5 Changes between 1.0.1h and 1.0.2 [xx XXX xxxx]
7 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8 for corner cases. (Certain input points at infinity could lead to
9 bogus results, with non-infinity inputs mapped to infinity too.)
12 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
13 This covers AES, SHA256/512 and GHASH. "Initial" means that most
14 common cases are optimized and there still is room for further
15 improvements. Vector Permutation AES for Altivec is also added.
18 *) Add support for little-endian ppc64 Linux target.
21 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
22 SHA1, SHA256 and GHASH. "Initial" means that most common cases
23 are optimized and there still is room for further improvements.
24 Both 32- and 64-bit modes are supported.
25 [Andy Polyakov, Ard Biesheuvel (Linaro)]
27 *) Improved ARMv7 NEON support.
30 *) Support for SPARC Architecture 2011 crypto extensions, first
31 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
32 SHA256/512, MD5, GHASH and modular exponentiation.
33 [Andy Polyakov, David Miller]
35 *) Accelerated modular exponentiation for Intel processors, a.k.a.
37 [Shay Gueron (Intel Corp)]
39 *) Support for new and upcoming Intel processors, including AVX2,
40 BMI and SHA ISA extensions. This includes additional "stitched"
41 implementations, AESNI-SHA256 and GCM, and multi-buffer support
44 This work was sponsored by Intel Corp.
47 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
48 this fixes a limiation in previous versions of OpenSSL.
51 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
52 MGF1 digest and OAEP label.
55 *) Add EVP support for key wrapping algorithms, to avoid problems with
56 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
57 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
58 algorithms and include tests cases.
61 *) Add functions to allocate and set the fields of an ECDSA_METHOD
63 [Douglas E. Engert, Steve Henson]
65 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
66 difference in days and seconds between two tm or ASN1_TIME structures.
69 *) Add -rev test option to s_server to just reverse order of characters
70 received by client and send back to server. Also prints an abbreviated
71 summary of the connection parameters.
74 *) New option -brief for s_client and s_server to print out a brief summary
75 of connection parameters.
78 *) Add callbacks for arbitrary TLS extensions.
79 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
81 *) New option -crl_download in several openssl utilities to download CRLs
82 from CRLDP extension in certificates.
85 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
88 *) New function X509_CRL_diff to generate a delta CRL from the difference
89 of two full CRLs. Add support to "crl" utility.
92 *) New functions to set lookup_crls function and to retrieve
93 X509_STORE from X509_STORE_CTX.
96 *) Print out deprecated issuer and subject unique ID fields in
100 *) Extend OCSP I/O functions so they can be used for simple general purpose
101 HTTP as well as OCSP. New wrapper function which can be used to download
102 CRLs using the OCSP API.
105 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
108 *) SSL_CONF* functions. These provide a common framework for application
109 configuration using configuration files or command lines.
112 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
113 message callback and prints the results. Needs compile time option
114 "enable-ssl-trace". New options to s_client and s_server to enable
118 *) New ctrl and macro to retrieve supported points extensions.
119 Print out extension in s_server and s_client.
122 *) New functions to retrieve certificate signature and signature
126 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
130 *) New Suite B modes for TLS code. These use and enforce the requirements
131 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
132 only use Suite B curves. The Suite B modes can be set by using the
133 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
136 *) New chain verification flags for Suite B levels of security. Check
137 algorithms are acceptable when flags are set in X509_verify_cert.
140 *) Make tls1_check_chain return a set of flags indicating checks passed
141 by a certificate chain. Add additional tests to handle client
142 certificates: checks for matching certificate type and issuer name
146 *) If an attempt is made to use a signature algorithm not in the peer
147 preference list abort the handshake. If client has no suitable
148 signature algorithms in response to a certificate request do not
152 *) If server EC tmp key is not in client preference list abort handshake.
155 *) Add support for certificate stores in CERT structure. This makes it
156 possible to have different stores per SSL structure or one store in
157 the parent SSL_CTX. Include distint stores for certificate chain
158 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
159 to build and store a certificate chain in CERT structure: returing
160 an error if the chain cannot be built: this will allow applications
161 to test if a chain is correctly configured.
163 Note: if the CERT based stores are not set then the parent SSL_CTX
164 store is used to retain compatibility with existing behaviour.
168 *) New function ssl_set_client_disabled to set a ciphersuite disabled
169 mask based on the current session, check mask when sending client
170 hello and checking the requested ciphersuite.
173 *) New ctrls to retrieve and set certificate types in a certificate
174 request message. Print out received values in s_client. If certificate
175 types is not set with custom values set sensible values based on
176 supported signature algorithms.
179 *) Support for distinct client and server supported signature algorithms.
182 *) Add certificate callback. If set this is called whenever a certificate
183 is required by client or server. An application can decide which
184 certificate chain to present based on arbitrary criteria: for example
185 supported signature algorithms. Add very simple example to s_server.
186 This fixes many of the problems and restrictions of the existing client
187 certificate callback: for example you can now clear an existing
188 certificate and specify the whole chain.
191 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
192 the certificate can be used for (if anything). Set valid_flags field
193 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
194 to have similar checks in it.
196 Add new "cert_flags" field to CERT structure and include a "strict mode".
197 This enforces some TLS certificate requirements (such as only permitting
198 certificate signature algorithms contained in the supported algorithms
199 extension) which some implementations ignore: this option should be used
200 with caution as it could cause interoperability issues.
203 *) Update and tidy signature algorithm extension processing. Work out
204 shared signature algorithms based on preferences and peer algorithms
205 and print them out in s_client and s_server. Abort handshake if no
206 shared signature algorithms.
209 *) Add new functions to allow customised supported signature algorithms
210 for SSL and SSL_CTX structures. Add options to s_client and s_server
214 *) New function SSL_certs_clear() to delete all references to certificates
215 from an SSL structure. Before this once a certificate had been added
216 it couldn't be removed.
219 *) Integrate hostname, email address and IP address checking with certificate
220 verification. New verify options supporting checking in opensl utility.
223 *) Fixes and wildcard matching support to hostname and email checking
224 functions. Add manual page.
225 [Florian Weimer (Red Hat Product Security Team)]
227 *) New functions to check a hostname email or IP address against a
228 certificate. Add options x509 utility to print results of checks against
232 *) Fix OCSP checking.
233 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
235 *) Initial experimental support for explicitly trusted non-root CAs.
236 OpenSSL still tries to build a complete chain to a root but if an
237 intermediate CA has a trust setting included that is used. The first
238 setting is used: whether to trust (e.g., -addtrust option to the x509
242 *) Add -trusted_first option which attempts to find certificates in the
243 trusted store even if an untrusted chain is also supplied.
246 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
247 platform support for Linux and Android.
250 *) Support for linux-x32, ILP32 environment in x86_64 framework.
253 *) Experimental multi-implementation support for FIPS capable OpenSSL.
254 When in FIPS mode the approved implementations are used as normal,
255 when not in FIPS mode the internal unapproved versions are used instead.
256 This means that the FIPS capable OpenSSL isn't forced to use the
257 (often lower perfomance) FIPS implementations outside FIPS mode.
260 *) Transparently support X9.42 DH parameters when calling
261 PEM_read_bio_DHparameters. This means existing applications can handle
262 the new parameter format automatically.
265 *) Initial experimental support for X9.42 DH parameter format: mainly
266 to support use of 'q' parameter for RFC5114 parameters.
269 *) Add DH parameters from RFC5114 including test data to dhtest.
272 *) Support for automatic EC temporary key parameter selection. If enabled
273 the most preferred EC parameters are automatically used instead of
274 hardcoded fixed parameters. Now a server just has to call:
275 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
276 support ECDH and use the most appropriate parameters.
279 *) Enhance and tidy EC curve and point format TLS extension code. Use
280 static structures instead of allocation if default values are used.
281 New ctrls to set curves we wish to support and to retrieve shared curves.
282 Print out shared curves in s_server. New options to s_server and s_client
283 to set list of supported curves.
286 *) New ctrls to retrieve supported signature algorithms and
287 supported curve values as an array of NIDs. Extend openssl utility
288 to print out received values.
291 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
292 between NIDs and the more common NIST names such as "P-256". Enhance
293 ecparam utility and ECC method to recognise the NIST names for curves.
296 *) Enhance SSL/TLS certificate chain handling to support different
297 chains for each certificate instead of one chain in the parent SSL_CTX.
300 *) Support for fixed DH ciphersuite client authentication: where both
301 server and client use DH certificates with common parameters.
304 *) Support for fixed DH ciphersuites: those requiring DH server
308 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
310 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
311 handshake can force the use of weak keying material in OpenSSL
312 SSL/TLS clients and servers.
314 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
315 researching this issue. (CVE-2014-0224)
316 [KIKUCHI Masashi, Steve Henson]
318 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
319 OpenSSL DTLS client the code can be made to recurse eventually crashing
322 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
324 [Imre Rad, Steve Henson]
326 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
327 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
328 client or server. This is potentially exploitable to run arbitrary
329 code on a vulnerable client or server.
331 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
332 [Jüri Aedla, Steve Henson]
334 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
335 are subject to a denial of service attack.
337 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
338 this issue. (CVE-2014-3470)
339 [Felix Gröbert, Ivan Fratric, Steve Henson]
341 *) Harmonize version and its documentation. -f flag is used to display
343 [mancha <mancha1@zoho.com>]
345 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
347 [mancha <mancha1@zoho.com>]
349 *) Fix some double frees. These are not thought to be exploitable.
350 [mancha <mancha1@zoho.com>]
352 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
354 *) A missing bounds check in the handling of the TLS heartbeat extension
355 can be used to reveal up to 64k of memory to a connected client or
358 Thanks for Neel Mehta of Google Security for discovering this bug and to
359 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
360 preparing the fix (CVE-2014-0160)
361 [Adam Langley, Bodo Moeller]
363 *) Fix for the attack described in the paper "Recovering OpenSSL
364 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
365 by Yuval Yarom and Naomi Benger. Details can be obtained from:
366 http://eprint.iacr.org/2014/140
368 Thanks to Yuval Yarom and Naomi Benger for discovering this
369 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
370 [Yuval Yarom and Naomi Benger]
372 *) TLS pad extension: draft-agl-tls-padding-03
374 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
375 TLS client Hello record length value would otherwise be > 255 and
376 less that 512 pad with a dummy extension containing zeroes so it
377 is at least 512 bytes long.
379 [Adam Langley, Steve Henson]
381 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
383 *) Fix for TLS record tampering bug. A carefully crafted invalid
384 handshake could crash OpenSSL with a NULL pointer exception.
385 Thanks to Anton Johansson for reporting this issues.
388 *) Keep original DTLS digest and encryption contexts in retransmission
389 structures so we can use the previous session parameters if they need
390 to be resent. (CVE-2013-6450)
393 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
394 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
395 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
396 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
397 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
398 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
399 [Rob Stradling, Adam Langley]
401 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
403 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
404 supporting platforms or when small records were transferred.
405 [Andy Polyakov, Steve Henson]
407 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
409 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
411 This addresses the flaw in CBC record processing discovered by
412 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
413 at: http://www.isg.rhul.ac.uk/tls/
415 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
416 Security Group at Royal Holloway, University of London
417 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
418 Emilia Käsper for the initial patch.
420 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
422 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
423 ciphersuites which can be exploited in a denial of service attack.
424 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
425 and detecting this bug and to Wolfgang Ettlinger
426 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
430 *) Return an error when checking OCSP signatures when key is NULL.
431 This fixes a DoS attack. (CVE-2013-0166)
434 *) Make openssl verify return errors.
435 [Chris Palmer <palmer@google.com> and Ben Laurie]
437 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
438 the right response is stapled. Also change SSL_get_certificate()
439 so it returns the certificate actually sent.
440 See http://rt.openssl.org/Ticket/Display.html?id=2836.
441 [Rob Stradling <rob.stradling@comodo.com>]
443 *) Fix possible deadlock when decoding public keys.
446 *) Don't use TLS 1.0 record version number in initial client hello
450 Changes between 1.0.1b and 1.0.1c [10 May 2012]
452 *) Sanity check record length before skipping explicit IV in TLS
453 1.2, 1.1 and DTLS to fix DoS attack.
455 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
456 fuzzing as a service testing platform.
460 *) Initialise tkeylen properly when encrypting CMS messages.
461 Thanks to Solar Designer of Openwall for reporting this issue.
464 *) In FIPS mode don't try to use composite ciphers as they are not
468 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
470 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
471 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
472 mean any application compiled against OpenSSL 1.0.0 headers setting
473 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
474 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
475 0x10000000L Any application which was previously compiled against
476 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
477 will need to be recompiled as a result. Letting be results in
478 inability to disable specifically TLS 1.1 and in client context,
479 in unlike event, limit maximum offered version to TLS 1.0 [see below].
482 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
483 disable just protocol X, but all protocols above X *if* there are
484 protocols *below* X still enabled. In more practical terms it means
485 that if application wants to disable TLS1.0 in favor of TLS1.1 and
486 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
487 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
491 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
493 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
494 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
495 in CRYPTO_realloc_clean.
497 Thanks to Tavis Ormandy, Google Security Team, for discovering this
498 issue and to Adam Langley <agl@chromium.org> for fixing it.
500 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
502 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
505 *) Workarounds for some broken servers that "hang" if a client hello
506 record length exceeds 255 bytes.
508 1. Do not use record version number > TLS 1.0 in initial client
509 hello: some (but not all) hanging servers will now work.
510 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
511 the number of ciphers sent in the client hello. This should be
512 set to an even number, such as 50, for example by passing:
513 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
514 Most broken servers should now work.
515 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
516 TLS 1.2 client support entirely.
519 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
522 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
524 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
525 STRING form instead of a DigestInfo.
528 *) The format used for MDC2 RSA signatures is inconsistent between EVP
529 and the RSA_sign/RSA_verify functions. This was made more apparent when
530 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
531 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
532 the correct format in RSA_verify so both forms transparently work.
535 *) Some servers which support TLS 1.0 can choke if we initially indicate
536 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
537 encrypted premaster secret. As a workaround use the maximum pemitted
538 client version in client hello, this should keep such servers happy
539 and still work with previous versions of OpenSSL.
542 *) Add support for TLS/DTLS heartbeats.
543 [Robin Seggelmann <seggelmann@fh-muenster.de>]
545 *) Add support for SCTP.
546 [Robin Seggelmann <seggelmann@fh-muenster.de>]
548 *) Improved PRNG seeding for VOS.
549 [Paul Green <Paul.Green@stratus.com>]
551 *) Extensive assembler packs updates, most notably:
553 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
554 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
555 - x86_64: bit-sliced AES implementation;
556 - ARM: NEON support, contemporary platforms optimizations;
557 - s390x: z196 support;
558 - *: GHASH and GF(2^m) multiplication implementations;
562 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
563 (removal of unnecessary code)
564 [Peter Sylvester <peter.sylvester@edelweb.fr>]
566 *) Add TLS key material exporter from RFC 5705.
569 *) Add DTLS-SRTP negotiation from RFC 5764.
572 *) Add Next Protocol Negotiation,
573 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
574 disabled with a no-npn flag to config or Configure. Code donated
576 [Adam Langley <agl@google.com> and Ben Laurie]
578 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
579 NIST-P256, NIST-P521, with constant-time single point multiplication on
580 typical inputs. Compiler support for the nonstandard type __uint128_t is
581 required to use this (present in gcc 4.4 and later, for 64-bit builds).
582 Code made available under Apache License version 2.0.
584 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
585 line to include this in your build of OpenSSL, and run "make depend" (or
586 "make update"). This enables the following EC_METHODs:
588 EC_GFp_nistp224_method()
589 EC_GFp_nistp256_method()
590 EC_GFp_nistp521_method()
592 EC_GROUP_new_by_curve_name() will automatically use these (while
593 EC_GROUP_new_curve_GFp() currently prefers the more flexible
595 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
597 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
598 all platforms. Move ssize_t definition from e_os.h to the public
599 header file e_os2.h as it now appears in public header file cms.h
602 *) New -sigopt option to the ca, req and x509 utilities. Additional
603 signature parameters can be passed using this option and in
607 *) Add RSA PSS signing function. This will generate and set the
608 appropriate AlgorithmIdentifiers for PSS based on those in the
609 corresponding EVP_MD_CTX structure. No application support yet.
612 *) Support for companion algorithm specific ASN1 signing routines.
613 New function ASN1_item_sign_ctx() signs a pre-initialised
614 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
615 the appropriate parameters.
618 *) Add new algorithm specific ASN1 verification initialisation function
619 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
620 handling will be the same no matter what EVP_PKEY_METHOD is used.
621 Add a PSS handler to support verification of PSS signatures: checked
622 against a number of sample certificates.
625 *) Add signature printing for PSS. Add PSS OIDs.
626 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
628 *) Add algorithm specific signature printing. An individual ASN1 method
629 can now print out signatures instead of the standard hex dump.
631 More complex signatures (e.g. PSS) can print out more meaningful
632 information. Include DSA version that prints out the signature
636 *) Password based recipient info support for CMS library: implementing
640 *) Split password based encryption into PBES2 and PBKDF2 functions. This
641 neatly separates the code into cipher and PBE sections and is required
642 for some algorithms that split PBES2 into separate pieces (such as
646 *) Session-handling fixes:
647 - Fix handling of connections that are resuming with a session ID,
648 but also support Session Tickets.
649 - Fix a bug that suppressed issuing of a new ticket if the client
650 presented a ticket with an expired session.
651 - Try to set the ticket lifetime hint to something reasonable.
652 - Make tickets shorter by excluding irrelevant information.
653 - On the client side, don't ignore renewed tickets.
654 [Adam Langley, Bodo Moeller (Google)]
656 *) Fix PSK session representation.
659 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
661 This work was sponsored by Intel.
664 *) Add GCM support to TLS library. Some custom code is needed to split
665 the IV between the fixed (from PRF) and explicit (from TLS record)
666 portions. This adds all GCM ciphersuites supported by RFC5288 and
667 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
668 add a special AESGCM string for GCM only.
671 *) Expand range of ctrls for AES GCM. Permit setting invocation
672 field on decrypt and retrieval of invocation field only on encrypt.
675 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
676 As required by RFC5289 these ciphersuites cannot be used if for
677 versions of TLS earlier than 1.2.
680 *) For FIPS capable OpenSSL interpret a NULL default public key method
681 as unset and return the appopriate default but do *not* set the default.
682 This means we can return the appopriate method in applications that
683 swicth between FIPS and non-FIPS modes.
686 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
687 ENGINE is used then we cannot handle that in the FIPS module so we
688 keep original code iff non-FIPS operations are allowed.
691 *) Add -attime option to openssl utilities.
692 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
694 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
697 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
698 FIPS EC methods unconditionally for now.
701 *) New build option no-ec2m to disable characteristic 2 code.
704 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
705 all cases can be covered as some introduce binary incompatibilities.
708 *) Redirect RSA operations to FIPS module including keygen,
709 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
712 *) Add similar low level API blocking to ciphers.
715 *) Low level digest APIs are not approved in FIPS mode: any attempt
716 to use these will cause a fatal error. Applications that *really* want
717 to use them can use the private_* version instead.
720 *) Redirect cipher operations to FIPS module for FIPS builds.
723 *) Redirect digest operations to FIPS module for FIPS builds.
726 *) Update build system to add "fips" flag which will link in fipscanister.o
727 for static and shared library builds embedding a signature if needed.
730 *) Output TLS supported curves in preference order instead of numerical
731 order. This is currently hardcoded for the highest order curves first.
732 This should be configurable so applications can judge speed vs strength.
735 *) Add TLS v1.2 server support for client authentication.
738 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
742 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
743 FIPS modules versions.
746 *) Add TLS v1.2 client side support for client authentication. Keep cache
747 of handshake records longer as we don't know the hash algorithm to use
748 until after the certificate request message is received.
751 *) Initial TLS v1.2 client support. Add a default signature algorithms
752 extension including all the algorithms we support. Parse new signature
753 format in client key exchange. Relax some ECC signing restrictions for
754 TLS v1.2 as indicated in RFC5246.
757 *) Add server support for TLS v1.2 signature algorithms extension. Switch
758 to new signature format when needed using client digest preference.
759 All server ciphersuites should now work correctly in TLS v1.2. No client
760 support yet and no support for client certificates.
763 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
764 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
765 ciphersuites. At present only RSA key exchange ciphersuites work with
766 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
767 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
768 and version checking.
771 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
772 with this defined it will not be affected by any changes to ssl internal
773 structures. Add several utility functions to allow openssl application
774 to work with OPENSSL_NO_SSL_INTERN defined.
778 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
780 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
783 *) Permit abbreviated handshakes when renegotiating using the function
784 SSL_renegotiate_abbreviated().
785 [Robin Seggelmann <seggelmann@fh-muenster.de>]
787 *) Add call to ENGINE_register_all_complete() to
788 ENGINE_load_builtin_engines(), so some implementations get used
789 automatically instead of needing explicit application support.
792 *) Add support for TLS key exporter as described in RFC5705.
793 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
795 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
796 a few changes are required:
798 Add SSL_OP_NO_TLSv1_1 flag.
800 Update version checking logic to handle version 1.1.
801 Add explicit IV handling (ported from DTLS code).
802 Add command line options to s_client/s_server.
805 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
807 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
809 This addresses the flaw in CBC record processing discovered by
810 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
811 at: http://www.isg.rhul.ac.uk/tls/
813 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
814 Security Group at Royal Holloway, University of London
815 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
816 Emilia Käsper for the initial patch.
818 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
820 *) Return an error when checking OCSP signatures when key is NULL.
821 This fixes a DoS attack. (CVE-2013-0166)
824 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
825 the right response is stapled. Also change SSL_get_certificate()
826 so it returns the certificate actually sent.
827 See http://rt.openssl.org/Ticket/Display.html?id=2836.
829 [Rob Stradling <rob.stradling@comodo.com>]
831 *) Fix possible deadlock when decoding public keys.
834 Changes between 1.0.0i and 1.0.0j [10 May 2012]
836 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
839 *) Sanity check record length before skipping explicit IV in DTLS
842 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
843 fuzzing as a service testing platform.
847 *) Initialise tkeylen properly when encrypting CMS messages.
848 Thanks to Solar Designer of Openwall for reporting this issue.
851 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
853 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
854 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
855 in CRYPTO_realloc_clean.
857 Thanks to Tavis Ormandy, Google Security Team, for discovering this
858 issue and to Adam Langley <agl@chromium.org> for fixing it.
860 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
862 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
864 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
865 in CMS and PKCS7 code. When RSA decryption fails use a random key for
866 content decryption and always return the same error. Note: this attack
867 needs on average 2^20 messages so it only affects automated senders. The
868 old behaviour can be reenabled in the CMS code by setting the
869 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
870 an MMA defence is not necessary.
871 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
872 this issue. (CVE-2012-0884)
875 *) Fix CVE-2011-4619: make sure we really are receiving a
876 client hello before rejecting multiple SGC restarts. Thanks to
877 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
880 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
882 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
883 Thanks to Antonio Martin, Enterprise Secure Access Research and
884 Development, Cisco Systems, Inc. for discovering this bug and
885 preparing a fix. (CVE-2012-0050)
888 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
890 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
891 of the Vaudenay padding oracle attack on CBC mode encryption
892 which enables an efficient plaintext recovery attack against
893 the OpenSSL implementation of DTLS. Their attack exploits timing
894 differences arising during decryption processing. A research
895 paper describing this attack can be found at:
896 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
897 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
898 Security Group at Royal Holloway, University of London
899 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
900 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
901 for preparing the fix. (CVE-2011-4108)
902 [Robin Seggelmann, Michael Tuexen]
904 *) Clear bytes used for block padding of SSL 3.0 records.
906 [Adam Langley (Google)]
908 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
909 Kadianakis <desnacked@gmail.com> for discovering this issue and
910 Adam Langley for preparing the fix. (CVE-2011-4619)
911 [Adam Langley (Google)]
913 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
914 [Andrey Kulikov <amdeich@gmail.com>]
916 *) Prevent malformed RFC3779 data triggering an assertion failure.
917 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
918 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
919 [Rob Austein <sra@hactrn.net>]
921 *) Improved PRNG seeding for VOS.
922 [Paul Green <Paul.Green@stratus.com>]
924 *) Fix ssl_ciph.c set-up race.
925 [Adam Langley (Google)]
927 *) Fix spurious failures in ecdsatest.c.
928 [Emilia Käsper (Google)]
930 *) Fix the BIO_f_buffer() implementation (which was mixing different
931 interpretations of the '..._len' fields).
932 [Adam Langley (Google)]
934 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
935 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
936 threads won't reuse the same blinding coefficients.
938 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
939 lock to call BN_BLINDING_invert_ex, and avoids one use of
940 BN_BLINDING_update for each BN_BLINDING structure (previously,
941 the last update always remained unused).
942 [Emilia Käsper (Google)]
944 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
945 [Bob Buckholz (Google)]
947 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
949 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
950 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
951 [Kaspar Brand <ossl@velox.ch>]
953 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
954 for multi-threaded use of ECDH. (CVE-2011-3210)
955 [Adam Langley (Google)]
957 *) Fix x509_name_ex_d2i memory leak on bad inputs.
960 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
961 signature public key algorithm by using OID xref utilities instead.
962 Before this you could only use some ECC ciphersuites with SHA1 only.
965 *) Add protection against ECDSA timing attacks as mentioned in the paper
966 by Billy Bob Brumley and Nicola Tuveri, see:
968 http://eprint.iacr.org/2011/232.pdf
970 [Billy Bob Brumley and Nicola Tuveri]
972 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
974 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
975 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
977 *) Fix bug in string printing code: if *any* escaping is enabled we must
978 escape the escape character (backslash) or the resulting string is
982 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
984 *) Disable code workaround for ancient and obsolete Netscape browsers
985 and servers: an attacker can use it in a ciphersuite downgrade attack.
986 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
989 *) Fixed J-PAKE implementation error, originally discovered by
990 Sebastien Martini, further info and confirmation from Stefan
991 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
994 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
996 *) Fix extension code to avoid race conditions which can result in a buffer
997 overrun vulnerability: resumed sessions must not be modified as they can
998 be shared by multiple threads. CVE-2010-3864
1001 *) Fix WIN32 build system to correctly link an ENGINE directory into
1005 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1007 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1009 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1011 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1013 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1014 context. The operation can be customised via the ctrl mechanism in
1015 case ENGINEs want to include additional functionality.
1018 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1021 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1022 output hashes compatible with older versions of OpenSSL.
1023 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1025 *) Fix compression algorithm handling: if resuming a session use the
1026 compression algorithm of the resumed session instead of determining
1027 it from client hello again. Don't allow server to change algorithm.
1030 *) Add load_crls() function to apps tidying load_certs() too. Add option
1031 to verify utility to allow additional CRLs to be included.
1034 *) Update OCSP request code to permit adding custom headers to the request:
1035 some responders need this.
1038 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1040 [Julia Lawall <julia@diku.dk>]
1042 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1043 needlessly dereferenced structures, used obsolete functions and
1044 didn't handle all updated verify codes correctly.
1047 *) Disable MD2 in the default configuration.
1050 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1051 indicate the initial BIO being pushed or popped. This makes it possible
1052 to determine whether the BIO is the one explicitly called or as a result
1053 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1054 it handles reference counts correctly and doesn't zero out the I/O bio
1055 when it is not being explicitly popped. WARNING: applications which
1056 included workarounds for the old buggy behaviour will need to be modified
1057 or they could free up already freed BIOs.
1060 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1061 renaming to all platforms (within the 0.9.8 branch, this was
1062 done conditionally on Netware platforms to avoid a name clash).
1063 [Guenter <lists@gknw.net>]
1065 *) Add ECDHE and PSK support to DTLS.
1066 [Michael Tuexen <tuexen@fh-muenster.de>]
1068 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1072 *) Add "missing" function EVP_MD_flags() (without this the only way to
1073 retrieve a digest flags is by accessing the structure directly. Update
1074 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1075 or cipher is registered as in the "from" argument. Print out all
1076 registered digests in the dgst usage message instead of manually
1077 attempting to work them out.
1080 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1081 this allows the use of compression and extensions. Change default cipher
1082 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1083 by default unless an application cipher string requests it.
1086 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1087 key ids to find matching certificates and keys but some PKCS#12 files
1088 don't follow the (somewhat unwritten) rules and this strategy fails.
1089 Now just gather all certificates together and the first private key
1090 then look for the first certificate that matches the key.
1093 *) Support use of registered digest and cipher names for dgst and cipher
1094 commands instead of having to add each one as a special case. So now
1101 openssl dgst -sha256 foo
1103 and this works for ENGINE based algorithms too.
1107 *) Update Gost ENGINE to support parameter files.
1108 [Victor B. Wagner <vitus@cryptocom.ru>]
1110 *) Support GeneralizedTime in ca utility.
1111 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1113 *) Enhance the hash format used for certificate directory links. The new
1114 form uses the canonical encoding (meaning equivalent names will work
1115 even if they aren't identical) and uses SHA1 instead of MD5. This form
1116 is incompatible with the older format and as a result c_rehash should
1117 be used to rebuild symbolic links.
1120 *) Make PKCS#8 the default write format for private keys, replacing the
1121 traditional format. This form is standardised, more secure and doesn't
1122 include an implicit MD5 dependency.
1125 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1126 committed to OpenSSL should pass this lot as a minimum.
1129 *) Add session ticket override functionality for use by EAP-FAST.
1130 [Jouni Malinen <j@w1.fi>]
1132 *) Modify HMAC functions to return a value. Since these can be implemented
1133 in an ENGINE errors can occur.
1136 *) Type-checked OBJ_bsearch_ex.
1139 *) Type-checked OBJ_bsearch. Also some constification necessitated
1140 by type-checking. Still to come: TXT_DB, bsearch(?),
1141 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1145 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1146 seconds to a tm structure directly, instead of going through OS
1147 specific date routines. This avoids any issues with OS routines such
1148 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1149 and X509_time_adj_ex() to cover the extended range. The existing
1150 X509_time_adj() is still usable and will no longer have any date issues.
1153 *) Delta CRL support. New use deltas option which will attempt to locate
1154 and search any appropriate delta CRLs available.
1156 This work was sponsored by Google.
1159 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1160 code and add additional score elements. Validate alternate CRL paths
1161 as part of the CRL checking and indicate a new error "CRL path validation
1162 error" in this case. Applications wanting additional details can use
1163 the verify callback and check the new "parent" field. If this is not
1164 NULL CRL path validation is taking place. Existing applications wont
1165 see this because it requires extended CRL support which is off by
1168 This work was sponsored by Google.
1171 *) Support for freshest CRL extension.
1173 This work was sponsored by Google.
1176 *) Initial indirect CRL support. Currently only supported in the CRLs
1177 passed directly and not via lookup. Process certificate issuer
1178 CRL entry extension and lookup CRL entries by bother issuer name
1179 and serial number. Check and process CRL issuer entry in IDP extension.
1181 This work was sponsored by Google.
1184 *) Add support for distinct certificate and CRL paths. The CRL issuer
1185 certificate is validated separately in this case. Only enabled if
1186 an extended CRL support flag is set: this flag will enable additional
1187 CRL functionality in future.
1189 This work was sponsored by Google.
1192 *) Add support for policy mappings extension.
1194 This work was sponsored by Google.
1197 *) Fixes to pathlength constraint, self issued certificate handling,
1198 policy processing to align with RFC3280 and PKITS tests.
1200 This work was sponsored by Google.
1203 *) Support for name constraints certificate extension. DN, email, DNS
1204 and URI types are currently supported.
1206 This work was sponsored by Google.
1209 *) To cater for systems that provide a pointer-based thread ID rather
1210 than numeric, deprecate the current numeric thread ID mechanism and
1211 replace it with a structure and associated callback type. This
1212 mechanism allows a numeric "hash" to be extracted from a thread ID in
1213 either case, and on platforms where pointers are larger than 'long',
1214 mixing is done to help ensure the numeric 'hash' is usable even if it
1215 can't be guaranteed unique. The default mechanism is to use "&errno"
1216 as a pointer-based thread ID to distinguish between threads.
1218 Applications that want to provide their own thread IDs should now use
1219 CRYPTO_THREADID_set_callback() to register a callback that will call
1220 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1222 Note that ERR_remove_state() is now deprecated, because it is tied
1223 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1224 to free the current thread's error state should be replaced by
1225 ERR_remove_thread_state(NULL).
1227 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1228 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1229 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1230 application was previously providing a numeric thread callback that
1231 was inappropriate for distinguishing threads, then uniqueness might
1232 have been obtained with &errno that happened immediately in the
1233 intermediate development versions of OpenSSL; this is no longer the
1234 case, the numeric thread callback will now override the automatic use
1236 [Geoff Thorpe, with help from Bodo Moeller]
1238 *) Initial support for different CRL issuing certificates. This covers a
1239 simple case where the self issued certificates in the chain exist and
1240 the real CRL issuer is higher in the existing chain.
1242 This work was sponsored by Google.
1245 *) Removed effectively defunct crypto/store from the build.
1248 *) Revamp of STACK to provide stronger type-checking. Still to come:
1249 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1250 ASN1_STRING, CONF_VALUE.
1253 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1254 RAM on SSL connections. This option can save about 34k per idle SSL.
1257 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1258 STACK, TXT_DB, bsearch, qsort.
1261 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1262 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1263 support for data, signedData, compressedData, digestedData and
1264 encryptedData, envelopedData types included. Scripts to check against
1265 RFC4134 examples draft and interop and consistency checks of many
1266 content types and variants.
1269 *) Add options to enc utility to support use of zlib compression BIO.
1272 *) Extend mk1mf to support importing of options and assembly language
1273 files from Configure script, currently only included in VC-WIN32.
1274 The assembly language rules can now optionally generate the source
1275 files from the associated perl scripts.
1278 *) Implement remaining functionality needed to support GOST ciphersuites.
1279 Interop testing has been performed using CryptoPro implementations.
1280 [Victor B. Wagner <vitus@cryptocom.ru>]
1282 *) s390x assembler pack.
1285 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1289 *) Implement Opaque PRF Input TLS extension as specified in
1290 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1291 official specification yet and no extension type assignment by
1292 IANA exists, this extension (for now) will have to be explicitly
1293 enabled when building OpenSSL by providing the extension number
1294 to use. For example, specify an option
1296 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1298 to the "config" or "Configure" script to enable the extension,
1299 assuming extension number 0x9527 (which is a completely arbitrary
1300 and unofficial assignment based on the MD5 hash of the Internet
1301 Draft). Note that by doing so, you potentially lose
1302 interoperability with other TLS implementations since these might
1303 be using the same extension number for other purposes.
1305 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1306 opaque PRF input value to use in the handshake. This will create
1307 an interal copy of the length-'len' string at 'src', and will
1308 return non-zero for success.
1310 To get more control and flexibility, provide a callback function
1313 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1314 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1318 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1321 Callback function 'cb' will be called in handshakes, and is
1322 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1323 Argument 'arg' is for application purposes (the value as given to
1324 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1325 be provided to the callback function). The callback function
1326 has to return non-zero to report success: usually 1 to use opaque
1327 PRF input just if possible, or 2 to enforce use of the opaque PRF
1328 input. In the latter case, the library will abort the handshake
1329 if opaque PRF input is not successfully negotiated.
1331 Arguments 'peerinput' and 'len' given to the callback function
1332 will always be NULL and 0 in the case of a client. A server will
1333 see the client's opaque PRF input through these variables if
1334 available (NULL and 0 otherwise). Note that if the server
1335 provides an opaque PRF input, the length must be the same as the
1336 length of the client's opaque PRF input.
1338 Note that the callback function will only be called when creating
1339 a new session (session resumption can resume whatever was
1340 previously negotiated), and will not be called in SSL 2.0
1341 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1342 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1343 for applications that need to enforce opaque PRF input.
1347 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1350 [Victor B. Wagner <vitus@cryptocom.ru>]
1352 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1353 RFC4507bis. The encrypted ticket format is an encrypted encoded
1354 SSL_SESSION structure, that way new session features are automatically
1357 If a client application caches session in an SSL_SESSION structure
1358 support is transparent because tickets are now stored in the encoded
1361 The SSL_CTX structure automatically generates keys for ticket
1362 protection in servers so again support should be possible
1363 with no application modification.
1365 If a client or server wishes to disable RFC4507 support then the option
1366 SSL_OP_NO_TICKET can be set.
1368 Add a TLS extension debugging callback to allow the contents of any client
1369 or server extensions to be examined.
1371 This work was sponsored by Google.
1374 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1375 OpenSSL should now compile cleanly on gcc 4.2
1376 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1378 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1379 support including streaming MAC support: this is required for GOST
1380 ciphersuite support.
1381 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1383 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1384 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1385 to output in BER and PEM format.
1388 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1389 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1390 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1391 ENGINE support for HMAC keys which are unextractable. New -mac and
1392 -macopt options to dgst utility.
