5 Changes between 1.0.1f and 1.0.1g [xx XXX xxxx]
9 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
11 *) Fix for TLS record tampering bug. A carefully crafted invalid
12 handshake could crash OpenSSL with a NULL pointer exception.
13 Thanks to Anton Johansson for reporting this issues.
16 *) Keep original DTLS digest and encryption contexts in retransmission
17 structures so we can use the previous session parameters if they need
18 to be resent. (CVE-2013-6450)
21 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
22 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
23 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
24 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
25 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
26 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
27 [Rob Stradling, Adam Langley]
29 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
31 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
32 supporting platforms or when small records were transferred.
33 [Andy Polyakov, Steve Henson]
35 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
37 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
39 This addresses the flaw in CBC record processing discovered by
40 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
41 at: http://www.isg.rhul.ac.uk/tls/
43 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
44 Security Group at Royal Holloway, University of London
45 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
46 Emilia Käsper for the initial patch.
48 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
50 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
51 ciphersuites which can be exploited in a denial of service attack.
52 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
53 and detecting this bug and to Wolfgang Ettlinger
54 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
58 *) Return an error when checking OCSP signatures when key is NULL.
59 This fixes a DoS attack. (CVE-2013-0166)
62 *) Make openssl verify return errors.
63 [Chris Palmer <palmer@google.com> and Ben Laurie]
65 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
66 the right response is stapled. Also change SSL_get_certificate()
67 so it returns the certificate actually sent.
68 See http://rt.openssl.org/Ticket/Display.html?id=2836.
69 [Rob Stradling <rob.stradling@comodo.com>]
71 *) Fix possible deadlock when decoding public keys.
74 *) Don't use TLS 1.0 record version number in initial client hello
78 Changes between 1.0.1b and 1.0.1c [10 May 2012]
80 *) Sanity check record length before skipping explicit IV in TLS
81 1.2, 1.1 and DTLS to fix DoS attack.
83 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
84 fuzzing as a service testing platform.
88 *) Initialise tkeylen properly when encrypting CMS messages.
89 Thanks to Solar Designer of Openwall for reporting this issue.
92 *) In FIPS mode don't try to use composite ciphers as they are not
96 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
98 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
99 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
100 mean any application compiled against OpenSSL 1.0.0 headers setting
101 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
102 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
103 0x10000000L Any application which was previously compiled against
104 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
105 will need to be recompiled as a result. Letting be results in
106 inability to disable specifically TLS 1.1 and in client context,
107 in unlike event, limit maximum offered version to TLS 1.0 [see below].
110 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
111 disable just protocol X, but all protocols above X *if* there are
112 protocols *below* X still enabled. In more practical terms it means
113 that if application wants to disable TLS1.0 in favor of TLS1.1 and
114 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
115 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
119 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
121 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
122 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
123 in CRYPTO_realloc_clean.
125 Thanks to Tavis Ormandy, Google Security Team, for discovering this
126 issue and to Adam Langley <agl@chromium.org> for fixing it.
128 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
130 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
133 *) Workarounds for some broken servers that "hang" if a client hello
134 record length exceeds 255 bytes.
136 1. Do not use record version number > TLS 1.0 in initial client
137 hello: some (but not all) hanging servers will now work.
138 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
139 the number of ciphers sent in the client hello. This should be
140 set to an even number, such as 50, for example by passing:
141 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
142 Most broken servers should now work.
143 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
144 TLS 1.2 client support entirely.
147 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
150 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
152 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
153 STRING form instead of a DigestInfo.
156 *) The format used for MDC2 RSA signatures is inconsistent between EVP
157 and the RSA_sign/RSA_verify functions. This was made more apparent when
158 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
159 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
160 the correct format in RSA_verify so both forms transparently work.
163 *) Some servers which support TLS 1.0 can choke if we initially indicate
164 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
165 encrypted premaster secret. As a workaround use the maximum pemitted
166 client version in client hello, this should keep such servers happy
167 and still work with previous versions of OpenSSL.
170 *) Add support for TLS/DTLS heartbeats.
171 [Robin Seggelmann <seggelmann@fh-muenster.de>]
173 *) Add support for SCTP.
174 [Robin Seggelmann <seggelmann@fh-muenster.de>]
176 *) Improved PRNG seeding for VOS.
177 [Paul Green <Paul.Green@stratus.com>]
179 *) Extensive assembler packs updates, most notably:
181 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
182 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
183 - x86_64: bit-sliced AES implementation;
184 - ARM: NEON support, contemporary platforms optimizations;
185 - s390x: z196 support;
186 - *: GHASH and GF(2^m) multiplication implementations;
190 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
191 (removal of unnecessary code)
192 [Peter Sylvester <peter.sylvester@edelweb.fr>]
194 *) Add TLS key material exporter from RFC 5705.
197 *) Add DTLS-SRTP negotiation from RFC 5764.
200 *) Add Next Protocol Negotiation,
201 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
202 disabled with a no-npn flag to config or Configure. Code donated
204 [Adam Langley <agl@google.com> and Ben Laurie]
206 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
207 NIST-P256, NIST-P521, with constant-time single point multiplication on
208 typical inputs. Compiler support for the nonstandard type __uint128_t is
209 required to use this (present in gcc 4.4 and later, for 64-bit builds).
210 Code made available under Apache License version 2.0.
212 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
213 line to include this in your build of OpenSSL, and run "make depend" (or
214 "make update"). This enables the following EC_METHODs:
216 EC_GFp_nistp224_method()
217 EC_GFp_nistp256_method()
218 EC_GFp_nistp521_method()
220 EC_GROUP_new_by_curve_name() will automatically use these (while
221 EC_GROUP_new_curve_GFp() currently prefers the more flexible
223 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
225 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
226 all platforms. Move ssize_t definition from e_os.h to the public
227 header file e_os2.h as it now appears in public header file cms.h
230 *) New -sigopt option to the ca, req and x509 utilities. Additional
231 signature parameters can be passed using this option and in
235 *) Add RSA PSS signing function. This will generate and set the
236 appropriate AlgorithmIdentifiers for PSS based on those in the
237 corresponding EVP_MD_CTX structure. No application support yet.
240 *) Support for companion algorithm specific ASN1 signing routines.
241 New function ASN1_item_sign_ctx() signs a pre-initialised
242 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
243 the appropriate parameters.
246 *) Add new algorithm specific ASN1 verification initialisation function
247 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
248 handling will be the same no matter what EVP_PKEY_METHOD is used.
249 Add a PSS handler to support verification of PSS signatures: checked
250 against a number of sample certificates.
253 *) Add signature printing for PSS. Add PSS OIDs.
254 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
256 *) Add algorithm specific signature printing. An individual ASN1 method
257 can now print out signatures instead of the standard hex dump.
259 More complex signatures (e.g. PSS) can print out more meaningful
260 information. Include DSA version that prints out the signature
264 *) Password based recipient info support for CMS library: implementing
268 *) Split password based encryption into PBES2 and PBKDF2 functions. This
269 neatly separates the code into cipher and PBE sections and is required
270 for some algorithms that split PBES2 into separate pieces (such as
274 *) Session-handling fixes:
275 - Fix handling of connections that are resuming with a session ID,
276 but also support Session Tickets.
277 - Fix a bug that suppressed issuing of a new ticket if the client
278 presented a ticket with an expired session.
279 - Try to set the ticket lifetime hint to something reasonable.
280 - Make tickets shorter by excluding irrelevant information.
281 - On the client side, don't ignore renewed tickets.
282 [Adam Langley, Bodo Moeller (Google)]
284 *) Fix PSK session representation.
287 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
289 This work was sponsored by Intel.
292 *) Add GCM support to TLS library. Some custom code is needed to split
293 the IV between the fixed (from PRF) and explicit (from TLS record)
294 portions. This adds all GCM ciphersuites supported by RFC5288 and
295 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
296 add a special AESGCM string for GCM only.
299 *) Expand range of ctrls for AES GCM. Permit setting invocation
300 field on decrypt and retrieval of invocation field only on encrypt.
303 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
304 As required by RFC5289 these ciphersuites cannot be used if for
305 versions of TLS earlier than 1.2.
308 *) For FIPS capable OpenSSL interpret a NULL default public key method
309 as unset and return the appopriate default but do *not* set the default.
310 This means we can return the appopriate method in applications that
311 swicth between FIPS and non-FIPS modes.
314 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
315 ENGINE is used then we cannot handle that in the FIPS module so we
316 keep original code iff non-FIPS operations are allowed.
319 *) Add -attime option to openssl utilities.
320 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
322 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
325 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
326 FIPS EC methods unconditionally for now.
329 *) New build option no-ec2m to disable characteristic 2 code.
332 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
333 all cases can be covered as some introduce binary incompatibilities.
336 *) Redirect RSA operations to FIPS module including keygen,
337 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
340 *) Add similar low level API blocking to ciphers.
343 *) Low level digest APIs are not approved in FIPS mode: any attempt
344 to use these will cause a fatal error. Applications that *really* want
345 to use them can use the private_* version instead.
348 *) Redirect cipher operations to FIPS module for FIPS builds.
351 *) Redirect digest operations to FIPS module for FIPS builds.
354 *) Update build system to add "fips" flag which will link in fipscanister.o
355 for static and shared library builds embedding a signature if needed.
358 *) Output TLS supported curves in preference order instead of numerical
359 order. This is currently hardcoded for the highest order curves first.
360 This should be configurable so applications can judge speed vs strength.
363 *) Add TLS v1.2 server support for client authentication.
366 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
370 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
371 FIPS modules versions.
374 *) Add TLS v1.2 client side support for client authentication. Keep cache
375 of handshake records longer as we don't know the hash algorithm to use
376 until after the certificate request message is received.
379 *) Initial TLS v1.2 client support. Add a default signature algorithms
380 extension including all the algorithms we support. Parse new signature
381 format in client key exchange. Relax some ECC signing restrictions for
382 TLS v1.2 as indicated in RFC5246.
385 *) Add server support for TLS v1.2 signature algorithms extension. Switch
386 to new signature format when needed using client digest preference.
387 All server ciphersuites should now work correctly in TLS v1.2. No client
388 support yet and no support for client certificates.
391 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
392 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
393 ciphersuites. At present only RSA key exchange ciphersuites work with
394 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
395 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
396 and version checking.
