda.COM>.
8774 *) DES CBC did not update the IV. Weird.
8777 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8778 Changing the behaviour of the former might break existing programs --
8779 where IV updating is needed, des_ncbc_encrypt can be used.
8782 *) When bntest is run from "make test" it drives bc to check its
8783 calculations, as well as internally checking them. If an internal check
8784 fails, it needs to cause bc to give a non-zero result or make test carries
8785 on without noticing the failure. Fixed.
8788 *) DES library cleanups.
8791 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8792 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8793 ciphers. NOTE: although the key derivation function has been verified
8794 against some published test vectors it has not been extensively tested
8795 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8799 *) Instead of "mkdir -p", which is not fully portable, use new
8800 Perl script "util/mkdir-p.pl".
8803 *) Rewrite the way password based encryption (PBE) is handled. It used to
8804 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8805 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8806 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8807 the 'parameter' field of the AlgorithmIdentifier is passed to the
8808 underlying key generation function so it must do its own ASN1 parsing.
8809 This has also changed the EVP_PBE_CipherInit() function which now has a
8810 'parameter' argument instead of literal salt and iteration count values
8811 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8814 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8815 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8816 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8817 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8818 value was just used as a "magic string" and not used directly its
8819 value doesn't matter.
8822 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8826 *) "linux-sparc64" configuration (ultrapenguin).
8827 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8828 "linux-sparc" configuration.
8829 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8831 *) config now generates no-xxx options for missing ciphers.
8834 *) Support the EBCDIC character set (work in progress).
8835 File ebcdic.c not yet included because it has a different license.
8836 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8838 *) Support BS2000/OSD-POSIX.
8839 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8841 *) Make callbacks for key generation use void * instead of char *.
8844 *) Make S/MIME samples compile (not yet tested).
8847 *) Additional typesafe stacks.
8850 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8854 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8856 *) New configuration variant "sco5-gcc".
8858 *) Updated some demos.
8859 [Sean O Riordain, Wade Scholine]
8861 *) Add missing BIO_free at exit of pkcs12 application.
8864 *) Fix memory leak in conf.c.
8867 *) Updates for Win32 to assembler version of MD5.
8870 *) Set #! path to perl in apps/der_chop to where we found it
8871 instead of using a fixed path.
8874 *) SHA library changes for irix64-mips4-cc.
8877 *) Improvements for VMS support.
8881 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8883 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8884 This also avoids the problems with SC4.2 and unpatched SC5.
8885 [Andy Polyakov <appro@fy.chalmers.se>]
8887 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8888 These are required because of the typesafe stack would otherwise break
8889 existing code. If old code used a structure member which used to be STACK
8890 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8891 sk_num or sk_value it would produce an error because the num, data members
8892 are not present in STACK_OF. Now it just produces a warning. sk_set
8893 replaces the old method of assigning a value to sk_value
8894 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8895 that does this will no longer work (and should use sk_set instead) but
8896 this could be regarded as a "questionable" behaviour anyway.
8899 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8900 correctly handle encrypted S/MIME data.
8903 *) Change type of various DES function arguments from des_cblock
8904 (which means, in function argument declarations, pointer to char)
8905 to des_cblock * (meaning pointer to array with 8 char elements),
8906 which allows the compiler to do more typechecking; it was like
8907 that back in SSLeay, but with lots of ugly casts.
8909 Introduce new type const_des_cblock.
8912 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8913 problems: find RecipientInfo structure that matches recipient certificate
8914 and initialise the ASN1 structures properly based on passed cipher.
8917 *) Belatedly make the BN tests actually check the results.
8920 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8921 to and from BNs: it was completely broken. New compilation option
8922 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8923 key elements as negative integers.
8926 *) Reorganize and speed up MD5.
8927 [Andy Polyakov <appro@fy.chalmers.se>]
8930 [Richard Levitte <richard@levitte.org>]
8932 *) New option -out to asn1parse to allow the parsed structure to be
8933 output to a file. This is most useful when combined with the -strparse
8934 option to examine the output of things like OCTET STRINGS.
8937 *) Make SSL library a little more fool-proof by not requiring any longer
8938 that SSL_set_{accept,connect}_state be called before
8939 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8940 in many applications because usually everything *appeared* to work as
8941 intended anyway -- now it really works as intended).
8944 *) Move openssl.cnf out of lib/.