1395 *) New option -sigopt to dgst utility. Update dgst to use
1396 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1397 alternative signing paramaters such as X9.31 or PSS in the dgst
1401 *) Change ssl_cipher_apply_rule(), the internal function that does
1402 the work each time a ciphersuite string requests enabling
1403 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1404 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1405 the order of disabled ciphersuites such that those ciphersuites
1406 that most recently went from enabled to disabled not only stay
1407 in order with respect to each other, but also have higher priority
1408 than other disabled ciphersuites the next time ciphersuites are
1411 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1412 the same ciphersuites as with "HIGH" alone, but in a specific
1413 order where the PSK ciphersuites come first (since they are the
1414 most recently disabled ciphersuites when "HIGH" is parsed).
1416 Also, change ssl_create_cipher_list() (using this new
1417 funcionality) such that between otherwise identical
1418 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1422 *) Change ssl_create_cipher_list() so that it automatically
1423 arranges the ciphersuites in reasonable order before starting
1424 to process the rule string. Thus, the definition for "DEFAULT"
1425 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1426 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1427 This makes it much easier to arrive at a reasonable default order
1428 in applications for which anonymous ciphers are OK (meaning
1429 that you can't actually use DEFAULT).
1430 [Bodo Moeller; suggested by Victor Duchovni]
1432 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1433 processing) into multiple integers instead of setting
1434 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1435 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1436 (These masks as well as the individual bit definitions are hidden
1437 away into the non-exported interface ssl/ssl_locl.h, so this
1438 change to the definition of the SSL_CIPHER structure shouldn't
1439 affect applications.) This give us more bits for each of these
1440 categories, so there is no longer a need to coagulate AES128 and
1441 AES256 into a single algorithm bit, and to coagulate Camellia128
1442 and Camellia256 into a single algorithm bit, which has led to all
1445 Thus, among other things, the kludge introduced in 0.9.7m and
1446 0.9.8e for masking out AES256 independently of AES128 or masking
1447 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1449 With the change, we also introduce new ciphersuite aliases that
1450 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1454 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1455 Use the leftmost N bytes of the signature input if the input is
1456 larger than the prime q (with N being the size in bytes of q).
1459 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1460 it yet and it is largely untested.
1463 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1466 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1467 some compilers (gcc 4.2 and later) reject their use. Safestack is
1468 reimplemented. Update ASN1 to avoid use of legacy functions.
1471 *) Win32/64 targets are linked with Winsock2.
1474 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1475 to external functions. This can be used to increase CRL handling
1476 efficiency especially when CRLs are very large by (for example) storing
1477 the CRL revoked certificates in a database.
1480 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1481 new CRLs added to a directory can be used. New command line option
1482 -verify_return_error to s_client and s_server. This causes real errors
1483 to be returned by the verify callback instead of carrying on no matter
1484 what. This reflects the way a "real world" verify callback would behave.
1487 *) GOST engine, supporting several GOST algorithms and public key formats.
1488 Kindly donated by Cryptocom.
1491 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1492 partitioned by DP are handled but no indirect CRL or reason partitioning
1493 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1494 selected via a scoring technique which handles IDP and AKID in CRLs.
1497 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1498 will ultimately be used for all verify operations: this will remove the
1499 X509_STORE dependency on certificate verification and allow alternative
1500 lookup methods. X509_STORE based implementations of these two callbacks.
1503 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1504 Modify get_crl() to find a valid (unexpired) CRL if possible.
1507 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1508 this would be called X509_CRL_cmp() but that name is already used by
1509 a function that just compares CRL issuer names. Cache several CRL
1510 extensions in X509_CRL structure and cache CRLDP in X509.
1513 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1514 this maps equivalent X509_NAME structures into a consistent structure.
1515 Name comparison can then be performed rapidly using memcmp().
1518 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1522 *) Allow digests to supply their own micalg string for S/MIME type using
1523 the ctrl EVP_MD_CTRL_MICALG.
1526 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1527 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1528 ctrl. It can then customise the structure before and/or after signing
1532 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1533 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1534 to free up any added signature OIDs.
1537 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1538 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1539 digest and cipher tables. New options added to openssl utility:
1540 list-message-digest-algorithms and list-cipher-algorithms.
1543 *) Change the array representation of binary polynomials: the list
1544 of degrees of non-zero coefficients is now terminated with -1.
1545 Previously it was terminated with 0, which was also part of the
1546 value; thus, the array representation was not applicable to
1547 polynomials where t^0 has coefficient zero. This change makes
1548 the array representation useful in a more general context.
1551 *) Various modifications and fixes to SSL/TLS cipher string
1552 handling. For ECC, the code now distinguishes between fixed ECDH
1553 with RSA certificates on the one hand and with ECDSA certificates
1554 on the other hand, since these are separate ciphersuites. The
1555 unused code for Fortezza ciphersuites has been removed.
1557 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1558 (not "ECDHE"). For consistency with the code for DH
1559 certificates, use of ECDH certificates is now considered ECDH
1560 authentication, not RSA or ECDSA authentication (the latter is
1561 merely the CA's signing algorithm and not actively used in the
1564 The temporary ciphersuite alias "ECCdraft" is no longer
1565 available, and ECC ciphersuites are no longer excluded from "ALL"
1566 and "DEFAULT". The following aliases now exist for RFC 4492
1567 ciphersuites, most of these by analogy with the DH case:
1569 kECDHr - ECDH cert, signed with RSA
1570 kECDHe - ECDH cert, signed with ECDSA
1571 kECDH - ECDH cert (signed with either RSA or ECDSA)
1572 kEECDH - ephemeral ECDH
1573 ECDH - ECDH cert or ephemeral ECDH
1579 AECDH - anonymous ECDH
1580 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1584 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1585 Use correct micalg parameters depending on digest(s) in signed message.
1588 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1589 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1592 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1593 an engine to register a method. Add ENGINE lookups for methods and
1594 functional reference processing.
1597 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1598 EVP_{Sign,Verify}* which allow an application to customise the signature
1602 *) New -resign option to smime utility. This adds one or more signers
1603 to an existing PKCS#7 signedData structure. Also -md option to use an
1604 alternative message digest algorithm for signing.
1607 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1608 create PKCS7 structures containing multiple signers. Update smime
1609 application to support multiple signers.
1612 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1616 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1617 Reorganize PBE internals to lookup from a static table using NIDs,
1618 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1619 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1620 PRF which will be automatically used with PBES2.
1623 *) Replace the algorithm specific calls to generate keys in "req" with the
1627 *) Update PKCS#7 enveloped data routines to use new API. This is now
1628 supported by any public key method supporting the encrypt operation. A
1629 ctrl is added to allow the public key algorithm to examine or modify
1630 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1634 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1635 a default digest type to use. In most cases this will be SHA1 but some
1636 algorithms (such as GOST) need to specify an alternative digest. The
1637 return value indicates how strong the prefernce is 1 means optional and
1638 2 is mandatory (that is it is the only supported type). Modify
1639 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1640 use the default md. Update openssl utilities to use the default digest
1641 type for signing if it is not explicitly indicated.
1644 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1645 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1646 signing method from the key type. This effectively removes the link
1647 between digests and public key types.
1650 *) Add an OID cross reference table and utility functions. Its purpose is to
1651 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1652 rsaEncryption. This will allow some of the algorithm specific hackery
1653 needed to use the correct OID to be removed.
1656 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1657 structures for PKCS7_sign(). They are now set up by the relevant public
1661 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1664 *) Add support for key derivation (agreement) in the API, DH method and
1668 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1669 public and private key formats. As a side effect these add additional
1670 command line functionality not previously available: DSA signatures can be
1671 generated and verified using pkeyutl and DH key support and generation in
1676 [Oliver Tappe <zooey@hirschkaefer.de>]
1678 *) New make target "install_html_docs" installs HTML renditions of the
1680 [Oliver Tappe <zooey@hirschkaefer.de>]
1682 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1683 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1684 support key and parameter generation and add initial key generation
1685 functionality for RSA.
1688 *) Add functions for main EVP_PKEY_method operations. The undocumented
1689 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1690 EVP_PKEY_{encrypt,decrypt}_old.
1693 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1694 key API, doesn't do much yet.
1697 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1698 public key algorithms. New option to openssl utility:
1699 "list-public-key-algorithms" to print out info.
1702 *) Implement the Supported Elliptic Curves Extension for
1703 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1706 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1707 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1710 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1711 utilities such as rsa, dsa, dsaparam etc except they process any key
1715 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1716 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1717 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1721 *) Initial support for pluggable public key ASN1.
1722 De-spaghettify the public key ASN1 handling. Move public and private
1723 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1724 algorithm specific handling to a single module within the relevant
1725 algorithm directory. Add functions to allow (near) opaque processing
1726 of public and private key structures.
1729 *) Implement the Supported Point Formats Extension for
1730 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1733 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1734 for the psk identity [hint] and the psk callback functions to the
1735 SSL_SESSION, SSL and SSL_CTX structure.
1738 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1742 SSL_CTX_use_psk_identity_hint
1743 SSL_get_psk_identity_hint
1744 SSL_get_psk_identity
1745 SSL_use_psk_identity_hint
1747 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1749 *) Add RFC 3161 compliant time stamp request creation, response generation
1750 and response verification functionality.
1751 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1753 *) Add initial support for TLS extensions, specifically for the server_name
1754 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1755 have new members for a host name. The SSL data structure has an
1756 additional member SSL_CTX *initial_ctx so that new sessions can be
1757 stored in that context to allow for session resumption, even after the
1758 SSL has been switched to a new SSL_CTX in reaction to a client's
1759 server_name extension.
1761 New functions (subject to change):
1763 SSL_get_servername()
1764 SSL_get_servername_type()
1767 New CTRL codes and macros (subject to change):
1769 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1770 - SSL_CTX_set_tlsext_servername_callback()
1771 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1772 - SSL_CTX_set_tlsext_servername_arg()
1773 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1775 openssl s_client has a new '-servername ...' option.
1777 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1778 '-key2 ...', '-servername_fatal' (subject to change). This allows
1779 testing the HostName extension for a specific single host name ('-cert'
1780 and '-key' remain fallbacks for handshakes without HostName
1781 negotiation). If the unrecogninzed_name alert has to be sent, this by
1782 default is a warning; it becomes fatal with the '-servername_fatal'
1785 [Peter Sylvester, Remy Allais, Christophe Renou]
1787 *) Whirlpool hash implementation is added.
1790 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1791 bn(64,32). Because of instruction set limitations it doesn't have
1792 any negative impact on performance. This was done mostly in order
1793 to make it possible to share assembler modules, such as bn_mul_mont
1794 implementations, between 32- and 64-bit builds without hassle.
1797 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1798 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1802 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1803 dedicated Montgomery multiplication procedure, is introduced.
1804 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1805 "64-bit" performance on certain 32-bit targets.
1808 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1809 in SSL structures. New SSL ctrl to set maximum send fragment size.
1810 Save memory by seeting the I/O buffer sizes dynamically instead of
1811 using the maximum available value.
1814 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1815 in addition to the text details.
1818 *) Very, very preliminary EXPERIMENTAL support for printing of general
1819 ASN1 structures. This currently produces rather ugly output and doesn't
1820 handle several customised structures at all.
1823 *) Integrated support for PVK file format and some related formats such
1824 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1825 these in the 'rsa' and 'dsa' utilities.
1828 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1831 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1832 place for the (very old) "NETSCAPE" format certificates which are now
1833 handled using new ASN1 code equivalents.
1836 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1837 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1838 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1841 *) Modify CRL distribution points extension code to print out previously
1842 unsupported fields. Enhance extension setting code to allow setting of
1846 *) Add print and set support for Issuing Distribution Point CRL extension.
1849 *) Change 'Configure' script to enable Camellia by default.
1852 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1854 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1856 This addresses the flaw in CBC record processing discovered by
1857 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1858 at: http://www.isg.rhul.ac.uk/tls/
1860 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1861 Security Group at Royal Holloway, University of London
1862 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1863 Emilia Käsper for the initial patch.
1865 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1867 *) Return an error when checking OCSP signatures when key is NULL.
1868 This fixes a DoS attack. (CVE-2013-0166)
1871 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1872 the right response is stapled. Also change SSL_get_certificate()
1873 so it returns the certificate actually sent.
1874 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1875 (This is a backport)
1876 [Rob Stradling <rob.stradling@comodo.com>]
1878 *) Fix possible deadlock when decoding public keys.
1881 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1883 *) Sanity check record length before skipping explicit IV in DTLS
1886 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1887 fuzzing as a service testing platform.
1891 *) Initialise tkeylen properly when encrypting CMS messages.
1892 Thanks to Solar Designer of Openwall for reporting this issue.
1895 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1897 *) The fix for CVE-2012-2110 did not take into account that the
1898 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1899 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1900 rejecting negative len parameter. (CVE-2012-2131)
1901 [Tomas Hoger <thoger@redhat.com>]
1903 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1905 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1906 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1907 in CRYPTO_realloc_clean.
1909 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1910 issue and to Adam Langley <agl@chromium.org> for fixing it.
1912 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1914 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1916 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1917 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1918 content decryption and always return the same error. Note: this attack
1919 needs on average 2^20 messages so it only affects automated senders. The
1920 old behaviour can be reenabled in the CMS code by setting the
1921 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1922 an MMA defence is not necessary.
1923 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1924 this issue. (CVE-2012-0884)
1927 *) Fix CVE-2011-4619: make sure we really are receiving a
1928 client hello before rejecting multiple SGC restarts. Thanks to
1929 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1932 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1934 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1935 Thanks to Antonio Martin, Enterprise Secure Access Research and
1936 Development, Cisco Systems, Inc. for discovering this bug and
1937 preparing a fix. (CVE-2012-0050)
1940 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1942 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1943 of the Vaudenay padding oracle attack on CBC mode encryption
1944 which enables an efficient plaintext recovery attack against
1945 the OpenSSL implementation of DTLS. Their attack exploits timing
1946 differences arising during decryption processing. A research
1947 paper describing this attack can be found at:
1948 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1949 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1950 Security Group at Royal Holloway, University of London
1951 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1952 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1953 for preparing the fix. (CVE-2011-4108)
1954 [Robin Seggelmann, Michael Tuexen]
1956 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1957 [Ben Laurie, Kasper <ekasper@google.com>]
1959 *) Clear bytes used for block padding of SSL 3.0 records.
1961 [Adam Langley (Google)]
1963 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1964 Kadianakis <desnacked@gmail.com> for discovering this issue and
1965 Adam Langley for preparing the fix. (CVE-2011-4619)
1966 [Adam Langley (Google)]
1968 *) Prevent malformed RFC3779 data triggering an assertion failure.
1969 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1970 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1971 [Rob Austein <sra@hactrn.net>]
1973 *) Fix ssl_ciph.c set-up race.
1974 [Adam Langley (Google)]
1976 *) Fix spurious failures in ecdsatest.c.
1977 [Emilia Käsper (Google)]
1979 *) Fix the BIO_f_buffer() implementation (which was mixing different
1980 interpretations of the '..._len' fields).
1981 [Adam Langley (Google)]
1983 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1984 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1985 threads won't reuse the same blinding coefficients.
1987 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1988 lock to call BN_BLINDING_invert_ex, and avoids one use of
1989 BN_BLINDING_update for each BN_BLINDING structure (previously,
1990 the last update always remained unused).
1991 [Emilia Käsper (Google)]
1993 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1994 for multi-threaded use of ECDH.
1995 [Adam Langley (Google)]
1997 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2000 *) Add protection against ECDSA timing attacks as mentioned in the paper
2001 by Billy Bob Brumley and Nicola Tuveri, see:
2003 http://eprint.iacr.org/2011/232.pdf
2005 [Billy Bob Brumley and Nicola Tuveri]
2007 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2009 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2010 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2012 *) Fix bug in string printing code: if *any* escaping is enabled we must
2013 escape the escape character (backslash) or the resulting string is
2017 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2019 *) Disable code workaround for ancient and obsolete Netscape browsers
2020 and servers: an attacker can use it in a ciphersuite downgrade attack.
2021 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2024 *) Fixed J-PAKE implementation error, originally discovered by
2025 Sebastien Martini, further info and confirmation from Stefan
2026 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2029 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2031 *) Fix extension code to avoid race conditions which can result in a buffer
2032 overrun vulnerability: resumed sessions must not be modified as they can
2033 be shared by multiple threads. CVE-2010-3864
2036 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2039 *) Don't reencode certificate when calculating signature: cache and use
2040 the original encoding instead. This makes signature verification of
2041 some broken encodings work correctly.
2044 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2045 is also one of the inputs.
2046 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2048 *) Don't repeatedly append PBE algorithms to table if they already exist.
2049 Sort table on each new add. This effectively makes the table read only
2050 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2054 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2056 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2059 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2060 access or freeing data twice (CVE-2010-0742)
2061 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2063 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2064 common in certificates and some applications which only call
2065 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2069 Reduce copying into .apps and .test in makevms.com
2070 Don't try to use blank CA certificate in CA.com
2071 Allow use of C files from original directories in maketests.com
2072 [Steven M. Schweda" <sms@antinode.info>]
2074 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2076 *) When rejecting SSL/TLS records due to an incorrect version number, never
2077 update s->server with a new major version number. As of
2078 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2079 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2080 the previous behavior could result in a read attempt at NULL when
2081 receiving specific incorrect SSL/TLS records once record payload
2082 protection is active. (CVE-2010-0740)
2083 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2085 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2086 could be crashed if the relevant tables were not present (e.g. chrooted).
2087 [Tomas Hoger <thoger@redhat.com>]
2089 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2091 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2092 [Martin Olsson, Neel Mehta]
2094 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2095 accommodate for stack sorting, always a write lock!).
2098 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2099 excessive delays in the RAND_poll(): over a minute. As a workaround
2100 include a time check in the inner Heap32Next loop too.
2103 *) The code that handled flushing of data in SSL/TLS originally used the
2104 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2105 the problem outlined in PR#1949. The fix suggested there however can
2106 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2107 of Apache). So instead simplify the code to flush unconditionally.
2108 This should be fine since flushing with no data to flush is a no op.
2111 *) Handle TLS versions 2.0 and later properly and correctly use the
2112 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2113 off ancient servers have a habit of sticking around for a while...
2116 *) Modify compression code so it frees up structures without using the
2117 ex_data callbacks. This works around a problem where some applications
2118 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2119 restarting) then use compression (e.g. SSL with compression) later.
2120 This results in significant per-connection memory leaks and
2121 has caused some security issues including CVE-2008-1678 and
2125 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2126 change when encrypting or decrypting.
2129 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2130 connect and renegotiate with servers which do not support RI.
2131 Until RI is more widely deployed this option is enabled by default.
2134 *) Add "missing" ssl ctrls to clear options and mode.
2137 *) If client attempts to renegotiate and doesn't support RI respond with
2138 a no_renegotiation alert as required by RFC5746. Some renegotiating
2139 TLS clients will continue a connection gracefully when they receive
2140 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2141 waiting for a server hello which it will never receive. Now we treat a
2142 received no_renegotiation alert as a fatal error. This is because
2143 applications requesting a renegotiation might well expect it to succeed
2144 and would have no code in place to handle the server denying it so the
2145 only safe thing to do is to terminate the connection.
2148 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2149 peer supports secure renegotiation and 0 otherwise. Print out peer
2150 renegotiation support in s_client/s_server.
2153 *) Replace the highly broken and deprecated SPKAC certification method with
2154 the updated NID creation version. This should correctly handle UTF8.
2157 *) Implement RFC5746. Re-enable renegotiation but require the extension
2158 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2159 turns out to be a bad idea. It has been replaced by
2160 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2161 SSL_CTX_set_options(). This is really not recommended unless you
2162 know what you are doing.
2163 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2165 *) Fixes to stateless session resumption handling. Use initial_ctx when
2166 issuing and attempting to decrypt tickets in case it has changed during
2167 servername handling. Use a non-zero length session ID when attempting
2168 stateless session resumption: this makes it possible to determine if
2169 a resumption has occurred immediately after receiving server hello
2170 (several places in OpenSSL subtly assume this) instead of later in
2174 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2175 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2176 fixes for a few places where the return code is not checked
2178 [Julia Lawall <julia@diku.dk>]
2180 *) Add --strict-warnings option to Configure script to include devteam
2181 warnings in other configurations.
2184 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2185 makes it possible to install openssl libraries in locations which
2186 have names other than "lib", for example "/usr/lib64" which some
2188 [Steve Henson, based on patch from Jeremy Utley]
2190 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2191 X690 8.9.12 and can produce some misleading textual output of OIDs.
2192 [Steve Henson, reported by Dan Kaminsky]
2194 *) Delete MD2 from algorithm tables. This follows the recommendation in
2195 several standards that it is not used in new applications due to
2196 several cryptographic weaknesses. For binary compatibility reasons
2197 the MD2 API is still compiled in by default.
2200 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2204 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2205 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2207 [Guenter <lists@gknw.net>]
2209 *) Fix the server certificate chain building code to use X509_verify_cert(),
2210 it used to have an ad-hoc builder which was unable to cope with anything
2211 other than a simple chain.
2212 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2214 *) Don't check self signed certificate signatures in X509_verify_cert()
2215 by default (a flag can override this): it just wastes time without
2216 adding any security. As a useful side effect self signed root CAs
2217 with non-FIPS digests are now usable in FIPS mode.
2220 *) In dtls1_process_out_of_seq_message() the check if the current message
2221 is already buffered was missing. For every new message was memory
2222 allocated, allowing an attacker to perform an denial of service attack
2223 with sending out of seq handshake messages until there is no memory
2224 left. Additionally every future messege was buffered, even if the
2225 sequence number made no sense and would be part of another handshake.
2226 So only messages with sequence numbers less than 10 in advance will be
2227 buffered. (CVE-2009-1378)
2228 [Robin Seggelmann, discovered by Daniel Mentz]
2230 *) Records are buffered if they arrive with a future epoch to be
2231 processed after finishing the corresponding handshake. There is
2232 currently no limitation to this buffer allowing an attacker to perform
2233 a DOS attack with sending records with future epochs until there is no
2234 memory left. This patch adds the pqueue_size() function to detemine
2235 the size of a buffer and limits the record buffer to 100 entries.
2237 [Robin Seggelmann, discovered by Daniel Mentz]
2239 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2240 parent structure is freed. (CVE-2009-1379)
2243 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2244 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2247 [Ilya O. <vrghost@gmail.com>]
2249 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2251 *) Disable renegotiation completely - this fixes a severe security
2252 problem (CVE-2009-3555) at the cost of breaking all
2253 renegotiation. Renegotiation can be re-enabled by setting
2254 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2255 run-time. This is really not recommended unless you know what
2259 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2261 *) Don't set val to NULL when freeing up structures, it is freed up by
2262 underlying code. If sizeof(void *) > sizeof(long) this can result in
2263 zeroing past the valid field. (CVE-2009-0789)
2264 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2266 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2267 checked correctly. This would allow some invalid signed attributes to
2268 appear to verify correctly. (CVE-2009-0591)
2269 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2271 *) Reject UniversalString and BMPString types with invalid lengths. This
2272 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2273 a legal length. (CVE-2009-0590)
2276 *) Set S/MIME signing as the default purpose rather than setting it
2277 unconditionally. This allows applications to override it at the store
2281 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2282 to handle some structures.
2285 *) Improve efficiency of mem_gets: don't search whole buffer each time
2287 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2289 *) New -hex option for openssl rand.
2292 *) Print out UTF8String and NumericString when parsing ASN1.
2295 *) Support NumericString type for name components.
2298 *) Allow CC in the environment to override the automatically chosen
2299 compiler. Note that nothing is done to ensure flags work with the
2303 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2305 *) Properly check EVP_VerifyFinal() and similar return values
2307 [Ben Laurie, Bodo Moeller, Google Security Team]
2309 *) Enable TLS extensions by default.
2312 *) Allow the CHIL engine to be loaded, whether the application is
2313 multithreaded or not. (This does not release the developer from the
2314 obligation to set up the dynamic locking callbacks.)
2315 [Sander Temme <sander@temme.net>]
2317 *) Use correct exit code if there is an error in dgst command.
2318 [Steve Henson; problem pointed out by Roland Dirlewanger]
2320 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2321 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2324 *) Add experimental JPAKE support, including demo authentication in
2325 s_client and s_server.
2328 *) Set the comparison function in v3_addr_canonize().
2329 [Rob Austein <sra@hactrn.net>]
2331 *) Add support for XMPP STARTTLS in s_client.
2332 [Philip Paeps <philip@freebsd.org>]
2334 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2335 to ensure that even with this option, only ciphersuites in the
2336 server's preference list will be accepted. (Note that the option
2337 applies only when resuming a session, so the earlier behavior was
2338 just about the algorithm choice for symmetric cryptography.)
2341 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2343 *) Fix NULL pointer dereference if a DTLS server received
2344 ChangeCipherSpec as first record (CVE-2009-1386).
2347 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2348 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2351 *) The fix in 0.9.8c that supposedly got rid of unsafe
2352 double-checked locking was incomplete for RSA blinding,
2353 addressing just one layer of what turns out to have been
2354 doubly unsafe triple-checked locking.
2356 So now fix this for real by retiring the MONT_HELPER macro
2357 in crypto/rsa/rsa_eay.c.
2359 [Bodo Moeller; problem pointed out by Marius Schilder]
2361 *) Various precautionary measures:
2363 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2365 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2366 (NB: This would require knowledge of the secret session ticket key
2367 to exploit, in which case you'd be SOL either way.)
2369 - Change bn_nist.c so that it will properly handle input BIGNUMs
2370 outside the expected range.
2372 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2375 [Neel Mehta, Bodo Moeller]
2377 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2378 the load fails. Useful for distros.
2379 [Ben Laurie and the FreeBSD team]
2381 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2384 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2387 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2389 This work was sponsored by Logica.
2392 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2393 keystores. Support for SSL/TLS client authentication too.
2394 Not compiled unless enable-capieng specified to Configure.
2396 This work was sponsored by Logica.
2399 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2400 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2401 attribute creation routines such as certifcate requests and PKCS#12
2405 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2407 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2408 handshake which could lead to a cilent crash as found using the
2409 Codenomicon TLS test suite (CVE-2008-1672)
2410 [Steve Henson, Mark Cox]
2412 *) Fix double free in TLS server name extensions which could lead to
2413 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2416 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2418 Clear the error queue to ensure that error entries left from
2419 older function calls do not interfere with the correct operation.
2420 [Lutz Jaenicke, Erik de Castro Lopo]
2422 *) Remove root CA certificates of commercial CAs:
2424 The OpenSSL project does not recommend any specific CA and does not
2425 have any policy with respect to including or excluding any CA.
2426 Therefore it does not make any sense to ship an arbitrary selection
2427 of root CA certificates with the OpenSSL software.
2430 *) RSA OAEP patches to fix two separate invalid memory reads.
2431 The first one involves inputs when 'lzero' is greater than
2432 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2433 before the beginning of from). The second one involves inputs where
2434 the 'db' section contains nothing but zeroes (there is a one-byte
2435 invalid read after the end of 'db').
2436 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2438 *) Partial backport from 0.9.9-dev:
2440 Introduce bn_mul_mont (dedicated Montgomery multiplication
2441 procedure) as a candidate for BIGNUM assembler implementation.
2442 While 0.9.9-dev uses assembler for various architectures, only
2443 x86_64 is available by default here in the 0.9.8 branch, and
2444 32-bit x86 is available through a compile-time setting.
2446 To try the 32-bit x86 assembler implementation, use Configure
2447 option "enable-montasm" (which exists only for this backport).
2449 As "enable-montasm" for 32-bit x86 disclaims code stability
2450 anyway, in this constellation we activate additional code
2451 backported from 0.9.9-dev for further performance improvements,
2452 namely BN_from_montgomery_word. (To enable this otherwise,
2453 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2455 [Andy Polyakov (backport partially by Bodo Moeller)]
2457 *) Add TLS session ticket callback. This allows an application to set
2458 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2459 values. This is useful for key rollover for example where several key
2460 sets may exist with different names.
2463 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2464 This was broken until now in 0.9.8 releases, such that the only way
2465 a registered ENGINE could be used (assuming it initialises
2466 successfully on the host) was to explicitly set it as the default
2467 for the relevant algorithms. This is in contradiction with 0.9.7
2468 behaviour and the documentation. With this fix, when an ENGINE is
2469 registered into a given algorithm's table of implementations, the
2470 'uptodate' flag is reset so that auto-discovery will be used next
2471 time a new context for that algorithm attempts to select an
2473 [Ian Lister (tweaked by Geoff Thorpe)]
2475 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2476 implemention in the following ways:
2478 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2481 Lack of BER streaming support means one pass streaming processing is
2482 only supported if data is detached: setting the streaming flag is
2483 ignored for embedded content.
2485 CMS support is disabled by default and must be explicitly enabled
2486 with the enable-cms configuration option.
2489 *) Update the GMP engine glue to do direct copies between BIGNUM and
2490 mpz_t when openssl and GMP use the same limb size. Otherwise the
2491 existing "conversion via a text string export" trick is still used.
2492 [Paul Sheer <paulsheer@gmail.com>]
2494 *) Zlib compression BIO. This is a filter BIO which compressed and
2495 uncompresses any data passed through it.
2498 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2499 RFC3394 compatible AES key wrapping.
2502 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2503 sets string data without copying. X509_ALGOR_set0() and
2504 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2505 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2506 from an X509_ATTRIBUTE structure optionally checking it occurs only
2507 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2511 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2512 to get the expected BN_FLG_CONSTTIME behavior.
2513 [Bodo Moeller (Google)]
2517 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2518 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2519 - added some more tests to do_tests.pl
2520 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2521 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2522 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2523 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2524 - various changes to netware.pl to enable gcc-cross builds on Win32
2526 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2527 - various changes to fix missing prototype warnings
2528 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2529 - added AES, WHIRLPOOL and CPUID assembler code to build files
2530 - added missing AES assembler make rules to mk1mf.pl
2531 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2532 [Guenter Knauf <eflash@gmx.net>]
2534 *) Implement certificate status request TLS extension defined in RFC3546.
2535 A client can set the appropriate parameters and receive the encoded
2536 OCSP response via a callback. A server can query the supplied parameters
2537 and set the encoded OCSP response in the callback. Add simplified examples
2538 to s_client and s_server.
2541 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2543 *) Fix various bugs:
2544 + Binary incompatibility of ssl_ctx_st structure
2545 + DTLS interoperation with non-compliant servers
2546 + Don't call get_session_cb() without proposed session
2547 + Fix ia64 assembler code
2548 [Andy Polyakov, Steve Henson]
2550 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2552 *) DTLS Handshake overhaul. There were longstanding issues with
2553 OpenSSL DTLS implementation, which were making it impossible for
2554 RFC 4347 compliant client to communicate with OpenSSL server.
2555 Unfortunately just fixing these incompatibilities would "cut off"
2556 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2557 server keeps tolerating non RFC compliant syntax. The opposite is
2558 not true, 0.9.8f client can not communicate with earlier server.
2559 This update even addresses CVE-2007-4995.
2562 *) Changes to avoid need for function casts in OpenSSL: some compilers
2563 (gcc 4.2 and later) reject their use.
2564 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2567 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2568 RFC4507bis. The encrypted ticket format is an encrypted encoded
2569 SSL_SESSION structure, that way new session features are automatically
2572 If a client application caches session in an SSL_SESSION structure
2573 support is transparent because tickets are now stored in the encoded
2576 The SSL_CTX structure automatically generates keys for ticket
2577 protection in servers so again support should be possible
2578 with no application modification.
2580 If a client or server wishes to disable RFC4507 support then the option
2581 SSL_OP_NO_TICKET can be set.
2583 Add a TLS extension debugging callback to allow the contents of any client
2584 or server extensions to be examined.
2586 This work was sponsored by Google.
2589 *) Add initial support for TLS extensions, specifically for the server_name
2590 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2591 have new members for a host name. The SSL data structure has an
2592 additional member SSL_CTX *initial_ctx so that new sessions can be
2593 stored in that context to allow for session resumption, even after the
2594 SSL has been switched to a new SSL_CTX in reaction to a client's
2595 server_name extension.
2597 New functions (subject to change):
2599 SSL_get_servername()
2600 SSL_get_servername_type()
2603 New CTRL codes and macros (subject to change):
2605 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2606 - SSL_CTX_set_tlsext_servername_callback()
2607 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2608 - SSL_CTX_set_tlsext_servername_arg()
2609 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2611 openssl s_client has a new '-servername ...' option.
2613 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2614 '-key2 ...', '-servername_fatal' (subject to change). This allows
2615 testing the HostName extension for a specific single host name ('-cert'
2616 and '-key' remain fallbacks for handshakes without HostName
2617 negotiation). If the unrecogninzed_name alert has to be sent, this by
2618 default is a warning; it becomes fatal with the '-servername_fatal'
2621 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2623 *) Add AES and SSE2 assembly language support to VC++ build.
2626 *) Mitigate attack on final subtraction in Montgomery reduction.
2629 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2630 (which previously caused an internal error).
2633 *) Squeeze another 10% out of IGE mode when in != out.
2636 *) AES IGE mode speedup.
2637 [Dean Gaudet (Google)]
2639 *) Add the Korean symmetric 128-bit cipher SEED (see
2640 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2641 add SEED ciphersuites from RFC 4162:
2643 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2644 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2645 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2646 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2648 To minimize changes between patchlevels in the OpenSSL 0.9.8
2649 series, SEED remains excluded from compilation unless OpenSSL
2650 is configured with 'enable-seed'.
2651 [KISA, Bodo Moeller]
2653 *) Mitigate branch prediction attacks, which can be practical if a
2654 single processor is shared, allowing a spy process to extract
2655 information. For detailed background information, see
2656 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2657 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2658 and Necessary Software Countermeasures"). The core of the change
2659 are new versions BN_div_no_branch() and
2660 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2661 respectively, which are slower, but avoid the security-relevant
2662 conditional branches. These are automatically called by BN_div()
2663 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2664 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2665 remove a conditional branch.
2667 BN_FLG_CONSTTIME is the new name for the previous
2668 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2669 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2670 in the exponent causes BN_mod_exp_mont() to use the alternative
2671 implementation in BN_mod_exp_mont_consttime().) The old name
2672 remains as a deprecated alias.
2674 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2675 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2676 constant-time implementations for more than just exponentiation.
2677 Here too the old name is kept as a deprecated alias.
2679 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2680 the BN_BLINDING structure gets an independent copy of the
2681 modulus. This means that the previous "BIGNUM *m" argument to
2682 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2683 essentially becomes "const BIGNUM *m", although we can't actually
2684 change this in the header file before 0.9.9. It allows
2685 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2686 enable BN_FLG_CONSTTIME.
2688 [Matthew D Wood (Intel Corp)]
2690 *) In the SSL/TLS server implementation, be strict about session ID
2691 context matching (which matters if an application uses a single
2692 external cache for different purposes). Previously,
2693 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2694 set. This did ensure strict client verification, but meant that,
2695 with applications using a single external cache for quite
2696 different requirements, clients could circumvent ciphersuite
2697 restrictions for a given session ID context by starting a session
2698 in a different context.
2701 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2702 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2703 authentication-only ciphersuites.
2706 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2707 not complete and could lead to a possible single byte overflow
2708 (CVE-2007-5135) [Ben Laurie]
2710 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2712 *) Since AES128 and AES256 (and similarly Camellia128 and
2713 Camellia256) share a single mask bit in the logic of
2714 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2715 kludge to work properly if AES128 is available and AES256 isn't
2716 (or if Camellia128 is available and Camellia256 isn't).
2719 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2720 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2721 When a point or a seed is encoded in a BIT STRING, we need to
2722 prevent the removal of trailing zero bits to get the proper DER
2723 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2724 of a NamedBitList, for which trailing 0 bits need to be removed.)
2727 *) Have SSL/TLS server implementation tolerate "mismatched" record
2728 protocol version while receiving ClientHello even if the
2729 ClientHello is fragmented. (The server can't insist on the
2730 particular protocol version it has chosen before the ServerHello
2731 message has informed the client about his choice.)
2734 *) Add RFC 3779 support.
2735 [Rob Austein for ARIN, Ben Laurie]
2737 *) Load error codes if they are not already present instead of using a
2738 static variable. This allows them to be cleanly unloaded and reloaded.
2739 Improve header file function name parsing.
2742 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2743 or CAPABILITY handshake as required by RFCs.
2746 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2748 *) Introduce limits to prevent malicious keys being able to
2749 cause a denial of service. (CVE-2006-2940)
2750 [Steve Henson, Bodo Moeller]
2752 *) Fix ASN.1 parsing of certain invalid structures that can result
2753 in a denial of service. (CVE-2006-2937) [Steve Henson]
2755 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2756 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2758 *) Fix SSL client code which could crash if connecting to a
2759 malicious SSLv2 server. (CVE-2006-4343)
2760 [Tavis Ormandy and Will Drewry, Google Security Team]
2762 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2763 match only those. Before that, "AES256-SHA" would be interpreted
2764 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2765 the same strength classification in 0.9.7h) as we currently only
2766 have a single AES bit in the ciphersuite description bitmap.
