5 Changes between 1.0.1i and 1.0.2 [xx XXX xxxx]
7 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
8 (other platforms pending).
9 [Shay Gueron (Intel Corp), Andy Polyakov]
11 *) Add support for the SignedCertificateTimestampList certificate and
12 OCSP response extensions from RFC6962.
15 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
16 for corner cases. (Certain input points at infinity could lead to
17 bogus results, with non-infinity inputs mapped to infinity too.)
20 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
21 This covers AES, SHA256/512 and GHASH. "Initial" means that most
22 common cases are optimized and there still is room for further
23 improvements. Vector Permutation AES for Altivec is also added.
26 *) Add support for little-endian ppc64 Linux target.
29 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
30 SHA1, SHA256 and GHASH. "Initial" means that most common cases
31 are optimized and there still is room for further improvements.
32 Both 32- and 64-bit modes are supported.
33 [Andy Polyakov, Ard Biesheuvel (Linaro)]
35 *) Improved ARMv7 NEON support.
38 *) Support for SPARC Architecture 2011 crypto extensions, first
39 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
40 SHA256/512, MD5, GHASH and modular exponentiation.
41 [Andy Polyakov, David Miller]
43 *) Accelerated modular exponentiation for Intel processors, a.k.a.
45 [Shay Gueron (Intel Corp)]
47 *) Support for new and upcoming Intel processors, including AVX2,
48 BMI and SHA ISA extensions. This includes additional "stitched"
49 implementations, AESNI-SHA256 and GCM, and multi-buffer support
52 This work was sponsored by Intel Corp.
55 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
56 this fixes a limiation in previous versions of OpenSSL.
59 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
60 MGF1 digest and OAEP label.
63 *) Add EVP support for key wrapping algorithms, to avoid problems with
64 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
65 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
66 algorithms and include tests cases.
69 *) Add functions to allocate and set the fields of an ECDSA_METHOD
71 [Douglas E. Engert, Steve Henson]
73 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
74 difference in days and seconds between two tm or ASN1_TIME structures.
77 *) Add -rev test option to s_server to just reverse order of characters
78 received by client and send back to server. Also prints an abbreviated
79 summary of the connection parameters.
82 *) New option -brief for s_client and s_server to print out a brief summary
83 of connection parameters.
86 *) Add callbacks for arbitrary TLS extensions.
87 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
89 *) New option -crl_download in several openssl utilities to download CRLs
90 from CRLDP extension in certificates.
93 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
96 *) New function X509_CRL_diff to generate a delta CRL from the difference
97 of two full CRLs. Add support to "crl" utility.
100 *) New functions to set lookup_crls function and to retrieve
101 X509_STORE from X509_STORE_CTX.
104 *) Print out deprecated issuer and subject unique ID fields in
108 *) Extend OCSP I/O functions so they can be used for simple general purpose
109 HTTP as well as OCSP. New wrapper function which can be used to download
110 CRLs using the OCSP API.
113 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
116 *) SSL_CONF* functions. These provide a common framework for application
117 configuration using configuration files or command lines.
120 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
121 message callback and prints the results. Needs compile time option
122 "enable-ssl-trace". New options to s_client and s_server to enable
126 *) New ctrl and macro to retrieve supported points extensions.
127 Print out extension in s_server and s_client.
130 *) New functions to retrieve certificate signature and signature
134 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
138 *) New Suite B modes for TLS code. These use and enforce the requirements
139 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
140 only use Suite B curves. The Suite B modes can be set by using the
141 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
144 *) New chain verification flags for Suite B levels of security. Check
145 algorithms are acceptable when flags are set in X509_verify_cert.
148 *) Make tls1_check_chain return a set of flags indicating checks passed
149 by a certificate chain. Add additional tests to handle client
150 certificates: checks for matching certificate type and issuer name
154 *) If an attempt is made to use a signature algorithm not in the peer
155 preference list abort the handshake. If client has no suitable
156 signature algorithms in response to a certificate request do not
160 *) If server EC tmp key is not in client preference list abort handshake.
163 *) Add support for certificate stores in CERT structure. This makes it
164 possible to have different stores per SSL structure or one store in
165 the parent SSL_CTX. Include distint stores for certificate chain
166 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
167 to build and store a certificate chain in CERT structure: returing
168 an error if the chain cannot be built: this will allow applications
169 to test if a chain is correctly configured.
171 Note: if the CERT based stores are not set then the parent SSL_CTX
172 store is used to retain compatibility with existing behaviour.
176 *) New function ssl_set_client_disabled to set a ciphersuite disabled
177 mask based on the current session, check mask when sending client
178 hello and checking the requested ciphersuite.
181 *) New ctrls to retrieve and set certificate types in a certificate
182 request message. Print out received values in s_client. If certificate
183 types is not set with custom values set sensible values based on
184 supported signature algorithms.
187 *) Support for distinct client and server supported signature algorithms.
190 *) Add certificate callback. If set this is called whenever a certificate
191 is required by client or server. An application can decide which
192 certificate chain to present based on arbitrary criteria: for example
193 supported signature algorithms. Add very simple example to s_server.
194 This fixes many of the problems and restrictions of the existing client
195 certificate callback: for example you can now clear an existing
196 certificate and specify the whole chain.
199 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
200 the certificate can be used for (if anything). Set valid_flags field
201 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
202 to have similar checks in it.
204 Add new "cert_flags" field to CERT structure and include a "strict mode".
205 This enforces some TLS certificate requirements (such as only permitting
206 certificate signature algorithms contained in the supported algorithms
207 extension) which some implementations ignore: this option should be used
208 with caution as it could cause interoperability issues.
211 *) Update and tidy signature algorithm extension processing. Work out
212 shared signature algorithms based on preferences and peer algorithms
213 and print them out in s_client and s_server. Abort handshake if no
214 shared signature algorithms.
217 *) Add new functions to allow customised supported signature algorithms
218 for SSL and SSL_CTX structures. Add options to s_client and s_server
222 *) New function SSL_certs_clear() to delete all references to certificates
223 from an SSL structure. Before this once a certificate had been added
224 it couldn't be removed.
227 *) Integrate hostname, email address and IP address checking with certificate
228 verification. New verify options supporting checking in opensl utility.
231 *) Fixes and wildcard matching support to hostname and email checking
232 functions. Add manual page.
233 [Florian Weimer (Red Hat Product Security Team)]
235 *) New functions to check a hostname email or IP address against a
236 certificate. Add options x509 utility to print results of checks against
240 *) Fix OCSP checking.
241 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
243 *) Initial experimental support for explicitly trusted non-root CAs.
244 OpenSSL still tries to build a complete chain to a root but if an
245 intermediate CA has a trust setting included that is used. The first
246 setting is used: whether to trust (e.g., -addtrust option to the x509
250 *) Add -trusted_first option which attempts to find certificates in the
251 trusted store even if an untrusted chain is also supplied.
254 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
255 platform support for Linux and Android.
258 *) Support for linux-x32, ILP32 environment in x86_64 framework.
261 *) Experimental multi-implementation support for FIPS capable OpenSSL.
262 When in FIPS mode the approved implementations are used as normal,
263 when not in FIPS mode the internal unapproved versions are used instead.
264 This means that the FIPS capable OpenSSL isn't forced to use the
265 (often lower perfomance) FIPS implementations outside FIPS mode.
268 *) Transparently support X9.42 DH parameters when calling
269 PEM_read_bio_DHparameters. This means existing applications can handle
270 the new parameter format automatically.
273 *) Initial experimental support for X9.42 DH parameter format: mainly
274 to support use of 'q' parameter for RFC5114 parameters.
277 *) Add DH parameters from RFC5114 including test data to dhtest.
280 *) Support for automatic EC temporary key parameter selection. If enabled
281 the most preferred EC parameters are automatically used instead of
282 hardcoded fixed parameters. Now a server just has to call:
283 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
284 support ECDH and use the most appropriate parameters.
287 *) Enhance and tidy EC curve and point format TLS extension code. Use
288 static structures instead of allocation if default values are used.
289 New ctrls to set curves we wish to support and to retrieve shared curves.
290 Print out shared curves in s_server. New options to s_server and s_client
291 to set list of supported curves.
294 *) New ctrls to retrieve supported signature algorithms and
295 supported curve values as an array of NIDs. Extend openssl utility
296 to print out received values.
299 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
300 between NIDs and the more common NIST names such as "P-256". Enhance
301 ecparam utility and ECC method to recognise the NIST names for curves.
304 *) Enhance SSL/TLS certificate chain handling to support different
305 chains for each certificate instead of one chain in the parent SSL_CTX.
308 *) Support for fixed DH ciphersuite client authentication: where both
309 server and client use DH certificates with common parameters.
312 *) Support for fixed DH ciphersuites: those requiring DH server
316 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
318 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
319 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
320 X509_CINF_get_signature were reverted post internal team review.
322 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
324 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
325 SRP code can be overrun an internal buffer. Add sanity check that
326 g, A, B < N to SRP code.
328 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
329 Group for discovering this issue.
333 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
334 TLS 1.0 instead of higher protocol versions when the ClientHello message
335 is badly fragmented. This allows a man-in-the-middle attacker to force a
336 downgrade to TLS 1.0 even if both the server and the client support a
337 higher protocol version, by modifying the client's TLS records.
339 Thanks to David Benjamin and Adam Langley (Google) for discovering and
340 researching this issue.
344 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
345 to a denial of service attack. A malicious server can crash the client
346 with a null pointer dereference (read) by specifying an anonymous (EC)DH
347 ciphersuite and sending carefully crafted handshake messages.
349 Thanks to Felix Gröbert (Google) for discovering and researching this
354 *) By sending carefully crafted DTLS packets an attacker could cause openssl
355 to leak memory. This can be exploited through a Denial of Service attack.
356 Thanks to Adam Langley for discovering and researching this issue.
360 *) An attacker can force openssl to consume large amounts of memory whilst
361 processing DTLS handshake messages. This can be exploited through a
362 Denial of Service attack.
363 Thanks to Adam Langley for discovering and researching this issue.
367 *) An attacker can force an error condition which causes openssl to crash
368 whilst processing DTLS packets due to memory being freed twice. This
369 can be exploited through a Denial of Service attack.
370 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
375 *) If a multithreaded client connects to a malicious server using a resumed
376 session and the server sends an ec point format extension it could write
377 up to 255 bytes to freed memory.
379 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
384 *) A malicious server can crash an OpenSSL client with a null pointer
385 dereference (read) by specifying an SRP ciphersuite even though it was not
386 properly negotiated with the client. This can be exploited through a
387 Denial of Service attack.
389 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
390 discovering and researching this issue.
394 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
395 X509_name_oneline, X509_name_print_ex et al. to leak some information
396 from the stack. Applications may be affected if they echo pretty printing
397 output to the attacker.
399 Thanks to Ivan Fratric (Google) for discovering this issue.
401 [Emilia Käsper, and Steve Henson]
403 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
404 for corner cases. (Certain input points at infinity could lead to
405 bogus results, with non-infinity inputs mapped to infinity too.)
408 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
410 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
411 handshake can force the use of weak keying material in OpenSSL
412 SSL/TLS clients and servers.
414 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
415 researching this issue. (CVE-2014-0224)
416 [KIKUCHI Masashi, Steve Henson]
418 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
419 OpenSSL DTLS client the code can be made to recurse eventually crashing
422 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
424 [Imre Rad, Steve Henson]
426 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
427 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
428 client or server. This is potentially exploitable to run arbitrary
429 code on a vulnerable client or server.
431 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
432 [Jüri Aedla, Steve Henson]
434 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
435 are subject to a denial of service attack.
437 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
438 this issue. (CVE-2014-3470)
439 [Felix Gröbert, Ivan Fratric, Steve Henson]
441 *) Harmonize version and its documentation. -f flag is used to display
443 [mancha <mancha1@zoho.com>]
445 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
447 [mancha <mancha1@zoho.com>]
449 *) Fix some double frees. These are not thought to be exploitable.
450 [mancha <mancha1@zoho.com>]
452 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
454 *) A missing bounds check in the handling of the TLS heartbeat extension
455 can be used to reveal up to 64k of memory to a connected client or
458 Thanks for Neel Mehta of Google Security for discovering this bug and to
459 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
460 preparing the fix (CVE-2014-0160)
461 [Adam Langley, Bodo Moeller]
463 *) Fix for the attack described in the paper "Recovering OpenSSL
464 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
465 by Yuval Yarom and Naomi Benger. Details can be obtained from:
466 http://eprint.iacr.org/2014/140
468 Thanks to Yuval Yarom and Naomi Benger for discovering this
469 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
470 [Yuval Yarom and Naomi Benger]
472 *) TLS pad extension: draft-agl-tls-padding-03
474 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
475 TLS client Hello record length value would otherwise be > 255 and
476 less that 512 pad with a dummy extension containing zeroes so it
477 is at least 512 bytes long.
479 [Adam Langley, Steve Henson]
481 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
483 *) Fix for TLS record tampering bug. A carefully crafted invalid
484 handshake could crash OpenSSL with a NULL pointer exception.
485 Thanks to Anton Johansson for reporting this issues.
488 *) Keep original DTLS digest and encryption contexts in retransmission
489 structures so we can use the previous session parameters if they need
490 to be resent. (CVE-2013-6450)
493 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
494 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
495 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
496 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
497 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
498 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
499 [Rob Stradling, Adam Langley]
501 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
503 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
504 supporting platforms or when small records were transferred.
505 [Andy Polyakov, Steve Henson]
507 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
509 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
511 This addresses the flaw in CBC record processing discovered by
512 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
513 at: http://www.isg.rhul.ac.uk/tls/
515 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
516 Security Group at Royal Holloway, University of London
517 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
518 Emilia Käsper for the initial patch.
520 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
522 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
523 ciphersuites which can be exploited in a denial of service attack.
524 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
525 and detecting this bug and to Wolfgang Ettlinger
526 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
530 *) Return an error when checking OCSP signatures when key is NULL.
531 This fixes a DoS attack. (CVE-2013-0166)
534 *) Make openssl verify return errors.
535 [Chris Palmer <palmer@google.com> and Ben Laurie]
537 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
538 the right response is stapled. Also change SSL_get_certificate()
539 so it returns the certificate actually sent.
540 See http://rt.openssl.org/Ticket/Display.html?id=2836.
541 [Rob Stradling <rob.stradling@comodo.com>]
543 *) Fix possible deadlock when decoding public keys.
546 *) Don't use TLS 1.0 record version number in initial client hello
550 Changes between 1.0.1b and 1.0.1c [10 May 2012]
552 *) Sanity check record length before skipping explicit IV in TLS
553 1.2, 1.1 and DTLS to fix DoS attack.
555 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
556 fuzzing as a service testing platform.
560 *) Initialise tkeylen properly when encrypting CMS messages.
561 Thanks to Solar Designer of Openwall for reporting this issue.
564 *) In FIPS mode don't try to use composite ciphers as they are not
568 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
570 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
571 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
572 mean any application compiled against OpenSSL 1.0.0 headers setting
573 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
574 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
575 0x10000000L Any application which was previously compiled against
576 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
577 will need to be recompiled as a result. Letting be results in
578 inability to disable specifically TLS 1.1 and in client context,
579 in unlike event, limit maximum offered version to TLS 1.0 [see below].
582 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
583 disable just protocol X, but all protocols above X *if* there are
584 protocols *below* X still enabled. In more practical terms it means
585 that if application wants to disable TLS1.0 in favor of TLS1.1 and
586 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
587 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
591 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
593 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
594 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
595 in CRYPTO_realloc_clean.
597 Thanks to Tavis Ormandy, Google Security Team, for discovering this
598 issue and to Adam Langley <agl@chromium.org> for fixing it.
600 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
602 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
605 *) Workarounds for some broken servers that "hang" if a client hello
606 record length exceeds 255 bytes.
608 1. Do not use record version number > TLS 1.0 in initial client
609 hello: some (but not all) hanging servers will now work.
610 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
611 the number of ciphers sent in the client hello. This should be
612 set to an even number, such as 50, for example by passing:
613 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
614 Most broken servers should now work.
615 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
616 TLS 1.2 client support entirely.
619 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
622 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
624 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
625 STRING form instead of a DigestInfo.
628 *) The format used for MDC2 RSA signatures is inconsistent between EVP
629 and the RSA_sign/RSA_verify functions. This was made more apparent when
630 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
631 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
632 the correct format in RSA_verify so both forms transparently work.
635 *) Some servers which support TLS 1.0 can choke if we initially indicate
636 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
637 encrypted premaster secret. As a workaround use the maximum pemitted
638 client version in client hello, this should keep such servers happy
639 and still work with previous versions of OpenSSL.
642 *) Add support for TLS/DTLS heartbeats.
643 [Robin Seggelmann <seggelmann@fh-muenster.de>]
645 *) Add support for SCTP.
646 [Robin Seggelmann <seggelmann@fh-muenster.de>]
648 *) Improved PRNG seeding for VOS.
649 [Paul Green <Paul.Green@stratus.com>]
651 *) Extensive assembler packs updates, most notably:
653 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
654 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
655 - x86_64: bit-sliced AES implementation;
656 - ARM: NEON support, contemporary platforms optimizations;
657 - s390x: z196 support;
658 - *: GHASH and GF(2^m) multiplication implementations;
662 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
663 (removal of unnecessary code)
664 [Peter Sylvester <peter.sylvester@edelweb.fr>]
666 *) Add TLS key material exporter from RFC 5705.
669 *) Add DTLS-SRTP negotiation from RFC 5764.
672 *) Add Next Protocol Negotiation,
673 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
674 disabled with a no-npn flag to config or Configure. Code donated
676 [Adam Langley <agl@google.com> and Ben Laurie]
678 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
679 NIST-P256, NIST-P521, with constant-time single point multiplication on
680 typical inputs. Compiler support for the nonstandard type __uint128_t is
681 required to use this (present in gcc 4.4 and later, for 64-bit builds).
682 Code made available under Apache License version 2.0.
684 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
685 line to include this in your build of OpenSSL, and run "make depend" (or
686 "make update"). This enables the following EC_METHODs:
688 EC_GFp_nistp224_method()
689 EC_GFp_nistp256_method()
690 EC_GFp_nistp521_method()
692 EC_GROUP_new_by_curve_name() will automatically use these (while
693 EC_GROUP_new_curve_GFp() currently prefers the more flexible
695 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
697 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
698 all platforms. Move ssize_t definition from e_os.h to the public
699 header file e_os2.h as it now appears in public header file cms.h
702 *) New -sigopt option to the ca, req and x509 utilities. Additional
703 signature parameters can be passed using this option and in
707 *) Add RSA PSS signing function. This will generate and set the
708 appropriate AlgorithmIdentifiers for PSS based on those in the
709 corresponding EVP_MD_CTX structure. No application support yet.
712 *) Support for companion algorithm specific ASN1 signing routines.
713 New function ASN1_item_sign_ctx() signs a pre-initialised
714 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
715 the appropriate parameters.
718 *) Add new algorithm specific ASN1 verification initialisation function
719 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
720 handling will be the same no matter what EVP_PKEY_METHOD is used.
721 Add a PSS handler to support verification of PSS signatures: checked
722 against a number of sample certificates.
725 *) Add signature printing for PSS. Add PSS OIDs.
726 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
728 *) Add algorithm specific signature printing. An individual ASN1 method
729 can now print out signatures instead of the standard hex dump.
731 More complex signatures (e.g. PSS) can print out more meaningful
732 information. Include DSA version that prints out the signature
736 *) Password based recipient info support for CMS library: implementing
740 *) Split password based encryption into PBES2 and PBKDF2 functions. This
741 neatly separates the code into cipher and PBE sections and is required
742 for some algorithms that split PBES2 into separate pieces (such as
746 *) Session-handling fixes:
747 - Fix handling of connections that are resuming with a session ID,
748 but also support Session Tickets.
749 - Fix a bug that suppressed issuing of a new ticket if the client
750 presented a ticket with an expired session.
751 - Try to set the ticket lifetime hint to something reasonable.
752 - Make tickets shorter by excluding irrelevant information.
753 - On the client side, don't ignore renewed tickets.
754 [Adam Langley, Bodo Moeller (Google)]
756 *) Fix PSK session representation.
759 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
761 This work was sponsored by Intel.
764 *) Add GCM support to TLS library. Some custom code is needed to split
765 the IV between the fixed (from PRF) and explicit (from TLS record)
766 portions. This adds all GCM ciphersuites supported by RFC5288 and
767 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
768 add a special AESGCM string for GCM only.
771 *) Expand range of ctrls for AES GCM. Permit setting invocation
772 field on decrypt and retrieval of invocation field only on encrypt.
775 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
776 As required by RFC5289 these ciphersuites cannot be used if for
777 versions of TLS earlier than 1.2.
780 *) For FIPS capable OpenSSL interpret a NULL default public key method
781 as unset and return the appopriate default but do *not* set the default.
782 This means we can return the appopriate method in applications that
783 swicth between FIPS and non-FIPS modes.
786 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
787 ENGINE is used then we cannot handle that in the FIPS module so we
788 keep original code iff non-FIPS operations are allowed.
791 *) Add -attime option to openssl utilities.
792 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
794 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
797 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
798 FIPS EC methods unconditionally for now.
801 *) New build option no-ec2m to disable characteristic 2 code.
804 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
805 all cases can be covered as some introduce binary incompatibilities.
808 *) Redirect RSA operations to FIPS module including keygen,
809 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
812 *) Add similar low level API blocking to ciphers.
815 *) Low level digest APIs are not approved in FIPS mode: any attempt
816 to use these will cause a fatal error. Applications that *really* want
817 to use them can use the private_* version instead.
820 *) Redirect cipher operations to FIPS module for FIPS builds.
823 *) Redirect digest operations to FIPS module for FIPS builds.
826 *) Update build system to add "fips" flag which will link in fipscanister.o
827 for static and shared library builds embedding a signature if needed.
830 *) Output TLS supported curves in preference order instead of numerical
831 order. This is currently hardcoded for the highest order curves first.
832 This should be configurable so applications can judge speed vs strength.
835 *) Add TLS v1.2 server support for client authentication.
838 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
842 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
843 FIPS modules versions.
846 *) Add TLS v1.2 client side support for client authentication. Keep cache
847 of handshake records longer as we don't know the hash algorithm to use
848 until after the certificate request message is received.
851 *) Initial TLS v1.2 client support. Add a default signature algorithms
852 extension including all the algorithms we support. Parse new signature
853 format in client key exchange. Relax some ECC signing restrictions for
854 TLS v1.2 as indicated in RFC5246.
857 *) Add server support for TLS v1.2 signature algorithms extension. Switch
858 to new signature format when needed using client digest preference.
859 All server ciphersuites should now work correctly in TLS v1.2. No client
860 support yet and no support for client certificates.
863 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
864 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
865 ciphersuites. At present only RSA key exchange ciphersuites work with
866 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
867 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
868 and version checking.
871 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
872 with this defined it will not be affected by any changes to ssl internal
873 structures. Add several utility functions to allow openssl application
874 to work with OPENSSL_NO_SSL_INTERN defined.
878 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
880 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
883 *) Permit abbreviated handshakes when renegotiating using the function
884 SSL_renegotiate_abbreviated().
885 [Robin Seggelmann <seggelmann@fh-muenster.de>]
887 *) Add call to ENGINE_register_all_complete() to
888 ENGINE_load_builtin_engines(), so some implementations get used
889 automatically instead of needing explicit application support.
892 *) Add support for TLS key exporter as described in RFC5705.
893 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
895 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
896 a few changes are required:
898 Add SSL_OP_NO_TLSv1_1 flag.
900 Update version checking logic to handle version 1.1.
901 Add explicit IV handling (ported from DTLS code).
902 Add command line options to s_client/s_server.
905 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
907 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
909 This addresses the flaw in CBC record processing discovered by
910 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
911 at: http://www.isg.rhul.ac.uk/tls/
913 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
914 Security Group at Royal Holloway, University of London
915 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
916 Emilia Käsper for the initial patch.
918 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
920 *) Return an error when checking OCSP signatures when key is NULL.
921 This fixes a DoS attack. (CVE-2013-0166)
924 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
925 the right response is stapled. Also change SSL_get_certificate()
926 so it returns the certificate actually sent.
927 See http://rt.openssl.org/Ticket/Display.html?id=2836.
929 [Rob Stradling <rob.stradling@comodo.com>]
931 *) Fix possible deadlock when decoding public keys.
934 Changes between 1.0.0i and 1.0.0j [10 May 2012]
936 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
939 *) Sanity check record length before skipping explicit IV in DTLS
942 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
943 fuzzing as a service testing platform.
947 *) Initialise tkeylen properly when encrypting CMS messages.
948 Thanks to Solar Designer of Openwall for reporting this issue.
951 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
953 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
954 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
955 in CRYPTO_realloc_clean.
957 Thanks to Tavis Ormandy, Google Security Team, for discovering this
958 issue and to Adam Langley <agl@chromium.org> for fixing it.
960 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
962 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
964 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
965 in CMS and PKCS7 code. When RSA decryption fails use a random key for
966 content decryption and always return the same error. Note: this attack
967 needs on average 2^20 messages so it only affects automated senders. The
968 old behaviour can be reenabled in the CMS code by setting the
969 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
970 an MMA defence is not necessary.
971 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
972 this issue. (CVE-2012-0884)
975 *) Fix CVE-2011-4619: make sure we really are receiving a
976 client hello before rejecting multiple SGC restarts. Thanks to
977 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
980 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
982 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
983 Thanks to Antonio Martin, Enterprise Secure Access Research and
984 Development, Cisco Systems, Inc. for discovering this bug and
985 preparing a fix. (CVE-2012-0050)
988 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
990 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
991 of the Vaudenay padding oracle attack on CBC mode encryption
992 which enables an efficient plaintext recovery attack against
993 the OpenSSL implementation of DTLS. Their attack exploits timing
994 differences arising during decryption processing. A research
995 paper describing this attack can be found at:
996 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
997 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
998 Security Group at Royal Holloway, University of London
999 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1000 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1001 for preparing the fix. (CVE-2011-4108)
1002 [Robin Seggelmann, Michael Tuexen]
1004 *) Clear bytes used for block padding of SSL 3.0 records.
1006 [Adam Langley (Google)]
1008 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1009 Kadianakis <desnacked@gmail.com> for discovering this issue and
1010 Adam Langley for preparing the fix. (CVE-2011-4619)
1011 [Adam Langley (Google)]
1013 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1014 [Andrey Kulikov <amdeich@gmail.com>]
1016 *) Prevent malformed RFC3779 data triggering an assertion failure.
1017 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1018 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1019 [Rob Austein <sra@hactrn.net>]
1021 *) Improved PRNG seeding for VOS.
1022 [Paul Green <Paul.Green@stratus.com>]
1024 *) Fix ssl_ciph.c set-up race.
1025 [Adam Langley (Google)]
1027 *) Fix spurious failures in ecdsatest.c.
1028 [Emilia Käsper (Google)]
1030 *) Fix the BIO_f_buffer() implementation (which was mixing different
1031 interpretations of the '..._len' fields).
1032 [Adam Langley (Google)]
1034 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1035 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1036 threads won't reuse the same blinding coefficients.
1038 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1039 lock to call BN_BLINDING_invert_ex, and avoids one use of
1040 BN_BLINDING_update for each BN_BLINDING structure (previously,
1041 the last update always remained unused).
1042 [Emilia Käsper (Google)]
1044 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1045 [Bob Buckholz (Google)]
1047 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1049 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1050 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1051 [Kaspar Brand <ossl@velox.ch>]
1053 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1054 for multi-threaded use of ECDH. (CVE-2011-3210)
1055 [Adam Langley (Google)]
1057 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1060 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1061 signature public key algorithm by using OID xref utilities instead.
1062 Before this you could only use some ECC ciphersuites with SHA1 only.
1065 *) Add protection against ECDSA timing attacks as mentioned in the paper
1066 by Billy Bob Brumley and Nicola Tuveri, see:
1068 http://eprint.iacr.org/2011/232.pdf
1070 [Billy Bob Brumley and Nicola Tuveri]
1072 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1074 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1075 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1077 *) Fix bug in string printing code: if *any* escaping is enabled we must
1078 escape the escape character (backslash) or the resulting string is
1082 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1084 *) Disable code workaround for ancient and obsolete Netscape browsers
1085 and servers: an attacker can use it in a ciphersuite downgrade attack.
1086 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1089 *) Fixed J-PAKE implementation error, originally discovered by
1090 Sebastien Martini, further info and confirmation from Stefan
1091 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1094 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1096 *) Fix extension code to avoid race conditions which can result in a buffer
1097 overrun vulnerability: resumed sessions must not be modified as they can
1098 be shared by multiple threads. CVE-2010-3864
1101 *) Fix WIN32 build system to correctly link an ENGINE directory into
1105 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1107 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1109 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1111 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1113 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1114 context. The operation can be customised via the ctrl mechanism in
1115 case ENGINEs want to include additional functionality.
1118 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1121 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1122 output hashes compatible with older versions of OpenSSL.
1123 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1125 *) Fix compression algorithm handling: if resuming a session use the
1126 compression algorithm of the resumed session instead of determining
1127 it from client hello again. Don't allow server to change algorithm.
1130 *) Add load_crls() function to apps tidying load_certs() too. Add option
1131 to verify utility to allow additional CRLs to be included.
1134 *) Update OCSP request code to permit adding custom headers to the request:
1135 some responders need this.
1138 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1140 [Julia Lawall <julia@diku.dk>]
1142 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1143 needlessly dereferenced structures, used obsolete functions and
1144 didn't handle all updated verify codes correctly.
1147 *) Disable MD2 in the default configuration.
1150 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1151 indicate the initial BIO being pushed or popped. This makes it possible
1152 to determine whether the BIO is the one explicitly called or as a result
1153 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1154 it handles reference counts correctly and doesn't zero out the I/O bio
1155 when it is not being explicitly popped. WARNING: applications which
1156 included workarounds for the old buggy behaviour will need to be modified
1157 or they could free up already freed BIOs.
1160 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1161 renaming to all platforms (within the 0.9.8 branch, this was
1162 done conditionally on Netware platforms to avoid a name clash).
1163 [Guenter <lists@gknw.net>]
1165 *) Add ECDHE and PSK support to DTLS.
1166 [Michael Tuexen <tuexen@fh-muenster.de>]
1168 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1172 *) Add "missing" function EVP_MD_flags() (without this the only way to
1173 retrieve a digest flags is by accessing the structure directly. Update
1174 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1175 or cipher is registered as in the "from" argument. Print out all
1176 registered digests in the dgst usage message instead of manually
1177 attempting to work them out.
1180 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1181 this allows the use of compression and extensions. Change default cipher
1182 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1183 by default unless an application cipher string requests it.
1186 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1187 key ids to find matching certificates and keys but some PKCS#12 files
1188 don't follow the (somewhat unwritten) rules and this strategy fails.
1189 Now just gather all certificates together and the first private key
1190 then look for the first certificate that matches the key.
1193 *) Support use of registered digest and cipher names for dgst and cipher
1194 commands instead of having to add each one as a special case. So now
1201 openssl dgst -sha256 foo
1203 and this works for ENGINE based algorithms too.
1207 *) Update Gost ENGINE to support parameter files.
1208 [Victor B. Wagner <vitus@cryptocom.ru>]
1210 *) Support GeneralizedTime in ca utility.
1211 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1213 *) Enhance the hash format used for certificate directory links. The new
1214 form uses the canonical encoding (meaning equivalent names will work
1215 even if they aren't identical) and uses SHA1 instead of MD5. This form
1216 is incompatible with the older format and as a result c_rehash should
1217 be used to rebuild symbolic links.
1220 *) Make PKCS#8 the default write format for private keys, replacing the
1221 traditional format. This form is standardised, more secure and doesn't
1222 include an implicit MD5 dependency.
1225 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1226 committed to OpenSSL should pass this lot as a minimum.
1229 *) Add session ticket override functionality for use by EAP-FAST.
1230 [Jouni Malinen <j@w1.fi>]
1232 *) Modify HMAC functions to return a value. Since these can be implemented
1233 in an ENGINE errors can occur.
1236 *) Type-checked OBJ_bsearch_ex.
1239 *) Type-checked OBJ_bsearch. Also some constification necessitated
1240 by type-checking. Still to come: TXT_DB, bsearch(?),
1241 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1245 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1246 seconds to a tm structure directly, instead of going through OS
1247 specific date routines. This avoids any issues with OS routines such
1248 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1249 and X509_time_adj_ex() to cover the extended range. The existing
1250 X509_time_adj() is still usable and will no longer have any date issues.
1253 *) Delta CRL support. New use deltas option which will attempt to locate
1254 and search any appropriate delta CRLs available.
1256 This work was sponsored by Google.
1259 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1260 code and add additional score elements. Validate alternate CRL paths
1261 as part of the CRL checking and indicate a new error "CRL path validation
1262 error" in this case. Applications wanting additional details can use
1263 the verify callback and check the new "parent" field. If this is not
1264 NULL CRL path validation is taking place. Existing applications wont
1265 see this because it requires extended CRL support which is off by
1268 This work was sponsored by Google.
1271 *) Support for freshest CRL extension.
1273 This work was sponsored by Google.
1276 *) Initial indirect CRL support. Currently only supported in the CRLs
1277 passed directly and not via lookup. Process certificate issuer
1278 CRL entry extension and lookup CRL entries by bother issuer name
1279 and serial number. Check and process CRL issuer entry in IDP extension.
1281 This work was sponsored by Google.
1284 *) Add support for distinct certificate and CRL paths. The CRL issuer
1285 certificate is validated separately in this case. Only enabled if
1286 an extended CRL support flag is set: this flag will enable additional
1287 CRL functionality in future.
1289 This work was sponsored by Google.
1292 *) Add support for policy mappings extension.
1294 This work was sponsored by Google.
1297 *) Fixes to pathlength constraint, self issued certificate handling,
1298 policy processing to align with RFC3280 and PKITS tests.
1300 This work was sponsored by Google.
1303 *) Support for name constraints certificate extension. DN, email, DNS
1304 and URI types are currently supported.
1306 This work was sponsored by Google.
1309 *) To cater for systems that provide a pointer-based thread ID rather
1310 than numeric, deprecate the current numeric thread ID mechanism and
1311 replace it with a structure and associated callback type. This
1312 mechanism allows a numeric "hash" to be extracted from a thread ID in
1313 either case, and on platforms where pointers are larger than 'long',
1314 mixing is done to help ensure the numeric 'hash' is usable even if it
1315 can't be guaranteed unique. The default mechanism is to use "&errno"
1316 as a pointer-based thread ID to distinguish between threads.
1318 Applications that want to provide their own thread IDs should now use
1319 CRYPTO_THREADID_set_callback() to register a callback that will call
1320 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1322 Note that ERR_remove_state() is now deprecated, because it is tied
1323 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1324 to free the current thread's error state should be replaced by
1325 ERR_remove_thread_state(NULL).
1327 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1328 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1329 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1330 application was previously providing a numeric thread callback that
1331 was inappropriate for distinguishing threads, then uniqueness might
1332 have been obtained with &errno that happened immediately in the
1333 intermediate development versions of OpenSSL; this is no longer the
1334 case, the numeric thread callback will now override the automatic use
1336 [Geoff Thorpe, with help from Bodo Moeller]
1338 *) Initial support for different CRL issuing certificates. This covers a
1339 simple case where the self issued certificates in the chain exist and
1340 the real CRL issuer is higher in the existing chain.
1342 This work was sponsored by Google.
1345 *) Removed effectively defunct crypto/store from the build.
1348 *) Revamp of STACK to provide stronger type-checking. Still to come:
1349 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1350 ASN1_STRING, CONF_VALUE.
1353 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1354 RAM on SSL connections. This option can save about 34k per idle SSL.
1357 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1358 STACK, TXT_DB, bsearch, qsort.
1361 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1362 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1363 support for data, signedData, compressedData, digestedData and
1364 encryptedData, envelopedData types included. Scripts to check against
1365 RFC4134 examples draft and interop and consistency checks of many
1366 content types and variants.
1369 *) Add options to enc utility to support use of zlib compression BIO.
1372 *) Extend mk1mf to support importing of options and assembly language
1373 files from Configure script, currently only included in VC-WIN32.
1374 The assembly language rules can now optionally generate the source
1375 files from the associated perl scripts.
1378 *) Implement remaining functionality needed to support GOST ciphersuites.
1379 Interop testing has been performed using CryptoPro implementations.
1380 [Victor B. Wagner <vitus@cryptocom.ru>]
1382 *) s390x assembler pack.
1385 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1389 *) Implement Opaque PRF Input TLS extension as specified in
1390 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1391 official specification yet and no extension type assignment by
1392 IANA exists, this extension (for now) will have to be explicitly
1393 enabled when building OpenSSL by providing the extension number
1394 to use. For example, specify an option
1396 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1398 to the "config" or "Configure" script to enable the extension,
1399 assuming extension number 0x9527 (which is a completely arbitrary
1400 and unofficial assignment based on the MD5 hash of the Internet
1401 Draft). Note that by doing so, you potentially lose
1402 interoperability with other TLS implementations since these might
1403 be using the same extension number for other purposes.
1405 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1406 opaque PRF input value to use in the handshake. This will create
1407 an interal copy of the length-'len' string at 'src', and will
1408 return non-zero for success.