399 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
400 with this defined it will not be affected by any changes to ssl internal
401 structures. Add several utility functions to allow openssl application
402 to work with OPENSSL_NO_SSL_INTERN defined.
406 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
408 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
411 *) Permit abbreviated handshakes when renegotiating using the function
412 SSL_renegotiate_abbreviated().
413 [Robin Seggelmann <seggelmann@fh-muenster.de>]
415 *) Add call to ENGINE_register_all_complete() to
416 ENGINE_load_builtin_engines(), so some implementations get used
417 automatically instead of needing explicit application support.
420 *) Add support for TLS key exporter as described in RFC5705.
421 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
423 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
424 a few changes are required:
426 Add SSL_OP_NO_TLSv1_1 flag.
428 Update version checking logic to handle version 1.1.
429 Add explicit IV handling (ported from DTLS code).
430 Add command line options to s_client/s_server.
433 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
435 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
437 This addresses the flaw in CBC record processing discovered by
438 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
439 at: http://www.isg.rhul.ac.uk/tls/
441 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
442 Security Group at Royal Holloway, University of London
443 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
444 Emilia Käsper for the initial patch.
446 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
448 *) Return an error when checking OCSP signatures when key is NULL.
449 This fixes a DoS attack. (CVE-2013-0166)
452 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
453 the right response is stapled. Also change SSL_get_certificate()
454 so it returns the certificate actually sent.
455 See http://rt.openssl.org/Ticket/Display.html?id=2836.
457 [Rob Stradling <rob.stradling@comodo.com>]
459 *) Fix possible deadlock when decoding public keys.
462 Changes between 1.0.0i and 1.0.0j [10 May 2012]
464 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
467 *) Sanity check record length before skipping explicit IV in DTLS
470 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
471 fuzzing as a service testing platform.
475 *) Initialise tkeylen properly when encrypting CMS messages.
476 Thanks to Solar Designer of Openwall for reporting this issue.
479 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
481 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
482 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
483 in CRYPTO_realloc_clean.
485 Thanks to Tavis Ormandy, Google Security Team, for discovering this
486 issue and to Adam Langley <agl@chromium.org> for fixing it.
488 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
490 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
492 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
493 in CMS and PKCS7 code. When RSA decryption fails use a random key for
494 content decryption and always return the same error. Note: this attack
495 needs on average 2^20 messages so it only affects automated senders. The
496 old behaviour can be reenabled in the CMS code by setting the
497 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
498 an MMA defence is not necessary.
499 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
500 this issue. (CVE-2012-0884)
503 *) Fix CVE-2011-4619: make sure we really are receiving a
504 client hello before rejecting multiple SGC restarts. Thanks to
505 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
508 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
510 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
511 Thanks to Antonio Martin, Enterprise Secure Access Research and
512 Development, Cisco Systems, Inc. for discovering this bug and
513 preparing a fix. (CVE-2012-0050)
516 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
518 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
519 of the Vaudenay padding oracle attack on CBC mode encryption
520 which enables an efficient plaintext recovery attack against
521 the OpenSSL implementation of DTLS. Their attack exploits timing
522 differences arising during decryption processing. A research
523 paper describing this attack can be found at:
524 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
525 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
526 Security Group at Royal Holloway, University of London
527 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
528 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
529 for preparing the fix. (CVE-2011-4108)
530 [Robin Seggelmann, Michael Tuexen]
532 *) Clear bytes used for block padding of SSL 3.0 records.
534 [Adam Langley (Google)]
536 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
537 Kadianakis <desnacked@gmail.com> for discovering this issue and
538 Adam Langley for preparing the fix. (CVE-2011-4619)
539 [Adam Langley (Google)]
541 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
542 [Andrey Kulikov <amdeich@gmail.com>]
544 *) Prevent malformed RFC3779 data triggering an assertion failure.
545 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
546 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
547 [Rob Austein <sra@hactrn.net>]
549 *) Improved PRNG seeding for VOS.
550 [Paul Green <Paul.Green@stratus.com>]
552 *) Fix ssl_ciph.c set-up race.
553 [Adam Langley (Google)]
555 *) Fix spurious failures in ecdsatest.c.
556 [Emilia Käsper (Google)]
558 *) Fix the BIO_f_buffer() implementation (which was mixing different
559 interpretations of the '..._len' fields).
560 [Adam Langley (Google)]
562 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
563 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
564 threads won't reuse the same blinding coefficients.
566 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
567 lock to call BN_BLINDING_invert_ex, and avoids one use of
568 BN_BLINDING_update for each BN_BLINDING structure (previously,
569 the last update always remained unused).
570 [Emilia Käsper (Google)]
572 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
573 [Bob Buckholz (Google)]
575 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
577 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
578 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
579 [Kaspar Brand <ossl@velox.ch>]
581 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
582 for multi-threaded use of ECDH. (CVE-2011-3210)
583 [Adam Langley (Google)]
585 *) Fix x509_name_ex_d2i memory leak on bad inputs.
588 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
589 signature public key algorithm by using OID xref utilities instead.
590 Before this you could only use some ECC ciphersuites with SHA1 only.
593 *) Add protection against ECDSA timing attacks as mentioned in the paper
594 by Billy Bob Brumley and Nicola Tuveri, see:
596 http://eprint.iacr.org/2011/232.pdf
598 [Billy Bob Brumley and Nicola Tuveri]
600 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
602 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
603 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
605 *) Fix bug in string printing code: if *any* escaping is enabled we must
606 escape the escape character (backslash) or the resulting string is
610 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
612 *) Disable code workaround for ancient and obsolete Netscape browsers
613 and servers: an attacker can use it in a ciphersuite downgrade attack.
614 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
617 *) Fixed J-PAKE implementation error, originally discovered by
618 Sebastien Martini, further info and confirmation from Stefan
619 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
622 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
624 *) Fix extension code to avoid race conditions which can result in a buffer
625 overrun vulnerability: resumed sessions must not be modified as they can
626 be shared by multiple threads. CVE-2010-3864
629 *) Fix WIN32 build system to correctly link an ENGINE directory into
633 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
635 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
637 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
639 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
641 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
642 context. The operation can be customised via the ctrl mechanism in
643 case ENGINEs want to include additional functionality.
646 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
649 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
650 output hashes compatible with older versions of OpenSSL.
651 [Willy Weisz <weisz@vcpc.univie.ac.at>]
653 *) Fix compression algorithm handling: if resuming a session use the
654 compression algorithm of the resumed session instead of determining
655 it from client hello again. Don't allow server to change algorithm.
658 *) Add load_crls() function to apps tidying load_certs() too. Add option
659 to verify utility to allow additional CRLs to be included.
662 *) Update OCSP request code to permit adding custom headers to the request:
663 some responders need this.
666 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
668 [Julia Lawall <julia@diku.dk>]
670 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
671 needlessly dereferenced structures, used obsolete functions and
672 didn't handle all updated verify codes correctly.
675 *) Disable MD2 in the default configuration.
678 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
679 indicate the initial BIO being pushed or popped. This makes it possible
680 to determine whether the BIO is the one explicitly called or as a result
681 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
682 it handles reference counts correctly and doesn't zero out the I/O bio
683 when it is not being explicitly popped. WARNING: applications which
684 included workarounds for the old buggy behaviour will need to be modified
685 or they could free up already freed BIOs.
688 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
689 renaming to all platforms (within the 0.9.8 branch, this was
690 done conditionally on Netware platforms to avoid a name clash).
691 [Guenter <lists@gknw.net>]
693 *) Add ECDHE and PSK support to DTLS.
694 [Michael Tuexen <tuexen@fh-muenster.de>]
696 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
700 *) Add "missing" function EVP_MD_flags() (without this the only way to
701 retrieve a digest flags is by accessing the structure directly. Update
702 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
703 or cipher is registered as in the "from" argument. Print out all
704 registered digests in the dgst usage message instead of manually
705 attempting to work them out.
708 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
709 this allows the use of compression and extensions. Change default cipher
710 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
711 by default unless an application cipher string requests it.
714 *) Alter match criteria in PKCS12_parse(). It used to try to use local
715 key ids to find matching certificates and keys but some PKCS#12 files
716 don't follow the (somewhat unwritten) rules and this strategy fails.
717 Now just gather all certificates together and the first private key
718 then look for the first certificate that matches the key.
721 *) Support use of registered digest and cipher names for dgst and cipher
722 commands instead of having to add each one as a special case. So now
729 openssl dgst -sha256 foo
731 and this works for ENGINE based algorithms too.
735 *) Update Gost ENGINE to support parameter files.
736 [Victor B. Wagner <vitus@cryptocom.ru>]
738 *) Support GeneralizedTime in ca utility.
739 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
741 *) Enhance the hash format used for certificate directory links. The new
742 form uses the canonical encoding (meaning equivalent names will work
743 even if they aren't identical) and uses SHA1 instead of MD5. This form
744 is incompatible with the older format and as a result c_rehash should
745 be used to rebuild symbolic links.
748 *) Make PKCS#8 the default write format for private keys, replacing the
749 traditional format. This form is standardised, more secure and doesn't
750 include an implicit MD5 dependency.
753 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
754 committed to OpenSSL should pass this lot as a minimum.
757 *) Add session ticket override functionality for use by EAP-FAST.
758 [Jouni Malinen <j@w1.fi>]
760 *) Modify HMAC functions to return a value. Since these can be implemented
761 in an ENGINE errors can occur.
764 *) Type-checked OBJ_bsearch_ex.
767 *) Type-checked OBJ_bsearch. Also some constification necessitated
768 by type-checking. Still to come: TXT_DB, bsearch(?),
769 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
773 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
774 seconds to a tm structure directly, instead of going through OS
775 specific date routines. This avoids any issues with OS routines such
776 as the year 2038 bug. New *_adj() functions for ASN1 time structures
777 and X509_time_adj_ex() to cover the extended range. The existing
778 X509_time_adj() is still usable and will no longer have any date issues.
781 *) Delta CRL support. New use deltas option which will attempt to locate
782 and search any appropriate delta CRLs available.
784 This work was sponsored by Google.
787 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
788 code and add additional score elements. Validate alternate CRL paths
789 as part of the CRL checking and indicate a new error "CRL path validation
790 error" in this case. Applications wanting additional details can use
791 the verify callback and check the new "parent" field. If this is not
792 NULL CRL path validation is taking place. Existing applications wont
793 see this because it requires extended CRL support which is off by
796 This work was sponsored by Google.
799 *) Support for freshest CRL extension.