2767 That change, however, also applied to ciphersuite strings such as
2768 "RC4-MD5" that intentionally matched multiple ciphersuites --
2769 namely, SSL 2.0 ciphersuites in addition to the more common ones
2770 from SSL 3.0/TLS 1.0.
2772 So we change the selection algorithm again: Naming an explicit
2773 ciphersuite selects this one ciphersuite, and any other similar
2774 ciphersuite (same bitmap) from *other* protocol versions.
2775 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2776 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2778 Since SSL 2.0 does not have any ciphersuites for which the
2779 128/256 bit distinction would be relevant, this works for now.
2780 The proper fix will be to use different bits for AES128 and
2781 AES256, which would have avoided the problems from the beginning;
2782 however, bits are scarce, so we can only do this in a new release
2783 (not just a patchlevel) when we can change the SSL_CIPHER
2784 definition to split the single 'unsigned long mask' bitmap into
2785 multiple values to extend the available space.
2789 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2791 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2792 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2794 *) Add AES IGE and biIGE modes.
2797 *) Change the Unix randomness entropy gathering to use poll() when
2798 possible instead of select(), since the latter has some
2799 undesirable limitations.
2800 [Darryl Miles via Richard Levitte and Bodo Moeller]
2802 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2803 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2804 cannot be implicitly activated as part of, e.g., the "AES" alias.
2805 However, please upgrade to OpenSSL 0.9.9[-dev] for
2806 non-experimental use of the ECC ciphersuites to get TLS extension
2807 support, which is required for curve and point format negotiation
2808 to avoid potential handshake problems.
2811 *) Disable rogue ciphersuites:
2813 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2814 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2815 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2817 The latter two were purportedly from
2818 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2821 Also deactivate the remaining ciphersuites from
2822 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2823 unofficial, and the ID has long expired.
2826 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2827 dual-core machines) and other potential thread-safety issues.
2830 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2831 versions), which is now available for royalty-free use
2832 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2833 Also, add Camellia TLS ciphersuites from RFC 4132.
2835 To minimize changes between patchlevels in the OpenSSL 0.9.8
2836 series, Camellia remains excluded from compilation unless OpenSSL
2837 is configured with 'enable-camellia'.
2840 *) Disable the padding bug check when compression is in use. The padding
2841 bug check assumes the first packet is of even length, this is not
2842 necessarily true if compresssion is enabled and can result in false
2843 positives causing handshake failure. The actual bug test is ancient
2844 code so it is hoped that implementations will either have fixed it by
2845 now or any which still have the bug do not support compression.
2848 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2850 *) When applying a cipher rule check to see if string match is an explicit
2851 cipher suite and only match that one cipher suite if it is.
2854 *) Link in manifests for VC++ if needed.
2855 [Austin Ziegler <halostatue@gmail.com>]
2857 *) Update support for ECC-based TLS ciphersuites according to
2858 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2859 TLS extensions, which are supported starting with the 0.9.9
2860 branch, not in the OpenSSL 0.9.8 branch).
2863 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2864 opaque EVP_CIPHER_CTX handling.
2867 *) Fixes and enhancements to zlib compression code. We now only use
2868 "zlib1.dll" and use the default __cdecl calling convention on Win32
2869 to conform with the standards mentioned here:
2870 http://www.zlib.net/DLL_FAQ.txt
2871 Static zlib linking now works on Windows and the new --with-zlib-include
2872 --with-zlib-lib options to Configure can be used to supply the location
2873 of the headers and library. Gracefully handle case where zlib library
2877 *) Several fixes and enhancements to the OID generation code. The old code
2878 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2879 handle numbers larger than ULONG_MAX, truncated printing and had a
2880 non standard OBJ_obj2txt() behaviour.
2883 *) Add support for building of engines under engine/ as shared libraries
2884 under VC++ build system.
2887 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2888 Hopefully, we will not see any false combination of paths any more.
2891 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2893 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2894 (part of SSL_OP_ALL). This option used to disable the
2895 countermeasure against man-in-the-middle protocol-version
2896 rollback in the SSL 2.0 server implementation, which is a bad
2897 idea. (CVE-2005-2969)
2899 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2900 for Information Security, National Institute of Advanced Industrial
2901 Science and Technology [AIST], Japan)]
2903 *) Add two function to clear and return the verify parameter flags.
2906 *) Keep cipherlists sorted in the source instead of sorting them at
2907 runtime, thus removing the need for a lock.
2910 *) Avoid some small subgroup attacks in Diffie-Hellman.
2911 [Nick Mathewson and Ben Laurie]
2913 *) Add functions for well-known primes.
2916 *) Extended Windows CE support.
2917 [Satoshi Nakamura and Andy Polyakov]
2919 *) Initialize SSL_METHOD structures at compile time instead of during
2920 runtime, thus removing the need for a lock.
2923 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2924 attempting to decrypt each encrypted key in turn. Add support to
2928 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2930 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2933 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2936 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2937 key into the same file any more.
2940 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2943 *) Add -utf8 command line and config file option to 'ca'.
2944 [Stefan <stf@udoma.org]
2946 *) Removed the macro des_crypt(), as it seems to conflict with some
2947 libraries. Use DES_crypt().
2950 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2951 involves renaming the source and generated shared-libs for
2952 both. The engines will accept the corrected or legacy ids
2953 ('ncipher' and '4758_cca' respectively) when binding. NB,
2954 this only applies when building 'shared'.
2955 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2957 *) Add attribute functions to EVP_PKEY structure. Modify
2958 PKCS12_create() to recognize a CSP name attribute and
2959 use it. Make -CSP option work again in pkcs12 utility.
2962 *) Add new functionality to the bn blinding code:
2963 - automatic re-creation of the BN_BLINDING parameters after
2964 a fixed number of uses (currently 32)
2965 - add new function for parameter creation
2966 - introduce flags to control the update behaviour of the
2967 BN_BLINDING parameters
2968 - hide BN_BLINDING structure
2969 Add a second BN_BLINDING slot to the RSA structure to improve
2970 performance when a single RSA object is shared among several
2974 *) Add support for DTLS.
2975 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2977 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2978 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2981 *) Remove buggy and incompletet DH cert support from
2982 ssl/ssl_rsa.c and ssl/s3_both.c
2985 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2986 the apps/openssl applications.
2989 *) Compile clean with "-Wall -Wmissing-prototypes
2990 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2991 DEBUG_SAFESTACK must also be set.
2994 *) Change ./Configure so that certain algorithms can be disabled by default.
2995 The new counterpiece to "no-xxx" is "enable-xxx".
2997 The patented RC5 and MDC2 algorithms will now be disabled unless
2998 "enable-rc5" and "enable-mdc2", respectively, are specified.
3000 (IDEA remains enabled despite being patented. This is because IDEA
3001 is frequently required for interoperability, and there is no license
3002 fee for non-commercial use. As before, "no-idea" can be used to
3003 avoid this algorithm.)
3007 *) Add processing of proxy certificates (see RFC 3820). This work was
3008 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3009 EGEE (Enabling Grids for E-science in Europe).
3012 *) RC4 performance overhaul on modern architectures/implementations, such
3013 as Intel P4, IA-64 and AMD64.
3016 *) New utility extract-section.pl. This can be used specify an alternative
3017 section number in a pod file instead of having to treat each file as
3018 a separate case in Makefile. This can be done by adding two lines to the
3021 =for comment openssl_section:XXX
3023 The blank line is mandatory.
3027 *) New arguments -certform, -keyform and -pass for s_client and s_server
3028 to allow alternative format key and certificate files and passphrase
3032 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3033 update associated structures and add various utility functions.
3035 Add new policy related verify parameters, include policy checking in
3036 standard verify code. Enhance 'smime' application with extra parameters
3037 to support policy checking and print out.
3040 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3041 Nehemiah processors. These extensions support AES encryption in hardware
3042 as well as RNG (though RNG support is currently disabled).
3043 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3045 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3048 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3049 [Andy Polyakov and a number of other people]
3051 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3052 implementation contributed by IBM.
3053 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3055 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3056 exponent rather than 'unsigned long'. There is a corresponding change to
3057 the new 'rsa_keygen' element of the RSA_METHOD structure.
3058 [Jelte Jansen, Geoff Thorpe]
3060 *) Functionality for creating the initial serial number file is now
3061 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3063 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3064 number file to 1, which is bound to cause problems. To avoid
3065 the problems while respecting compatibility between different 0.9.7
3066 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3067 CA.pl for serial number initialization. With the new release 0.9.8,
3068 we can fix the problem directly in the 'ca' utility.)
3071 *) Reduced header interdepencies by declaring more opaque objects in
3072 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3073 give fewer recursive includes, which could break lazy source code - so
3074 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3075 developers should define this symbol when building and using openssl to
3076 ensure they track the recommended behaviour, interfaces, [etc], but
3077 backwards-compatible behaviour prevails when this isn't defined.
3080 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3083 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3084 This will generate a random key of the appropriate length based on the
3085 cipher context. The EVP_CIPHER can provide its own random key generation
3086 routine to support keys of a specific form. This is used in the des and
3087 3des routines to generate a key of the correct parity. Update S/MIME
3088 code to use new functions and hence generate correct parity DES keys.
3089 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3090 valid (weak or incorrect parity).
3093 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3094 as looking them up. This is useful when the verified structure may contain
3095 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3096 present unless the new PKCS7_NO_CRL flag is asserted.
3099 *) Extend ASN1 oid configuration module. It now additionally accepts the
3102 shortName = some long name, 1.2.3.4
3105 *) Reimplemented the BN_CTX implementation. There is now no more static
3106 limitation on the number of variables it can handle nor the depth of the
3107 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3108 information can now expand as required, and rather than having a single
3109 static array of bignums, BN_CTX now uses a linked-list of such arrays
3110 allowing it to expand on demand whilst maintaining the usefulness of
3111 BN_CTX's "bundling".
3114 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3115 to allow all RSA operations to function using a single BN_CTX.
3118 *) Preliminary support for certificate policy evaluation and checking. This
3119 is initially intended to pass the tests outlined in "Conformance Testing
3120 of Relying Party Client Certificate Path Processing Logic" v1.07.
3123 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3124 remained unused and not that useful. A variety of other little bignum
3125 tweaks and fixes have also been made continuing on from the audit (see
3129 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3130 associated ASN1, EVP and SSL functions and old ASN1 macros.
3133 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3134 and this should never fail. So the return value from the use of
3135 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3136 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3139 *) BN_CTX_get() should return zero-valued bignums, providing the same
3140 initialised value as BN_new().
3141 [Geoff Thorpe, suggested by Ulf Möller]
3143 *) Support for inhibitAnyPolicy certificate extension.
3146 *) An audit of the BIGNUM code is underway, for which debugging code is
3147 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3148 is considered valid when processing BIGNUMs, and causes execution to
3149 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3150 further steps are taken to deliberately pollute unused data in BIGNUM
3151 structures to try and expose faulty code further on. For now, openssl will
3152 (in its default mode of operation) continue to tolerate the inconsistent
3153 forms that it has tolerated in the past, but authors and packagers should
3154 consider trying openssl and their own applications when compiled with
3155 these debugging symbols defined. It will help highlight potential bugs in
3156 their own code, and will improve the test coverage for OpenSSL itself. At
3157 some point, these tighter rules will become openssl's default to improve
3158 maintainability, though the assert()s and other overheads will remain only
3159 in debugging configurations. See bn.h for more details.
3160 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3162 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3163 that can only be obtained through BN_CTX_new() (which implicitly
3164 initialises it). The presence of this function only made it possible
3165 to overwrite an existing structure (and cause memory leaks).
3168 *) Because of the callback-based approach for implementing LHASH as a
3169 template type, lh_insert() adds opaque objects to hash-tables and
3170 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3171 to clean up those corresponding objects before destroying the hash table
3172 (and losing the object pointers). So some over-zealous constifications in
3173 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3174 objects as "const" and the lh_doall[_arg] callback wrappers are not
3175 prototyped to have "const" restrictions on the object pointers they are
3176 given (and so aren't required to cast them away any more).
3179 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3180 (speed) prefers to use its own implementation. The two implementations
3181 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3182 its object type properly exposed (MS_TM) instead of casting to/from "char
3183 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3184 aren't necessarily the greatest nomenclatures - but this is what was used
3185 internally to the implementation so I've used that for now.
3188 *) Ensure that deprecated functions do not get compiled when
3189 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3190 the self-tests were still using deprecated key-generation functions so
3191 these have been updated also.
3194 *) Reorganise PKCS#7 code to separate the digest location functionality
3195 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3196 New function PKCS7_set_digest() to set the digest type for PKCS#7
3197 digestedData type. Add additional code to correctly generate the
3198 digestedData type and add support for this type in PKCS7 initialization
3202 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3203 structure of type "other".
3206 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3207 sure the loop does correctly stop and breaking ("division by zero")
3208 modulus operations are not performed. The (pre-generated) prime
3209 table crypto/bn/bn_prime.h was already correct, but it could not be
3210 re-generated on some platforms because of the "division by zero"
3211 situation in the script.
3212 [Ralf S. Engelschall]
3214 *) Update support for ECC-based TLS ciphersuites according to
3215 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3216 SHA-1 now is only used for "small" curves (where the
3217 representation of a field element takes up to 24 bytes); for
3218 larger curves, the field element resulting from ECDH is directly
3219 used as premaster secret.
3220 [Douglas Stebila (Sun Microsystems Laboratories)]
3222 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3223 curve secp160r1 to the tests.
3224 [Douglas Stebila (Sun Microsystems Laboratories)]
3226 *) Add the possibility to load symbols globally with DSO.
3227 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3229 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3230 control of the error stack.
3233 *) Add support for STORE in ENGINE.
3236 *) Add the STORE type. The intention is to provide a common interface
3237 to certificate and key stores, be they simple file-based stores, or
3238 HSM-type store, or LDAP stores, or...
3239 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3242 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3243 pass a list of arguments to any function as well as provide a way
3244 for a function to pass data back to the caller.
3247 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3248 works like BUF_strdup() but can be used to duplicate a portion of
3249 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3253 *) Add the function sk_find_ex() which works like sk_find(), but will
3254 return an index to an element even if an exact match couldn't be
3255 found. The index is guaranteed to point at the element where the
3256 searched-for key would be inserted to preserve sorting order.
3259 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3260 takes an extra flags argument for optional functionality. Currently,
3261 the following flags are defined:
3263 OBJ_BSEARCH_VALUE_ON_NOMATCH
3264 This one gets OBJ_bsearch_ex() to return a pointer to the first
3265 element where the comparing function returns a negative or zero
3268 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3269 This one gets OBJ_bsearch_ex() to return a pointer to the first
3270 element where the comparing function returns zero. This is useful
3271 if there are more than one element where the comparing function
3275 *) Make it possible to create self-signed certificates with 'openssl ca'
3276 in such a way that the self-signed certificate becomes part of the
3277 CA database and uses the same mechanisms for serial number generation
3278 as all other certificate signing. The new flag '-selfsign' enables
3279 this functionality. Adapt CA.sh and CA.pl.in.
3282 *) Add functionality to check the public key of a certificate request
3283 against a given private. This is useful to check that a certificate
3284 request can be signed by that key (self-signing).
3287 *) Make it possible to have multiple active certificates with the same
3288 subject in the CA index file. This is done only if the keyword
3289 'unique_subject' is set to 'no' in the main CA section (default
3290 if 'CA_default') of the configuration file. The value is saved
3291 with the database itself in a separate index attribute file,
3292 named like the index file with '.attr' appended to the name.
3295 *) Generate muti valued AVAs using '+' notation in config files for
3299 *) Support for nameConstraints certificate extension.
3302 *) Support for policyConstraints certificate extension.
3305 *) Support for policyMappings certificate extension.
3308 *) Make sure the default DSA_METHOD implementation only uses its
3309 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3310 and change its own handlers to be NULL so as to remove unnecessary
3311 indirection. This lets alternative implementations fallback to the
3312 default implementation more easily.
3315 *) Support for directoryName in GeneralName related extensions
3319 *) Make it possible to link applications using Makefile.shared.
3320 Make that possible even when linking against static libraries!
3323 *) Support for single pass processing for S/MIME signing. This now
3324 means that S/MIME signing can be done from a pipe, in addition
3325 cleartext signing (multipart/signed type) is effectively streaming
3326 and the signed data does not need to be all held in memory.
3328 This is done with a new flag PKCS7_STREAM. When this flag is set
3329 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3330 is done after the data is output (and digests calculated) in
3331 SMIME_write_PKCS7().
3334 *) Add full support for -rpath/-R, both in shared libraries and
3335 applications, at least on the platforms where it's known how
3339 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3340 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3341 will now compute a table of multiples of the generator that
3342 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3343 faster (notably in the case of a single point multiplication,
3344 scalar * generator).
3345 [Nils Larsch, Bodo Moeller]
3347 *) IPv6 support for certificate extensions. The various extensions
3348 which use the IP:a.b.c.d can now take IPv6 addresses using the
3349 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3353 *) Added an ENGINE that implements RSA by performing private key
3354 exponentiations with the GMP library. The conversions to and from
3355 GMP's mpz_t format aren't optimised nor are any montgomery forms
3356 cached, and on x86 it appears OpenSSL's own performance has caught up.
3357 However there are likely to be other architectures where GMP could
3358 provide a boost. This ENGINE is not built in by default, but it can be
3359 specified at Configure time and should be accompanied by the necessary
3360 linker additions, eg;
3361 ./config -DOPENSSL_USE_GMP -lgmp
3364 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3365 testing availability of engines with "-t" - the old behaviour is
3366 produced by increasing the feature's verbosity with "-tt".
3369 *) ECDSA routines: under certain error conditions uninitialized BN objects
3370 could be freed. Solution: make sure initialization is performed early
3371 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3375 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3376 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3377 software implementations. For DSA and DH, parameter generation can
3378 also be overriden by providing the appropriate method callbacks.
3381 *) Change the "progress" mechanism used in key-generation and
3382 primality testing to functions that take a new BN_GENCB pointer in
3383 place of callback/argument pairs. The new API functions have "_ex"
3384 postfixes and the older functions are reimplemented as wrappers for
3385 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3386 declarations of the old functions to help (graceful) attempts to
3387 migrate to the new functions. Also, the new key-generation API
3388 functions operate on a caller-supplied key-structure and return
3389 success/failure rather than returning a key or NULL - this is to
3390 help make "keygen" another member function of RSA_METHOD etc.
3392 Example for using the new callback interface:
3394 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3398 BN_GENCB_set(&my_cb, my_callback, my_arg);
3400 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3401 /* For the meaning of a, b in calls to my_callback(), see the
3402 * documentation of the function that calls the callback.
3403 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3404 * my_callback should return 1 if it wants BN_is_prime_ex()
3405 * to continue, or 0 to stop.
3410 *) Change the ZLIB compression method to be stateful, and make it
3411 available to TLS with the number defined in
3412 draft-ietf-tls-compression-04.txt.
3415 *) Add the ASN.1 structures and functions for CertificatePair, which
3416 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3418 CertificatePair ::= SEQUENCE {
3419 forward [0] Certificate OPTIONAL,
3420 reverse [1] Certificate OPTIONAL,
3421 -- at least one of the pair shall be present -- }
3423 Also implement the PEM functions to read and write certificate
3424 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3426 This needed to be defined, mostly for the sake of the LDAP
3427 attribute crossCertificatePair, but may prove useful elsewhere as
3431 *) Make it possible to inhibit symlinking of shared libraries in
3432 Makefile.shared, for Cygwin's sake.
3435 *) Extend the BIGNUM API by creating a function
3436 void BN_set_negative(BIGNUM *a, int neg);
3437 and a macro that behave like
3438 int BN_is_negative(const BIGNUM *a);
3440 to avoid the need to access 'a->neg' directly in applications.
3443 *) Implement fast modular reduction for pseudo-Mersenne primes
3444 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3445 EC_GROUP_new_curve_GFp() will now automatically use this
3447 [Nils Larsch <nla@trustcenter.de>]
3449 *) Add new lock type (CRYPTO_LOCK_BN).
3452 *) Change the ENGINE framework to automatically load engines
3453 dynamically from specific directories unless they could be
3454 found to already be built in or loaded. Move all the
3455 current engines except for the cryptodev one to a new
3457 The engines in engines/ are built as shared libraries if
3458 the "shared" options was given to ./Configure or ./config.
3459 Otherwise, they are inserted in libcrypto.a.
3460 /usr/local/ssl/engines is the default directory for dynamic
3461 engines, but that can be overriden at configure time through
3462 the usual use of --prefix and/or --openssldir, and at run
3463 time with the environment variable OPENSSL_ENGINES.
3464 [Geoff Thorpe and Richard Levitte]
3466 *) Add Makefile.shared, a helper makefile to build shared
3467 libraries. Addapt Makefile.org.
3470 *) Add version info to Win32 DLLs.
3471 [Peter 'Luna' Runestig" <peter@runestig.com>]
3473 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3474 can be added using this API to created arbitrary PKCS#12
3475 files while avoiding the low level API.
3477 New options to PKCS12_create(), key or cert can be NULL and
3478 will then be omitted from the output file. The encryption
3479 algorithm NIDs can be set to -1 for no encryption, the mac
3480 iteration count can be set to 0 to omit the mac.
3482 Enhance pkcs12 utility by making the -nokeys and -nocerts
3483 options work when creating a PKCS#12 file. New option -nomac
3484 to omit the mac, NONE can be set for an encryption algorithm.
3485 New code is modified to use the enhanced PKCS12_create()
3486 instead of the low level API.
3489 *) Extend ASN1 encoder to support indefinite length constructed
3490 encoding. This can output sequences tags and octet strings in
3491 this form. Modify pk7_asn1.c to support indefinite length
3492 encoding. This is experimental and needs additional code to
3493 be useful, such as an ASN1 bio and some enhanced streaming
3496 Extend template encode functionality so that tagging is passed
3497 down to the template encoder.
3500 *) Let 'openssl req' fail if an argument to '-newkey' is not
3501 recognized instead of using RSA as a default.
3504 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3505 As these are not official, they are not included in "ALL";
3506 the "ECCdraft" ciphersuite group alias can be used to select them.
3507 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3509 *) Add ECDH engine support.
3510 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3512 *) Add ECDH in new directory crypto/ecdh/.
3513 [Douglas Stebila (Sun Microsystems Laboratories)]
3515 *) Let BN_rand_range() abort with an error after 100 iterations
3516 without success (which indicates a broken PRNG).
3519 *) Change BN_mod_sqrt() so that it verifies that the input value
3520 is really the square of the return value. (Previously,
3521 BN_mod_sqrt would show GIGO behaviour.)
3524 *) Add named elliptic curves over binary fields from X9.62, SECG,
3525 and WAP/WTLS; add OIDs that were still missing.
3527 [Sheueling Chang Shantz and Douglas Stebila
3528 (Sun Microsystems Laboratories)]
3530 *) Extend the EC library for elliptic curves over binary fields
3531 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3534 EC_GF2m_simple_method
3538 EC_GROUP_new_curve_GF2m
3539 EC_GROUP_set_curve_GF2m
3540 EC_GROUP_get_curve_GF2m
3541 EC_POINT_set_affine_coordinates_GF2m
3542 EC_POINT_get_affine_coordinates_GF2m
3543 EC_POINT_set_compressed_coordinates_GF2m
3545 Point compression for binary fields is disabled by default for
3546 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3549 As binary polynomials are represented as BIGNUMs, various members
3550 of the EC_GROUP and EC_POINT data structures can be shared
3551 between the implementations for prime fields and binary fields;
3552 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3553 are essentially identical to their ..._GFp counterparts.
3554 (For simplicity, the '..._GFp' prefix has been dropped from
3555 various internal method names.)
3557 An internal 'field_div' method (similar to 'field_mul' and
3558 'field_sqr') has been added; this is used only for binary fields.
3560 [Sheueling Chang Shantz and Douglas Stebila
3561 (Sun Microsystems Laboratories)]
3563 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3564 through methods ('mul', 'precompute_mult').
3566 The generic implementations (now internally called 'ec_wNAF_mul'
3567 and 'ec_wNAF_precomputed_mult') remain the default if these
3568 methods are undefined.
3570 [Sheueling Chang Shantz and Douglas Stebila
3571 (Sun Microsystems Laboratories)]
3573 *) New function EC_GROUP_get_degree, which is defined through
3574 EC_METHOD. For curves over prime fields, this returns the bit
3575 length of the modulus.
3577 [Sheueling Chang Shantz and Douglas Stebila
3578 (Sun Microsystems Laboratories)]
3580 *) New functions EC_GROUP_dup, EC_POINT_dup.
3581 (These simply call ..._new and ..._copy).
3583 [Sheueling Chang Shantz and Douglas Stebila
3584 (Sun Microsystems Laboratories)]
3586 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3587 Polynomials are represented as BIGNUMs (where the sign bit is not
3588 used) in the following functions [macros]:
3591 BN_GF2m_sub [= BN_GF2m_add]
3592 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3593 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3594 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3596 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3597 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3598 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3599 BN_GF2m_cmp [= BN_ucmp]
3601 (Note that only the 'mod' functions are actually for fields GF(2^m).
3602 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3604 For some functions, an the irreducible polynomial defining a
3605 field can be given as an 'unsigned int[]' with strictly
3606 decreasing elements giving the indices of those bits that are set;
3607 i.e., p[] represents the polynomial
3608 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3610 p[0] > p[1] > ... > p[k] = 0.
3611 This applies to the following functions:
3616 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3617 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3619 BN_GF2m_mod_sqrt_arr
3620 BN_GF2m_mod_solve_quad_arr
3624 Conversion can be performed by the following functions:
3629 bntest.c has additional tests for binary polynomial arithmetic.
3631 Two implementations for BN_GF2m_mod_div() are available.
3632 The default algorithm simply uses BN_GF2m_mod_inv() and
3633 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3634 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3635 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3637 [Sheueling Chang Shantz and Douglas Stebila
3638 (Sun Microsystems Laboratories)]
3640 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3641 functionality is disabled at compile-time.
3642 [Douglas Stebila <douglas.stebila@sun.com>]
3644 *) Change default behaviour of 'openssl asn1parse' so that more
3645 information is visible when viewing, e.g., a certificate:
3647 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3648 mode the content of non-printable OCTET STRINGs is output in a
3649 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3650 avoid the appearance of a printable string.
3651 [Nils Larsch <nla@trustcenter.de>]
3653 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3655 EC_GROUP_set_asn1_flag()
3656 EC_GROUP_get_asn1_flag()
3657 EC_GROUP_set_point_conversion_form()
3658 EC_GROUP_get_point_conversion_form()
3659 These control ASN1 encoding details:
3660 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3661 has been set to OPENSSL_EC_NAMED_CURVE.
3662 - Points are encoded in uncompressed form by default; options for
3663 asn1_for are as for point2oct, namely
3664 POINT_CONVERSION_COMPRESSED
3665 POINT_CONVERSION_UNCOMPRESSED
3666 POINT_CONVERSION_HYBRID
3668 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3671 EC_GROUP_get0_seed()
3672 EC_GROUP_get_seed_len()
3673 This is used only for ASN1 purposes (so far).
3674 [Nils Larsch <nla@trustcenter.de>]
3676 *) Add 'field_type' member to EC_METHOD, which holds the NID
3677 of the appropriate field type OID. The new function
3678 EC_METHOD_get_field_type() returns this value.
3679 [Nils Larsch <nla@trustcenter.de>]
3684 EC_POINT_point2hex()
3685 EC_POINT_hex2point()
3686 providing useful interfaces to EC_POINT_point2oct() and
3687 EC_POINT_oct2point().
3688 [Nils Larsch <nla@trustcenter.de>]
3690 *) Change internals of the EC library so that the functions
3691 EC_GROUP_set_generator()
3692 EC_GROUP_get_generator()
3693 EC_GROUP_get_order()
3694 EC_GROUP_get_cofactor()
3695 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3696 to methods, which would lead to unnecessary code duplication when
3697 adding different types of curves.
3698 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3700 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3701 arithmetic, and such that modified wNAFs are generated
3702 (which avoid length expansion in many cases).
3705 *) Add a function EC_GROUP_check_discriminant() (defined via
3706 EC_METHOD) that verifies that the curve discriminant is non-zero.
3708 Add a function EC_GROUP_check() that makes some sanity tests
3709 on a EC_GROUP, its generator and order. This includes
3710 EC_GROUP_check_discriminant().
3711 [Nils Larsch <nla@trustcenter.de>]
3713 *) Add ECDSA in new directory crypto/ecdsa/.
3715 Add applications 'openssl ecparam' and 'openssl ecdsa'
3716 (these are based on 'openssl dsaparam' and 'openssl dsa').
3718 ECDSA support is also included in various other files across the
3719 library. Most notably,
3720 - 'openssl req' now has a '-newkey ecdsa:file' option;
3721 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3722 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3723 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3724 them suitable for ECDSA where domain parameters must be
3725 extracted before the specific public key;
3726 - ECDSA engine support has been added.
3727 [Nils Larsch <nla@trustcenter.de>]
3729 *) Include some named elliptic curves, and add OIDs from X9.62,
3730 SECG, and WAP/WTLS. Each curve can be obtained from the new
3732 EC_GROUP_new_by_curve_name(),
3733 and the list of available named curves can be obtained with
3734 EC_get_builtin_curves().
3735 Also add a 'curve_name' member to EC_GROUP objects, which can be
3737 EC_GROUP_set_curve_name()
3738 EC_GROUP_get_curve_name()
3739 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3741 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3742 was actually never needed) and in BN_mul(). The removal in BN_mul()
3743 required a small change in bn_mul_part_recursive() and the addition
3744 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3745 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3746 bn_sub_words() and bn_add_words() except they take arrays with
3750 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3752 *) Cleanse PEM buffers before freeing them since they may contain
3754 [Benjamin Bennett <ben@psc.edu>]
3756 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3757 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3758 authentication-only ciphersuites.
3761 *) Since AES128 and AES256 share a single mask bit in the logic of
3762 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3763 kludge to work properly if AES128 is available and AES256 isn't.
3766 *) Expand security boundary to match 1.1.1 module.
3769 *) Remove redundant features: hash file source, editing of test vectors
3770 modify fipsld to use external fips_premain.c signature.
3773 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3774 run algorithm test programs.
3777 *) Make algorithm test programs more tolerant of whitespace.
3780 *) Have SSL/TLS server implementation tolerate "mismatched" record
3781 protocol version while receiving ClientHello even if the
3782 ClientHello is fragmented. (The server can't insist on the
3783 particular protocol version it has chosen before the ServerHello
3784 message has informed the client about his choice.)
3787 *) Load error codes if they are not already present instead of using a
3788 static variable. This allows them to be cleanly unloaded and reloaded.
3791 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3793 *) Introduce limits to prevent malicious keys being able to
3794 cause a denial of service. (CVE-2006-2940)
3795 [Steve Henson, Bodo Moeller]
3797 *) Fix ASN.1 parsing of certain invalid structures that can result
3798 in a denial of service. (CVE-2006-2937) [Steve Henson]
3800 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3801 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3803 *) Fix SSL client code which could crash if connecting to a
3804 malicious SSLv2 server. (CVE-2006-4343)
3805 [Tavis Ormandy and Will Drewry, Google Security Team]
3807 *) Change ciphersuite string processing so that an explicit
3808 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3809 will no longer include "AES128-SHA"), and any other similar
3810 ciphersuite (same bitmap) from *other* protocol versions (so that
3811 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3812 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3813 changes from 0.9.8b and 0.9.8d.
3816 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3818 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3819 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3821 *) Change the Unix randomness entropy gathering to use poll() when
3822 possible instead of select(), since the latter has some
3823 undesirable limitations.
3824 [Darryl Miles via Richard Levitte and Bodo Moeller]
3826 *) Disable rogue ciphersuites:
3828 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3829 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3830 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3832 The latter two were purportedly from
3833 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3836 Also deactive the remaining ciphersuites from
3837 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3838 unofficial, and the ID has long expired.
3841 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3842 dual-core machines) and other potential thread-safety issues.
3845 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3847 *) Adapt fipsld and the build system to link against the validated FIPS
3848 module in FIPS mode.
3851 *) Fixes for VC++ 2005 build under Windows.
3854 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3855 from a Windows bash shell such as MSYS. It is autodetected from the
3856 "config" script when run from a VC++ environment. Modify standard VC++
3857 build to use fipscanister.o from the GNU make build.
3860 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3862 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3863 The value now differs depending on if you build for FIPS or not.
3864 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3865 safely run with a non-FIPSed libcrypto, as it may crash because of
3866 the difference induced by this change.
3869 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3871 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3872 (part of SSL_OP_ALL). This option used to disable the
3873 countermeasure against man-in-the-middle protocol-version
3874 rollback in the SSL 2.0 server implementation, which is a bad
3875 idea. (CVE-2005-2969)
3877 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3878 for Information Security, National Institute of Advanced Industrial
3879 Science and Technology [AIST], Japan)]
3881 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3882 mainly for FIPS compliance and not fully integrated at this stage.
3885 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3886 the exponentiation using a fixed-length exponent. (Otherwise,
3887 the information leaked through timing could expose the secret key
3888 after many signatures; cf. Bleichenbacher's attack on DSA with
3892 *) Make a new fixed-window mod_exp implementation the default for
3893 RSA, DSA, and DH private-key operations so that the sequence of
3894 squares and multiplies and the memory access pattern are
3895 independent of the particular secret key. This will mitigate
3896 cache-timing and potential related attacks.
3898 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3899 and this is automatically used by BN_mod_exp_mont() if the new flag
3900 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3901 will use this BN flag for private exponents unless the flag
3902 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3903 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3905 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3907 *) Change the client implementation for SSLv23_method() and
3908 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3909 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3910 (Previously, the SSL 2.0 backwards compatible Client Hello
3911 message format would be used even with SSL_OP_NO_SSLv2.)
3914 *) Add support for smime-type MIME parameter in S/MIME messages which some
3918 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3919 a threadsafe manner. Modify rsa code to use new function and add calls
3920 to dsa and dh code (which had race conditions before).
3923 *) Include the fixed error library code in the C error file definitions
3924 instead of fixing them up at runtime. This keeps the error code
3925 structures constant.
3928 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3930 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3933 *) Fixes for newer kerberos headers. NB: the casts are needed because
3934 the 'length' field is signed on one version and unsigned on another
3935 with no (?) obvious way to tell the difference, without these VC++
3936 complains. Also the "definition" of FAR (blank) is no longer included
3937 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3938 some needed definitions.
3941 *) Undo Cygwin change.
3944 *) Added support for proxy certificates according to RFC 3820.
3945 Because they may be a security thread to unaware applications,
3946 they must be explicitely allowed in run-time. See
3947 docs/HOWTO/proxy_certificates.txt for further information.
3950 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3952 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3953 server and client random values. Previously
3954 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3955 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3957 This change has negligible security impact because:
3959 1. Server and client random values still have 24 bytes of pseudo random
3962 2. Server and client random values are sent in the clear in the initial
3965 3. The master secret is derived using the premaster secret (48 bytes in
3966 size for static RSA ciphersuites) as well as client server and random
3969 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3972 [Stephen Henson, reported by UK NISCC]
3974 *) Use Windows randomness collection on Cygwin.
3977 *) Fix hang in EGD/PRNGD query when communication socket is closed
3978 prematurely by EGD/PRNGD.
3979 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3981 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3984 *) Back-port of selected performance improvements from development
3985 branch, as well as improved support for PowerPC platforms.
3988 *) Add lots of checks for memory allocation failure, error codes to indicate
3989 failure and freeing up memory if a failure occurs.
3990 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3992 *) Add new -passin argument to dgst.
3995 *) Perform some character comparisons of different types in X509_NAME_cmp:
3996 this is needed for some certificates that reencode DNs into UTF8Strings
3997 (in violation of RFC3280) and can't or wont issue name rollover
4001 *) Make an explicit check during certificate validation to see that
4002 the CA setting in each certificate on the chain is correct. As a
4003 side effect always do the following basic checks on extensions,
4004 not just when there's an associated purpose to the check:
4006 - if there is an unhandled critical extension (unless the user
4007 has chosen to ignore this fault)
4008 - if the path length has been exceeded (if one is set at all)
4009 - that certain extensions fit the associated purpose (if one has
4013 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4015 *) Avoid a race condition when CRLs are checked in a multi threaded
4016 environment. This would happen due to the reordering of the revoked
4017 entries during signature checking and serial number lookup. Now the
4018 encoding is cached and the serial number sort performed under a lock.
4019 Add new STACK function sk_is_sorted().
4022 *) Add Delta CRL to the extension code.
4025 *) Various fixes to s3_pkt.c so alerts are sent properly.