1410 To get more control and flexibility, provide a callback function
1413 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1414 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1418 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1421 Callback function 'cb' will be called in handshakes, and is
1422 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1423 Argument 'arg' is for application purposes (the value as given to
1424 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1425 be provided to the callback function). The callback function
1426 has to return non-zero to report success: usually 1 to use opaque
1427 PRF input just if possible, or 2 to enforce use of the opaque PRF
1428 input. In the latter case, the library will abort the handshake
1429 if opaque PRF input is not successfully negotiated.
1431 Arguments 'peerinput' and 'len' given to the callback function
1432 will always be NULL and 0 in the case of a client. A server will
1433 see the client's opaque PRF input through these variables if
1434 available (NULL and 0 otherwise). Note that if the server
1435 provides an opaque PRF input, the length must be the same as the
1436 length of the client's opaque PRF input.
1438 Note that the callback function will only be called when creating
1439 a new session (session resumption can resume whatever was
1440 previously negotiated), and will not be called in SSL 2.0
1441 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1442 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1443 for applications that need to enforce opaque PRF input.
1447 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1450 [Victor B. Wagner <vitus@cryptocom.ru>]
1452 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1453 RFC4507bis. The encrypted ticket format is an encrypted encoded
1454 SSL_SESSION structure, that way new session features are automatically
1457 If a client application caches session in an SSL_SESSION structure
1458 support is transparent because tickets are now stored in the encoded
1461 The SSL_CTX structure automatically generates keys for ticket
1462 protection in servers so again support should be possible
1463 with no application modification.
1465 If a client or server wishes to disable RFC4507 support then the option
1466 SSL_OP_NO_TICKET can be set.
1468 Add a TLS extension debugging callback to allow the contents of any client
1469 or server extensions to be examined.
1471 This work was sponsored by Google.
1474 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1475 OpenSSL should now compile cleanly on gcc 4.2
1476 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1478 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1479 support including streaming MAC support: this is required for GOST
1480 ciphersuite support.
1481 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1483 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1484 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1485 to output in BER and PEM format.
1488 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1489 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1490 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1491 ENGINE support for HMAC keys which are unextractable. New -mac and
1492 -macopt options to dgst utility.
1495 *) New option -sigopt to dgst utility. Update dgst to use
1496 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1497 alternative signing paramaters such as X9.31 or PSS in the dgst
1501 *) Change ssl_cipher_apply_rule(), the internal function that does
1502 the work each time a ciphersuite string requests enabling
1503 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1504 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1505 the order of disabled ciphersuites such that those ciphersuites
1506 that most recently went from enabled to disabled not only stay
1507 in order with respect to each other, but also have higher priority
1508 than other disabled ciphersuites the next time ciphersuites are
1511 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1512 the same ciphersuites as with "HIGH" alone, but in a specific
1513 order where the PSK ciphersuites come first (since they are the
1514 most recently disabled ciphersuites when "HIGH" is parsed).
1516 Also, change ssl_create_cipher_list() (using this new
1517 funcionality) such that between otherwise identical
1518 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1522 *) Change ssl_create_cipher_list() so that it automatically
1523 arranges the ciphersuites in reasonable order before starting
1524 to process the rule string. Thus, the definition for "DEFAULT"
1525 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1526 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1527 This makes it much easier to arrive at a reasonable default order
1528 in applications for which anonymous ciphers are OK (meaning
1529 that you can't actually use DEFAULT).
1530 [Bodo Moeller; suggested by Victor Duchovni]
1532 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1533 processing) into multiple integers instead of setting
1534 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1535 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1536 (These masks as well as the individual bit definitions are hidden
1537 away into the non-exported interface ssl/ssl_locl.h, so this
1538 change to the definition of the SSL_CIPHER structure shouldn't
1539 affect applications.) This give us more bits for each of these
1540 categories, so there is no longer a need to coagulate AES128 and
1541 AES256 into a single algorithm bit, and to coagulate Camellia128
1542 and Camellia256 into a single algorithm bit, which has led to all
1545 Thus, among other things, the kludge introduced in 0.9.7m and
1546 0.9.8e for masking out AES256 independently of AES128 or masking
1547 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1549 With the change, we also introduce new ciphersuite aliases that
1550 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1554 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1555 Use the leftmost N bytes of the signature input if the input is
1556 larger than the prime q (with N being the size in bytes of q).
1559 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1560 it yet and it is largely untested.
1563 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1566 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1567 some compilers (gcc 4.2 and later) reject their use. Safestack is
1568 reimplemented. Update ASN1 to avoid use of legacy functions.
1571 *) Win32/64 targets are linked with Winsock2.
1574 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1575 to external functions. This can be used to increase CRL handling
1576 efficiency especially when CRLs are very large by (for example) storing
1577 the CRL revoked certificates in a database.
1580 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1581 new CRLs added to a directory can be used. New command line option
1582 -verify_return_error to s_client and s_server. This causes real errors
1583 to be returned by the verify callback instead of carrying on no matter
1584 what. This reflects the way a "real world" verify callback would behave.
1587 *) GOST engine, supporting several GOST algorithms and public key formats.
1588 Kindly donated by Cryptocom.
1591 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1592 partitioned by DP are handled but no indirect CRL or reason partitioning
1593 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1594 selected via a scoring technique which handles IDP and AKID in CRLs.
1597 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1598 will ultimately be used for all verify operations: this will remove the
1599 X509_STORE dependency on certificate verification and allow alternative
1600 lookup methods. X509_STORE based implementations of these two callbacks.
1603 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1604 Modify get_crl() to find a valid (unexpired) CRL if possible.
1607 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1608 this would be called X509_CRL_cmp() but that name is already used by
1609 a function that just compares CRL issuer names. Cache several CRL
1610 extensions in X509_CRL structure and cache CRLDP in X509.
1613 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1614 this maps equivalent X509_NAME structures into a consistent structure.
1615 Name comparison can then be performed rapidly using memcmp().
1618 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1622 *) Allow digests to supply their own micalg string for S/MIME type using
1623 the ctrl EVP_MD_CTRL_MICALG.
1626 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1627 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1628 ctrl. It can then customise the structure before and/or after signing
1632 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1633 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1634 to free up any added signature OIDs.
1637 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1638 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1639 digest and cipher tables. New options added to openssl utility:
1640 list-message-digest-algorithms and list-cipher-algorithms.
1643 *) Change the array representation of binary polynomials: the list
1644 of degrees of non-zero coefficients is now terminated with -1.
1645 Previously it was terminated with 0, which was also part of the
1646 value; thus, the array representation was not applicable to
1647 polynomials where t^0 has coefficient zero. This change makes
1648 the array representation useful in a more general context.
1651 *) Various modifications and fixes to SSL/TLS cipher string
1652 handling. For ECC, the code now distinguishes between fixed ECDH
1653 with RSA certificates on the one hand and with ECDSA certificates
1654 on the other hand, since these are separate ciphersuites. The
1655 unused code for Fortezza ciphersuites has been removed.
1657 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1658 (not "ECDHE"). For consistency with the code for DH
1659 certificates, use of ECDH certificates is now considered ECDH
1660 authentication, not RSA or ECDSA authentication (the latter is
1661 merely the CA's signing algorithm and not actively used in the
1664 The temporary ciphersuite alias "ECCdraft" is no longer
1665 available, and ECC ciphersuites are no longer excluded from "ALL"
1666 and "DEFAULT". The following aliases now exist for RFC 4492
1667 ciphersuites, most of these by analogy with the DH case:
1669 kECDHr - ECDH cert, signed with RSA
1670 kECDHe - ECDH cert, signed with ECDSA
1671 kECDH - ECDH cert (signed with either RSA or ECDSA)
1672 kEECDH - ephemeral ECDH
1673 ECDH - ECDH cert or ephemeral ECDH
1679 AECDH - anonymous ECDH
1680 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1684 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1685 Use correct micalg parameters depending on digest(s) in signed message.
1688 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1689 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1692 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1693 an engine to register a method. Add ENGINE lookups for methods and
1694 functional reference processing.
1697 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1698 EVP_{Sign,Verify}* which allow an application to customise the signature
1702 *) New -resign option to smime utility. This adds one or more signers
1703 to an existing PKCS#7 signedData structure. Also -md option to use an
1704 alternative message digest algorithm for signing.
1707 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1708 create PKCS7 structures containing multiple signers. Update smime
1709 application to support multiple signers.
1712 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1716 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1717 Reorganize PBE internals to lookup from a static table using NIDs,
1718 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1719 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1720 PRF which will be automatically used with PBES2.
1723 *) Replace the algorithm specific calls to generate keys in "req" with the
1727 *) Update PKCS#7 enveloped data routines to use new API. This is now
1728 supported by any public key method supporting the encrypt operation. A
1729 ctrl is added to allow the public key algorithm to examine or modify
1730 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1734 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1735 a default digest type to use. In most cases this will be SHA1 but some
1736 algorithms (such as GOST) need to specify an alternative digest. The
1737 return value indicates how strong the prefernce is 1 means optional and
1738 2 is mandatory (that is it is the only supported type). Modify
1739 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1740 use the default md. Update openssl utilities to use the default digest
1741 type for signing if it is not explicitly indicated.
1744 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1745 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1746 signing method from the key type. This effectively removes the link
1747 between digests and public key types.
1750 *) Add an OID cross reference table and utility functions. Its purpose is to
1751 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1752 rsaEncryption. This will allow some of the algorithm specific hackery
1753 needed to use the correct OID to be removed.
1756 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1757 structures for PKCS7_sign(). They are now set up by the relevant public
1761 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1764 *) Add support for key derivation (agreement) in the API, DH method and
1768 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1769 public and private key formats. As a side effect these add additional
1770 command line functionality not previously available: DSA signatures can be
1771 generated and verified using pkeyutl and DH key support and generation in
1776 [Oliver Tappe <zooey@hirschkaefer.de>]
1778 *) New make target "install_html_docs" installs HTML renditions of the
1780 [Oliver Tappe <zooey@hirschkaefer.de>]
1782 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1783 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1784 support key and parameter generation and add initial key generation
1785 functionality for RSA.
1788 *) Add functions for main EVP_PKEY_method operations. The undocumented
1789 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1790 EVP_PKEY_{encrypt,decrypt}_old.
1793 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1794 key API, doesn't do much yet.
1797 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1798 public key algorithms. New option to openssl utility:
1799 "list-public-key-algorithms" to print out info.
1802 *) Implement the Supported Elliptic Curves Extension for
1803 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1806 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1807 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1810 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1811 utilities such as rsa, dsa, dsaparam etc except they process any key
1815 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1816 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1817 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1821 *) Initial support for pluggable public key ASN1.
1822 De-spaghettify the public key ASN1 handling. Move public and private
1823 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1824 algorithm specific handling to a single module within the relevant
1825 algorithm directory. Add functions to allow (near) opaque processing
1826 of public and private key structures.
1829 *) Implement the Supported Point Formats Extension for
1830 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1833 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1834 for the psk identity [hint] and the psk callback functions to the
1835 SSL_SESSION, SSL and SSL_CTX structure.
1838 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1842 SSL_CTX_use_psk_identity_hint
1843 SSL_get_psk_identity_hint
1844 SSL_get_psk_identity
1845 SSL_use_psk_identity_hint
1847 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1849 *) Add RFC 3161 compliant time stamp request creation, response generation
1850 and response verification functionality.
1851 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1853 *) Add initial support for TLS extensions, specifically for the server_name
1854 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1855 have new members for a host name. The SSL data structure has an
1856 additional member SSL_CTX *initial_ctx so that new sessions can be
1857 stored in that context to allow for session resumption, even after the
1858 SSL has been switched to a new SSL_CTX in reaction to a client's
1859 server_name extension.
1861 New functions (subject to change):
1863 SSL_get_servername()
1864 SSL_get_servername_type()
1867 New CTRL codes and macros (subject to change):
1869 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1870 - SSL_CTX_set_tlsext_servername_callback()
1871 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1872 - SSL_CTX_set_tlsext_servername_arg()
1873 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1875 openssl s_client has a new '-servername ...' option.
1877 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1878 '-key2 ...', '-servername_fatal' (subject to change). This allows
1879 testing the HostName extension for a specific single host name ('-cert'
1880 and '-key' remain fallbacks for handshakes without HostName
1881 negotiation). If the unrecogninzed_name alert has to be sent, this by
1882 default is a warning; it becomes fatal with the '-servername_fatal'
1885 [Peter Sylvester, Remy Allais, Christophe Renou]
1887 *) Whirlpool hash implementation is added.
1890 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1891 bn(64,32). Because of instruction set limitations it doesn't have
1892 any negative impact on performance. This was done mostly in order
1893 to make it possible to share assembler modules, such as bn_mul_mont
1894 implementations, between 32- and 64-bit builds without hassle.
1897 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1898 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1902 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1903 dedicated Montgomery multiplication procedure, is introduced.
1904 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1905 "64-bit" performance on certain 32-bit targets.
1908 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1909 in SSL structures. New SSL ctrl to set maximum send fragment size.
1910 Save memory by seeting the I/O buffer sizes dynamically instead of
1911 using the maximum available value.
1914 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1915 in addition to the text details.
1918 *) Very, very preliminary EXPERIMENTAL support for printing of general
1919 ASN1 structures. This currently produces rather ugly output and doesn't
1920 handle several customised structures at all.
1923 *) Integrated support for PVK file format and some related formats such
1924 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1925 these in the 'rsa' and 'dsa' utilities.
1928 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1931 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1932 place for the (very old) "NETSCAPE" format certificates which are now
1933 handled using new ASN1 code equivalents.
1936 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1937 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1938 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1941 *) Modify CRL distribution points extension code to print out previously
1942 unsupported fields. Enhance extension setting code to allow setting of
1946 *) Add print and set support for Issuing Distribution Point CRL extension.
1949 *) Change 'Configure' script to enable Camellia by default.
1952 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1954 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1956 This addresses the flaw in CBC record processing discovered by
1957 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1958 at: http://www.isg.rhul.ac.uk/tls/
1960 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1961 Security Group at Royal Holloway, University of London
1962 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1963 Emilia Käsper for the initial patch.
1965 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1967 *) Return an error when checking OCSP signatures when key is NULL.
1968 This fixes a DoS attack. (CVE-2013-0166)
1971 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1972 the right response is stapled. Also change SSL_get_certificate()
1973 so it returns the certificate actually sent.
1974 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1975 (This is a backport)
1976 [Rob Stradling <rob.stradling@comodo.com>]
1978 *) Fix possible deadlock when decoding public keys.
1981 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1983 *) Sanity check record length before skipping explicit IV in DTLS
1986 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1987 fuzzing as a service testing platform.
1991 *) Initialise tkeylen properly when encrypting CMS messages.
1992 Thanks to Solar Designer of Openwall for reporting this issue.
1995 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1997 *) The fix for CVE-2012-2110 did not take into account that the
1998 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1999 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2000 rejecting negative len parameter. (CVE-2012-2131)
2001 [Tomas Hoger <thoger@redhat.com>]
2003 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2005 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2006 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2007 in CRYPTO_realloc_clean.
2009 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2010 issue and to Adam Langley <agl@chromium.org> for fixing it.
2012 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2014 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2016 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2017 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2018 content decryption and always return the same error. Note: this attack
2019 needs on average 2^20 messages so it only affects automated senders. The
2020 old behaviour can be reenabled in the CMS code by setting the
2021 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2022 an MMA defence is not necessary.
2023 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2024 this issue. (CVE-2012-0884)
2027 *) Fix CVE-2011-4619: make sure we really are receiving a
2028 client hello before rejecting multiple SGC restarts. Thanks to
2029 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2032 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2034 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2035 Thanks to Antonio Martin, Enterprise Secure Access Research and
2036 Development, Cisco Systems, Inc. for discovering this bug and
2037 preparing a fix. (CVE-2012-0050)
2040 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2042 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2043 of the Vaudenay padding oracle attack on CBC mode encryption
2044 which enables an efficient plaintext recovery attack against
2045 the OpenSSL implementation of DTLS. Their attack exploits timing
2046 differences arising during decryption processing. A research
2047 paper describing this attack can be found at:
2048 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2049 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2050 Security Group at Royal Holloway, University of London
2051 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2052 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2053 for preparing the fix. (CVE-2011-4108)
2054 [Robin Seggelmann, Michael Tuexen]
2056 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2057 [Ben Laurie, Kasper <ekasper@google.com>]
2059 *) Clear bytes used for block padding of SSL 3.0 records.
2061 [Adam Langley (Google)]
2063 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2064 Kadianakis <desnacked@gmail.com> for discovering this issue and
2065 Adam Langley for preparing the fix. (CVE-2011-4619)
2066 [Adam Langley (Google)]
2068 *) Prevent malformed RFC3779 data triggering an assertion failure.
2069 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2070 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2071 [Rob Austein <sra@hactrn.net>]
2073 *) Fix ssl_ciph.c set-up race.
2074 [Adam Langley (Google)]
2076 *) Fix spurious failures in ecdsatest.c.
2077 [Emilia Käsper (Google)]
2079 *) Fix the BIO_f_buffer() implementation (which was mixing different
2080 interpretations of the '..._len' fields).
2081 [Adam Langley (Google)]
2083 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2084 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2085 threads won't reuse the same blinding coefficients.
2087 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2088 lock to call BN_BLINDING_invert_ex, and avoids one use of
2089 BN_BLINDING_update for each BN_BLINDING structure (previously,
2090 the last update always remained unused).
2091 [Emilia Käsper (Google)]
2093 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2094 for multi-threaded use of ECDH.
2095 [Adam Langley (Google)]
2097 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2100 *) Add protection against ECDSA timing attacks as mentioned in the paper
2101 by Billy Bob Brumley and Nicola Tuveri, see:
2103 http://eprint.iacr.org/2011/232.pdf
2105 [Billy Bob Brumley and Nicola Tuveri]
2107 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2109 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2110 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2112 *) Fix bug in string printing code: if *any* escaping is enabled we must
2113 escape the escape character (backslash) or the resulting string is
2117 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2119 *) Disable code workaround for ancient and obsolete Netscape browsers
2120 and servers: an attacker can use it in a ciphersuite downgrade attack.
2121 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2124 *) Fixed J-PAKE implementation error, originally discovered by
2125 Sebastien Martini, further info and confirmation from Stefan
2126 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2129 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2131 *) Fix extension code to avoid race conditions which can result in a buffer
2132 overrun vulnerability: resumed sessions must not be modified as they can
2133 be shared by multiple threads. CVE-2010-3864
2136 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2139 *) Don't reencode certificate when calculating signature: cache and use
2140 the original encoding instead. This makes signature verification of
2141 some broken encodings work correctly.
2144 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2145 is also one of the inputs.
2146 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2148 *) Don't repeatedly append PBE algorithms to table if they already exist.
2149 Sort table on each new add. This effectively makes the table read only
2150 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2154 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2156 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2159 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2160 access or freeing data twice (CVE-2010-0742)
2161 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2163 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2164 common in certificates and some applications which only call
2165 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2169 Reduce copying into .apps and .test in makevms.com
2170 Don't try to use blank CA certificate in CA.com
2171 Allow use of C files from original directories in maketests.com
2172 [Steven M. Schweda" <sms@antinode.info>]
2174 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2176 *) When rejecting SSL/TLS records due to an incorrect version number, never
2177 update s->server with a new major version number. As of
2178 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2179 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2180 the previous behavior could result in a read attempt at NULL when
2181 receiving specific incorrect SSL/TLS records once record payload
2182 protection is active. (CVE-2010-0740)
2183 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2185 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2186 could be crashed if the relevant tables were not present (e.g. chrooted).
2187 [Tomas Hoger <thoger@redhat.com>]
2189 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2191 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2192 [Martin Olsson, Neel Mehta]
2194 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2195 accommodate for stack sorting, always a write lock!).
2198 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2199 excessive delays in the RAND_poll(): over a minute. As a workaround
2200 include a time check in the inner Heap32Next loop too.
2203 *) The code that handled flushing of data in SSL/TLS originally used the
2204 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2205 the problem outlined in PR#1949. The fix suggested there however can
2206 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2207 of Apache). So instead simplify the code to flush unconditionally.
2208 This should be fine since flushing with no data to flush is a no op.
2211 *) Handle TLS versions 2.0 and later properly and correctly use the
2212 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2213 off ancient servers have a habit of sticking around for a while...
2216 *) Modify compression code so it frees up structures without using the
2217 ex_data callbacks. This works around a problem where some applications
2218 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2219 restarting) then use compression (e.g. SSL with compression) later.
2220 This results in significant per-connection memory leaks and
2221 has caused some security issues including CVE-2008-1678 and
2225 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2226 change when encrypting or decrypting.
2229 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2230 connect and renegotiate with servers which do not support RI.
2231 Until RI is more widely deployed this option is enabled by default.
2234 *) Add "missing" ssl ctrls to clear options and mode.
2237 *) If client attempts to renegotiate and doesn't support RI respond with
2238 a no_renegotiation alert as required by RFC5746. Some renegotiating
2239 TLS clients will continue a connection gracefully when they receive
2240 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2241 waiting for a server hello which it will never receive. Now we treat a
2242 received no_renegotiation alert as a fatal error. This is because
2243 applications requesting a renegotiation might well expect it to succeed
2244 and would have no code in place to handle the server denying it so the
2245 only safe thing to do is to terminate the connection.
2248 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2249 peer supports secure renegotiation and 0 otherwise. Print out peer
2250 renegotiation support in s_client/s_server.
2253 *) Replace the highly broken and deprecated SPKAC certification method with
2254 the updated NID creation version. This should correctly handle UTF8.
2257 *) Implement RFC5746. Re-enable renegotiation but require the extension
2258 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2259 turns out to be a bad idea. It has been replaced by
2260 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2261 SSL_CTX_set_options(). This is really not recommended unless you
2262 know what you are doing.
2263 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2265 *) Fixes to stateless session resumption handling. Use initial_ctx when
2266 issuing and attempting to decrypt tickets in case it has changed during
2267 servername handling. Use a non-zero length session ID when attempting
2268 stateless session resumption: this makes it possible to determine if
2269 a resumption has occurred immediately after receiving server hello
2270 (several places in OpenSSL subtly assume this) instead of later in
2274 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2275 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2276 fixes for a few places where the return code is not checked
2278 [Julia Lawall <julia@diku.dk>]
2280 *) Add --strict-warnings option to Configure script to include devteam
2281 warnings in other configurations.
2284 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2285 makes it possible to install openssl libraries in locations which
2286 have names other than "lib", for example "/usr/lib64" which some
2288 [Steve Henson, based on patch from Jeremy Utley]
2290 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2291 X690 8.9.12 and can produce some misleading textual output of OIDs.
2292 [Steve Henson, reported by Dan Kaminsky]
2294 *) Delete MD2 from algorithm tables. This follows the recommendation in
2295 several standards that it is not used in new applications due to
2296 several cryptographic weaknesses. For binary compatibility reasons
2297 the MD2 API is still compiled in by default.
2300 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2304 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2305 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2307 [Guenter <lists@gknw.net>]
2309 *) Fix the server certificate chain building code to use X509_verify_cert(),
2310 it used to have an ad-hoc builder which was unable to cope with anything
2311 other than a simple chain.
2312 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2314 *) Don't check self signed certificate signatures in X509_verify_cert()
2315 by default (a flag can override this): it just wastes time without
2316 adding any security. As a useful side effect self signed root CAs
2317 with non-FIPS digests are now usable in FIPS mode.
2320 *) In dtls1_process_out_of_seq_message() the check if the current message
2321 is already buffered was missing. For every new message was memory
2322 allocated, allowing an attacker to perform an denial of service attack
2323 with sending out of seq handshake messages until there is no memory
2324 left. Additionally every future messege was buffered, even if the
2325 sequence number made no sense and would be part of another handshake.
2326 So only messages with sequence numbers less than 10 in advance will be
2327 buffered. (CVE-2009-1378)
2328 [Robin Seggelmann, discovered by Daniel Mentz]
2330 *) Records are buffered if they arrive with a future epoch to be
2331 processed after finishing the corresponding handshake. There is
2332 currently no limitation to this buffer allowing an attacker to perform
2333 a DOS attack with sending records with future epochs until there is no
2334 memory left. This patch adds the pqueue_size() function to detemine
2335 the size of a buffer and limits the record buffer to 100 entries.
2337 [Robin Seggelmann, discovered by Daniel Mentz]
2339 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2340 parent structure is freed. (CVE-2009-1379)
2343 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2344 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2347 [Ilya O. <vrghost@gmail.com>]
2349 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2351 *) Disable renegotiation completely - this fixes a severe security
2352 problem (CVE-2009-3555) at the cost of breaking all
2353 renegotiation. Renegotiation can be re-enabled by setting
2354 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2355 run-time. This is really not recommended unless you know what
2359 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2361 *) Don't set val to NULL when freeing up structures, it is freed up by
2362 underlying code. If sizeof(void *) > sizeof(long) this can result in
2363 zeroing past the valid field. (CVE-2009-0789)
2364 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2366 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2367 checked correctly. This would allow some invalid signed attributes to
2368 appear to verify correctly. (CVE-2009-0591)
2369 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2371 *) Reject UniversalString and BMPString types with invalid lengths. This
2372 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2373 a legal length. (CVE-2009-0590)
2376 *) Set S/MIME signing as the default purpose rather than setting it
2377 unconditionally. This allows applications to override it at the store
2381 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2382 to handle some structures.
2385 *) Improve efficiency of mem_gets: don't search whole buffer each time
2387 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2389 *) New -hex option for openssl rand.
2392 *) Print out UTF8String and NumericString when parsing ASN1.
2395 *) Support NumericString type for name components.
2398 *) Allow CC in the environment to override the automatically chosen
2399 compiler. Note that nothing is done to ensure flags work with the
2403 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2405 *) Properly check EVP_VerifyFinal() and similar return values
2407 [Ben Laurie, Bodo Moeller, Google Security Team]
2409 *) Enable TLS extensions by default.
2412 *) Allow the CHIL engine to be loaded, whether the application is
2413 multithreaded or not. (This does not release the developer from the
2414 obligation to set up the dynamic locking callbacks.)
2415 [Sander Temme <sander@temme.net>]
2417 *) Use correct exit code if there is an error in dgst command.
2418 [Steve Henson; problem pointed out by Roland Dirlewanger]
2420 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2421 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2424 *) Add experimental JPAKE support, including demo authentication in
2425 s_client and s_server.
2428 *) Set the comparison function in v3_addr_canonize().
2429 [Rob Austein <sra@hactrn.net>]
2431 *) Add support for XMPP STARTTLS in s_client.
2432 [Philip Paeps <philip@freebsd.org>]
2434 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2435 to ensure that even with this option, only ciphersuites in the
2436 server's preference list will be accepted. (Note that the option
2437 applies only when resuming a session, so the earlier behavior was
2438 just about the algorithm choice for symmetric cryptography.)
2441 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2443 *) Fix NULL pointer dereference if a DTLS server received
2444 ChangeCipherSpec as first record (CVE-2009-1386).
2447 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2448 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2451 *) The fix in 0.9.8c that supposedly got rid of unsafe
2452 double-checked locking was incomplete for RSA blinding,
2453 addressing just one layer of what turns out to have been
2454 doubly unsafe triple-checked locking.
2456 So now fix this for real by retiring the MONT_HELPER macro
2457 in crypto/rsa/rsa_eay.c.
2459 [Bodo Moeller; problem pointed out by Marius Schilder]
2461 *) Various precautionary measures:
2463 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2465 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2466 (NB: This would require knowledge of the secret session ticket key
2467 to exploit, in which case you'd be SOL either way.)
2469 - Change bn_nist.c so that it will properly handle input BIGNUMs
2470 outside the expected range.
2472 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2475 [Neel Mehta, Bodo Moeller]
2477 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2478 the load fails. Useful for distros.
2479 [Ben Laurie and the FreeBSD team]
2481 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2484 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2487 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2489 This work was sponsored by Logica.
2492 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2493 keystores. Support for SSL/TLS client authentication too.
2494 Not compiled unless enable-capieng specified to Configure.
2496 This work was sponsored by Logica.
2499 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2500 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2501 attribute creation routines such as certifcate requests and PKCS#12
2505 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2507 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2508 handshake which could lead to a cilent crash as found using the
2509 Codenomicon TLS test suite (CVE-2008-1672)
2510 [Steve Henson, Mark Cox]
2512 *) Fix double free in TLS server name extensions which could lead to
2513 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2516 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2518 Clear the error queue to ensure that error entries left from
2519 older function calls do not interfere with the correct operation.
2520 [Lutz Jaenicke, Erik de Castro Lopo]
2522 *) Remove root CA certificates of commercial CAs:
2524 The OpenSSL project does not recommend any specific CA and does not
2525 have any policy with respect to including or excluding any CA.
2526 Therefore it does not make any sense to ship an arbitrary selection
2527 of root CA certificates with the OpenSSL software.
2530 *) RSA OAEP patches to fix two separate invalid memory reads.
2531 The first one involves inputs when 'lzero' is greater than
2532 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2533 before the beginning of from). The second one involves inputs where
2534 the 'db' section contains nothing but zeroes (there is a one-byte
2535 invalid read after the end of 'db').
2536 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2538 *) Partial backport from 0.9.9-dev:
2540 Introduce bn_mul_mont (dedicated Montgomery multiplication
2541 procedure) as a candidate for BIGNUM assembler implementation.
2542 While 0.9.9-dev uses assembler for various architectures, only
2543 x86_64 is available by default here in the 0.9.8 branch, and
2544 32-bit x86 is available through a compile-time setting.
2546 To try the 32-bit x86 assembler implementation, use Configure
2547 option "enable-montasm" (which exists only for this backport).
2549 As "enable-montasm" for 32-bit x86 disclaims code stability
2550 anyway, in this constellation we activate additional code
2551 backported from 0.9.9-dev for further performance improvements,
2552 namely BN_from_montgomery_word. (To enable this otherwise,
2553 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2555 [Andy Polyakov (backport partially by Bodo Moeller)]
2557 *) Add TLS session ticket callback. This allows an application to set
2558 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2559 values. This is useful for key rollover for example where several key
2560 sets may exist with different names.
2563 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2564 This was broken until now in 0.9.8 releases, such that the only way
2565 a registered ENGINE could be used (assuming it initialises
2566 successfully on the host) was to explicitly set it as the default
2567 for the relevant algorithms. This is in contradiction with 0.9.7
2568 behaviour and the documentation. With this fix, when an ENGINE is
2569 registered into a given algorithm's table of implementations, the
2570 'uptodate' flag is reset so that auto-discovery will be used next
2571 time a new context for that algorithm attempts to select an
2573 [Ian Lister (tweaked by Geoff Thorpe)]
2575 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2576 implemention in the following ways:
2578 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2581 Lack of BER streaming support means one pass streaming processing is
2582 only supported if data is detached: setting the streaming flag is
2583 ignored for embedded content.
2585 CMS support is disabled by default and must be explicitly enabled
2586 with the enable-cms configuration option.
2589 *) Update the GMP engine glue to do direct copies between BIGNUM and
2590 mpz_t when openssl and GMP use the same limb size. Otherwise the
2591 existing "conversion via a text string export" trick is still used.
2592 [Paul Sheer <paulsheer@gmail.com>]
2594 *) Zlib compression BIO. This is a filter BIO which compressed and
2595 uncompresses any data passed through it.
2598 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2599 RFC3394 compatible AES key wrapping.
2602 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2603 sets string data without copying. X509_ALGOR_set0() and
2604 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2605 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2606 from an X509_ATTRIBUTE structure optionally checking it occurs only
2607 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2611 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2612 to get the expected BN_FLG_CONSTTIME behavior.
2613 [Bodo Moeller (Google)]
2617 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2618 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2619 - added some more tests to do_tests.pl
2620 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2621 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2622 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2623 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2624 - various changes to netware.pl to enable gcc-cross builds on Win32
2626 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2627 - various changes to fix missing prototype warnings
2628 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2629 - added AES, WHIRLPOOL and CPUID assembler code to build files
2630 - added missing AES assembler make rules to mk1mf.pl
2631 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2632 [Guenter Knauf <eflash@gmx.net>]
2634 *) Implement certificate status request TLS extension defined in RFC3546.
2635 A client can set the appropriate parameters and receive the encoded
2636 OCSP response via a callback. A server can query the supplied parameters
2637 and set the encoded OCSP response in the callback. Add simplified examples
2638 to s_client and s_server.
2641 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2643 *) Fix various bugs:
2644 + Binary incompatibility of ssl_ctx_st structure
2645 + DTLS interoperation with non-compliant servers
2646 + Don't call get_session_cb() without proposed session
2647 + Fix ia64 assembler code
2648 [Andy Polyakov, Steve Henson]
2650 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2652 *) DTLS Handshake overhaul. There were longstanding issues with
2653 OpenSSL DTLS implementation, which were making it impossible for
2654 RFC 4347 compliant client to communicate with OpenSSL server.
2655 Unfortunately just fixing these incompatibilities would "cut off"
2656 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2657 server keeps tolerating non RFC compliant syntax. The opposite is
2658 not true, 0.9.8f client can not communicate with earlier server.
2659 This update even addresses CVE-2007-4995.
2662 *) Changes to avoid need for function casts in OpenSSL: some compilers
2663 (gcc 4.2 and later) reject their use.
2664 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2667 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2668 RFC4507bis. The encrypted ticket format is an encrypted encoded
2669 SSL_SESSION structure, that way new session features are automatically
2672 If a client application caches session in an SSL_SESSION structure
2673 support is transparent because tickets are now stored in the encoded
2676 The SSL_CTX structure automatically generates keys for ticket
2677 protection in servers so again support should be possible
2678 with no application modification.
2680 If a client or server wishes to disable RFC4507 support then the option
2681 SSL_OP_NO_TICKET can be set.
2683 Add a TLS extension debugging callback to allow the contents of any client
2684 or server extensions to be examined.
2686 This work was sponsored by Google.
2689 *) Add initial support for TLS extensions, specifically for the server_name
2690 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2691 have new members for a host name. The SSL data structure has an
2692 additional member SSL_CTX *initial_ctx so that new sessions can be
2693 stored in that context to allow for session resumption, even after the
2694 SSL has been switched to a new SSL_CTX in reaction to a client's
2695 server_name extension.
2697 New functions (subject to change):
2699 SSL_get_servername()
2700 SSL_get_servername_type()
2703 New CTRL codes and macros (subject to change):
2705 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2706 - SSL_CTX_set_tlsext_servername_callback()
2707 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2708 - SSL_CTX_set_tlsext_servername_arg()
2709 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2711 openssl s_client has a new '-servername ...' option.
2713 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2714 '-key2 ...', '-servername_fatal' (subject to change). This allows
2715 testing the HostName extension for a specific single host name ('-cert'
2716 and '-key' remain fallbacks for handshakes without HostName
2717 negotiation). If the unrecogninzed_name alert has to be sent, this by
2718 default is a warning; it becomes fatal with the '-servername_fatal'
2721 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2723 *) Add AES and SSE2 assembly language support to VC++ build.
2726 *) Mitigate attack on final subtraction in Montgomery reduction.
2729 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2730 (which previously caused an internal error).
2733 *) Squeeze another 10% out of IGE mode when in != out.
2736 *) AES IGE mode speedup.
2737 [Dean Gaudet (Google)]
2739 *) Add the Korean symmetric 128-bit cipher SEED (see
2740 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2741 add SEED ciphersuites from RFC 4162:
2743 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2744 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2745 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2746 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2748 To minimize changes between patchlevels in the OpenSSL 0.9.8
2749 series, SEED remains excluded from compilation unless OpenSSL
2750 is configured with 'enable-seed'.
2751 [KISA, Bodo Moeller]
2753 *) Mitigate branch prediction attacks, which can be practical if a
2754 single processor is shared, allowing a spy process to extract
2755 information. For detailed background information, see
2756 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2757 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2758 and Necessary Software Countermeasures"). The core of the change
2759 are new versions BN_div_no_branch() and
2760 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2761 respectively, which are slower, but avoid the security-relevant
2762 conditional branches. These are automatically called by BN_div()
2763 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2764 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2765 remove a conditional branch.
2767 BN_FLG_CONSTTIME is the new name for the previous
2768 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2769 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2770 in the exponent causes BN_mod_exp_mont() to use the alternative
2771 implementation in BN_mod_exp_mont_consttime().) The old name
2772 remains as a deprecated alias.
2774 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2775 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2776 constant-time implementations for more than just exponentiation.
2777 Here too the old name is kept as a deprecated alias.
2779 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2780 the BN_BLINDING structure gets an independent copy of the
2781 modulus. This means that the previous "BIGNUM *m" argument to
2782 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2783 essentially becomes "const BIGNUM *m", although we can't actually
2784 change this in the header file before 0.9.9. It allows
2785 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2786 enable BN_FLG_CONSTTIME.
2788 [Matthew D Wood (Intel Corp)]
2790 *) In the SSL/TLS server implementation, be strict about session ID
2791 context matching (which matters if an application uses a single
2792 external cache for different purposes). Previously,
2793 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2794 set. This did ensure strict client verification, but meant that,
2795 with applications using a single external cache for quite
2796 different requirements, clients could circumvent ciphersuite
2797 restrictions for a given session ID context by starting a session
2798 in a different context.