8947 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8948 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8949 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8950 [Ralf S. Engelschall]
8952 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8953 handle PKCS#7 enveloped data properly.
8954 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8956 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8957 copying pointers. The cert_st handling is changed by this in
8958 various ways (and thus what used to be known as ctx->default_cert
8959 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8960 any longer when s->cert does not give us what we need).
8961 ssl_cert_instantiate becomes obsolete by this change.
8962 As soon as we've got the new code right (possibly it already is?),
8963 we have solved a couple of bugs of the earlier code where s->cert
8964 was used as if it could not have been shared with other SSL structures.
8966 Note that using the SSL API in certain dirty ways now will result
8967 in different behaviour than observed with earlier library versions:
8968 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8969 does not influence s as it used to.
8971 In order to clean up things more thoroughly, inside SSL_SESSION
8972 we don't use CERT any longer, but a new structure SESS_CERT
8973 that holds per-session data (if available); currently, this is
8974 the peer's certificate chain and, for clients, the server's certificate
8975 and temporary key. CERT holds only those values that can have
8976 meaningful defaults in an SSL_CTX.
8979 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8980 from the internal representation. Various PKCS#7 fixes: remove some
8981 evil casts and set the enc_dig_alg field properly based on the signing
8985 *) Allow PKCS#12 password to be set from the command line or the
8986 environment. Let 'ca' get its config file name from the environment
8987 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8991 *) Allow certificate policies extension to use an IA5STRING for the
8992 organization field. This is contrary to the PKIX definition but
8993 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8997 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8998 without disallowing inline assembler and the like for non-pedantic builds.
9001 *) Support Borland C++ builder.
9002 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9007 *) SHA-1 cleanups and performance enhancements.
9008 [Andy Polyakov <appro@fy.chalmers.se>]
9010 *) Sparc v8plus assembler for the bignum library.
9011 [Andy Polyakov <appro@fy.chalmers.se>]
9013 *) Accept any -xxx and +xxx compiler options in Configure.
9016 *) Update HPUX configuration.
9019 *) Add missing sk_<type>_unshift() function to safestack.h
9020 [Ralf S. Engelschall]
9022 *) New function SSL_CTX_use_certificate_chain_file that sets the
9023 "extra_cert"s in addition to the certificate. (This makes sense
9024 only for "PEM" format files, as chains as a whole are not
9028 *) Support verify_depth from the SSL API.
9029 x509_vfy.c had what can be considered an off-by-one-error:
9030 Its depth (which was not part of the external interface)
9031 was actually counting the number of certificates in a chain;
9032 now it really counts the depth.
9035 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9036 instead of X509err, which often resulted in confusing error
9037 messages since the error codes are not globally unique
9038 (e.g. an alleged error in ssl3_accept when a certificate
9039 didn't match the private key).
9041 *) New function SSL_CTX_set_session_id_context that allows to set a default
9042 value (so that you don't need SSL_set_session_id_context for each
9043 connection using the SSL_CTX).
9046 *) OAEP decoding bug fix.
9049 *) Support INSTALL_PREFIX for package builders, as proposed by
9053 *) New Configure options "threads" and "no-threads". For systems
9054 where the proper compiler options are known (currently Solaris
9055 and Linux), "threads" is the default.
9058 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9061 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9062 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9063 such as /usr/local/bin.
9066 *) "make linux-shared" to build shared libraries.
9067 [Niels Poppe <niels@netbox.org>]
9069 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9072 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9073 extension adding in x509 utility.
9076 *) Remove NOPROTO sections and error code comments.
9079 *) Partial rewrite of the DEF file generator to now parse the ANSI
9083 *) New Configure options --prefix=DIR and --openssldir=DIR.
9086 *) Complete rewrite of the error code script(s). It is all now handled
9087 by one script at the top level which handles error code gathering,
9088 header rewriting and C source file generation. It should be much better
9089 than the old method: it now uses a modified version of Ulf's parser to
9090 read the ANSI prototypes in all header files (thus the old K&R definitions
9091 aren't needed for error creation any more) and do a better job of
9092 translating function codes into names. The old 'ASN1 error code imbedded
9093 in a comment' is no longer necessary and it doesn't use .err files which
9094 have now been deleted. Also the error code call doesn't have to appear all
9095 on one line (which resulted in some large lines...).
9098 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9101 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9102 0 (which usually indicates a closed connection), but continue reading.