4026 [David Holmes <d.holmes@f5.com>]
4028 *) Reduce the chances of duplicate issuer name and serial numbers (in
4029 violation of RFC3280) using the OpenSSL certificate creation utilities.
4030 This is done by creating a random 64 bit value for the initial serial
4031 number when a serial number file is created or when a self signed
4032 certificate is created using 'openssl req -x509'. The initial serial
4033 number file is created using 'openssl x509 -next_serial' in CA.pl
4034 rather than being initialized to 1.
4037 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4039 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4040 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4041 [Joe Orton, Steve Henson]
4043 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4045 [Joe Orton, Steve Henson]
4047 *) Make it possible to have multiple active certificates with the same
4048 subject in the CA index file. This is done only if the keyword
4049 'unique_subject' is set to 'no' in the main CA section (default
4050 if 'CA_default') of the configuration file. The value is saved
4051 with the database itself in a separate index attribute file,
4052 named like the index file with '.attr' appended to the name.
4055 *) X509 verify fixes. Disable broken certificate workarounds when
4056 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4057 keyUsage extension present. Don't accept CRLs with unhandled critical
4058 extensions: since verify currently doesn't process CRL extensions this
4059 rejects a CRL with *any* critical extensions. Add new verify error codes
4063 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4064 A clarification of RFC2560 will require the use of OCTET STRINGs and
4065 some implementations cannot handle the current raw format. Since OpenSSL
4066 copies and compares OCSP nonces as opaque blobs without any attempt at
4067 parsing them this should not create any compatibility issues.
4070 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4071 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4072 this HMAC (and other) operations are several times slower than OpenSSL
4076 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4077 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4079 *) Use the correct content when signing type "other".
4082 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4084 *) Fix various bugs revealed by running the NISCC test suite:
4086 Stop out of bounds reads in the ASN1 code when presented with
4087 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4089 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4091 If verify callback ignores invalid public key errors don't try to check
4092 certificate signature with the NULL public key.
4096 *) New -ignore_err option in ocsp application to stop the server
4097 exiting on the first error in a request.
4100 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4101 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4105 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4106 extra data after the compression methods not only for TLS 1.0
4107 but also for SSL 3.0 (as required by the specification).
4108 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4110 *) Change X509_certificate_type() to mark the key as exported/exportable
4111 when it's 512 *bits* long, not 512 bytes.
4114 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4115 blocks during encryption.
4118 *) Various fixes to base64 BIO and non blocking I/O. On write
4119 flushes were not handled properly if the BIO retried. On read
4120 data was not being buffered properly and had various logic bugs.
4121 This also affects blocking I/O when the data being decoded is a
4125 *) Various S/MIME bugfixes and compatibility changes:
4126 output correct application/pkcs7 MIME type if
4127 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4128 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4129 of files as .eml work). Correctly handle very long lines in MIME
4133 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4135 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4136 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4137 a protocol version number mismatch like a decryption error
4138 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4141 *) Turn on RSA blinding by default in the default implementation
4142 to avoid a timing attack. Applications that don't want it can call
4143 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4144 They would be ill-advised to do so in most cases.
4145 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4147 *) Change RSA blinding code so that it works when the PRNG is not
4148 seeded (in this case, the secret RSA exponent is abused as
4149 an unpredictable seed -- if it is not unpredictable, there
4150 is no point in blinding anyway). Make RSA blinding thread-safe
4151 by remembering the creator's thread ID in rsa->blinding and
4152 having all other threads use local one-time blinding factors
4153 (this requires more computation than sharing rsa->blinding, but
4154 avoids excessive locking; and if an RSA object is not shared
4155 between threads, blinding will still be very fast).
4158 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4159 ENGINE as defaults for all supported algorithms irrespective of
4160 the 'flags' parameter. 'flags' is now honoured, so applications
4161 should make sure they are passing it correctly.
4164 *) Target "mingw" now allows native Windows code to be generated in
4165 the Cygwin environment as well as with the MinGW compiler.
4168 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4170 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4171 via timing by performing a MAC computation even if incorrrect
4172 block cipher padding has been found. This is a countermeasure
4173 against active attacks where the attacker has to distinguish
4174 between bad padding and a MAC verification error. (CVE-2003-0078)
4176 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4177 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4178 Martin Vuagnoux (EPFL, Ilion)]
4180 *) Make the no-err option work as intended. The intention with no-err
4181 is not to have the whole error stack handling routines removed from
4182 libcrypto, it's only intended to remove all the function name and
4183 reason texts, thereby removing some of the footprint that may not
4184 be interesting if those errors aren't displayed anyway.
4186 NOTE: it's still possible for any application or module to have it's
4187 own set of error texts inserted. The routines are there, just not
4188 used by default when no-err is given.
4191 *) Add support for FreeBSD on IA64.
4192 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4194 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4195 Kerberos function mit_des_cbc_cksum(). Before this change,
4196 the value returned by DES_cbc_cksum() was like the one from
4197 mit_des_cbc_cksum(), except the bytes were swapped.
4198 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4200 *) Allow an application to disable the automatic SSL chain building.
4201 Before this a rather primitive chain build was always performed in
4202 ssl3_output_cert_chain(): an application had no way to send the
4203 correct chain if the automatic operation produced an incorrect result.
4205 Now the chain builder is disabled if either:
4207 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4209 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4211 The reasoning behind this is that an application would not want the
4212 auto chain building to take place if extra chain certificates are
4213 present and it might also want a means of sending no additional
4214 certificates (for example the chain has two certificates and the
4218 *) Add the possibility to build without the ENGINE framework.
4219 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4221 *) Under Win32 gmtime() can return NULL: check return value in
4222 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4225 *) DSA routines: under certain error conditions uninitialized BN objects
4226 could be freed. Solution: make sure initialization is performed early
4227 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4228 Nils Larsch <nla@trustcenter.de> via PR#459)
4231 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4232 checked on reconnect on the client side, therefore session resumption
4233 could still fail with a "ssl session id is different" error. This
4234 behaviour is masked when SSL_OP_ALL is used due to
4235 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4236 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4237 followup to PR #377.
4240 *) IA-32 assembler support enhancements: unified ELF targets, support
4241 for SCO/Caldera platforms, fix for Cygwin shared build.
4244 *) Add support for FreeBSD on sparc64. As a consequence, support for
4245 FreeBSD on non-x86 processors is separate from x86 processors on
4246 the config script, much like the NetBSD support.
4247 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4249 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4251 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4254 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4255 code (06) was taken as the first octet of the session ID and the last
4256 octet was ignored consequently. As a result SSLv2 client side session
4257 caching could not have worked due to the session ID mismatch between
4259 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4263 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4264 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4268 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4269 seems that in spite of existing for more than a year, many application
4270 author have done nothing to provide the necessary callbacks, which
4271 means that this particular engine will not work properly anywhere.
4272 This is a very unfortunate situation which forces us, in the name
4273 of usability, to give the hw_ncipher.c a static lock, which is part
4275 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4276 appear in 0.9.8 or later. We EXPECT application authors to have
4277 dealt properly with this when 0.9.8 is released (unless we actually
4278 make such changes in the libcrypto locking code that changes will
4279 have to be made anyway).
4282 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4283 octets have been read, EOF or an error occurs. Without this change
4284 some truncated ASN1 structures will not produce an error.
4287 *) Disable Heimdal support, since it hasn't been fully implemented.
4288 Still give the possibility to force the use of Heimdal, but with
4289 warnings and a request that patches get sent to openssl-dev.
4292 *) Add the VC-CE target, introduce the WINCE sysname, and add
4293 INSTALL.WCE and appropriate conditionals to make it build.
4294 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4296 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4297 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4298 edit numbers of the version.
4299 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4301 *) Introduce safe string copy and catenation functions
4302 (BUF_strlcpy() and BUF_strlcat()).
4303 [Ben Laurie (CHATS) and Richard Levitte]
4305 *) Avoid using fixed-size buffers for one-line DNs.
4306 [Ben Laurie (CHATS)]
4308 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4309 resizing buffers containing secrets, and use where appropriate.
4310 [Ben Laurie (CHATS)]
4312 *) Avoid using fixed size buffers for configuration file location.
4313 [Ben Laurie (CHATS)]
4315 *) Avoid filename truncation for various CA files.
4316 [Ben Laurie (CHATS)]
4318 *) Use sizeof in preference to magic numbers.
4319 [Ben Laurie (CHATS)]
4321 *) Avoid filename truncation in cert requests.
4322 [Ben Laurie (CHATS)]
4324 *) Add assertions to check for (supposedly impossible) buffer
4326 [Ben Laurie (CHATS)]
4328 *) Don't cache truncated DNS entries in the local cache (this could
4329 potentially lead to a spoofing attack).
4330 [Ben Laurie (CHATS)]
4332 *) Fix various buffers to be large enough for hex/decimal
4333 representations in a platform independent manner.
4334 [Ben Laurie (CHATS)]
4336 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4337 resizing buffers containing secrets, and use where appropriate.
4338 [Ben Laurie (CHATS)]
4340 *) Add BIO_indent() to avoid much slightly worrying code to do
4342 [Ben Laurie (CHATS)]
4344 *) Convert sprintf()/BIO_puts() to BIO_printf().
4345 [Ben Laurie (CHATS)]
4347 *) buffer_gets() could terminate with the buffer only half
4349 [Ben Laurie (CHATS)]
4351 *) Add assertions to prevent user-supplied crypto functions from
4352 overflowing internal buffers by having large block sizes, etc.
4353 [Ben Laurie (CHATS)]
4355 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4357 [Ben Laurie (CHATS)]
4359 *) Eliminate unused copy of key in RC4.
4360 [Ben Laurie (CHATS)]
4362 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4363 [Ben Laurie (CHATS)]
4365 *) Fix off-by-one error in EGD path.
4366 [Ben Laurie (CHATS)]
4368 *) If RANDFILE path is too long, ignore instead of truncating.
4369 [Ben Laurie (CHATS)]
4371 *) Eliminate unused and incorrectly sized X.509 structure
4373 [Ben Laurie (CHATS)]
4375 *) Eliminate unused and dangerous function knumber().
4376 [Ben Laurie (CHATS)]
4378 *) Eliminate unused and dangerous structure, KSSL_ERR.
4379 [Ben Laurie (CHATS)]
4381 *) Protect against overlong session ID context length in an encoded
4382 session object. Since these are local, this does not appear to be
4384 [Ben Laurie (CHATS)]
4386 *) Change from security patch (see 0.9.6e below) that did not affect
4387 the 0.9.6 release series:
4389 Remote buffer overflow in SSL3 protocol - an attacker could
4390 supply an oversized master key in Kerberos-enabled versions.
4392 [Ben Laurie (CHATS)]
4394 *) Change the SSL kerb5 codes to match RFC 2712.
4397 *) Make -nameopt work fully for req and add -reqopt switch.
4398 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4400 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4401 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4403 *) Make sure tests can be performed even if the corresponding algorithms
4404 have been removed entirely. This was also the last step to make
4405 OpenSSL compilable with DJGPP under all reasonable conditions.
4406 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4408 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4409 to allow version independent disabling of normally unselected ciphers,
4410 which may be activated as a side-effect of selecting a single cipher.
4412 (E.g., cipher list string "RSA" enables ciphersuites that are left
4413 out of "ALL" because they do not provide symmetric encryption.
4414 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4415 [Lutz Jaenicke, Bodo Moeller]
4417 *) Add appropriate support for separate platform-dependent build
4418 directories. The recommended way to make a platform-dependent
4419 build directory is the following (tested on Linux), maybe with
4422 # Place yourself outside of the OpenSSL source tree. In
4423 # this example, the environment variable OPENSSL_SOURCE
4424 # is assumed to contain the absolute OpenSSL source directory.
4425 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4426 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4427 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4428 mkdir -p `dirname $F`
4429 ln -s $OPENSSL_SOURCE/$F $F
4432 To be absolutely sure not to disturb the source tree, a "make clean"
4433 is a good thing. If it isn't successfull, don't worry about it,
4434 it probably means the source directory is very clean.
4437 *) Make sure any ENGINE control commands make local copies of string
4438 pointers passed to them whenever necessary. Otherwise it is possible
4439 the caller may have overwritten (or deallocated) the original string
4440 data when a later ENGINE operation tries to use the stored values.
4441 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4443 *) Improve diagnostics in file reading and command-line digests.
4444 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4446 *) Add AES modes CFB and OFB to the object database. Correct an
4447 error in AES-CFB decryption.
4450 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4451 allows existing EVP_CIPHER_CTX structures to be reused after
4452 calling EVP_*Final(). This behaviour is used by encryption
4453 BIOs and some applications. This has the side effect that
4454 applications must explicitly clean up cipher contexts with
4455 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4458 *) Check the values of dna and dnb in bn_mul_recursive before calling
4459 bn_mul_comba (a non zero value means the a or b arrays do not contain
4460 n2 elements) and fallback to bn_mul_normal if either is not zero.
4463 *) Fix escaping of non-ASCII characters when using the -subj option
4464 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4467 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4468 form for "surname", serialNumber has no short form.
4469 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4470 therefore remove "mail" short name for "internet 7".
4471 The OID for unique identifiers in X509 certificates is
4472 x500UniqueIdentifier, not uniqueIdentifier.
4473 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4476 *) Add an "init" command to the ENGINE config module and auto initialize
4477 ENGINEs. Without any "init" command the ENGINE will be initialized
4478 after all ctrl commands have been executed on it. If init=1 the
4479 ENGINE is initailized at that point (ctrls before that point are run
4480 on the uninitialized ENGINE and after on the initialized one). If
4481 init=0 then the ENGINE will not be iniatialized at all.
4484 *) Fix the 'app_verify_callback' interface so that the user-defined
4485 argument is actually passed to the callback: In the
4486 SSL_CTX_set_cert_verify_callback() prototype, the callback
4487 declaration has been changed from
4490 int (*cb)(X509_STORE_CTX *,void *);
4491 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4492 i=s->ctx->app_verify_callback(&ctx)
4493 has been changed into
4494 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4496 To update applications using SSL_CTX_set_cert_verify_callback(),
4497 a dummy argument can be added to their callback functions.
4498 [D. K. Smetters <smetters@parc.xerox.com>]
4500 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4501 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4503 *) Add and OPENSSL_LOAD_CONF define which will cause
4504 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4505 This allows older applications to transparently support certain
4506 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4507 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4508 load the config file and OPENSSL_add_all_algorithms_conf() which will
4509 always load it have also been added.
4512 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4513 Adjust NIDs and EVP layer.
4514 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4516 *) Config modules support in openssl utility.
4518 Most commands now load modules from the config file,
4519 though in a few (such as version) this isn't done
4520 because it couldn't be used for anything.
4522 In the case of ca and req the config file used is
4523 the same as the utility itself: that is the -config
4524 command line option can be used to specify an
4528 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4529 use "openssl_conf" if filename is NULL use default openssl config file.
4532 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4533 config section name. Add a new flag to tolerate a missing config file
4534 and move code to CONF_modules_load_file().
4537 *) Support for crypto accelerator cards from Accelerated Encryption
4538 Processing, www.aep.ie. (Use engine 'aep')
4539 The support was copied from 0.9.6c [engine] and adapted/corrected
4540 to work with the new engine framework.
4541 [AEP Inc. and Richard Levitte]
4543 *) Support for SureWare crypto accelerator cards from Baltimore
4544 Technologies. (Use engine 'sureware')
4545 The support was copied from 0.9.6c [engine] and adapted
4546 to work with the new engine framework.
4549 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4550 make the newer ENGINE framework commands for the CHIL engine work.
4551 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4553 *) Make it possible to produce shared libraries on ReliantUNIX.
4554 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4556 *) Add the configuration target debug-linux-ppro.
4557 Make 'openssl rsa' use the general key loading routines
4558 implemented in apps.c, and make those routines able to
4559 handle the key format FORMAT_NETSCAPE and the variant
4561 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4563 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4564 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4566 *) Add -keyform to rsautl, and document -engine.
4567 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4569 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4570 BIO_R_NO_SUCH_FILE error code rather than the generic
4571 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4574 *) Add new functions
4576 ERR_peek_last_error_line
4577 ERR_peek_last_error_line_data.
4578 These are similar to
4581 ERR_peek_error_line_data,
4582 but report on the latest error recorded rather than the first one
4583 still in the error queue.
4584 [Ben Laurie, Bodo Moeller]
4586 *) default_algorithms option in ENGINE config module. This allows things
4588 default_algorithms = ALL
4589 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4592 *) Prelminary ENGINE config module.
4595 *) New experimental application configuration code.
4598 *) Change the AES code to follow the same name structure as all other
4599 symmetric ciphers, and behave the same way. Move everything to
4600 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4601 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4603 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4604 [Ben Laurie and Theo de Raadt]
4606 *) Add option to output public keys in req command.
4607 [Massimiliano Pala madwolf@openca.org]
4609 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4610 (up to about 10% better than before for P-192 and P-224).
4613 *) New functions/macros
4615 SSL_CTX_set_msg_callback(ctx, cb)
4616 SSL_CTX_set_msg_callback_arg(ctx, arg)
4617 SSL_set_msg_callback(ssl, cb)
4618 SSL_set_msg_callback_arg(ssl, arg)
4620 to request calling a callback function
4622 void cb(int write_p, int version, int content_type,
4623 const void *buf, size_t len, SSL *ssl, void *arg)
4625 whenever a protocol message has been completely received
4626 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4627 protocol version according to which the SSL library interprets
4628 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4629 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4630 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4631 specification (change_cipher_spec(20), alert(21), handshake(22)).
4632 'buf' and 'len' point to the actual message, 'ssl' to the
4633 SSL object, and 'arg' is the application-defined value set by
4634 SSL[_CTX]_set_msg_callback_arg().
4636 'openssl s_client' and 'openssl s_server' have new '-msg' options
4637 to enable a callback that displays all protocol messages.
4640 *) Change the shared library support so shared libraries are built as
4641 soon as the corresponding static library is finished, and thereby get
4642 openssl and the test programs linked against the shared library.
4643 This still only happens when the keyword "shard" has been given to
4644 the configuration scripts.
4646 NOTE: shared library support is still an experimental thing, and
4647 backward binary compatibility is still not guaranteed.
4648 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4650 *) Add support for Subject Information Access extension.
4651 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4653 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4654 additional bytes when new memory had to be allocated, not just
4655 when reusing an existing buffer.
4658 *) New command line and configuration option 'utf8' for the req command.
4659 This allows field values to be specified as UTF8 strings.
4662 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4663 runs for the former and machine-readable output for the latter.
4666 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4667 of the e-mail address in the DN (i.e., it will go into a certificate
4668 extension only). The new configuration file option 'email_in_dn = no'
4669 has the same effect.
4670 [Massimiliano Pala madwolf@openca.org]
4672 *) Change all functions with names starting with des_ to be starting
4673 with DES_ instead. Add wrappers that are compatible with libdes,
4674 but are named _ossl_old_des_*. Finally, add macros that map the
4675 des_* symbols to the corresponding _ossl_old_des_* if libdes
4676 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4677 desired, the des_* symbols will be mapped to DES_*, with one
4680 Since we provide two compatibility mappings, the user needs to
4681 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4682 compatibility is desired. The default (i.e., when that macro
4683 isn't defined) is OpenSSL 0.9.6c compatibility.
4685 There are also macros that enable and disable the support of old
4686 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4687 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4688 are defined, the default will apply: to support the old des routines.
4690 In either case, one must include openssl/des.h to get the correct
4691 definitions. Do not try to just include openssl/des_old.h, that
4694 NOTE: This is a major break of an old API into a new one. Software
4695 authors are encouraged to switch to the DES_ style functions. Some
4696 time in the future, des_old.h and the libdes compatibility functions
4697 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4698 default), and then completely removed.
4701 *) Test for certificates which contain unsupported critical extensions.
4702 If such a certificate is found during a verify operation it is
4703 rejected by default: this behaviour can be overridden by either
4704 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4705 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4706 X509_supported_extension() has also been added which returns 1 if a
4707 particular extension is supported.
4710 *) Modify the behaviour of EVP cipher functions in similar way to digests
4711 to retain compatibility with existing code.
4714 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4715 compatibility with existing code. In particular the 'ctx' parameter does
4716 not have to be to be initialized before the call to EVP_DigestInit() and
4717 it is tidied up after a call to EVP_DigestFinal(). New function
4718 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4719 EVP_MD_CTX_copy() changed to not require the destination to be
4720 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4721 requires the destination to be valid.
4723 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4724 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4727 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4728 so that complete 'Handshake' protocol structures are kept in memory
4729 instead of overwriting 'msg_type' and 'length' with 'body' data.
4732 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4733 [Massimo Santin via Richard Levitte]
4735 *) Major restructuring to the underlying ENGINE code. This includes
4736 reduction of linker bloat, separation of pure "ENGINE" manipulation
4737 (initialisation, etc) from functionality dealing with implementations
4738 of specific crypto iterfaces. This change also introduces integrated
4739 support for symmetric ciphers and digest implementations - so ENGINEs
4740 can now accelerate these by providing EVP_CIPHER and EVP_MD
4741 implementations of their own. This is detailed in crypto/engine/README
4742 as it couldn't be adequately described here. However, there are a few
4743 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4744 were changed in the original introduction of ENGINE code have now
4745 reverted back - the hooking from this code to ENGINE is now a good
4746 deal more passive and at run-time, operations deal directly with
4747 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4748 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4749 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4750 they were not being used by the framework as there is no concept of a
4751 BIGNUM_METHOD and they could not be generalised to the new
4752 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4753 ENGINE_cpy() has been removed as it cannot be consistently defined in
4757 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4760 *) Change mkdef.pl to sort symbols that get the same entry number,
4761 and make sure the automatically generated functions ERR_load_*
4762 become part of libeay.num as well.
4765 *) New function SSL_renegotiate_pending(). This returns true once
4766 renegotiation has been requested (either SSL_renegotiate() call
4767 or HelloRequest/ClientHello receveived from the peer) and becomes
4768 false once a handshake has been completed.
4769 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4770 sends a HelloRequest, but does not ensure that a handshake takes
4771 place. SSL_renegotiate_pending() is useful for checking if the
4772 client has followed the request.)
4775 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4776 By default, clients may request session resumption even during
4777 renegotiation (if session ID contexts permit); with this option,
4778 session resumption is possible only in the first handshake.
4780 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4781 more bits available for options that should not be part of
4782 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4785 *) Add some demos for certificate and certificate request creation.
4788 *) Make maximum certificate chain size accepted from the peer application
4789 settable (SSL*_get/set_max_cert_list()), as proposed by
4790 "Douglas E. Engert" <deengert@anl.gov>.
4793 *) Add support for shared libraries for Unixware-7
4794 (Boyd Lynn Gerber <gerberb@zenez.com>).
4797 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4798 be done prior to destruction. Use this to unload error strings from
4799 ENGINEs that load their own error strings. NB: This adds two new API
4800 functions to "get" and "set" this destroy handler in an ENGINE.
4803 *) Alter all existing ENGINE implementations (except "openssl" and
4804 "openbsd") to dynamically instantiate their own error strings. This
4805 makes them more flexible to be built both as statically-linked ENGINEs
4806 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4807 Also, add stub code to each that makes building them as self-contained
4808 shared-libraries easier (see README.ENGINE).
4811 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4812 implementations into applications that are completely implemented in
4813 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4814 commands that can be used to configure what shared-library to load and
4815 to control aspects of the way it is handled. Also, made an update to
4816 the README.ENGINE file that brings its information up-to-date and
4817 provides some information and instructions on the "dynamic" ENGINE
4818 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4821 *) Make it possible to unload ranges of ERR strings with a new
4822 "ERR_unload_strings" function.
4825 *) Add a copy() function to EVP_MD.
4828 *) Make EVP_MD routines take a context pointer instead of just the
4829 md_data void pointer.
4832 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4833 that the digest can only process a single chunk of data
4834 (typically because it is provided by a piece of
4835 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4836 is only going to provide a single chunk of data, and hence the
4837 framework needn't accumulate the data for oneshot drivers.
4840 *) As with "ERR", make it possible to replace the underlying "ex_data"
4841 functions. This change also alters the storage and management of global
4842 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4843 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4844 index counters. The API functions that use this state have been changed
4845 to take a "class_index" rather than pointers to the class's local STACK
4846 and counter, and there is now an API function to dynamically create new
4847 classes. This centralisation allows us to (a) plug a lot of the
4848 thread-safety problems that existed, and (b) makes it possible to clean
4849 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4850 such data would previously have always leaked in application code and
4851 workarounds were in place to make the memory debugging turn a blind eye
4852 to it. Application code that doesn't use this new function will still
4853 leak as before, but their memory debugging output will announce it now
4854 rather than letting it slide.
4856 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4857 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4858 has a return value to indicate success or failure.
4861 *) Make it possible to replace the underlying "ERR" functions such that the
4862 global state (2 LHASH tables and 2 locks) is only used by the "default"
4863 implementation. This change also adds two functions to "get" and "set"
4864 the implementation prior to it being automatically set the first time
4865 any other ERR function takes place. Ie. an application can call "get",
4866 pass the return value to a module it has just loaded, and that module
4867 can call its own "set" function using that value. This means the
4868 module's "ERR" operations will use (and modify) the error state in the
4869 application and not in its own statically linked copy of OpenSSL code.
4872 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4873 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4874 the operation, and provides a more encapsulated way for external code
4875 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4876 to use these functions rather than manually incrementing the counts.
4878 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4881 *) Add EVP test program.
4884 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4887 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4888 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4889 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4890 These allow a CRL to be built without having to access X509_CRL fields
4891 directly. Modify 'ca' application to use new functions.
4894 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4895 bug workarounds. Rollback attack detection is a security feature.
4896 The problem will only arise on OpenSSL servers when TLSv1 is not
4897 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4898 Software authors not wanting to support TLSv1 will have special reasons
4899 for their choice and can explicitly enable this option.
4900 [Bodo Moeller, Lutz Jaenicke]
4902 *) Rationalise EVP so it can be extended: don't include a union of
4903 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4904 (similar to those existing for EVP_CIPHER_CTX).
4909 EVP_MD_CTX_init(&md); /* new function call */
4910 EVP_DigestInit(&md, EVP_sha1());
4911 EVP_DigestUpdate(&md, in, len);
4912 EVP_DigestFinal(&md, out, NULL);
4913 EVP_MD_CTX_cleanup(&md); /* new function call */
4917 *) Make DES key schedule conform to the usual scheme, as well as
4918 correcting its structure. This means that calls to DES functions
4919 now have to pass a pointer to a des_key_schedule instead of a
4920 plain des_key_schedule (which was actually always a pointer
4923 des_key_schedule ks;
4925 des_set_key_checked(..., &ks);
4926 des_ncbc_encrypt(..., &ks, ...);
4928 (Note that a later change renames 'des_...' into 'DES_...'.)
4931 *) Initial reduction of linker bloat: the use of some functions, such as
4932 PEM causes large amounts of unused functions to be linked in due to
4933 poor organisation. For example pem_all.c contains every PEM function
4934 which has a knock on effect of linking in large amounts of (unused)
4935 ASN1 code. Grouping together similar functions and splitting unrelated
4936 functions prevents this.
4939 *) Cleanup of EVP macros.
4942 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4943 correct _ecb suffix.
4946 *) Add initial OCSP responder support to ocsp application. The
4947 revocation information is handled using the text based index
4948 use by the ca application. The responder can either handle
4949 requests generated internally, supplied in files (for example
4950 via a CGI script) or using an internal minimal server.
4953 *) Add configuration choices to get zlib compression for TLS.
4956 *) Changes to Kerberos SSL for RFC 2712 compliance:
4957 1. Implemented real KerberosWrapper, instead of just using
4958 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4959 2. Implemented optional authenticator field of KerberosWrapper.
4961 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4962 and authenticator structs; see crypto/krb5/.
4964 Generalized Kerberos calls to support multiple Kerberos libraries.
4965 [Vern Staats <staatsvr@asc.hpc.mil>,
4966 Jeffrey Altman <jaltman@columbia.edu>
4967 via Richard Levitte]
4969 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4970 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4971 values for each of the key sizes rather than having just
4972 parameters (and 'speed' generating keys each time).
4975 *) Speed up EVP routines.
4978 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4979 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4980 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4981 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4983 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4984 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4985 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4988 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4990 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4993 *) Added the OS2-EMX target.
4994 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4996 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4997 to support NCONF routines in extension code. New function CONF_set_nconf()
4998 to allow functions which take an NCONF to also handle the old LHASH
4999 structure: this means that the old CONF compatible routines can be
5000 retained (in particular wrt extensions) without having to duplicate the
5001 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5004 *) Enhance the general user interface with mechanisms for inner control
5005 and with possibilities to have yes/no kind of prompts.
5008 *) Change all calls to low level digest routines in the library and
5009 applications to use EVP. Add missing calls to HMAC_cleanup() and
5010 don't assume HMAC_CTX can be copied using memcpy().
5011 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5013 *) Add the possibility to control engines through control names but with
5014 arbitrary arguments instead of just a string.
5015 Change the key loaders to take a UI_METHOD instead of a callback
5016 function pointer. NOTE: this breaks binary compatibility with earlier
5017 versions of OpenSSL [engine].
5018 Adapt the nCipher code for these new conditions and add a card insertion
5022 *) Enhance the general user interface with mechanisms to better support
5023 dialog box interfaces, application-defined prompts, the possibility
5024 to use defaults (for example default passwords from somewhere else)
5025 and interrupts/cancellations.
5028 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5029 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5032 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5033 tidy up some unnecessarily weird code in 'sk_new()').
5034 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5036 *) Change the key loading routines for ENGINEs to use the same kind
5037 callback (pem_password_cb) as all other routines that need this
5041 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5042 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5043 than this minimum value is recommended.
5046 *) New random seeder for OpenVMS, using the system process statistics
5047 that are easily reachable.
5050 *) Windows apparently can't transparently handle global
5051 variables defined in DLLs. Initialisations such as:
5053 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5055 wont compile. This is used by the any applications that need to
5056 declare their own ASN1 modules. This was fixed by adding the option
5057 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5058 needed for static libraries under Win32.
5061 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5062 setting of purpose and trust fields. New X509_STORE trust and
5063 purpose functions and tidy up setting in other SSL functions.
5066 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5067 structure. These are inherited by X509_STORE_CTX when it is
5068 initialised. This allows various defaults to be set in the
5069 X509_STORE structure (such as flags for CRL checking and custom
5070 purpose or trust settings) for functions which only use X509_STORE_CTX
5071 internally such as S/MIME.
5073 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5074 trust settings if they are not set in X509_STORE. This allows X509_STORE
5075 purposes and trust (in S/MIME for example) to override any set by default.
5077 Add command line options for CRL checking to smime, s_client and s_server
5081 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5082 are set then the CRL is looked up in the X509_STORE structure and
5083 its validity and signature checked, then if the certificate is found
5084 in the CRL the verify fails with a revoked error.
5086 Various new CRL related callbacks added to X509_STORE_CTX structure.
5088 Command line options added to 'verify' application to support this.
5090 This needs some additional work, such as being able to handle multiple
5091 CRLs with different times, extension based lookup (rather than just
5092 by subject name) and ultimately more complete V2 CRL extension
5096 *) Add a general user interface API (crypto/ui/). This is designed
5097 to replace things like des_read_password and friends (backward
5098 compatibility functions using this new API are provided).
5099 The purpose is to remove prompting functions from the DES code
5100 section as well as provide for prompting through dialog boxes in
5101 a window system and the like.
5104 *) Add "ex_data" support to ENGINE so implementations can add state at a
5105 per-structure level rather than having to store it globally.
5108 *) Make it possible for ENGINE structures to be copied when retrieved by
5109 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5110 This causes the "original" ENGINE structure to act like a template,
5111 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5112 operational state can be localised to each ENGINE structure, despite the
5113 fact they all share the same "methods". New ENGINE structures returned in
5114 this case have no functional references and the return value is the single
5115 structural reference. This matches the single structural reference returned
5116 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5120 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5121 needs to match any other type at all we need to manually clear the
5125 *) Changes to the "openssl engine" utility to include;
5126 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5127 about an ENGINE's available control commands.
5128 - executing control commands from command line arguments using the
5129 '-pre' and '-post' switches. '-post' is only used if '-t' is
5130 specified and the ENGINE is successfully initialised. The syntax for
5131 the individual commands are colon-separated, for example;
5132 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5135 *) New dynamic control command support for ENGINEs. ENGINEs can now
5136 declare their own commands (numbers), names (strings), descriptions,
5137 and input types for run-time discovery by calling applications. A
5138 subset of these commands are implicitly classed as "executable"
5139 depending on their input type, and only these can be invoked through
5140 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5141 can be based on user input, config files, etc). The distinction is
5142 that "executable" commands cannot return anything other than a boolean
5143 result and can only support numeric or string input, whereas some
5144 discoverable commands may only be for direct use through
5145 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5146 pointers, or other custom uses. The "executable" commands are to
5147 support parameterisations of ENGINE behaviour that can be
5148 unambiguously defined by ENGINEs and used consistently across any
5149 OpenSSL-based application. Commands have been added to all the
5150 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5151 control over shared-library paths without source code alterations.
5154 *) Changed all ENGINE implementations to dynamically allocate their
5155 ENGINEs rather than declaring them statically. Apart from this being
5156 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5157 this also allows the implementations to compile without using the
5158 internal engine_int.h header.
5161 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5162 'const' value. Any code that should be able to modify a RAND_METHOD
5163 should already have non-const pointers to it (ie. they should only
5164 modify their own ones).
5167 *) Made a variety of little tweaks to the ENGINE code.
5168 - "atalla" and "ubsec" string definitions were moved from header files
5169 to C code. "nuron" string definitions were placed in variables
5170 rather than hard-coded - allowing parameterisation of these values
5171 later on via ctrl() commands.
5172 - Removed unused "#if 0"'d code.
5173 - Fixed engine list iteration code so it uses ENGINE_free() to release
5174 structural references.
5175 - Constified the RAND_METHOD element of ENGINE structures.
5176 - Constified various get/set functions as appropriate and added
5177 missing functions (including a catch-all ENGINE_cpy that duplicates
5178 all ENGINE values onto a new ENGINE except reference counts/state).
5179 - Removed NULL parameter checks in get/set functions. Setting a method
5180 or function to NULL is a way of cancelling out a previously set
5181 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5182 and doesn't justify the extra error symbols and code.
5183 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5184 flags from engine_int.h to engine.h.
5185 - Changed prototypes for ENGINE handler functions (init(), finish(),
5186 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5189 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5190 to the algorithm using long division. The binary algorithm can be
5191 used only if the modulus is odd. On 32-bit systems, it is faster
5192 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5193 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5194 up to 450 bits. In 64-bit environments, the binary algorithm
5195 appears to be advantageous for much longer moduli; here we use it
5196 for moduli up to 2048 bits.
5199 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5200 could not support the combine flag in choice fields.
5203 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5204 extensions from a certificate request to the certificate.
5207 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5208 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5209 file: this allows the display of the certificate about to be
5210 signed to be customised, to allow certain fields to be included
5211 or excluded and extension details. The old system didn't display
5212 multicharacter strings properly, omitted fields not in the policy
5213 and couldn't display additional details such as extensions.
5216 *) Function EC_POINTs_mul for multiple scalar multiplication
5217 of an arbitrary number of elliptic curve points
5218 \sum scalars[i]*points[i],
5219 optionally including the generator defined for the EC_GROUP:
5220 scalar*generator + \sum scalars[i]*points[i].
5222 EC_POINT_mul is a simple wrapper function for the typical case
5223 that the point list has just one item (besides the optional
5227 *) First EC_METHODs for curves over GF(p):
5229 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5230 operations and provides various method functions that can also
5231 operate with faster implementations of modular arithmetic.
5233 EC_GFp_mont_method() reuses most functions that are part of
5234 EC_GFp_simple_method, but uses Montgomery arithmetic.
5236 [Bodo Moeller; point addition and point doubling
5237 implementation directly derived from source code provided by
5238 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5240 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5241 crypto/ec/ec_lib.c):
5243 Curves are EC_GROUP objects (with an optional group generator)
5244 based on EC_METHODs that are built into the library.
5246 Points are EC_POINT objects based on EC_GROUP objects.
5248 Most of the framework would be able to handle curves over arbitrary
5249 finite fields, but as there are no obvious types for fields other
5250 than GF(p), some functions are limited to that for now.
5253 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5254 that the file contains a complete HTTP response.
5257 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5258 change the def and num file printf format specifier from "%-40sXXX"
5259 to "%-39s XXX". The latter will always guarantee a space after the
5260 field while the former will cause them to run together if the field
5261 is 40 of more characters long.
5264 *) Constify the cipher and digest 'method' functions and structures
5265 and modify related functions to take constant EVP_MD and EVP_CIPHER
5269 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5270 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5273 *) Modify EVP_Digest*() routines so they now return values. Although the
5274 internal software routines can never fail additional hardware versions
5278 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5280 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5281 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5284 ERR_R_NESTED_ASN1_ERROR
5286 ERR_R_MISSING_ASN1_EOS
5287 were 4 .. 9, conflicting with
5288 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5290 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5291 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5293 Add new error code 'ERR_R_INTERNAL_ERROR'.