2801 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2802 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2803 authentication-only ciphersuites.
2806 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2807 not complete and could lead to a possible single byte overflow
2808 (CVE-2007-5135) [Ben Laurie]
2810 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2812 *) Since AES128 and AES256 (and similarly Camellia128 and
2813 Camellia256) share a single mask bit in the logic of
2814 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2815 kludge to work properly if AES128 is available and AES256 isn't
2816 (or if Camellia128 is available and Camellia256 isn't).
2819 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2820 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2821 When a point or a seed is encoded in a BIT STRING, we need to
2822 prevent the removal of trailing zero bits to get the proper DER
2823 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2824 of a NamedBitList, for which trailing 0 bits need to be removed.)
2827 *) Have SSL/TLS server implementation tolerate "mismatched" record
2828 protocol version while receiving ClientHello even if the
2829 ClientHello is fragmented. (The server can't insist on the
2830 particular protocol version it has chosen before the ServerHello
2831 message has informed the client about his choice.)
2834 *) Add RFC 3779 support.
2835 [Rob Austein for ARIN, Ben Laurie]
2837 *) Load error codes if they are not already present instead of using a
2838 static variable. This allows them to be cleanly unloaded and reloaded.
2839 Improve header file function name parsing.
2842 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2843 or CAPABILITY handshake as required by RFCs.
2846 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2848 *) Introduce limits to prevent malicious keys being able to
2849 cause a denial of service. (CVE-2006-2940)
2850 [Steve Henson, Bodo Moeller]
2852 *) Fix ASN.1 parsing of certain invalid structures that can result
2853 in a denial of service. (CVE-2006-2937) [Steve Henson]
2855 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2856 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2858 *) Fix SSL client code which could crash if connecting to a
2859 malicious SSLv2 server. (CVE-2006-4343)
2860 [Tavis Ormandy and Will Drewry, Google Security Team]
2862 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2863 match only those. Before that, "AES256-SHA" would be interpreted
2864 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2865 the same strength classification in 0.9.7h) as we currently only
2866 have a single AES bit in the ciphersuite description bitmap.
2867 That change, however, also applied to ciphersuite strings such as
2868 "RC4-MD5" that intentionally matched multiple ciphersuites --
2869 namely, SSL 2.0 ciphersuites in addition to the more common ones
2870 from SSL 3.0/TLS 1.0.
2872 So we change the selection algorithm again: Naming an explicit
2873 ciphersuite selects this one ciphersuite, and any other similar
2874 ciphersuite (same bitmap) from *other* protocol versions.
2875 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2876 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2878 Since SSL 2.0 does not have any ciphersuites for which the
2879 128/256 bit distinction would be relevant, this works for now.
2880 The proper fix will be to use different bits for AES128 and
2881 AES256, which would have avoided the problems from the beginning;
2882 however, bits are scarce, so we can only do this in a new release
2883 (not just a patchlevel) when we can change the SSL_CIPHER
2884 definition to split the single 'unsigned long mask' bitmap into
2885 multiple values to extend the available space.
2889 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2891 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2892 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2894 *) Add AES IGE and biIGE modes.
2897 *) Change the Unix randomness entropy gathering to use poll() when
2898 possible instead of select(), since the latter has some
2899 undesirable limitations.
2900 [Darryl Miles via Richard Levitte and Bodo Moeller]
2902 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2903 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2904 cannot be implicitly activated as part of, e.g., the "AES" alias.
2905 However, please upgrade to OpenSSL 0.9.9[-dev] for
2906 non-experimental use of the ECC ciphersuites to get TLS extension
2907 support, which is required for curve and point format negotiation
2908 to avoid potential handshake problems.
2911 *) Disable rogue ciphersuites:
2913 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2914 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2915 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2917 The latter two were purportedly from
2918 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2921 Also deactivate the remaining ciphersuites from
2922 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2923 unofficial, and the ID has long expired.
2926 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2927 dual-core machines) and other potential thread-safety issues.
2930 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2931 versions), which is now available for royalty-free use
2932 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2933 Also, add Camellia TLS ciphersuites from RFC 4132.
2935 To minimize changes between patchlevels in the OpenSSL 0.9.8
2936 series, Camellia remains excluded from compilation unless OpenSSL
2937 is configured with 'enable-camellia'.
2940 *) Disable the padding bug check when compression is in use. The padding
2941 bug check assumes the first packet is of even length, this is not
2942 necessarily true if compresssion is enabled and can result in false
2943 positives causing handshake failure. The actual bug test is ancient
2944 code so it is hoped that implementations will either have fixed it by
2945 now or any which still have the bug do not support compression.
2948 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2950 *) When applying a cipher rule check to see if string match is an explicit
2951 cipher suite and only match that one cipher suite if it is.
2954 *) Link in manifests for VC++ if needed.
2955 [Austin Ziegler <halostatue@gmail.com>]
2957 *) Update support for ECC-based TLS ciphersuites according to
2958 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2959 TLS extensions, which are supported starting with the 0.9.9
2960 branch, not in the OpenSSL 0.9.8 branch).
2963 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2964 opaque EVP_CIPHER_CTX handling.
2967 *) Fixes and enhancements to zlib compression code. We now only use
2968 "zlib1.dll" and use the default __cdecl calling convention on Win32
2969 to conform with the standards mentioned here:
2970 http://www.zlib.net/DLL_FAQ.txt
2971 Static zlib linking now works on Windows and the new --with-zlib-include
2972 --with-zlib-lib options to Configure can be used to supply the location
2973 of the headers and library. Gracefully handle case where zlib library
2977 *) Several fixes and enhancements to the OID generation code. The old code
2978 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2979 handle numbers larger than ULONG_MAX, truncated printing and had a
2980 non standard OBJ_obj2txt() behaviour.
2983 *) Add support for building of engines under engine/ as shared libraries
2984 under VC++ build system.
2987 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2988 Hopefully, we will not see any false combination of paths any more.
2991 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2993 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2994 (part of SSL_OP_ALL). This option used to disable the
2995 countermeasure against man-in-the-middle protocol-version
2996 rollback in the SSL 2.0 server implementation, which is a bad
2997 idea. (CVE-2005-2969)
2999 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3000 for Information Security, National Institute of Advanced Industrial
3001 Science and Technology [AIST], Japan)]
3003 *) Add two function to clear and return the verify parameter flags.
3006 *) Keep cipherlists sorted in the source instead of sorting them at
3007 runtime, thus removing the need for a lock.
3010 *) Avoid some small subgroup attacks in Diffie-Hellman.
3011 [Nick Mathewson and Ben Laurie]
3013 *) Add functions for well-known primes.
3016 *) Extended Windows CE support.
3017 [Satoshi Nakamura and Andy Polyakov]
3019 *) Initialize SSL_METHOD structures at compile time instead of during
3020 runtime, thus removing the need for a lock.
3023 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3024 attempting to decrypt each encrypted key in turn. Add support to
3028 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3030 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3033 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3036 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3037 key into the same file any more.
3040 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3043 *) Add -utf8 command line and config file option to 'ca'.
3044 [Stefan <stf@udoma.org]
3046 *) Removed the macro des_crypt(), as it seems to conflict with some
3047 libraries. Use DES_crypt().
3050 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3051 involves renaming the source and generated shared-libs for
3052 both. The engines will accept the corrected or legacy ids
3053 ('ncipher' and '4758_cca' respectively) when binding. NB,
3054 this only applies when building 'shared'.
3055 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3057 *) Add attribute functions to EVP_PKEY structure. Modify
3058 PKCS12_create() to recognize a CSP name attribute and
3059 use it. Make -CSP option work again in pkcs12 utility.
3062 *) Add new functionality to the bn blinding code:
3063 - automatic re-creation of the BN_BLINDING parameters after
3064 a fixed number of uses (currently 32)
3065 - add new function for parameter creation
3066 - introduce flags to control the update behaviour of the
3067 BN_BLINDING parameters
3068 - hide BN_BLINDING structure
3069 Add a second BN_BLINDING slot to the RSA structure to improve
3070 performance when a single RSA object is shared among several
3074 *) Add support for DTLS.
3075 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3077 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3078 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3081 *) Remove buggy and incompletet DH cert support from
3082 ssl/ssl_rsa.c and ssl/s3_both.c
3085 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3086 the apps/openssl applications.
3089 *) Compile clean with "-Wall -Wmissing-prototypes
3090 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3091 DEBUG_SAFESTACK must also be set.
3094 *) Change ./Configure so that certain algorithms can be disabled by default.
3095 The new counterpiece to "no-xxx" is "enable-xxx".
3097 The patented RC5 and MDC2 algorithms will now be disabled unless
3098 "enable-rc5" and "enable-mdc2", respectively, are specified.
3100 (IDEA remains enabled despite being patented. This is because IDEA
3101 is frequently required for interoperability, and there is no license
3102 fee for non-commercial use. As before, "no-idea" can be used to
3103 avoid this algorithm.)
3107 *) Add processing of proxy certificates (see RFC 3820). This work was
3108 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3109 EGEE (Enabling Grids for E-science in Europe).
3112 *) RC4 performance overhaul on modern architectures/implementations, such
3113 as Intel P4, IA-64 and AMD64.
3116 *) New utility extract-section.pl. This can be used specify an alternative
3117 section number in a pod file instead of having to treat each file as
3118 a separate case in Makefile. This can be done by adding two lines to the
3121 =for comment openssl_section:XXX
3123 The blank line is mandatory.
3127 *) New arguments -certform, -keyform and -pass for s_client and s_server
3128 to allow alternative format key and certificate files and passphrase
3132 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3133 update associated structures and add various utility functions.
3135 Add new policy related verify parameters, include policy checking in
3136 standard verify code. Enhance 'smime' application with extra parameters
3137 to support policy checking and print out.
3140 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3141 Nehemiah processors. These extensions support AES encryption in hardware
3142 as well as RNG (though RNG support is currently disabled).
3143 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3145 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3148 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3149 [Andy Polyakov and a number of other people]
3151 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3152 implementation contributed by IBM.
3153 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3155 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3156 exponent rather than 'unsigned long'. There is a corresponding change to
3157 the new 'rsa_keygen' element of the RSA_METHOD structure.
3158 [Jelte Jansen, Geoff Thorpe]
3160 *) Functionality for creating the initial serial number file is now
3161 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3163 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3164 number file to 1, which is bound to cause problems. To avoid
3165 the problems while respecting compatibility between different 0.9.7
3166 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3167 CA.pl for serial number initialization. With the new release 0.9.8,
3168 we can fix the problem directly in the 'ca' utility.)
3171 *) Reduced header interdepencies by declaring more opaque objects in
3172 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3173 give fewer recursive includes, which could break lazy source code - so
3174 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3175 developers should define this symbol when building and using openssl to
3176 ensure they track the recommended behaviour, interfaces, [etc], but
3177 backwards-compatible behaviour prevails when this isn't defined.
3180 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3183 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3184 This will generate a random key of the appropriate length based on the
3185 cipher context. The EVP_CIPHER can provide its own random key generation
3186 routine to support keys of a specific form. This is used in the des and
3187 3des routines to generate a key of the correct parity. Update S/MIME
3188 code to use new functions and hence generate correct parity DES keys.
3189 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3190 valid (weak or incorrect parity).
3193 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3194 as looking them up. This is useful when the verified structure may contain
3195 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3196 present unless the new PKCS7_NO_CRL flag is asserted.
3199 *) Extend ASN1 oid configuration module. It now additionally accepts the
3202 shortName = some long name, 1.2.3.4
3205 *) Reimplemented the BN_CTX implementation. There is now no more static
3206 limitation on the number of variables it can handle nor the depth of the
3207 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3208 information can now expand as required, and rather than having a single
3209 static array of bignums, BN_CTX now uses a linked-list of such arrays
3210 allowing it to expand on demand whilst maintaining the usefulness of
3211 BN_CTX's "bundling".
3214 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3215 to allow all RSA operations to function using a single BN_CTX.
3218 *) Preliminary support for certificate policy evaluation and checking. This
3219 is initially intended to pass the tests outlined in "Conformance Testing
3220 of Relying Party Client Certificate Path Processing Logic" v1.07.
3223 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3224 remained unused and not that useful. A variety of other little bignum
3225 tweaks and fixes have also been made continuing on from the audit (see
3229 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3230 associated ASN1, EVP and SSL functions and old ASN1 macros.
3233 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3234 and this should never fail. So the return value from the use of
3235 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3236 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3239 *) BN_CTX_get() should return zero-valued bignums, providing the same
3240 initialised value as BN_new().
3241 [Geoff Thorpe, suggested by Ulf Möller]
3243 *) Support for inhibitAnyPolicy certificate extension.
3246 *) An audit of the BIGNUM code is underway, for which debugging code is
3247 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3248 is considered valid when processing BIGNUMs, and causes execution to
3249 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3250 further steps are taken to deliberately pollute unused data in BIGNUM
3251 structures to try and expose faulty code further on. For now, openssl will
3252 (in its default mode of operation) continue to tolerate the inconsistent
3253 forms that it has tolerated in the past, but authors and packagers should
3254 consider trying openssl and their own applications when compiled with
3255 these debugging symbols defined. It will help highlight potential bugs in
3256 their own code, and will improve the test coverage for OpenSSL itself. At
3257 some point, these tighter rules will become openssl's default to improve
3258 maintainability, though the assert()s and other overheads will remain only
3259 in debugging configurations. See bn.h for more details.
3260 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3262 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3263 that can only be obtained through BN_CTX_new() (which implicitly
3264 initialises it). The presence of this function only made it possible
3265 to overwrite an existing structure (and cause memory leaks).
3268 *) Because of the callback-based approach for implementing LHASH as a
3269 template type, lh_insert() adds opaque objects to hash-tables and
3270 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3271 to clean up those corresponding objects before destroying the hash table
3272 (and losing the object pointers). So some over-zealous constifications in
3273 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3274 objects as "const" and the lh_doall[_arg] callback wrappers are not
3275 prototyped to have "const" restrictions on the object pointers they are
3276 given (and so aren't required to cast them away any more).
3279 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3280 (speed) prefers to use its own implementation. The two implementations
3281 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3282 its object type properly exposed (MS_TM) instead of casting to/from "char
3283 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3284 aren't necessarily the greatest nomenclatures - but this is what was used
3285 internally to the implementation so I've used that for now.
3288 *) Ensure that deprecated functions do not get compiled when
3289 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3290 the self-tests were still using deprecated key-generation functions so
3291 these have been updated also.
3294 *) Reorganise PKCS#7 code to separate the digest location functionality
3295 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3296 New function PKCS7_set_digest() to set the digest type for PKCS#7
3297 digestedData type. Add additional code to correctly generate the
3298 digestedData type and add support for this type in PKCS7 initialization
3302 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3303 structure of type "other".
3306 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3307 sure the loop does correctly stop and breaking ("division by zero")
3308 modulus operations are not performed. The (pre-generated) prime
3309 table crypto/bn/bn_prime.h was already correct, but it could not be
3310 re-generated on some platforms because of the "division by zero"
3311 situation in the script.
3312 [Ralf S. Engelschall]
3314 *) Update support for ECC-based TLS ciphersuites according to
3315 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3316 SHA-1 now is only used for "small" curves (where the
3317 representation of a field element takes up to 24 bytes); for
3318 larger curves, the field element resulting from ECDH is directly
3319 used as premaster secret.
3320 [Douglas Stebila (Sun Microsystems Laboratories)]
3322 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3323 curve secp160r1 to the tests.
3324 [Douglas Stebila (Sun Microsystems Laboratories)]
3326 *) Add the possibility to load symbols globally with DSO.
3327 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3329 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3330 control of the error stack.
3333 *) Add support for STORE in ENGINE.
3336 *) Add the STORE type. The intention is to provide a common interface
3337 to certificate and key stores, be they simple file-based stores, or
3338 HSM-type store, or LDAP stores, or...
3339 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3342 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3343 pass a list of arguments to any function as well as provide a way
3344 for a function to pass data back to the caller.
3347 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3348 works like BUF_strdup() but can be used to duplicate a portion of
3349 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3353 *) Add the function sk_find_ex() which works like sk_find(), but will
3354 return an index to an element even if an exact match couldn't be
3355 found. The index is guaranteed to point at the element where the
3356 searched-for key would be inserted to preserve sorting order.
3359 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3360 takes an extra flags argument for optional functionality. Currently,
3361 the following flags are defined:
3363 OBJ_BSEARCH_VALUE_ON_NOMATCH
3364 This one gets OBJ_bsearch_ex() to return a pointer to the first
3365 element where the comparing function returns a negative or zero
3368 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3369 This one gets OBJ_bsearch_ex() to return a pointer to the first
3370 element where the comparing function returns zero. This is useful
3371 if there are more than one element where the comparing function
3375 *) Make it possible to create self-signed certificates with 'openssl ca'
3376 in such a way that the self-signed certificate becomes part of the
3377 CA database and uses the same mechanisms for serial number generation
3378 as all other certificate signing. The new flag '-selfsign' enables
3379 this functionality. Adapt CA.sh and CA.pl.in.
3382 *) Add functionality to check the public key of a certificate request
3383 against a given private. This is useful to check that a certificate
3384 request can be signed by that key (self-signing).
3387 *) Make it possible to have multiple active certificates with the same
3388 subject in the CA index file. This is done only if the keyword
3389 'unique_subject' is set to 'no' in the main CA section (default
3390 if 'CA_default') of the configuration file. The value is saved
3391 with the database itself in a separate index attribute file,
3392 named like the index file with '.attr' appended to the name.
3395 *) Generate muti valued AVAs using '+' notation in config files for
3399 *) Support for nameConstraints certificate extension.
3402 *) Support for policyConstraints certificate extension.
3405 *) Support for policyMappings certificate extension.
3408 *) Make sure the default DSA_METHOD implementation only uses its
3409 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3410 and change its own handlers to be NULL so as to remove unnecessary
3411 indirection. This lets alternative implementations fallback to the
3412 default implementation more easily.
3415 *) Support for directoryName in GeneralName related extensions
3419 *) Make it possible to link applications using Makefile.shared.
3420 Make that possible even when linking against static libraries!
3423 *) Support for single pass processing for S/MIME signing. This now
3424 means that S/MIME signing can be done from a pipe, in addition
3425 cleartext signing (multipart/signed type) is effectively streaming
3426 and the signed data does not need to be all held in memory.
3428 This is done with a new flag PKCS7_STREAM. When this flag is set
3429 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3430 is done after the data is output (and digests calculated) in
3431 SMIME_write_PKCS7().
3434 *) Add full support for -rpath/-R, both in shared libraries and
3435 applications, at least on the platforms where it's known how
3439 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3440 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3441 will now compute a table of multiples of the generator that
3442 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3443 faster (notably in the case of a single point multiplication,
3444 scalar * generator).
3445 [Nils Larsch, Bodo Moeller]
3447 *) IPv6 support for certificate extensions. The various extensions
3448 which use the IP:a.b.c.d can now take IPv6 addresses using the
3449 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3453 *) Added an ENGINE that implements RSA by performing private key
3454 exponentiations with the GMP library. The conversions to and from
3455 GMP's mpz_t format aren't optimised nor are any montgomery forms
3456 cached, and on x86 it appears OpenSSL's own performance has caught up.
3457 However there are likely to be other architectures where GMP could
3458 provide a boost. This ENGINE is not built in by default, but it can be
3459 specified at Configure time and should be accompanied by the necessary
3460 linker additions, eg;
3461 ./config -DOPENSSL_USE_GMP -lgmp
3464 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3465 testing availability of engines with "-t" - the old behaviour is
3466 produced by increasing the feature's verbosity with "-tt".
3469 *) ECDSA routines: under certain error conditions uninitialized BN objects
3470 could be freed. Solution: make sure initialization is performed early
3471 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3475 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3476 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3477 software implementations. For DSA and DH, parameter generation can
3478 also be overriden by providing the appropriate method callbacks.
3481 *) Change the "progress" mechanism used in key-generation and
3482 primality testing to functions that take a new BN_GENCB pointer in
3483 place of callback/argument pairs. The new API functions have "_ex"
3484 postfixes and the older functions are reimplemented as wrappers for
3485 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3486 declarations of the old functions to help (graceful) attempts to
3487 migrate to the new functions. Also, the new key-generation API
3488 functions operate on a caller-supplied key-structure and return
3489 success/failure rather than returning a key or NULL - this is to
3490 help make "keygen" another member function of RSA_METHOD etc.
3492 Example for using the new callback interface:
3494 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3498 BN_GENCB_set(&my_cb, my_callback, my_arg);
3500 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3501 /* For the meaning of a, b in calls to my_callback(), see the
3502 * documentation of the function that calls the callback.
3503 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3504 * my_callback should return 1 if it wants BN_is_prime_ex()
3505 * to continue, or 0 to stop.
3510 *) Change the ZLIB compression method to be stateful, and make it
3511 available to TLS with the number defined in
3512 draft-ietf-tls-compression-04.txt.
3515 *) Add the ASN.1 structures and functions for CertificatePair, which
3516 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3518 CertificatePair ::= SEQUENCE {
3519 forward [0] Certificate OPTIONAL,
3520 reverse [1] Certificate OPTIONAL,
3521 -- at least one of the pair shall be present -- }
3523 Also implement the PEM functions to read and write certificate
3524 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3526 This needed to be defined, mostly for the sake of the LDAP
3527 attribute crossCertificatePair, but may prove useful elsewhere as
3531 *) Make it possible to inhibit symlinking of shared libraries in
3532 Makefile.shared, for Cygwin's sake.
3535 *) Extend the BIGNUM API by creating a function
3536 void BN_set_negative(BIGNUM *a, int neg);
3537 and a macro that behave like
3538 int BN_is_negative(const BIGNUM *a);
3540 to avoid the need to access 'a->neg' directly in applications.
3543 *) Implement fast modular reduction for pseudo-Mersenne primes
3544 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3545 EC_GROUP_new_curve_GFp() will now automatically use this
3547 [Nils Larsch <nla@trustcenter.de>]
3549 *) Add new lock type (CRYPTO_LOCK_BN).
3552 *) Change the ENGINE framework to automatically load engines
3553 dynamically from specific directories unless they could be
3554 found to already be built in or loaded. Move all the
3555 current engines except for the cryptodev one to a new
3557 The engines in engines/ are built as shared libraries if
3558 the "shared" options was given to ./Configure or ./config.
3559 Otherwise, they are inserted in libcrypto.a.
3560 /usr/local/ssl/engines is the default directory for dynamic
3561 engines, but that can be overriden at configure time through
3562 the usual use of --prefix and/or --openssldir, and at run
3563 time with the environment variable OPENSSL_ENGINES.
3564 [Geoff Thorpe and Richard Levitte]
3566 *) Add Makefile.shared, a helper makefile to build shared
3567 libraries. Addapt Makefile.org.
3570 *) Add version info to Win32 DLLs.
3571 [Peter 'Luna' Runestig" <peter@runestig.com>]
3573 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3574 can be added using this API to created arbitrary PKCS#12
3575 files while avoiding the low level API.
3577 New options to PKCS12_create(), key or cert can be NULL and
3578 will then be omitted from the output file. The encryption
3579 algorithm NIDs can be set to -1 for no encryption, the mac
3580 iteration count can be set to 0 to omit the mac.
3582 Enhance pkcs12 utility by making the -nokeys and -nocerts
3583 options work when creating a PKCS#12 file. New option -nomac
3584 to omit the mac, NONE can be set for an encryption algorithm.
3585 New code is modified to use the enhanced PKCS12_create()
3586 instead of the low level API.
3589 *) Extend ASN1 encoder to support indefinite length constructed
3590 encoding. This can output sequences tags and octet strings in
3591 this form. Modify pk7_asn1.c to support indefinite length
3592 encoding. This is experimental and needs additional code to
3593 be useful, such as an ASN1 bio and some enhanced streaming
3596 Extend template encode functionality so that tagging is passed
3597 down to the template encoder.
3600 *) Let 'openssl req' fail if an argument to '-newkey' is not
3601 recognized instead of using RSA as a default.
3604 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3605 As these are not official, they are not included in "ALL";
3606 the "ECCdraft" ciphersuite group alias can be used to select them.
3607 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3609 *) Add ECDH engine support.
3610 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3612 *) Add ECDH in new directory crypto/ecdh/.
3613 [Douglas Stebila (Sun Microsystems Laboratories)]
3615 *) Let BN_rand_range() abort with an error after 100 iterations
3616 without success (which indicates a broken PRNG).
3619 *) Change BN_mod_sqrt() so that it verifies that the input value
3620 is really the square of the return value. (Previously,
3621 BN_mod_sqrt would show GIGO behaviour.)
3624 *) Add named elliptic curves over binary fields from X9.62, SECG,
3625 and WAP/WTLS; add OIDs that were still missing.
3627 [Sheueling Chang Shantz and Douglas Stebila
3628 (Sun Microsystems Laboratories)]
3630 *) Extend the EC library for elliptic curves over binary fields
3631 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3634 EC_GF2m_simple_method
3638 EC_GROUP_new_curve_GF2m
3639 EC_GROUP_set_curve_GF2m
3640 EC_GROUP_get_curve_GF2m
3641 EC_POINT_set_affine_coordinates_GF2m
3642 EC_POINT_get_affine_coordinates_GF2m
3643 EC_POINT_set_compressed_coordinates_GF2m
3645 Point compression for binary fields is disabled by default for
3646 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3649 As binary polynomials are represented as BIGNUMs, various members
3650 of the EC_GROUP and EC_POINT data structures can be shared
3651 between the implementations for prime fields and binary fields;
3652 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3653 are essentially identical to their ..._GFp counterparts.
3654 (For simplicity, the '..._GFp' prefix has been dropped from
3655 various internal method names.)
3657 An internal 'field_div' method (similar to 'field_mul' and
3658 'field_sqr') has been added; this is used only for binary fields.
3660 [Sheueling Chang Shantz and Douglas Stebila
3661 (Sun Microsystems Laboratories)]
3663 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3664 through methods ('mul', 'precompute_mult').
3666 The generic implementations (now internally called 'ec_wNAF_mul'
3667 and 'ec_wNAF_precomputed_mult') remain the default if these
3668 methods are undefined.
3670 [Sheueling Chang Shantz and Douglas Stebila
3671 (Sun Microsystems Laboratories)]
3673 *) New function EC_GROUP_get_degree, which is defined through
3674 EC_METHOD. For curves over prime fields, this returns the bit
3675 length of the modulus.
3677 [Sheueling Chang Shantz and Douglas Stebila
3678 (Sun Microsystems Laboratories)]
3680 *) New functions EC_GROUP_dup, EC_POINT_dup.
3681 (These simply call ..._new and ..._copy).
3683 [Sheueling Chang Shantz and Douglas Stebila
3684 (Sun Microsystems Laboratories)]
3686 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3687 Polynomials are represented as BIGNUMs (where the sign bit is not
3688 used) in the following functions [macros]:
3691 BN_GF2m_sub [= BN_GF2m_add]
3692 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3693 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3694 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3696 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3697 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3698 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3699 BN_GF2m_cmp [= BN_ucmp]
3701 (Note that only the 'mod' functions are actually for fields GF(2^m).
3702 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3704 For some functions, an the irreducible polynomial defining a
3705 field can be given as an 'unsigned int[]' with strictly
3706 decreasing elements giving the indices of those bits that are set;
3707 i.e., p[] represents the polynomial
3708 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3710 p[0] > p[1] > ... > p[k] = 0.
3711 This applies to the following functions:
3716 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3717 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3719 BN_GF2m_mod_sqrt_arr
3720 BN_GF2m_mod_solve_quad_arr
3724 Conversion can be performed by the following functions:
3729 bntest.c has additional tests for binary polynomial arithmetic.
3731 Two implementations for BN_GF2m_mod_div() are available.
3732 The default algorithm simply uses BN_GF2m_mod_inv() and
3733 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3734 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3735 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3737 [Sheueling Chang Shantz and Douglas Stebila
3738 (Sun Microsystems Laboratories)]
3740 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3741 functionality is disabled at compile-time.
3742 [Douglas Stebila <douglas.stebila@sun.com>]
3744 *) Change default behaviour of 'openssl asn1parse' so that more
3745 information is visible when viewing, e.g., a certificate:
3747 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3748 mode the content of non-printable OCTET STRINGs is output in a
3749 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3750 avoid the appearance of a printable string.
3751 [Nils Larsch <nla@trustcenter.de>]
3753 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3755 EC_GROUP_set_asn1_flag()
3756 EC_GROUP_get_asn1_flag()
3757 EC_GROUP_set_point_conversion_form()
3758 EC_GROUP_get_point_conversion_form()
3759 These control ASN1 encoding details:
3760 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3761 has been set to OPENSSL_EC_NAMED_CURVE.
3762 - Points are encoded in uncompressed form by default; options for
3763 asn1_for are as for point2oct, namely
3764 POINT_CONVERSION_COMPRESSED
3765 POINT_CONVERSION_UNCOMPRESSED
3766 POINT_CONVERSION_HYBRID
3768 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3771 EC_GROUP_get0_seed()
3772 EC_GROUP_get_seed_len()
3773 This is used only for ASN1 purposes (so far).
3774 [Nils Larsch <nla@trustcenter.de>]
3776 *) Add 'field_type' member to EC_METHOD, which holds the NID
3777 of the appropriate field type OID. The new function
3778 EC_METHOD_get_field_type() returns this value.
3779 [Nils Larsch <nla@trustcenter.de>]
3784 EC_POINT_point2hex()
3785 EC_POINT_hex2point()
3786 providing useful interfaces to EC_POINT_point2oct() and
3787 EC_POINT_oct2point().
3788 [Nils Larsch <nla@trustcenter.de>]
3790 *) Change internals of the EC library so that the functions
3791 EC_GROUP_set_generator()
3792 EC_GROUP_get_generator()
3793 EC_GROUP_get_order()
3794 EC_GROUP_get_cofactor()
3795 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3796 to methods, which would lead to unnecessary code duplication when
3797 adding different types of curves.
3798 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3800 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3801 arithmetic, and such that modified wNAFs are generated
3802 (which avoid length expansion in many cases).
3805 *) Add a function EC_GROUP_check_discriminant() (defined via
3806 EC_METHOD) that verifies that the curve discriminant is non-zero.
3808 Add a function EC_GROUP_check() that makes some sanity tests
3809 on a EC_GROUP, its generator and order. This includes
3810 EC_GROUP_check_discriminant().
3811 [Nils Larsch <nla@trustcenter.de>]
3813 *) Add ECDSA in new directory crypto/ecdsa/.
3815 Add applications 'openssl ecparam' and 'openssl ecdsa'
3816 (these are based on 'openssl dsaparam' and 'openssl dsa').
3818 ECDSA support is also included in various other files across the
3819 library. Most notably,
3820 - 'openssl req' now has a '-newkey ecdsa:file' option;
3821 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3822 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3823 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3824 them suitable for ECDSA where domain parameters must be
3825 extracted before the specific public key;
3826 - ECDSA engine support has been added.
3827 [Nils Larsch <nla@trustcenter.de>]
3829 *) Include some named elliptic curves, and add OIDs from X9.62,
3830 SECG, and WAP/WTLS. Each curve can be obtained from the new
3832 EC_GROUP_new_by_curve_name(),
3833 and the list of available named curves can be obtained with
3834 EC_get_builtin_curves().
3835 Also add a 'curve_name' member to EC_GROUP objects, which can be
3837 EC_GROUP_set_curve_name()
3838 EC_GROUP_get_curve_name()
3839 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3841 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3842 was actually never needed) and in BN_mul(). The removal in BN_mul()
3843 required a small change in bn_mul_part_recursive() and the addition
3844 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3845 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3846 bn_sub_words() and bn_add_words() except they take arrays with
3850 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3852 *) Cleanse PEM buffers before freeing them since they may contain
3854 [Benjamin Bennett <ben@psc.edu>]
3856 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3857 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3858 authentication-only ciphersuites.
3861 *) Since AES128 and AES256 share a single mask bit in the logic of
3862 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3863 kludge to work properly if AES128 is available and AES256 isn't.
3866 *) Expand security boundary to match 1.1.1 module.
3869 *) Remove redundant features: hash file source, editing of test vectors
3870 modify fipsld to use external fips_premain.c signature.
3873 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3874 run algorithm test programs.
3877 *) Make algorithm test programs more tolerant of whitespace.
3880 *) Have SSL/TLS server implementation tolerate "mismatched" record
3881 protocol version while receiving ClientHello even if the
3882 ClientHello is fragmented. (The server can't insist on the
3883 particular protocol version it has chosen before the ServerHello
3884 message has informed the client about his choice.)
3887 *) Load error codes if they are not already present instead of using a
3888 static variable. This allows them to be cleanly unloaded and reloaded.
3891 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3893 *) Introduce limits to prevent malicious keys being able to
3894 cause a denial of service. (CVE-2006-2940)
3895 [Steve Henson, Bodo Moeller]
3897 *) Fix ASN.1 parsing of certain invalid structures that can result
3898 in a denial of service. (CVE-2006-2937) [Steve Henson]
3900 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3901 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3903 *) Fix SSL client code which could crash if connecting to a
3904 malicious SSLv2 server. (CVE-2006-4343)
3905 [Tavis Ormandy and Will Drewry, Google Security Team]
3907 *) Change ciphersuite string processing so that an explicit
3908 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3909 will no longer include "AES128-SHA"), and any other similar
3910 ciphersuite (same bitmap) from *other* protocol versions (so that
3911 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3912 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3913 changes from 0.9.8b and 0.9.8d.
3916 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3918 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3919 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3921 *) Change the Unix randomness entropy gathering to use poll() when
3922 possible instead of select(), since the latter has some
3923 undesirable limitations.
3924 [Darryl Miles via Richard Levitte and Bodo Moeller]
3926 *) Disable rogue ciphersuites:
3928 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3929 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3930 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3932 The latter two were purportedly from
3933 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3936 Also deactive the remaining ciphersuites from
3937 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3938 unofficial, and the ID has long expired.
3941 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3942 dual-core machines) and other potential thread-safety issues.
3945 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3947 *) Adapt fipsld and the build system to link against the validated FIPS
3948 module in FIPS mode.
3951 *) Fixes for VC++ 2005 build under Windows.
3954 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3955 from a Windows bash shell such as MSYS. It is autodetected from the
3956 "config" script when run from a VC++ environment. Modify standard VC++
3957 build to use fipscanister.o from the GNU make build.
3960 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3962 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3963 The value now differs depending on if you build for FIPS or not.
3964 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3965 safely run with a non-FIPSed libcrypto, as it may crash because of
3966 the difference induced by this change.
3969 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3971 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3972 (part of SSL_OP_ALL). This option used to disable the
3973 countermeasure against man-in-the-middle protocol-version
3974 rollback in the SSL 2.0 server implementation, which is a bad
3975 idea. (CVE-2005-2969)
3977 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3978 for Information Security, National Institute of Advanced Industrial
3979 Science and Technology [AIST], Japan)]
3981 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3982 mainly for FIPS compliance and not fully integrated at this stage.
3985 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3986 the exponentiation using a fixed-length exponent. (Otherwise,
3987 the information leaked through timing could expose the secret key
3988 after many signatures; cf. Bleichenbacher's attack on DSA with
3992 *) Make a new fixed-window mod_exp implementation the default for
3993 RSA, DSA, and DH private-key operations so that the sequence of
3994 squares and multiplies and the memory access pattern are
3995 independent of the particular secret key. This will mitigate
3996 cache-timing and potential related attacks.
3998 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3999 and this is automatically used by BN_mod_exp_mont() if the new flag
4000 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4001 will use this BN flag for private exponents unless the flag
4002 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4003 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4005 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4007 *) Change the client implementation for SSLv23_method() and
4008 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4009 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4010 (Previously, the SSL 2.0 backwards compatible Client Hello
4011 message format would be used even with SSL_OP_NO_SSLv2.)
4014 *) Add support for smime-type MIME parameter in S/MIME messages which some
4018 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4019 a threadsafe manner. Modify rsa code to use new function and add calls
4020 to dsa and dh code (which had race conditions before).
4023 *) Include the fixed error library code in the C error file definitions
4024 instead of fixing them up at runtime. This keeps the error code
4025 structures constant.
4028 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4030 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4033 *) Fixes for newer kerberos headers. NB: the casts are needed because
4034 the 'length' field is signed on one version and unsigned on another
4035 with no (?) obvious way to tell the difference, without these VC++
4036 complains. Also the "definition" of FAR (blank) is no longer included
4037 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4038 some needed definitions.
4041 *) Undo Cygwin change.
4044 *) Added support for proxy certificates according to RFC 3820.
4045 Because they may be a security thread to unaware applications,
4046 they must be explicitely allowed in run-time. See
4047 docs/HOWTO/proxy_certificates.txt for further information.
4050 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4052 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4053 server and client random values. Previously
4054 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4055 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4057 This change has negligible security impact because:
4059 1. Server and client random values still have 24 bytes of pseudo random
4062 2. Server and client random values are sent in the clear in the initial
4065 3. The master secret is derived using the premaster secret (48 bytes in
4066 size for static RSA ciphersuites) as well as client server and random
4069 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4072 [Stephen Henson, reported by UK NISCC]
4074 *) Use Windows randomness collection on Cygwin.