9105 *) Fix some race conditions.
9108 *) Add support for CRL distribution points extension. Add Certificate
9109 Policies and CRL distribution points documentation.
9112 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9115 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9116 8 of keying material. Merlin has also confirmed interop with this fix
9117 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9118 [Merlin Hughes <merlin@baltimore.ie>]
9120 *) Fix lots of warnings.
9121 [Richard Levitte <levitte@stacken.kth.se>]
9123 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9124 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9125 [Richard Levitte <levitte@stacken.kth.se>]
9127 *) Fix problems with sizeof(long) == 8.
9128 [Andy Polyakov <appro@fy.chalmers.se>]
9130 *) Change functions to ANSI C.
9133 *) Fix typos in error codes.
9134 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9136 *) Remove defunct assembler files from Configure.
9139 *) SPARC v8 assembler BIGNUM implementation.
9140 [Andy Polyakov <appro@fy.chalmers.se>]
9142 *) Support for Certificate Policies extension: both print and set.
9143 Various additions to support the r2i method this uses.
9146 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9147 return a const string when you are expecting an allocated buffer.
9150 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9151 types DirectoryString and DisplayText.
9154 *) Add code to allow r2i extensions to access the configuration database,
9155 add an LHASH database driver and add several ctx helper functions.
9158 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9159 fail when they extended the size of a BIGNUM.
9162 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9163 support typesafe stack.
9166 *) Fix typo in SSL_[gs]et_options().
9167 [Nils Frostberg <nils@medcom.se>]
9169 *) Delete various functions and files that belonged to the (now obsolete)
9170 old X509V3 handling code.
9173 *) New Configure option "rsaref".
9176 *) Don't auto-generate pem.h.
9179 *) Introduce type-safe ASN.1 SETs.
9182 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9183 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9185 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9186 that links with OpenSSL (well at least cause lots of warnings), but fear
9187 not: the conversion is trivial, and it eliminates loads of evil casts. A
9188 few STACKed things have been converted already. Feel free to convert more.
9189 In the fullness of time, I'll do away with the STACK type altogether.
9192 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9193 specified in <certfile> by updating the entry in the index.txt file.
9194 This way one no longer has to edit the index.txt file manually for
9195 revoking a certificate. The -revoke option does the gory details now.
9196 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9198 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9199 `-text' option at all and this way the `-noout -text' combination was
9200 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9201 [Ralf S. Engelschall]
9203 *) Make sure a corresponding plain text error message exists for the
9204 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9205 verify callback function determined that a certificate was revoked.
9206 [Ralf S. Engelschall]
9208 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9209 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9210 all available cipers including rc5, which was forgotten until now.
9211 In order to let the testing shell script know which algorithms
9212 are available, a new (up to now undocumented) command
9213 "openssl list-cipher-commands" is used.
9216 *) Bugfix: s_client occasionally would sleep in select() when
9217 it should have checked SSL_pending() first.
9220 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9221 the raw DSA values prior to ASN.1 encoding.
9224 *) Tweaks to Configure
9225 [Niels Poppe <niels@netbox.org>]
9227 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9231 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9234 *) New config option to avoid instructions that are illegal on the 80386.
9235 The default code is faster, but requires at least a 486.
9238 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9239 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9240 same as SSL2_VERSION anyway.
9243 *) New "-showcerts" option for s_client.
9246 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9247 application. Various cleanups and fixes.
9250 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9251 modify error routines to work internally. Add error codes and PBE init
9252 to library startup routines.
9255 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9256 packing functions to asn1 and evp. Changed function names and error
9257 codes along the way.
9260 *) PKCS12 integration: and so it begins... First of several patches to
9261 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9262 objects to objects.h
9265 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9266 and display support for Thawte strong extranet extension.
9269 *) Add LinuxPPC support.
9270 [Jeff Dubrule <igor@pobox.org>]
9272 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9273 bn_div_words in alpha.s.
9274 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9276 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9277 OAEP isn't supported when OpenSSL is built with RSAref.
9278 [Ulf Moeller <ulf@fitug.de>]
9280 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9281 so they no longer are missing under -DNOPROTO.
9282 [Soren S. Jorvang <soren@t.dk>]
9285 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9287 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9288 doesn't work when the session is reused. Coming soon!