5296 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5300 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5301 sets the subject name for a new request or supersedes the
5302 subject name in a given request. Formats that can be parsed are
5303 'CN=Some Name, OU=myOU, C=IT'
5305 'CN=Some Name/OU=myOU/C=IT'.
5307 Add options '-batch' and '-verbose' to 'openssl req'.
5308 [Massimiliano Pala <madwolf@hackmasters.net>]
5310 *) Introduce the possibility to access global variables through
5311 functions on platform were that's the best way to handle exporting
5312 global variables in shared libraries. To enable this functionality,
5313 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5314 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5315 is normally done by Configure or something similar).
5317 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5318 in the source file (foo.c) like this:
5320 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5321 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5323 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5324 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5326 OPENSSL_DECLARE_GLOBAL(int,foo);
5327 #define foo OPENSSL_GLOBAL_REF(foo)
5328 OPENSSL_DECLARE_GLOBAL(double,bar);
5329 #define bar OPENSSL_GLOBAL_REF(bar)
5331 The #defines are very important, and therefore so is including the
5332 header file everywhere where the defined globals are used.
5334 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5335 of ASN.1 items, but that structure is a bit different.
5337 The largest change is in util/mkdef.pl which has been enhanced with
5338 better and easier to understand logic to choose which symbols should
5339 go into the Windows .def files as well as a number of fixes and code
5340 cleanup (among others, algorithm keywords are now sorted
5341 lexicographically to avoid constant rewrites).
5344 *) In BN_div() keep a copy of the sign of 'num' before writing the
5345 result to 'rm' because if rm==num the value will be overwritten
5346 and produce the wrong result if 'num' is negative: this caused
5347 problems with BN_mod() and BN_nnmod().
5350 *) Function OCSP_request_verify(). This checks the signature on an
5351 OCSP request and verifies the signer certificate. The signer
5352 certificate is just checked for a generic purpose and OCSP request
5356 *) Add OCSP_check_validity() function to check the validity of OCSP
5357 responses. OCSP responses are prepared in real time and may only
5358 be a few seconds old. Simply checking that the current time lies
5359 between thisUpdate and nextUpdate max reject otherwise valid responses
5360 caused by either OCSP responder or client clock inaccuracy. Instead
5361 we allow thisUpdate and nextUpdate to fall within a certain period of
5362 the current time. The age of the response can also optionally be
5363 checked. Two new options -validity_period and -status_age added to
5367 *) If signature or public key algorithm is unrecognized print out its
5368 OID rather that just UNKNOWN.
5371 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5372 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5373 ID to be generated from the issuer certificate alone which can then be
5374 passed to OCSP_id_issuer_cmp().
5377 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5378 ASN1 modules to export functions returning ASN1_ITEM pointers
5379 instead of the ASN1_ITEM structures themselves. This adds several
5380 new macros which allow the underlying ASN1 function/structure to
5381 be accessed transparently. As a result code should not use ASN1_ITEM
5382 references directly (such as &X509_it) but instead use the relevant
5383 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5384 use of the new ASN1 code on platforms where exporting structures
5385 is problematical (for example in shared libraries) but exporting
5386 functions returning pointers to structures is not.
5389 *) Add support for overriding the generation of SSL/TLS session IDs.
5390 These callbacks can be registered either in an SSL_CTX or per SSL.
5391 The purpose of this is to allow applications to control, if they wish,
5392 the arbitrary values chosen for use as session IDs, particularly as it
5393 can be useful for session caching in multiple-server environments. A
5394 command-line switch for testing this (and any client code that wishes
5395 to use such a feature) has been added to "s_server".
5396 [Geoff Thorpe, Lutz Jaenicke]
5398 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5399 of the form '#if defined(...) || defined(...) || ...' and
5400 '#if !defined(...) && !defined(...) && ...'. This also avoids
5401 the growing number of special cases it was previously handling.
5404 *) Make all configuration macros available for application by making
5405 sure they are available in opensslconf.h, by giving them names starting
5406 with "OPENSSL_" to avoid conflicts with other packages and by making
5407 sure e_os2.h will cover all platform-specific cases together with
5409 Additionally, it is now possible to define configuration/platform-
5410 specific names (called "system identities"). In the C code, these
5411 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5412 macro with the name beginning with "OPENSSL_SYS_", which is determined
5413 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5417 *) New option -set_serial to 'req' and 'x509' this allows the serial
5418 number to use to be specified on the command line. Previously self
5419 signed certificates were hard coded with serial number 0 and the
5420 CA options of 'x509' had to use a serial number in a file which was
5424 *) New options to 'ca' utility to support V2 CRL entry extensions.
5425 Currently CRL reason, invalidity date and hold instruction are
5426 supported. Add new CRL extensions to V3 code and some new objects.
5429 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5430 disable standard block padding (aka PKCS#5 padding) in the EVP
5431 API, which was previously mandatory. This means that the data is
5432 not padded in any way and so the total length much be a multiple
5433 of the block size, otherwise an error occurs.
5436 *) Initial (incomplete) OCSP SSL support.
5439 *) New function OCSP_parse_url(). This splits up a URL into its host,
5440 port and path components: primarily to parse OCSP URLs. New -url
5441 option to ocsp utility.
5444 *) New nonce behavior. The return value of OCSP_check_nonce() now
5445 reflects the various checks performed. Applications can decide
5446 whether to tolerate certain situations such as an absent nonce
5447 in a response when one was present in a request: the ocsp application
5448 just prints out a warning. New function OCSP_add1_basic_nonce()
5449 this is to allow responders to include a nonce in a response even if
5450 the request is nonce-less.
5453 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5454 skipped when using openssl x509 multiple times on a single input file,
5455 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5458 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5459 set string type: to handle setting ASN1_TIME structures. Fix ca
5460 utility to correctly initialize revocation date of CRLs.
5463 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5464 the clients preferred ciphersuites and rather use its own preferences.
5465 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5466 Internet Explorer by ensuring unchanged hash method during stepup.
5467 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5470 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5471 to aes and add a new 'exist' option to print out symbols that don't
5475 *) Additional options to ocsp utility to allow flags to be set and
5476 additional certificates supplied.
5479 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5480 OCSP client a number of certificate to only verify the response
5484 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5485 handle the new API. Currently only ECB, CBC modes supported. Add new
5488 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5489 Encryption Standard (AES) Ciphersuites for Transport Layer
5490 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5491 not enabled by default and were not part of the "ALL" ciphersuite
5492 alias because they were not yet official; they could be
5493 explicitly requested by specifying the "AESdraft" ciphersuite
5494 group alias. In the final release of OpenSSL 0.9.7, the group
5495 alias is called "AES" and is part of "ALL".)
5496 [Ben Laurie, Steve Henson, Bodo Moeller]
5498 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5499 request to response.
5502 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5503 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5504 extract information from a certificate request. OCSP_response_create()
5505 creates a response and optionally adds a basic response structure.
5506 OCSP_basic_add1_status() adds a complete single response to a basic
5507 response and returns the OCSP_SINGLERESP structure just added (to allow
5508 extensions to be included for example). OCSP_basic_add1_cert() adds a
5509 certificate to a basic response and OCSP_basic_sign() signs a basic
5510 response with various flags. New helper functions ASN1_TIME_check()
5511 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5512 (converts ASN1_TIME to GeneralizedTime).
5515 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5516 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5517 structure from a certificate. X509_pubkey_digest() digests the public_key
5518 contents: this is used in various key identifiers.
5521 *) Make sk_sort() tolerate a NULL argument.
5522 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5524 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5525 passed by the function are trusted implicitly. If any of them signed the
5526 response then it is assumed to be valid and is not verified.
5529 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5530 to data. This was previously part of the PKCS7 ASN1 code. This
5531 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5532 [Steve Henson, reported by Kenneth R. Robinette
5533 <support@securenetterm.com>]
5535 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5536 routines: without these tracing memory leaks is very painful.
5537 Fix leaks in PKCS12 and PKCS7 routines.
5540 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5541 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5542 effectively meant GeneralizedTime would never be used. Now it
5543 is initialised to -1 but X509_time_adj() now has to check the value
5544 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5545 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5546 [Steve Henson, reported by Kenneth R. Robinette
5547 <support@securenetterm.com>]
5549 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5550 result in a zero length in the ASN1_INTEGER structure which was
5551 not consistent with the structure when d2i_ASN1_INTEGER() was used
5552 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5553 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5554 where it did not print out a minus for negative ASN1_INTEGER.
5557 *) Add summary printout to ocsp utility. The various functions which
5558 convert status values to strings have been renamed to:
5559 OCSP_response_status_str(), OCSP_cert_status_str() and
5560 OCSP_crl_reason_str() and are no longer static. New options
5561 to verify nonce values and to disable verification. OCSP response
5562 printout format cleaned up.
5565 *) Add additional OCSP certificate checks. These are those specified
5566 in RFC2560. This consists of two separate checks: the CA of the
5567 certificate being checked must either be the OCSP signer certificate
5568 or the issuer of the OCSP signer certificate. In the latter case the
5569 OCSP signer certificate must contain the OCSP signing extended key
5570 usage. This check is performed by attempting to match the OCSP
5571 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5572 in the OCSP_CERTID structures of the response.
5575 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5576 and related routines. This uses the standard OpenSSL certificate
5577 verify routines to perform initial checks (just CA validity) and
5578 to obtain the certificate chain. Then additional checks will be
5579 performed on the chain. Currently the root CA is checked to see
5580 if it is explicitly trusted for OCSP signing. This is used to set
5581 a root CA as a global signing root: that is any certificate that
5582 chains to that CA is an acceptable OCSP signing certificate.
5585 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5586 extensions from a separate configuration file.
5587 As when reading extensions from the main configuration file,
5588 the '-extensions ...' option may be used for specifying the
5590 [Massimiliano Pala <madwolf@comune.modena.it>]
5592 *) New OCSP utility. Allows OCSP requests to be generated or
5593 read. The request can be sent to a responder and the output
5594 parsed, outputed or printed in text form. Not complete yet:
5595 still needs to check the OCSP response validity.
5598 *) New subcommands for 'openssl ca':
5599 'openssl ca -status <serial>' prints the status of the cert with
5600 the given serial number (according to the index file).
5601 'openssl ca -updatedb' updates the expiry status of certificates
5603 [Massimiliano Pala <madwolf@comune.modena.it>]
5605 *) New '-newreq-nodes' command option to CA.pl. This is like
5606 '-newreq', but calls 'openssl req' with the '-nodes' option
5607 so that the resulting key is not encrypted.
5608 [Damien Miller <djm@mindrot.org>]
5610 *) New configuration for the GNU Hurd.
5611 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5613 *) Initial code to implement OCSP basic response verify. This
5614 is currently incomplete. Currently just finds the signer's
5615 certificate and verifies the signature on the response.
5618 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5619 value of OPENSSLDIR. This is available via the new '-d' option
5620 to 'openssl version', and is also included in 'openssl version -a'.
5623 *) Allowing defining memory allocation callbacks that will be given
5624 file name and line number information in additional arguments
5625 (a const char* and an int). The basic functionality remains, as
5626 well as the original possibility to just replace malloc(),
5627 realloc() and free() by functions that do not know about these
5628 additional arguments. To register and find out the current
5629 settings for extended allocation functions, the following
5630 functions are provided:
5632 CRYPTO_set_mem_ex_functions
5633 CRYPTO_set_locked_mem_ex_functions
5634 CRYPTO_get_mem_ex_functions
5635 CRYPTO_get_locked_mem_ex_functions
5637 These work the same way as CRYPTO_set_mem_functions and friends.
5638 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5639 extended allocation function is enabled.
5640 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5641 a conventional allocation function is enabled.
5642 [Richard Levitte, Bodo Moeller]
5644 *) Finish off removing the remaining LHASH function pointer casts.
5645 There should no longer be any prototype-casting required when using
5646 the LHASH abstraction, and any casts that remain are "bugs". See
5647 the callback types and macros at the head of lhash.h for details
5648 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5651 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5652 If /dev/[u]random devices are not available or do not return enough
5653 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5655 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5656 /etc/entropy will be queried once each in this sequence, quering stops
5657 when enough entropy was collected without querying more sockets.
5660 *) Change the Unix RAND_poll() variant to be able to poll several
5661 random devices, as specified by DEVRANDOM, until a sufficient amount
5662 of data has been collected. We spend at most 10 ms on each file
5663 (select timeout) and read in non-blocking mode. DEVRANDOM now
5664 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5665 (previously it was just the string "/dev/urandom"), so on typical
5666 platforms the 10 ms delay will never occur.
5667 Also separate out the Unix variant to its own file, rand_unix.c.
5668 For VMS, there's a currently-empty rand_vms.c.
5671 *) Move OCSP client related routines to ocsp_cl.c. These
5672 provide utility functions which an application needing
5673 to issue a request to an OCSP responder and analyse the
5674 response will typically need: as opposed to those which an
5675 OCSP responder itself would need which will be added later.
5677 OCSP_request_sign() signs an OCSP request with an API similar
5678 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5679 response. OCSP_response_get1_basic() extracts basic response
5680 from response. OCSP_resp_find_status(): finds and extracts status
5681 information from an OCSP_CERTID structure (which will be created
5682 when the request structure is built). These are built from lower
5683 level functions which work on OCSP_SINGLERESP structures but
5684 wont normally be used unless the application wishes to examine
5685 extensions in the OCSP response for example.
5687 Replace nonce routines with a pair of functions.
5688 OCSP_request_add1_nonce() adds a nonce value and optionally
5689 generates a random value. OCSP_check_nonce() checks the
5690 validity of the nonce in an OCSP response.
5693 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5694 This doesn't copy the supplied OCSP_CERTID and avoids the
5695 need to free up the newly created id. Change return type
5696 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5697 This can then be used to add extensions to the request.
5698 Deleted OCSP_request_new(), since most of its functionality
5699 is now in OCSP_REQUEST_new() (and the case insensitive name
5700 clash) apart from the ability to set the request name which
5701 will be added elsewhere.
5704 *) Update OCSP API. Remove obsolete extensions argument from
5705 various functions. Extensions are now handled using the new
5706 OCSP extension code. New simple OCSP HTTP function which
5707 can be used to send requests and parse the response.
5710 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5711 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5712 uses the special reorder version of SET OF to sort the attributes
5713 and reorder them to match the encoded order. This resolves a long
5714 standing problem: a verify on a PKCS7 structure just after signing
5715 it used to fail because the attribute order did not match the
5716 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5717 it uses the received order. This is necessary to tolerate some broken
5718 software that does not order SET OF. This is handled by encoding
5719 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5720 to produce the required SET OF.
5723 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5724 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5725 files to get correct declarations of the ASN.1 item variables.
5728 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5729 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5730 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5731 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5732 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5733 ASN1_ITEM and no wrapper functions.
5736 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5737 replace the old function pointer based I/O routines. Change most of
5738 the *_d2i_bio() and *_d2i_fp() functions to use these.
5741 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5742 lines, recognice more "algorithms" that can be deselected, and make
5743 it complain about algorithm deselection that isn't recognised.
5746 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5747 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5748 to use new functions. Add NO_ASN1_OLD which can be set to remove
5749 some old style ASN1 functions: this can be used to determine if old
5750 code will still work when these eventually go away.
5753 *) New extension functions for OCSP structures, these follow the
5754 same conventions as certificates and CRLs.
5757 *) New function X509V3_add1_i2d(). This automatically encodes and
5758 adds an extension. Its behaviour can be customised with various
5759 flags to append, replace or delete. Various wrappers added for
5760 certifcates and CRLs.
5763 *) Fix to avoid calling the underlying ASN1 print routine when
5764 an extension cannot be parsed. Correct a typo in the
5765 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5768 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5769 entries for variables.
5772 *) Add functionality to apps/openssl.c for detecting locking
5773 problems: As the program is single-threaded, all we have
5774 to do is register a locking callback using an array for
5775 storing which locks are currently held by the program.
5778 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5779 SSL_get_ex_data_X509_STORE_idx(), which is used in
5780 ssl_verify_cert_chain() and thus can be called at any time
5781 during TLS/SSL handshakes so that thread-safety is essential.
5782 Unfortunately, the ex_data design is not at all suited
5783 for multi-threaded use, so it probably should be abolished.
5786 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5787 [Broadcom, tweaked and integrated by Geoff Thorpe]
5789 *) Move common extension printing code to new function
5790 X509V3_print_extensions(). Reorganise OCSP print routines and
5791 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5794 *) New function X509_signature_print() to remove duplication in some
5798 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5799 set (this was treated exactly the same as SET OF previously). This
5800 is used to reorder the STACK representing the structure to match the
5801 encoding. This will be used to get round a problem where a PKCS7
5802 structure which was signed could not be verified because the STACK
5803 order did not reflect the encoded order.
5806 *) Reimplement the OCSP ASN1 module using the new code.
5809 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5810 for its ASN1 operations. The old style function pointers still exist
5811 for now but they will eventually go away.
5814 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5815 completely replaces the old ASN1 functionality with a table driven
5816 encoder and decoder which interprets an ASN1_ITEM structure describing
5817 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5818 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5819 has also been converted to the new form.
5822 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5823 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5824 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5825 for negative moduli.
5828 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5829 of not touching the result's sign bit.
5832 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5836 *) Changed the LHASH code to use prototypes for callbacks, and created
5837 macros to declare and implement thin (optionally static) functions
5838 that provide type-safety and avoid function pointer casting for the
5839 type-specific callbacks.
5842 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5844 [Veers Staats <staatsvr@asc.hpc.mil>,
5845 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5847 *) Reformat the FAQ so the different questions and answers can be divided
5848 in sections depending on the subject.
5851 *) Have the zlib compression code load ZLIB.DLL dynamically under
5855 *) New function BN_mod_sqrt for computing square roots modulo a prime
5856 (using the probabilistic Tonelli-Shanks algorithm unless
5857 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5858 be handled deterministically).
5859 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5861 *) Make BN_mod_inverse faster by explicitly handling small quotients
5862 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5863 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5866 *) New function BN_kronecker.
5869 *) Fix BN_gcd so that it works on negative inputs; the result is
5870 positive unless both parameters are zero.
5871 Previously something reasonably close to an infinite loop was
5872 possible because numbers could be growing instead of shrinking
5873 in the implementation of Euclid's algorithm.
5876 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5877 sign of the number in question.
5879 Fix BN_is_word(a,w) to work correctly for w == 0.
5881 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5882 because its test if the absolute value of 'a' equals 'w'.
5883 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5884 it exists mostly for use in the implementations of BN_is_zero(),
5885 BN_is_one(), and BN_is_word().
5888 *) New function BN_swap.
5891 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5892 the exponentiation functions are more likely to produce reasonable
5893 results on negative inputs.
5896 *) Change BN_mod_mul so that the result is always non-negative.
5897 Previously, it could be negative if one of the factors was negative;
5898 I don't think anyone really wanted that behaviour.
5901 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5902 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5903 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5904 and add new functions:
5913 BN_mod_lshift1_quick
5917 These functions always generate non-negative results.
5919 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5920 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5922 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5923 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5924 be reduced modulo m.
5925 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5928 The following entry accidentily appeared in the CHANGES file
5929 distributed with OpenSSL 0.9.7. The modifications described in
5930 it do *not* apply to OpenSSL 0.9.7.
5932 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5933 was actually never needed) and in BN_mul(). The removal in BN_mul()
5934 required a small change in bn_mul_part_recursive() and the addition
5935 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5936 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5937 bn_sub_words() and bn_add_words() except they take arrays with
5942 *) In 'openssl passwd', verify passwords read from the terminal
5943 unless the '-salt' option is used (which usually means that
5944 verification would just waste user's time since the resulting
5945 hash is going to be compared with some given password hash)
5946 or the new '-noverify' option is used.
5948 This is an incompatible change, but it does not affect
5949 non-interactive use of 'openssl passwd' (passwords on the command
5950 line, '-stdin' option, '-in ...' option) and thus should not
5954 *) Remove all references to RSAref, since there's no more need for it.
5957 *) Make DSO load along a path given through an environment variable
5958 (SHLIB_PATH) with shl_load().
5961 *) Constify the ENGINE code as a result of BIGNUM constification.
5962 Also constify the RSA code and most things related to it. In a
5963 few places, most notable in the depth of the ASN.1 code, ugly
5964 casts back to non-const were required (to be solved at a later
5968 *) Make it so the openssl application has all engines loaded by default.
5971 *) Constify the BIGNUM routines a little more.
5974 *) Add the following functions:
5976 ENGINE_load_cswift()
5978 ENGINE_load_atalla()
5980 ENGINE_load_builtin_engines()
5982 That way, an application can itself choose if external engines that
5983 are built-in in OpenSSL shall ever be used or not. The benefit is
5984 that applications won't have to be linked with libdl or other dso
5985 libraries unless it's really needed.
5987 Changed 'openssl engine' to load all engines on demand.
5988 Changed the engine header files to avoid the duplication of some
5989 declarations (they differed!).
5992 *) 'openssl engine' can now list capabilities.
5995 *) Better error reporting in 'openssl engine'.
5998 *) Never call load_dh_param(NULL) in s_server.
6001 *) Add engine application. It can currently list engines by name and
6002 identity, and test if they are actually available.
6005 *) Improve RPM specification file by forcing symbolic linking and making
6006 sure the installed documentation is also owned by root.root.
6007 [Damien Miller <djm@mindrot.org>]
6009 *) Give the OpenSSL applications more possibilities to make use of
6010 keys (public as well as private) handled by engines.
6013 *) Add OCSP code that comes from CertCo.
6016 *) Add VMS support for the Rijndael code.
6019 *) Added untested support for Nuron crypto accelerator.
6022 *) Add support for external cryptographic devices. This code was
6023 previously distributed separately as the "engine" branch.
6024 [Geoff Thorpe, Richard Levitte]
6026 *) Rework the filename-translation in the DSO code. It is now possible to
6027 have far greater control over how a "name" is turned into a filename
6028 depending on the operating environment and any oddities about the
6029 different shared library filenames on each system.
6032 *) Support threads on FreeBSD-elf in Configure.
6035 *) Fix for SHA1 assembly problem with MASM: it produces
6036 warnings about corrupt line number information when assembling
6037 with debugging information. This is caused by the overlapping
6039 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6042 NCONF_get_number() has no error checking at all. As a replacement,
6043 NCONF_get_number_e() is defined (_e for "error checking") and is
6044 promoted strongly. The old NCONF_get_number is kept around for
6045 binary backward compatibility.
6046 Make it possible for methods to load from something other than a BIO,
6047 by providing a function pointer that is given a name instead of a BIO.
6048 For example, this could be used to load configuration data from an
6052 *) Fix for non blocking accept BIOs. Added new I/O special reason
6053 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6054 with non blocking I/O was not possible because no retry code was
6055 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6059 *) Added the beginnings of Rijndael support.
6062 *) Fix for bug in DirectoryString mask setting. Add support for
6063 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6064 to allow certificate printing to more controllable, additional
6065 'certopt' option to 'x509' to allow new printing options to be
6069 *) Clean old EAY MD5 hack from e_os.h.
6072 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6074 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6075 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6076 [Joe Orton, Steve Henson]
6078 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6080 *) Fix additional bug revealed by the NISCC test suite:
6082 Stop bug triggering large recursion when presented with
6083 certain ASN.1 tags (CVE-2003-0851)
6086 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6088 *) Fix various bugs revealed by running the NISCC test suite:
6090 Stop out of bounds reads in the ASN1 code when presented with
6091 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6093 If verify callback ignores invalid public key errors don't try to check
6094 certificate signature with the NULL public key.
6098 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6099 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6103 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6104 extra data after the compression methods not only for TLS 1.0
6105 but also for SSL 3.0 (as required by the specification).
6106 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6108 *) Change X509_certificate_type() to mark the key as exported/exportable
6109 when it's 512 *bits* long, not 512 bytes.
6112 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6114 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6115 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6116 a protocol version number mismatch like a decryption error
6117 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6120 *) Turn on RSA blinding by default in the default implementation
6121 to avoid a timing attack. Applications that don't want it can call
6122 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6123 They would be ill-advised to do so in most cases.
6124 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6126 *) Change RSA blinding code so that it works when the PRNG is not
6127 seeded (in this case, the secret RSA exponent is abused as
6128 an unpredictable seed -- if it is not unpredictable, there
6129 is no point in blinding anyway). Make RSA blinding thread-safe
6130 by remembering the creator's thread ID in rsa->blinding and
6131 having all other threads use local one-time blinding factors
6132 (this requires more computation than sharing rsa->blinding, but
6133 avoids excessive locking; and if an RSA object is not shared
6134 between threads, blinding will still be very fast).
6137 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6139 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6140 via timing by performing a MAC computation even if incorrrect
6141 block cipher padding has been found. This is a countermeasure
6142 against active attacks where the attacker has to distinguish
6143 between bad padding and a MAC verification error. (CVE-2003-0078)
6145 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6146 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6147 Martin Vuagnoux (EPFL, Ilion)]
6149 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6151 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6152 memory from it's contents. This is done with a counter that will
6153 place alternating values in each byte. This can be used to solve
6154 two issues: 1) the removal of calls to memset() by highly optimizing
6155 compilers, and 2) cleansing with other values than 0, since those can
6156 be read through on certain media, for example a swap space on disk.
6159 *) Bugfix: client side session caching did not work with external caching,
6160 because the session->cipher setting was not restored when reloading
6161 from the external cache. This problem was masked, when
6162 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6163 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6166 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6167 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6168 [Zeev Lieber <zeev-l@yahoo.com>]
6170 *) Undo an undocumented change introduced in 0.9.6e which caused
6171 repeated calls to OpenSSL_add_all_ciphers() and
6172 OpenSSL_add_all_digests() to be ignored, even after calling
6176 *) Change the default configuration reader to deal with last line not
6177 being properly terminated.
6180 *) Change X509_NAME_cmp() so it applies the special rules on handling
6181 DN values that are of type PrintableString, as well as RDNs of type
6182 emailAddress where the value has the type ia5String.
6183 [stefank@valicert.com via Richard Levitte]
6185 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6186 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6187 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6188 the bitwise-OR of the two for use by the majority of applications
6189 wanting this behaviour, and update the docs. The documented
6190 behaviour and actual behaviour were inconsistent and had been
6191 changing anyway, so this is more a bug-fix than a behavioural
6193 [Geoff Thorpe, diagnosed by Nadav Har'El]
6195 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6196 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6199 *) Fix initialization code race conditions in
6200 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6201 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6202 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6203 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6204 ssl2_get_cipher_by_char(),
6205 ssl3_get_cipher_by_char().
6206 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6208 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6209 the cached sessions are flushed, as the remove_cb() might use ex_data
6210 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6211 (see [openssl.org #212]).
6212 [Geoff Thorpe, Lutz Jaenicke]
6214 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6215 length, instead of the encoding length to d2i_ASN1_OBJECT.
6218 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6220 *) [In 0.9.6g-engine release:]
6221 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6222 [Lynn Gazis <lgazis@rainbow.com>]
6224 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6226 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6227 and get fix the header length calculation.
6228 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6229 Alon Kantor <alonk@checkpoint.com> (and others),
6232 *) Use proper error handling instead of 'assertions' in buffer
6233 overflow checks added in 0.9.6e. This prevents DoS (the
6234 assertions could call abort()).
6235 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6237 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6239 *) Add various sanity checks to asn1_get_length() to reject
6240 the ASN1 length bytes if they exceed sizeof(long), will appear
6241 negative or the content length exceeds the length of the
6243 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6245 *) Fix cipher selection routines: ciphers without encryption had no flags
6246 for the cipher strength set and where therefore not handled correctly
6247 by the selection routines (PR #130).
6250 *) Fix EVP_dsa_sha macro.
6254 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6255 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6256 that was added in OpenSSL 0.9.6d.
6258 As the countermeasure turned out to be incompatible with some
6259 broken SSL implementations, the new option is part of SSL_OP_ALL.
6260 SSL_OP_ALL is usually employed when compatibility with weird SSL
6261 implementations is desired (e.g. '-bugs' option to 's_client' and
6262 's_server'), so the new option is automatically set in many
6266 *) Changes in security patch:
6268 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6269 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6270 Air Force Materiel Command, USAF, under agreement number
6273 *) Add various sanity checks to asn1_get_length() to reject
6274 the ASN1 length bytes if they exceed sizeof(long), will appear
6275 negative or the content length exceeds the length of the
6276 supplied buffer. (CVE-2002-0659)
6277 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6279 *) Assertions for various potential buffer overflows, not known to
6281 [Ben Laurie (CHATS)]
6283 *) Various temporary buffers to hold ASCII versions of integers were
6284 too small for 64 bit platforms. (CVE-2002-0655)
6285 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6287 *) Remote buffer overflow in SSL3 protocol - an attacker could
6288 supply an oversized session ID to a client. (CVE-2002-0656)
6289 [Ben Laurie (CHATS)]
6291 *) Remote buffer overflow in SSL2 protocol - an attacker could
6292 supply an oversized client master key. (CVE-2002-0656)
6293 [Ben Laurie (CHATS)]
6295 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6297 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6298 encoded as NULL) with id-dsa-with-sha1.
6299 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6301 *) Check various X509_...() return values in apps/req.c.
6302 [Nils Larsch <nla@trustcenter.de>]
6304 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6305 an end-of-file condition would erronously be flagged, when the CRLF
6306 was just at the end of a processed block. The bug was discovered when
6307 processing data through a buffering memory BIO handing the data to a
6308 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6309 <ptsekov@syntrex.com> and Nedelcho Stanev.
6312 *) Implement a countermeasure against a vulnerability recently found
6313 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6314 before application data chunks to avoid the use of known IVs
6315 with data potentially chosen by the attacker.
6318 *) Fix length checks in ssl3_get_client_hello().
6321 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6322 to prevent ssl3_read_internal() from incorrectly assuming that
6323 ssl3_read_bytes() found application data while handshake
6324 processing was enabled when in fact s->s3->in_read_app_data was
6325 merely automatically cleared during the initial handshake.
6326 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6328 *) Fix object definitions for Private and Enterprise: they were not
6329 recognized in their shortname (=lowercase) representation. Extend
6330 obj_dat.pl to issue an error when using undefined keywords instead
6331 of silently ignoring the problem (Svenning Sorensen
6332 <sss@sss.dnsalias.net>).
6335 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6336 generators, i.e. generators other than 2 and 5. (Previously, the
6337 code did not properly initialise the 'add' and 'rem' values to
6338 BN_generate_prime().)
6340 In the new general case, we do not insist that 'generator' is
6341 actually a primitive root: This requirement is rather pointless;
6342 a generator of the order-q subgroup is just as good, if not
6346 *) Map new X509 verification errors to alerts. Discovered and submitted by
6347 Tom Wu <tom@arcot.com>.
6350 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6351 returning non-zero before the data has been completely received
6352 when using non-blocking I/O.
6353 [Bodo Moeller; problem pointed out by John Hughes]
6355 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6356 [Ben Laurie, Lutz Jaenicke]
6358 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6359 Yoram Zahavi <YoramZ@gilian.com>).
6362 *) Add information about CygWin 1.3 and on, and preserve proper
6363 configuration for the versions before that.
6364 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6366 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6367 check whether we deal with a copy of a session and do not delete from
6368 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6369 <izhar@checkpoint.com>.
6372 *) Do not store session data into the internal session cache, if it
6373 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6374 flag is set). Proposed by Aslam <aslam@funk.com>.
6377 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6381 *) [In 0.9.6d-engine release:]
6382 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6383 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6385 *) Add the configuration target linux-s390x.
6386 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6388 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6389 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6390 variable as an indication that a ClientHello message has been
6391 received. As the flag value will be lost between multiple
6392 invocations of ssl3_accept when using non-blocking I/O, the
6393 function may not be aware that a handshake has actually taken
6394 place, thus preventing a new session from being added to the
6397 To avoid this problem, we now set s->new_session to 2 instead of
6398 using a local variable.
6399 [Lutz Jaenicke, Bodo Moeller]
6401 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6402 if the SSL_R_LENGTH_MISMATCH error is detected.
6403 [Geoff Thorpe, Bodo Moeller]
6405 *) New 'shared_ldflag' column in Configure platform table.
6408 *) Fix EVP_CIPHER_mode macro.
6409 ["Dan S. Camper" <dan@bti.net>]
6411 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6412 type, we must throw them away by setting rr->length to 0.
6413 [D P Chang <dpc@qualys.com>]
6415 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6417 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6418 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6419 worked incorrectly for those cases where range = 10..._2 and
6420 3*range is two bits longer than range.)
6423 *) Only add signing time to PKCS7 structures if it is not already
6427 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6428 OBJ_ld_ce should be OBJ_id_ce.
6429 Also some ip-pda OIDs in crypto/objects/objects.txt were
6430 incorrect (cf. RFC 3039).
6431 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6433 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6434 returns early because it has nothing to do.
6435 [Andy Schneider <andy.schneider@bjss.co.uk>]
6437 *) [In 0.9.6c-engine release:]
6438 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6439 [Andy Schneider <andy.schneider@bjss.co.uk>]
6441 *) [In 0.9.6c-engine release:]
6442 Add support for Cryptographic Appliance's keyserver technology.
6443 (Use engine 'keyclient')
6444 [Cryptographic Appliances and Geoff Thorpe]
6446 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6447 is called via tools/c89.sh because arguments have to be
6448 rearranged (all '-L' options must appear before the first object
6450 [Richard Shapiro <rshapiro@abinitio.com>]
6452 *) [In 0.9.6c-engine release:]
6453 Add support for Broadcom crypto accelerator cards, backported
6455 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6457 *) [In 0.9.6c-engine release:]
6458 Add support for SureWare crypto accelerator cards from
6459 Baltimore Technologies. (Use engine 'sureware')
6460 [Baltimore Technologies and Mark Cox]
6462 *) [In 0.9.6c-engine release:]
6463 Add support for crypto accelerator cards from Accelerated
6464 Encryption Processing, www.aep.ie. (Use engine 'aep')
6465 [AEP Inc. and Mark Cox]
6467 *) Add a configuration entry for gcc on UnixWare.
6468 [Gary Benson <gbenson@redhat.com>]
6470 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6471 messages are stored in a single piece (fixed-length part and
6472 variable-length part combined) and fix various bugs found on the way.
6475 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6476 instead. BIO_gethostbyname() does not know what timeouts are
6477 appropriate, so entries would stay in cache even when they have
6479 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6481 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6482 faced with a pathologically small ClientHello fragment that does
6483 not contain client_version: Instead of aborting with an error,
6484 simply choose the highest available protocol version (i.e.,
6485 TLS 1.0 unless it is disabled). In practice, ClientHello
6486 messages are never sent like this, but this change gives us
6487 strictly correct behaviour at least for TLS.
6490 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6491 never resets s->method to s->ctx->method when called from within
6492 one of the SSL handshake functions.
6493 [Bodo Moeller; problem pointed out by Niko Baric]
6495 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6496 (sent using the client's version number) if client_version is
6497 smaller than the protocol version in use. Also change
6498 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6499 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6500 the client will at least see that alert.
6503 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6507 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6508 client receives HelloRequest while in a handshake.
6509 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6511 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6512 should end in 'break', not 'goto end' which circuments various
6513 cleanups done in state SSL_ST_OK. But session related stuff
6514 must be disabled for SSL_ST_OK in the case that we just sent a
6517 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6518 before just sending a HelloRequest.
6519 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6521 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6522 reveal whether illegal block cipher padding was found or a MAC
6523 verification error occured. (Neither SSLerr() codes nor alerts
6524 are directly visible to potential attackers, but the information
6525 may leak via logfiles.)
6527 Similar changes are not required for the SSL 2.0 implementation
6528 because the number of padding bytes is sent in clear for SSL 2.0,
6529 and the extra bytes are just ignored. However ssl/s2_pkt.c
6530 failed to verify that the purported number of padding bytes is in
6534 *) Add OpenUNIX-8 support including shared libraries
6535 (Boyd Lynn Gerber <gerberb@zenez.com>).
6538 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6539 'wristwatch attack' using huge encoding parameters (cf.
6540 James H. Manger's CRYPTO 2001 paper). Note that the
6541 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6542 encoding parameters and hence was not vulnerable.
6545 *) BN_sqr() bug fix.
6546 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6548 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6549 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6550 followed by modular reduction.
6551 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6553 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6554 equivalent based on BN_pseudo_rand() instead of BN_rand().
6557 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6558 This function was broken, as the check for a new client hello message
6559 to handle SGC did not allow these large messages.
6560 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6563 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6566 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6567 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6570 *) Rework the configuration and shared library support for Tru64 Unix.