4077 *) Fix hang in EGD/PRNGD query when communication socket is closed
4078 prematurely by EGD/PRNGD.
4079 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4081 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4084 *) Back-port of selected performance improvements from development
4085 branch, as well as improved support for PowerPC platforms.
4088 *) Add lots of checks for memory allocation failure, error codes to indicate
4089 failure and freeing up memory if a failure occurs.
4090 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4092 *) Add new -passin argument to dgst.
4095 *) Perform some character comparisons of different types in X509_NAME_cmp:
4096 this is needed for some certificates that reencode DNs into UTF8Strings
4097 (in violation of RFC3280) and can't or wont issue name rollover
4101 *) Make an explicit check during certificate validation to see that
4102 the CA setting in each certificate on the chain is correct. As a
4103 side effect always do the following basic checks on extensions,
4104 not just when there's an associated purpose to the check:
4106 - if there is an unhandled critical extension (unless the user
4107 has chosen to ignore this fault)
4108 - if the path length has been exceeded (if one is set at all)
4109 - that certain extensions fit the associated purpose (if one has
4113 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4115 *) Avoid a race condition when CRLs are checked in a multi threaded
4116 environment. This would happen due to the reordering of the revoked
4117 entries during signature checking and serial number lookup. Now the
4118 encoding is cached and the serial number sort performed under a lock.
4119 Add new STACK function sk_is_sorted().
4122 *) Add Delta CRL to the extension code.
4125 *) Various fixes to s3_pkt.c so alerts are sent properly.
4126 [David Holmes <d.holmes@f5.com>]
4128 *) Reduce the chances of duplicate issuer name and serial numbers (in
4129 violation of RFC3280) using the OpenSSL certificate creation utilities.
4130 This is done by creating a random 64 bit value for the initial serial
4131 number when a serial number file is created or when a self signed
4132 certificate is created using 'openssl req -x509'. The initial serial
4133 number file is created using 'openssl x509 -next_serial' in CA.pl
4134 rather than being initialized to 1.
4137 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4139 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4140 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4141 [Joe Orton, Steve Henson]
4143 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4145 [Joe Orton, Steve Henson]
4147 *) Make it possible to have multiple active certificates with the same
4148 subject in the CA index file. This is done only if the keyword
4149 'unique_subject' is set to 'no' in the main CA section (default
4150 if 'CA_default') of the configuration file. The value is saved
4151 with the database itself in a separate index attribute file,
4152 named like the index file with '.attr' appended to the name.
4155 *) X509 verify fixes. Disable broken certificate workarounds when
4156 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4157 keyUsage extension present. Don't accept CRLs with unhandled critical
4158 extensions: since verify currently doesn't process CRL extensions this
4159 rejects a CRL with *any* critical extensions. Add new verify error codes
4163 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4164 A clarification of RFC2560 will require the use of OCTET STRINGs and
4165 some implementations cannot handle the current raw format. Since OpenSSL
4166 copies and compares OCSP nonces as opaque blobs without any attempt at
4167 parsing them this should not create any compatibility issues.
4170 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4171 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4172 this HMAC (and other) operations are several times slower than OpenSSL
4176 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4177 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4179 *) Use the correct content when signing type "other".
4182 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4184 *) Fix various bugs revealed by running the NISCC test suite:
4186 Stop out of bounds reads in the ASN1 code when presented with
4187 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4189 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4191 If verify callback ignores invalid public key errors don't try to check
4192 certificate signature with the NULL public key.
4196 *) New -ignore_err option in ocsp application to stop the server
4197 exiting on the first error in a request.
4200 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4201 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4205 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4206 extra data after the compression methods not only for TLS 1.0
4207 but also for SSL 3.0 (as required by the specification).
4208 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4210 *) Change X509_certificate_type() to mark the key as exported/exportable
4211 when it's 512 *bits* long, not 512 bytes.
4214 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4215 blocks during encryption.
4218 *) Various fixes to base64 BIO and non blocking I/O. On write
4219 flushes were not handled properly if the BIO retried. On read
4220 data was not being buffered properly and had various logic bugs.
4221 This also affects blocking I/O when the data being decoded is a
4225 *) Various S/MIME bugfixes and compatibility changes:
4226 output correct application/pkcs7 MIME type if
4227 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4228 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4229 of files as .eml work). Correctly handle very long lines in MIME
4233 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4235 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4236 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4237 a protocol version number mismatch like a decryption error
4238 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4241 *) Turn on RSA blinding by default in the default implementation
4242 to avoid a timing attack. Applications that don't want it can call
4243 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4244 They would be ill-advised to do so in most cases.
4245 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4247 *) Change RSA blinding code so that it works when the PRNG is not
4248 seeded (in this case, the secret RSA exponent is abused as
4249 an unpredictable seed -- if it is not unpredictable, there
4250 is no point in blinding anyway). Make RSA blinding thread-safe
4251 by remembering the creator's thread ID in rsa->blinding and
4252 having all other threads use local one-time blinding factors
4253 (this requires more computation than sharing rsa->blinding, but
4254 avoids excessive locking; and if an RSA object is not shared
4255 between threads, blinding will still be very fast).
4258 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4259 ENGINE as defaults for all supported algorithms irrespective of
4260 the 'flags' parameter. 'flags' is now honoured, so applications
4261 should make sure they are passing it correctly.
4264 *) Target "mingw" now allows native Windows code to be generated in
4265 the Cygwin environment as well as with the MinGW compiler.
4268 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4270 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4271 via timing by performing a MAC computation even if incorrrect
4272 block cipher padding has been found. This is a countermeasure
4273 against active attacks where the attacker has to distinguish
4274 between bad padding and a MAC verification error. (CVE-2003-0078)
4276 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4277 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4278 Martin Vuagnoux (EPFL, Ilion)]
4280 *) Make the no-err option work as intended. The intention with no-err
4281 is not to have the whole error stack handling routines removed from
4282 libcrypto, it's only intended to remove all the function name and
4283 reason texts, thereby removing some of the footprint that may not
4284 be interesting if those errors aren't displayed anyway.
4286 NOTE: it's still possible for any application or module to have it's
4287 own set of error texts inserted. The routines are there, just not
4288 used by default when no-err is given.
4291 *) Add support for FreeBSD on IA64.
4292 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4294 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4295 Kerberos function mit_des_cbc_cksum(). Before this change,
4296 the value returned by DES_cbc_cksum() was like the one from
4297 mit_des_cbc_cksum(), except the bytes were swapped.
4298 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4300 *) Allow an application to disable the automatic SSL chain building.
4301 Before this a rather primitive chain build was always performed in
4302 ssl3_output_cert_chain(): an application had no way to send the
4303 correct chain if the automatic operation produced an incorrect result.
4305 Now the chain builder is disabled if either:
4307 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4309 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4311 The reasoning behind this is that an application would not want the
4312 auto chain building to take place if extra chain certificates are
4313 present and it might also want a means of sending no additional
4314 certificates (for example the chain has two certificates and the
4318 *) Add the possibility to build without the ENGINE framework.
4319 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4321 *) Under Win32 gmtime() can return NULL: check return value in
4322 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4325 *) DSA routines: under certain error conditions uninitialized BN objects
4326 could be freed. Solution: make sure initialization is performed early
4327 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4328 Nils Larsch <nla@trustcenter.de> via PR#459)
4331 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4332 checked on reconnect on the client side, therefore session resumption
4333 could still fail with a "ssl session id is different" error. This
4334 behaviour is masked when SSL_OP_ALL is used due to
4335 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4336 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4337 followup to PR #377.
4340 *) IA-32 assembler support enhancements: unified ELF targets, support
4341 for SCO/Caldera platforms, fix for Cygwin shared build.
4344 *) Add support for FreeBSD on sparc64. As a consequence, support for
4345 FreeBSD on non-x86 processors is separate from x86 processors on
4346 the config script, much like the NetBSD support.
4347 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4349 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4351 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4354 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4355 code (06) was taken as the first octet of the session ID and the last
4356 octet was ignored consequently. As a result SSLv2 client side session
4357 caching could not have worked due to the session ID mismatch between
4359 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4363 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4364 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4368 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4369 seems that in spite of existing for more than a year, many application
4370 author have done nothing to provide the necessary callbacks, which
4371 means that this particular engine will not work properly anywhere.
4372 This is a very unfortunate situation which forces us, in the name
4373 of usability, to give the hw_ncipher.c a static lock, which is part
4375 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4376 appear in 0.9.8 or later. We EXPECT application authors to have
4377 dealt properly with this when 0.9.8 is released (unless we actually
4378 make such changes in the libcrypto locking code that changes will
4379 have to be made anyway).
4382 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4383 octets have been read, EOF or an error occurs. Without this change
4384 some truncated ASN1 structures will not produce an error.
4387 *) Disable Heimdal support, since it hasn't been fully implemented.
4388 Still give the possibility to force the use of Heimdal, but with
4389 warnings and a request that patches get sent to openssl-dev.
4392 *) Add the VC-CE target, introduce the WINCE sysname, and add
4393 INSTALL.WCE and appropriate conditionals to make it build.
4394 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4396 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4397 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4398 edit numbers of the version.
4399 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4401 *) Introduce safe string copy and catenation functions
4402 (BUF_strlcpy() and BUF_strlcat()).
4403 [Ben Laurie (CHATS) and Richard Levitte]
4405 *) Avoid using fixed-size buffers for one-line DNs.
4406 [Ben Laurie (CHATS)]
4408 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4409 resizing buffers containing secrets, and use where appropriate.
4410 [Ben Laurie (CHATS)]
4412 *) Avoid using fixed size buffers for configuration file location.
4413 [Ben Laurie (CHATS)]
4415 *) Avoid filename truncation for various CA files.
4416 [Ben Laurie (CHATS)]
4418 *) Use sizeof in preference to magic numbers.
4419 [Ben Laurie (CHATS)]
4421 *) Avoid filename truncation in cert requests.
4422 [Ben Laurie (CHATS)]
4424 *) Add assertions to check for (supposedly impossible) buffer
4426 [Ben Laurie (CHATS)]
4428 *) Don't cache truncated DNS entries in the local cache (this could
4429 potentially lead to a spoofing attack).
4430 [Ben Laurie (CHATS)]
4432 *) Fix various buffers to be large enough for hex/decimal
4433 representations in a platform independent manner.
4434 [Ben Laurie (CHATS)]
4436 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4437 resizing buffers containing secrets, and use where appropriate.
4438 [Ben Laurie (CHATS)]
4440 *) Add BIO_indent() to avoid much slightly worrying code to do
4442 [Ben Laurie (CHATS)]
4444 *) Convert sprintf()/BIO_puts() to BIO_printf().
4445 [Ben Laurie (CHATS)]
4447 *) buffer_gets() could terminate with the buffer only half
4449 [Ben Laurie (CHATS)]
4451 *) Add assertions to prevent user-supplied crypto functions from
4452 overflowing internal buffers by having large block sizes, etc.
4453 [Ben Laurie (CHATS)]
4455 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4457 [Ben Laurie (CHATS)]
4459 *) Eliminate unused copy of key in RC4.
4460 [Ben Laurie (CHATS)]
4462 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4463 [Ben Laurie (CHATS)]
4465 *) Fix off-by-one error in EGD path.
4466 [Ben Laurie (CHATS)]
4468 *) If RANDFILE path is too long, ignore instead of truncating.
4469 [Ben Laurie (CHATS)]
4471 *) Eliminate unused and incorrectly sized X.509 structure
4473 [Ben Laurie (CHATS)]
4475 *) Eliminate unused and dangerous function knumber().
4476 [Ben Laurie (CHATS)]
4478 *) Eliminate unused and dangerous structure, KSSL_ERR.
4479 [Ben Laurie (CHATS)]
4481 *) Protect against overlong session ID context length in an encoded
4482 session object. Since these are local, this does not appear to be
4484 [Ben Laurie (CHATS)]
4486 *) Change from security patch (see 0.9.6e below) that did not affect
4487 the 0.9.6 release series:
4489 Remote buffer overflow in SSL3 protocol - an attacker could
4490 supply an oversized master key in Kerberos-enabled versions.
4492 [Ben Laurie (CHATS)]
4494 *) Change the SSL kerb5 codes to match RFC 2712.
4497 *) Make -nameopt work fully for req and add -reqopt switch.
4498 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4500 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4501 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4503 *) Make sure tests can be performed even if the corresponding algorithms
4504 have been removed entirely. This was also the last step to make
4505 OpenSSL compilable with DJGPP under all reasonable conditions.
4506 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4508 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4509 to allow version independent disabling of normally unselected ciphers,
4510 which may be activated as a side-effect of selecting a single cipher.
4512 (E.g., cipher list string "RSA" enables ciphersuites that are left
4513 out of "ALL" because they do not provide symmetric encryption.
4514 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4515 [Lutz Jaenicke, Bodo Moeller]
4517 *) Add appropriate support for separate platform-dependent build
4518 directories. The recommended way to make a platform-dependent
4519 build directory is the following (tested on Linux), maybe with
4522 # Place yourself outside of the OpenSSL source tree. In
4523 # this example, the environment variable OPENSSL_SOURCE
4524 # is assumed to contain the absolute OpenSSL source directory.
4525 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4526 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4527 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4528 mkdir -p `dirname $F`
4529 ln -s $OPENSSL_SOURCE/$F $F
4532 To be absolutely sure not to disturb the source tree, a "make clean"
4533 is a good thing. If it isn't successfull, don't worry about it,
4534 it probably means the source directory is very clean.
4537 *) Make sure any ENGINE control commands make local copies of string
4538 pointers passed to them whenever necessary. Otherwise it is possible
4539 the caller may have overwritten (or deallocated) the original string
4540 data when a later ENGINE operation tries to use the stored values.
4541 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4543 *) Improve diagnostics in file reading and command-line digests.
4544 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4546 *) Add AES modes CFB and OFB to the object database. Correct an
4547 error in AES-CFB decryption.
4550 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4551 allows existing EVP_CIPHER_CTX structures to be reused after
4552 calling EVP_*Final(). This behaviour is used by encryption
4553 BIOs and some applications. This has the side effect that
4554 applications must explicitly clean up cipher contexts with
4555 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4558 *) Check the values of dna and dnb in bn_mul_recursive before calling
4559 bn_mul_comba (a non zero value means the a or b arrays do not contain
4560 n2 elements) and fallback to bn_mul_normal if either is not zero.
4563 *) Fix escaping of non-ASCII characters when using the -subj option
4564 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4567 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4568 form for "surname", serialNumber has no short form.
4569 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4570 therefore remove "mail" short name for "internet 7".
4571 The OID for unique identifiers in X509 certificates is
4572 x500UniqueIdentifier, not uniqueIdentifier.
4573 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4576 *) Add an "init" command to the ENGINE config module and auto initialize
4577 ENGINEs. Without any "init" command the ENGINE will be initialized
4578 after all ctrl commands have been executed on it. If init=1 the
4579 ENGINE is initailized at that point (ctrls before that point are run
4580 on the uninitialized ENGINE and after on the initialized one). If
4581 init=0 then the ENGINE will not be iniatialized at all.
4584 *) Fix the 'app_verify_callback' interface so that the user-defined
4585 argument is actually passed to the callback: In the
4586 SSL_CTX_set_cert_verify_callback() prototype, the callback
4587 declaration has been changed from
4590 int (*cb)(X509_STORE_CTX *,void *);
4591 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4592 i=s->ctx->app_verify_callback(&ctx)
4593 has been changed into
4594 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4596 To update applications using SSL_CTX_set_cert_verify_callback(),
4597 a dummy argument can be added to their callback functions.
4598 [D. K. Smetters <smetters@parc.xerox.com>]
4600 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4601 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4603 *) Add and OPENSSL_LOAD_CONF define which will cause
4604 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4605 This allows older applications to transparently support certain
4606 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4607 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4608 load the config file and OPENSSL_add_all_algorithms_conf() which will
4609 always load it have also been added.
4612 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4613 Adjust NIDs and EVP layer.
4614 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4616 *) Config modules support in openssl utility.
4618 Most commands now load modules from the config file,
4619 though in a few (such as version) this isn't done
4620 because it couldn't be used for anything.
4622 In the case of ca and req the config file used is
4623 the same as the utility itself: that is the -config
4624 command line option can be used to specify an
4628 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4629 use "openssl_conf" if filename is NULL use default openssl config file.
4632 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4633 config section name. Add a new flag to tolerate a missing config file
4634 and move code to CONF_modules_load_file().
4637 *) Support for crypto accelerator cards from Accelerated Encryption
4638 Processing, www.aep.ie. (Use engine 'aep')
4639 The support was copied from 0.9.6c [engine] and adapted/corrected
4640 to work with the new engine framework.
4641 [AEP Inc. and Richard Levitte]
4643 *) Support for SureWare crypto accelerator cards from Baltimore
4644 Technologies. (Use engine 'sureware')
4645 The support was copied from 0.9.6c [engine] and adapted
4646 to work with the new engine framework.
4649 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4650 make the newer ENGINE framework commands for the CHIL engine work.
4651 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4653 *) Make it possible to produce shared libraries on ReliantUNIX.
4654 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4656 *) Add the configuration target debug-linux-ppro.
4657 Make 'openssl rsa' use the general key loading routines
4658 implemented in apps.c, and make those routines able to
4659 handle the key format FORMAT_NETSCAPE and the variant
4661 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4663 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4664 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4666 *) Add -keyform to rsautl, and document -engine.
4667 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4669 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4670 BIO_R_NO_SUCH_FILE error code rather than the generic
4671 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4674 *) Add new functions
4676 ERR_peek_last_error_line
4677 ERR_peek_last_error_line_data.
4678 These are similar to
4681 ERR_peek_error_line_data,
4682 but report on the latest error recorded rather than the first one
4683 still in the error queue.
4684 [Ben Laurie, Bodo Moeller]
4686 *) default_algorithms option in ENGINE config module. This allows things
4688 default_algorithms = ALL
4689 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4692 *) Prelminary ENGINE config module.
4695 *) New experimental application configuration code.
4698 *) Change the AES code to follow the same name structure as all other
4699 symmetric ciphers, and behave the same way. Move everything to
4700 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4701 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4703 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4704 [Ben Laurie and Theo de Raadt]
4706 *) Add option to output public keys in req command.
4707 [Massimiliano Pala madwolf@openca.org]
4709 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4710 (up to about 10% better than before for P-192 and P-224).
4713 *) New functions/macros
4715 SSL_CTX_set_msg_callback(ctx, cb)
4716 SSL_CTX_set_msg_callback_arg(ctx, arg)
4717 SSL_set_msg_callback(ssl, cb)
4718 SSL_set_msg_callback_arg(ssl, arg)
4720 to request calling a callback function
4722 void cb(int write_p, int version, int content_type,
4723 const void *buf, size_t len, SSL *ssl, void *arg)
4725 whenever a protocol message has been completely received
4726 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4727 protocol version according to which the SSL library interprets
4728 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4729 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4730 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4731 specification (change_cipher_spec(20), alert(21), handshake(22)).
4732 'buf' and 'len' point to the actual message, 'ssl' to the
4733 SSL object, and 'arg' is the application-defined value set by
4734 SSL[_CTX]_set_msg_callback_arg().
4736 'openssl s_client' and 'openssl s_server' have new '-msg' options
4737 to enable a callback that displays all protocol messages.
4740 *) Change the shared library support so shared libraries are built as
4741 soon as the corresponding static library is finished, and thereby get
4742 openssl and the test programs linked against the shared library.
4743 This still only happens when the keyword "shard" has been given to
4744 the configuration scripts.
4746 NOTE: shared library support is still an experimental thing, and
4747 backward binary compatibility is still not guaranteed.
4748 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4750 *) Add support for Subject Information Access extension.
4751 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4753 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4754 additional bytes when new memory had to be allocated, not just
4755 when reusing an existing buffer.
4758 *) New command line and configuration option 'utf8' for the req command.
4759 This allows field values to be specified as UTF8 strings.
4762 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4763 runs for the former and machine-readable output for the latter.
4766 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4767 of the e-mail address in the DN (i.e., it will go into a certificate
4768 extension only). The new configuration file option 'email_in_dn = no'
4769 has the same effect.
4770 [Massimiliano Pala madwolf@openca.org]
4772 *) Change all functions with names starting with des_ to be starting
4773 with DES_ instead. Add wrappers that are compatible with libdes,
4774 but are named _ossl_old_des_*. Finally, add macros that map the
4775 des_* symbols to the corresponding _ossl_old_des_* if libdes
4776 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4777 desired, the des_* symbols will be mapped to DES_*, with one
4780 Since we provide two compatibility mappings, the user needs to
4781 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4782 compatibility is desired. The default (i.e., when that macro
4783 isn't defined) is OpenSSL 0.9.6c compatibility.
4785 There are also macros that enable and disable the support of old
4786 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4787 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4788 are defined, the default will apply: to support the old des routines.
4790 In either case, one must include openssl/des.h to get the correct
4791 definitions. Do not try to just include openssl/des_old.h, that
4794 NOTE: This is a major break of an old API into a new one. Software
4795 authors are encouraged to switch to the DES_ style functions. Some
4796 time in the future, des_old.h and the libdes compatibility functions
4797 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4798 default), and then completely removed.
4801 *) Test for certificates which contain unsupported critical extensions.
4802 If such a certificate is found during a verify operation it is
4803 rejected by default: this behaviour can be overridden by either
4804 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4805 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4806 X509_supported_extension() has also been added which returns 1 if a
4807 particular extension is supported.
4810 *) Modify the behaviour of EVP cipher functions in similar way to digests
4811 to retain compatibility with existing code.
4814 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4815 compatibility with existing code. In particular the 'ctx' parameter does
4816 not have to be to be initialized before the call to EVP_DigestInit() and
4817 it is tidied up after a call to EVP_DigestFinal(). New function
4818 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4819 EVP_MD_CTX_copy() changed to not require the destination to be
4820 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4821 requires the destination to be valid.
4823 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4824 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4827 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4828 so that complete 'Handshake' protocol structures are kept in memory
4829 instead of overwriting 'msg_type' and 'length' with 'body' data.
4832 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4833 [Massimo Santin via Richard Levitte]
4835 *) Major restructuring to the underlying ENGINE code. This includes
4836 reduction of linker bloat, separation of pure "ENGINE" manipulation
4837 (initialisation, etc) from functionality dealing with implementations
4838 of specific crypto iterfaces. This change also introduces integrated
4839 support for symmetric ciphers and digest implementations - so ENGINEs
4840 can now accelerate these by providing EVP_CIPHER and EVP_MD
4841 implementations of their own. This is detailed in crypto/engine/README
4842 as it couldn't be adequately described here. However, there are a few
4843 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4844 were changed in the original introduction of ENGINE code have now
4845 reverted back - the hooking from this code to ENGINE is now a good
4846 deal more passive and at run-time, operations deal directly with
4847 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4848 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4849 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4850 they were not being used by the framework as there is no concept of a
4851 BIGNUM_METHOD and they could not be generalised to the new
4852 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4853 ENGINE_cpy() has been removed as it cannot be consistently defined in
4857 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4860 *) Change mkdef.pl to sort symbols that get the same entry number,
4861 and make sure the automatically generated functions ERR_load_*
4862 become part of libeay.num as well.
4865 *) New function SSL_renegotiate_pending(). This returns true once
4866 renegotiation has been requested (either SSL_renegotiate() call
4867 or HelloRequest/ClientHello receveived from the peer) and becomes
4868 false once a handshake has been completed.
4869 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4870 sends a HelloRequest, but does not ensure that a handshake takes
4871 place. SSL_renegotiate_pending() is useful for checking if the
4872 client has followed the request.)
4875 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4876 By default, clients may request session resumption even during
4877 renegotiation (if session ID contexts permit); with this option,
4878 session resumption is possible only in the first handshake.
4880 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4881 more bits available for options that should not be part of
4882 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4885 *) Add some demos for certificate and certificate request creation.
4888 *) Make maximum certificate chain size accepted from the peer application
4889 settable (SSL*_get/set_max_cert_list()), as proposed by
4890 "Douglas E. Engert" <deengert@anl.gov>.
4893 *) Add support for shared libraries for Unixware-7
4894 (Boyd Lynn Gerber <gerberb@zenez.com>).
4897 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4898 be done prior to destruction. Use this to unload error strings from
4899 ENGINEs that load their own error strings. NB: This adds two new API
4900 functions to "get" and "set" this destroy handler in an ENGINE.
4903 *) Alter all existing ENGINE implementations (except "openssl" and
4904 "openbsd") to dynamically instantiate their own error strings. This
4905 makes them more flexible to be built both as statically-linked ENGINEs
4906 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4907 Also, add stub code to each that makes building them as self-contained
4908 shared-libraries easier (see README.ENGINE).
4911 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4912 implementations into applications that are completely implemented in
4913 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4914 commands that can be used to configure what shared-library to load and
4915 to control aspects of the way it is handled. Also, made an update to
4916 the README.ENGINE file that brings its information up-to-date and
4917 provides some information and instructions on the "dynamic" ENGINE
4918 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4921 *) Make it possible to unload ranges of ERR strings with a new
4922 "ERR_unload_strings" function.
4925 *) Add a copy() function to EVP_MD.
4928 *) Make EVP_MD routines take a context pointer instead of just the
4929 md_data void pointer.
4932 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4933 that the digest can only process a single chunk of data
4934 (typically because it is provided by a piece of
4935 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4936 is only going to provide a single chunk of data, and hence the
4937 framework needn't accumulate the data for oneshot drivers.
4940 *) As with "ERR", make it possible to replace the underlying "ex_data"
4941 functions. This change also alters the storage and management of global
4942 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4943 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4944 index counters. The API functions that use this state have been changed
4945 to take a "class_index" rather than pointers to the class's local STACK
4946 and counter, and there is now an API function to dynamically create new
4947 classes. This centralisation allows us to (a) plug a lot of the
4948 thread-safety problems that existed, and (b) makes it possible to clean
4949 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4950 such data would previously have always leaked in application code and
4951 workarounds were in place to make the memory debugging turn a blind eye
4952 to it. Application code that doesn't use this new function will still
4953 leak as before, but their memory debugging output will announce it now
4954 rather than letting it slide.
4956 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4957 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4958 has a return value to indicate success or failure.
4961 *) Make it possible to replace the underlying "ERR" functions such that the
4962 global state (2 LHASH tables and 2 locks) is only used by the "default"
4963 implementation. This change also adds two functions to "get" and "set"
4964 the implementation prior to it being automatically set the first time
4965 any other ERR function takes place. Ie. an application can call "get",
4966 pass the return value to a module it has just loaded, and that module
4967 can call its own "set" function using that value. This means the
4968 module's "ERR" operations will use (and modify) the error state in the
4969 application and not in its own statically linked copy of OpenSSL code.
4972 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4973 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4974 the operation, and provides a more encapsulated way for external code
4975 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4976 to use these functions rather than manually incrementing the counts.
4978 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4981 *) Add EVP test program.
4984 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4987 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4988 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4989 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4990 These allow a CRL to be built without having to access X509_CRL fields
4991 directly. Modify 'ca' application to use new functions.
4994 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4995 bug workarounds. Rollback attack detection is a security feature.
4996 The problem will only arise on OpenSSL servers when TLSv1 is not
4997 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4998 Software authors not wanting to support TLSv1 will have special reasons
4999 for their choice and can explicitly enable this option.
5000 [Bodo Moeller, Lutz Jaenicke]
5002 *) Rationalise EVP so it can be extended: don't include a union of
5003 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5004 (similar to those existing for EVP_CIPHER_CTX).
5009 EVP_MD_CTX_init(&md); /* new function call */
5010 EVP_DigestInit(&md, EVP_sha1());
5011 EVP_DigestUpdate(&md, in, len);
5012 EVP_DigestFinal(&md, out, NULL);
5013 EVP_MD_CTX_cleanup(&md); /* new function call */
5017 *) Make DES key schedule conform to the usual scheme, as well as
5018 correcting its structure. This means that calls to DES functions
5019 now have to pass a pointer to a des_key_schedule instead of a
5020 plain des_key_schedule (which was actually always a pointer
5023 des_key_schedule ks;
5025 des_set_key_checked(..., &ks);
5026 des_ncbc_encrypt(..., &ks, ...);
5028 (Note that a later change renames 'des_...' into 'DES_...'.)
5031 *) Initial reduction of linker bloat: the use of some functions, such as
5032 PEM causes large amounts of unused functions to be linked in due to
5033 poor organisation. For example pem_all.c contains every PEM function
5034 which has a knock on effect of linking in large amounts of (unused)
5035 ASN1 code. Grouping together similar functions and splitting unrelated
5036 functions prevents this.
5039 *) Cleanup of EVP macros.
5042 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5043 correct _ecb suffix.
5046 *) Add initial OCSP responder support to ocsp application. The
5047 revocation information is handled using the text based index
5048 use by the ca application. The responder can either handle
5049 requests generated internally, supplied in files (for example
5050 via a CGI script) or using an internal minimal server.
5053 *) Add configuration choices to get zlib compression for TLS.
5056 *) Changes to Kerberos SSL for RFC 2712 compliance:
5057 1. Implemented real KerberosWrapper, instead of just using
5058 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5059 2. Implemented optional authenticator field of KerberosWrapper.
5061 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5062 and authenticator structs; see crypto/krb5/.
5064 Generalized Kerberos calls to support multiple Kerberos libraries.
5065 [Vern Staats <staatsvr@asc.hpc.mil>,
5066 Jeffrey Altman <jaltman@columbia.edu>
5067 via Richard Levitte]
5069 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5070 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5071 values for each of the key sizes rather than having just
5072 parameters (and 'speed' generating keys each time).
5075 *) Speed up EVP routines.
5078 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5079 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5080 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5081 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5083 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5084 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5085 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5088 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5090 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5093 *) Added the OS2-EMX target.
5094 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5096 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5097 to support NCONF routines in extension code. New function CONF_set_nconf()
5098 to allow functions which take an NCONF to also handle the old LHASH
5099 structure: this means that the old CONF compatible routines can be
5100 retained (in particular wrt extensions) without having to duplicate the
5101 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5104 *) Enhance the general user interface with mechanisms for inner control
5105 and with possibilities to have yes/no kind of prompts.
5108 *) Change all calls to low level digest routines in the library and
5109 applications to use EVP. Add missing calls to HMAC_cleanup() and
5110 don't assume HMAC_CTX can be copied using memcpy().
5111 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5113 *) Add the possibility to control engines through control names but with
5114 arbitrary arguments instead of just a string.
5115 Change the key loaders to take a UI_METHOD instead of a callback
5116 function pointer. NOTE: this breaks binary compatibility with earlier
5117 versions of OpenSSL [engine].
5118 Adapt the nCipher code for these new conditions and add a card insertion
5122 *) Enhance the general user interface with mechanisms to better support
5123 dialog box interfaces, application-defined prompts, the possibility
5124 to use defaults (for example default passwords from somewhere else)
5125 and interrupts/cancellations.
5128 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5129 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5132 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5133 tidy up some unnecessarily weird code in 'sk_new()').
5134 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5136 *) Change the key loading routines for ENGINEs to use the same kind
5137 callback (pem_password_cb) as all other routines that need this
5141 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5142 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5143 than this minimum value is recommended.
5146 *) New random seeder for OpenVMS, using the system process statistics
5147 that are easily reachable.
5150 *) Windows apparently can't transparently handle global
5151 variables defined in DLLs. Initialisations such as:
5153 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5155 wont compile. This is used by the any applications that need to
5156 declare their own ASN1 modules. This was fixed by adding the option
5157 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5158 needed for static libraries under Win32.
5161 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5162 setting of purpose and trust fields. New X509_STORE trust and
5163 purpose functions and tidy up setting in other SSL functions.
5166 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5167 structure. These are inherited by X509_STORE_CTX when it is
5168 initialised. This allows various defaults to be set in the
5169 X509_STORE structure (such as flags for CRL checking and custom
5170 purpose or trust settings) for functions which only use X509_STORE_CTX
5171 internally such as S/MIME.
5173 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5174 trust settings if they are not set in X509_STORE. This allows X509_STORE
5175 purposes and trust (in S/MIME for example) to override any set by default.
5177 Add command line options for CRL checking to smime, s_client and s_server
5181 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5182 are set then the CRL is looked up in the X509_STORE structure and
5183 its validity and signature checked, then if the certificate is found
5184 in the CRL the verify fails with a revoked error.
5186 Various new CRL related callbacks added to X509_STORE_CTX structure.
5188 Command line options added to 'verify' application to support this.
5190 This needs some additional work, such as being able to handle multiple
5191 CRLs with different times, extension based lookup (rather than just
5192 by subject name) and ultimately more complete V2 CRL extension
5196 *) Add a general user interface API (crypto/ui/). This is designed
5197 to replace things like des_read_password and friends (backward
5198 compatibility functions using this new API are provided).
5199 The purpose is to remove prompting functions from the DES code
5200 section as well as provide for prompting through dialog boxes in
5201 a window system and the like.
5204 *) Add "ex_data" support to ENGINE so implementations can add state at a
5205 per-structure level rather than having to store it globally.
5208 *) Make it possible for ENGINE structures to be copied when retrieved by
5209 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5210 This causes the "original" ENGINE structure to act like a template,
5211 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5212 operational state can be localised to each ENGINE structure, despite the
5213 fact they all share the same "methods". New ENGINE structures returned in
5214 this case have no functional references and the return value is the single
5215 structural reference. This matches the single structural reference returned
5216 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5220 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5221 needs to match any other type at all we need to manually clear the
5225 *) Changes to the "openssl engine" utility to include;
5226 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5227 about an ENGINE's available control commands.
5228 - executing control commands from command line arguments using the
5229 '-pre' and '-post' switches. '-post' is only used if '-t' is
5230 specified and the ENGINE is successfully initialised. The syntax for
5231 the individual commands are colon-separated, for example;
5232 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5235 *) New dynamic control command support for ENGINEs. ENGINEs can now
5236 declare their own commands (numbers), names (strings), descriptions,
5237 and input types for run-time discovery by calling applications. A
5238 subset of these commands are implicitly classed as "executable"
5239 depending on their input type, and only these can be invoked through
5240 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5241 can be based on user input, config files, etc). The distinction is
5242 that "executable" commands cannot return anything other than a boolean
5243 result and can only support numeric or string input, whereas some
5244 discoverable commands may only be for direct use through
5245 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5246 pointers, or other custom uses. The "executable" commands are to
5247 support parameterisations of ENGINE behaviour that can be
5248 unambiguously defined by ENGINEs and used consistently across any
5249 OpenSSL-based application. Commands have been added to all the
5250 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5251 control over shared-library paths without source code alterations.
5254 *) Changed all ENGINE implementations to dynamically allocate their
5255 ENGINEs rather than declaring them statically. Apart from this being
5256 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5257 this also allows the implementations to compile without using the
5258 internal engine_int.h header.
5261 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5262 'const' value. Any code that should be able to modify a RAND_METHOD
5263 should already have non-const pointers to it (ie. they should only
5264 modify their own ones).
5267 *) Made a variety of little tweaks to the ENGINE code.
5268 - "atalla" and "ubsec" string definitions were moved from header files
5269 to C code. "nuron" string definitions were placed in variables
5270 rather than hard-coded - allowing parameterisation of these values
5271 later on via ctrl() commands.
5272 - Removed unused "#if 0"'d code.
5273 - Fixed engine list iteration code so it uses ENGINE_free() to release
5274 structural references.
5275 - Constified the RAND_METHOD element of ENGINE structures.
5276 - Constified various get/set functions as appropriate and added
5277 missing functions (including a catch-all ENGINE_cpy that duplicates
5278 all ENGINE values onto a new ENGINE except reference counts/state).
5279 - Removed NULL parameter checks in get/set functions. Setting a method
5280 or function to NULL is a way of cancelling out a previously set
5281 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5282 and doesn't justify the extra error symbols and code.
5283 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5284 flags from engine_int.h to engine.h.
5285 - Changed prototypes for ENGINE handler functions (init(), finish(),
5286 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5289 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5290 to the algorithm using long division. The binary algorithm can be
5291 used only if the modulus is odd. On 32-bit systems, it is faster
5292 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5293 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5294 up to 450 bits. In 64-bit environments, the binary algorithm
5295 appears to be advantageous for much longer moduli; here we use it
5296 for moduli up to 2048 bits.
5299 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5300 could not support the combine flag in choice fields.
5303 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5304 extensions from a certificate request to the certificate.
5307 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5308 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5309 file: this allows the display of the certificate about to be
5310 signed to be customised, to allow certain fields to be included
5311 or excluded and extension details. The old system didn't display
5312 multicharacter strings properly, omitted fields not in the policy
5313 and couldn't display additional details such as extensions.
5316 *) Function EC_POINTs_mul for multiple scalar multiplication
5317 of an arbitrary number of elliptic curve points
5318 \sum scalars[i]*points[i],
5319 optionally including the generator defined for the EC_GROUP:
5320 scalar*generator + \sum scalars[i]*points[i].
5322 EC_POINT_mul is a simple wrapper function for the typical case
5323 that the point list has just one item (besides the optional
5327 *) First EC_METHODs for curves over GF(p):
5329 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5330 operations and provides various method functions that can also
5331 operate with faster implementations of modular arithmetic.