9291 *) Fix a security hole, that allows sessions to be reused in the wrong
9292 context thus bypassing client cert protection! All software that uses
9293 client certs and session caches in multiple contexts NEEDS PATCHING to
9294 allow session reuse! A fuller solution is in the works.
9295 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9297 *) Some more source tree cleanups (removed obsolete files
9298 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9299 permission on "config" script to be executable) and a fix for the INSTALL
9301 [Ulf Moeller <ulf@fitug.de>]
9303 *) Remove some legacy and erroneous uses of malloc, free instead of
9305 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9307 *) Make rsa_oaep_test return non-zero on error.
9308 [Ulf Moeller <ulf@fitug.de>]
9310 *) Add support for native Solaris shared libraries. Configure
9311 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9312 if someone would make that last step automatic.
9313 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9315 *) ctx_size was not built with the right compiler during "make links". Fixed.
9318 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9319 except NULL ciphers". This means the default cipher list will no longer
9320 enable NULL ciphers. They need to be specifically enabled e.g. with
9321 the string "DEFAULT:eNULL".
9324 *) Fix to RSA private encryption routines: if p < q then it would
9325 occasionally produce an invalid result. This will only happen with
9326 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9329 *) Be less restrictive and allow also `perl util/perlpath.pl
9330 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9331 because this way one can also use an interpreter named `perl5' (which is
9332 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9333 installed as `perl').
9334 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9336 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9337 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9339 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9340 advapi32.lib to Win32 build and change the pem test comparision
9341 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9342 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9343 and crypto/des/ede_cbcm_enc.c.
9346 *) DES quad checksum was broken on big-endian architectures. Fixed.
9349 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9350 Win32 test batch file so it (might) work again. The Win32 test batch file
9351 is horrible: I feel ill....
9354 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9355 in e_os.h. Audit of header files to check ANSI and non ANSI
9356 sections: 10 functions were absent from non ANSI section and not exported
9357 from Windows DLLs. Fixed up libeay.num for new functions.
9360 *) Make `openssl version' output lines consistent.
9361 [Ralf S. Engelschall]
9363 *) Fix Win32 symbol export lists for BIO functions: Added
9364 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9365 to ms/libeay{16,32}.def.
9366 [Ralf S. Engelschall]
9368 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9369 fine under Unix and passes some trivial tests I've now added. But the
9370 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9371 added to make sure no one expects that this stuff really works in the
9372 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9373 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9375 [Ralf S. Engelschall]
9377 *) Fix the generation of two part addresses in perl.
9378 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9380 *) Add config entry for Linux on MIPS.
9381 [John Tobey <jtobey@channel1.com>]
9383 *) Make links whenever Configure is run, unless we are on Windoze.
9386 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9387 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9391 *) Add a useful kludge to allow package maintainers to specify compiler and
9392 other platforms details on the command line without having to patch the
9393 Configure script everytime: One now can use ``perl Configure
9394 <id>:<details>'', i.e. platform ids are allowed to have details appended
9395 to them (seperated by colons). This is treated as there would be a static
9396 pre-configured entry in Configure's %table under key <id> with value
9397 <details> and ``perl Configure <id>'' is called. So, when you want to
9398 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9399 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9400 now, which overrides the FreeBSD-elf entry on-the-fly.
9401 [Ralf S. Engelschall]
9403 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9406 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9407 on the `perl Configure ...' command line. This way one can compile
9408 OpenSSL libraries with Position Independent Code (PIC) which is needed
9409 for linking it into DSOs.
9410 [Ralf S. Engelschall]
9412 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9416 *) Cleaned up the LICENSE document: The official contact for any license
9417 questions now is the OpenSSL core team under openssl-core@openssl.org.
9418 And add a paragraph about the dual-license situation to make sure people
9419 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9420 to the OpenSSL toolkit.
9421 [Ralf S. Engelschall]
9423 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9424 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9425 Additonally cleaned up the `make links' target: Remove unnecessary
9426 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9427 to speed processing and no longer clutter the display with confusing
9428 stuff. Instead only the actually done links are displayed.