801 This work was sponsored by Google.
804 *) Initial indirect CRL support. Currently only supported in the CRLs
805 passed directly and not via lookup. Process certificate issuer
806 CRL entry extension and lookup CRL entries by bother issuer name
807 and serial number. Check and process CRL issuer entry in IDP extension.
809 This work was sponsored by Google.
812 *) Add support for distinct certificate and CRL paths. The CRL issuer
813 certificate is validated separately in this case. Only enabled if
814 an extended CRL support flag is set: this flag will enable additional
815 CRL functionality in future.
817 This work was sponsored by Google.
820 *) Add support for policy mappings extension.
822 This work was sponsored by Google.
825 *) Fixes to pathlength constraint, self issued certificate handling,
826 policy processing to align with RFC3280 and PKITS tests.
828 This work was sponsored by Google.
831 *) Support for name constraints certificate extension. DN, email, DNS
832 and URI types are currently supported.
834 This work was sponsored by Google.
837 *) To cater for systems that provide a pointer-based thread ID rather
838 than numeric, deprecate the current numeric thread ID mechanism and
839 replace it with a structure and associated callback type. This
840 mechanism allows a numeric "hash" to be extracted from a thread ID in
841 either case, and on platforms where pointers are larger than 'long',
842 mixing is done to help ensure the numeric 'hash' is usable even if it
843 can't be guaranteed unique. The default mechanism is to use "&errno"
844 as a pointer-based thread ID to distinguish between threads.
846 Applications that want to provide their own thread IDs should now use
847 CRYPTO_THREADID_set_callback() to register a callback that will call
848 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
850 Note that ERR_remove_state() is now deprecated, because it is tied
851 to the assumption that thread IDs are numeric. ERR_remove_state(0)
852 to free the current thread's error state should be replaced by
853 ERR_remove_thread_state(NULL).
855 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
856 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
857 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
858 application was previously providing a numeric thread callback that
859 was inappropriate for distinguishing threads, then uniqueness might
860 have been obtained with &errno that happened immediately in the
861 intermediate development versions of OpenSSL; this is no longer the
862 case, the numeric thread callback will now override the automatic use
864 [Geoff Thorpe, with help from Bodo Moeller]
866 *) Initial support for different CRL issuing certificates. This covers a
867 simple case where the self issued certificates in the chain exist and
868 the real CRL issuer is higher in the existing chain.
870 This work was sponsored by Google.
873 *) Removed effectively defunct crypto/store from the build.
876 *) Revamp of STACK to provide stronger type-checking. Still to come:
877 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
878 ASN1_STRING, CONF_VALUE.
881 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
882 RAM on SSL connections. This option can save about 34k per idle SSL.
885 *) Revamp of LHASH to provide stronger type-checking. Still to come:
886 STACK, TXT_DB, bsearch, qsort.
889 *) Initial support for Cryptographic Message Syntax (aka CMS) based
890 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
891 support for data, signedData, compressedData, digestedData and
892 encryptedData, envelopedData types included. Scripts to check against
893 RFC4134 examples draft and interop and consistency checks of many
894 content types and variants.
897 *) Add options to enc utility to support use of zlib compression BIO.
900 *) Extend mk1mf to support importing of options and assembly language
901 files from Configure script, currently only included in VC-WIN32.
902 The assembly language rules can now optionally generate the source
903 files from the associated perl scripts.
906 *) Implement remaining functionality needed to support GOST ciphersuites.
907 Interop testing has been performed using CryptoPro implementations.
908 [Victor B. Wagner <vitus@cryptocom.ru>]
910 *) s390x assembler pack.
913 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
917 *) Implement Opaque PRF Input TLS extension as specified in
918 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
919 official specification yet and no extension type assignment by
920 IANA exists, this extension (for now) will have to be explicitly
921 enabled when building OpenSSL by providing the extension number
922 to use. For example, specify an option
924 -DTLSEXT_TYPE_opaque_prf_input=0x9527
926 to the "config" or "Configure" script to enable the extension,
927 assuming extension number 0x9527 (which is a completely arbitrary
928 and unofficial assignment based on the MD5 hash of the Internet
929 Draft). Note that by doing so, you potentially lose
930 interoperability with other TLS implementations since these might
931 be using the same extension number for other purposes.
933 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
934 opaque PRF input value to use in the handshake. This will create
935 an interal copy of the length-'len' string at 'src', and will
936 return non-zero for success.
938 To get more control and flexibility, provide a callback function
941 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
942 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
946 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
949 Callback function 'cb' will be called in handshakes, and is
950 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
951 Argument 'arg' is for application purposes (the value as given to
952 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
953 be provided to the callback function). The callback function
954 has to return non-zero to report success: usually 1 to use opaque
955 PRF input just if possible, or 2 to enforce use of the opaque PRF
956 input. In the latter case, the library will abort the handshake
957 if opaque PRF input is not successfully negotiated.
959 Arguments 'peerinput' and 'len' given to the callback function
960 will always be NULL and 0 in the case of a client. A server will
961 see the client's opaque PRF input through these variables if
962 available (NULL and 0 otherwise). Note that if the server
963 provides an opaque PRF input, the length must be the same as the
964 length of the client's opaque PRF input.
966 Note that the callback function will only be called when creating
967 a new session (session resumption can resume whatever was
968 previously negotiated), and will not be called in SSL 2.0
969 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
970 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
971 for applications that need to enforce opaque PRF input.
975 *) Update ssl code to support digests other than SHA1+MD5 for handshake
978 [Victor B. Wagner <vitus@cryptocom.ru>]
980 *) Add RFC4507 support to OpenSSL. This includes the corrections in
981 RFC4507bis. The encrypted ticket format is an encrypted encoded
982 SSL_SESSION structure, that way new session features are automatically
985 If a client application caches session in an SSL_SESSION structure
986 support is transparent because tickets are now stored in the encoded
989 The SSL_CTX structure automatically generates keys for ticket
990 protection in servers so again support should be possible
991 with no application modification.
993 If a client or server wishes to disable RFC4507 support then the option
994 SSL_OP_NO_TICKET can be set.
996 Add a TLS extension debugging callback to allow the contents of any client
997 or server extensions to be examined.
999 This work was sponsored by Google.
1002 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1003 OpenSSL should now compile cleanly on gcc 4.2
1004 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1006 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1007 support including streaming MAC support: this is required for GOST
1008 ciphersuite support.
1009 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1011 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1012 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1013 to output in BER and PEM format.
1016 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1017 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1018 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1019 ENGINE support for HMAC keys which are unextractable. New -mac and
1020 -macopt options to dgst utility.
1023 *) New option -sigopt to dgst utility. Update dgst to use
1024 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1025 alternative signing paramaters such as X9.31 or PSS in the dgst
1029 *) Change ssl_cipher_apply_rule(), the internal function that does
1030 the work each time a ciphersuite string requests enabling
1031 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1032 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1033 the order of disabled ciphersuites such that those ciphersuites
1034 that most recently went from enabled to disabled not only stay
1035 in order with respect to each other, but also have higher priority
1036 than other disabled ciphersuites the next time ciphersuites are
1039 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1040 the same ciphersuites as with "HIGH" alone, but in a specific
1041 order where the PSK ciphersuites come first (since they are the
1042 most recently disabled ciphersuites when "HIGH" is parsed).
1044 Also, change ssl_create_cipher_list() (using this new
1045 funcionality) such that between otherwise identical
1046 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1050 *) Change ssl_create_cipher_list() so that it automatically
1051 arranges the ciphersuites in reasonable order before starting
1052 to process the rule string. Thus, the definition for "DEFAULT"
1053 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1054 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1055 This makes it much easier to arrive at a reasonable default order
1056 in applications for which anonymous ciphers are OK (meaning
1057 that you can't actually use DEFAULT).
1058 [Bodo Moeller; suggested by Victor Duchovni]
1060 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1061 processing) into multiple integers instead of setting
1062 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1063 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1064 (These masks as well as the individual bit definitions are hidden
1065 away into the non-exported interface ssl/ssl_locl.h, so this
1066 change to the definition of the SSL_CIPHER structure shouldn't
1067 affect applications.) This give us more bits for each of these
1068 categories, so there is no longer a need to coagulate AES128 and
1069 AES256 into a single algorithm bit, and to coagulate Camellia128
1070 and Camellia256 into a single algorithm bit, which has led to all
1073 Thus, among other things, the kludge introduced in 0.9.7m and
1074 0.9.8e for masking out AES256 independently of AES128 or masking
1075 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1077 With the change, we also introduce new ciphersuite aliases that
1078 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1082 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1083 Use the leftmost N bytes of the signature input if the input is
1084 larger than the prime q (with N being the size in bytes of q).
1087 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1088 it yet and it is largely untested.
1091 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1094 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1095 some compilers (gcc 4.2 and later) reject their use. Safestack is
1096 reimplemented. Update ASN1 to avoid use of legacy functions.
1099 *) Win32/64 targets are linked with Winsock2.
1102 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1103 to external functions. This can be used to increase CRL handling
1104 efficiency especially when CRLs are very large by (for example) storing
1105 the CRL revoked certificates in a database.
1108 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1109 new CRLs added to a directory can be used. New command line option
1110 -verify_return_error to s_client and s_server. This causes real errors
1111 to be returned by the verify callback instead of carrying on no matter
1112 what. This reflects the way a "real world" verify callback would behave.
1115 *) GOST engine, supporting several GOST algorithms and public key formats.
1116 Kindly donated by Cryptocom.
1119 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1120 partitioned by DP are handled but no indirect CRL or reason partitioning
1121 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1122 selected via a scoring technique which handles IDP and AKID in CRLs.
1125 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1126 will ultimately be used for all verify operations: this will remove the
1127 X509_STORE dependency on certificate verification and allow alternative
1128 lookup methods. X509_STORE based implementations of these two callbacks.
1131 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1132 Modify get_crl() to find a valid (unexpired) CRL if possible.
1135 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1136 this would be called X509_CRL_cmp() but that name is already used by
1137 a function that just compares CRL issuer names. Cache several CRL
1138 extensions in X509_CRL structure and cache CRLDP in X509.