5333 EC_GFp_mont_method() reuses most functions that are part of
5334 EC_GFp_simple_method, but uses Montgomery arithmetic.
5336 [Bodo Moeller; point addition and point doubling
5337 implementation directly derived from source code provided by
5338 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5340 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5341 crypto/ec/ec_lib.c):
5343 Curves are EC_GROUP objects (with an optional group generator)
5344 based on EC_METHODs that are built into the library.
5346 Points are EC_POINT objects based on EC_GROUP objects.
5348 Most of the framework would be able to handle curves over arbitrary
5349 finite fields, but as there are no obvious types for fields other
5350 than GF(p), some functions are limited to that for now.
5353 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5354 that the file contains a complete HTTP response.
5357 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5358 change the def and num file printf format specifier from "%-40sXXX"
5359 to "%-39s XXX". The latter will always guarantee a space after the
5360 field while the former will cause them to run together if the field
5361 is 40 of more characters long.
5364 *) Constify the cipher and digest 'method' functions and structures
5365 and modify related functions to take constant EVP_MD and EVP_CIPHER
5369 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5370 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5373 *) Modify EVP_Digest*() routines so they now return values. Although the
5374 internal software routines can never fail additional hardware versions
5378 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5380 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5381 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5384 ERR_R_NESTED_ASN1_ERROR
5386 ERR_R_MISSING_ASN1_EOS
5387 were 4 .. 9, conflicting with
5388 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5390 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5391 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5393 Add new error code 'ERR_R_INTERNAL_ERROR'.
5396 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5400 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5401 sets the subject name for a new request or supersedes the
5402 subject name in a given request. Formats that can be parsed are
5403 'CN=Some Name, OU=myOU, C=IT'
5405 'CN=Some Name/OU=myOU/C=IT'.
5407 Add options '-batch' and '-verbose' to 'openssl req'.
5408 [Massimiliano Pala <madwolf@hackmasters.net>]
5410 *) Introduce the possibility to access global variables through
5411 functions on platform were that's the best way to handle exporting
5412 global variables in shared libraries. To enable this functionality,
5413 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5414 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5415 is normally done by Configure or something similar).
5417 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5418 in the source file (foo.c) like this:
5420 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5421 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5423 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5424 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5426 OPENSSL_DECLARE_GLOBAL(int,foo);
5427 #define foo OPENSSL_GLOBAL_REF(foo)
5428 OPENSSL_DECLARE_GLOBAL(double,bar);
5429 #define bar OPENSSL_GLOBAL_REF(bar)
5431 The #defines are very important, and therefore so is including the
5432 header file everywhere where the defined globals are used.
5434 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5435 of ASN.1 items, but that structure is a bit different.
5437 The largest change is in util/mkdef.pl which has been enhanced with
5438 better and easier to understand logic to choose which symbols should
5439 go into the Windows .def files as well as a number of fixes and code
5440 cleanup (among others, algorithm keywords are now sorted
5441 lexicographically to avoid constant rewrites).
5444 *) In BN_div() keep a copy of the sign of 'num' before writing the
5445 result to 'rm' because if rm==num the value will be overwritten
5446 and produce the wrong result if 'num' is negative: this caused
5447 problems with BN_mod() and BN_nnmod().
5450 *) Function OCSP_request_verify(). This checks the signature on an
5451 OCSP request and verifies the signer certificate. The signer
5452 certificate is just checked for a generic purpose and OCSP request
5456 *) Add OCSP_check_validity() function to check the validity of OCSP
5457 responses. OCSP responses are prepared in real time and may only
5458 be a few seconds old. Simply checking that the current time lies
5459 between thisUpdate and nextUpdate max reject otherwise valid responses
5460 caused by either OCSP responder or client clock inaccuracy. Instead
5461 we allow thisUpdate and nextUpdate to fall within a certain period of
5462 the current time. The age of the response can also optionally be
5463 checked. Two new options -validity_period and -status_age added to
5467 *) If signature or public key algorithm is unrecognized print out its
5468 OID rather that just UNKNOWN.
5471 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5472 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5473 ID to be generated from the issuer certificate alone which can then be
5474 passed to OCSP_id_issuer_cmp().
5477 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5478 ASN1 modules to export functions returning ASN1_ITEM pointers
5479 instead of the ASN1_ITEM structures themselves. This adds several
5480 new macros which allow the underlying ASN1 function/structure to
5481 be accessed transparently. As a result code should not use ASN1_ITEM
5482 references directly (such as &X509_it) but instead use the relevant
5483 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5484 use of the new ASN1 code on platforms where exporting structures
5485 is problematical (for example in shared libraries) but exporting
5486 functions returning pointers to structures is not.
5489 *) Add support for overriding the generation of SSL/TLS session IDs.
5490 These callbacks can be registered either in an SSL_CTX or per SSL.
5491 The purpose of this is to allow applications to control, if they wish,
5492 the arbitrary values chosen for use as session IDs, particularly as it
5493 can be useful for session caching in multiple-server environments. A
5494 command-line switch for testing this (and any client code that wishes
5495 to use such a feature) has been added to "s_server".
5496 [Geoff Thorpe, Lutz Jaenicke]
5498 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5499 of the form '#if defined(...) || defined(...) || ...' and
5500 '#if !defined(...) && !defined(...) && ...'. This also avoids
5501 the growing number of special cases it was previously handling.
5504 *) Make all configuration macros available for application by making
5505 sure they are available in opensslconf.h, by giving them names starting
5506 with "OPENSSL_" to avoid conflicts with other packages and by making
5507 sure e_os2.h will cover all platform-specific cases together with
5509 Additionally, it is now possible to define configuration/platform-
5510 specific names (called "system identities"). In the C code, these
5511 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5512 macro with the name beginning with "OPENSSL_SYS_", which is determined
5513 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5517 *) New option -set_serial to 'req' and 'x509' this allows the serial
5518 number to use to be specified on the command line. Previously self
5519 signed certificates were hard coded with serial number 0 and the
5520 CA options of 'x509' had to use a serial number in a file which was
5524 *) New options to 'ca' utility to support V2 CRL entry extensions.
5525 Currently CRL reason, invalidity date and hold instruction are
5526 supported. Add new CRL extensions to V3 code and some new objects.
5529 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5530 disable standard block padding (aka PKCS#5 padding) in the EVP
5531 API, which was previously mandatory. This means that the data is
5532 not padded in any way and so the total length much be a multiple
5533 of the block size, otherwise an error occurs.
5536 *) Initial (incomplete) OCSP SSL support.
5539 *) New function OCSP_parse_url(). This splits up a URL into its host,
5540 port and path components: primarily to parse OCSP URLs. New -url
5541 option to ocsp utility.
5544 *) New nonce behavior. The return value of OCSP_check_nonce() now
5545 reflects the various checks performed. Applications can decide
5546 whether to tolerate certain situations such as an absent nonce
5547 in a response when one was present in a request: the ocsp application
5548 just prints out a warning. New function OCSP_add1_basic_nonce()
5549 this is to allow responders to include a nonce in a response even if
5550 the request is nonce-less.
5553 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5554 skipped when using openssl x509 multiple times on a single input file,
5555 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5558 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5559 set string type: to handle setting ASN1_TIME structures. Fix ca
5560 utility to correctly initialize revocation date of CRLs.
5563 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5564 the clients preferred ciphersuites and rather use its own preferences.
5565 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5566 Internet Explorer by ensuring unchanged hash method during stepup.
5567 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5570 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5571 to aes and add a new 'exist' option to print out symbols that don't
5575 *) Additional options to ocsp utility to allow flags to be set and
5576 additional certificates supplied.
5579 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5580 OCSP client a number of certificate to only verify the response
5584 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5585 handle the new API. Currently only ECB, CBC modes supported. Add new
5588 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5589 Encryption Standard (AES) Ciphersuites for Transport Layer
5590 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5591 not enabled by default and were not part of the "ALL" ciphersuite
5592 alias because they were not yet official; they could be
5593 explicitly requested by specifying the "AESdraft" ciphersuite
5594 group alias. In the final release of OpenSSL 0.9.7, the group
5595 alias is called "AES" and is part of "ALL".)
5596 [Ben Laurie, Steve Henson, Bodo Moeller]
5598 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5599 request to response.
5602 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5603 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5604 extract information from a certificate request. OCSP_response_create()
5605 creates a response and optionally adds a basic response structure.
5606 OCSP_basic_add1_status() adds a complete single response to a basic
5607 response and returns the OCSP_SINGLERESP structure just added (to allow
5608 extensions to be included for example). OCSP_basic_add1_cert() adds a
5609 certificate to a basic response and OCSP_basic_sign() signs a basic
5610 response with various flags. New helper functions ASN1_TIME_check()
5611 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5612 (converts ASN1_TIME to GeneralizedTime).
5615 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5616 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5617 structure from a certificate. X509_pubkey_digest() digests the public_key
5618 contents: this is used in various key identifiers.
5621 *) Make sk_sort() tolerate a NULL argument.
5622 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5624 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5625 passed by the function are trusted implicitly. If any of them signed the
5626 response then it is assumed to be valid and is not verified.
5629 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5630 to data. This was previously part of the PKCS7 ASN1 code. This
5631 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5632 [Steve Henson, reported by Kenneth R. Robinette
5633 <support@securenetterm.com>]
5635 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5636 routines: without these tracing memory leaks is very painful.
5637 Fix leaks in PKCS12 and PKCS7 routines.
5640 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5641 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5642 effectively meant GeneralizedTime would never be used. Now it
5643 is initialised to -1 but X509_time_adj() now has to check the value
5644 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5645 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5646 [Steve Henson, reported by Kenneth R. Robinette
5647 <support@securenetterm.com>]
5649 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5650 result in a zero length in the ASN1_INTEGER structure which was
5651 not consistent with the structure when d2i_ASN1_INTEGER() was used
5652 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5653 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5654 where it did not print out a minus for negative ASN1_INTEGER.
5657 *) Add summary printout to ocsp utility. The various functions which
5658 convert status values to strings have been renamed to:
5659 OCSP_response_status_str(), OCSP_cert_status_str() and
5660 OCSP_crl_reason_str() and are no longer static. New options
5661 to verify nonce values and to disable verification. OCSP response
5662 printout format cleaned up.
5665 *) Add additional OCSP certificate checks. These are those specified
5666 in RFC2560. This consists of two separate checks: the CA of the
5667 certificate being checked must either be the OCSP signer certificate
5668 or the issuer of the OCSP signer certificate. In the latter case the
5669 OCSP signer certificate must contain the OCSP signing extended key
5670 usage. This check is performed by attempting to match the OCSP
5671 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5672 in the OCSP_CERTID structures of the response.
5675 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5676 and related routines. This uses the standard OpenSSL certificate
5677 verify routines to perform initial checks (just CA validity) and
5678 to obtain the certificate chain. Then additional checks will be
5679 performed on the chain. Currently the root CA is checked to see
5680 if it is explicitly trusted for OCSP signing. This is used to set
5681 a root CA as a global signing root: that is any certificate that
5682 chains to that CA is an acceptable OCSP signing certificate.
5685 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5686 extensions from a separate configuration file.
5687 As when reading extensions from the main configuration file,
5688 the '-extensions ...' option may be used for specifying the
5690 [Massimiliano Pala <madwolf@comune.modena.it>]
5692 *) New OCSP utility. Allows OCSP requests to be generated or
5693 read. The request can be sent to a responder and the output
5694 parsed, outputed or printed in text form. Not complete yet:
5695 still needs to check the OCSP response validity.
5698 *) New subcommands for 'openssl ca':
5699 'openssl ca -status <serial>' prints the status of the cert with
5700 the given serial number (according to the index file).
5701 'openssl ca -updatedb' updates the expiry status of certificates
5703 [Massimiliano Pala <madwolf@comune.modena.it>]
5705 *) New '-newreq-nodes' command option to CA.pl. This is like
5706 '-newreq', but calls 'openssl req' with the '-nodes' option
5707 so that the resulting key is not encrypted.
5708 [Damien Miller <djm@mindrot.org>]
5710 *) New configuration for the GNU Hurd.
5711 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5713 *) Initial code to implement OCSP basic response verify. This
5714 is currently incomplete. Currently just finds the signer's
5715 certificate and verifies the signature on the response.
5718 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5719 value of OPENSSLDIR. This is available via the new '-d' option
5720 to 'openssl version', and is also included in 'openssl version -a'.
5723 *) Allowing defining memory allocation callbacks that will be given
5724 file name and line number information in additional arguments
5725 (a const char* and an int). The basic functionality remains, as
5726 well as the original possibility to just replace malloc(),
5727 realloc() and free() by functions that do not know about these
5728 additional arguments. To register and find out the current
5729 settings for extended allocation functions, the following
5730 functions are provided:
5732 CRYPTO_set_mem_ex_functions
5733 CRYPTO_set_locked_mem_ex_functions
5734 CRYPTO_get_mem_ex_functions
5735 CRYPTO_get_locked_mem_ex_functions
5737 These work the same way as CRYPTO_set_mem_functions and friends.
5738 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5739 extended allocation function is enabled.
5740 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5741 a conventional allocation function is enabled.
5742 [Richard Levitte, Bodo Moeller]
5744 *) Finish off removing the remaining LHASH function pointer casts.
5745 There should no longer be any prototype-casting required when using
5746 the LHASH abstraction, and any casts that remain are "bugs". See
5747 the callback types and macros at the head of lhash.h for details
5748 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5751 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5752 If /dev/[u]random devices are not available or do not return enough
5753 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5755 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5756 /etc/entropy will be queried once each in this sequence, quering stops
5757 when enough entropy was collected without querying more sockets.
5760 *) Change the Unix RAND_poll() variant to be able to poll several
5761 random devices, as specified by DEVRANDOM, until a sufficient amount
5762 of data has been collected. We spend at most 10 ms on each file
5763 (select timeout) and read in non-blocking mode. DEVRANDOM now
5764 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5765 (previously it was just the string "/dev/urandom"), so on typical
5766 platforms the 10 ms delay will never occur.
5767 Also separate out the Unix variant to its own file, rand_unix.c.
5768 For VMS, there's a currently-empty rand_vms.c.
5771 *) Move OCSP client related routines to ocsp_cl.c. These
5772 provide utility functions which an application needing
5773 to issue a request to an OCSP responder and analyse the
5774 response will typically need: as opposed to those which an
5775 OCSP responder itself would need which will be added later.
5777 OCSP_request_sign() signs an OCSP request with an API similar
5778 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5779 response. OCSP_response_get1_basic() extracts basic response
5780 from response. OCSP_resp_find_status(): finds and extracts status
5781 information from an OCSP_CERTID structure (which will be created
5782 when the request structure is built). These are built from lower
5783 level functions which work on OCSP_SINGLERESP structures but
5784 wont normally be used unless the application wishes to examine
5785 extensions in the OCSP response for example.
5787 Replace nonce routines with a pair of functions.
5788 OCSP_request_add1_nonce() adds a nonce value and optionally
5789 generates a random value. OCSP_check_nonce() checks the
5790 validity of the nonce in an OCSP response.
5793 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5794 This doesn't copy the supplied OCSP_CERTID and avoids the
5795 need to free up the newly created id. Change return type
5796 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5797 This can then be used to add extensions to the request.
5798 Deleted OCSP_request_new(), since most of its functionality
5799 is now in OCSP_REQUEST_new() (and the case insensitive name
5800 clash) apart from the ability to set the request name which
5801 will be added elsewhere.
5804 *) Update OCSP API. Remove obsolete extensions argument from
5805 various functions. Extensions are now handled using the new
5806 OCSP extension code. New simple OCSP HTTP function which
5807 can be used to send requests and parse the response.
5810 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5811 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5812 uses the special reorder version of SET OF to sort the attributes
5813 and reorder them to match the encoded order. This resolves a long
5814 standing problem: a verify on a PKCS7 structure just after signing
5815 it used to fail because the attribute order did not match the
5816 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5817 it uses the received order. This is necessary to tolerate some broken
5818 software that does not order SET OF. This is handled by encoding
5819 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5820 to produce the required SET OF.
5823 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5824 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5825 files to get correct declarations of the ASN.1 item variables.
5828 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5829 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5830 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5831 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5832 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5833 ASN1_ITEM and no wrapper functions.
5836 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5837 replace the old function pointer based I/O routines. Change most of
5838 the *_d2i_bio() and *_d2i_fp() functions to use these.
5841 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5842 lines, recognice more "algorithms" that can be deselected, and make
5843 it complain about algorithm deselection that isn't recognised.
5846 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5847 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5848 to use new functions. Add NO_ASN1_OLD which can be set to remove
5849 some old style ASN1 functions: this can be used to determine if old
5850 code will still work when these eventually go away.
5853 *) New extension functions for OCSP structures, these follow the
5854 same conventions as certificates and CRLs.
5857 *) New function X509V3_add1_i2d(). This automatically encodes and
5858 adds an extension. Its behaviour can be customised with various
5859 flags to append, replace or delete. Various wrappers added for
5860 certifcates and CRLs.
5863 *) Fix to avoid calling the underlying ASN1 print routine when
5864 an extension cannot be parsed. Correct a typo in the
5865 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5868 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5869 entries for variables.
5872 *) Add functionality to apps/openssl.c for detecting locking
5873 problems: As the program is single-threaded, all we have
5874 to do is register a locking callback using an array for
5875 storing which locks are currently held by the program.
5878 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5879 SSL_get_ex_data_X509_STORE_idx(), which is used in
5880 ssl_verify_cert_chain() and thus can be called at any time
5881 during TLS/SSL handshakes so that thread-safety is essential.
5882 Unfortunately, the ex_data design is not at all suited
5883 for multi-threaded use, so it probably should be abolished.
5886 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5887 [Broadcom, tweaked and integrated by Geoff Thorpe]
5889 *) Move common extension printing code to new function
5890 X509V3_print_extensions(). Reorganise OCSP print routines and
5891 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5894 *) New function X509_signature_print() to remove duplication in some
5898 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5899 set (this was treated exactly the same as SET OF previously). This
5900 is used to reorder the STACK representing the structure to match the
5901 encoding. This will be used to get round a problem where a PKCS7
5902 structure which was signed could not be verified because the STACK
5903 order did not reflect the encoded order.
5906 *) Reimplement the OCSP ASN1 module using the new code.
5909 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5910 for its ASN1 operations. The old style function pointers still exist
5911 for now but they will eventually go away.
5914 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5915 completely replaces the old ASN1 functionality with a table driven
5916 encoder and decoder which interprets an ASN1_ITEM structure describing
5917 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5918 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5919 has also been converted to the new form.
5922 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5923 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5924 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5925 for negative moduli.
5928 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5929 of not touching the result's sign bit.
5932 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5936 *) Changed the LHASH code to use prototypes for callbacks, and created
5937 macros to declare and implement thin (optionally static) functions
5938 that provide type-safety and avoid function pointer casting for the
5939 type-specific callbacks.
5942 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5944 [Veers Staats <staatsvr@asc.hpc.mil>,
5945 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5947 *) Reformat the FAQ so the different questions and answers can be divided
5948 in sections depending on the subject.
5951 *) Have the zlib compression code load ZLIB.DLL dynamically under
5955 *) New function BN_mod_sqrt for computing square roots modulo a prime
5956 (using the probabilistic Tonelli-Shanks algorithm unless
5957 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5958 be handled deterministically).
5959 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5961 *) Make BN_mod_inverse faster by explicitly handling small quotients
5962 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5963 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5966 *) New function BN_kronecker.
5969 *) Fix BN_gcd so that it works on negative inputs; the result is
5970 positive unless both parameters are zero.
5971 Previously something reasonably close to an infinite loop was
5972 possible because numbers could be growing instead of shrinking
5973 in the implementation of Euclid's algorithm.
5976 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5977 sign of the number in question.
5979 Fix BN_is_word(a,w) to work correctly for w == 0.
5981 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5982 because its test if the absolute value of 'a' equals 'w'.
5983 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5984 it exists mostly for use in the implementations of BN_is_zero(),
5985 BN_is_one(), and BN_is_word().
5988 *) New function BN_swap.
5991 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5992 the exponentiation functions are more likely to produce reasonable
5993 results on negative inputs.
5996 *) Change BN_mod_mul so that the result is always non-negative.
5997 Previously, it could be negative if one of the factors was negative;
5998 I don't think anyone really wanted that behaviour.
6001 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6002 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6003 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6004 and add new functions:
6013 BN_mod_lshift1_quick
6017 These functions always generate non-negative results.
6019 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6020 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6022 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6023 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6024 be reduced modulo m.
6025 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6028 The following entry accidentily appeared in the CHANGES file
6029 distributed with OpenSSL 0.9.7. The modifications described in
6030 it do *not* apply to OpenSSL 0.9.7.
6032 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6033 was actually never needed) and in BN_mul(). The removal in BN_mul()
6034 required a small change in bn_mul_part_recursive() and the addition
6035 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6036 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6037 bn_sub_words() and bn_add_words() except they take arrays with
6042 *) In 'openssl passwd', verify passwords read from the terminal
6043 unless the '-salt' option is used (which usually means that
6044 verification would just waste user's time since the resulting
6045 hash is going to be compared with some given password hash)
6046 or the new '-noverify' option is used.
6048 This is an incompatible change, but it does not affect
6049 non-interactive use of 'openssl passwd' (passwords on the command
6050 line, '-stdin' option, '-in ...' option) and thus should not
6054 *) Remove all references to RSAref, since there's no more need for it.
6057 *) Make DSO load along a path given through an environment variable
6058 (SHLIB_PATH) with shl_load().
6061 *) Constify the ENGINE code as a result of BIGNUM constification.
6062 Also constify the RSA code and most things related to it. In a
6063 few places, most notable in the depth of the ASN.1 code, ugly
6064 casts back to non-const were required (to be solved at a later
6068 *) Make it so the openssl application has all engines loaded by default.
6071 *) Constify the BIGNUM routines a little more.
6074 *) Add the following functions:
6076 ENGINE_load_cswift()
6078 ENGINE_load_atalla()
6080 ENGINE_load_builtin_engines()
6082 That way, an application can itself choose if external engines that
6083 are built-in in OpenSSL shall ever be used or not. The benefit is
6084 that applications won't have to be linked with libdl or other dso
6085 libraries unless it's really needed.
6087 Changed 'openssl engine' to load all engines on demand.
6088 Changed the engine header files to avoid the duplication of some
6089 declarations (they differed!).
6092 *) 'openssl engine' can now list capabilities.
6095 *) Better error reporting in 'openssl engine'.
6098 *) Never call load_dh_param(NULL) in s_server.
6101 *) Add engine application. It can currently list engines by name and
6102 identity, and test if they are actually available.
6105 *) Improve RPM specification file by forcing symbolic linking and making
6106 sure the installed documentation is also owned by root.root.
6107 [Damien Miller <djm@mindrot.org>]
6109 *) Give the OpenSSL applications more possibilities to make use of
6110 keys (public as well as private) handled by engines.
6113 *) Add OCSP code that comes from CertCo.
6116 *) Add VMS support for the Rijndael code.
6119 *) Added untested support for Nuron crypto accelerator.
6122 *) Add support for external cryptographic devices. This code was
6123 previously distributed separately as the "engine" branch.
6124 [Geoff Thorpe, Richard Levitte]
6126 *) Rework the filename-translation in the DSO code. It is now possible to
6127 have far greater control over how a "name" is turned into a filename
6128 depending on the operating environment and any oddities about the
6129 different shared library filenames on each system.
6132 *) Support threads on FreeBSD-elf in Configure.
6135 *) Fix for SHA1 assembly problem with MASM: it produces
6136 warnings about corrupt line number information when assembling
6137 with debugging information. This is caused by the overlapping
6139 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6142 NCONF_get_number() has no error checking at all. As a replacement,
6143 NCONF_get_number_e() is defined (_e for "error checking") and is
6144 promoted strongly. The old NCONF_get_number is kept around for
6145 binary backward compatibility.
6146 Make it possible for methods to load from something other than a BIO,
6147 by providing a function pointer that is given a name instead of a BIO.
6148 For example, this could be used to load configuration data from an
6152 *) Fix for non blocking accept BIOs. Added new I/O special reason
6153 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6154 with non blocking I/O was not possible because no retry code was
6155 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6159 *) Added the beginnings of Rijndael support.
6162 *) Fix for bug in DirectoryString mask setting. Add support for
6163 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6164 to allow certificate printing to more controllable, additional
6165 'certopt' option to 'x509' to allow new printing options to be
6169 *) Clean old EAY MD5 hack from e_os.h.
6172 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6174 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6175 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6176 [Joe Orton, Steve Henson]
6178 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6180 *) Fix additional bug revealed by the NISCC test suite:
6182 Stop bug triggering large recursion when presented with
6183 certain ASN.1 tags (CVE-2003-0851)
6186 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6188 *) Fix various bugs revealed by running the NISCC test suite:
6190 Stop out of bounds reads in the ASN1 code when presented with
6191 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6193 If verify callback ignores invalid public key errors don't try to check
6194 certificate signature with the NULL public key.
6198 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6199 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6203 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6204 extra data after the compression methods not only for TLS 1.0
6205 but also for SSL 3.0 (as required by the specification).
6206 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6208 *) Change X509_certificate_type() to mark the key as exported/exportable
6209 when it's 512 *bits* long, not 512 bytes.
6212 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6214 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6215 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6216 a protocol version number mismatch like a decryption error
6217 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6220 *) Turn on RSA blinding by default in the default implementation
6221 to avoid a timing attack. Applications that don't want it can call
6222 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6223 They would be ill-advised to do so in most cases.
6224 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6226 *) Change RSA blinding code so that it works when the PRNG is not
6227 seeded (in this case, the secret RSA exponent is abused as
6228 an unpredictable seed -- if it is not unpredictable, there
6229 is no point in blinding anyway). Make RSA blinding thread-safe
6230 by remembering the creator's thread ID in rsa->blinding and
6231 having all other threads use local one-time blinding factors
6232 (this requires more computation than sharing rsa->blinding, but
6233 avoids excessive locking; and if an RSA object is not shared
6234 between threads, blinding will still be very fast).
6237 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6239 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6240 via timing by performing a MAC computation even if incorrrect
6241 block cipher padding has been found. This is a countermeasure
6242 against active attacks where the attacker has to distinguish
6243 between bad padding and a MAC verification error. (CVE-2003-0078)
6245 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6246 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6247 Martin Vuagnoux (EPFL, Ilion)]
6249 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6251 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6252 memory from it's contents. This is done with a counter that will
6253 place alternating values in each byte. This can be used to solve
6254 two issues: 1) the removal of calls to memset() by highly optimizing
6255 compilers, and 2) cleansing with other values than 0, since those can
6256 be read through on certain media, for example a swap space on disk.
6259 *) Bugfix: client side session caching did not work with external caching,
6260 because the session->cipher setting was not restored when reloading
6261 from the external cache. This problem was masked, when
6262 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6263 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6266 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6267 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6268 [Zeev Lieber <zeev-l@yahoo.com>]
6270 *) Undo an undocumented change introduced in 0.9.6e which caused
6271 repeated calls to OpenSSL_add_all_ciphers() and
6272 OpenSSL_add_all_digests() to be ignored, even after calling
6276 *) Change the default configuration reader to deal with last line not
6277 being properly terminated.
6280 *) Change X509_NAME_cmp() so it applies the special rules on handling
6281 DN values that are of type PrintableString, as well as RDNs of type
6282 emailAddress where the value has the type ia5String.
6283 [stefank@valicert.com via Richard Levitte]
6285 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6286 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6287 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6288 the bitwise-OR of the two for use by the majority of applications
6289 wanting this behaviour, and update the docs. The documented
6290 behaviour and actual behaviour were inconsistent and had been
6291 changing anyway, so this is more a bug-fix than a behavioural
6293 [Geoff Thorpe, diagnosed by Nadav Har'El]
6295 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6296 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6299 *) Fix initialization code race conditions in
6300 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6301 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6302 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6303 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6304 ssl2_get_cipher_by_char(),
6305 ssl3_get_cipher_by_char().
6306 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6308 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6309 the cached sessions are flushed, as the remove_cb() might use ex_data
6310 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6311 (see [openssl.org #212]).
6312 [Geoff Thorpe, Lutz Jaenicke]
6314 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6315 length, instead of the encoding length to d2i_ASN1_OBJECT.
6318 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6320 *) [In 0.9.6g-engine release:]
6321 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6322 [Lynn Gazis <lgazis@rainbow.com>]
6324 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6326 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6327 and get fix the header length calculation.
6328 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6329 Alon Kantor <alonk@checkpoint.com> (and others),
6332 *) Use proper error handling instead of 'assertions' in buffer
6333 overflow checks added in 0.9.6e. This prevents DoS (the
6334 assertions could call abort()).
6335 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6337 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6339 *) Add various sanity checks to asn1_get_length() to reject
6340 the ASN1 length bytes if they exceed sizeof(long), will appear
6341 negative or the content length exceeds the length of the
6343 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6345 *) Fix cipher selection routines: ciphers without encryption had no flags
6346 for the cipher strength set and where therefore not handled correctly
6347 by the selection routines (PR #130).
6350 *) Fix EVP_dsa_sha macro.
6354 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6355 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6356 that was added in OpenSSL 0.9.6d.
6358 As the countermeasure turned out to be incompatible with some
6359 broken SSL implementations, the new option is part of SSL_OP_ALL.
6360 SSL_OP_ALL is usually employed when compatibility with weird SSL
6361 implementations is desired (e.g. '-bugs' option to 's_client' and
6362 's_server'), so the new option is automatically set in many
6366 *) Changes in security patch:
6368 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6369 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6370 Air Force Materiel Command, USAF, under agreement number
6373 *) Add various sanity checks to asn1_get_length() to reject
6374 the ASN1 length bytes if they exceed sizeof(long), will appear
6375 negative or the content length exceeds the length of the
6376 supplied buffer. (CVE-2002-0659)
6377 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6379 *) Assertions for various potential buffer overflows, not known to
6381 [Ben Laurie (CHATS)]
6383 *) Various temporary buffers to hold ASCII versions of integers were
6384 too small for 64 bit platforms. (CVE-2002-0655)
6385 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6387 *) Remote buffer overflow in SSL3 protocol - an attacker could
6388 supply an oversized session ID to a client. (CVE-2002-0656)
6389 [Ben Laurie (CHATS)]
6391 *) Remote buffer overflow in SSL2 protocol - an attacker could
6392 supply an oversized client master key. (CVE-2002-0656)
6393 [Ben Laurie (CHATS)]
6395 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6397 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6398 encoded as NULL) with id-dsa-with-sha1.
6399 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6401 *) Check various X509_...() return values in apps/req.c.
6402 [Nils Larsch <nla@trustcenter.de>]
6404 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6405 an end-of-file condition would erronously be flagged, when the CRLF
6406 was just at the end of a processed block. The bug was discovered when
6407 processing data through a buffering memory BIO handing the data to a
6408 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6409 <ptsekov@syntrex.com> and Nedelcho Stanev.
6412 *) Implement a countermeasure against a vulnerability recently found
6413 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6414 before application data chunks to avoid the use of known IVs
6415 with data potentially chosen by the attacker.
6418 *) Fix length checks in ssl3_get_client_hello().
6421 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6422 to prevent ssl3_read_internal() from incorrectly assuming that
6423 ssl3_read_bytes() found application data while handshake
6424 processing was enabled when in fact s->s3->in_read_app_data was
6425 merely automatically cleared during the initial handshake.
6426 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6428 *) Fix object definitions for Private and Enterprise: they were not
6429 recognized in their shortname (=lowercase) representation. Extend
6430 obj_dat.pl to issue an error when using undefined keywords instead
6431 of silently ignoring the problem (Svenning Sorensen
6432 <sss@sss.dnsalias.net>).
6435 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6436 generators, i.e. generators other than 2 and 5. (Previously, the
6437 code did not properly initialise the 'add' and 'rem' values to
6438 BN_generate_prime().)
6440 In the new general case, we do not insist that 'generator' is
6441 actually a primitive root: This requirement is rather pointless;
6442 a generator of the order-q subgroup is just as good, if not
6446 *) Map new X509 verification errors to alerts. Discovered and submitted by
6447 Tom Wu <tom@arcot.com>.
6450 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6451 returning non-zero before the data has been completely received
6452 when using non-blocking I/O.
6453 [Bodo Moeller; problem pointed out by John Hughes]
6455 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6456 [Ben Laurie, Lutz Jaenicke]
6458 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6459 Yoram Zahavi <YoramZ@gilian.com>).
6462 *) Add information about CygWin 1.3 and on, and preserve proper
6463 configuration for the versions before that.
6464 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6466 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6467 check whether we deal with a copy of a session and do not delete from
6468 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6469 <izhar@checkpoint.com>.
6472 *) Do not store session data into the internal session cache, if it
6473 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6474 flag is set). Proposed by Aslam <aslam@funk.com>.
6477 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6481 *) [In 0.9.6d-engine release:]
6482 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6483 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6485 *) Add the configuration target linux-s390x.
6486 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6488 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6489 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6490 variable as an indication that a ClientHello message has been
6491 received. As the flag value will be lost between multiple
6492 invocations of ssl3_accept when using non-blocking I/O, the
6493 function may not be aware that a handshake has actually taken
6494 place, thus preventing a new session from being added to the
6497 To avoid this problem, we now set s->new_session to 2 instead of
6498 using a local variable.
6499 [Lutz Jaenicke, Bodo Moeller]
6501 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6502 if the SSL_R_LENGTH_MISMATCH error is detected.
6503 [Geoff Thorpe, Bodo Moeller]
6505 *) New 'shared_ldflag' column in Configure platform table.
6508 *) Fix EVP_CIPHER_mode macro.
6509 ["Dan S. Camper" <dan@bti.net>]
6511 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6512 type, we must throw them away by setting rr->length to 0.
6513 [D P Chang <dpc@qualys.com>]
6515 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6517 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6518 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6519 worked incorrectly for those cases where range = 10..._2 and
6520 3*range is two bits longer than range.)
6523 *) Only add signing time to PKCS7 structures if it is not already
6527 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6528 OBJ_ld_ce should be OBJ_id_ce.
6529 Also some ip-pda OIDs in crypto/objects/objects.txt were
6530 incorrect (cf. RFC 3039).
6531 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6533 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6534 returns early because it has nothing to do.
6535 [Andy Schneider <andy.schneider@bjss.co.uk>]
6537 *) [In 0.9.6c-engine release:]
6538 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6539 [Andy Schneider <andy.schneider@bjss.co.uk>]
6541 *) [In 0.9.6c-engine release:]
6542 Add support for Cryptographic Appliance's keyserver technology.
6543 (Use engine 'keyclient')
6544 [Cryptographic Appliances and Geoff Thorpe]
6546 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6547 is called via tools/c89.sh because arguments have to be
6548 rearranged (all '-L' options must appear before the first object
6550 [Richard Shapiro <rshapiro@abinitio.com>]
6552 *) [In 0.9.6c-engine release:]
6553 Add support for Broadcom crypto accelerator cards, backported
6555 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6557 *) [In 0.9.6c-engine release:]
6558 Add support for SureWare crypto accelerator cards from
6559 Baltimore Technologies. (Use engine 'sureware')
6560 [Baltimore Technologies and Mark Cox]
6562 *) [In 0.9.6c-engine release:]
6563 Add support for crypto accelerator cards from Accelerated
6564 Encryption Processing, www.aep.ie. (Use engine 'aep')
6565 [AEP Inc. and Mark Cox]
6567 *) Add a configuration entry for gcc on UnixWare.
6568 [Gary Benson <gbenson@redhat.com>]
6570 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6571 messages are stored in a single piece (fixed-length part and
6572 variable-length part combined) and fix various bugs found on the way.
6575 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6576 instead. BIO_gethostbyname() does not know what timeouts are
6577 appropriate, so entries would stay in cache even when they have
6579 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6581 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6582 faced with a pathologically small ClientHello fragment that does
6583 not contain client_version: Instead of aborting with an error,
6584 simply choose the highest available protocol version (i.e.,
6585 TLS 1.0 unless it is disabled). In practice, ClientHello
6586 messages are never sent like this, but this change gives us
6587 strictly correct behaviour at least for TLS.
6590 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6591 never resets s->method to s->ctx->method when called from within
6592 one of the SSL handshake functions.
6593 [Bodo Moeller; problem pointed out by Niko Baric]
6595 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6596 (sent using the client's version number) if client_version is
6597 smaller than the protocol version in use. Also change
6598 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6599 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6600 the client will at least see that alert.
6603 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6607 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6608 client receives HelloRequest while in a handshake.
6609 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6611 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6612 should end in 'break', not 'goto end' which circuments various
6613 cleanups done in state SSL_ST_OK. But session related stuff
6614 must be disabled for SSL_ST_OK in the case that we just sent a
6617 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6618 before just sending a HelloRequest.
6619 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6621 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6622 reveal whether illegal block cipher padding was found or a MAC
6623 verification error occured. (Neither SSLerr() codes nor alerts
6624 are directly visible to potential attackers, but the information
6625 may leak via logfiles.)