9429 [Ralf S. Engelschall]
9431 *) Permit null encryption ciphersuites, used for authentication only. It used
9432 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9433 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9437 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9438 signed attributes when verifying signatures (this would break them),
9439 the detached data encoding was wrong and public keys obtained using
9440 X509_get_pubkey() weren't freed.
9443 *) Add text documentation for the BUFFER functions. Also added a work around
9444 to a Win95 console bug. This was triggered by the password read stuff: the
9445 last character typed gets carried over to the next fread(). If you were
9446 generating a new cert request using 'req' for example then the last
9447 character of the passphrase would be CR which would then enter the first
9451 *) Added the new `Includes OpenSSL Cryptography Software' button as
9452 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9453 button and can be used by applications based on OpenSSL to show the
9454 relationship to the OpenSSL project.
9455 [Ralf S. Engelschall]
9457 *) Remove confusing variables in function signatures in files
9458 ssl/ssl_lib.c and ssl/ssl.h.
9459 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9461 *) Don't install bss_file.c under PREFIX/include/
9462 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9464 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9465 functions that return function pointers and has support for NT specific
9466 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9467 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9468 unsigned to signed types: this was killing the Win32 compile.
9471 *) Add new certificate file to stack functions,
9472 SSL_add_dir_cert_subjects_to_stack() and
9473 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9474 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9475 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9476 This means that Apache-SSL and similar packages don't have to mess around
9477 to add as many CAs as they want to the preferred list.
9480 *) Experiment with doxygen documentation. Currently only partially applied to
9482 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9483 openssl.doxy as the configuration file.
9486 *) Get rid of remaining C++-style comments which strict C compilers hate.
9487 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9489 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9490 compiled in by default: it has problems with large keys.
9493 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9494 DH private keys and/or callback functions which directly correspond to
9495 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9496 is needed for applications which have to configure certificates on a
9497 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9499 For the RSA certificate situation is makes no difference, but
9500 for the DSA certificate situation this fixes the "no shared cipher"
9501 problem where the OpenSSL cipher selection procedure failed because the
9502 temporary keys were not overtaken from the context and the API provided
9503 no way to reconfigure them.
9504 The new functions now let applications reconfigure the stuff and they
9505 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9506 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9507 non-public-API function ssl_cert_instantiate() is used as a helper
9508 function and also to reduce code redundancy inside ssl_rsa.c.
9509 [Ralf S. Engelschall]
9511 *) Move s_server -dcert and -dkey options out of the undocumented feature
9512 area because they are useful for the DSA situation and should be
9513 recognized by the users.
9514 [Ralf S. Engelschall]
9516 *) Fix the cipher decision scheme for export ciphers: the export bits are
9517 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9518 SSL_EXP_MASK. So, the original variable has to be used instead of the
9519 already masked variable.
9520 [Richard Levitte <levitte@stacken.kth.se>]
9522 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9523 [Richard Levitte <levitte@stacken.kth.se>]
9525 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9526 from `int' to `unsigned int' because it's a length and initialized by
9527 EVP_DigestFinal() which expects an `unsigned int *'.
9528 [Richard Levitte <levitte@stacken.kth.se>]
9530 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9531 script. Instead use the usual Shell->Perl transition trick.
9532 [Ralf S. Engelschall]
9534 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9535 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9536 -noout -modulus' as it's already the case for `openssl rsa -noout
9537 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9538 currently the public key is printed (a decision which was already done by
9539 `openssl dsa -modulus' in the past) which serves a similar purpose.
9540 Additionally the NO_RSA no longer completely removes the whole -modulus
9541 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9543 [Ralf S. Engelschall]
9545 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9546 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9547 [Arne Ansper <arne@ats.cyber.ee>]
9549 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9550 to be added. Now both 'req' and 'ca' can use new objects defined in the
9554 *) Add cool BIO that does syslog (or event log on NT).
9555 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9557 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9558 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9559 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9560 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9563 *) Add preliminary config info for new extension code.
9566 *) Make RSA_NO_PADDING really use no padding.
9567 [Ulf Moeller <ulf@fitug.de>]
9569 *) Generate errors when private/public key check is done.
9572 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9573 for some CRL extensions and new objects added.
9576 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9577 key usage extension and fuller support for authority key id.
9580 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9581 padding method for RSA, which is recommended for new applications in PKCS
9582 #1 v2.0 (RFC 2437, October 1998).
9583 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9584 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9585 against Bleichbacher's attack on RSA.