6571 The configuration part makes use of modern compiler features and
6572 still retains old compiler behavior for those that run older versions
6573 of the OS. The shared library support part includes a variant that
6574 uses the RPATH feature, and is available through the special
6575 configuration target "alpha-cc-rpath", which will never be selected
6577 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6579 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6580 with the same message size as in ssl3_get_certificate_request().
6581 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6582 messages might inadvertently be reject as too long.
6583 [Petr Lampa <lampa@fee.vutbr.cz>]
6585 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6588 *) Modified SSL library such that the verify_callback that has been set
6589 specificly for an SSL object with SSL_set_verify() is actually being
6590 used. Before the change, a verify_callback set with this function was
6591 ignored and the verify_callback() set in the SSL_CTX at the time of
6592 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6593 to allow the necessary settings.
6596 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6597 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6598 done automatically (in contradiction to the requirements of the C
6599 standard). This made problems when used from OpenSSH.
6602 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6603 dh->length and always used
6605 BN_rand_range(priv_key, dh->p).
6607 BN_rand_range() is not necessary for Diffie-Hellman, and this
6608 specific range makes Diffie-Hellman unnecessarily inefficient if
6609 dh->length (recommended exponent length) is much smaller than the
6610 length of dh->p. We could use BN_rand_range() if the order of
6611 the subgroup was stored in the DH structure, but we only have
6616 BN_rand(priv_key, l, ...)
6618 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6624 RSA_eay_public_encrypt
6625 RSA_eay_private_decrypt
6626 RSA_eay_private_encrypt (signing)
6627 RSA_eay_public_decrypt (signature verification)
6629 (default implementations for RSA_public_encrypt,
6630 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6631 always reject numbers >= n.
6634 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6635 to synchronize access to 'locking_thread'. This is necessary on
6636 systems where access to 'locking_thread' (an 'unsigned long'
6637 variable) is not atomic.
6640 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6641 *before* setting the 'crypto_lock_rand' flag. The previous code had
6642 a race condition if 0 is a valid thread ID.
6643 [Travis Vitek <vitek@roguewave.com>]
6645 *) Add support for shared libraries under Irix.
6646 [Albert Chin-A-Young <china@thewrittenword.com>]
6648 *) Add configuration option to build on Linux on both big-endian and
6650 [Ralf Baechle <ralf@uni-koblenz.de>]
6652 *) Add the possibility to create shared libraries on HP-UX.
6655 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6657 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6658 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6659 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6660 PRNG state recovery was possible based on the output of
6661 one PRNG request appropriately sized to gain knowledge on
6662 'md' followed by enough consecutive 1-byte PRNG requests
6663 to traverse all of 'state'.
6665 1. When updating 'md_local' (the current thread's copy of 'md')
6666 during PRNG output generation, hash all of the previous
6667 'md_local' value, not just the half used for PRNG output.
6669 2. Make the number of bytes from 'state' included into the hash
6670 independent from the number of PRNG bytes requested.
6672 The first measure alone would be sufficient to avoid
6673 Markku-Juhani's attack. (Actually it had never occurred
6674 to me that the half of 'md_local' used for chaining was the
6675 half from which PRNG output bytes were taken -- I had always
6676 assumed that the secret half would be used.) The second
6677 measure makes sure that additional data from 'state' is never
6678 mixed into 'md_local' in small portions; this heuristically
6679 further strengthens the PRNG.
6682 *) Fix crypto/bn/asm/mips3.s.
6685 *) When only the key is given to "enc", the IV is undefined. Print out
6686 an error message in this case.
6689 *) Handle special case when X509_NAME is empty in X509 printing routines.
6692 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6693 positive and less than q.
6696 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6697 used: it isn't thread safe and the add_lock_callback should handle
6699 [Paul Rose <Paul.Rose@bridge.com>]
6701 *) Verify that incoming data obeys the block size in
6702 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6706 [Ulf Möller, Bodo Möller]
6708 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6709 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6710 when fixing the server behaviour for backwards-compatible 'client
6711 hello' messages. (Note that the attack is impractical against
6712 SSL 3.0 and TLS 1.0 anyway because length and version checking
6713 means that the probability of guessing a valid ciphertext is
6714 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6717 Before 0.9.5, the countermeasure (hide the error by generating a
6718 random 'decryption result') did not work properly because
6719 ERR_clear_error() was missing, meaning that SSL_get_error() would
6720 detect the supposedly ignored error.
6722 Both problems are now fixed.
6725 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6726 (previously it was 1024).
6729 *) Fix for compatibility mode trust settings: ignore trust settings
6730 unless some valid trust or reject settings are present.
6733 *) Fix for blowfish EVP: its a variable length cipher.
6736 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6737 parameters in DSA public key structures and return an error in the
6738 DSA routines if parameters are absent.
6741 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6742 in the current directory if neither $RANDFILE nor $HOME was set.
6743 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6744 caused some confusion to Windows users who haven't defined $HOME.
6745 Thus RAND_file_name() is changed again: e_os.h can define a
6746 DEFAULT_HOME, which will be used if $HOME is not set.
6747 For Windows, we use "C:"; on other platforms, we still require
6748 environment variables.
6750 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6751 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6752 having multiple threads call RAND_poll() concurrently.
6755 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6756 combination of a flag and a thread ID variable.
6757 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6758 flag), *other* threads can enter ssleay_add_bytes without obeying
6759 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6760 that they do not hold after the first thread unsets add_do_not_lock).
6763 *) Change bctest again: '-x' expressions are not available in all
6767 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6769 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6770 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6772 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6773 the default extension for executables, if any. Also, make the perl
6774 scripts that use symlink() to test if it really exists and use "cp"
6775 if it doesn't. All this made OpenSSL compilable and installable in
6779 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6780 If SEQUENCE is length is indefinite just set c->slen to the total
6781 amount of data available.
6782 [Steve Henson, reported by shige@FreeBSD.org]
6783 [This change does not apply to 0.9.7.]
6785 *) Change bctest to avoid here-documents inside command substitution
6786 (workaround for FreeBSD /bin/sh bug).
6787 For compatibility with Ultrix, avoid shell functions (introduced
6788 in the bctest version that searches along $PATH).
6791 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6792 with des_encrypt() defined on some operating systems, like Solaris
6796 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6797 On the Importance of Eliminating Errors in Cryptographic
6798 Computations, J. Cryptology 14 (2001) 2, 101-119,
6799 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6802 *) MIPS assembler BIGNUM division bug fix.
6805 *) Disabled incorrect Alpha assembler code.
6808 *) Fix PKCS#7 decode routines so they correctly update the length
6809 after reading an EOC for the EXPLICIT tag.
6811 [This change does not apply to 0.9.7.]
6813 *) Fix bug in PKCS#12 key generation routines. This was triggered
6814 if a 3DES key was generated with a 0 initial byte. Include
6815 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6816 (but broken) behaviour.
6819 *) Enhance bctest to search for a working bc along $PATH and print
6821 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6823 *) Fix memory leaks in err.c: free err_data string if necessary;
6824 don't write to the wrong index in ERR_set_error_data.
6827 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6831 *) Replace rdtsc with _emit statements for VC++ version 5.
6832 [Jeremy Cooper <jeremy@baymoo.org>]
6834 *) Make it possible to reuse SSLv2 sessions.
6837 *) In copy_email() check for >= 0 as a return value for
6838 X509_NAME_get_index_by_NID() since 0 is a valid index.
6839 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6841 *) Avoid coredump with unsupported or invalid public keys by checking if
6842 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6843 PKCS7_verify() fails with non detached data.
6846 *) Don't use getenv in library functions when run as setuid/setgid.
6847 New function OPENSSL_issetugid().
6850 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6851 due to incorrect handling of multi-threading:
6853 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6855 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6857 3. Count how many times MemCheck_off() has been called so that
6858 nested use can be treated correctly. This also avoids
6859 inband-signalling in the previous code (which relied on the
6860 assumption that thread ID 0 is impossible).
6863 *) Add "-rand" option also to s_client and s_server.
6866 *) Fix CPU detection on Irix 6.x.
6867 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6868 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6870 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6873 [This change does not apply to 0.9.7.]
6875 *) Use the cached encoding of an X509_NAME structure rather than
6876 copying it. This is apparently the reason for the libsafe "errors"
6877 but the code is actually correct.
6880 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6881 Bleichenbacher's DSA attack.
6882 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6883 to be set and top=0 forces the highest bit to be set; top=-1 is new
6884 and leaves the highest bit random.
6885 [Ulf Moeller, Bodo Moeller]
6887 *) In the NCONF_...-based implementations for CONF_... queries
6888 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6889 a temporary CONF structure with the data component set to NULL
6890 (which gives segmentation faults in lh_retrieve).
6891 Instead, use NULL for the CONF pointer in CONF_get_string and
6892 CONF_get_number (which may use environment variables) and directly
6893 return NULL from CONF_get_section.
6896 *) Fix potential buffer overrun for EBCDIC.
6899 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6900 keyUsage if basicConstraints absent for a CA.
6903 *) Make SMIME_write_PKCS7() write mail header values with a format that
6904 is more generally accepted (no spaces before the semicolon), since
6905 some programs can't parse those values properly otherwise. Also make
6906 sure BIO's that break lines after each write do not create invalid
6910 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6911 macros previously used would not encode an empty SEQUENCE OF
6912 and break the signature.
6914 [This change does not apply to 0.9.7.]
6916 *) Zero the premaster secret after deriving the master secret in
6920 *) Add some EVP_add_digest_alias registrations (as found in
6921 OpenSSL_add_all_digests()) to SSL_library_init()
6922 aka OpenSSL_add_ssl_algorithms(). This provides improved
6923 compatibility with peers using X.509 certificates
6924 with unconventional AlgorithmIdentifier OIDs.
6927 *) Fix for Irix with NO_ASM.
6928 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6930 *) ./config script fixes.
6931 [Ulf Moeller, Richard Levitte]
6933 *) Fix 'openssl passwd -1'.
6936 *) Change PKCS12_key_gen_asc() so it can cope with non null
6937 terminated strings whose length is passed in the passlen
6938 parameter, for example from PEM callbacks. This was done
6939 by adding an extra length parameter to asc2uni().
6940 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6942 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6943 call failed, free the DSA structure.
6946 *) Fix to uni2asc() to cope with zero length Unicode strings.
6947 These are present in some PKCS#12 files.
6950 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6951 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6952 when writing a 32767 byte record.
6953 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6955 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6956 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6958 (RSA objects have a reference count access to which is protected
6959 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6960 so they are meant to be shared between threads.)
6961 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6962 "Reddie, Steven" <Steven.Reddie@ca.com>]
6964 *) Fix a deadlock in CRYPTO_mem_leaks().
6967 *) Use better test patterns in bntest.
6970 *) rand_win.c fix for Borland C.
6973 *) BN_rshift bugfix for n == 0.
6976 *) Add a 'bctest' script that checks for some known 'bc' bugs
6977 so that 'make test' does not abort just because 'bc' is broken.
6980 *) Store verify_result within SSL_SESSION also for client side to
6981 avoid potential security hole. (Re-used sessions on the client side
6982 always resulted in verify_result==X509_V_OK, not using the original
6983 result of the server certificate verification.)
6986 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6987 SSL3_RT_APPLICATION_DATA, return 0.
6988 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6992 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6993 releases, have been re-implemented by renaming the previous
6994 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6995 and ssl3_read_internal, respectively, and adding 'peek' parameters
6996 to them. The new ssl[23]_{read,peek} functions are calls to
6997 ssl[23]_read_internal with the 'peek' flag set appropriately.
6998 A 'peek' parameter has also been added to ssl3_read_bytes, which
6999 does the actual work for ssl3_read_internal.
7002 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7003 the method-specific "init()" handler. Also clean up ex_data after
7004 calling the method-specific "finish()" handler. Previously, this was
7005 happening the other way round.
7008 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7009 The previous value, 12, was not always sufficient for BN_mod_exp().
7012 *) Make sure that shared libraries get the internal name engine with
7013 the full version number and not just 0. This should mark the
7014 shared libraries as not backward compatible. Of course, this should
7015 be changed again when we can guarantee backward binary compatibility.
7018 *) Fix typo in get_cert_by_subject() in by_dir.c
7019 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7021 *) Rework the system to generate shared libraries:
7023 - Make note of the expected extension for the shared libraries and
7024 if there is a need for symbolic links from for example libcrypto.so.0
7025 to libcrypto.so.0.9.7. There is extended info in Configure for
7028 - Make as few rebuilds of the shared libraries as possible.
7030 - Still avoid linking the OpenSSL programs with the shared libraries.
7032 - When installing, install the shared libraries separately from the
7036 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7038 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7039 and not in SSL_clear because the latter is also used by the
7040 accept/connect functions; previously, the settings made by
7041 SSL_set_read_ahead would be lost during the handshake.
7042 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7044 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7045 Previously, it would create entries for disableed algorithms no
7049 *) Added several new manual pages for SSL_* function.
7052 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7054 *) In ssl23_get_client_hello, generate an error message when faced
7055 with an initial SSL 3.0/TLS record that is too small to contain the
7056 first two bytes of the ClientHello message, i.e. client_version.
7057 (Note that this is a pathologic case that probably has never happened
7058 in real life.) The previous approach was to use the version number
7059 from the record header as a substitute; but our protocol choice
7060 should not depend on that one because it is not authenticated
7061 by the Finished messages.
7064 *) More robust randomness gathering functions for Windows.
7065 [Jeffrey Altman <jaltman@columbia.edu>]
7067 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7068 not set then we don't setup the error code for issuer check errors
7069 to avoid possibly overwriting other errors which the callback does
7070 handle. If an application does set the flag then we assume it knows
7071 what it is doing and can handle the new informational codes
7075 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7076 a general "ANY" type, as such it should be able to decode anything
7077 including tagged types. However it didn't check the class so it would
7078 wrongly interpret tagged types in the same way as their universal
7079 counterpart and unknown types were just rejected. Changed so that the
7080 tagged and unknown types are handled in the same way as a SEQUENCE:
7081 that is the encoding is stored intact. There is also a new type
7082 "V_ASN1_OTHER" which is used when the class is not universal, in this
7083 case we have no idea what the actual type is so we just lump them all
7087 *) On VMS, stdout may very well lead to a file that is written to
7088 in a record-oriented fashion. That means that every write() will
7089 write a separate record, which will be read separately by the
7090 programs trying to read from it. This can be very confusing.
7092 The solution is to put a BIO filter in the way that will buffer
7093 text until a linefeed is reached, and then write everything a
7094 line at a time, so every record written will be an actual line,
7095 not chunks of lines and not (usually doesn't happen, but I've
7096 seen it once) several lines in one record. BIO_f_linebuffer() is
7099 Currently, it's a VMS-only method, because that's where it has
7100 been tested well enough.
7103 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7104 it can return incorrect results.
7105 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7106 but it was in 0.9.6-beta[12].)
7109 *) Disable the check for content being present when verifying detached
7110 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7111 include zero length content when signing messages.
7114 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7115 BIO_ctrl (for BIO pairs).
7118 *) Add DSO method for VMS.
7121 *) Bug fix: Montgomery multiplication could produce results with the
7125 *) Add RPM specification openssl.spec and modify it to build three
7126 packages. The default package contains applications, application
7127 documentation and run-time libraries. The devel package contains
7128 include files, static libraries and function documentation. The
7129 doc package contains the contents of the doc directory. The original
7130 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7133 *) Add a large number of documentation files for many SSL routines.
7134 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7136 *) Add a configuration entry for Sony News 4.
7137 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7139 *) Don't set the two most significant bits to one when generating a
7140 random number < q in the DSA library.
7143 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7144 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7145 the underlying transport is blocking) if a handshake took place.
7146 (The default behaviour is needed by applications such as s_client
7147 and s_server that use select() to determine when to use SSL_read;
7148 but for applications that know in advance when to expect data, it
7149 just makes things more complicated.)
7152 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7156 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7157 work better on such systems.
7158 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7160 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7161 Update PKCS12_parse() so it copies the friendlyName and the
7162 keyid to the certificates aux info.
7165 *) Fix bug in PKCS7_verify() which caused an infinite loop
7166 if there was more than one signature.
7167 [Sven Uszpelkat <su@celocom.de>]
7169 *) Major change in util/mkdef.pl to include extra information
7170 about each symbol, as well as presentig variables as well
7171 as functions. This change means that there's n more need
7172 to rebuild the .num files when some algorithms are excluded.
7175 *) Allow the verify time to be set by an application,
7176 rather than always using the current time.
7179 *) Phase 2 verify code reorganisation. The certificate
7180 verify code now looks up an issuer certificate by a
7181 number of criteria: subject name, authority key id
7182 and key usage. It also verifies self signed certificates
7183 by the same criteria. The main comparison function is
7184 X509_check_issued() which performs these checks.
7186 Lot of changes were necessary in order to support this
7187 without completely rewriting the lookup code.
7189 Authority and subject key identifier are now cached.
7191 The LHASH 'certs' is X509_STORE has now been replaced
7192 by a STACK_OF(X509_OBJECT). This is mainly because an
7193 LHASH can't store or retrieve multiple objects with
7194 the same hash value.
7196 As a result various functions (which were all internal
7197 use only) have changed to handle the new X509_STORE
7198 structure. This will break anything that messed round
7199 with X509_STORE internally.
7201 The functions X509_STORE_add_cert() now checks for an
7202 exact match, rather than just subject name.
7204 The X509_STORE API doesn't directly support the retrieval
7205 of multiple certificates matching a given criteria, however
7206 this can be worked round by performing a lookup first
7207 (which will fill the cache with candidate certificates)
7208 and then examining the cache for matches. This is probably
7209 the best we can do without throwing out X509_LOOKUP
7210 entirely (maybe later...).
7212 The X509_VERIFY_CTX structure has been enhanced considerably.
7214 All certificate lookup operations now go via a get_issuer()
7215 callback. Although this currently uses an X509_STORE it
7216 can be replaced by custom lookups. This is a simple way
7217 to bypass the X509_STORE hackery necessary to make this
7218 work and makes it possible to use more efficient techniques
7219 in future. A very simple version which uses a simple
7220 STACK for its trusted certificate store is also provided
7221 using X509_STORE_CTX_trusted_stack().
7223 The verify_cb() and verify() callbacks now have equivalents
7224 in the X509_STORE_CTX structure.
7226 X509_STORE_CTX also has a 'flags' field which can be used
7227 to customise the verify behaviour.
7230 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7231 excludes S/MIME capabilities.
7234 *) When a certificate request is read in keep a copy of the
7235 original encoding of the signed data and use it when outputing
7236 again. Signatures then use the original encoding rather than
7237 a decoded, encoded version which may cause problems if the
7238 request is improperly encoded.
7241 *) For consistency with other BIO_puts implementations, call
7242 buffer_write(b, ...) directly in buffer_puts instead of calling
7245 In BIO_puts, increment b->num_write as in BIO_write.
7246 [Peter.Sylvester@EdelWeb.fr]
7248 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7249 BN_zero, we may not return a BIGNUM with an array consisting of
7253 *) Avoid calling abort() from within the library when problems are
7254 detected, except if preprocessor symbols have been defined
7255 (such as REF_CHECK, BN_DEBUG etc.).
7258 *) New openssl application 'rsautl'. This utility can be
7259 used for low level RSA operations. DER public key
7260 BIO/fp routines also added.
7263 *) New Configure entry and patches for compiling on QNX 4.
7264 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7266 *) A demo state-machine implementation was sponsored by
7267 Nuron (http://www.nuron.com/) and is now available in
7268 demos/state_machine.
7271 *) New options added to the 'dgst' utility for signature
7272 generation and verification.
7275 *) Unrecognized PKCS#7 content types are now handled via a
7276 catch all ASN1_TYPE structure. This allows unsupported
7277 types to be stored as a "blob" and an application can
7278 encode and decode it manually.
7281 *) Fix various signed/unsigned issues to make a_strex.c
7283 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7285 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7286 length if passed a buffer. ASN1_INTEGER_to_BN failed
7287 if passed a NULL BN and its argument was negative.
7288 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7290 *) Modification to PKCS#7 encoding routines to output definite
7291 length encoding. Since currently the whole structures are in
7292 memory there's not real point in using indefinite length
7293 constructed encoding. However if OpenSSL is compiled with
7294 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7297 *) Added BIO_vprintf() and BIO_vsnprintf().
7300 *) Added more prefixes to parse for in the the strings written
7301 through a logging bio, to cover all the levels that are available
7302 through syslog. The prefixes are now:
7304 PANIC, EMERG, EMR => LOG_EMERG
7305 ALERT, ALR => LOG_ALERT
7306 CRIT, CRI => LOG_CRIT
7307 ERROR, ERR => LOG_ERR
7308 WARNING, WARN, WAR => LOG_WARNING
7309 NOTICE, NOTE, NOT => LOG_NOTICE
7310 INFO, INF => LOG_INFO
7311 DEBUG, DBG => LOG_DEBUG
7313 and as before, if none of those prefixes are present at the
7314 beginning of the string, LOG_ERR is chosen.
7316 On Win32, the LOG_* levels are mapped according to this:
7318 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7319 LOG_WARNING => EVENTLOG_WARNING_TYPE
7320 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7324 *) Made it possible to reconfigure with just the configuration
7325 argument "reconf" or "reconfigure". The command line arguments
7326 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7327 and are retrieved from there when reconfiguring.
7331 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7333 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7336 *) The obj_dat.pl script was messing up the sorting of object
7337 names. The reason was that it compared the quoted version
7338 of strings as a result "OCSP" > "OCSP Signing" because
7339 " > SPACE. Changed script to store unquoted versions of
7340 names and add quotes on output. It was also omitting some
7341 names from the lookup table if they were given a default
7342 value (that is if SN is missing it is given the same
7343 value as LN and vice versa), these are now added on the
7344 grounds that if an object has a name we should be able to
7345 look it up. Finally added warning output when duplicate
7346 short or long names are found.
7349 *) Changes needed for Tandem NSK.
7350 [Scott Uroff <scott@xypro.com>]
7352 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7353 RSA_padding_check_SSLv23(), special padding was never detected
7354 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7355 version rollback attacks was not effective.
7357 In s23_clnt.c, don't use special rollback-attack detection padding
7358 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7359 client; similarly, in s23_srvr.c, don't do the rollback check if
7360 SSL 2.0 is the only protocol enabled in the server.
7363 *) Make it possible to get hexdumps of unprintable data with 'openssl
7364 asn1parse'. By implication, the functions ASN1_parse_dump() and
7365 BIO_dump_indent() are added.
7368 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7369 these print out strings and name structures based on various
7370 flags including RFC2253 support and proper handling of
7371 multibyte characters. Added options to the 'x509' utility
7372 to allow the various flags to be set.
7375 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7376 Also change the functions X509_cmp_current_time() and
7377 X509_gmtime_adj() work with an ASN1_TIME structure,
7378 this will enable certificates using GeneralizedTime in validity
7379 dates to be checked.
7382 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7383 negative public key encodings) on by default,
7384 NO_NEG_PUBKEY_BUG can be set to disable it.
7387 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7388 content octets. An i2c_ASN1_OBJECT is unnecessary because
7389 the encoding can be trivially obtained from the structure.
7392 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7393 not read locks (CRYPTO_r_[un]lock).
7396 *) A first attempt at creating official support for shared
7397 libraries through configuration. I've kept it so the
7398 default is static libraries only, and the OpenSSL programs
7399 are always statically linked for now, but there are
7400 preparations for dynamic linking in place.
7401 This has been tested on Linux and Tru64.
7404 *) Randomness polling function for Win9x, as described in:
7405 Peter Gutmann, Software Generation of Practically Strong
7409 *) Fix so PRNG is seeded in req if using an already existing
7413 *) New options to smime application. -inform and -outform
7414 allow alternative formats for the S/MIME message including
7415 PEM and DER. The -content option allows the content to be
7416 specified separately. This should allow things like Netscape
7417 form signing output easier to verify.
7420 *) Fix the ASN1 encoding of tags using the 'long form'.
7423 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7424 STRING types. These convert content octets to and from the
7425 underlying type. The actual tag and length octets are
7426 already assumed to have been read in and checked. These
7427 are needed because all other string types have virtually
7428 identical handling apart from the tag. By having versions
7429 of the ASN1 functions that just operate on content octets
7430 IMPLICIT tagging can be handled properly. It also allows
7431 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7432 and ASN1_INTEGER are identical apart from the tag.
7435 *) Change the handling of OID objects as follows:
7437 - New object identifiers are inserted in objects.txt, following
7438 the syntax given in objects.README.
7439 - objects.pl is used to process obj_mac.num and create a new
7441 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7444 This is currently kind of a hack, and the perl code in objects.pl
7445 isn't very elegant, but it works as I intended. The simplest way
7446 to check that it worked correctly is to look in obj_dat.h and
7447 check the array nid_objs and make sure the objects haven't moved
7448 around (this is important!). Additions are OK, as well as
7449 consistent name changes.
7452 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7455 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7456 The given file adds to whatever has already been seeded into the
7457 random pool through the RANDFILE configuration file option or
7458 environment variable, or the default random state file.
7461 *) mkstack.pl now sorts each macro group into lexical order.
7462 Previously the output order depended on the order the files
7463 appeared in the directory, resulting in needless rewriting
7467 *) Patches to make OpenSSL compile under Win32 again. Mostly
7468 work arounds for the VC++ problem that it treats func() as
7469 func(void). Also stripped out the parts of mkdef.pl that
7470 added extra typesafe functions: these no longer exist.
7473 *) Reorganisation of the stack code. The macros are now all
7474 collected in safestack.h . Each macro is defined in terms of
7475 a "stack macro" of the form SKM_<name>(type, a, b). The
7476 DEBUG_SAFESTACK is now handled in terms of function casts,
7477 this has the advantage of retaining type safety without the
7478 use of additional functions. If DEBUG_SAFESTACK is not defined
7479 then the non typesafe macros are used instead. Also modified the
7480 mkstack.pl script to handle the new form. Needs testing to see
7481 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7482 the default if no major problems. Similar behaviour for ASN1_SET_OF
7483 and PKCS12_STACK_OF.
7486 *) When some versions of IIS use the 'NET' form of private key the
7487 key derivation algorithm is different. Normally MD5(password) is
7488 used as a 128 bit RC4 key. In the modified case
7489 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7490 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7491 as the old Netscape_RSA functions except they have an additional
7492 'sgckey' parameter which uses the modified algorithm. Also added
7493 an -sgckey command line option to the rsa utility. Thanks to
7494 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7495 algorithm to openssl-dev.
7498 *) The evp_local.h macros were using 'c.##kname' which resulted in
7499 invalid expansion on some systems (SCO 5.0.5 for example).
7500 Corrected to 'c.kname'.
7501 [Phillip Porch <root@theporch.com>]
7503 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7504 a STACK of email addresses from a certificate or request, these look
7505 in the subject name and the subject alternative name extensions and
7506 omit any duplicate addresses.
7509 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7510 This makes DSA verification about 2 % faster.
7513 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7514 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7515 plus overhead for 1024 bit moduli).
7516 This makes exponentiations about 0.5 % faster for 1024 bit
7517 exponents (as measured by "openssl speed rsa2048").
7520 *) Rename memory handling macros to avoid conflicts with other
7522 Malloc => OPENSSL_malloc
7523 Malloc_locked => OPENSSL_malloc_locked
7524 Realloc => OPENSSL_realloc
7525 Free => OPENSSL_free
7528 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7529 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7532 *) CygWin32 support.
7533 [John Jarvie <jjarvie@newsguy.com>]
7535 *) The type-safe stack code has been rejigged. It is now only compiled
7536 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7537 by default all type-specific stack functions are "#define"d back to
7538 standard stack functions. This results in more streamlined output
7539 but retains the type-safety checking possibilities of the original
7543 *) The STACK code has been cleaned up, and certain type declarations
7544 that didn't make a lot of sense have been brought in line. This has
7545 also involved a cleanup of sorts in safestack.h to more correctly
7546 map type-safe stack functions onto their plain stack counterparts.
7547 This work has also resulted in a variety of "const"ifications of
7548 lots of the code, especially "_cmp" operations which should normally
7549 be prototyped with "const" parameters anyway.
7552 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7553 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7554 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7555 where all of 'md' is used each time the PRNG is used, but 'state'
7556 is used only indexed by a cyclic counter. As entropy may not be
7557 well distributed from the beginning, 'md' is important as a
7558 chaining variable. However, the output function chains only half
7559 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7560 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7561 in all of 'state' being rewritten, with the new values depending
7562 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7565 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7566 the handshake is continued after ssl_verify_cert_chain();
7567 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7568 can lead to 'unexplainable' connection aborts later.
7569 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7571 *) Major EVP API cipher revision.
7572 Add hooks for extra EVP features. This allows various cipher
7573 parameters to be set in the EVP interface. Support added for variable
7574 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7575 setting of RC2 and RC5 parameters.
7577 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7580 Remove lots of duplicated code from the EVP library. For example *every*
7581 cipher init() function handles the 'iv' in the same way according to the
7582 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7583 for CFB and OFB modes they zero ctx->num.
7585 New functionality allows removal of S/MIME code RC2 hack.
7587 Most of the routines have the same form and so can be declared in terms
7590 By shifting this to the top level EVP_CipherInit() it can be removed from
7591 all individual ciphers. If the cipher wants to handle IVs or keys
7592 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7595 Change lots of functions like EVP_EncryptUpdate() to now return a
7596 value: although software versions of the algorithms cannot fail
7597 any installed hardware versions can.
7600 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7601 this option is set, tolerate broken clients that send the negotiated
7602 protocol version number instead of the requested protocol version
7606 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7607 i.e. non-zero for export ciphersuites, zero otherwise.
7608 Previous versions had this flag inverted, inconsistent with
7609 rsa_tmp_cb (..._TMP_RSA_CB).
7610 [Bodo Moeller; problem reported by Amit Chopra]
7612 *) Add missing DSA library text string. Work around for some IIS
7613 key files with invalid SEQUENCE encoding.
7616 *) Add a document (doc/standards.txt) that list all kinds of standards
7617 and so on that are implemented in OpenSSL.
7620 *) Enhance c_rehash script. Old version would mishandle certificates
7621 with the same subject name hash and wouldn't handle CRLs at all.
7622 Added -fingerprint option to crl utility, to support new c_rehash
7626 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7629 *) Fix for SSL server purpose checking. Server checking was
7630 rejecting certificates which had extended key usage present
7631 but no ssl client purpose.
7632 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7634 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7635 is a little unclear about how a blank password is handled.
7636 Since the password in encoded as a BMPString with terminating
7637 double NULL a zero length password would end up as just the
7638 double NULL. However no password at all is different and is
7639 handled differently in the PKCS#12 key generation code. NS
7640 treats a blank password as zero length. MSIE treats it as no
7641 password on export: but it will try both on import. We now do
7642 the same: PKCS12_parse() tries zero length and no password if
7643 the password is set to "" or NULL (NULL is now a valid password:
7644 it wasn't before) as does the pkcs12 application.
7647 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7648 perror when PEM_read_bio_X509_REQ fails, the error message must
7649 be obtained from the error queue.
7652 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7653 it in ERR_remove_state if appropriate, and change ERR_get_state
7654 accordingly to avoid race conditions (this is necessary because
7655 thread_hash is no longer constant once set).
7658 *) Bugfix for linux-elf makefile.one.
7661 *) RSA_get_default_method() will now cause a default
7662 RSA_METHOD to be chosen if one doesn't exist already.
7663 Previously this was only set during a call to RSA_new()
7664 or RSA_new_method(NULL) meaning it was possible for
7665 RSA_get_default_method() to return NULL.
7668 *) Added native name translation to the existing DSO code
7669 that will convert (if the flag to do so is set) filenames
7670 that are sufficiently small and have no path information
7671 into a canonical native form. Eg. "blah" converted to
7672 "libblah.so" or "blah.dll" etc.
7675 *) New function ERR_error_string_n(e, buf, len) which is like
7676 ERR_error_string(e, buf), but writes at most 'len' bytes
7677 including the 0 terminator. For ERR_error_string_n, 'buf'
7679 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7681 *) CONF library reworked to become more general. A new CONF
7682 configuration file reader "class" is implemented as well as a
7683 new functions (NCONF_*, for "New CONF") to handle it. The now
7684 old CONF_* functions are still there, but are reimplemented to
7685 work in terms of the new functions. Also, a set of functions
7686 to handle the internal storage of the configuration data is
7687 provided to make it easier to write new configuration file
7688 reader "classes" (I can definitely see something reading a
7689 configuration file in XML format, for example), called _CONF_*,
7690 or "the configuration storage API"...
7692 The new configuration file reading functions are:
7694 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7695 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7697 NCONF_default, NCONF_WIN32
7699 NCONF_dump_fp, NCONF_dump_bio
7701 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7702 NCONF_new creates a new CONF object. This works in the same way
7703 as other interfaces in OpenSSL, like the BIO interface.
7704 NCONF_dump_* dump the internal storage of the configuration file,
7705 which is useful for debugging. All other functions take the same
7706 arguments as the old CONF_* functions wth the exception of the
7707 first that must be a `CONF *' instead of a `LHASH *'.
7709 To make it easer to use the new classes with the old CONF_* functions,
7710 the function CONF_set_default_method is provided.
7713 *) Add '-tls1' option to 'openssl ciphers', which was already
7714 mentioned in the documentation but had not been implemented.
7715 (This option is not yet really useful because even the additional
7716 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7719 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7720 OpenSSL-based applications) load shared libraries and bind to
7721 them in a portable way.
7722 [Geoff Thorpe, with contributions from Richard Levitte]
7724 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7726 *) Make sure _lrotl and _lrotr are only used with MSVC.
7728 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7729 (the default implementation of RAND_status).
7731 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7732 to '-clrext' (= clear extensions), as intended and documented.
7733 [Bodo Moeller; inconsistency pointed out by Michael Attili
7736 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7737 was larger than the MD block size.
7738 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7740 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7741 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7742 using the passed key: if the passed key was a private key the result
7743 of X509_print(), for example, would be to print out all the private key
7747 *) des_quad_cksum() byte order bug fix.
7748 [Ulf Möller, using the problem description in krb4-0.9.7, where
7749 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7751 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7753 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7755 *) For easily testing in shell scripts whether some command
7756 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7757 returns with exit code 0 iff no command of the given name is available.
7758 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7759 the output goes to stdout and nothing is printed to stderr.
7760 Additional arguments are always ignored.
7762 Since for each cipher there is a command of the same name,
7763 the 'no-cipher' compilation switches can be tested this way.
7765 ('openssl no-XXX' is not able to detect pseudo-commands such
7766 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7769 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7772 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7773 is set; it will be thrown away anyway because each handshake creates
7775 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7776 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7777 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7778 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7781 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7782 'Q' and 'R' lose their special meanings (quit/renegotiate).
7783 This is part of what -quiet does; unlike -quiet, -ign_eof
7784 does not suppress any output.
7787 *) Add compatibility options to the purpose and trust code. The
7788 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7789 accepts a certificate or CA, this was the previous behaviour,
7790 with all the associated security issues.
7792 X509_TRUST_COMPAT is the old trust behaviour: only and
7793 automatically trust self signed roots in certificate store. A
7794 new trust setting X509_TRUST_DEFAULT is used to specify that
7795 a purpose has no associated trust setting and it should instead
7796 use the value in the default purpose.
7799 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7800 and fix a memory leak.
7803 *) In util/mkerr.pl (which implements 'make errors'), preserve
7804 reason strings from the previous version of the .c file, as
7805 the default to have only downcase letters (and digits) in
7806 automatically generated reasons codes is not always appropriate.
7809 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7810 using strerror. Previously, ERR_reason_error_string() returned
7811 library names as reason strings for SYSerr; but SYSerr is a special
7812 case where small numbers are errno values, not library numbers.
7815 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7816 converts DSA parameters into DH parameters. (When creating parameters,
7817 DSA_generate_parameters is used.)
7820 *) Include 'length' (recommended exponent length) in C code generated
7821 by 'openssl dhparam -C'.
7824 *) The second argument to set_label in perlasm was already being used
7825 so couldn't be used as a "file scope" flag. Moved to third argument
7829 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7830 instead of RAND_bytes for encryption IVs and salts.
7833 *) Include RAND_status() into RAND_METHOD instead of implementing
7834 it only for md_rand.c Otherwise replacing the PRNG by calling
7835 RAND_set_rand_method would be impossible.
7838 *) Don't let DSA_generate_key() enter an infinite loop if the random
7839 number generation fails.
7842 *) New 'rand' application for creating pseudo-random output.
7845 *) Added configuration support for Linux/IA64
7846 [Rolf Haberrecker <rolf@suse.de>]
7848 *) Assembler module support for Mingw32.
7851 *) Shared library support for HPUX (in shlib/).
7852 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7854 *) Shared library support for Solaris gcc.
7855 [Lutz Behnke <behnke@trustcenter.de>]
7857 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7859 *) PKCS7_encrypt() was adding text MIME headers twice because they
7860 were added manually and by SMIME_crlf_copy().
7863 *) In bntest.c don't call BN_rand with zero bits argument.