6627 Similar changes are not required for the SSL 2.0 implementation
6628 because the number of padding bytes is sent in clear for SSL 2.0,
6629 and the extra bytes are just ignored. However ssl/s2_pkt.c
6630 failed to verify that the purported number of padding bytes is in
6634 *) Add OpenUNIX-8 support including shared libraries
6635 (Boyd Lynn Gerber <gerberb@zenez.com>).
6638 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6639 'wristwatch attack' using huge encoding parameters (cf.
6640 James H. Manger's CRYPTO 2001 paper). Note that the
6641 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6642 encoding parameters and hence was not vulnerable.
6645 *) BN_sqr() bug fix.
6646 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6648 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6649 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6650 followed by modular reduction.
6651 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6653 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6654 equivalent based on BN_pseudo_rand() instead of BN_rand().
6657 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6658 This function was broken, as the check for a new client hello message
6659 to handle SGC did not allow these large messages.
6660 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6663 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6666 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6667 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6670 *) Rework the configuration and shared library support for Tru64 Unix.
6671 The configuration part makes use of modern compiler features and
6672 still retains old compiler behavior for those that run older versions
6673 of the OS. The shared library support part includes a variant that
6674 uses the RPATH feature, and is available through the special
6675 configuration target "alpha-cc-rpath", which will never be selected
6677 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6679 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6680 with the same message size as in ssl3_get_certificate_request().
6681 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6682 messages might inadvertently be reject as too long.
6683 [Petr Lampa <lampa@fee.vutbr.cz>]
6685 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6688 *) Modified SSL library such that the verify_callback that has been set
6689 specificly for an SSL object with SSL_set_verify() is actually being
6690 used. Before the change, a verify_callback set with this function was
6691 ignored and the verify_callback() set in the SSL_CTX at the time of
6692 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6693 to allow the necessary settings.
6696 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6697 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6698 done automatically (in contradiction to the requirements of the C
6699 standard). This made problems when used from OpenSSH.
6702 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6703 dh->length and always used
6705 BN_rand_range(priv_key, dh->p).
6707 BN_rand_range() is not necessary for Diffie-Hellman, and this
6708 specific range makes Diffie-Hellman unnecessarily inefficient if
6709 dh->length (recommended exponent length) is much smaller than the
6710 length of dh->p. We could use BN_rand_range() if the order of
6711 the subgroup was stored in the DH structure, but we only have
6716 BN_rand(priv_key, l, ...)
6718 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6724 RSA_eay_public_encrypt
6725 RSA_eay_private_decrypt
6726 RSA_eay_private_encrypt (signing)
6727 RSA_eay_public_decrypt (signature verification)
6729 (default implementations for RSA_public_encrypt,
6730 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6731 always reject numbers >= n.
6734 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6735 to synchronize access to 'locking_thread'. This is necessary on
6736 systems where access to 'locking_thread' (an 'unsigned long'
6737 variable) is not atomic.
6740 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6741 *before* setting the 'crypto_lock_rand' flag. The previous code had
6742 a race condition if 0 is a valid thread ID.
6743 [Travis Vitek <vitek@roguewave.com>]
6745 *) Add support for shared libraries under Irix.
6746 [Albert Chin-A-Young <china@thewrittenword.com>]
6748 *) Add configuration option to build on Linux on both big-endian and
6750 [Ralf Baechle <ralf@uni-koblenz.de>]
6752 *) Add the possibility to create shared libraries on HP-UX.
6755 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6757 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6758 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6759 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6760 PRNG state recovery was possible based on the output of
6761 one PRNG request appropriately sized to gain knowledge on
6762 'md' followed by enough consecutive 1-byte PRNG requests
6763 to traverse all of 'state'.
6765 1. When updating 'md_local' (the current thread's copy of 'md')
6766 during PRNG output generation, hash all of the previous
6767 'md_local' value, not just the half used for PRNG output.
6769 2. Make the number of bytes from 'state' included into the hash
6770 independent from the number of PRNG bytes requested.
6772 The first measure alone would be sufficient to avoid
6773 Markku-Juhani's attack. (Actually it had never occurred
6774 to me that the half of 'md_local' used for chaining was the
6775 half from which PRNG output bytes were taken -- I had always
6776 assumed that the secret half would be used.) The second
6777 measure makes sure that additional data from 'state' is never
6778 mixed into 'md_local' in small portions; this heuristically
6779 further strengthens the PRNG.
6782 *) Fix crypto/bn/asm/mips3.s.
6785 *) When only the key is given to "enc", the IV is undefined. Print out
6786 an error message in this case.
6789 *) Handle special case when X509_NAME is empty in X509 printing routines.
6792 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6793 positive and less than q.
6796 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6797 used: it isn't thread safe and the add_lock_callback should handle
6799 [Paul Rose <Paul.Rose@bridge.com>]
6801 *) Verify that incoming data obeys the block size in
6802 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6806 [Ulf Möller, Bodo Möller]
6808 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6809 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6810 when fixing the server behaviour for backwards-compatible 'client
6811 hello' messages. (Note that the attack is impractical against
6812 SSL 3.0 and TLS 1.0 anyway because length and version checking
6813 means that the probability of guessing a valid ciphertext is
6814 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6817 Before 0.9.5, the countermeasure (hide the error by generating a
6818 random 'decryption result') did not work properly because
6819 ERR_clear_error() was missing, meaning that SSL_get_error() would
6820 detect the supposedly ignored error.
6822 Both problems are now fixed.
6825 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6826 (previously it was 1024).
6829 *) Fix for compatibility mode trust settings: ignore trust settings
6830 unless some valid trust or reject settings are present.
6833 *) Fix for blowfish EVP: its a variable length cipher.
6836 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6837 parameters in DSA public key structures and return an error in the
6838 DSA routines if parameters are absent.
6841 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6842 in the current directory if neither $RANDFILE nor $HOME was set.
6843 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6844 caused some confusion to Windows users who haven't defined $HOME.
6845 Thus RAND_file_name() is changed again: e_os.h can define a
6846 DEFAULT_HOME, which will be used if $HOME is not set.
6847 For Windows, we use "C:"; on other platforms, we still require
6848 environment variables.
6850 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6851 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6852 having multiple threads call RAND_poll() concurrently.
6855 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6856 combination of a flag and a thread ID variable.
6857 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6858 flag), *other* threads can enter ssleay_add_bytes without obeying
6859 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6860 that they do not hold after the first thread unsets add_do_not_lock).
6863 *) Change bctest again: '-x' expressions are not available in all
6867 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6869 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6870 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6872 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6873 the default extension for executables, if any. Also, make the perl
6874 scripts that use symlink() to test if it really exists and use "cp"
6875 if it doesn't. All this made OpenSSL compilable and installable in
6879 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6880 If SEQUENCE is length is indefinite just set c->slen to the total
6881 amount of data available.
6882 [Steve Henson, reported by shige@FreeBSD.org]
6883 [This change does not apply to 0.9.7.]
6885 *) Change bctest to avoid here-documents inside command substitution
6886 (workaround for FreeBSD /bin/sh bug).
6887 For compatibility with Ultrix, avoid shell functions (introduced
6888 in the bctest version that searches along $PATH).
6891 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6892 with des_encrypt() defined on some operating systems, like Solaris
6896 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6897 On the Importance of Eliminating Errors in Cryptographic
6898 Computations, J. Cryptology 14 (2001) 2, 101-119,
6899 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6902 *) MIPS assembler BIGNUM division bug fix.
6905 *) Disabled incorrect Alpha assembler code.
6908 *) Fix PKCS#7 decode routines so they correctly update the length
6909 after reading an EOC for the EXPLICIT tag.
6911 [This change does not apply to 0.9.7.]
6913 *) Fix bug in PKCS#12 key generation routines. This was triggered
6914 if a 3DES key was generated with a 0 initial byte. Include
6915 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6916 (but broken) behaviour.
6919 *) Enhance bctest to search for a working bc along $PATH and print
6921 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6923 *) Fix memory leaks in err.c: free err_data string if necessary;
6924 don't write to the wrong index in ERR_set_error_data.
6927 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6931 *) Replace rdtsc with _emit statements for VC++ version 5.
6932 [Jeremy Cooper <jeremy@baymoo.org>]
6934 *) Make it possible to reuse SSLv2 sessions.
6937 *) In copy_email() check for >= 0 as a return value for
6938 X509_NAME_get_index_by_NID() since 0 is a valid index.
6939 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6941 *) Avoid coredump with unsupported or invalid public keys by checking if
6942 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6943 PKCS7_verify() fails with non detached data.
6946 *) Don't use getenv in library functions when run as setuid/setgid.
6947 New function OPENSSL_issetugid().
6950 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6951 due to incorrect handling of multi-threading:
6953 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6955 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6957 3. Count how many times MemCheck_off() has been called so that
6958 nested use can be treated correctly. This also avoids
6959 inband-signalling in the previous code (which relied on the
6960 assumption that thread ID 0 is impossible).
6963 *) Add "-rand" option also to s_client and s_server.
6966 *) Fix CPU detection on Irix 6.x.
6967 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6968 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6970 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6973 [This change does not apply to 0.9.7.]
6975 *) Use the cached encoding of an X509_NAME structure rather than
6976 copying it. This is apparently the reason for the libsafe "errors"
6977 but the code is actually correct.
6980 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6981 Bleichenbacher's DSA attack.
6982 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6983 to be set and top=0 forces the highest bit to be set; top=-1 is new
6984 and leaves the highest bit random.
6985 [Ulf Moeller, Bodo Moeller]
6987 *) In the NCONF_...-based implementations for CONF_... queries
6988 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6989 a temporary CONF structure with the data component set to NULL
6990 (which gives segmentation faults in lh_retrieve).
6991 Instead, use NULL for the CONF pointer in CONF_get_string and
6992 CONF_get_number (which may use environment variables) and directly
6993 return NULL from CONF_get_section.
6996 *) Fix potential buffer overrun for EBCDIC.
6999 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7000 keyUsage if basicConstraints absent for a CA.
7003 *) Make SMIME_write_PKCS7() write mail header values with a format that
7004 is more generally accepted (no spaces before the semicolon), since
7005 some programs can't parse those values properly otherwise. Also make
7006 sure BIO's that break lines after each write do not create invalid
7010 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7011 macros previously used would not encode an empty SEQUENCE OF
7012 and break the signature.
7014 [This change does not apply to 0.9.7.]
7016 *) Zero the premaster secret after deriving the master secret in
7020 *) Add some EVP_add_digest_alias registrations (as found in
7021 OpenSSL_add_all_digests()) to SSL_library_init()
7022 aka OpenSSL_add_ssl_algorithms(). This provides improved
7023 compatibility with peers using X.509 certificates
7024 with unconventional AlgorithmIdentifier OIDs.
7027 *) Fix for Irix with NO_ASM.
7028 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7030 *) ./config script fixes.
7031 [Ulf Moeller, Richard Levitte]
7033 *) Fix 'openssl passwd -1'.
7036 *) Change PKCS12_key_gen_asc() so it can cope with non null
7037 terminated strings whose length is passed in the passlen
7038 parameter, for example from PEM callbacks. This was done
7039 by adding an extra length parameter to asc2uni().
7040 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7042 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7043 call failed, free the DSA structure.
7046 *) Fix to uni2asc() to cope with zero length Unicode strings.
7047 These are present in some PKCS#12 files.
7050 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7051 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7052 when writing a 32767 byte record.
7053 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7055 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7056 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7058 (RSA objects have a reference count access to which is protected
7059 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7060 so they are meant to be shared between threads.)
7061 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7062 "Reddie, Steven" <Steven.Reddie@ca.com>]
7064 *) Fix a deadlock in CRYPTO_mem_leaks().
7067 *) Use better test patterns in bntest.
7070 *) rand_win.c fix for Borland C.
7073 *) BN_rshift bugfix for n == 0.
7076 *) Add a 'bctest' script that checks for some known 'bc' bugs
7077 so that 'make test' does not abort just because 'bc' is broken.
7080 *) Store verify_result within SSL_SESSION also for client side to
7081 avoid potential security hole. (Re-used sessions on the client side
7082 always resulted in verify_result==X509_V_OK, not using the original
7083 result of the server certificate verification.)
7086 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7087 SSL3_RT_APPLICATION_DATA, return 0.
7088 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7092 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7093 releases, have been re-implemented by renaming the previous
7094 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7095 and ssl3_read_internal, respectively, and adding 'peek' parameters
7096 to them. The new ssl[23]_{read,peek} functions are calls to
7097 ssl[23]_read_internal with the 'peek' flag set appropriately.
7098 A 'peek' parameter has also been added to ssl3_read_bytes, which
7099 does the actual work for ssl3_read_internal.
7102 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7103 the method-specific "init()" handler. Also clean up ex_data after
7104 calling the method-specific "finish()" handler. Previously, this was
7105 happening the other way round.
7108 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7109 The previous value, 12, was not always sufficient for BN_mod_exp().
7112 *) Make sure that shared libraries get the internal name engine with
7113 the full version number and not just 0. This should mark the
7114 shared libraries as not backward compatible. Of course, this should
7115 be changed again when we can guarantee backward binary compatibility.
7118 *) Fix typo in get_cert_by_subject() in by_dir.c
7119 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7121 *) Rework the system to generate shared libraries:
7123 - Make note of the expected extension for the shared libraries and
7124 if there is a need for symbolic links from for example libcrypto.so.0
7125 to libcrypto.so.0.9.7. There is extended info in Configure for
7128 - Make as few rebuilds of the shared libraries as possible.
7130 - Still avoid linking the OpenSSL programs with the shared libraries.
7132 - When installing, install the shared libraries separately from the
7136 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7138 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7139 and not in SSL_clear because the latter is also used by the
7140 accept/connect functions; previously, the settings made by
7141 SSL_set_read_ahead would be lost during the handshake.
7142 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7144 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7145 Previously, it would create entries for disableed algorithms no
7149 *) Added several new manual pages for SSL_* function.
7152 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7154 *) In ssl23_get_client_hello, generate an error message when faced
7155 with an initial SSL 3.0/TLS record that is too small to contain the
7156 first two bytes of the ClientHello message, i.e. client_version.
7157 (Note that this is a pathologic case that probably has never happened
7158 in real life.) The previous approach was to use the version number
7159 from the record header as a substitute; but our protocol choice
7160 should not depend on that one because it is not authenticated
7161 by the Finished messages.
7164 *) More robust randomness gathering functions for Windows.
7165 [Jeffrey Altman <jaltman@columbia.edu>]
7167 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7168 not set then we don't setup the error code for issuer check errors
7169 to avoid possibly overwriting other errors which the callback does
7170 handle. If an application does set the flag then we assume it knows
7171 what it is doing and can handle the new informational codes
7175 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7176 a general "ANY" type, as such it should be able to decode anything
7177 including tagged types. However it didn't check the class so it would
7178 wrongly interpret tagged types in the same way as their universal
7179 counterpart and unknown types were just rejected. Changed so that the
7180 tagged and unknown types are handled in the same way as a SEQUENCE:
7181 that is the encoding is stored intact. There is also a new type
7182 "V_ASN1_OTHER" which is used when the class is not universal, in this
7183 case we have no idea what the actual type is so we just lump them all
7187 *) On VMS, stdout may very well lead to a file that is written to
7188 in a record-oriented fashion. That means that every write() will
7189 write a separate record, which will be read separately by the
7190 programs trying to read from it. This can be very confusing.
7192 The solution is to put a BIO filter in the way that will buffer
7193 text until a linefeed is reached, and then write everything a
7194 line at a time, so every record written will be an actual line,
7195 not chunks of lines and not (usually doesn't happen, but I've
7196 seen it once) several lines in one record. BIO_f_linebuffer() is
7199 Currently, it's a VMS-only method, because that's where it has
7200 been tested well enough.
7203 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7204 it can return incorrect results.
7205 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7206 but it was in 0.9.6-beta[12].)
7209 *) Disable the check for content being present when verifying detached
7210 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7211 include zero length content when signing messages.
7214 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7215 BIO_ctrl (for BIO pairs).
7218 *) Add DSO method for VMS.
7221 *) Bug fix: Montgomery multiplication could produce results with the
7225 *) Add RPM specification openssl.spec and modify it to build three
7226 packages. The default package contains applications, application
7227 documentation and run-time libraries. The devel package contains
7228 include files, static libraries and function documentation. The
7229 doc package contains the contents of the doc directory. The original
7230 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7233 *) Add a large number of documentation files for many SSL routines.
7234 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7236 *) Add a configuration entry for Sony News 4.
7237 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7239 *) Don't set the two most significant bits to one when generating a
7240 random number < q in the DSA library.
7243 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7244 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7245 the underlying transport is blocking) if a handshake took place.
7246 (The default behaviour is needed by applications such as s_client
7247 and s_server that use select() to determine when to use SSL_read;
7248 but for applications that know in advance when to expect data, it
7249 just makes things more complicated.)
7252 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7256 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7257 work better on such systems.
7258 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7260 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7261 Update PKCS12_parse() so it copies the friendlyName and the
7262 keyid to the certificates aux info.
7265 *) Fix bug in PKCS7_verify() which caused an infinite loop
7266 if there was more than one signature.
7267 [Sven Uszpelkat <su@celocom.de>]
7269 *) Major change in util/mkdef.pl to include extra information
7270 about each symbol, as well as presentig variables as well
7271 as functions. This change means that there's n more need
7272 to rebuild the .num files when some algorithms are excluded.
7275 *) Allow the verify time to be set by an application,
7276 rather than always using the current time.
7279 *) Phase 2 verify code reorganisation. The certificate
7280 verify code now looks up an issuer certificate by a
7281 number of criteria: subject name, authority key id
7282 and key usage. It also verifies self signed certificates
7283 by the same criteria. The main comparison function is
7284 X509_check_issued() which performs these checks.
7286 Lot of changes were necessary in order to support this
7287 without completely rewriting the lookup code.
7289 Authority and subject key identifier are now cached.
7291 The LHASH 'certs' is X509_STORE has now been replaced
7292 by a STACK_OF(X509_OBJECT). This is mainly because an
7293 LHASH can't store or retrieve multiple objects with
7294 the same hash value.
7296 As a result various functions (which were all internal
7297 use only) have changed to handle the new X509_STORE
7298 structure. This will break anything that messed round
7299 with X509_STORE internally.
7301 The functions X509_STORE_add_cert() now checks for an
7302 exact match, rather than just subject name.
7304 The X509_STORE API doesn't directly support the retrieval
7305 of multiple certificates matching a given criteria, however
7306 this can be worked round by performing a lookup first
7307 (which will fill the cache with candidate certificates)
7308 and then examining the cache for matches. This is probably
7309 the best we can do without throwing out X509_LOOKUP
7310 entirely (maybe later...).
7312 The X509_VERIFY_CTX structure has been enhanced considerably.
7314 All certificate lookup operations now go via a get_issuer()
7315 callback. Although this currently uses an X509_STORE it
7316 can be replaced by custom lookups. This is a simple way
7317 to bypass the X509_STORE hackery necessary to make this
7318 work and makes it possible to use more efficient techniques
7319 in future. A very simple version which uses a simple
7320 STACK for its trusted certificate store is also provided
7321 using X509_STORE_CTX_trusted_stack().
7323 The verify_cb() and verify() callbacks now have equivalents
7324 in the X509_STORE_CTX structure.
7326 X509_STORE_CTX also has a 'flags' field which can be used
7327 to customise the verify behaviour.
7330 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7331 excludes S/MIME capabilities.
7334 *) When a certificate request is read in keep a copy of the
7335 original encoding of the signed data and use it when outputing
7336 again. Signatures then use the original encoding rather than
7337 a decoded, encoded version which may cause problems if the
7338 request is improperly encoded.
7341 *) For consistency with other BIO_puts implementations, call
7342 buffer_write(b, ...) directly in buffer_puts instead of calling
7345 In BIO_puts, increment b->num_write as in BIO_write.
7346 [Peter.Sylvester@EdelWeb.fr]
7348 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7349 BN_zero, we may not return a BIGNUM with an array consisting of
7353 *) Avoid calling abort() from within the library when problems are
7354 detected, except if preprocessor symbols have been defined
7355 (such as REF_CHECK, BN_DEBUG etc.).
7358 *) New openssl application 'rsautl'. This utility can be
7359 used for low level RSA operations. DER public key
7360 BIO/fp routines also added.
7363 *) New Configure entry and patches for compiling on QNX 4.
7364 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7366 *) A demo state-machine implementation was sponsored by
7367 Nuron (http://www.nuron.com/) and is now available in
7368 demos/state_machine.
7371 *) New options added to the 'dgst' utility for signature
7372 generation and verification.
7375 *) Unrecognized PKCS#7 content types are now handled via a
7376 catch all ASN1_TYPE structure. This allows unsupported
7377 types to be stored as a "blob" and an application can
7378 encode and decode it manually.
7381 *) Fix various signed/unsigned issues to make a_strex.c
7383 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7385 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7386 length if passed a buffer. ASN1_INTEGER_to_BN failed
7387 if passed a NULL BN and its argument was negative.
7388 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7390 *) Modification to PKCS#7 encoding routines to output definite
7391 length encoding. Since currently the whole structures are in
7392 memory there's not real point in using indefinite length
7393 constructed encoding. However if OpenSSL is compiled with
7394 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7397 *) Added BIO_vprintf() and BIO_vsnprintf().
7400 *) Added more prefixes to parse for in the the strings written
7401 through a logging bio, to cover all the levels that are available
7402 through syslog. The prefixes are now:
7404 PANIC, EMERG, EMR => LOG_EMERG
7405 ALERT, ALR => LOG_ALERT
7406 CRIT, CRI => LOG_CRIT
7407 ERROR, ERR => LOG_ERR
7408 WARNING, WARN, WAR => LOG_WARNING
7409 NOTICE, NOTE, NOT => LOG_NOTICE
7410 INFO, INF => LOG_INFO
7411 DEBUG, DBG => LOG_DEBUG
7413 and as before, if none of those prefixes are present at the
7414 beginning of the string, LOG_ERR is chosen.
7416 On Win32, the LOG_* levels are mapped according to this:
7418 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7419 LOG_WARNING => EVENTLOG_WARNING_TYPE
7420 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7424 *) Made it possible to reconfigure with just the configuration
7425 argument "reconf" or "reconfigure". The command line arguments
7426 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7427 and are retrieved from there when reconfiguring.
7431 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7433 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7436 *) The obj_dat.pl script was messing up the sorting of object
7437 names. The reason was that it compared the quoted version
7438 of strings as a result "OCSP" > "OCSP Signing" because
7439 " > SPACE. Changed script to store unquoted versions of
7440 names and add quotes on output. It was also omitting some
7441 names from the lookup table if they were given a default
7442 value (that is if SN is missing it is given the same
7443 value as LN and vice versa), these are now added on the
7444 grounds that if an object has a name we should be able to
7445 look it up. Finally added warning output when duplicate
7446 short or long names are found.
7449 *) Changes needed for Tandem NSK.
7450 [Scott Uroff <scott@xypro.com>]
7452 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7453 RSA_padding_check_SSLv23(), special padding was never detected
7454 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7455 version rollback attacks was not effective.
7457 In s23_clnt.c, don't use special rollback-attack detection padding
7458 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7459 client; similarly, in s23_srvr.c, don't do the rollback check if
7460 SSL 2.0 is the only protocol enabled in the server.
7463 *) Make it possible to get hexdumps of unprintable data with 'openssl
7464 asn1parse'. By implication, the functions ASN1_parse_dump() and
7465 BIO_dump_indent() are added.
7468 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7469 these print out strings and name structures based on various
7470 flags including RFC2253 support and proper handling of
7471 multibyte characters. Added options to the 'x509' utility
7472 to allow the various flags to be set.
7475 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7476 Also change the functions X509_cmp_current_time() and
7477 X509_gmtime_adj() work with an ASN1_TIME structure,
7478 this will enable certificates using GeneralizedTime in validity
7479 dates to be checked.
7482 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7483 negative public key encodings) on by default,
7484 NO_NEG_PUBKEY_BUG can be set to disable it.
7487 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7488 content octets. An i2c_ASN1_OBJECT is unnecessary because
7489 the encoding can be trivially obtained from the structure.
7492 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7493 not read locks (CRYPTO_r_[un]lock).
7496 *) A first attempt at creating official support for shared
7497 libraries through configuration. I've kept it so the
7498 default is static libraries only, and the OpenSSL programs
7499 are always statically linked for now, but there are
7500 preparations for dynamic linking in place.
7501 This has been tested on Linux and Tru64.
7504 *) Randomness polling function for Win9x, as described in:
7505 Peter Gutmann, Software Generation of Practically Strong
7509 *) Fix so PRNG is seeded in req if using an already existing
7513 *) New options to smime application. -inform and -outform
7514 allow alternative formats for the S/MIME message including
7515 PEM and DER. The -content option allows the content to be
7516 specified separately. This should allow things like Netscape
7517 form signing output easier to verify.
7520 *) Fix the ASN1 encoding of tags using the 'long form'.
7523 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7524 STRING types. These convert content octets to and from the
7525 underlying type. The actual tag and length octets are
7526 already assumed to have been read in and checked. These
7527 are needed because all other string types have virtually
7528 identical handling apart from the tag. By having versions
7529 of the ASN1 functions that just operate on content octets
7530 IMPLICIT tagging can be handled properly. It also allows
7531 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7532 and ASN1_INTEGER are identical apart from the tag.
7535 *) Change the handling of OID objects as follows:
7537 - New object identifiers are inserted in objects.txt, following
7538 the syntax given in objects.README.
7539 - objects.pl is used to process obj_mac.num and create a new
7541 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7544 This is currently kind of a hack, and the perl code in objects.pl
7545 isn't very elegant, but it works as I intended. The simplest way
7546 to check that it worked correctly is to look in obj_dat.h and
7547 check the array nid_objs and make sure the objects haven't moved
7548 around (this is important!). Additions are OK, as well as
7549 consistent name changes.
7552 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7555 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7556 The given file adds to whatever has already been seeded into the
7557 random pool through the RANDFILE configuration file option or
7558 environment variable, or the default random state file.
7561 *) mkstack.pl now sorts each macro group into lexical order.
7562 Previously the output order depended on the order the files
7563 appeared in the directory, resulting in needless rewriting
7567 *) Patches to make OpenSSL compile under Win32 again. Mostly
7568 work arounds for the VC++ problem that it treats func() as
7569 func(void). Also stripped out the parts of mkdef.pl that
7570 added extra typesafe functions: these no longer exist.
7573 *) Reorganisation of the stack code. The macros are now all
7574 collected in safestack.h . Each macro is defined in terms of
7575 a "stack macro" of the form SKM_<name>(type, a, b). The
7576 DEBUG_SAFESTACK is now handled in terms of function casts,
7577 this has the advantage of retaining type safety without the
7578 use of additional functions. If DEBUG_SAFESTACK is not defined
7579 then the non typesafe macros are used instead. Also modified the
7580 mkstack.pl script to handle the new form. Needs testing to see
7581 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7582 the default if no major problems. Similar behaviour for ASN1_SET_OF
7583 and PKCS12_STACK_OF.
7586 *) When some versions of IIS use the 'NET' form of private key the
7587 key derivation algorithm is different. Normally MD5(password) is
7588 used as a 128 bit RC4 key. In the modified case
7589 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7590 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7591 as the old Netscape_RSA functions except they have an additional
7592 'sgckey' parameter which uses the modified algorithm. Also added
7593 an -sgckey command line option to the rsa utility. Thanks to
7594 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7595 algorithm to openssl-dev.
7598 *) The evp_local.h macros were using 'c.##kname' which resulted in
7599 invalid expansion on some systems (SCO 5.0.5 for example).
7600 Corrected to 'c.kname'.
7601 [Phillip Porch <root@theporch.com>]
7603 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7604 a STACK of email addresses from a certificate or request, these look
7605 in the subject name and the subject alternative name extensions and
7606 omit any duplicate addresses.
7609 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7610 This makes DSA verification about 2 % faster.
7613 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7614 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7615 plus overhead for 1024 bit moduli).
7616 This makes exponentiations about 0.5 % faster for 1024 bit
7617 exponents (as measured by "openssl speed rsa2048").
7620 *) Rename memory handling macros to avoid conflicts with other
7622 Malloc => OPENSSL_malloc
7623 Malloc_locked => OPENSSL_malloc_locked
7624 Realloc => OPENSSL_realloc
7625 Free => OPENSSL_free
7628 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7629 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7632 *) CygWin32 support.
7633 [John Jarvie <jjarvie@newsguy.com>]
7635 *) The type-safe stack code has been rejigged. It is now only compiled
7636 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7637 by default all type-specific stack functions are "#define"d back to
7638 standard stack functions. This results in more streamlined output
7639 but retains the type-safety checking possibilities of the original
7643 *) The STACK code has been cleaned up, and certain type declarations
7644 that didn't make a lot of sense have been brought in line. This has
7645 also involved a cleanup of sorts in safestack.h to more correctly
7646 map type-safe stack functions onto their plain stack counterparts.
7647 This work has also resulted in a variety of "const"ifications of
7648 lots of the code, especially "_cmp" operations which should normally
7649 be prototyped with "const" parameters anyway.
7652 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7653 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7654 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7655 where all of 'md' is used each time the PRNG is used, but 'state'
7656 is used only indexed by a cyclic counter. As entropy may not be
7657 well distributed from the beginning, 'md' is important as a
7658 chaining variable. However, the output function chains only half
7659 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7660 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7661 in all of 'state' being rewritten, with the new values depending
7662 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7665 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7666 the handshake is continued after ssl_verify_cert_chain();
7667 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7668 can lead to 'unexplainable' connection aborts later.
7669 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7671 *) Major EVP API cipher revision.
7672 Add hooks for extra EVP features. This allows various cipher
7673 parameters to be set in the EVP interface. Support added for variable
7674 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7675 setting of RC2 and RC5 parameters.
7677 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7680 Remove lots of duplicated code from the EVP library. For example *every*
7681 cipher init() function handles the 'iv' in the same way according to the
7682 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7683 for CFB and OFB modes they zero ctx->num.
7685 New functionality allows removal of S/MIME code RC2 hack.
7687 Most of the routines have the same form and so can be declared in terms
7690 By shifting this to the top level EVP_CipherInit() it can be removed from
7691 all individual ciphers. If the cipher wants to handle IVs or keys
7692 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7695 Change lots of functions like EVP_EncryptUpdate() to now return a
7696 value: although software versions of the algorithms cannot fail
7697 any installed hardware versions can.
7700 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7701 this option is set, tolerate broken clients that send the negotiated
7702 protocol version number instead of the requested protocol version
7706 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7707 i.e. non-zero for export ciphersuites, zero otherwise.
7708 Previous versions had this flag inverted, inconsistent with
7709 rsa_tmp_cb (..._TMP_RSA_CB).
7710 [Bodo Moeller; problem reported by Amit Chopra]
7712 *) Add missing DSA library text string. Work around for some IIS
7713 key files with invalid SEQUENCE encoding.
7716 *) Add a document (doc/standards.txt) that list all kinds of standards
7717 and so on that are implemented in OpenSSL.
7720 *) Enhance c_rehash script. Old version would mishandle certificates
7721 with the same subject name hash and wouldn't handle CRLs at all.
7722 Added -fingerprint option to crl utility, to support new c_rehash
7726 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7729 *) Fix for SSL server purpose checking. Server checking was
7730 rejecting certificates which had extended key usage present
7731 but no ssl client purpose.
7732 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7734 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7735 is a little unclear about how a blank password is handled.
7736 Since the password in encoded as a BMPString with terminating
7737 double NULL a zero length password would end up as just the
7738 double NULL. However no password at all is different and is
7739 handled differently in the PKCS#12 key generation code. NS
7740 treats a blank password as zero length. MSIE treats it as no
7741 password on export: but it will try both on import. We now do
7742 the same: PKCS12_parse() tries zero length and no password if
7743 the password is set to "" or NULL (NULL is now a valid password:
7744 it wasn't before) as does the pkcs12 application.
7747 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7748 perror when PEM_read_bio_X509_REQ fails, the error message must
7749 be obtained from the error queue.
7752 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7753 it in ERR_remove_state if appropriate, and change ERR_get_state
7754 accordingly to avoid race conditions (this is necessary because
7755 thread_hash is no longer constant once set).
7758 *) Bugfix for linux-elf makefile.one.
7761 *) RSA_get_default_method() will now cause a default
7762 RSA_METHOD to be chosen if one doesn't exist already.
7763 Previously this was only set during a call to RSA_new()
7764 or RSA_new_method(NULL) meaning it was possible for
7765 RSA_get_default_method() to return NULL.
7768 *) Added native name translation to the existing DSO code
7769 that will convert (if the flag to do so is set) filenames
7770 that are sufficiently small and have no path information
7771 into a canonical native form. Eg. "blah" converted to
7772 "libblah.so" or "blah.dll" etc.
7775 *) New function ERR_error_string_n(e, buf, len) which is like
7776 ERR_error_string(e, buf), but writes at most 'len' bytes
7777 including the 0 terminator. For ERR_error_string_n, 'buf'
7779 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7781 *) CONF library reworked to become more general. A new CONF
7782 configuration file reader "class" is implemented as well as a
7783 new functions (NCONF_*, for "New CONF") to handle it. The now
7784 old CONF_* functions are still there, but are reimplemented to
7785 work in terms of the new functions. Also, a set of functions
7786 to handle the internal storage of the configuration data is
7787 provided to make it easier to write new configuration file
7788 reader "classes" (I can definitely see something reading a
7789 configuration file in XML format, for example), called _CONF_*,
7790 or "the configuration storage API"...
7792 The new configuration file reading functions are:
7794 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7795 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7797 NCONF_default, NCONF_WIN32
7799 NCONF_dump_fp, NCONF_dump_bio
7801 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7802 NCONF_new creates a new CONF object. This works in the same way
7803 as other interfaces in OpenSSL, like the BIO interface.
7804 NCONF_dump_* dump the internal storage of the configuration file,
7805 which is useful for debugging. All other functions take the same
7806 arguments as the old CONF_* functions wth the exception of the
7807 first that must be a `CONF *' instead of a `LHASH *'.
7809 To make it easer to use the new classes with the old CONF_* functions,
7810 the function CONF_set_default_method is provided.
7813 *) Add '-tls1' option to 'openssl ciphers', which was already
7814 mentioned in the documentation but had not been implemented.
7815 (This option is not yet really useful because even the additional
7816 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7819 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7820 OpenSSL-based applications) load shared libraries and bind to
7821 them in a portable way.
7822 [Geoff Thorpe, with contributions from Richard Levitte]
7824 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7826 *) Make sure _lrotl and _lrotr are only used with MSVC.
7828 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7829 (the default implementation of RAND_status).
7831 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7832 to '-clrext' (= clear extensions), as intended and documented.
7833 [Bodo Moeller; inconsistency pointed out by Michael Attili
7836 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7837 was larger than the MD block size.
7838 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7840 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7841 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7842 using the passed key: if the passed key was a private key the result
7843 of X509_print(), for example, would be to print out all the private key
7847 *) des_quad_cksum() byte order bug fix.
7848 [Ulf Möller, using the problem description in krb4-0.9.7, where
7849 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7851 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7853 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7855 *) For easily testing in shell scripts whether some command
7856 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7857 returns with exit code 0 iff no command of the given name is available.
7858 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7859 the output goes to stdout and nothing is printed to stderr.
7860 Additional arguments are always ignored.
7862 Since for each cipher there is a command of the same name,
7863 the 'no-cipher' compilation switches can be tested this way.
7865 ('openssl no-XXX' is not able to detect pseudo-commands such
7866 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7869 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7872 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7873 is set; it will be thrown away anyway because each handshake creates
7875 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7876 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7877 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7878 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7881 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7882 'Q' and 'R' lose their special meanings (quit/renegotiate).
7883 This is part of what -quiet does; unlike -quiet, -ign_eof
7884 does not suppress any output.
7887 *) Add compatibility options to the purpose and trust code. The
7888 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7889 accepts a certificate or CA, this was the previous behaviour,
7890 with all the associated security issues.
7892 X509_TRUST_COMPAT is the old trust behaviour: only and
7893 automatically trust self signed roots in certificate store. A
7894 new trust setting X509_TRUST_DEFAULT is used to specify that
7895 a purpose has no associated trust setting and it should instead
7896 use the value in the default purpose.
7899 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7900 and fix a memory leak.
7903 *) In util/mkerr.pl (which implements 'make errors'), preserve
7904 reason strings from the previous version of the .c file, as
7905 the default to have only downcase letters (and digits) in
7906 automatically generated reasons codes is not always appropriate.
7909 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7910 using strerror. Previously, ERR_reason_error_string() returned
7911 library names as reason strings for SYSerr; but SYSerr is a special
7912 case where small numbers are errno values, not library numbers.
7915 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7916 converts DSA parameters into DH parameters. (When creating parameters,
7917 DSA_generate_parameters is used.)
7920 *) Include 'length' (recommended exponent length) in C code generated
7921 by 'openssl dhparam -C'.
7924 *) The second argument to set_label in perlasm was already being used
7925 so couldn't be used as a "file scope" flag. Moved to third argument
7929 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7930 instead of RAND_bytes for encryption IVs and salts.
7933 *) Include RAND_status() into RAND_METHOD instead of implementing
7934 it only for md_rand.c Otherwise replacing the PRNG by calling
7935 RAND_set_rand_method would be impossible.
7938 *) Don't let DSA_generate_key() enter an infinite loop if the random
7939 number generation fails.