9586 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9589 *) Updates to the new SSL compression code
9590 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9592 *) Fix so that the version number in the master secret, when passed
9593 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9594 (because the server will not accept higher), that the version number
9595 is 0x03,0x01, not 0x03,0x00
9596 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9598 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9599 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9600 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9603 *) Support for RAW extensions where an arbitrary extension can be
9604 created by including its DER encoding. See apps/openssl.cnf for
9608 *) Make sure latest Perl versions don't interpret some generated C array
9609 code as Perl array code in the crypto/err/err_genc.pl script.
9610 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9612 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9613 not many people have the assembler. Various Win32 compilation fixes and
9614 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9618 *) Modify configure script 'Configure' to automatically create crypto/date.h
9619 file under Win32 and also build pem.h from pem.org. New script
9620 util/mkfiles.pl to create the MINFO file on environments that can't do a
9621 'make files': perl util/mkfiles.pl >MINFO should work.
9624 *) Major rework of DES function declarations, in the pursuit of correctness
9625 and purity. As a result, many evil casts evaporated, and some weirdness,
9626 too. You may find this causes warnings in your code. Zapping your evil
9627 casts will probably fix them. Mostly.
9630 *) Fix for a typo in asn1.h. Bug fix to object creation script
9631 obj_dat.pl. It considered a zero in an object definition to mean
9632 "end of object": none of the objects in objects.h have any zeros
9633 so it wasn't spotted.
9634 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9636 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9637 Masking (CBCM). In the absence of test vectors, the best I have been able
9638 to do is check that the decrypt undoes the encrypt, so far. Send me test
9639 vectors if you have them.
9642 *) Correct calculation of key length for export ciphers (too much space was
9643 allocated for null ciphers). This has not been tested!
9646 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9647 message is now correct (it understands "crypto" and "ssl" on its
9648 command line). There is also now an "update" option. This will update
9649 the util/ssleay.num and util/libeay.num files with any new functions.
9651 perl util/mkdef.pl crypto ssl update
9652 it will update them.
9655 *) Overhauled the Perl interface (perl/*):
9656 - ported BN stuff to OpenSSL's different BN library
9657 - made the perl/ source tree CVS-aware
9658 - renamed the package from SSLeay to OpenSSL (the files still contain
9659 their history because I've copied them in the repository)
9660 - removed obsolete files (the test scripts will be replaced
9661 by better Test::Harness variants in the future)
9662 [Ralf S. Engelschall]
9664 *) First cut for a very conservative source tree cleanup:
9665 1. merge various obsolete readme texts into doc/ssleay.txt
9666 where we collect the old documents and readme texts.
9667 2. remove the first part of files where I'm already sure that we no
9668 longer need them because of three reasons: either they are just temporary
9669 files which were left by Eric or they are preserved original files where
9670 I've verified that the diff is also available in the CVS via "cvs diff
9671 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9672 the crypto/md/ stuff).
9673 [Ralf S. Engelschall]
9675 *) More extension code. Incomplete support for subject and issuer alt
9676 name, issuer and authority key id. Change the i2v function parameters
9677 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9678 what that's for :-) Fix to ASN1 macro which messed up
9679 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9682 *) Preliminary support for ENUMERATED type. This is largely copied from the
9686 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9687 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9689 *) Make sure `make rehash' target really finds the `openssl' program.
9690 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9692 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9693 like to hear about it if this slows down other processors.
9696 *) Add CygWin32 platform information to Configure script.
9697 [Alan Batie <batie@aahz.jf.intel.com>]
9699 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9700 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9702 *) New program nseq to manipulate netscape certificate sequences
9705 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9709 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9710 but the BN code had some problems that would cause failures when
9711 doing certificate verification and some other functions.
9712 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9714 *) Add ASN1 and PEM code to support netscape certificate sequences.
9717 *) Add ASN1 and PEM code to support netscape certificate sequences.
9720 *) Add several PKIX and private extended key usage OIDs.
9723 *) Modify the 'ca' program to handle the new extension code. Modify
9724 openssl.cnf for new extension format, add comments.
9727 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9728 and add a sample to openssl.cnf so req -x509 now adds appropriate
9732 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9733 error code, add initial support to X509_print() and x509 application.
9736 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9737 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9738 stuff is currently isolated and isn't even compiled yet.
9741 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9742 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9743 Removed the versions check from X509 routines when loading extensions:
9744 this allows certain broken certificates that don't set the version
9745 properly to be processed.
9748 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9749 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9750 can still be regenerated with "make depend".
9753 *) Spelling mistake in C version of CAST-128.