7864 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7866 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7867 case was implemented. This caused BN_div_recp() to fail occasionally.
7870 *) Add an optional second argument to the set_label() in the perl
7871 assembly language builder. If this argument exists and is set
7872 to 1 it signals that the assembler should use a symbol whose
7873 scope is the entire file, not just the current function. This
7874 is needed with MASM which uses the format label:: for this scope.
7875 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7877 *) Change the ASN1 types so they are typedefs by default. Before
7878 almost all types were #define'd to ASN1_STRING which was causing
7879 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7883 *) Change names of new functions to the new get1/get0 naming
7884 convention: After 'get1', the caller owns a reference count
7885 and has to call ..._free; 'get0' returns a pointer to some
7886 data structure without incrementing reference counters.
7887 (Some of the existing 'get' functions increment a reference
7888 counter, some don't.)
7889 Similarly, 'set1' and 'add1' functions increase reference
7890 counters or duplicate objects.
7893 *) Allow for the possibility of temp RSA key generation failure:
7894 the code used to assume it always worked and crashed on failure.
7897 *) Fix potential buffer overrun problem in BIO_printf().
7898 [Ulf Möller, using public domain code by Patrick Powell; problem
7899 pointed out by David Sacerdote <das33@cornell.edu>]
7901 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7902 RAND_egd() and RAND_status(). In the command line application,
7903 the EGD socket can be specified like a seed file using RANDFILE
7907 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7908 Some CAs (e.g. Verisign) distribute certificates in this form.
7911 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7912 list to exclude them. This means that no special compilation option
7913 is needed to use anonymous DH: it just needs to be included in the
7917 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7918 EVP_MD_type. The old functionality is available in a new macro called
7919 EVP_MD_md(). Change code that uses it and update docs.
7922 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7923 where the 'void *' argument is replaced by a function pointer argument.
7924 Previously 'void *' was abused to point to functions, which works on
7925 many platforms, but is not correct. As these functions are usually
7926 called by macros defined in OpenSSL header files, most source code
7927 should work without changes.
7930 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7931 sections with information on -D... compiler switches used for
7932 compiling the library so that applications can see them. To enable
7933 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7934 must be defined. E.g.,
7935 #define OPENSSL_ALGORITHM_DEFINES
7936 #include <openssl/opensslconf.h>
7937 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7938 [Richard Levitte, Ulf and Bodo Möller]
7940 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7944 *) Change the 'other' type in certificate aux info to a STACK_OF
7945 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7946 the required ASN1 format: arbitrary types determined by an OID.
7949 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7950 argument to 'req'. This is not because the function is newer or
7951 better than others it just uses the work 'NEW' in the certificate
7952 request header lines. Some software needs this.
7955 *) Reorganise password command line arguments: now passwords can be
7956 obtained from various sources. Delete the PEM_cb function and make
7957 it the default behaviour: i.e. if the callback is NULL and the
7958 usrdata argument is not NULL interpret it as a null terminated pass
7959 phrase. If usrdata and the callback are NULL then the pass phrase
7960 is prompted for as usual.
7963 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7964 the support is automatically enabled. The resulting binaries will
7965 autodetect the card and use it if present.
7966 [Ben Laurie and Compaq Inc.]
7968 *) Work around for Netscape hang bug. This sends certificate request
7969 and server done in one record. Since this is perfectly legal in the
7970 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7971 the bugs/SSLv3 entry for more info.
7974 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7977 *) Add -rand argument to smime and pkcs12 applications and read/write
7981 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7984 *) Add command line password options to the remaining applications.
7987 *) Bug fix for BN_div_recp() for numerators with an even number of
7991 *) More tests in bntest.c, and changed test_bn output.
7994 *) ./config recognizes MacOS X now.
7997 *) Bug fix for BN_div() when the first words of num and divsor are
7998 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8001 *) Add support for various broken PKCS#8 formats, and command line
8002 options to produce them.
8005 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8006 get temporary BIGNUMs from a BN_CTX.
8009 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8013 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8014 include a #define from the old name to the new. The original intent
8015 was that statically linked binaries could for example just call
8016 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8017 link with digests. This never worked becayse SSLeay_add_all_digests()
8018 and SSLeay_add_all_ciphers() were in the same source file so calling
8019 one would link with the other. They are now in separate source files.
8022 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8025 *) Use a less unusual form of the Miller-Rabin primality test (it used
8026 a binary algorithm for exponentiation integrated into the Miller-Rabin
8027 loop, our standard modexp algorithms are faster).
8030 *) Support for the EBCDIC character set completed.
8031 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8033 *) Source code cleanups: use const where appropriate, eliminate casts,
8034 use void * instead of char * in lhash.
8037 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8038 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8039 this the server could overwrite ephemeral keys that the client
8043 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8044 using 50 iterations of the Rabin-Miller test.
8046 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8047 iterations of the Rabin-Miller test as required by the appendix
8048 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8049 As BN_is_prime_fasttest includes trial division, DSA parameter
8050 generation becomes much faster.
8052 This implies a change for the callback functions in DSA_is_prime
8053 and DSA_generate_parameters: The callback function is called once
8054 for each positive witness in the Rabin-Miller test, not just
8055 occasionally in the inner loop; and the parameters to the
8056 callback function now provide an iteration count for the outer
8057 loop rather than for the current invocation of the inner loop.
8058 DSA_generate_parameters additionally can call the callback
8059 function with an 'iteration count' of -1, meaning that a
8060 candidate has passed the trial division test (when q is generated
8061 from an application-provided seed, trial division is skipped).
8064 *) New function BN_is_prime_fasttest that optionally does trial
8065 division before starting the Rabin-Miller test and has
8066 an additional BN_CTX * argument (whereas BN_is_prime always
8067 has to allocate at least one BN_CTX).
8068 'callback(1, -1, cb_arg)' is called when a number has passed the
8069 trial division stage.
8072 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8076 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8079 *) New function BN_pseudo_rand().
8082 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8083 bignum version of BN_from_montgomery() with the working code from
8084 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8088 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8089 made it impossible to use the same SSL_SESSION data structure in
8090 SSL2 clients in multiple threads.
8093 *) The return value of RAND_load_file() no longer counts bytes obtained
8094 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8095 to seed the PRNG (previously an explicit byte count was required).
8096 [Ulf Möller, Bodo Möller]
8098 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8099 used (char *) instead of (void *) and had casts all over the place.
8102 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8105 *) Retain source code compatibility for BN_prime_checks macro:
8106 BN_is_prime(..., BN_prime_checks, ...) now uses
8107 BN_prime_checks_for_size to determine the appropriate number of
8108 Rabin-Miller iterations.
8111 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8112 DH_CHECK_P_NOT_SAFE_PRIME.
8113 (Check if this is true? OpenPGP calls them "strong".)
8116 *) Merge the functionality of "dh" and "gendh" programs into a new program
8117 "dhparam". The old programs are retained for now but will handle DH keys
8118 (instead of parameters) in future.
8121 *) Make the ciphers, s_server and s_client programs check the return values
8122 when a new cipher list is set.
8125 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8126 ciphers. Before when the 56bit ciphers were enabled the sorting was
8129 The syntax for the cipher sorting has been extended to support sorting by
8130 cipher-strength (using the strength_bits hard coded in the tables).
8131 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8133 Fix a bug in the cipher-command parser: when supplying a cipher command
8134 string with an "undefined" symbol (neither command nor alphanumeric
8135 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8136 an error is flagged.
8138 Due to the strength-sorting extension, the code of the
8139 ssl_create_cipher_list() function was completely rearranged. I hope that
8140 the readability was also increased :-)
8141 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8143 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8144 for the first serial number and places 2 in the serial number file. This
8145 avoids problems when the root CA is created with serial number zero and
8146 the first user certificate has the same issuer name and serial number
8150 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8151 the new code. Add documentation for this stuff.
8154 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8155 X509_*() to X509at_*() on the grounds that they don't handle X509
8156 structures and behave in an analagous way to the X509v3 functions:
8157 they shouldn't be called directly but wrapper functions should be used
8160 So we also now have some wrapper functions that call the X509at functions
8161 when passed certificate requests. (TO DO: similar things can be done with
8162 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8163 things. Some of these need some d2i or i2d and print functionality
8164 because they handle more complex structures.)
8167 *) Add missing #ifndefs that caused missing symbols when building libssl
8168 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8169 NO_RSA in ssl/s2*.c.
8170 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8172 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8173 has a return value which indicates the quality of the random data
8174 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8175 error queue. New function RAND_pseudo_bytes() generates output that is
8176 guaranteed to be unique but not unpredictable. RAND_add is like
8177 RAND_seed, but takes an extra argument for an entropy estimate
8178 (RAND_seed always assumes full entropy).
8181 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8182 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8183 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8184 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8185 false-positive rate of at most 2^-80 for random input.
8188 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8191 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8192 in the 0.9.5 release), this returns the chain
8193 from an X509_CTX structure with a dup of the stack and all
8194 the X509 reference counts upped: so the stack will exist
8195 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8198 Also make SSL_SESSION_print() print out the verify return
8202 *) Add manpage for the pkcs12 command. Also change the default
8203 behaviour so MAC iteration counts are used unless the new
8204 -nomaciter option is used. This improves file security and
8205 only older versions of MSIE (4.0 for example) need it.
8208 *) Honor the no-xxx Configure options when creating .DEF files.
8211 *) Add PKCS#10 attributes to field table: challengePassword,
8212 unstructuredName and unstructuredAddress. These are taken from
8213 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8214 international characters are used.
8216 More changes to X509_ATTRIBUTE code: allow the setting of types
8217 based on strings. Remove the 'loc' parameter when adding
8218 attributes because these will be a SET OF encoding which is sorted
8222 *) Initial changes to the 'req' utility to allow request generation
8223 automation. This will allow an application to just generate a template
8224 file containing all the field values and have req construct the
8227 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8228 used all over the place including certificate requests and PKCS#7
8229 structures. They are currently handled manually where necessary with
8230 some primitive wrappers for PKCS#7. The new functions behave in a
8231 manner analogous to the X509 extension functions: they allow
8232 attributes to be looked up by NID and added.
8234 Later something similar to the X509V3 code would be desirable to
8235 automatically handle the encoding, decoding and printing of the
8236 more complex types. The string types like challengePassword can
8237 be handled by the string table functions.
8239 Also modified the multi byte string table handling. Now there is
8240 a 'global mask' which masks out certain types. The table itself
8241 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8242 is useful when for example there is only one permissible type
8243 (as in countryName) and using the mask might result in no valid
8247 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8248 SSL_get_peer_finished to allow applications to obtain the latest
8249 Finished messages sent to the peer or expected from the peer,
8250 respectively. (SSL_get_peer_finished is usually the Finished message
8251 actually received from the peer, otherwise the protocol will be aborted.)
8253 As the Finished message are message digests of the complete handshake
8254 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8255 be used for external authentication procedures when the authentication
8256 provided by SSL/TLS is not desired or is not enough.
8259 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8260 the host supports BWX extension and if Compaq C is present on the
8261 $PATH. Just exploiting of the BWX extension results in 20-30%
8262 performance kick for some algorithms, e.g. DES and RC4 to mention
8263 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8267 *) Add support for MS "fast SGC". This is arguably a violation of the
8268 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8269 weak crypto and after checking the certificate is SGC a second one
8270 with strong crypto. MS SGC stops the first handshake after receiving
8271 the server certificate message and sends a second client hello. Since
8272 a server will typically do all the time consuming operations before
8273 expecting any further messages from the client (server key exchange
8274 is the most expensive) there is little difference between the two.
8276 To get OpenSSL to support MS SGC we have to permit a second client
8277 hello message after we have sent server done. In addition we have to
8278 reset the MAC if we do get this second client hello.
8281 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8282 if a DER encoded private key is RSA or DSA traditional format. Changed
8283 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8284 format DER encoded private key. Newer code should use PKCS#8 format which
8285 has the key type encoded in the ASN1 structure. Added DER private key
8286 support to pkcs8 application.
8289 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8290 ciphersuites has been selected (as required by the SSL 3/TLS 1
8291 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8292 is set, we interpret this as a request to violate the specification
8293 (the worst that can happen is a handshake failure, and 'correct'
8294 behaviour would result in a handshake failure anyway).
8297 *) In SSL_CTX_add_session, take into account that there might be multiple
8298 SSL_SESSION structures with the same session ID (e.g. when two threads
8299 concurrently obtain them from an external cache).
8300 The internal cache can handle only one SSL_SESSION with a given ID,
8301 so if there's a conflict, we now throw out the old one to achieve
8305 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8306 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8307 some routines that use cipher OIDs: some ciphers do not have OIDs
8308 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8312 *) Simplify the trust setting structure and code. Now we just have
8313 two sequences of OIDs for trusted and rejected settings. These will
8314 typically have values the same as the extended key usage extension
8315 and any application specific purposes.
8317 The trust checking code now has a default behaviour: it will just
8318 check for an object with the same NID as the passed id. Functions can
8319 be provided to override either the default behaviour or the behaviour
8320 for a given id. SSL client, server and email already have functions
8321 in place for compatibility: they check the NID and also return "trusted"
8322 if the certificate is self signed.
8325 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8326 traditional format into an EVP_PKEY structure.
8329 *) Add a password callback function PEM_cb() which either prompts for
8330 a password if usr_data is NULL or otherwise assumes it is a null
8331 terminated password. Allow passwords to be passed on command line
8332 environment or config files in a few more utilities.
8335 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8336 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8337 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8338 Update documentation.
8341 *) Support for ASN1 "NULL" type. This could be handled before by using
8342 ASN1_TYPE but there wasn't any function that would try to read a NULL
8343 and produce an error if it couldn't. For compatibility we also have
8344 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8345 don't allocate anything because they don't need to.
8348 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8350 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8352 *) Rebuild of the memory allocation routines used by OpenSSL code and
8353 possibly others as well. The purpose is to make an interface that
8354 provide hooks so anyone can build a separate set of allocation and
8355 deallocation routines to be used by OpenSSL, for example memory
8356 pool implementations, or something else, which was previously hard
8357 since Malloc(), Realloc() and Free() were defined as macros having
8358 the values malloc, realloc and free, respectively (except for Win32
8359 compilations). The same is provided for memory debugging code.
8360 OpenSSL already comes with functionality to find memory leaks, but
8361 this gives people a chance to debug other memory problems.
8363 With these changes, a new set of functions and macros have appeared:
8365 CRYPTO_set_mem_debug_functions() [F]
8366 CRYPTO_get_mem_debug_functions() [F]
8367 CRYPTO_dbg_set_options() [F]
8368 CRYPTO_dbg_get_options() [F]
8369 CRYPTO_malloc_debug_init() [M]
8371 The memory debug functions are NULL by default, unless the library
8372 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8373 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8374 gives the standard debugging functions that come with OpenSSL) or
8375 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8376 provided by the library user) must be used. When the standard
8377 debugging functions are used, CRYPTO_dbg_set_options can be used to
8378 request additional information:
8379 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8380 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8382 Also, things like CRYPTO_set_mem_functions will always give the
8383 expected result (the new set of functions is used for allocation
8384 and deallocation) at all times, regardless of platform and compiler
8387 To finish it up, some functions that were never use in any other
8388 way than through macros have a new API and new semantic:
8391 CRYPTO_dbg_realloc()
8394 All macros of value have retained their old syntax.
8395 [Richard Levitte and Bodo Moeller]
8397 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8398 ordering of SMIMECapabilities wasn't in "strength order" and there
8399 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8403 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8404 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8405 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8407 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8408 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8409 functionality to handle multipart/signed properly) and a utility
8410 called 'smime' to call all this stuff. This is based on code I
8411 originally wrote for Celo who have kindly allowed it to be
8412 included in OpenSSL.
8415 *) Add variants des_set_key_checked and des_set_key_unchecked of
8416 des_set_key (aka des_key_sched). Global variable des_check_key
8417 decides which of these is called by des_set_key; this way
8418 des_check_key behaves as it always did, but applications and
8419 the library itself, which was buggy for des_check_key == 1,
8420 have a cleaner way to pick the version they need.
8423 *) New function PKCS12_newpass() which changes the password of a
8427 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8428 dynamic mix. In both cases the ids can be used as an index into the
8429 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8430 functions so they accept a list of the field values and the
8431 application doesn't need to directly manipulate the X509_TRUST
8435 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8439 *) Modify the way the V3 extension code looks up extensions. This now
8440 works in a similar way to the object code: we have some "standard"
8441 extensions in a static table which is searched with OBJ_bsearch()
8442 and the application can add dynamic ones if needed. The file
8443 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8444 updated whenever a new extension is added to the core code and kept
8445 in ext_nid order. There is a simple program 'tabtest.c' which checks
8446 this. New extensions are not added too often so this file can readily
8447 be maintained manually.
8449 There are two big advantages in doing things this way. The extensions
8450 can be looked up immediately and no longer need to be "added" using
8451 X509V3_add_standard_extensions(): this function now does nothing.
8452 [Side note: I get *lots* of email saying the extension code doesn't
8453 work because people forget to call this function]
8454 Also no dynamic allocation is done unless new extensions are added:
8455 so if we don't add custom extensions there is no need to call
8456 X509V3_EXT_cleanup().
8459 *) Modify enc utility's salting as follows: make salting the default. Add a
8460 magic header, so unsalted files fail gracefully instead of just decrypting
8461 to garbage. This is because not salting is a big security hole, so people
8462 should be discouraged from doing it.
8465 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8466 digest to be passed on the command line but it only used this
8467 parameter when signing a certificate. Modified so all relevant
8468 operations are affected by the digest parameter including the
8469 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8470 DSA key was used because it didn't fix the digest.
8473 *) Initial certificate chain verify code. Currently tests the untrusted
8474 certificates for consistency with the verify purpose (which is set
8475 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8477 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8478 this is because it will reject chains with invalid extensions whereas
8479 every previous version of OpenSSL and SSLeay made no checks at all.
8481 Trust code: checks the root CA for the relevant trust settings. Trust
8482 settings have an initial value consistent with the verify purpose: e.g.
8483 if the verify purpose is for SSL client use it expects the CA to be
8484 trusted for SSL client use. However the default value can be changed to
8485 permit custom trust settings: one example of this would be to only trust
8486 certificates from a specific "secure" set of CAs.
8488 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8489 which should be used for version portability: especially since the
8490 verify structure is likely to change more often now.
8492 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8493 to set them. If not set then assume SSL clients will verify SSL servers
8496 Two new options to the verify program: -untrusted allows a set of
8497 untrusted certificates to be passed in and -purpose which sets the
8498 intended purpose of the certificate. If a purpose is set then the
8499 new chain verify code is used to check extension consistency.
8502 *) Support for the authority information access extension.
8505 *) Modify RSA and DSA PEM read routines to transparently handle
8506 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8507 public keys in a format compatible with certificate
8508 SubjectPublicKeyInfo structures. Unfortunately there were already
8509 functions called *_PublicKey_* which used various odd formats so
8510 these are retained for compatibility: however the DSA variants were
8511 never in a public release so they have been deleted. Changed dsa/rsa
8512 utilities to handle the new format: note no releases ever handled public
8513 keys so we should be OK.
8515 The primary motivation for this change is to avoid the same fiasco
8516 that dogs private keys: there are several incompatible private key
8517 formats some of which are standard and some OpenSSL specific and
8518 require various evil hacks to allow partial transparent handling and
8519 even then it doesn't work with DER formats. Given the option anything
8520 other than PKCS#8 should be dumped: but the other formats have to
8521 stay in the name of compatibility.
8523 With public keys and the benefit of hindsight one standard format
8524 is used which works with EVP_PKEY, RSA or DSA structures: though
8525 it clearly returns an error if you try to read the wrong kind of key.
8527 Added a -pubkey option to the 'x509' utility to output the public key.
8528 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8529 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8530 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8531 that do the same as the EVP_PKEY_assign_*() except they up the
8532 reference count of the added key (they don't "swallow" the
8536 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8537 CRLs would fail if the file contained no certificates or no CRLs:
8538 added a new function to read in both types and return the number
8539 read: this means that if none are read it will be an error. The
8540 DER versions of the certificate and CRL reader would always fail
8541 because it isn't possible to mix certificates and CRLs in DER format
8542 without choking one or the other routine. Changed this to just read
8543 a certificate: this is the best we can do. Also modified the code
8544 in apps/verify.c to take notice of return codes: it was previously
8545 attempting to read in certificates from NULL pointers and ignoring
8546 any errors: this is one reason why the cert and CRL reader seemed
8547 to work. It doesn't check return codes from the default certificate
8548 routines: these may well fail if the certificates aren't installed.
8551 *) Code to support otherName option in GeneralName.
8554 *) First update to verify code. Change the verify utility
8555 so it warns if it is passed a self signed certificate:
8556 for consistency with the normal behaviour. X509_verify
8557 has been modified to it will now verify a self signed
8558 certificate if *exactly* the same certificate appears
8559 in the store: it was previously impossible to trust a
8560 single self signed certificate. This means that:
8561 openssl verify ss.pem
8562 now gives a warning about a self signed certificate but
8563 openssl verify -CAfile ss.pem ss.pem
8567 *) For servers, store verify_result in SSL_SESSION data structure
8568 (and add it to external session representation).
8569 This is needed when client certificate verifications fails,
8570 but an application-provided verification callback (set by
8571 SSL_CTX_set_cert_verify_callback) allows accepting the session
8572 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8573 but returns 1): When the session is reused, we have to set
8574 ssl->verify_result to the appropriate error code to avoid
8576 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8578 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8579 case in PKCS7_dataInit() where the signed PKCS7 structure
8580 didn't contain any existing data because it was being created.
8581 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8583 *) Add a salt to the key derivation routines in enc.c. This
8584 forms the first 8 bytes of the encrypted file. Also add a
8585 -S option to allow a salt to be input on the command line.
8588 *) New function X509_cmp(). Oddly enough there wasn't a function
8589 to compare two certificates. We do this by working out the SHA1
8590 hash and comparing that. X509_cmp() will be needed by the trust
8594 *) SSL_get1_session() is like SSL_get_session(), but increments
8595 the reference count in the SSL_SESSION returned.
8596 [Geoff Thorpe <geoff@eu.c2.net>]
8598 *) Fix for 'req': it was adding a null to request attributes.
8599 Also change the X509_LOOKUP and X509_INFO code to handle
8600 certificate auxiliary information.
8603 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8607 *) Add the possibility to add extra information to the memory leak
8608 detecting output, to form tracebacks, showing from where each
8609 allocation was originated: CRYPTO_push_info("constant string") adds
8610 the string plus current file name and line number to a per-thread
8611 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8612 is like calling CYRPTO_pop_info() until the stack is empty.
8613 Also updated memory leak detection code to be multi-thread-safe.
8616 *) Add options -text and -noout to pkcs7 utility and delete the
8617 encryption options which never did anything. Update docs.
8620 *) Add options to some of the utilities to allow the pass phrase
8621 to be included on either the command line (not recommended on
8622 OSes like Unix) or read from the environment. Update the
8623 manpages and fix a few bugs.
8626 *) Add a few manpages for some of the openssl commands.
8629 *) Fix the -revoke option in ca. It was freeing up memory twice,
8630 leaking and not finding already revoked certificates.
8633 *) Extensive changes to support certificate auxiliary information.
8634 This involves the use of X509_CERT_AUX structure and X509_AUX
8635 functions. An X509_AUX function such as PEM_read_X509_AUX()
8636 can still read in a certificate file in the usual way but it
8637 will also read in any additional "auxiliary information". By
8638 doing things this way a fair degree of compatibility can be
8639 retained: existing certificates can have this information added
8640 using the new 'x509' options.
8642 Current auxiliary information includes an "alias" and some trust
8643 settings. The trust settings will ultimately be used in enhanced
8644 certificate chain verification routines: currently a certificate
8645 can only be trusted if it is self signed and then it is trusted
8649 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8650 The problem was that one of the replacement routines had not been working
8651 since SSLeay releases. For now the offending routine has been replaced
8652 with non-optimised assembler. Even so, this now gives around 95%
8653 performance improvement for 1024 bit RSA signs.
8656 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8657 handling. Most clients have the effective key size in bits equal to
8658 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8659 A few however don't do this and instead use the size of the decrypted key
8660 to determine the RC2 key length and the AlgorithmIdentifier to determine
8661 the effective key length. In this case the effective key length can still
8662 be 40 bits but the key length can be 168 bits for example. This is fixed
8663 by manually forcing an RC2 key into the EVP_PKEY structure because the
8664 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8665 the key length and effective key length are equal.
8668 *) Add a bunch of functions that should simplify the creation of
8669 X509_NAME structures. Now you should be able to do:
8670 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8671 and have it automatically work out the correct field type and fill in
8672 the structures. The more adventurous can try:
8673 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8674 and it will (hopefully) work out the correct multibyte encoding.
8677 *) Change the 'req' utility to use the new field handling and multibyte
8678 copy routines. Before the DN field creation was handled in an ad hoc
8679 way in req, ca, and x509 which was rather broken and didn't support
8680 BMPStrings or UTF8Strings. Since some software doesn't implement
8681 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8682 using the dirstring_type option. See the new comment in the default
8683 openssl.cnf for more info.
8686 *) Make crypto/rand/md_rand.c more robust:
8687 - Assure unique random numbers after fork().
8688 - Make sure that concurrent threads access the global counter and
8689 md serializably so that we never lose entropy in them
8690 or use exactly the same state in multiple threads.
8691 Access to the large state is not always serializable because
8692 the additional locking could be a performance killer, and
8693 md should be large enough anyway.
8696 *) New file apps/app_rand.c with commonly needed functionality
8697 for handling the random seed file.
8699 Use the random seed file in some applications that previously did not:
8701 dsaparam -genkey (which also ignored its '-rand' option),
8704 x509 (when signing).
8705 Except on systems with /dev/urandom, it is crucial to have a random
8706 seed file at least for key creation, DSA signing, and for DH exchanges;
8707 for RSA signatures we could do without one.
8709 gendh and gendsa (unlike genrsa) used to read only the first byte
8710 of each file listed in the '-rand' option. The function as previously
8711 found in genrsa is now in app_rand.c and is used by all programs
8712 that support '-rand'.
8715 *) In RAND_write_file, use mode 0600 for creating files;
8716 don't just chmod when it may be too late.
8719 *) Report an error from X509_STORE_load_locations
8720 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8723 *) New function ASN1_mbstring_copy() this copies a string in either
8724 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8725 into an ASN1_STRING type. A mask of permissible types is passed
8726 and it chooses the "minimal" type to use or an error if not type
8730 *) Add function equivalents to the various macros in asn1.h. The old
8731 macros are retained with an M_ prefix. Code inside the library can
8732 use the M_ macros. External code (including the openssl utility)
8733 should *NOT* in order to be "shared library friendly".
8736 *) Add various functions that can check a certificate's extensions
8737 to see if it usable for various purposes such as SSL client,
8738 server or S/MIME and CAs of these types. This is currently
8739 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8740 verification. Also added a -purpose flag to x509 utility to
8741 print out all the purposes.
8744 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8748 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8749 for, obtain and decode and extension and obtain its critical flag.
8750 This allows all the necessary extension code to be handled in a
8751 single function call.
8754 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8755 platforms. See crypto/rc4/rc4_enc.c for further details.
8758 *) New -noout option to asn1parse. This causes no output to be produced
8759 its main use is when combined with -strparse and -out to extract data
8760 from a file (which may not be in ASN.1 format).
8763 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8764 when producing the local key id.
8765 [Richard Levitte <levitte@stacken.kth.se>]
8767 *) New option -dhparam in s_server. This allows a DH parameter file to be
8768 stated explicitly. If it is not stated then it tries the first server
8769 certificate file. The previous behaviour hard coded the filename
8773 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8774 a public key to be input or output. For example:
8775 openssl rsa -in key.pem -pubout -out pubkey.pem
8776 Also added necessary DSA public key functions to handle this.
8779 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8780 in the message. This was handled by allowing
8781 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8782 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8784 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8785 to the end of the strings whereas this didn't. This would cause problems
8786 if strings read with d2i_ASN1_bytes() were later modified.
8787 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8789 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8790 data and it contains EOF it will end up returning an error. This is
8791 caused by input 46 bytes long. The cause is due to the way base64
8792 BIOs find the start of base64 encoded data. They do this by trying a
8793 trial decode on each line until they find one that works. When they
8794 do a flag is set and it starts again knowing it can pass all the
8795 data directly through the decoder. Unfortunately it doesn't reset
8796 the context it uses. This means that if EOF is reached an attempt
8797 is made to pass two EOFs through the context and this causes the
8798 resulting error. This can also cause other problems as well. As is
8799 usual with these problems it takes *ages* to find and the fix is
8800 trivial: move one line.
8801 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8803 *) Ugly workaround to get s_client and s_server working under Windows. The
8804 old code wouldn't work because it needed to select() on sockets and the
8805 tty (for keypresses and to see if data could be written). Win32 only
8806 supports select() on sockets so we select() with a 1s timeout on the
8807 sockets and then see if any characters are waiting to be read, if none
8808 are present then we retry, we also assume we can always write data to
8809 the tty. This isn't nice because the code then blocks until we've
8810 received a complete line of data and it is effectively polling the
8811 keyboard at 1s intervals: however it's quite a bit better than not
8812 working at all :-) A dedicated Windows application might handle this
8813 with an event loop for example.
8816 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8817 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8818 will be called when RSA_sign() and RSA_verify() are used. This is useful
8819 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8820 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8821 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8822 This necessitated the support of an extra signature type NID_md5_sha1
8823 for SSL signatures and modifications to the SSL library to use it instead
8824 of calling RSA_public_decrypt() and RSA_private_encrypt().
8827 *) Add new -verify -CAfile and -CApath options to the crl program, these
8828 will lookup a CRL issuers certificate and verify the signature in a
8829 similar way to the verify program. Tidy up the crl program so it
8830 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8831 less strict. It will now permit CRL extensions even if it is not
8832 a V2 CRL: this will allow it to tolerate some broken CRLs.
8835 *) Initialize all non-automatic variables each time one of the openssl
8836 sub-programs is started (this is necessary as they may be started
8837 multiple times from the "OpenSSL>" prompt).
8838 [Lennart Bang, Bodo Moeller]
8840 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8841 removing all other RSA functionality (this is what NO_RSA does). This
8842 is so (for example) those in the US can disable those operations covered
8843 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8847 *) Non-copying interface to BIO pairs.
8848 (still largely untested)
8851 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8852 ASCII string. This was handled independently in various places before.
8855 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8856 UTF8 strings a character at a time.
8859 *) Use client_version from client hello to select the protocol
8860 (s23_srvr.c) and for RSA client key exchange verification
8861 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8864 *) Add various utility functions to handle SPKACs, these were previously
8865 handled by poking round in the structure internals. Added new function
8866 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8867 print, verify and generate SPKACs. Based on an original idea from
8868 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8871 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8874 *) Allow the config file extension section to be overwritten on the
8875 command line. Based on an original idea from Massimiliano Pala
8876 <madwolf@comune.modena.it>. The new option is called -extensions
8877 and can be applied to ca, req and x509. Also -reqexts to override
8878 the request extensions in req and -crlexts to override the crl extensions
8882 *) Add new feature to the SPKAC handling in ca. Now you can include
8883 the same field multiple times by preceding it by "XXXX." for example:
8886 this is the same syntax as used in the req config file.
8889 *) Allow certificate extensions to be added to certificate requests. These
8890 are specified in a 'req_extensions' option of the req section of the
8891 config file. They can be printed out with the -text option to req but
8892 are otherwise ignored at present.
8895 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8896 data read consists of only the final block it would not decrypted because
8897 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8898 A misplaced 'break' also meant the decrypted final block might not be
8899 copied until the next read.
8902 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8903 a few extra parameters to the DH structure: these will be useful if
8904 for example we want the value of 'q' or implement X9.42 DH.
8907 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8908 provides hooks that allow the default DSA functions or functions on a
8909 "per key" basis to be replaced. This allows hardware acceleration and
8910 hardware key storage to be handled without major modification to the
8911 library. Also added low level modexp hooks and CRYPTO_EX structure and
8912 associated functions.
8915 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8916 as "read only": it can't be written to and the buffer it points to will
8917 not be freed. Reading from a read only BIO is much more efficient than
8918 a normal memory BIO. This was added because there are several times when
8919 an area of memory needs to be read from a BIO. The previous method was
8920 to create a memory BIO and write the data to it, this results in two
8921 copies of the data and an O(n^2) reading algorithm. There is a new
8922 function BIO_new_mem_buf() which creates a read only memory BIO from
8923 an area of memory. Also modified the PKCS#7 routines to use read only
8927 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8928 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8929 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8930 but a retry condition occured while trying to read the rest.
8933 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8934 NID_pkcs7_encrypted by default: this was wrong since this should almost
8935 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8936 the encrypted data type: this is a more sensible place to put it and it
8937 allows the PKCS#12 code to be tidied up that duplicated this
8941 *) Changed obj_dat.pl script so it takes its input and output files on
8942 the command line. This should avoid shell escape redirection problems
8946 *) Initial support for certificate extension requests, these are included
8947 in things like Xenroll certificate requests. Included functions to allow
8948 extensions to be obtained and added.
8951 *) -crlf option to s_client and s_server for sending newlines as
8952 CRLF (as required by many protocols).
8955 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8957 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8958 [Ralf S. Engelschall]
8960 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8961 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8963 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8967 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8968 DH parameters/keys (q is lost during that conversion, but the resulting
8969 DH parameters contain its length).
8971 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8972 much faster than DH_generate_parameters (which creates parameters
8973 where p = 2*q + 1), and also the smaller q makes DH computations
8974 much more efficient (160-bit exponentiation instead of 1024-bit
8975 exponentiation); so this provides a convenient way to support DHE
8976 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8977 utter importance to use
8978 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8980 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8981 when such DH parameters are used, because otherwise small subgroup
8982 attacks may become possible!
8985 *) Avoid memory leak in i2d_DHparams.
8988 *) Allow the -k option to be used more than once in the enc program:
8989 this allows the same encrypted message to be read by multiple recipients.
8992 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8993 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8994 it will always use the numerical form of the OID, even if it has a short
8998 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8999 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9000 otherwise bn_mod_exp was called. In the case of hardware keys for example
9001 no private key components need be present and it might store extra data
9002 in the RSA structure, which cannot be accessed from bn_mod_exp.
9003 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9004 private key operations.
9007 *) Added support for SPARC Linux.
9010 *) pem_password_cb function type incompatibly changed from
9011 typedef int pem_password_cb(char *buf, int size, int rwflag);
9013 ....(char *buf, int size, int rwflag, void *userdata);
9014 so that applications can pass data to their callbacks:
9015 The PEM[_ASN1]_{read,write}... functions and macros now take an
9016 additional void * argument, which is just handed through whenever
9017 the password callback is called.
9018 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9020 New function SSL_CTX_set_default_passwd_cb_userdata.
9022 Compatibility note: As many C implementations push function arguments
9023 onto the stack in reverse order, the new library version is likely to
9024 interoperate with programs that have been compiled with the old
9025 pem_password_cb definition (PEM_whatever takes some data that
9026 happens to be on the stack as its last argument, and the callback
9027 just ignores this garbage); but there is no guarantee whatsoever that
9030 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9031 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9032 problems not only on Windows, but also on some Unix platforms.
9033 To avoid problematic command lines, these definitions are now in an
9034 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9035 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9038 *) MIPS III/IV assembler module is reimplemented.
9041 *) More DES library cleanups: remove references to srand/rand and
9042 delete an unused file.
9045 *) Add support for the the free Netwide assembler (NASM) under Win32,
9046 since not many people have MASM (ml) and it can be hard to obtain.
9047 This is currently experimental but it seems to work OK and pass all
9048 the tests. Check out INSTALL.W32 for info.
9051 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9052 without temporary keys kept an extra copy of the server key,
9053 and connections with temporary keys did not free everything in case
9057 *) New function RSA_check_key and new openssl rsa option -check
9058 for verifying the consistency of RSA keys.
9059 [Ulf Moeller, Bodo Moeller]
9061 *) Various changes to make Win32 compile work:
9062 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9063 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9064 comparison" warnings.
9065 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9068 *) Add a debugging option to PKCS#5 v2 key generation function: when
9069 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9070 derived keys are printed to stderr.
9073 *) Copy the flags in ASN1_STRING_dup().
9074 [Roman E. Pavlov <pre@mo.msk.ru>]
9076 *) The x509 application mishandled signing requests containing DSA
9077 keys when the signing key was also DSA and the parameters didn't match.
9079 It was supposed to omit the parameters when they matched the signing key:
9080 the verifying software was then supposed to automatically use the CA's
9081 parameters if they were absent from the end user certificate.