7942 *) New 'rand' application for creating pseudo-random output.
7945 *) Added configuration support for Linux/IA64
7946 [Rolf Haberrecker <rolf@suse.de>]
7948 *) Assembler module support for Mingw32.
7951 *) Shared library support for HPUX (in shlib/).
7952 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7954 *) Shared library support for Solaris gcc.
7955 [Lutz Behnke <behnke@trustcenter.de>]
7957 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7959 *) PKCS7_encrypt() was adding text MIME headers twice because they
7960 were added manually and by SMIME_crlf_copy().
7963 *) In bntest.c don't call BN_rand with zero bits argument.
7964 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7966 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7967 case was implemented. This caused BN_div_recp() to fail occasionally.
7970 *) Add an optional second argument to the set_label() in the perl
7971 assembly language builder. If this argument exists and is set
7972 to 1 it signals that the assembler should use a symbol whose
7973 scope is the entire file, not just the current function. This
7974 is needed with MASM which uses the format label:: for this scope.
7975 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7977 *) Change the ASN1 types so they are typedefs by default. Before
7978 almost all types were #define'd to ASN1_STRING which was causing
7979 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7983 *) Change names of new functions to the new get1/get0 naming
7984 convention: After 'get1', the caller owns a reference count
7985 and has to call ..._free; 'get0' returns a pointer to some
7986 data structure without incrementing reference counters.
7987 (Some of the existing 'get' functions increment a reference
7988 counter, some don't.)
7989 Similarly, 'set1' and 'add1' functions increase reference
7990 counters or duplicate objects.
7993 *) Allow for the possibility of temp RSA key generation failure:
7994 the code used to assume it always worked and crashed on failure.
7997 *) Fix potential buffer overrun problem in BIO_printf().
7998 [Ulf Möller, using public domain code by Patrick Powell; problem
7999 pointed out by David Sacerdote <das33@cornell.edu>]
8001 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8002 RAND_egd() and RAND_status(). In the command line application,
8003 the EGD socket can be specified like a seed file using RANDFILE
8007 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8008 Some CAs (e.g. Verisign) distribute certificates in this form.
8011 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8012 list to exclude them. This means that no special compilation option
8013 is needed to use anonymous DH: it just needs to be included in the
8017 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8018 EVP_MD_type. The old functionality is available in a new macro called
8019 EVP_MD_md(). Change code that uses it and update docs.
8022 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8023 where the 'void *' argument is replaced by a function pointer argument.
8024 Previously 'void *' was abused to point to functions, which works on
8025 many platforms, but is not correct. As these functions are usually
8026 called by macros defined in OpenSSL header files, most source code
8027 should work without changes.
8030 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8031 sections with information on -D... compiler switches used for
8032 compiling the library so that applications can see them. To enable
8033 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8034 must be defined. E.g.,
8035 #define OPENSSL_ALGORITHM_DEFINES
8036 #include <openssl/opensslconf.h>
8037 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8038 [Richard Levitte, Ulf and Bodo Möller]
8040 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8044 *) Change the 'other' type in certificate aux info to a STACK_OF
8045 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8046 the required ASN1 format: arbitrary types determined by an OID.
8049 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8050 argument to 'req'. This is not because the function is newer or
8051 better than others it just uses the work 'NEW' in the certificate
8052 request header lines. Some software needs this.
8055 *) Reorganise password command line arguments: now passwords can be
8056 obtained from various sources. Delete the PEM_cb function and make
8057 it the default behaviour: i.e. if the callback is NULL and the
8058 usrdata argument is not NULL interpret it as a null terminated pass
8059 phrase. If usrdata and the callback are NULL then the pass phrase
8060 is prompted for as usual.
8063 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8064 the support is automatically enabled. The resulting binaries will
8065 autodetect the card and use it if present.
8066 [Ben Laurie and Compaq Inc.]
8068 *) Work around for Netscape hang bug. This sends certificate request
8069 and server done in one record. Since this is perfectly legal in the
8070 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8071 the bugs/SSLv3 entry for more info.
8074 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8077 *) Add -rand argument to smime and pkcs12 applications and read/write
8081 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8084 *) Add command line password options to the remaining applications.
8087 *) Bug fix for BN_div_recp() for numerators with an even number of
8091 *) More tests in bntest.c, and changed test_bn output.
8094 *) ./config recognizes MacOS X now.
8097 *) Bug fix for BN_div() when the first words of num and divsor are
8098 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8101 *) Add support for various broken PKCS#8 formats, and command line
8102 options to produce them.
8105 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8106 get temporary BIGNUMs from a BN_CTX.
8109 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8113 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8114 include a #define from the old name to the new. The original intent
8115 was that statically linked binaries could for example just call
8116 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8117 link with digests. This never worked becayse SSLeay_add_all_digests()
8118 and SSLeay_add_all_ciphers() were in the same source file so calling
8119 one would link with the other. They are now in separate source files.
8122 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8125 *) Use a less unusual form of the Miller-Rabin primality test (it used
8126 a binary algorithm for exponentiation integrated into the Miller-Rabin
8127 loop, our standard modexp algorithms are faster).
8130 *) Support for the EBCDIC character set completed.
8131 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8133 *) Source code cleanups: use const where appropriate, eliminate casts,
8134 use void * instead of char * in lhash.
8137 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8138 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8139 this the server could overwrite ephemeral keys that the client
8143 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8144 using 50 iterations of the Rabin-Miller test.
8146 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8147 iterations of the Rabin-Miller test as required by the appendix
8148 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8149 As BN_is_prime_fasttest includes trial division, DSA parameter
8150 generation becomes much faster.
8152 This implies a change for the callback functions in DSA_is_prime
8153 and DSA_generate_parameters: The callback function is called once
8154 for each positive witness in the Rabin-Miller test, not just
8155 occasionally in the inner loop; and the parameters to the
8156 callback function now provide an iteration count for the outer
8157 loop rather than for the current invocation of the inner loop.
8158 DSA_generate_parameters additionally can call the callback
8159 function with an 'iteration count' of -1, meaning that a
8160 candidate has passed the trial division test (when q is generated
8161 from an application-provided seed, trial division is skipped).
8164 *) New function BN_is_prime_fasttest that optionally does trial
8165 division before starting the Rabin-Miller test and has
8166 an additional BN_CTX * argument (whereas BN_is_prime always
8167 has to allocate at least one BN_CTX).
8168 'callback(1, -1, cb_arg)' is called when a number has passed the
8169 trial division stage.
8172 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8176 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8179 *) New function BN_pseudo_rand().
8182 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8183 bignum version of BN_from_montgomery() with the working code from
8184 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8188 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8189 made it impossible to use the same SSL_SESSION data structure in
8190 SSL2 clients in multiple threads.
8193 *) The return value of RAND_load_file() no longer counts bytes obtained
8194 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8195 to seed the PRNG (previously an explicit byte count was required).
8196 [Ulf Möller, Bodo Möller]
8198 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8199 used (char *) instead of (void *) and had casts all over the place.
8202 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8205 *) Retain source code compatibility for BN_prime_checks macro:
8206 BN_is_prime(..., BN_prime_checks, ...) now uses
8207 BN_prime_checks_for_size to determine the appropriate number of
8208 Rabin-Miller iterations.
8211 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8212 DH_CHECK_P_NOT_SAFE_PRIME.
8213 (Check if this is true? OpenPGP calls them "strong".)
8216 *) Merge the functionality of "dh" and "gendh" programs into a new program
8217 "dhparam". The old programs are retained for now but will handle DH keys
8218 (instead of parameters) in future.
8221 *) Make the ciphers, s_server and s_client programs check the return values
8222 when a new cipher list is set.
8225 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8226 ciphers. Before when the 56bit ciphers were enabled the sorting was
8229 The syntax for the cipher sorting has been extended to support sorting by
8230 cipher-strength (using the strength_bits hard coded in the tables).
8231 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8233 Fix a bug in the cipher-command parser: when supplying a cipher command
8234 string with an "undefined" symbol (neither command nor alphanumeric
8235 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8236 an error is flagged.
8238 Due to the strength-sorting extension, the code of the
8239 ssl_create_cipher_list() function was completely rearranged. I hope that
8240 the readability was also increased :-)
8241 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8243 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8244 for the first serial number and places 2 in the serial number file. This
8245 avoids problems when the root CA is created with serial number zero and
8246 the first user certificate has the same issuer name and serial number
8250 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8251 the new code. Add documentation for this stuff.
8254 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8255 X509_*() to X509at_*() on the grounds that they don't handle X509
8256 structures and behave in an analagous way to the X509v3 functions:
8257 they shouldn't be called directly but wrapper functions should be used
8260 So we also now have some wrapper functions that call the X509at functions
8261 when passed certificate requests. (TO DO: similar things can be done with
8262 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8263 things. Some of these need some d2i or i2d and print functionality
8264 because they handle more complex structures.)
8267 *) Add missing #ifndefs that caused missing symbols when building libssl
8268 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8269 NO_RSA in ssl/s2*.c.
8270 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8272 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8273 has a return value which indicates the quality of the random data
8274 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8275 error queue. New function RAND_pseudo_bytes() generates output that is
8276 guaranteed to be unique but not unpredictable. RAND_add is like
8277 RAND_seed, but takes an extra argument for an entropy estimate
8278 (RAND_seed always assumes full entropy).
8281 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8282 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8283 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8284 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8285 false-positive rate of at most 2^-80 for random input.
8288 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8291 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8292 in the 0.9.5 release), this returns the chain
8293 from an X509_CTX structure with a dup of the stack and all
8294 the X509 reference counts upped: so the stack will exist
8295 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8298 Also make SSL_SESSION_print() print out the verify return
8302 *) Add manpage for the pkcs12 command. Also change the default
8303 behaviour so MAC iteration counts are used unless the new
8304 -nomaciter option is used. This improves file security and
8305 only older versions of MSIE (4.0 for example) need it.
8308 *) Honor the no-xxx Configure options when creating .DEF files.
8311 *) Add PKCS#10 attributes to field table: challengePassword,
8312 unstructuredName and unstructuredAddress. These are taken from
8313 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8314 international characters are used.
8316 More changes to X509_ATTRIBUTE code: allow the setting of types
8317 based on strings. Remove the 'loc' parameter when adding
8318 attributes because these will be a SET OF encoding which is sorted
8322 *) Initial changes to the 'req' utility to allow request generation
8323 automation. This will allow an application to just generate a template
8324 file containing all the field values and have req construct the
8327 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8328 used all over the place including certificate requests and PKCS#7
8329 structures. They are currently handled manually where necessary with
8330 some primitive wrappers for PKCS#7. The new functions behave in a
8331 manner analogous to the X509 extension functions: they allow
8332 attributes to be looked up by NID and added.
8334 Later something similar to the X509V3 code would be desirable to
8335 automatically handle the encoding, decoding and printing of the
8336 more complex types. The string types like challengePassword can
8337 be handled by the string table functions.
8339 Also modified the multi byte string table handling. Now there is
8340 a 'global mask' which masks out certain types. The table itself
8341 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8342 is useful when for example there is only one permissible type
8343 (as in countryName) and using the mask might result in no valid
8347 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8348 SSL_get_peer_finished to allow applications to obtain the latest
8349 Finished messages sent to the peer or expected from the peer,
8350 respectively. (SSL_get_peer_finished is usually the Finished message
8351 actually received from the peer, otherwise the protocol will be aborted.)
8353 As the Finished message are message digests of the complete handshake
8354 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8355 be used for external authentication procedures when the authentication
8356 provided by SSL/TLS is not desired or is not enough.
8359 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8360 the host supports BWX extension and if Compaq C is present on the
8361 $PATH. Just exploiting of the BWX extension results in 20-30%
8362 performance kick for some algorithms, e.g. DES and RC4 to mention
8363 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8367 *) Add support for MS "fast SGC". This is arguably a violation of the
8368 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8369 weak crypto and after checking the certificate is SGC a second one
8370 with strong crypto. MS SGC stops the first handshake after receiving
8371 the server certificate message and sends a second client hello. Since
8372 a server will typically do all the time consuming operations before
8373 expecting any further messages from the client (server key exchange
8374 is the most expensive) there is little difference between the two.
8376 To get OpenSSL to support MS SGC we have to permit a second client
8377 hello message after we have sent server done. In addition we have to
8378 reset the MAC if we do get this second client hello.
8381 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8382 if a DER encoded private key is RSA or DSA traditional format. Changed
8383 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8384 format DER encoded private key. Newer code should use PKCS#8 format which
8385 has the key type encoded in the ASN1 structure. Added DER private key
8386 support to pkcs8 application.
8389 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8390 ciphersuites has been selected (as required by the SSL 3/TLS 1
8391 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8392 is set, we interpret this as a request to violate the specification
8393 (the worst that can happen is a handshake failure, and 'correct'
8394 behaviour would result in a handshake failure anyway).
8397 *) In SSL_CTX_add_session, take into account that there might be multiple
8398 SSL_SESSION structures with the same session ID (e.g. when two threads
8399 concurrently obtain them from an external cache).
8400 The internal cache can handle only one SSL_SESSION with a given ID,
8401 so if there's a conflict, we now throw out the old one to achieve
8405 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8406 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8407 some routines that use cipher OIDs: some ciphers do not have OIDs
8408 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8412 *) Simplify the trust setting structure and code. Now we just have
8413 two sequences of OIDs for trusted and rejected settings. These will
8414 typically have values the same as the extended key usage extension
8415 and any application specific purposes.
8417 The trust checking code now has a default behaviour: it will just
8418 check for an object with the same NID as the passed id. Functions can
8419 be provided to override either the default behaviour or the behaviour
8420 for a given id. SSL client, server and email already have functions
8421 in place for compatibility: they check the NID and also return "trusted"
8422 if the certificate is self signed.
8425 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8426 traditional format into an EVP_PKEY structure.
8429 *) Add a password callback function PEM_cb() which either prompts for
8430 a password if usr_data is NULL or otherwise assumes it is a null
8431 terminated password. Allow passwords to be passed on command line
8432 environment or config files in a few more utilities.
8435 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8436 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8437 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8438 Update documentation.
8441 *) Support for ASN1 "NULL" type. This could be handled before by using
8442 ASN1_TYPE but there wasn't any function that would try to read a NULL
8443 and produce an error if it couldn't. For compatibility we also have
8444 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8445 don't allocate anything because they don't need to.
8448 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8450 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8452 *) Rebuild of the memory allocation routines used by OpenSSL code and
8453 possibly others as well. The purpose is to make an interface that
8454 provide hooks so anyone can build a separate set of allocation and
8455 deallocation routines to be used by OpenSSL, for example memory
8456 pool implementations, or something else, which was previously hard
8457 since Malloc(), Realloc() and Free() were defined as macros having
8458 the values malloc, realloc and free, respectively (except for Win32
8459 compilations). The same is provided for memory debugging code.
8460 OpenSSL already comes with functionality to find memory leaks, but
8461 this gives people a chance to debug other memory problems.
8463 With these changes, a new set of functions and macros have appeared:
8465 CRYPTO_set_mem_debug_functions() [F]
8466 CRYPTO_get_mem_debug_functions() [F]
8467 CRYPTO_dbg_set_options() [F]
8468 CRYPTO_dbg_get_options() [F]
8469 CRYPTO_malloc_debug_init() [M]
8471 The memory debug functions are NULL by default, unless the library
8472 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8473 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8474 gives the standard debugging functions that come with OpenSSL) or
8475 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8476 provided by the library user) must be used. When the standard
8477 debugging functions are used, CRYPTO_dbg_set_options can be used to
8478 request additional information:
8479 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8480 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8482 Also, things like CRYPTO_set_mem_functions will always give the
8483 expected result (the new set of functions is used for allocation
8484 and deallocation) at all times, regardless of platform and compiler
8487 To finish it up, some functions that were never use in any other
8488 way than through macros have a new API and new semantic:
8491 CRYPTO_dbg_realloc()
8494 All macros of value have retained their old syntax.
8495 [Richard Levitte and Bodo Moeller]
8497 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8498 ordering of SMIMECapabilities wasn't in "strength order" and there
8499 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8503 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8504 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8505 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8507 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8508 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8509 functionality to handle multipart/signed properly) and a utility
8510 called 'smime' to call all this stuff. This is based on code I
8511 originally wrote for Celo who have kindly allowed it to be
8512 included in OpenSSL.
8515 *) Add variants des_set_key_checked and des_set_key_unchecked of
8516 des_set_key (aka des_key_sched). Global variable des_check_key
8517 decides which of these is called by des_set_key; this way
8518 des_check_key behaves as it always did, but applications and
8519 the library itself, which was buggy for des_check_key == 1,
8520 have a cleaner way to pick the version they need.
8523 *) New function PKCS12_newpass() which changes the password of a
8527 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8528 dynamic mix. In both cases the ids can be used as an index into the
8529 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8530 functions so they accept a list of the field values and the
8531 application doesn't need to directly manipulate the X509_TRUST
8535 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8539 *) Modify the way the V3 extension code looks up extensions. This now
8540 works in a similar way to the object code: we have some "standard"
8541 extensions in a static table which is searched with OBJ_bsearch()
8542 and the application can add dynamic ones if needed. The file
8543 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8544 updated whenever a new extension is added to the core code and kept
8545 in ext_nid order. There is a simple program 'tabtest.c' which checks
8546 this. New extensions are not added too often so this file can readily
8547 be maintained manually.
8549 There are two big advantages in doing things this way. The extensions
8550 can be looked up immediately and no longer need to be "added" using
8551 X509V3_add_standard_extensions(): this function now does nothing.
8552 [Side note: I get *lots* of email saying the extension code doesn't
8553 work because people forget to call this function]
8554 Also no dynamic allocation is done unless new extensions are added:
8555 so if we don't add custom extensions there is no need to call
8556 X509V3_EXT_cleanup().
8559 *) Modify enc utility's salting as follows: make salting the default. Add a
8560 magic header, so unsalted files fail gracefully instead of just decrypting
8561 to garbage. This is because not salting is a big security hole, so people
8562 should be discouraged from doing it.
8565 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8566 digest to be passed on the command line but it only used this
8567 parameter when signing a certificate. Modified so all relevant
8568 operations are affected by the digest parameter including the
8569 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8570 DSA key was used because it didn't fix the digest.
8573 *) Initial certificate chain verify code. Currently tests the untrusted
8574 certificates for consistency with the verify purpose (which is set
8575 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8577 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8578 this is because it will reject chains with invalid extensions whereas
8579 every previous version of OpenSSL and SSLeay made no checks at all.
8581 Trust code: checks the root CA for the relevant trust settings. Trust
8582 settings have an initial value consistent with the verify purpose: e.g.
8583 if the verify purpose is for SSL client use it expects the CA to be
8584 trusted for SSL client use. However the default value can be changed to
8585 permit custom trust settings: one example of this would be to only trust
8586 certificates from a specific "secure" set of CAs.
8588 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8589 which should be used for version portability: especially since the
8590 verify structure is likely to change more often now.
8592 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8593 to set them. If not set then assume SSL clients will verify SSL servers
8596 Two new options to the verify program: -untrusted allows a set of
8597 untrusted certificates to be passed in and -purpose which sets the
8598 intended purpose of the certificate. If a purpose is set then the
8599 new chain verify code is used to check extension consistency.
8602 *) Support for the authority information access extension.
8605 *) Modify RSA and DSA PEM read routines to transparently handle
8606 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8607 public keys in a format compatible with certificate
8608 SubjectPublicKeyInfo structures. Unfortunately there were already
8609 functions called *_PublicKey_* which used various odd formats so
8610 these are retained for compatibility: however the DSA variants were
8611 never in a public release so they have been deleted. Changed dsa/rsa
8612 utilities to handle the new format: note no releases ever handled public
8613 keys so we should be OK.
8615 The primary motivation for this change is to avoid the same fiasco
8616 that dogs private keys: there are several incompatible private key
8617 formats some of which are standard and some OpenSSL specific and
8618 require various evil hacks to allow partial transparent handling and
8619 even then it doesn't work with DER formats. Given the option anything
8620 other than PKCS#8 should be dumped: but the other formats have to
8621 stay in the name of compatibility.
8623 With public keys and the benefit of hindsight one standard format
8624 is used which works with EVP_PKEY, RSA or DSA structures: though
8625 it clearly returns an error if you try to read the wrong kind of key.
8627 Added a -pubkey option to the 'x509' utility to output the public key.
8628 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8629 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8630 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8631 that do the same as the EVP_PKEY_assign_*() except they up the
8632 reference count of the added key (they don't "swallow" the
8636 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8637 CRLs would fail if the file contained no certificates or no CRLs:
8638 added a new function to read in both types and return the number
8639 read: this means that if none are read it will be an error. The
8640 DER versions of the certificate and CRL reader would always fail
8641 because it isn't possible to mix certificates and CRLs in DER format
8642 without choking one or the other routine. Changed this to just read
8643 a certificate: this is the best we can do. Also modified the code
8644 in apps/verify.c to take notice of return codes: it was previously
8645 attempting to read in certificates from NULL pointers and ignoring
8646 any errors: this is one reason why the cert and CRL reader seemed
8647 to work. It doesn't check return codes from the default certificate
8648 routines: these may well fail if the certificates aren't installed.
8651 *) Code to support otherName option in GeneralName.
8654 *) First update to verify code. Change the verify utility
8655 so it warns if it is passed a self signed certificate:
8656 for consistency with the normal behaviour. X509_verify
8657 has been modified to it will now verify a self signed
8658 certificate if *exactly* the same certificate appears
8659 in the store: it was previously impossible to trust a
8660 single self signed certificate. This means that:
8661 openssl verify ss.pem
8662 now gives a warning about a self signed certificate but
8663 openssl verify -CAfile ss.pem ss.pem
8667 *) For servers, store verify_result in SSL_SESSION data structure
8668 (and add it to external session representation).
8669 This is needed when client certificate verifications fails,
8670 but an application-provided verification callback (set by
8671 SSL_CTX_set_cert_verify_callback) allows accepting the session
8672 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8673 but returns 1): When the session is reused, we have to set
8674 ssl->verify_result to the appropriate error code to avoid
8676 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8678 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8679 case in PKCS7_dataInit() where the signed PKCS7 structure
8680 didn't contain any existing data because it was being created.
8681 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8683 *) Add a salt to the key derivation routines in enc.c. This
8684 forms the first 8 bytes of the encrypted file. Also add a
8685 -S option to allow a salt to be input on the command line.
8688 *) New function X509_cmp(). Oddly enough there wasn't a function
8689 to compare two certificates. We do this by working out the SHA1
8690 hash and comparing that. X509_cmp() will be needed by the trust
8694 *) SSL_get1_session() is like SSL_get_session(), but increments
8695 the reference count in the SSL_SESSION returned.
8696 [Geoff Thorpe <geoff@eu.c2.net>]
8698 *) Fix for 'req': it was adding a null to request attributes.
8699 Also change the X509_LOOKUP and X509_INFO code to handle
8700 certificate auxiliary information.
8703 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8707 *) Add the possibility to add extra information to the memory leak
8708 detecting output, to form tracebacks, showing from where each
8709 allocation was originated: CRYPTO_push_info("constant string") adds
8710 the string plus current file name and line number to a per-thread
8711 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8712 is like calling CYRPTO_pop_info() until the stack is empty.
8713 Also updated memory leak detection code to be multi-thread-safe.
8716 *) Add options -text and -noout to pkcs7 utility and delete the
8717 encryption options which never did anything. Update docs.
8720 *) Add options to some of the utilities to allow the pass phrase
8721 to be included on either the command line (not recommended on
8722 OSes like Unix) or read from the environment. Update the
8723 manpages and fix a few bugs.
8726 *) Add a few manpages for some of the openssl commands.
8729 *) Fix the -revoke option in ca. It was freeing up memory twice,
8730 leaking and not finding already revoked certificates.
8733 *) Extensive changes to support certificate auxiliary information.
8734 This involves the use of X509_CERT_AUX structure and X509_AUX
8735 functions. An X509_AUX function such as PEM_read_X509_AUX()
8736 can still read in a certificate file in the usual way but it
8737 will also read in any additional "auxiliary information". By
8738 doing things this way a fair degree of compatibility can be
8739 retained: existing certificates can have this information added
8740 using the new 'x509' options.
8742 Current auxiliary information includes an "alias" and some trust
8743 settings. The trust settings will ultimately be used in enhanced
8744 certificate chain verification routines: currently a certificate
8745 can only be trusted if it is self signed and then it is trusted
8749 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8750 The problem was that one of the replacement routines had not been working
8751 since SSLeay releases. For now the offending routine has been replaced
8752 with non-optimised assembler. Even so, this now gives around 95%
8753 performance improvement for 1024 bit RSA signs.
8756 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8757 handling. Most clients have the effective key size in bits equal to
8758 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8759 A few however don't do this and instead use the size of the decrypted key
8760 to determine the RC2 key length and the AlgorithmIdentifier to determine
8761 the effective key length. In this case the effective key length can still
8762 be 40 bits but the key length can be 168 bits for example. This is fixed
8763 by manually forcing an RC2 key into the EVP_PKEY structure because the
8764 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8765 the key length and effective key length are equal.
8768 *) Add a bunch of functions that should simplify the creation of
8769 X509_NAME structures. Now you should be able to do:
8770 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8771 and have it automatically work out the correct field type and fill in
8772 the structures. The more adventurous can try:
8773 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8774 and it will (hopefully) work out the correct multibyte encoding.
8777 *) Change the 'req' utility to use the new field handling and multibyte
8778 copy routines. Before the DN field creation was handled in an ad hoc
8779 way in req, ca, and x509 which was rather broken and didn't support
8780 BMPStrings or UTF8Strings. Since some software doesn't implement
8781 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8782 using the dirstring_type option. See the new comment in the default
8783 openssl.cnf for more info.
8786 *) Make crypto/rand/md_rand.c more robust:
8787 - Assure unique random numbers after fork().
8788 - Make sure that concurrent threads access the global counter and
8789 md serializably so that we never lose entropy in them
8790 or use exactly the same state in multiple threads.
8791 Access to the large state is not always serializable because
8792 the additional locking could be a performance killer, and
8793 md should be large enough anyway.
8796 *) New file apps/app_rand.c with commonly needed functionality
8797 for handling the random seed file.
8799 Use the random seed file in some applications that previously did not:
8801 dsaparam -genkey (which also ignored its '-rand' option),
8804 x509 (when signing).
8805 Except on systems with /dev/urandom, it is crucial to have a random
8806 seed file at least for key creation, DSA signing, and for DH exchanges;
8807 for RSA signatures we could do without one.
8809 gendh and gendsa (unlike genrsa) used to read only the first byte
8810 of each file listed in the '-rand' option. The function as previously
8811 found in genrsa is now in app_rand.c and is used by all programs
8812 that support '-rand'.
8815 *) In RAND_write_file, use mode 0600 for creating files;
8816 don't just chmod when it may be too late.
8819 *) Report an error from X509_STORE_load_locations
8820 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8823 *) New function ASN1_mbstring_copy() this copies a string in either
8824 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8825 into an ASN1_STRING type. A mask of permissible types is passed
8826 and it chooses the "minimal" type to use or an error if not type
8830 *) Add function equivalents to the various macros in asn1.h. The old
8831 macros are retained with an M_ prefix. Code inside the library can
8832 use the M_ macros. External code (including the openssl utility)
8833 should *NOT* in order to be "shared library friendly".
8836 *) Add various functions that can check a certificate's extensions
8837 to see if it usable for various purposes such as SSL client,
8838 server or S/MIME and CAs of these types. This is currently
8839 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8840 verification. Also added a -purpose flag to x509 utility to
8841 print out all the purposes.
8844 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8848 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8849 for, obtain and decode and extension and obtain its critical flag.
8850 This allows all the necessary extension code to be handled in a
8851 single function call.
8854 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8855 platforms. See crypto/rc4/rc4_enc.c for further details.
8858 *) New -noout option to asn1parse. This causes no output to be produced
8859 its main use is when combined with -strparse and -out to extract data
8860 from a file (which may not be in ASN.1 format).
8863 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8864 when producing the local key id.
8865 [Richard Levitte <levitte@stacken.kth.se>]
8867 *) New option -dhparam in s_server. This allows a DH parameter file to be
8868 stated explicitly. If it is not stated then it tries the first server
8869 certificate file. The previous behaviour hard coded the filename
8873 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8874 a public key to be input or output. For example:
8875 openssl rsa -in key.pem -pubout -out pubkey.pem
8876 Also added necessary DSA public key functions to handle this.
8879 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8880 in the message. This was handled by allowing
8881 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8882 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8884 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8885 to the end of the strings whereas this didn't. This would cause problems
8886 if strings read with d2i_ASN1_bytes() were later modified.
8887 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8889 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8890 data and it contains EOF it will end up returning an error. This is
8891 caused by input 46 bytes long. The cause is due to the way base64
8892 BIOs find the start of base64 encoded data. They do this by trying a
8893 trial decode on each line until they find one that works. When they
8894 do a flag is set and it starts again knowing it can pass all the
8895 data directly through the decoder. Unfortunately it doesn't reset
8896 the context it uses. This means that if EOF is reached an attempt
8897 is made to pass two EOFs through the context and this causes the
8898 resulting error. This can also cause other problems as well. As is
8899 usual with these problems it takes *ages* to find and the fix is
8900 trivial: move one line.
8901 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8903 *) Ugly workaround to get s_client and s_server working under Windows. The
8904 old code wouldn't work because it needed to select() on sockets and the
8905 tty (for keypresses and to see if data could be written). Win32 only
8906 supports select() on sockets so we select() with a 1s timeout on the
8907 sockets and then see if any characters are waiting to be read, if none
8908 are present then we retry, we also assume we can always write data to
8909 the tty. This isn't nice because the code then blocks until we've
8910 received a complete line of data and it is effectively polling the
8911 keyboard at 1s intervals: however it's quite a bit better than not
8912 working at all :-) A dedicated Windows application might handle this
8913 with an event loop for example.
8916 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8917 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8918 will be called when RSA_sign() and RSA_verify() are used. This is useful
8919 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8920 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8921 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8922 This necessitated the support of an extra signature type NID_md5_sha1
8923 for SSL signatures and modifications to the SSL library to use it instead
8924 of calling RSA_public_decrypt() and RSA_private_encrypt().
8927 *) Add new -verify -CAfile and -CApath options to the crl program, these
8928 will lookup a CRL issuers certificate and verify the signature in a
8929 similar way to the verify program. Tidy up the crl program so it
8930 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8931 less strict. It will now permit CRL extensions even if it is not
8932 a V2 CRL: this will allow it to tolerate some broken CRLs.
8935 *) Initialize all non-automatic variables each time one of the openssl
8936 sub-programs is started (this is necessary as they may be started
8937 multiple times from the "OpenSSL>" prompt).
8938 [Lennart Bang, Bodo Moeller]
8940 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8941 removing all other RSA functionality (this is what NO_RSA does). This
8942 is so (for example) those in the US can disable those operations covered
8943 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8947 *) Non-copying interface to BIO pairs.
8948 (still largely untested)
8951 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8952 ASCII string. This was handled independently in various places before.
8955 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8956 UTF8 strings a character at a time.
8959 *) Use client_version from client hello to select the protocol
8960 (s23_srvr.c) and for RSA client key exchange verification
8961 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8964 *) Add various utility functions to handle SPKACs, these were previously
8965 handled by poking round in the structure internals. Added new function
8966 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8967 print, verify and generate SPKACs. Based on an original idea from
8968 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8971 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8974 *) Allow the config file extension section to be overwritten on the
8975 command line. Based on an original idea from Massimiliano Pala
8976 <madwolf@comune.modena.it>. The new option is called -extensions
8977 and can be applied to ca, req and x509. Also -reqexts to override
8978 the request extensions in req and -crlexts to override the crl extensions
8982 *) Add new feature to the SPKAC handling in ca. Now you can include
8983 the same field multiple times by preceding it by "XXXX." for example:
8986 this is the same syntax as used in the req config file.
8989 *) Allow certificate extensions to be added to certificate requests. These
8990 are specified in a 'req_extensions' option of the req section of the
8991 config file. They can be printed out with the -text option to req but
8992 are otherwise ignored at present.
8995 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8996 data read consists of only the final block it would not decrypted because
8997 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8998 A misplaced 'break' also meant the decrypted final block might not be
8999 copied until the next read.
9002 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9003 a few extra parameters to the DH structure: these will be useful if
9004 for example we want the value of 'q' or implement X9.42 DH.
9007 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9008 provides hooks that allow the default DSA functions or functions on a
9009 "per key" basis to be replaced. This allows hardware acceleration and
9010 hardware key storage to be handled without major modification to the
9011 library. Also added low level modexp hooks and CRYPTO_EX structure and
9012 associated functions.
9015 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9016 as "read only": it can't be written to and the buffer it points to will
9017 not be freed. Reading from a read only BIO is much more efficient than
9018 a normal memory BIO. This was added because there are several times when
9019 an area of memory needs to be read from a BIO. The previous method was
9020 to create a memory BIO and write the data to it, this results in two
9021 copies of the data and an O(n^2) reading algorithm. There is a new
9022 function BIO_new_mem_buf() which creates a read only memory BIO from
9023 an area of memory. Also modified the PKCS#7 routines to use read only
9027 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9028 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9029 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9030 but a retry condition occured while trying to read the rest.
9033 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9034 NID_pkcs7_encrypted by default: this was wrong since this should almost
9035 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9036 the encrypted data type: this is a more sensible place to put it and it
9037 allows the PKCS#12 code to be tidied up that duplicated this
9041 *) Changed obj_dat.pl script so it takes its input and output files on
9042 the command line. This should avoid shell escape redirection problems
9046 *) Initial support for certificate extension requests, these are included
9047 in things like Xenroll certificate requests. Included functions to allow
9048 extensions to be obtained and added.
9051 *) -crlf option to s_client and s_server for sending newlines as
9052 CRLF (as required by many protocols).
9055 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9057 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9058 [Ralf S. Engelschall]
9060 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9061 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9063 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9067 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9068 DH parameters/keys (q is lost during that conversion, but the resulting
9069 DH parameters contain its length).
9071 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9072 much faster than DH_generate_parameters (which creates parameters
9073 where p = 2*q + 1), and also the smaller q makes DH computations
9074 much more efficient (160-bit exponentiation instead of 1024-bit
9075 exponentiation); so this provides a convenient way to support DHE
9076 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9077 utter importance to use
9078 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9080 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9081 when such DH parameters are used, because otherwise small subgroup
9082 attacks may become possible!
9085 *) Avoid memory leak in i2d_DHparams.
9088 *) Allow the -k option to be used more than once in the enc program:
9089 this allows the same encrypted message to be read by multiple recipients.
9092 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9093 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9094 it will always use the numerical form of the OID, even if it has a short
9098 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9099 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9100 otherwise bn_mod_exp was called. In the case of hardware keys for example
9101 no private key components need be present and it might store extra data
9102 in the RSA structure, which cannot be accessed from bn_mod_exp.
9103 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9104 private key operations.
9107 *) Added support for SPARC Linux.
9110 *) pem_password_cb function type incompatibly changed from
9111 typedef int pem_password_cb(char *buf, int size, int rwflag);
9113 ....(char *buf, int size, int rwflag, void *userdata);
9114 so that applications can pass data to their callbacks:
9115 The PEM[_ASN1]_{read,write}... functions and macros now take an
9116 additional void * argument, which is just handed through whenever
9117 the password callback is called.
9118 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9120 New function SSL_CTX_set_default_passwd_cb_userdata.
9122 Compatibility note: As many C implementations push function arguments
9123 onto the stack in reverse order, the new library version is likely to
9124 interoperate with programs that have been compiled with the old
9125 pem_password_cb definition (PEM_whatever takes some data that
9126 happens to be on the stack as its last argument, and the callback
9127 just ignores this garbage); but there is no guarantee whatsoever that
9130 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9131 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9132 problems not only on Windows, but also on some Unix platforms.
9133 To avoid problematic command lines, these definitions are now in an
9134 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9135 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9138 *) MIPS III/IV assembler module is reimplemented.
9141 *) More DES library cleanups: remove references to srand/rand and
9142 delete an unused file.
9145 *) Add support for the the free Netwide assembler (NASM) under Win32,
9146 since not many people have MASM (ml) and it can be hard to obtain.
9147 This is currently experimental but it seems to work OK and pass all
9148 the tests. Check out INSTALL.W32 for info.
9151 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9152 without temporary keys kept an extra copy of the server key,
9153 and connections with temporary keys did not free everything in case
9157 *) New function RSA_check_key and new openssl rsa option -check
9158 for verifying the consistency of RSA keys.
9159 [Ulf Moeller, Bodo Moeller]
9161 *) Various changes to make Win32 compile work:
9162 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9163 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9164 comparison" warnings.
9165 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9168 *) Add a debugging option to PKCS#5 v2 key generation function: when
9169 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9170 derived keys are printed to stderr.
9173 *) Copy the flags in ASN1_STRING_dup().
9174 [Roman E. Pavlov <pre@mo.msk.ru>]
9176 *) The x509 application mishandled signing requests containing DSA
9177 keys when the signing key was also DSA and the parameters didn't match.