1141 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1142 this maps equivalent X509_NAME structures into a consistent structure.
1143 Name comparison can then be performed rapidly using memcmp().
1146 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1150 *) Allow digests to supply their own micalg string for S/MIME type using
1151 the ctrl EVP_MD_CTRL_MICALG.
1154 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1155 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1156 ctrl. It can then customise the structure before and/or after signing
1160 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1161 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1162 to free up any added signature OIDs.
1165 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1166 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1167 digest and cipher tables. New options added to openssl utility:
1168 list-message-digest-algorithms and list-cipher-algorithms.
1171 *) Change the array representation of binary polynomials: the list
1172 of degrees of non-zero coefficients is now terminated with -1.
1173 Previously it was terminated with 0, which was also part of the
1174 value; thus, the array representation was not applicable to
1175 polynomials where t^0 has coefficient zero. This change makes
1176 the array representation useful in a more general context.
1179 *) Various modifications and fixes to SSL/TLS cipher string
1180 handling. For ECC, the code now distinguishes between fixed ECDH
1181 with RSA certificates on the one hand and with ECDSA certificates
1182 on the other hand, since these are separate ciphersuites. The
1183 unused code for Fortezza ciphersuites has been removed.
1185 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1186 (not "ECDHE"). For consistency with the code for DH
1187 certificates, use of ECDH certificates is now considered ECDH
1188 authentication, not RSA or ECDSA authentication (the latter is
1189 merely the CA's signing algorithm and not actively used in the
1192 The temporary ciphersuite alias "ECCdraft" is no longer
1193 available, and ECC ciphersuites are no longer excluded from "ALL"
1194 and "DEFAULT". The following aliases now exist for RFC 4492
1195 ciphersuites, most of these by analogy with the DH case:
1197 kECDHr - ECDH cert, signed with RSA
1198 kECDHe - ECDH cert, signed with ECDSA
1199 kECDH - ECDH cert (signed with either RSA or ECDSA)
1200 kEECDH - ephemeral ECDH
1201 ECDH - ECDH cert or ephemeral ECDH
1207 AECDH - anonymous ECDH
1208 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1212 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1213 Use correct micalg parameters depending on digest(s) in signed message.
1216 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1217 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1220 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1221 an engine to register a method. Add ENGINE lookups for methods and
1222 functional reference processing.
1225 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1226 EVP_{Sign,Verify}* which allow an application to customise the signature
1230 *) New -resign option to smime utility. This adds one or more signers
1231 to an existing PKCS#7 signedData structure. Also -md option to use an
1232 alternative message digest algorithm for signing.
1235 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1236 create PKCS7 structures containing multiple signers. Update smime
1237 application to support multiple signers.
1240 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1244 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1245 Reorganize PBE internals to lookup from a static table using NIDs,
1246 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1247 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1248 PRF which will be automatically used with PBES2.
1251 *) Replace the algorithm specific calls to generate keys in "req" with the
1255 *) Update PKCS#7 enveloped data routines to use new API. This is now
1256 supported by any public key method supporting the encrypt operation. A
1257 ctrl is added to allow the public key algorithm to examine or modify
1258 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1262 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1263 a default digest type to use. In most cases this will be SHA1 but some
1264 algorithms (such as GOST) need to specify an alternative digest. The
1265 return value indicates how strong the prefernce is 1 means optional and
1266 2 is mandatory (that is it is the only supported type). Modify
1267 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1268 use the default md. Update openssl utilities to use the default digest
1269 type for signing if it is not explicitly indicated.
1272 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1273 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1274 signing method from the key type. This effectively removes the link
1275 between digests and public key types.
1278 *) Add an OID cross reference table and utility functions. Its purpose is to
1279 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1280 rsaEncryption. This will allow some of the algorithm specific hackery
1281 needed to use the correct OID to be removed.
1284 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1285 structures for PKCS7_sign(). They are now set up by the relevant public
1289 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1292 *) Add support for key derivation (agreement) in the API, DH method and
1296 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1297 public and private key formats. As a side effect these add additional
1298 command line functionality not previously available: DSA signatures can be
1299 generated and verified using pkeyutl and DH key support and generation in
1304 [Oliver Tappe <zooey@hirschkaefer.de>]
1306 *) New make target "install_html_docs" installs HTML renditions of the
1308 [Oliver Tappe <zooey@hirschkaefer.de>]
1310 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1311 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1312 support key and parameter generation and add initial key generation
1313 functionality for RSA.
1316 *) Add functions for main EVP_PKEY_method operations. The undocumented
1317 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1318 EVP_PKEY_{encrypt,decrypt}_old.
1321 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1322 key API, doesn't do much yet.
1325 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1326 public key algorithms. New option to openssl utility:
1327 "list-public-key-algorithms" to print out info.
1330 *) Implement the Supported Elliptic Curves Extension for
1331 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1334 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1335 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1338 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1339 utilities such as rsa, dsa, dsaparam etc except they process any key
1343 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1344 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1345 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1349 *) Initial support for pluggable public key ASN1.
1350 De-spaghettify the public key ASN1 handling. Move public and private
1351 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1352 algorithm specific handling to a single module within the relevant
1353 algorithm directory. Add functions to allow (near) opaque processing
1354 of public and private key structures.
1357 *) Implement the Supported Point Formats Extension for
1358 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1361 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1362 for the psk identity [hint] and the psk callback functions to the
1363 SSL_SESSION, SSL and SSL_CTX structure.
1366 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1370 SSL_CTX_use_psk_identity_hint
1371 SSL_get_psk_identity_hint
1372 SSL_get_psk_identity
1373 SSL_use_psk_identity_hint
1375 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1377 *) Add RFC 3161 compliant time stamp request creation, response generation
1378 and response verification functionality.
1379 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1381 *) Add initial support for TLS extensions, specifically for the server_name
1382 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1383 have new members for a host name. The SSL data structure has an
1384 additional member SSL_CTX *initial_ctx so that new sessions can be
1385 stored in that context to allow for session resumption, even after the
1386 SSL has been switched to a new SSL_CTX in reaction to a client's
1387 server_name extension.
1389 New functions (subject to change):
1391 SSL_get_servername()
1392 SSL_get_servername_type()
1395 New CTRL codes and macros (subject to change):
1397 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1398 - SSL_CTX_set_tlsext_servername_callback()
1399 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1400 - SSL_CTX_set_tlsext_servername_arg()
1401 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1403 openssl s_client has a new '-servername ...' option.
1405 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1406 '-key2 ...', '-servername_fatal' (subject to change). This allows
1407 testing the HostName extension for a specific single host name ('-cert'
1408 and '-key' remain fallbacks for handshakes without HostName
1409 negotiation). If the unrecogninzed_name alert has to be sent, this by
1410 default is a warning; it becomes fatal with the '-servername_fatal'
1413 [Peter Sylvester, Remy Allais, Christophe Renou]
1415 *) Whirlpool hash implementation is added.
1418 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1419 bn(64,32). Because of instruction set limitations it doesn't have
1420 any negative impact on performance. This was done mostly in order
1421 to make it possible to share assembler modules, such as bn_mul_mont
1422 implementations, between 32- and 64-bit builds without hassle.
1425 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1426 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1430 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1431 dedicated Montgomery multiplication procedure, is introduced.
1432 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1433 "64-bit" performance on certain 32-bit targets.
1436 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1437 in SSL structures. New SSL ctrl to set maximum send fragment size.
1438 Save memory by seeting the I/O buffer sizes dynamically instead of
1439 using the maximum available value.
1442 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1443 in addition to the text details.
1446 *) Very, very preliminary EXPERIMENTAL support for printing of general
1447 ASN1 structures. This currently produces rather ugly output and doesn't
1448 handle several customised structures at all.
1451 *) Integrated support for PVK file format and some related formats such
1452 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1453 these in the 'rsa' and 'dsa' utilities.
1456 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1459 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1460 place for the (very old) "NETSCAPE" format certificates which are now
1461 handled using new ASN1 code equivalents.
1464 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1465 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1466 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1469 *) Modify CRL distribution points extension code to print out previously
1470 unsupported fields. Enhance extension setting code to allow setting of
1474 *) Add print and set support for Issuing Distribution Point CRL extension.
1477 *) Change 'Configure' script to enable Camellia by default.
1480 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1482 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1484 This addresses the flaw in CBC record processing discovered by
1485 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1486 at: http://www.isg.rhul.ac.uk/tls/
1488 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1489 Security Group at Royal Holloway, University of London
1490 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1491 Emilia Käsper for the initial patch.
1493 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1495 *) Return an error when checking OCSP signatures when key is NULL.
1496 This fixes a DoS attack. (CVE-2013-0166)
1499 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1500 the right response is stapled. Also change SSL_get_certificate()
1501 so it returns the certificate actually sent.
1502 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1503 (This is a backport)
1504 [Rob Stradling <rob.stradling@comodo.com>]
1506 *) Fix possible deadlock when decoding public keys.
1509 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1511 *) Sanity check record length before skipping explicit IV in DTLS
1514 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1515 fuzzing as a service testing platform.
1519 *) Initialise tkeylen properly when encrypting CMS messages.
1520 Thanks to Solar Designer of Openwall for reporting this issue.
1523 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1525 *) The fix for CVE-2012-2110 did not take into account that the
1526 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1527 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1528 rejecting negative len parameter. (CVE-2012-2131)
1529 [Tomas Hoger <thoger@redhat.com>]
1531 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1533 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1534 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1535 in CRYPTO_realloc_clean.
1537 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1538 issue and to Adam Langley <agl@chromium.org> for fixing it.
1540 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1542 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1544 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1545 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1546 content decryption and always return the same error. Note: this attack
1547 needs on average 2^20 messages so it only affects automated senders. The
1548 old behaviour can be reenabled in the CMS code by setting the
1549 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1550 an MMA defence is not necessary.