9083 Omitting parameters is no longer recommended. The test was also
9084 the wrong way round! This was probably due to unusual behaviour in
9085 EVP_cmp_parameters() which returns 1 if the parameters match.
9086 This meant that parameters were omitted when they *didn't* match and
9087 the certificate was useless. Certificates signed with 'ca' didn't have
9089 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9091 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9092 The interface is as follows:
9093 Applications can use
9094 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9095 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9096 "off" is now the default.
9097 The library internally uses
9098 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9099 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9100 to disable memory-checking temporarily.
9102 Some inconsistent states that previously were possible (and were
9103 even the default) are now avoided.
9105 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9106 with each memory chunk allocated; this is occasionally more helpful
9107 than just having a counter.
9109 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9111 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9115 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9116 which largely parallels "options", but is for changing API behaviour,
9117 whereas "options" are about protocol behaviour.
9118 Initial "mode" flags are:
9120 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9121 a single record has been written.
9122 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9123 retries use the same buffer location.
9124 (But all of the contents must be
9128 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9131 *) Fix problems with no-hmac etc.
9132 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9134 *) New functions RSA_get_default_method(), RSA_set_method() and
9135 RSA_get_method(). These allows replacement of RSA_METHODs without having
9136 to mess around with the internals of an RSA structure.
9139 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9140 Also really enable memory leak checks in openssl.c and in some
9142 [Chad C. Mulligan, Bodo Moeller]
9144 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9145 up the length of negative integers. This has now been simplified to just
9146 store the length when it is first determined and use it later, rather
9147 than trying to keep track of where data is copied and updating it to
9149 [Steve Henson, reported by Brien Wheeler
9150 <bwheeler@authentica-security.com>]
9152 *) Add a new function PKCS7_signatureVerify. This allows the verification
9153 of a PKCS#7 signature but with the signing certificate passed to the
9154 function itself. This contrasts with PKCS7_dataVerify which assumes the
9155 certificate is present in the PKCS#7 structure. This isn't always the
9156 case: certificates can be omitted from a PKCS#7 structure and be
9157 distributed by "out of band" means (such as a certificate database).
9160 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9161 function prototypes in pem.h, also change util/mkdef.pl to add the
9162 necessary function names.
9165 *) mk1mf.pl (used by Windows builds) did not properly read the
9166 options set by Configure in the top level Makefile, and Configure
9167 was not even able to write more than one option correctly.
9168 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9171 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9172 file to be loaded from a BIO or FILE pointer. The BIO version will
9173 for example allow memory BIOs to contain config info.
9176 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9177 Whoever hopes to achieve shared-library compatibility across versions
9178 must use this, not the compile-time macro.
9179 (Exercise 0.9.4: Which is the minimum library version required by
9181 Note: All this applies only to multi-threaded programs, others don't
9185 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9186 through a BIO pair triggered the default case, i.e.
9187 SSLerr(...,SSL_R_UNKNOWN_STATE).
9190 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9191 can use the SSL library even if none of the specific BIOs is
9195 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9196 for the encoded length.
9197 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9199 *) Add initial documentation of the X509V3 functions.
9202 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9203 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9204 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9205 secure PKCS#8 private key format with a high iteration count.
9208 *) Fix determination of Perl interpreter: A perl or perl5
9209 _directory_ in $PATH was also accepted as the interpreter.
9210 [Ralf S. Engelschall]
9212 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9213 wrong with it but it was very old and did things like calling
9214 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9218 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9219 to use the new extension code.
9222 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9223 with macros. This should make it easier to change their form, add extra
9224 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9228 *) Add to configuration table a new entry that can specify an alternative
9229 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9230 according to Mark Crispin <MRC@Panda.COM>.
9234 *) DES CBC did not update the IV. Weird.
9237 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9238 Changing the behaviour of the former might break existing programs --
9239 where IV updating is needed, des_ncbc_encrypt can be used.
9242 *) When bntest is run from "make test" it drives bc to check its
9243 calculations, as well as internally checking them. If an internal check
9244 fails, it needs to cause bc to give a non-zero result or make test carries
9245 on without noticing the failure. Fixed.
9248 *) DES library cleanups.
9251 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9252 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9253 ciphers. NOTE: although the key derivation function has been verified
9254 against some published test vectors it has not been extensively tested
9255 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9259 *) Instead of "mkdir -p", which is not fully portable, use new
9260 Perl script "util/mkdir-p.pl".
9263 *) Rewrite the way password based encryption (PBE) is handled. It used to
9264 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9265 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9266 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9267 the 'parameter' field of the AlgorithmIdentifier is passed to the
9268 underlying key generation function so it must do its own ASN1 parsing.
9269 This has also changed the EVP_PBE_CipherInit() function which now has a
9270 'parameter' argument instead of literal salt and iteration count values
9271 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9274 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9275 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9276 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9277 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9278 value was just used as a "magic string" and not used directly its
9279 value doesn't matter.
9282 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9286 *) "linux-sparc64" configuration (ultrapenguin).
9287 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9288 "linux-sparc" configuration.
9289 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9291 *) config now generates no-xxx options for missing ciphers.
9294 *) Support the EBCDIC character set (work in progress).
9295 File ebcdic.c not yet included because it has a different license.
9296 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9298 *) Support BS2000/OSD-POSIX.
9299 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9301 *) Make callbacks for key generation use void * instead of char *.
9304 *) Make S/MIME samples compile (not yet tested).
9307 *) Additional typesafe stacks.
9310 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9314 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9316 *) New configuration variant "sco5-gcc".
9318 *) Updated some demos.
9319 [Sean O Riordain, Wade Scholine]
9321 *) Add missing BIO_free at exit of pkcs12 application.
9324 *) Fix memory leak in conf.c.
9327 *) Updates for Win32 to assembler version of MD5.
9330 *) Set #! path to perl in apps/der_chop to where we found it
9331 instead of using a fixed path.
9334 *) SHA library changes for irix64-mips4-cc.
9337 *) Improvements for VMS support.
9341 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9343 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9344 This also avoids the problems with SC4.2 and unpatched SC5.
9345 [Andy Polyakov <appro@fy.chalmers.se>]
9347 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9348 These are required because of the typesafe stack would otherwise break
9349 existing code. If old code used a structure member which used to be STACK
9350 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9351 sk_num or sk_value it would produce an error because the num, data members
9352 are not present in STACK_OF. Now it just produces a warning. sk_set
9353 replaces the old method of assigning a value to sk_value
9354 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9355 that does this will no longer work (and should use sk_set instead) but
9356 this could be regarded as a "questionable" behaviour anyway.
9359 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9360 correctly handle encrypted S/MIME data.
9363 *) Change type of various DES function arguments from des_cblock
9364 (which means, in function argument declarations, pointer to char)
9365 to des_cblock * (meaning pointer to array with 8 char elements),
9366 which allows the compiler to do more typechecking; it was like
9367 that back in SSLeay, but with lots of ugly casts.
9369 Introduce new type const_des_cblock.
9372 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9373 problems: find RecipientInfo structure that matches recipient certificate
9374 and initialise the ASN1 structures properly based on passed cipher.
9377 *) Belatedly make the BN tests actually check the results.
9380 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9381 to and from BNs: it was completely broken. New compilation option
9382 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9383 key elements as negative integers.
9386 *) Reorganize and speed up MD5.
9387 [Andy Polyakov <appro@fy.chalmers.se>]
9390 [Richard Levitte <richard@levitte.org>]
9392 *) New option -out to asn1parse to allow the parsed structure to be
9393 output to a file. This is most useful when combined with the -strparse
9394 option to examine the output of things like OCTET STRINGS.
9397 *) Make SSL library a little more fool-proof by not requiring any longer
9398 that SSL_set_{accept,connect}_state be called before
9399 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9400 in many applications because usually everything *appeared* to work as
9401 intended anyway -- now it really works as intended).
9404 *) Move openssl.cnf out of lib/.
9407 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9408 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9409 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9410 [Ralf S. Engelschall]
9412 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9413 handle PKCS#7 enveloped data properly.
9414 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9416 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9417 copying pointers. The cert_st handling is changed by this in
9418 various ways (and thus what used to be known as ctx->default_cert
9419 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9420 any longer when s->cert does not give us what we need).
9421 ssl_cert_instantiate becomes obsolete by this change.
9422 As soon as we've got the new code right (possibly it already is?),
9423 we have solved a couple of bugs of the earlier code where s->cert
9424 was used as if it could not have been shared with other SSL structures.
9426 Note that using the SSL API in certain dirty ways now will result
9427 in different behaviour than observed with earlier library versions:
9428 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9429 does not influence s as it used to.
9431 In order to clean up things more thoroughly, inside SSL_SESSION
9432 we don't use CERT any longer, but a new structure SESS_CERT
9433 that holds per-session data (if available); currently, this is
9434 the peer's certificate chain and, for clients, the server's certificate
9435 and temporary key. CERT holds only those values that can have
9436 meaningful defaults in an SSL_CTX.
9439 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9440 from the internal representation. Various PKCS#7 fixes: remove some
9441 evil casts and set the enc_dig_alg field properly based on the signing
9445 *) Allow PKCS#12 password to be set from the command line or the
9446 environment. Let 'ca' get its config file name from the environment
9447 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9451 *) Allow certificate policies extension to use an IA5STRING for the
9452 organization field. This is contrary to the PKIX definition but
9453 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9457 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9458 without disallowing inline assembler and the like for non-pedantic builds.
9461 *) Support Borland C++ builder.
9462 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9467 *) SHA-1 cleanups and performance enhancements.
9468 [Andy Polyakov <appro@fy.chalmers.se>]
9470 *) Sparc v8plus assembler for the bignum library.
9471 [Andy Polyakov <appro@fy.chalmers.se>]
9473 *) Accept any -xxx and +xxx compiler options in Configure.
9476 *) Update HPUX configuration.
9479 *) Add missing sk_<type>_unshift() function to safestack.h
9480 [Ralf S. Engelschall]
9482 *) New function SSL_CTX_use_certificate_chain_file that sets the
9483 "extra_cert"s in addition to the certificate. (This makes sense
9484 only for "PEM" format files, as chains as a whole are not
9488 *) Support verify_depth from the SSL API.
9489 x509_vfy.c had what can be considered an off-by-one-error:
9490 Its depth (which was not part of the external interface)
9491 was actually counting the number of certificates in a chain;
9492 now it really counts the depth.
9495 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9496 instead of X509err, which often resulted in confusing error
9497 messages since the error codes are not globally unique
9498 (e.g. an alleged error in ssl3_accept when a certificate
9499 didn't match the private key).
9501 *) New function SSL_CTX_set_session_id_context that allows to set a default
9502 value (so that you don't need SSL_set_session_id_context for each
9503 connection using the SSL_CTX).
9506 *) OAEP decoding bug fix.
9509 *) Support INSTALL_PREFIX for package builders, as proposed by
9513 *) New Configure options "threads" and "no-threads". For systems
9514 where the proper compiler options are known (currently Solaris
9515 and Linux), "threads" is the default.
9518 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9521 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9522 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9523 such as /usr/local/bin.
9526 *) "make linux-shared" to build shared libraries.
9527 [Niels Poppe <niels@netbox.org>]
9529 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9532 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9533 extension adding in x509 utility.
9536 *) Remove NOPROTO sections and error code comments.
9539 *) Partial rewrite of the DEF file generator to now parse the ANSI
9543 *) New Configure options --prefix=DIR and --openssldir=DIR.
9546 *) Complete rewrite of the error code script(s). It is all now handled
9547 by one script at the top level which handles error code gathering,
9548 header rewriting and C source file generation. It should be much better
9549 than the old method: it now uses a modified version of Ulf's parser to
9550 read the ANSI prototypes in all header files (thus the old K&R definitions
9551 aren't needed for error creation any more) and do a better job of
9552 translating function codes into names. The old 'ASN1 error code imbedded
9553 in a comment' is no longer necessary and it doesn't use .err files which
9554 have now been deleted. Also the error code call doesn't have to appear all
9555 on one line (which resulted in some large lines...).
9558 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9561 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9562 0 (which usually indicates a closed connection), but continue reading.
9565 *) Fix some race conditions.
9568 *) Add support for CRL distribution points extension. Add Certificate
9569 Policies and CRL distribution points documentation.
9572 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9575 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9576 8 of keying material. Merlin has also confirmed interop with this fix
9577 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9578 [Merlin Hughes <merlin@baltimore.ie>]
9580 *) Fix lots of warnings.
9581 [Richard Levitte <levitte@stacken.kth.se>]
9583 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9584 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9585 [Richard Levitte <levitte@stacken.kth.se>]
9587 *) Fix problems with sizeof(long) == 8.
9588 [Andy Polyakov <appro@fy.chalmers.se>]
9590 *) Change functions to ANSI C.
9593 *) Fix typos in error codes.
9594 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9596 *) Remove defunct assembler files from Configure.
9599 *) SPARC v8 assembler BIGNUM implementation.
9600 [Andy Polyakov <appro@fy.chalmers.se>]
9602 *) Support for Certificate Policies extension: both print and set.
9603 Various additions to support the r2i method this uses.
9606 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9607 return a const string when you are expecting an allocated buffer.
9610 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9611 types DirectoryString and DisplayText.
9614 *) Add code to allow r2i extensions to access the configuration database,
9615 add an LHASH database driver and add several ctx helper functions.
9618 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9619 fail when they extended the size of a BIGNUM.
9622 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9623 support typesafe stack.
9626 *) Fix typo in SSL_[gs]et_options().
9627 [Nils Frostberg <nils@medcom.se>]
9629 *) Delete various functions and files that belonged to the (now obsolete)
9630 old X509V3 handling code.
9633 *) New Configure option "rsaref".
9636 *) Don't auto-generate pem.h.
9639 *) Introduce type-safe ASN.1 SETs.
9642 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9643 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9645 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9646 that links with OpenSSL (well at least cause lots of warnings), but fear
9647 not: the conversion is trivial, and it eliminates loads of evil casts. A
9648 few STACKed things have been converted already. Feel free to convert more.
9649 In the fullness of time, I'll do away with the STACK type altogether.
9652 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9653 specified in <certfile> by updating the entry in the index.txt file.
9654 This way one no longer has to edit the index.txt file manually for
9655 revoking a certificate. The -revoke option does the gory details now.
9656 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9658 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9659 `-text' option at all and this way the `-noout -text' combination was
9660 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9661 [Ralf S. Engelschall]
9663 *) Make sure a corresponding plain text error message exists for the
9664 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9665 verify callback function determined that a certificate was revoked.
9666 [Ralf S. Engelschall]
9668 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9669 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9670 all available cipers including rc5, which was forgotten until now.
9671 In order to let the testing shell script know which algorithms
9672 are available, a new (up to now undocumented) command
9673 "openssl list-cipher-commands" is used.
9676 *) Bugfix: s_client occasionally would sleep in select() when
9677 it should have checked SSL_pending() first.
9680 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9681 the raw DSA values prior to ASN.1 encoding.
9684 *) Tweaks to Configure
9685 [Niels Poppe <niels@netbox.org>]
9687 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9691 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9694 *) New config option to avoid instructions that are illegal on the 80386.
9695 The default code is faster, but requires at least a 486.
9698 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9699 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9700 same as SSL2_VERSION anyway.
9703 *) New "-showcerts" option for s_client.
9706 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9707 application. Various cleanups and fixes.
9710 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9711 modify error routines to work internally. Add error codes and PBE init
9712 to library startup routines.
9715 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9716 packing functions to asn1 and evp. Changed function names and error
9717 codes along the way.
9720 *) PKCS12 integration: and so it begins... First of several patches to
9721 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9722 objects to objects.h
9725 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9726 and display support for Thawte strong extranet extension.
9729 *) Add LinuxPPC support.
9730 [Jeff Dubrule <igor@pobox.org>]
9732 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9733 bn_div_words in alpha.s.
9734 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9736 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9737 OAEP isn't supported when OpenSSL is built with RSAref.
9738 [Ulf Moeller <ulf@fitug.de>]
9740 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9741 so they no longer are missing under -DNOPROTO.
9742 [Soren S. Jorvang <soren@t.dk>]
9745 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9747 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9748 doesn't work when the session is reused. Coming soon!
9751 *) Fix a security hole, that allows sessions to be reused in the wrong
9752 context thus bypassing client cert protection! All software that uses
9753 client certs and session caches in multiple contexts NEEDS PATCHING to
9754 allow session reuse! A fuller solution is in the works.
9755 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9757 *) Some more source tree cleanups (removed obsolete files
9758 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9759 permission on "config" script to be executable) and a fix for the INSTALL
9761 [Ulf Moeller <ulf@fitug.de>]
9763 *) Remove some legacy and erroneous uses of malloc, free instead of
9765 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9767 *) Make rsa_oaep_test return non-zero on error.
9768 [Ulf Moeller <ulf@fitug.de>]
9770 *) Add support for native Solaris shared libraries. Configure
9771 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9772 if someone would make that last step automatic.
9773 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9775 *) ctx_size was not built with the right compiler during "make links". Fixed.
9778 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9779 except NULL ciphers". This means the default cipher list will no longer
9780 enable NULL ciphers. They need to be specifically enabled e.g. with
9781 the string "DEFAULT:eNULL".
9784 *) Fix to RSA private encryption routines: if p < q then it would
9785 occasionally produce an invalid result. This will only happen with
9786 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9789 *) Be less restrictive and allow also `perl util/perlpath.pl
9790 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9791 because this way one can also use an interpreter named `perl5' (which is
9792 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9793 installed as `perl').
9794 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9796 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9797 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9799 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9800 advapi32.lib to Win32 build and change the pem test comparision
9801 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9802 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9803 and crypto/des/ede_cbcm_enc.c.
9806 *) DES quad checksum was broken on big-endian architectures. Fixed.
9809 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9810 Win32 test batch file so it (might) work again. The Win32 test batch file
9811 is horrible: I feel ill....
9814 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9815 in e_os.h. Audit of header files to check ANSI and non ANSI
9816 sections: 10 functions were absent from non ANSI section and not exported
9817 from Windows DLLs. Fixed up libeay.num for new functions.
9820 *) Make `openssl version' output lines consistent.
9821 [Ralf S. Engelschall]
9823 *) Fix Win32 symbol export lists for BIO functions: Added
9824 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9825 to ms/libeay{16,32}.def.
9826 [Ralf S. Engelschall]
9828 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9829 fine under Unix and passes some trivial tests I've now added. But the
9830 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9831 added to make sure no one expects that this stuff really works in the
9832 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9833 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9835 [Ralf S. Engelschall]
9837 *) Fix the generation of two part addresses in perl.
9838 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9840 *) Add config entry for Linux on MIPS.
9841 [John Tobey <jtobey@channel1.com>]
9843 *) Make links whenever Configure is run, unless we are on Windoze.
9846 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9847 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9851 *) Add a useful kludge to allow package maintainers to specify compiler and
9852 other platforms details on the command line without having to patch the
9853 Configure script everytime: One now can use ``perl Configure
9854 <id>:<details>'', i.e. platform ids are allowed to have details appended
9855 to them (seperated by colons). This is treated as there would be a static
9856 pre-configured entry in Configure's %table under key <id> with value
9857 <details> and ``perl Configure <id>'' is called. So, when you want to
9858 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9859 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9860 now, which overrides the FreeBSD-elf entry on-the-fly.
9861 [Ralf S. Engelschall]
9863 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9866 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9867 on the `perl Configure ...' command line. This way one can compile
9868 OpenSSL libraries with Position Independent Code (PIC) which is needed
9869 for linking it into DSOs.
9870 [Ralf S. Engelschall]
9872 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9876 *) Cleaned up the LICENSE document: The official contact for any license
9877 questions now is the OpenSSL core team under openssl-core@openssl.org.
9878 And add a paragraph about the dual-license situation to make sure people
9879 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9880 to the OpenSSL toolkit.
9881 [Ralf S. Engelschall]
9883 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9884 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9885 Additonally cleaned up the `make links' target: Remove unnecessary
9886 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9887 to speed processing and no longer clutter the display with confusing
9888 stuff. Instead only the actually done links are displayed.
9889 [Ralf S. Engelschall]
9891 *) Permit null encryption ciphersuites, used for authentication only. It used
9892 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9893 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9897 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9898 signed attributes when verifying signatures (this would break them),
9899 the detached data encoding was wrong and public keys obtained using
9900 X509_get_pubkey() weren't freed.
9903 *) Add text documentation for the BUFFER functions. Also added a work around
9904 to a Win95 console bug. This was triggered by the password read stuff: the
9905 last character typed gets carried over to the next fread(). If you were
9906 generating a new cert request using 'req' for example then the last
9907 character of the passphrase would be CR which would then enter the first
9911 *) Added the new `Includes OpenSSL Cryptography Software' button as
9912 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9913 button and can be used by applications based on OpenSSL to show the
9914 relationship to the OpenSSL project.
9915 [Ralf S. Engelschall]
9917 *) Remove confusing variables in function signatures in files
9918 ssl/ssl_lib.c and ssl/ssl.h.
9919 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9921 *) Don't install bss_file.c under PREFIX/include/
9922 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9924 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9925 functions that return function pointers and has support for NT specific
9926 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9927 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9928 unsigned to signed types: this was killing the Win32 compile.
9931 *) Add new certificate file to stack functions,
9932 SSL_add_dir_cert_subjects_to_stack() and
9933 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9934 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9935 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9936 This means that Apache-SSL and similar packages don't have to mess around
9937 to add as many CAs as they want to the preferred list.
9940 *) Experiment with doxygen documentation. Currently only partially applied to
9942 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9943 openssl.doxy as the configuration file.
9946 *) Get rid of remaining C++-style comments which strict C compilers hate.
9947 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9949 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9950 compiled in by default: it has problems with large keys.
9953 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9954 DH private keys and/or callback functions which directly correspond to
9955 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9956 is needed for applications which have to configure certificates on a
9957 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9959 For the RSA certificate situation is makes no difference, but
9960 for the DSA certificate situation this fixes the "no shared cipher"
9961 problem where the OpenSSL cipher selection procedure failed because the
9962 temporary keys were not overtaken from the context and the API provided
9963 no way to reconfigure them.
9964 The new functions now let applications reconfigure the stuff and they
9965 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9966 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9967 non-public-API function ssl_cert_instantiate() is used as a helper
9968 function and also to reduce code redundancy inside ssl_rsa.c.
9969 [Ralf S. Engelschall]
9971 *) Move s_server -dcert and -dkey options out of the undocumented feature
9972 area because they are useful for the DSA situation and should be
9973 recognized by the users.
9974 [Ralf S. Engelschall]
9976 *) Fix the cipher decision scheme for export ciphers: the export bits are
9977 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9978 SSL_EXP_MASK. So, the original variable has to be used instead of the
9979 already masked variable.
9980 [Richard Levitte <levitte@stacken.kth.se>]
9982 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9983 [Richard Levitte <levitte@stacken.kth.se>]
9985 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9986 from `int' to `unsigned int' because it's a length and initialized by
9987 EVP_DigestFinal() which expects an `unsigned int *'.
9988 [Richard Levitte <levitte@stacken.kth.se>]
9990 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9991 script. Instead use the usual Shell->Perl transition trick.
9992 [Ralf S. Engelschall]
9994 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9995 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9996 -noout -modulus' as it's already the case for `openssl rsa -noout
9997 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9998 currently the public key is printed (a decision which was already done by
9999 `openssl dsa -modulus' in the past) which serves a similar purpose.
10000 Additionally the NO_RSA no longer completely removes the whole -modulus
10001 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10003 [Ralf S. Engelschall]
10005 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10006 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10007 [Arne Ansper <arne@ats.cyber.ee>]
10009 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10010 to be added. Now both 'req' and 'ca' can use new objects defined in the
10014 *) Add cool BIO that does syslog (or event log on NT).
10015 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10017 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10018 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10019 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10020 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10023 *) Add preliminary config info for new extension code.
10026 *) Make RSA_NO_PADDING really use no padding.
10027 [Ulf Moeller <ulf@fitug.de>]
10029 *) Generate errors when private/public key check is done.
10032 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10033 for some CRL extensions and new objects added.
10036 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10037 key usage extension and fuller support for authority key id.
10040 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10041 padding method for RSA, which is recommended for new applications in PKCS
10042 #1 v2.0 (RFC 2437, October 1998).
10043 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10044 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10045 against Bleichbacher's attack on RSA.
10046 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10049 *) Updates to the new SSL compression code
10050 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10052 *) Fix so that the version number in the master secret, when passed
10053 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10054 (because the server will not accept higher), that the version number
10055 is 0x03,0x01, not 0x03,0x00
10056 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10058 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10059 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10060 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10063 *) Support for RAW extensions where an arbitrary extension can be
10064 created by including its DER encoding. See apps/openssl.cnf for
10068 *) Make sure latest Perl versions don't interpret some generated C array
10069 code as Perl array code in the crypto/err/err_genc.pl script.
10070 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10072 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10073 not many people have the assembler. Various Win32 compilation fixes and
10074 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10075 build instructions.
10078 *) Modify configure script 'Configure' to automatically create crypto/date.h
10079 file under Win32 and also build pem.h from pem.org. New script
10080 util/mkfiles.pl to create the MINFO file on environments that can't do a
10081 'make files': perl util/mkfiles.pl >MINFO should work.
10084 *) Major rework of DES function declarations, in the pursuit of correctness
10085 and purity. As a result, many evil casts evaporated, and some weirdness,
10086 too. You may find this causes warnings in your code. Zapping your evil
10087 casts will probably fix them. Mostly.
10090 *) Fix for a typo in asn1.h. Bug fix to object creation script
10091 obj_dat.pl. It considered a zero in an object definition to mean
10092 "end of object": none of the objects in objects.h have any zeros
10093 so it wasn't spotted.
10094 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10096 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10097 Masking (CBCM). In the absence of test vectors, the best I have been able
10098 to do is check that the decrypt undoes the encrypt, so far. Send me test
10099 vectors if you have them.
10102 *) Correct calculation of key length for export ciphers (too much space was
10103 allocated for null ciphers). This has not been tested!
10106 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10107 message is now correct (it understands "crypto" and "ssl" on its
10108 command line). There is also now an "update" option. This will update
10109 the util/ssleay.num and util/libeay.num files with any new functions.
10111 perl util/mkdef.pl crypto ssl update
10112 it will update them.
10115 *) Overhauled the Perl interface (perl/*):
10116 - ported BN stuff to OpenSSL's different BN library
10117 - made the perl/ source tree CVS-aware
10118 - renamed the package from SSLeay to OpenSSL (the files still contain
10119 their history because I've copied them in the repository)
10120 - removed obsolete files (the test scripts will be replaced
10121 by better Test::Harness variants in the future)
10122 [Ralf S. Engelschall]
10124 *) First cut for a very conservative source tree cleanup:
10125 1. merge various obsolete readme texts into doc/ssleay.txt
10126 where we collect the old documents and readme texts.
10127 2. remove the first part of files where I'm already sure that we no
10128 longer need them because of three reasons: either they are just temporary
10129 files which were left by Eric or they are preserved original files where
10130 I've verified that the diff is also available in the CVS via "cvs diff
10131 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10132 the crypto/md/ stuff).
10133 [Ralf S. Engelschall]
10135 *) More extension code. Incomplete support for subject and issuer alt
10136 name, issuer and authority key id. Change the i2v function parameters
10137 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10138 what that's for :-) Fix to ASN1 macro which messed up
10139 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10142 *) Preliminary support for ENUMERATED type. This is largely copied from the
10146 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10147 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10149 *) Make sure `make rehash' target really finds the `openssl' program.
10150 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10152 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10153 like to hear about it if this slows down other processors.
10156 *) Add CygWin32 platform information to Configure script.
10157 [Alan Batie <batie@aahz.jf.intel.com>]
10159 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10160 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10162 *) New program nseq to manipulate netscape certificate sequences
10165 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10169 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10170 but the BN code had some problems that would cause failures when
10171 doing certificate verification and some other functions.
10172 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10174 *) Add ASN1 and PEM code to support netscape certificate sequences.
10177 *) Add ASN1 and PEM code to support netscape certificate sequences.
10180 *) Add several PKIX and private extended key usage OIDs.
10183 *) Modify the 'ca' program to handle the new extension code. Modify
10184 openssl.cnf for new extension format, add comments.
10187 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10188 and add a sample to openssl.cnf so req -x509 now adds appropriate
10192 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10193 error code, add initial support to X509_print() and x509 application.
10196 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10197 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10198 stuff is currently isolated and isn't even compiled yet.
10201 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10202 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10203 Removed the versions check from X509 routines when loading extensions:
10204 this allows certain broken certificates that don't set the version
10205 properly to be processed.
10208 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10209 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10210 can still be regenerated with "make depend".
10213 *) Spelling mistake in C version of CAST-128.
10214 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10216 *) Changes to the error generation code. The perl script err-code.pl
10217 now reads in the old error codes and retains the old numbers, only
10218 adding new ones if necessary. It also only changes the .err files if new
10219 codes are added. The makefiles have been modified to only insert errors
10220 when needed (to avoid needlessly modifying header files). This is done
10221 by only inserting errors if the .err file is newer than the auto generated
10222 C file. To rebuild all the error codes from scratch (the old behaviour)
10223 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10224 or delete all the .err files.
10227 *) CAST-128 was incorrectly implemented for short keys. The C version has
10228 been fixed, but is untested. The assembler versions are also fixed, but
10229 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10230 to regenerate it if needed.
10231 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10232 Hagino <itojun@kame.net>]
10234 *) File was opened incorrectly in randfile.c.
10235 [Ulf Möller <ulf@fitug.de>]
10237 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10238 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10239 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10240 al: it's just almost always a UTCTime. Note this patch adds new error
10241 codes so do a "make errors" if there are problems.
10244 *) Correct Linux 1 recognition in config.
10245 [Ulf Möller <ulf@fitug.de>]
10247 *) Remove pointless MD5 hash when using DSA keys in ca.
10248 [Anonymous <nobody@replay.com>]
10250 *) Generate an error if given an empty string as a cert directory. Also
10251 generate an error if handed NULL (previously returned 0 to indicate an
10252 error, but didn't set one).
10253 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10255 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10258 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10259 parameters. This was causing a warning which killed off the Win32 compile.
10262 *) Remove C++ style comments from crypto/bn/bn_local.h.
10263 [Neil Costigan <neil.costigan@celocom.com>]
10265 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10266 based on a text string, looking up short and long names and finally
10267 "dot" format. The "dot" format stuff didn't work. Added new function
10268 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10269 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10270 OID is not part of the table.
10273 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10274 X509_LOOKUP_by_alias().
10277 *) Sort openssl functions by name.
10280 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10281 encryption from sample DSA keys (in case anyone is interested the password
10285 *) Make _all_ *_free functions accept a NULL pointer.
10286 [Frans Heymans <fheymans@isaserver.be>]
10288 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10290 [Anonymous <nobody@replay.com>]
10292 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10293 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10295 *) Don't blow it for numeric -newkey arguments to apps/req.
10296 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10298 *) Temp key "for export" tests were wrong in s3_srvr.c.
10299 [Anonymous <nobody@replay.com>]
10301 *) Add prototype for temp key callback functions
10302 SSL_CTX_set_tmp_{rsa,dh}_callback().
10305 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10306 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10309 *) X509_name_add_entry() freed the wrong thing after an error.
10310 [Arne Ansper <arne@ats.cyber.ee>]
10312 *) rsa_eay.c would attempt to free a NULL context.
10313 [Arne Ansper <arne@ats.cyber.ee>]
10315 *) BIO_s_socket() had a broken should_retry() on Windoze.
10316 [Arne Ansper <arne@ats.cyber.ee>]
10318 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10319 [Arne Ansper <arne@ats.cyber.ee>]
10321 *) Make sure the already existing X509_STORE->depth variable is initialized
10322 in X509_STORE_new(), but document the fact that this variable is still
10323 unused in the certificate verification process.
10324 [Ralf S. Engelschall]
10326 *) Fix the various library and apps files to free up pkeys obtained from
10327 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10330 *) Fix reference counting in X509_PUBKEY_get(). This makes
10331 demos/maurice/example2.c work, amongst others, probably.
10332 [Steve Henson and Ben Laurie]
10334 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10335 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10336 are no longer created. This way we have a single and consistent command
10337 line interface `openssl <command>', similar to `cvs <command>'.
10338 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10340 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10341 BIT STRING wrapper always have zero unused bits.
10344 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10347 *) Make the top-level INSTALL documentation easier to understand.
10350 *) Makefiles updated to exit if an error occurs in a sub-directory
10351 make (including if user presses ^C) [Paul Sutton]
10353 *) Make Montgomery context stuff explicit in RSA data structure.
10356 *) Fix build order of pem and err to allow for generated pem.h.
10359 *) Fix renumbering bug in X509_NAME_delete_entry().
10362 *) Enhanced the err-ins.pl script so it makes the error library number
10363 global and can add a library name. This is needed for external ASN1 and
10364 other error libraries.
10367 *) Fixed sk_insert which never worked properly.
10370 *) Fix ASN1 macros so they can handle indefinite length construted
10371 EXPLICIT tags. Some non standard certificates use these: they can now
10375 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10376 into a single doc/ssleay.txt bundle. This way the information is still
10377 preserved but no longer messes up this directory. Now it's new room for
10378 the new set of documenation files.
10379 [Ralf S. Engelschall]
10381 *) SETs were incorrectly DER encoded. This was a major pain, because they
10382 shared code with SEQUENCEs, which aren't coded the same. This means that
10383 almost everything to do with SETs or SEQUENCEs has either changed name or
10384 number of arguments.
10385 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10387 *) Fix test data to work with the above.
10390 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10391 was already fixed by Eric for 0.9.1 it seems.
10392 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10394 *) Autodetect FreeBSD3.
10397 *) Fix various bugs in Configure. This affects the following platforms:
10401 unixware-2.0-pentium
10405 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10406 before they are needed.
10409 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10413 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10415 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10416 changed SSLeay to OpenSSL in version strings.
10417 [Ralf S. Engelschall]
10419 *) Some fixups to the top-level documents.
10422 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10423 because the symlink to include/ was missing.
10424 [Ralf S. Engelschall]
10426 *) Incorporated the popular no-RSA/DSA-only patches
10427 which allow to compile a RSA-free SSLeay.
10428 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10430 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10431 when "ssleay" is still not found.
10432 [Ralf S. Engelschall]
10434 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10435 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10437 *) Updated the README file.
10438 [Ralf S. Engelschall]
10440 *) Added various .cvsignore files in the CVS repository subdirs
10441 to make a "cvs update" really silent.
10442 [Ralf S. Engelschall]
10444 *) Recompiled the error-definition header files and added
10445 missing symbols to the Win32 linker tables.
10446 [Ralf S. Engelschall]
10448 *) Cleaned up the top-level documents;
10449 o new files: CHANGES and LICENSE
10450 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10451 o merged COPYRIGHT into LICENSE
10452 o removed obsolete TODO file
10453 o renamed MICROSOFT to INSTALL.W32
10454 [Ralf S. Engelschall]
10456 *) Removed dummy files from the 0.9.1b source tree:
10457 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10458 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10459 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10460 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10461 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10462 [Ralf S. Engelschall]
10464 *) Added various platform portability fixes.
10467 *) The Genesis of the OpenSSL rpject:
10468 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10469 Young and Tim J. Hudson created while they were working for C2Net until
10471 [The OpenSSL Project]
10474 Changes between 0.9.0b and 0.9.1b [not released]
10476 *) Updated a few CA certificates under certs/
10479 *) Changed some BIGNUM api stuff.
10482 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10483 DGUX x86, Linux Alpha, etc.
10486 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10487 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10491 *) Add -strparse option to asn1pars program which parses nested
10493 [Dr Stephen Henson <shenson@bigfoot.com>]
10495 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10498 *) DSA fix for "ca" program.
10501 *) Added "-genkey" option to "dsaparam" program.
10504 *) Added RIPE MD160 (rmd160) message digest.
10507 *) Added -a (all) option to "ssleay version" command.
10510 *) Added PLATFORM define which is the id given to Configure.
10513 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10516 *) Extended the ASN.1 parser routines.
10519 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10522 *) Added a BN_CTX to the BN library.
10525 *) Fixed the weak key values in DES library
10528 *) Changed API in EVP library for cipher aliases.
10531 *) Added support for RC2/64bit cipher.
10534 *) Converted the lhash library to the crypto/mem.c functions.
10537 *) Added more recognized ASN.1 object ids.
10540 *) Added more RSA padding checks for SSL/TLS.
10543 *) Added BIO proxy/filter functionality.
10546 *) Added extra_certs to SSL_CTX which can be used
10547 send extra CA certificates to the client in the CA cert chain sending
10548 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10551 *) Now Fortezza is denied in the authentication phase because
10552 this is key exchange mechanism is not supported by SSLeay at all.
10555 *) Additional PKCS1 checks.
10558 *) Support the string "TLSv1" for all TLS v1 ciphers.
10561 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10562 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10565 *) Fixed a few memory leaks.
10568 *) Fixed various code and comment typos.
10571 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10572 bytes sent in the client random.
10573 [Edward Bishop <ebishop@spyglass.com>]