9179 It was supposed to omit the parameters when they matched the signing key:
9180 the verifying software was then supposed to automatically use the CA's
9181 parameters if they were absent from the end user certificate.
9183 Omitting parameters is no longer recommended. The test was also
9184 the wrong way round! This was probably due to unusual behaviour in
9185 EVP_cmp_parameters() which returns 1 if the parameters match.
9186 This meant that parameters were omitted when they *didn't* match and
9187 the certificate was useless. Certificates signed with 'ca' didn't have
9189 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9191 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9192 The interface is as follows:
9193 Applications can use
9194 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9195 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9196 "off" is now the default.
9197 The library internally uses
9198 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9199 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9200 to disable memory-checking temporarily.
9202 Some inconsistent states that previously were possible (and were
9203 even the default) are now avoided.
9205 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9206 with each memory chunk allocated; this is occasionally more helpful
9207 than just having a counter.
9209 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9211 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9215 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9216 which largely parallels "options", but is for changing API behaviour,
9217 whereas "options" are about protocol behaviour.
9218 Initial "mode" flags are:
9220 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9221 a single record has been written.
9222 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9223 retries use the same buffer location.
9224 (But all of the contents must be
9228 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9231 *) Fix problems with no-hmac etc.
9232 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9234 *) New functions RSA_get_default_method(), RSA_set_method() and
9235 RSA_get_method(). These allows replacement of RSA_METHODs without having
9236 to mess around with the internals of an RSA structure.
9239 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9240 Also really enable memory leak checks in openssl.c and in some
9242 [Chad C. Mulligan, Bodo Moeller]
9244 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9245 up the length of negative integers. This has now been simplified to just
9246 store the length when it is first determined and use it later, rather
9247 than trying to keep track of where data is copied and updating it to
9249 [Steve Henson, reported by Brien Wheeler
9250 <bwheeler@authentica-security.com>]
9252 *) Add a new function PKCS7_signatureVerify. This allows the verification
9253 of a PKCS#7 signature but with the signing certificate passed to the
9254 function itself. This contrasts with PKCS7_dataVerify which assumes the
9255 certificate is present in the PKCS#7 structure. This isn't always the
9256 case: certificates can be omitted from a PKCS#7 structure and be
9257 distributed by "out of band" means (such as a certificate database).
9260 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9261 function prototypes in pem.h, also change util/mkdef.pl to add the
9262 necessary function names.
9265 *) mk1mf.pl (used by Windows builds) did not properly read the
9266 options set by Configure in the top level Makefile, and Configure
9267 was not even able to write more than one option correctly.
9268 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9271 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9272 file to be loaded from a BIO or FILE pointer. The BIO version will
9273 for example allow memory BIOs to contain config info.
9276 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9277 Whoever hopes to achieve shared-library compatibility across versions
9278 must use this, not the compile-time macro.
9279 (Exercise 0.9.4: Which is the minimum library version required by
9281 Note: All this applies only to multi-threaded programs, others don't
9285 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9286 through a BIO pair triggered the default case, i.e.
9287 SSLerr(...,SSL_R_UNKNOWN_STATE).
9290 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9291 can use the SSL library even if none of the specific BIOs is
9295 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9296 for the encoded length.
9297 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9299 *) Add initial documentation of the X509V3 functions.
9302 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9303 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9304 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9305 secure PKCS#8 private key format with a high iteration count.
9308 *) Fix determination of Perl interpreter: A perl or perl5
9309 _directory_ in $PATH was also accepted as the interpreter.
9310 [Ralf S. Engelschall]
9312 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9313 wrong with it but it was very old and did things like calling
9314 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9318 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9319 to use the new extension code.
9322 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9323 with macros. This should make it easier to change their form, add extra
9324 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9328 *) Add to configuration table a new entry that can specify an alternative
9329 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9330 according to Mark Crispin <MRC@Panda.COM>.
9334 *) DES CBC did not update the IV. Weird.
9337 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9338 Changing the behaviour of the former might break existing programs --
9339 where IV updating is needed, des_ncbc_encrypt can be used.
9342 *) When bntest is run from "make test" it drives bc to check its
9343 calculations, as well as internally checking them. If an internal check
9344 fails, it needs to cause bc to give a non-zero result or make test carries
9345 on without noticing the failure. Fixed.
9348 *) DES library cleanups.
9351 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9352 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9353 ciphers. NOTE: although the key derivation function has been verified
9354 against some published test vectors it has not been extensively tested
9355 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9359 *) Instead of "mkdir -p", which is not fully portable, use new
9360 Perl script "util/mkdir-p.pl".
9363 *) Rewrite the way password based encryption (PBE) is handled. It used to
9364 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9365 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9366 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9367 the 'parameter' field of the AlgorithmIdentifier is passed to the
9368 underlying key generation function so it must do its own ASN1 parsing.
9369 This has also changed the EVP_PBE_CipherInit() function which now has a
9370 'parameter' argument instead of literal salt and iteration count values
9371 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9374 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9375 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9376 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9377 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9378 value was just used as a "magic string" and not used directly its
9379 value doesn't matter.
9382 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9386 *) "linux-sparc64" configuration (ultrapenguin).
9387 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9388 "linux-sparc" configuration.
9389 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9391 *) config now generates no-xxx options for missing ciphers.
9394 *) Support the EBCDIC character set (work in progress).
9395 File ebcdic.c not yet included because it has a different license.
9396 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9398 *) Support BS2000/OSD-POSIX.
9399 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9401 *) Make callbacks for key generation use void * instead of char *.
9404 *) Make S/MIME samples compile (not yet tested).
9407 *) Additional typesafe stacks.
9410 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9414 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9416 *) New configuration variant "sco5-gcc".
9418 *) Updated some demos.
9419 [Sean O Riordain, Wade Scholine]
9421 *) Add missing BIO_free at exit of pkcs12 application.
9424 *) Fix memory leak in conf.c.
9427 *) Updates for Win32 to assembler version of MD5.
9430 *) Set #! path to perl in apps/der_chop to where we found it
9431 instead of using a fixed path.
9434 *) SHA library changes for irix64-mips4-cc.
9437 *) Improvements for VMS support.
9441 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9443 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9444 This also avoids the problems with SC4.2 and unpatched SC5.
9445 [Andy Polyakov <appro@fy.chalmers.se>]
9447 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9448 These are required because of the typesafe stack would otherwise break
9449 existing code. If old code used a structure member which used to be STACK
9450 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9451 sk_num or sk_value it would produce an error because the num, data members
9452 are not present in STACK_OF. Now it just produces a warning. sk_set
9453 replaces the old method of assigning a value to sk_value
9454 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9455 that does this will no longer work (and should use sk_set instead) but
9456 this could be regarded as a "questionable" behaviour anyway.
9459 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9460 correctly handle encrypted S/MIME data.
9463 *) Change type of various DES function arguments from des_cblock
9464 (which means, in function argument declarations, pointer to char)
9465 to des_cblock * (meaning pointer to array with 8 char elements),
9466 which allows the compiler to do more typechecking; it was like
9467 that back in SSLeay, but with lots of ugly casts.
9469 Introduce new type const_des_cblock.
9472 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9473 problems: find RecipientInfo structure that matches recipient certificate
9474 and initialise the ASN1 structures properly based on passed cipher.
9477 *) Belatedly make the BN tests actually check the results.
9480 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9481 to and from BNs: it was completely broken. New compilation option
9482 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9483 key elements as negative integers.
9486 *) Reorganize and speed up MD5.
9487 [Andy Polyakov <appro@fy.chalmers.se>]
9490 [Richard Levitte <richard@levitte.org>]
9492 *) New option -out to asn1parse to allow the parsed structure to be
9493 output to a file. This is most useful when combined with the -strparse
9494 option to examine the output of things like OCTET STRINGS.
9497 *) Make SSL library a little more fool-proof by not requiring any longer
9498 that SSL_set_{accept,connect}_state be called before
9499 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9500 in many applications because usually everything *appeared* to work as
9501 intended anyway -- now it really works as intended).
9504 *) Move openssl.cnf out of lib/.
9507 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9508 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9509 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9510 [Ralf S. Engelschall]
9512 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9513 handle PKCS#7 enveloped data properly.
9514 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9516 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9517 copying pointers. The cert_st handling is changed by this in
9518 various ways (and thus what used to be known as ctx->default_cert
9519 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9520 any longer when s->cert does not give us what we need).
9521 ssl_cert_instantiate becomes obsolete by this change.
9522 As soon as we've got the new code right (possibly it already is?),
9523 we have solved a couple of bugs of the earlier code where s->cert
9524 was used as if it could not have been shared with other SSL structures.
9526 Note that using the SSL API in certain dirty ways now will result
9527 in different behaviour than observed with earlier library versions:
9528 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9529 does not influence s as it used to.
9531 In order to clean up things more thoroughly, inside SSL_SESSION
9532 we don't use CERT any longer, but a new structure SESS_CERT
9533 that holds per-session data (if available); currently, this is
9534 the peer's certificate chain and, for clients, the server's certificate
9535 and temporary key. CERT holds only those values that can have
9536 meaningful defaults in an SSL_CTX.
9539 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9540 from the internal representation. Various PKCS#7 fixes: remove some
9541 evil casts and set the enc_dig_alg field properly based on the signing
9545 *) Allow PKCS#12 password to be set from the command line or the
9546 environment. Let 'ca' get its config file name from the environment
9547 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9551 *) Allow certificate policies extension to use an IA5STRING for the
9552 organization field. This is contrary to the PKIX definition but
9553 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9557 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9558 without disallowing inline assembler and the like for non-pedantic builds.
9561 *) Support Borland C++ builder.
9562 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9567 *) SHA-1 cleanups and performance enhancements.
9568 [Andy Polyakov <appro@fy.chalmers.se>]
9570 *) Sparc v8plus assembler for the bignum library.
9571 [Andy Polyakov <appro@fy.chalmers.se>]
9573 *) Accept any -xxx and +xxx compiler options in Configure.
9576 *) Update HPUX configuration.
9579 *) Add missing sk_<type>_unshift() function to safestack.h
9580 [Ralf S. Engelschall]
9582 *) New function SSL_CTX_use_certificate_chain_file that sets the
9583 "extra_cert"s in addition to the certificate. (This makes sense
9584 only for "PEM" format files, as chains as a whole are not
9588 *) Support verify_depth from the SSL API.
9589 x509_vfy.c had what can be considered an off-by-one-error:
9590 Its depth (which was not part of the external interface)
9591 was actually counting the number of certificates in a chain;
9592 now it really counts the depth.
9595 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9596 instead of X509err, which often resulted in confusing error
9597 messages since the error codes are not globally unique
9598 (e.g. an alleged error in ssl3_accept when a certificate
9599 didn't match the private key).
9601 *) New function SSL_CTX_set_session_id_context that allows to set a default
9602 value (so that you don't need SSL_set_session_id_context for each
9603 connection using the SSL_CTX).
9606 *) OAEP decoding bug fix.
9609 *) Support INSTALL_PREFIX for package builders, as proposed by
9613 *) New Configure options "threads" and "no-threads". For systems
9614 where the proper compiler options are known (currently Solaris
9615 and Linux), "threads" is the default.
9618 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9621 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9622 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9623 such as /usr/local/bin.
9626 *) "make linux-shared" to build shared libraries.
9627 [Niels Poppe <niels@netbox.org>]
9629 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9632 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9633 extension adding in x509 utility.
9636 *) Remove NOPROTO sections and error code comments.
9639 *) Partial rewrite of the DEF file generator to now parse the ANSI
9643 *) New Configure options --prefix=DIR and --openssldir=DIR.
9646 *) Complete rewrite of the error code script(s). It is all now handled
9647 by one script at the top level which handles error code gathering,
9648 header rewriting and C source file generation. It should be much better
9649 than the old method: it now uses a modified version of Ulf's parser to
9650 read the ANSI prototypes in all header files (thus the old K&R definitions
9651 aren't needed for error creation any more) and do a better job of
9652 translating function codes into names. The old 'ASN1 error code imbedded
9653 in a comment' is no longer necessary and it doesn't use .err files which
9654 have now been deleted. Also the error code call doesn't have to appear all
9655 on one line (which resulted in some large lines...).
9658 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9661 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9662 0 (which usually indicates a closed connection), but continue reading.
9665 *) Fix some race conditions.
9668 *) Add support for CRL distribution points extension. Add Certificate
9669 Policies and CRL distribution points documentation.
9672 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9675 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9676 8 of keying material. Merlin has also confirmed interop with this fix
9677 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9678 [Merlin Hughes <merlin@baltimore.ie>]
9680 *) Fix lots of warnings.
9681 [Richard Levitte <levitte@stacken.kth.se>]
9683 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9684 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9685 [Richard Levitte <levitte@stacken.kth.se>]
9687 *) Fix problems with sizeof(long) == 8.
9688 [Andy Polyakov <appro@fy.chalmers.se>]
9690 *) Change functions to ANSI C.
9693 *) Fix typos in error codes.
9694 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9696 *) Remove defunct assembler files from Configure.
9699 *) SPARC v8 assembler BIGNUM implementation.
9700 [Andy Polyakov <appro@fy.chalmers.se>]
9702 *) Support for Certificate Policies extension: both print and set.
9703 Various additions to support the r2i method this uses.
9706 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9707 return a const string when you are expecting an allocated buffer.
9710 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9711 types DirectoryString and DisplayText.
9714 *) Add code to allow r2i extensions to access the configuration database,
9715 add an LHASH database driver and add several ctx helper functions.
9718 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9719 fail when they extended the size of a BIGNUM.
9722 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9723 support typesafe stack.
9726 *) Fix typo in SSL_[gs]et_options().
9727 [Nils Frostberg <nils@medcom.se>]
9729 *) Delete various functions and files that belonged to the (now obsolete)
9730 old X509V3 handling code.
9733 *) New Configure option "rsaref".
9736 *) Don't auto-generate pem.h.
9739 *) Introduce type-safe ASN.1 SETs.
9742 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9743 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9745 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9746 that links with OpenSSL (well at least cause lots of warnings), but fear
9747 not: the conversion is trivial, and it eliminates loads of evil casts. A
9748 few STACKed things have been converted already. Feel free to convert more.
9749 In the fullness of time, I'll do away with the STACK type altogether.
9752 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9753 specified in <certfile> by updating the entry in the index.txt file.
9754 This way one no longer has to edit the index.txt file manually for
9755 revoking a certificate. The -revoke option does the gory details now.
9756 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9758 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9759 `-text' option at all and this way the `-noout -text' combination was
9760 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9761 [Ralf S. Engelschall]
9763 *) Make sure a corresponding plain text error message exists for the
9764 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9765 verify callback function determined that a certificate was revoked.
9766 [Ralf S. Engelschall]
9768 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9769 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9770 all available cipers including rc5, which was forgotten until now.
9771 In order to let the testing shell script know which algorithms
9772 are available, a new (up to now undocumented) command
9773 "openssl list-cipher-commands" is used.
9776 *) Bugfix: s_client occasionally would sleep in select() when
9777 it should have checked SSL_pending() first.
9780 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9781 the raw DSA values prior to ASN.1 encoding.
9784 *) Tweaks to Configure
9785 [Niels Poppe <niels@netbox.org>]
9787 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9791 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9794 *) New config option to avoid instructions that are illegal on the 80386.
9795 The default code is faster, but requires at least a 486.
9798 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9799 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9800 same as SSL2_VERSION anyway.
9803 *) New "-showcerts" option for s_client.
9806 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9807 application. Various cleanups and fixes.
9810 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9811 modify error routines to work internally. Add error codes and PBE init
9812 to library startup routines.
9815 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9816 packing functions to asn1 and evp. Changed function names and error
9817 codes along the way.
9820 *) PKCS12 integration: and so it begins... First of several patches to
9821 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9822 objects to objects.h
9825 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9826 and display support for Thawte strong extranet extension.
9829 *) Add LinuxPPC support.
9830 [Jeff Dubrule <igor@pobox.org>]
9832 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9833 bn_div_words in alpha.s.
9834 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9836 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9837 OAEP isn't supported when OpenSSL is built with RSAref.
9838 [Ulf Moeller <ulf@fitug.de>]
9840 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9841 so they no longer are missing under -DNOPROTO.
9842 [Soren S. Jorvang <soren@t.dk>]
9845 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9847 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9848 doesn't work when the session is reused. Coming soon!
9851 *) Fix a security hole, that allows sessions to be reused in the wrong
9852 context thus bypassing client cert protection! All software that uses
9853 client certs and session caches in multiple contexts NEEDS PATCHING to
9854 allow session reuse! A fuller solution is in the works.
9855 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9857 *) Some more source tree cleanups (removed obsolete files
9858 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9859 permission on "config" script to be executable) and a fix for the INSTALL
9861 [Ulf Moeller <ulf@fitug.de>]
9863 *) Remove some legacy and erroneous uses of malloc, free instead of
9865 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9867 *) Make rsa_oaep_test return non-zero on error.
9868 [Ulf Moeller <ulf@fitug.de>]
9870 *) Add support for native Solaris shared libraries. Configure
9871 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9872 if someone would make that last step automatic.
9873 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9875 *) ctx_size was not built with the right compiler during "make links". Fixed.
9878 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9879 except NULL ciphers". This means the default cipher list will no longer
9880 enable NULL ciphers. They need to be specifically enabled e.g. with
9881 the string "DEFAULT:eNULL".
9884 *) Fix to RSA private encryption routines: if p < q then it would
9885 occasionally produce an invalid result. This will only happen with
9886 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9889 *) Be less restrictive and allow also `perl util/perlpath.pl
9890 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9891 because this way one can also use an interpreter named `perl5' (which is
9892 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9893 installed as `perl').
9894 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9896 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9897 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9899 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9900 advapi32.lib to Win32 build and change the pem test comparision
9901 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9902 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9903 and crypto/des/ede_cbcm_enc.c.
9906 *) DES quad checksum was broken on big-endian architectures. Fixed.
9909 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9910 Win32 test batch file so it (might) work again. The Win32 test batch file
9911 is horrible: I feel ill....
9914 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9915 in e_os.h. Audit of header files to check ANSI and non ANSI
9916 sections: 10 functions were absent from non ANSI section and not exported
9917 from Windows DLLs. Fixed up libeay.num for new functions.
9920 *) Make `openssl version' output lines consistent.
9921 [Ralf S. Engelschall]
9923 *) Fix Win32 symbol export lists for BIO functions: Added
9924 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9925 to ms/libeay{16,32}.def.
9926 [Ralf S. Engelschall]
9928 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9929 fine under Unix and passes some trivial tests I've now added. But the
9930 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9931 added to make sure no one expects that this stuff really works in the
9932 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9933 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9935 [Ralf S. Engelschall]
9937 *) Fix the generation of two part addresses in perl.
9938 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9940 *) Add config entry for Linux on MIPS.
9941 [John Tobey <jtobey@channel1.com>]
9943 *) Make links whenever Configure is run, unless we are on Windoze.
9946 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9947 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9951 *) Add a useful kludge to allow package maintainers to specify compiler and
9952 other platforms details on the command line without having to patch the
9953 Configure script everytime: One now can use ``perl Configure
9954 <id>:<details>'', i.e. platform ids are allowed to have details appended
9955 to them (seperated by colons). This is treated as there would be a static
9956 pre-configured entry in Configure's %table under key <id> with value
9957 <details> and ``perl Configure <id>'' is called. So, when you want to
9958 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9959 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9960 now, which overrides the FreeBSD-elf entry on-the-fly.
9961 [Ralf S. Engelschall]
9963 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9966 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9967 on the `perl Configure ...' command line. This way one can compile
9968 OpenSSL libraries with Position Independent Code (PIC) which is needed
9969 for linking it into DSOs.
9970 [Ralf S. Engelschall]
9972 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9976 *) Cleaned up the LICENSE document: The official contact for any license
9977 questions now is the OpenSSL core team under openssl-core@openssl.org.
9978 And add a paragraph about the dual-license situation to make sure people
9979 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9980 to the OpenSSL toolkit.
9981 [Ralf S. Engelschall]
9983 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9984 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9985 Additonally cleaned up the `make links' target: Remove unnecessary
9986 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9987 to speed processing and no longer clutter the display with confusing
9988 stuff. Instead only the actually done links are displayed.
9989 [Ralf S. Engelschall]
9991 *) Permit null encryption ciphersuites, used for authentication only. It used
9992 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9993 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9997 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9998 signed attributes when verifying signatures (this would break them),
9999 the detached data encoding was wrong and public keys obtained using
10000 X509_get_pubkey() weren't freed.
10003 *) Add text documentation for the BUFFER functions. Also added a work around
10004 to a Win95 console bug. This was triggered by the password read stuff: the
10005 last character typed gets carried over to the next fread(). If you were
10006 generating a new cert request using 'req' for example then the last
10007 character of the passphrase would be CR which would then enter the first
10011 *) Added the new `Includes OpenSSL Cryptography Software' button as
10012 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10013 button and can be used by applications based on OpenSSL to show the
10014 relationship to the OpenSSL project.
10015 [Ralf S. Engelschall]
10017 *) Remove confusing variables in function signatures in files
10018 ssl/ssl_lib.c and ssl/ssl.h.
10019 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10021 *) Don't install bss_file.c under PREFIX/include/
10022 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10024 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10025 functions that return function pointers and has support for NT specific
10026 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10027 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10028 unsigned to signed types: this was killing the Win32 compile.
10031 *) Add new certificate file to stack functions,
10032 SSL_add_dir_cert_subjects_to_stack() and
10033 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10034 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10035 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10036 This means that Apache-SSL and similar packages don't have to mess around
10037 to add as many CAs as they want to the preferred list.
10040 *) Experiment with doxygen documentation. Currently only partially applied to
10042 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10043 openssl.doxy as the configuration file.
10046 *) Get rid of remaining C++-style comments which strict C compilers hate.
10047 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10049 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10050 compiled in by default: it has problems with large keys.
10053 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10054 DH private keys and/or callback functions which directly correspond to
10055 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10056 is needed for applications which have to configure certificates on a
10057 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10059 For the RSA certificate situation is makes no difference, but
10060 for the DSA certificate situation this fixes the "no shared cipher"
10061 problem where the OpenSSL cipher selection procedure failed because the
10062 temporary keys were not overtaken from the context and the API provided
10063 no way to reconfigure them.
10064 The new functions now let applications reconfigure the stuff and they
10065 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10066 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10067 non-public-API function ssl_cert_instantiate() is used as a helper
10068 function and also to reduce code redundancy inside ssl_rsa.c.
10069 [Ralf S. Engelschall]
10071 *) Move s_server -dcert and -dkey options out of the undocumented feature
10072 area because they are useful for the DSA situation and should be
10073 recognized by the users.
10074 [Ralf S. Engelschall]
10076 *) Fix the cipher decision scheme for export ciphers: the export bits are
10077 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10078 SSL_EXP_MASK. So, the original variable has to be used instead of the
10079 already masked variable.
10080 [Richard Levitte <levitte@stacken.kth.se>]
10082 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10083 [Richard Levitte <levitte@stacken.kth.se>]
10085 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10086 from `int' to `unsigned int' because it's a length and initialized by
10087 EVP_DigestFinal() which expects an `unsigned int *'.
10088 [Richard Levitte <levitte@stacken.kth.se>]
10090 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10091 script. Instead use the usual Shell->Perl transition trick.
10092 [Ralf S. Engelschall]
10094 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10095 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10096 -noout -modulus' as it's already the case for `openssl rsa -noout
10097 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10098 currently the public key is printed (a decision which was already done by
10099 `openssl dsa -modulus' in the past) which serves a similar purpose.
10100 Additionally the NO_RSA no longer completely removes the whole -modulus
10101 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10103 [Ralf S. Engelschall]
10105 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10106 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10107 [Arne Ansper <arne@ats.cyber.ee>]
10109 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10110 to be added. Now both 'req' and 'ca' can use new objects defined in the
10114 *) Add cool BIO that does syslog (or event log on NT).
10115 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10117 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10118 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10119 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10120 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10123 *) Add preliminary config info for new extension code.
10126 *) Make RSA_NO_PADDING really use no padding.
10127 [Ulf Moeller <ulf@fitug.de>]
10129 *) Generate errors when private/public key check is done.
10132 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10133 for some CRL extensions and new objects added.
10136 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10137 key usage extension and fuller support for authority key id.
10140 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10141 padding method for RSA, which is recommended for new applications in PKCS
10142 #1 v2.0 (RFC 2437, October 1998).
10143 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10144 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10145 against Bleichbacher's attack on RSA.
10146 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10149 *) Updates to the new SSL compression code
10150 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10152 *) Fix so that the version number in the master secret, when passed
10153 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10154 (because the server will not accept higher), that the version number
10155 is 0x03,0x01, not 0x03,0x00
10156 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10158 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10159 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10160 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10163 *) Support for RAW extensions where an arbitrary extension can be
10164 created by including its DER encoding. See apps/openssl.cnf for
10168 *) Make sure latest Perl versions don't interpret some generated C array
10169 code as Perl array code in the crypto/err/err_genc.pl script.
10170 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10172 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10173 not many people have the assembler. Various Win32 compilation fixes and
10174 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10175 build instructions.
10178 *) Modify configure script 'Configure' to automatically create crypto/date.h
10179 file under Win32 and also build pem.h from pem.org. New script
10180 util/mkfiles.pl to create the MINFO file on environments that can't do a
10181 'make files': perl util/mkfiles.pl >MINFO should work.
10184 *) Major rework of DES function declarations, in the pursuit of correctness
10185 and purity. As a result, many evil casts evaporated, and some weirdness,
10186 too. You may find this causes warnings in your code. Zapping your evil
10187 casts will probably fix them. Mostly.
10190 *) Fix for a typo in asn1.h. Bug fix to object creation script
10191 obj_dat.pl. It considered a zero in an object definition to mean
10192 "end of object": none of the objects in objects.h have any zeros
10193 so it wasn't spotted.
10194 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10196 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10197 Masking (CBCM). In the absence of test vectors, the best I have been able
10198 to do is check that the decrypt undoes the encrypt, so far. Send me test
10199 vectors if you have them.
10202 *) Correct calculation of key length for export ciphers (too much space was
10203 allocated for null ciphers). This has not been tested!
10206 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10207 message is now correct (it understands "crypto" and "ssl" on its
10208 command line). There is also now an "update" option. This will update
10209 the util/ssleay.num and util/libeay.num files with any new functions.
10211 perl util/mkdef.pl crypto ssl update
10212 it will update them.
10215 *) Overhauled the Perl interface (perl/*):
10216 - ported BN stuff to OpenSSL's different BN library
10217 - made the perl/ source tree CVS-aware
10218 - renamed the package from SSLeay to OpenSSL (the files still contain
10219 their history because I've copied them in the repository)
10220 - removed obsolete files (the test scripts will be replaced
10221 by better Test::Harness variants in the future)
10222 [Ralf S. Engelschall]
10224 *) First cut for a very conservative source tree cleanup:
10225 1. merge various obsolete readme texts into doc/ssleay.txt
10226 where we collect the old documents and readme texts.
10227 2. remove the first part of files where I'm already sure that we no
10228 longer need them because of three reasons: either they are just temporary
10229 files which were left by Eric or they are preserved original files where
10230 I've verified that the diff is also available in the CVS via "cvs diff
10231 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10232 the crypto/md/ stuff).
10233 [Ralf S. Engelschall]
10235 *) More extension code. Incomplete support for subject and issuer alt
10236 name, issuer and authority key id. Change the i2v function parameters
10237 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10238 what that's for :-) Fix to ASN1 macro which messed up
10239 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10242 *) Preliminary support for ENUMERATED type. This is largely copied from the
10246 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10247 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10249 *) Make sure `make rehash' target really finds the `openssl' program.
10250 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10252 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10253 like to hear about it if this slows down other processors.
10256 *) Add CygWin32 platform information to Configure script.
10257 [Alan Batie <batie@aahz.jf.intel.com>]
10259 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10260 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10262 *) New program nseq to manipulate netscape certificate sequences
10265 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10269 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10270 but the BN code had some problems that would cause failures when
10271 doing certificate verification and some other functions.
10272 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10274 *) Add ASN1 and PEM code to support netscape certificate sequences.
10277 *) Add ASN1 and PEM code to support netscape certificate sequences.
10280 *) Add several PKIX and private extended key usage OIDs.
10283 *) Modify the 'ca' program to handle the new extension code. Modify
10284 openssl.cnf for new extension format, add comments.
10287 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10288 and add a sample to openssl.cnf so req -x509 now adds appropriate
10292 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10293 error code, add initial support to X509_print() and x509 application.
10296 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10297 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10298 stuff is currently isolated and isn't even compiled yet.
10301 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10302 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10303 Removed the versions check from X509 routines when loading extensions:
10304 this allows certain broken certificates that don't set the version
10305 properly to be processed.
10308 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10309 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10310 can still be regenerated with "make depend".
10313 *) Spelling mistake in C version of CAST-128.
10314 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10316 *) Changes to the error generation code. The perl script err-code.pl
10317 now reads in the old error codes and retains the old numbers, only
10318 adding new ones if necessary. It also only changes the .err files if new
10319 codes are added. The makefiles have been modified to only insert errors
10320 when needed (to avoid needlessly modifying header files). This is done
10321 by only inserting errors if the .err file is newer than the auto generated
10322 C file. To rebuild all the error codes from scratch (the old behaviour)
10323 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10324 or delete all the .err files.
10327 *) CAST-128 was incorrectly implemented for short keys. The C version has
10328 been fixed, but is untested. The assembler versions are also fixed, but
10329 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10330 to regenerate it if needed.
10331 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10332 Hagino <itojun@kame.net>]
10334 *) File was opened incorrectly in randfile.c.
10335 [Ulf Möller <ulf@fitug.de>]
10337 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10338 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10339 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10340 al: it's just almost always a UTCTime. Note this patch adds new error
10341 codes so do a "make errors" if there are problems.
10344 *) Correct Linux 1 recognition in config.
10345 [Ulf Möller <ulf@fitug.de>]
10347 *) Remove pointless MD5 hash when using DSA keys in ca.
10348 [Anonymous <nobody@replay.com>]
10350 *) Generate an error if given an empty string as a cert directory. Also
10351 generate an error if handed NULL (previously returned 0 to indicate an
10352 error, but didn't set one).
10353 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10355 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10358 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10359 parameters. This was causing a warning which killed off the Win32 compile.
10362 *) Remove C++ style comments from crypto/bn/bn_local.h.
10363 [Neil Costigan <neil.costigan@celocom.com>]
10365 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10366 based on a text string, looking up short and long names and finally
10367 "dot" format. The "dot" format stuff didn't work. Added new function
10368 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10369 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10370 OID is not part of the table.
10373 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10374 X509_LOOKUP_by_alias().
10377 *) Sort openssl functions by name.
10380 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10381 encryption from sample DSA keys (in case anyone is interested the password
10385 *) Make _all_ *_free functions accept a NULL pointer.
10386 [Frans Heymans <fheymans@isaserver.be>]
10388 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10390 [Anonymous <nobody@replay.com>]
10392 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10393 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10395 *) Don't blow it for numeric -newkey arguments to apps/req.
10396 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10398 *) Temp key "for export" tests were wrong in s3_srvr.c.
10399 [Anonymous <nobody@replay.com>]
10401 *) Add prototype for temp key callback functions
10402 SSL_CTX_set_tmp_{rsa,dh}_callback().
10405 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10406 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10409 *) X509_name_add_entry() freed the wrong thing after an error.
10410 [Arne Ansper <arne@ats.cyber.ee>]
10412 *) rsa_eay.c would attempt to free a NULL context.
10413 [Arne Ansper <arne@ats.cyber.ee>]
10415 *) BIO_s_socket() had a broken should_retry() on Windoze.
10416 [Arne Ansper <arne@ats.cyber.ee>]
10418 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10419 [Arne Ansper <arne@ats.cyber.ee>]
10421 *) Make sure the already existing X509_STORE->depth variable is initialized
10422 in X509_STORE_new(), but document the fact that this variable is still
10423 unused in the certificate verification process.
10424 [Ralf S. Engelschall]
10426 *) Fix the various library and apps files to free up pkeys obtained from
10427 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10430 *) Fix reference counting in X509_PUBKEY_get(). This makes
10431 demos/maurice/example2.c work, amongst others, probably.
10432 [Steve Henson and Ben Laurie]
10434 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10435 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10436 are no longer created. This way we have a single and consistent command
10437 line interface `openssl <command>', similar to `cvs <command>'.
10438 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10440 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10441 BIT STRING wrapper always have zero unused bits.
10444 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10447 *) Make the top-level INSTALL documentation easier to understand.
10450 *) Makefiles updated to exit if an error occurs in a sub-directory
10451 make (including if user presses ^C) [Paul Sutton]
10453 *) Make Montgomery context stuff explicit in RSA data structure.
10456 *) Fix build order of pem and err to allow for generated pem.h.
10459 *) Fix renumbering bug in X509_NAME_delete_entry().
10462 *) Enhanced the err-ins.pl script so it makes the error library number
10463 global and can add a library name. This is needed for external ASN1 and
10464 other error libraries.
10467 *) Fixed sk_insert which never worked properly.
10470 *) Fix ASN1 macros so they can handle indefinite length construted
10471 EXPLICIT tags. Some non standard certificates use these: they can now
10475 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10476 into a single doc/ssleay.txt bundle. This way the information is still
10477 preserved but no longer messes up this directory. Now it's new room for
10478 the new set of documenation files.
10479 [Ralf S. Engelschall]
10481 *) SETs were incorrectly DER encoded. This was a major pain, because they
10482 shared code with SEQUENCEs, which aren't coded the same. This means that
10483 almost everything to do with SETs or SEQUENCEs has either changed name or
10484 number of arguments.
10485 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10487 *) Fix test data to work with the above.
10490 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10491 was already fixed by Eric for 0.9.1 it seems.
10492 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10494 *) Autodetect FreeBSD3.
10497 *) Fix various bugs in Configure. This affects the following platforms:
10501 unixware-2.0-pentium
10505 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10506 before they are needed.
10509 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10513 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10515 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10516 changed SSLeay to OpenSSL in version strings.
10517 [Ralf S. Engelschall]
10519 *) Some fixups to the top-level documents.
10522 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10523 because the symlink to include/ was missing.
10524 [Ralf S. Engelschall]
10526 *) Incorporated the popular no-RSA/DSA-only patches
10527 which allow to compile a RSA-free SSLeay.
10528 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10530 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10531 when "ssleay" is still not found.
10532 [Ralf S. Engelschall]
10534 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10535 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10537 *) Updated the README file.
10538 [Ralf S. Engelschall]
10540 *) Added various .cvsignore files in the CVS repository subdirs
10541 to make a "cvs update" really silent.
10542 [Ralf S. Engelschall]
10544 *) Recompiled the error-definition header files and added
10545 missing symbols to the Win32 linker tables.
10546 [Ralf S. Engelschall]
10548 *) Cleaned up the top-level documents;
10549 o new files: CHANGES and LICENSE
10550 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10551 o merged COPYRIGHT into LICENSE
10552 o removed obsolete TODO file
10553 o renamed MICROSOFT to INSTALL.W32
10554 [Ralf S. Engelschall]
10556 *) Removed dummy files from the 0.9.1b source tree:
10557 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10558 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10559 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10560 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10561 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10562 [Ralf S. Engelschall]
10564 *) Added various platform portability fixes.
10567 *) The Genesis of the OpenSSL rpject:
10568 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10569 Young and Tim J. Hudson created while they were working for C2Net until
10571 [The OpenSSL Project]
10574 Changes between 0.9.0b and 0.9.1b [not released]
10576 *) Updated a few CA certificates under certs/
10579 *) Changed some BIGNUM api stuff.
10582 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10583 DGUX x86, Linux Alpha, etc.
10586 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10587 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10591 *) Add -strparse option to asn1pars program which parses nested
10593 [Dr Stephen Henson <shenson@bigfoot.com>]
10595 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10598 *) DSA fix for "ca" program.
10601 *) Added "-genkey" option to "dsaparam" program.
10604 *) Added RIPE MD160 (rmd160) message digest.
10607 *) Added -a (all) option to "ssleay version" command.
10610 *) Added PLATFORM define which is the id given to Configure.
10613 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10616 *) Extended the ASN.1 parser routines.
10619 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10622 *) Added a BN_CTX to the BN library.
10625 *) Fixed the weak key values in DES library
10628 *) Changed API in EVP library for cipher aliases.
10631 *) Added support for RC2/64bit cipher.
10634 *) Converted the lhash library to the crypto/mem.c functions.
10637 *) Added more recognized ASN.1 object ids.
10640 *) Added more RSA padding checks for SSL/TLS.
10643 *) Added BIO proxy/filter functionality.
10646 *) Added extra_certs to SSL_CTX which can be used
10647 send extra CA certificates to the client in the CA cert chain sending
10648 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10651 *) Now Fortezza is denied in the authentication phase because
10652 this is key exchange mechanism is not supported by SSLeay at all.
10655 *) Additional PKCS1 checks.
10658 *) Support the string "TLSv1" for all TLS v1 ciphers.
10661 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10662 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10665 *) Fixed a few memory leaks.
10668 *) Fixed various code and comment typos.
10671 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10672 bytes sent in the client random.
10673 [Edward Bishop <ebishop@spyglass.com>]