1551 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1552 this issue. (CVE-2012-0884)
1555 *) Fix CVE-2011-4619: make sure we really are receiving a
1556 client hello before rejecting multiple SGC restarts. Thanks to
1557 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1560 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1562 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1563 Thanks to Antonio Martin, Enterprise Secure Access Research and
1564 Development, Cisco Systems, Inc. for discovering this bug and
1565 preparing a fix. (CVE-2012-0050)
1568 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1570 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1571 of the Vaudenay padding oracle attack on CBC mode encryption
1572 which enables an efficient plaintext recovery attack against
1573 the OpenSSL implementation of DTLS. Their attack exploits timing
1574 differences arising during decryption processing. A research
1575 paper describing this attack can be found at:
1576 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1577 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1578 Security Group at Royal Holloway, University of London
1579 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1580 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1581 for preparing the fix. (CVE-2011-4108)
1582 [Robin Seggelmann, Michael Tuexen]
1584 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1585 [Ben Laurie, Kasper <ekasper@google.com>]
1587 *) Clear bytes used for block padding of SSL 3.0 records.
1589 [Adam Langley (Google)]
1591 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1592 Kadianakis <desnacked@gmail.com> for discovering this issue and
1593 Adam Langley for preparing the fix. (CVE-2011-4619)
1594 [Adam Langley (Google)]
1596 *) Prevent malformed RFC3779 data triggering an assertion failure.
1597 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1598 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1599 [Rob Austein <sra@hactrn.net>]
1601 *) Fix ssl_ciph.c set-up race.
1602 [Adam Langley (Google)]
1604 *) Fix spurious failures in ecdsatest.c.
1605 [Emilia Käsper (Google)]
1607 *) Fix the BIO_f_buffer() implementation (which was mixing different
1608 interpretations of the '..._len' fields).
1609 [Adam Langley (Google)]
1611 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1612 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1613 threads won't reuse the same blinding coefficients.
1615 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1616 lock to call BN_BLINDING_invert_ex, and avoids one use of
1617 BN_BLINDING_update for each BN_BLINDING structure (previously,
1618 the last update always remained unused).
1619 [Emilia Käsper (Google)]
1621 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1622 for multi-threaded use of ECDH.
1623 [Adam Langley (Google)]
1625 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1628 *) Add protection against ECDSA timing attacks as mentioned in the paper
1629 by Billy Bob Brumley and Nicola Tuveri, see:
1631 http://eprint.iacr.org/2011/232.pdf
1633 [Billy Bob Brumley and Nicola Tuveri]
1635 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1637 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1638 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1640 *) Fix bug in string printing code: if *any* escaping is enabled we must
1641 escape the escape character (backslash) or the resulting string is
1645 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1647 *) Disable code workaround for ancient and obsolete Netscape browsers
1648 and servers: an attacker can use it in a ciphersuite downgrade attack.
1649 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1652 *) Fixed J-PAKE implementation error, originally discovered by
1653 Sebastien Martini, further info and confirmation from Stefan
1654 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1657 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1659 *) Fix extension code to avoid race conditions which can result in a buffer
1660 overrun vulnerability: resumed sessions must not be modified as they can
1661 be shared by multiple threads. CVE-2010-3864
1664 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1667 *) Don't reencode certificate when calculating signature: cache and use
1668 the original encoding instead. This makes signature verification of
1669 some broken encodings work correctly.
1672 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1673 is also one of the inputs.
1674 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1676 *) Don't repeatedly append PBE algorithms to table if they already exist.
1677 Sort table on each new add. This effectively makes the table read only
1678 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1682 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1684 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1687 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1688 access or freeing data twice (CVE-2010-0742)
1689 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1691 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1692 common in certificates and some applications which only call
1693 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1697 Reduce copying into .apps and .test in makevms.com
1698 Don't try to use blank CA certificate in CA.com
1699 Allow use of C files from original directories in maketests.com
1700 [Steven M. Schweda" <sms@antinode.info>]
1702 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1704 *) When rejecting SSL/TLS records due to an incorrect version number, never
1705 update s->server with a new major version number. As of
1706 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1707 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1708 the previous behavior could result in a read attempt at NULL when
1709 receiving specific incorrect SSL/TLS records once record payload
1710 protection is active. (CVE-2010-0740)
1711 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1713 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1714 could be crashed if the relevant tables were not present (e.g. chrooted).
1715 [Tomas Hoger <thoger@redhat.com>]
1717 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1719 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1720 [Martin Olsson, Neel Mehta]
1722 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1723 accommodate for stack sorting, always a write lock!).
1726 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1727 excessive delays in the RAND_poll(): over a minute. As a workaround
1728 include a time check in the inner Heap32Next loop too.
1731 *) The code that handled flushing of data in SSL/TLS originally used the
1732 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1733 the problem outlined in PR#1949. The fix suggested there however can
1734 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1735 of Apache). So instead simplify the code to flush unconditionally.
1736 This should be fine since flushing with no data to flush is a no op.
1739 *) Handle TLS versions 2.0 and later properly and correctly use the
1740 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1741 off ancient servers have a habit of sticking around for a while...
1744 *) Modify compression code so it frees up structures without using the
1745 ex_data callbacks. This works around a problem where some applications
1746 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1747 restarting) then use compression (e.g. SSL with compression) later.
1748 This results in significant per-connection memory leaks and
1749 has caused some security issues including CVE-2008-1678 and
1753 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1754 change when encrypting or decrypting.
1757 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1758 connect and renegotiate with servers which do not support RI.
1759 Until RI is more widely deployed this option is enabled by default.
1762 *) Add "missing" ssl ctrls to clear options and mode.
1765 *) If client attempts to renegotiate and doesn't support RI respond with
1766 a no_renegotiation alert as required by RFC5746. Some renegotiating
1767 TLS clients will continue a connection gracefully when they receive
1768 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1769 waiting for a server hello which it will never receive. Now we treat a
1770 received no_renegotiation alert as a fatal error. This is because
1771 applications requesting a renegotiation might well expect it to succeed
1772 and would have no code in place to handle the server denying it so the
1773 only safe thing to do is to terminate the connection.
1776 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1777 peer supports secure renegotiation and 0 otherwise. Print out peer
1778 renegotiation support in s_client/s_server.
1781 *) Replace the highly broken and deprecated SPKAC certification method with
1782 the updated NID creation version. This should correctly handle UTF8.
1785 *) Implement RFC5746. Re-enable renegotiation but require the extension
1786 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1787 turns out to be a bad idea. It has been replaced by
1788 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1789 SSL_CTX_set_options(). This is really not recommended unless you
1790 know what you are doing.
1791 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1793 *) Fixes to stateless session resumption handling. Use initial_ctx when
1794 issuing and attempting to decrypt tickets in case it has changed during
1795 servername handling. Use a non-zero length session ID when attempting
1796 stateless session resumption: this makes it possible to determine if
1797 a resumption has occurred immediately after receiving server hello
1798 (several places in OpenSSL subtly assume this) instead of later in
1802 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1803 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1804 fixes for a few places where the return code is not checked
1806 [Julia Lawall <julia@diku.dk>]
1808 *) Add --strict-warnings option to Configure script to include devteam
1809 warnings in other configurations.
1812 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1813 makes it possible to install openssl libraries in locations which
1814 have names other than "lib", for example "/usr/lib64" which some
1816 [Steve Henson, based on patch from Jeremy Utley]
1818 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1819 X690 8.9.12 and can produce some misleading textual output of OIDs.
1820 [Steve Henson, reported by Dan Kaminsky]
1822 *) Delete MD2 from algorithm tables. This follows the recommendation in
1823 several standards that it is not used in new applications due to
1824 several cryptographic weaknesses. For binary compatibility reasons
1825 the MD2 API is still compiled in by default.
1828 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1832 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1833 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1835 [Guenter <lists@gknw.net>]
1837 *) Fix the server certificate chain building code to use X509_verify_cert(),
1838 it used to have an ad-hoc builder which was unable to cope with anything
1839 other than a simple chain.
1840 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1842 *) Don't check self signed certificate signatures in X509_verify_cert()
1843 by default (a flag can override this): it just wastes time without
1844 adding any security. As a useful side effect self signed root CAs
1845 with non-FIPS digests are now usable in FIPS mode.
1848 *) In dtls1_process_out_of_seq_message() the check if the current message
1849 is already buffered was missing. For every new message was memory
1850 allocated, allowing an attacker to perform an denial of service attack
1851 with sending out of seq handshake messages until there is no memory
1852 left. Additionally every future messege was buffered, even if the
1853 sequence number made no sense and would be part of another handshake.
1854 So only messages with sequence numbers less than 10 in advance will be
1855 buffered. (CVE-2009-1378)
1856 [Robin Seggelmann, discovered by Daniel Mentz]
1858 *) Records are buffered if they arrive with a future epoch to be
1859 processed after finishing the corresponding handshake. There is
1860 currently no limitation to this buffer allowing an attacker to perform
1861 a DOS attack with sending records with future epochs until there is no
1862 memory left. This patch adds the pqueue_size() function to detemine
1863 the size of a buffer and limits the record buffer to 100 entries.
1865 [Robin Seggelmann, discovered by Daniel Mentz]
1867 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1868 parent structure is freed. (CVE-2009-1379)
1871 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1872 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1875 [Ilya O. <vrghost@gmail.com>]
1877 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1879 *) Disable renegotiation completely - this fixes a severe security
1880 problem (CVE-2009-3555) at the cost of breaking all
1881 renegotiation. Renegotiation can be re-enabled by setting
1882 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1883 run-time. This is really not recommended unless you know what
1887 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1889 *) Don't set val to NULL when freeing up structures, it is freed up by
1890 underlying code. If sizeof(void *) > sizeof(long) this can result in
1891 zeroing past the valid field. (CVE-2009-0789)
1892 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1894 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1895 checked correctly. This would allow some invalid signed attributes to
1896 appear to verify correctly. (CVE-2009-0591)
1897 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1899 *) Reject UniversalString and BMPString types with invalid lengths. This
1900 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1901 a legal length. (CVE-2009-0590)
1904 *) Set S/MIME signing as the default purpose rather than setting it
1905 unconditionally. This allows applications to override it at the store
1909 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1910 to handle some structures.
1913 *) Improve efficiency of mem_gets: don't search whole buffer each time
1915 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1917 *) New -hex option for openssl rand.
1920 *) Print out UTF8String and NumericString when parsing ASN1.
1923 *) Support NumericString type for name components.
1926 *) Allow CC in the environment to override the automatically chosen
1927 compiler. Note that nothing is done to ensure flags work with the
1931 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1933 *) Properly check EVP_VerifyFinal() and similar return values
1935 [Ben Laurie, Bodo Moeller, Google Security Team]
1937 *) Enable TLS extensions by default.