9754 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9756 *) Changes to the error generation code. The perl script err-code.pl
9757 now reads in the old error codes and retains the old numbers, only
9758 adding new ones if necessary. It also only changes the .err files if new
9759 codes are added. The makefiles have been modified to only insert errors
9760 when needed (to avoid needlessly modifying header files). This is done
9761 by only inserting errors if the .err file is newer than the auto generated
9762 C file. To rebuild all the error codes from scratch (the old behaviour)
9763 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9764 or delete all the .err files.
9767 *) CAST-128 was incorrectly implemented for short keys. The C version has
9768 been fixed, but is untested. The assembler versions are also fixed, but
9769 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9770 to regenerate it if needed.
9771 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9772 Hagino <itojun@kame.net>]
9774 *) File was opened incorrectly in randfile.c.
9775 [Ulf Möller <ulf@fitug.de>]
9777 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9778 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9779 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9780 al: it's just almost always a UTCTime. Note this patch adds new error
9781 codes so do a "make errors" if there are problems.
9784 *) Correct Linux 1 recognition in config.
9785 [Ulf Möller <ulf@fitug.de>]
9787 *) Remove pointless MD5 hash when using DSA keys in ca.
9788 [Anonymous <nobody@replay.com>]
9790 *) Generate an error if given an empty string as a cert directory. Also
9791 generate an error if handed NULL (previously returned 0 to indicate an
9792 error, but didn't set one).
9793 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9795 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9798 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9799 parameters. This was causing a warning which killed off the Win32 compile.
9802 *) Remove C++ style comments from crypto/bn/bn_local.h.
9803 [Neil Costigan <neil.costigan@celocom.com>]
9805 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9806 based on a text string, looking up short and long names and finally
9807 "dot" format. The "dot" format stuff didn't work. Added new function
9808 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9809 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9810 OID is not part of the table.
9813 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9814 X509_LOOKUP_by_alias().
9817 *) Sort openssl functions by name.
9820 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9821 encryption from sample DSA keys (in case anyone is interested the password
9825 *) Make _all_ *_free functions accept a NULL pointer.
9826 [Frans Heymans <fheymans@isaserver.be>]
9828 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9830 [Anonymous <nobody@replay.com>]
9832 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9833 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9835 *) Don't blow it for numeric -newkey arguments to apps/req.
9836 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9838 *) Temp key "for export" tests were wrong in s3_srvr.c.
9839 [Anonymous <nobody@replay.com>]
9841 *) Add prototype for temp key callback functions
9842 SSL_CTX_set_tmp_{rsa,dh}_callback().
9845 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9846 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9849 *) X509_name_add_entry() freed the wrong thing after an error.
9850 [Arne Ansper <arne@ats.cyber.ee>]
9852 *) rsa_eay.c would attempt to free a NULL context.
9853 [Arne Ansper <arne@ats.cyber.ee>]
9855 *) BIO_s_socket() had a broken should_retry() on Windoze.
9856 [Arne Ansper <arne@ats.cyber.ee>]
9858 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9859 [Arne Ansper <arne@ats.cyber.ee>]
9861 *) Make sure the already existing X509_STORE->depth variable is initialized
9862 in X509_STORE_new(), but document the fact that this variable is still
9863 unused in the certificate verification process.
9864 [Ralf S. Engelschall]
9866 *) Fix the various library and apps files to free up pkeys obtained from
9867 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9870 *) Fix reference counting in X509_PUBKEY_get(). This makes
9871 demos/maurice/example2.c work, amongst others, probably.
9872 [Steve Henson and Ben Laurie]
9874 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9875 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9876 are no longer created. This way we have a single and consistent command
9877 line interface `openssl <command>', similar to `cvs <command>'.
9878 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9880 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9881 BIT STRING wrapper always have zero unused bits.
9884 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9887 *) Make the top-level INSTALL documentation easier to understand.
9890 *) Makefiles updated to exit if an error occurs in a sub-directory
9891 make (including if user presses ^C) [Paul Sutton]
9893 *) Make Montgomery context stuff explicit in RSA data structure.
9896 *) Fix build order of pem and err to allow for generated pem.h.
9899 *) Fix renumbering bug in X509_NAME_delete_entry().
9902 *) Enhanced the err-ins.pl script so it makes the error library number
9903 global and can add a library name. This is needed for external ASN1 and
9904 other error libraries.