1940 *) Allow the CHIL engine to be loaded, whether the application is
1941 multithreaded or not. (This does not release the developer from the
1942 obligation to set up the dynamic locking callbacks.)
1943 [Sander Temme <sander@temme.net>]
1945 *) Use correct exit code if there is an error in dgst command.
1946 [Steve Henson; problem pointed out by Roland Dirlewanger]
1948 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1949 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1952 *) Add experimental JPAKE support, including demo authentication in
1953 s_client and s_server.
1956 *) Set the comparison function in v3_addr_canonize().
1957 [Rob Austein <sra@hactrn.net>]
1959 *) Add support for XMPP STARTTLS in s_client.
1960 [Philip Paeps <philip@freebsd.org>]
1962 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1963 to ensure that even with this option, only ciphersuites in the
1964 server's preference list will be accepted. (Note that the option
1965 applies only when resuming a session, so the earlier behavior was
1966 just about the algorithm choice for symmetric cryptography.)
1969 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1971 *) Fix NULL pointer dereference if a DTLS server received
1972 ChangeCipherSpec as first record (CVE-2009-1386).
1975 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1976 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1979 *) The fix in 0.9.8c that supposedly got rid of unsafe
1980 double-checked locking was incomplete for RSA blinding,
1981 addressing just one layer of what turns out to have been
1982 doubly unsafe triple-checked locking.
1984 So now fix this for real by retiring the MONT_HELPER macro
1985 in crypto/rsa/rsa_eay.c.
1987 [Bodo Moeller; problem pointed out by Marius Schilder]
1989 *) Various precautionary measures:
1991 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1993 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1994 (NB: This would require knowledge of the secret session ticket key
1995 to exploit, in which case you'd be SOL either way.)
1997 - Change bn_nist.c so that it will properly handle input BIGNUMs
1998 outside the expected range.
2000 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2003 [Neel Mehta, Bodo Moeller]
2005 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2006 the load fails. Useful for distros.
2007 [Ben Laurie and the FreeBSD team]
2009 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2012 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2015 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2017 This work was sponsored by Logica.
2020 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2021 keystores. Support for SSL/TLS client authentication too.
2022 Not compiled unless enable-capieng specified to Configure.
2024 This work was sponsored by Logica.
2027 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2028 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2029 attribute creation routines such as certifcate requests and PKCS#12
2033 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2035 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2036 handshake which could lead to a cilent crash as found using the
2037 Codenomicon TLS test suite (CVE-2008-1672)
2038 [Steve Henson, Mark Cox]
2040 *) Fix double free in TLS server name extensions which could lead to
2041 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2044 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2046 Clear the error queue to ensure that error entries left from
2047 older function calls do not interfere with the correct operation.
2048 [Lutz Jaenicke, Erik de Castro Lopo]
2050 *) Remove root CA certificates of commercial CAs:
2052 The OpenSSL project does not recommend any specific CA and does not
2053 have any policy with respect to including or excluding any CA.
2054 Therefore it does not make any sense to ship an arbitrary selection
2055 of root CA certificates with the OpenSSL software.
2058 *) RSA OAEP patches to fix two separate invalid memory reads.
2059 The first one involves inputs when 'lzero' is greater than
2060 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2061 before the beginning of from). The second one involves inputs where
2062 the 'db' section contains nothing but zeroes (there is a one-byte
2063 invalid read after the end of 'db').
2064 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2066 *) Partial backport from 0.9.9-dev:
2068 Introduce bn_mul_mont (dedicated Montgomery multiplication
2069 procedure) as a candidate for BIGNUM assembler implementation.
2070 While 0.9.9-dev uses assembler for various architectures, only
2071 x86_64 is available by default here in the 0.9.8 branch, and
2072 32-bit x86 is available through a compile-time setting.
2074 To try the 32-bit x86 assembler implementation, use Configure
2075 option "enable-montasm" (which exists only for this backport).
2077 As "enable-montasm" for 32-bit x86 disclaims code stability
2078 anyway, in this constellation we activate additional code
2079 backported from 0.9.9-dev for further performance improvements,
2080 namely BN_from_montgomery_word. (To enable this otherwise,
2081 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2083 [Andy Polyakov (backport partially by Bodo Moeller)]
2085 *) Add TLS session ticket callback. This allows an application to set
2086 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2087 values. This is useful for key rollover for example where several key
2088 sets may exist with different names.
2091 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2092 This was broken until now in 0.9.8 releases, such that the only way
2093 a registered ENGINE could be used (assuming it initialises
2094 successfully on the host) was to explicitly set it as the default
2095 for the relevant algorithms. This is in contradiction with 0.9.7
2096 behaviour and the documentation. With this fix, when an ENGINE is
2097 registered into a given algorithm's table of implementations, the
2098 'uptodate' flag is reset so that auto-discovery will be used next
2099 time a new context for that algorithm attempts to select an
2101 [Ian Lister (tweaked by Geoff Thorpe)]
2103 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2104 implemention in the following ways:
2106 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2109 Lack of BER streaming support means one pass streaming processing is
2110 only supported if data is detached: setting the streaming flag is
2111 ignored for embedded content.
2113 CMS support is disabled by default and must be explicitly enabled
2114 with the enable-cms configuration option.
2117 *) Update the GMP engine glue to do direct copies between BIGNUM and
2118 mpz_t when openssl and GMP use the same limb size. Otherwise the
2119 existing "conversion via a text string export" trick is still used.
2120 [Paul Sheer <paulsheer@gmail.com>]
2122 *) Zlib compression BIO. This is a filter BIO which compressed and
2123 uncompresses any data passed through it.
2126 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2127 RFC3394 compatible AES key wrapping.
2130 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2131 sets string data without copying. X509_ALGOR_set0() and
2132 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2133 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2134 from an X509_ATTRIBUTE structure optionally checking it occurs only
2135 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2139 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2140 to get the expected BN_FLG_CONSTTIME behavior.
2141 [Bodo Moeller (Google)]
2145 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2146 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2147 - added some more tests to do_tests.pl
2148 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2149 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2150 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2151 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2152 - various changes to netware.pl to enable gcc-cross builds on Win32
2154 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2155 - various changes to fix missing prototype warnings
2156 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2157 - added AES, WHIRLPOOL and CPUID assembler code to build files
2158 - added missing AES assembler make rules to mk1mf.pl
2159 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2160 [Guenter Knauf <eflash@gmx.net>]
2162 *) Implement certificate status request TLS extension defined in RFC3546.
2163 A client can set the appropriate parameters and receive the encoded
2164 OCSP response via a callback. A server can query the supplied parameters
2165 and set the encoded OCSP response in the callback. Add simplified examples
2166 to s_client and s_server.
2169 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2171 *) Fix various bugs:
2172 + Binary incompatibility of ssl_ctx_st structure
2173 + DTLS interoperation with non-compliant servers
2174 + Don't call get_session_cb() without proposed session
2175 + Fix ia64 assembler code
2176 [Andy Polyakov, Steve Henson]
2178 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2180 *) DTLS Handshake overhaul. There were longstanding issues with
2181 OpenSSL DTLS implementation, which were making it impossible for
2182 RFC 4347 compliant client to communicate with OpenSSL server.
2183 Unfortunately just fixing these incompatibilities would "cut off"
2184 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2185 server keeps tolerating non RFC compliant syntax. The opposite is
2186 not true, 0.9.8f client can not communicate with earlier server.
2187 This update even addresses CVE-2007-4995.
2190 *) Changes to avoid need for function casts in OpenSSL: some compilers
2191 (gcc 4.2 and later) reject their use.
2192 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2195 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2196 RFC4507bis. The encrypted ticket format is an encrypted encoded
2197 SSL_SESSION structure, that way new session features are automatically
2200 If a client application caches session in an SSL_SESSION structure
2201 support is transparent because tickets are now stored in the encoded
2204 The SSL_CTX structure automatically generates keys for ticket
2205 protection in servers so again support should be possible
2206 with no application modification.
2208 If a client or server wishes to disable RFC4507 support then the option
2209 SSL_OP_NO_TICKET can be set.
2211 Add a TLS extension debugging callback to allow the contents of any client
2212 or server extensions to be examined.
2214 This work was sponsored by Google.
2217 *) Add initial support for TLS extensions, specifically for the server_name
2218 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2219 have new members for a host name. The SSL data structure has an
2220 additional member SSL_CTX *initial_ctx so that new sessions can be
2221 stored in that context to allow for session resumption, even after the
2222 SSL has been switched to a new SSL_CTX in reaction to a client's
2223 server_name extension.
2225 New functions (subject to change):
2227 SSL_get_servername()
2228 SSL_get_servername_type()
2231 New CTRL codes and macros (subject to change):
2233 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2234 - SSL_CTX_set_tlsext_servername_callback()
2235 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2236 - SSL_CTX_set_tlsext_servername_arg()
2237 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2239 openssl s_client has a new '-servername ...' option.
2241 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2242 '-key2 ...', '-servername_fatal' (subject to change). This allows
2243 testing the HostName extension for a specific single host name ('-cert'
2244 and '-key' remain fallbacks for handshakes without HostName
2245 negotiation). If the unrecogninzed_name alert has to be sent, this by
2246 default is a warning; it becomes fatal with the '-servername_fatal'
2249 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2251 *) Add AES and SSE2 assembly language support to VC++ build.
2254 *) Mitigate attack on final subtraction in Montgomery reduction.
2257 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2258 (which previously caused an internal error).
2261 *) Squeeze another 10% out of IGE mode when in != out.
2264 *) AES IGE mode speedup.
2265 [Dean Gaudet (Google)]
2267 *) Add the Korean symmetric 128-bit cipher SEED (see
2268 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2269 add SEED ciphersuites from RFC 4162:
2271 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2272 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2273 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2274 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2276 To minimize changes between patchlevels in the OpenSSL 0.9.8
2277 series, SEED remains excluded from compilation unless OpenSSL
2278 is configured with 'enable-seed'.
2279 [KISA, Bodo Moeller]
2281 *) Mitigate branch prediction attacks, which can be practical if a
2282 single processor is shared, allowing a spy process to extract
2283 information. For detailed background information, see
2284 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2285 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2286 and Necessary Software Countermeasures"). The core of the change
2287 are new versions BN_div_no_branch() and
2288 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2289 respectively, which are slower, but avoid the security-relevant
2290 conditional branches. These are automatically called by BN_div()
2291 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2292 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2293 remove a conditional branch.
2295 BN_FLG_CONSTTIME is the new name for the previous
2296 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2297 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2298 in the exponent causes BN_mod_exp_mont() to use the alternative
2299 implementation in BN_mod_exp_mont_consttime().) The old name
2300 remains as a deprecated alias.
2302 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2303 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2304 constant-time implementations for more than just exponentiation.
2305 Here too the old name is kept as a deprecated alias.
2307 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2308 the BN_BLINDING structure gets an independent copy of the
2309 modulus. This means that the previous "BIGNUM *m" argument to
2310 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2311 essentially becomes "const BIGNUM *m", although we can't actually
2312 change this in the header file before 0.9.9. It allows
2313 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2314 enable BN_FLG_CONSTTIME.
2316 [Matthew D Wood (Intel Corp)]
2318 *) In the SSL/TLS server implementation, be strict about session ID
2319 context matching (which matters if an application uses a single
2320 external cache for different purposes). Previously,
2321 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2322 set. This did ensure strict client verification, but meant that,
2323 with applications using a single external cache for quite
2324 different requirements, clients could circumvent ciphersuite
2325 restrictions for a given session ID context by starting a session
2326 in a different context.
2329 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2330 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2331 authentication-only ciphersuites.
2334 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2335 not complete and could lead to a possible single byte overflow
2336 (CVE-2007-5135) [Ben Laurie]
2338 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2340 *) Since AES128 and AES256 (and similarly Camellia128 and
2341 Camellia256) share a single mask bit in the logic of
2342 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2343 kludge to work properly if AES128 is available and AES256 isn't
2344 (or if Camellia128 is available and Camellia256 isn't).
2347 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2348 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2349 When a point or a seed is encoded in a BIT STRING, we need to
2350 prevent the removal of trailing zero bits to get the proper DER
2351 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2352 of a NamedBitList, for which trailing 0 bits need to be removed.)
2355 *) Have SSL/TLS server implementation tolerate "mismatched" record
2356 protocol version while receiving ClientHello even if the
2357 ClientHello is fragmented. (The server can't insist on the
2358 particular protocol version it has chosen before the ServerHello
2359 message has informed the client about his choice.)
2362 *) Add RFC 3779 support.
2363 [Rob Austein for ARIN, Ben Laurie]
2365 *) Load error codes if they are not already present instead of using a
2366 static variable. This allows them to be cleanly unloaded and reloaded.
2367 Improve header file function name parsing.
2370 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2371 or CAPABILITY handshake as required by RFCs.
2374 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2376 *) Introduce limits to prevent malicious keys being able to
2377 cause a denial of service. (CVE-2006-2940)
2378 [Steve Henson, Bodo Moeller]
2380 *) Fix ASN.1 parsing of certain invalid structures that can result
2381 in a denial of service. (CVE-2006-2937) [Steve Henson]
2383 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2384 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2386 *) Fix SSL client code which could crash if connecting to a
2387 malicious SSLv2 server. (CVE-2006-4343)
2388 [Tavis Ormandy and Will Drewry, Google Security Team]
2390 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2391 match only those. Before that, "AES256-SHA" would be interpreted
2392 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2393 the same strength classification in 0.9.7h) as we currently only
2394 have a single AES bit in the ciphersuite description bitmap.
2395 That change, however, also applied to ciphersuite strings such as
2396 "RC4-MD5" that intentionally matched multiple ciphersuites --
2397 namely, SSL 2.0 ciphersuites in addition to the more common ones
2398 from SSL 3.0/TLS 1.0.
2400 So we change the selection algorithm again: Naming an explicit
2401 ciphersuite selects this one ciphersuite, and any other similar
2402 ciphersuite (same bitmap) from *other* protocol versions.
2403 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2404 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2406 Since SSL 2.0 does not have any ciphersuites for which the
2407 128/256 bit distinction would be relevant, this works for now.
2408 The proper fix will be to use different bits for AES128 and
2409 AES256, which would have avoided the problems from the beginning;
2410 however, bits are scarce, so we can only do this in a new release
2411 (not just a patchlevel) when we can change the SSL_CIPHER
2412 definition to split the single 'unsigned long mask' bitmap into
2413 multiple values to extend the available space.
2417 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2419 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2420 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2422 *) Add AES IGE and biIGE modes.
2425 *) Change the Unix randomness entropy gathering to use poll() when
2426 possible instead of select(), since the latter has some
2427 undesirable limitations.
2428 [Darryl Miles via Richard Levitte and Bodo Moeller]
2430 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2431 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2432 cannot be implicitly activated as part of, e.g., the "AES" alias.
2433 However, please upgrade to OpenSSL 0.9.9[-dev] for
2434 non-experimental use of the ECC ciphersuites to get TLS extension
2435 support, which is required for curve and point format negotiation
2436 to avoid potential handshake problems.
2439 *) Disable rogue ciphersuites:
2441 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2442 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2443 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2445 The latter two were purportedly from
2446 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2449 Also deactivate the remaining ciphersuites from
2450 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2451 unofficial, and the ID has long expired.
2454 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2455 dual-core machines) and other potential thread-safety issues.
2458 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2459 versions), which is now available for royalty-free use
2460 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2461 Also, add Camellia TLS ciphersuites from RFC 4132.
2463 To minimize changes between patchlevels in the OpenSSL 0.9.8
2464 series, Camellia remains excluded from compilation unless OpenSSL
2465 is configured with 'enable-camellia'.
2468 *) Disable the padding bug check when compression is in use. The padding
2469 bug check assumes the first packet is of even length, this is not
2470 necessarily true if compresssion is enabled and can result in false
2471 positives causing handshake failure. The actual bug test is ancient
2472 code so it is hoped that implementations will either have fixed it by
2473 now or any which still have the bug do not support compression.
2476 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2478 *) When applying a cipher rule check to see if string match is an explicit
2479 cipher suite and only match that one cipher suite if it is.
2482 *) Link in manifests for VC++ if needed.
2483 [Austin Ziegler <halostatue@gmail.com>]
2485 *) Update support for ECC-based TLS ciphersuites according to
2486 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2487 TLS extensions, which are supported starting with the 0.9.9
2488 branch, not in the OpenSSL 0.9.8 branch).
2491 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2492 opaque EVP_CIPHER_CTX handling.
2495 *) Fixes and enhancements to zlib compression code. We now only use
2496 "zlib1.dll" and use the default __cdecl calling convention on Win32
2497 to conform with the standards mentioned here:
2498 http://www.zlib.net/DLL_FAQ.txt
2499 Static zlib linking now works on Windows and the new --with-zlib-include
2500 --with-zlib-lib options to Configure can be used to supply the location
2501 of the headers and library. Gracefully handle case where zlib library
2505 *) Several fixes and enhancements to the OID generation code. The old code
2506 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2507 handle numbers larger than ULONG_MAX, truncated printing and had a
2508 non standard OBJ_obj2txt() behaviour.
2511 *) Add support for building of engines under engine/ as shared libraries
2512 under VC++ build system.
2515 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2516 Hopefully, we will not see any false combination of paths any more.
2519 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2521 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2522 (part of SSL_OP_ALL). This option used to disable the
2523 countermeasure against man-in-the-middle protocol-version
2524 rollback in the SSL 2.0 server implementation, which is a bad
2525 idea. (CVE-2005-2969)
2527 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2528 for Information Security, National Institute of Advanced Industrial
2529 Science and Technology [AIST], Japan)]
2531 *) Add two function to clear and return the verify parameter flags.
2534 *) Keep cipherlists sorted in the source instead of sorting them at
2535 runtime, thus removing the need for a lock.
2538 *) Avoid some small subgroup attacks in Diffie-Hellman.
2539 [Nick Mathewson and Ben Laurie]
2541 *) Add functions for well-known primes.
2544 *) Extended Windows CE support.
2545 [Satoshi Nakamura and Andy Polyakov]
2547 *) Initialize SSL_METHOD structures at compile time instead of during
2548 runtime, thus removing the need for a lock.
2551 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2552 attempting to decrypt each encrypted key in turn. Add support to
2556 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2558 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2561 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2564 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2565 key into the same file any more.
2568 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2571 *) Add -utf8 command line and config file option to 'ca'.
2572 [Stefan <stf@udoma.org]
2574 *) Removed the macro des_crypt(), as it seems to conflict with some
2575 libraries. Use DES_crypt().
2578 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2579 involves renaming the source and generated shared-libs for
2580 both. The engines will accept the corrected or legacy ids
2581 ('ncipher' and '4758_cca' respectively) when binding. NB,
2582 this only applies when building 'shared'.
2583 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2585 *) Add attribute functions to EVP_PKEY structure. Modify
2586 PKCS12_create() to recognize a CSP name attribute and
2587 use it. Make -CSP option work again in pkcs12 utility.
2590 *) Add new functionality to the bn blinding code:
2591 - automatic re-creation of the BN_BLINDING parameters after
2592 a fixed number of uses (currently 32)
2593 - add new function for parameter creation
2594 - introduce flags to control the update behaviour of the
2595 BN_BLINDING parameters
2596 - hide BN_BLINDING structure
2597 Add a second BN_BLINDING slot to the RSA structure to improve
2598 performance when a single RSA object is shared among several
2602 *) Add support for DTLS.
2603 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2605 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2606 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2609 *) Remove buggy and incompletet DH cert support from
2610 ssl/ssl_rsa.c and ssl/s3_both.c
2613 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2614 the apps/openssl applications.
2617 *) Compile clean with "-Wall -Wmissing-prototypes
2618 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2619 DEBUG_SAFESTACK must also be set.
2622 *) Change ./Configure so that certain algorithms can be disabled by default.
2623 The new counterpiece to "no-xxx" is "enable-xxx".
2625 The patented RC5 and MDC2 algorithms will now be disabled unless
2626 "enable-rc5" and "enable-mdc2", respectively, are specified.
2628 (IDEA remains enabled despite being patented. This is because IDEA
2629 is frequently required for interoperability, and there is no license
2630 fee for non-commercial use. As before, "no-idea" can be used to
2631 avoid this algorithm.)