9907 *) Fixed sk_insert which never worked properly.
9910 *) Fix ASN1 macros so they can handle indefinite length construted
9911 EXPLICIT tags. Some non standard certificates use these: they can now
9915 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9916 into a single doc/ssleay.txt bundle. This way the information is still
9917 preserved but no longer messes up this directory. Now it's new room for
9918 the new set of documenation files.
9919 [Ralf S. Engelschall]
9921 *) SETs were incorrectly DER encoded. This was a major pain, because they
9922 shared code with SEQUENCEs, which aren't coded the same. This means that
9923 almost everything to do with SETs or SEQUENCEs has either changed name or
9924 number of arguments.
9925 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9927 *) Fix test data to work with the above.
9930 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9931 was already fixed by Eric for 0.9.1 it seems.
9932 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9934 *) Autodetect FreeBSD3.
9937 *) Fix various bugs in Configure. This affects the following platforms:
9941 unixware-2.0-pentium
9945 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9946 before they are needed.
9949 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9953 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9955 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9956 changed SSLeay to OpenSSL in version strings.
9957 [Ralf S. Engelschall]
9959 *) Some fixups to the top-level documents.
9962 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9963 because the symlink to include/ was missing.
9964 [Ralf S. Engelschall]
9966 *) Incorporated the popular no-RSA/DSA-only patches
9967 which allow to compile a RSA-free SSLeay.
9968 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9970 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9971 when "ssleay" is still not found.
9972 [Ralf S. Engelschall]
9974 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9975 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9977 *) Updated the README file.
9978 [Ralf S. Engelschall]
9980 *) Added various .cvsignore files in the CVS repository subdirs
9981 to make a "cvs update" really silent.
9982 [Ralf S. Engelschall]
9984 *) Recompiled the error-definition header files and added
9985 missing symbols to the Win32 linker tables.
9986 [Ralf S. Engelschall]
9988 *) Cleaned up the top-level documents;
9989 o new files: CHANGES and LICENSE
9990 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9991 o merged COPYRIGHT into LICENSE
9992 o removed obsolete TODO file
9993 o renamed MICROSOFT to INSTALL.W32
9994 [Ralf S. Engelschall]
9996 *) Removed dummy files from the 0.9.1b source tree:
9997 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9998 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9999 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10000 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10001 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10002 [Ralf S. Engelschall]
10004 *) Added various platform portability fixes.
10007 *) The Genesis of the OpenSSL rpject:
10008 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10009 Young and Tim J. Hudson created while they were working for C2Net until
10011 [The OpenSSL Project]
10014 Changes between 0.9.0b and 0.9.1b [not released]
10016 *) Updated a few CA certificates under certs/
10019 *) Changed some BIGNUM api stuff.
10022 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10023 DGUX x86, Linux Alpha, etc.
10026 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10027 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10031 *) Add -strparse option to asn1pars program which parses nested
10033 [Dr Stephen Henson <shenson@bigfoot.com>]
10035 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10038 *) DSA fix for "ca" program.
10041 *) Added "-genkey" option to "dsaparam" program.
10044 *) Added RIPE MD160 (rmd160) message digest.
10047 *) Added -a (all) option to "ssleay version" command.
10050 *) Added PLATFORM define which is the id given to Configure.
10053 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10056 *) Extended the ASN.1 parser routines.
10059 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10062 *) Added a BN_CTX to the BN library.
10065 *) Fixed the weak key values in DES library
10068 *) Changed API in EVP library for cipher aliases.
10071 *) Added support for RC2/64bit cipher.
10074 *) Converted the lhash library to the crypto/mem.c functions.
10077 *) Added more recognized ASN.1 object ids.
10080 *) Added more RSA padding checks for SSL/TLS.
10083 *) Added BIO proxy/filter functionality.
10086 *) Added extra_certs to SSL_CTX which can be used
10087 send extra CA certificates to the client in the CA cert chain sending
10088 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10091 *) Now Fortezza is denied in the authentication phase because
10092 this is key exchange mechanism is not supported by SSLeay at all.
10095 *) Additional PKCS1 checks.
10098 *) Support the string "TLSv1" for all TLS v1 ciphers.
10101 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10102 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10105 *) Fixed a few memory leaks.
10108 *) Fixed various code and comment typos.
10111 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10112 bytes sent in the client random.
10113 [Edward Bishop <ebishop@spyglass.com>]