2635 *) Add processing of proxy certificates (see RFC 3820). This work was
2636 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2637 EGEE (Enabling Grids for E-science in Europe).
2640 *) RC4 performance overhaul on modern architectures/implementations, such
2641 as Intel P4, IA-64 and AMD64.
2644 *) New utility extract-section.pl. This can be used specify an alternative
2645 section number in a pod file instead of having to treat each file as
2646 a separate case in Makefile. This can be done by adding two lines to the
2649 =for comment openssl_section:XXX
2651 The blank line is mandatory.
2655 *) New arguments -certform, -keyform and -pass for s_client and s_server
2656 to allow alternative format key and certificate files and passphrase
2660 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2661 update associated structures and add various utility functions.
2663 Add new policy related verify parameters, include policy checking in
2664 standard verify code. Enhance 'smime' application with extra parameters
2665 to support policy checking and print out.
2668 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2669 Nehemiah processors. These extensions support AES encryption in hardware
2670 as well as RNG (though RNG support is currently disabled).
2671 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2673 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2676 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2677 [Andy Polyakov and a number of other people]
2679 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2680 implementation contributed by IBM.
2681 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2683 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2684 exponent rather than 'unsigned long'. There is a corresponding change to
2685 the new 'rsa_keygen' element of the RSA_METHOD structure.
2686 [Jelte Jansen, Geoff Thorpe]
2688 *) Functionality for creating the initial serial number file is now
2689 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2691 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2692 number file to 1, which is bound to cause problems. To avoid
2693 the problems while respecting compatibility between different 0.9.7
2694 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2695 CA.pl for serial number initialization. With the new release 0.9.8,
2696 we can fix the problem directly in the 'ca' utility.)
2699 *) Reduced header interdepencies by declaring more opaque objects in
2700 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2701 give fewer recursive includes, which could break lazy source code - so
2702 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2703 developers should define this symbol when building and using openssl to
2704 ensure they track the recommended behaviour, interfaces, [etc], but
2705 backwards-compatible behaviour prevails when this isn't defined.
2708 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2711 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2712 This will generate a random key of the appropriate length based on the
2713 cipher context. The EVP_CIPHER can provide its own random key generation
2714 routine to support keys of a specific form. This is used in the des and
2715 3des routines to generate a key of the correct parity. Update S/MIME
2716 code to use new functions and hence generate correct parity DES keys.
2717 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2718 valid (weak or incorrect parity).
2721 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2722 as looking them up. This is useful when the verified structure may contain
2723 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2724 present unless the new PKCS7_NO_CRL flag is asserted.
2727 *) Extend ASN1 oid configuration module. It now additionally accepts the
2730 shortName = some long name, 1.2.3.4
2733 *) Reimplemented the BN_CTX implementation. There is now no more static
2734 limitation on the number of variables it can handle nor the depth of the
2735 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2736 information can now expand as required, and rather than having a single
2737 static array of bignums, BN_CTX now uses a linked-list of such arrays
2738 allowing it to expand on demand whilst maintaining the usefulness of
2739 BN_CTX's "bundling".
2742 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2743 to allow all RSA operations to function using a single BN_CTX.
2746 *) Preliminary support for certificate policy evaluation and checking. This
2747 is initially intended to pass the tests outlined in "Conformance Testing
2748 of Relying Party Client Certificate Path Processing Logic" v1.07.
2751 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2752 remained unused and not that useful. A variety of other little bignum
2753 tweaks and fixes have also been made continuing on from the audit (see
2757 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2758 associated ASN1, EVP and SSL functions and old ASN1 macros.
2761 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2762 and this should never fail. So the return value from the use of
2763 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2764 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2767 *) BN_CTX_get() should return zero-valued bignums, providing the same
2768 initialised value as BN_new().
2769 [Geoff Thorpe, suggested by Ulf Möller]
2771 *) Support for inhibitAnyPolicy certificate extension.
2774 *) An audit of the BIGNUM code is underway, for which debugging code is
2775 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2776 is considered valid when processing BIGNUMs, and causes execution to
2777 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2778 further steps are taken to deliberately pollute unused data in BIGNUM
2779 structures to try and expose faulty code further on. For now, openssl will
2780 (in its default mode of operation) continue to tolerate the inconsistent
2781 forms that it has tolerated in the past, but authors and packagers should
2782 consider trying openssl and their own applications when compiled with
2783 these debugging symbols defined. It will help highlight potential bugs in
2784 their own code, and will improve the test coverage for OpenSSL itself. At
2785 some point, these tighter rules will become openssl's default to improve
2786 maintainability, though the assert()s and other overheads will remain only
2787 in debugging configurations. See bn.h for more details.
2788 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2790 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2791 that can only be obtained through BN_CTX_new() (which implicitly
2792 initialises it). The presence of this function only made it possible
2793 to overwrite an existing structure (and cause memory leaks).
2796 *) Because of the callback-based approach for implementing LHASH as a
2797 template type, lh_insert() adds opaque objects to hash-tables and
2798 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2799 to clean up those corresponding objects before destroying the hash table
2800 (and losing the object pointers). So some over-zealous constifications in
2801 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2802 objects as "const" and the lh_doall[_arg] callback wrappers are not
2803 prototyped to have "const" restrictions on the object pointers they are
2804 given (and so aren't required to cast them away any more).
2807 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2808 (speed) prefers to use its own implementation. The two implementations
2809 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2810 its object type properly exposed (MS_TM) instead of casting to/from "char
2811 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2812 aren't necessarily the greatest nomenclatures - but this is what was used
2813 internally to the implementation so I've used that for now.
2816 *) Ensure that deprecated functions do not get compiled when
2817 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2818 the self-tests were still using deprecated key-generation functions so
2819 these have been updated also.
2822 *) Reorganise PKCS#7 code to separate the digest location functionality
2823 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2824 New function PKCS7_set_digest() to set the digest type for PKCS#7
2825 digestedData type. Add additional code to correctly generate the
2826 digestedData type and add support for this type in PKCS7 initialization
2830 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2831 structure of type "other".
2834 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2835 sure the loop does correctly stop and breaking ("division by zero")
2836 modulus operations are not performed. The (pre-generated) prime
2837 table crypto/bn/bn_prime.h was already correct, but it could not be
2838 re-generated on some platforms because of the "division by zero"
2839 situation in the script.
2840 [Ralf S. Engelschall]
2842 *) Update support for ECC-based TLS ciphersuites according to
2843 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2844 SHA-1 now is only used for "small" curves (where the
2845 representation of a field element takes up to 24 bytes); for
2846 larger curves, the field element resulting from ECDH is directly
2847 used as premaster secret.
2848 [Douglas Stebila (Sun Microsystems Laboratories)]
2850 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2851 curve secp160r1 to the tests.
2852 [Douglas Stebila (Sun Microsystems Laboratories)]
2854 *) Add the possibility to load symbols globally with DSO.
2855 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2857 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2858 control of the error stack.
2861 *) Add support for STORE in ENGINE.
2864 *) Add the STORE type. The intention is to provide a common interface
2865 to certificate and key stores, be they simple file-based stores, or
2866 HSM-type store, or LDAP stores, or...
2867 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2870 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2871 pass a list of arguments to any function as well as provide a way
2872 for a function to pass data back to the caller.
2875 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2876 works like BUF_strdup() but can be used to duplicate a portion of
2877 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2881 *) Add the function sk_find_ex() which works like sk_find(), but will
2882 return an index to an element even if an exact match couldn't be
2883 found. The index is guaranteed to point at the element where the
2884 searched-for key would be inserted to preserve sorting order.
2887 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2888 takes an extra flags argument for optional functionality. Currently,
2889 the following flags are defined:
2891 OBJ_BSEARCH_VALUE_ON_NOMATCH
2892 This one gets OBJ_bsearch_ex() to return a pointer to the first
2893 element where the comparing function returns a negative or zero
2896 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2897 This one gets OBJ_bsearch_ex() to return a pointer to the first
2898 element where the comparing function returns zero. This is useful
2899 if there are more than one element where the comparing function
2903 *) Make it possible to create self-signed certificates with 'openssl ca'
2904 in such a way that the self-signed certificate becomes part of the
2905 CA database and uses the same mechanisms for serial number generation
2906 as all other certificate signing. The new flag '-selfsign' enables
2907 this functionality. Adapt CA.sh and CA.pl.in.
2910 *) Add functionality to check the public key of a certificate request
2911 against a given private. This is useful to check that a certificate
2912 request can be signed by that key (self-signing).
2915 *) Make it possible to have multiple active certificates with the same
2916 subject in the CA index file. This is done only if the keyword
2917 'unique_subject' is set to 'no' in the main CA section (default
2918 if 'CA_default') of the configuration file. The value is saved
2919 with the database itself in a separate index attribute file,
2920 named like the index file with '.attr' appended to the name.
2923 *) Generate muti valued AVAs using '+' notation in config files for
2927 *) Support for nameConstraints certificate extension.
2930 *) Support for policyConstraints certificate extension.
2933 *) Support for policyMappings certificate extension.
2936 *) Make sure the default DSA_METHOD implementation only uses its
2937 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2938 and change its own handlers to be NULL so as to remove unnecessary
2939 indirection. This lets alternative implementations fallback to the
2940 default implementation more easily.
2943 *) Support for directoryName in GeneralName related extensions
2947 *) Make it possible to link applications using Makefile.shared.
2948 Make that possible even when linking against static libraries!
2951 *) Support for single pass processing for S/MIME signing. This now
2952 means that S/MIME signing can be done from a pipe, in addition
2953 cleartext signing (multipart/signed type) is effectively streaming
2954 and the signed data does not need to be all held in memory.
2956 This is done with a new flag PKCS7_STREAM. When this flag is set
2957 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2958 is done after the data is output (and digests calculated) in
2959 SMIME_write_PKCS7().
2962 *) Add full support for -rpath/-R, both in shared libraries and
2963 applications, at least on the platforms where it's known how
2967 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2968 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2969 will now compute a table of multiples of the generator that
2970 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2971 faster (notably in the case of a single point multiplication,
2972 scalar * generator).
2973 [Nils Larsch, Bodo Moeller]
2975 *) IPv6 support for certificate extensions. The various extensions
2976 which use the IP:a.b.c.d can now take IPv6 addresses using the
2977 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2981 *) Added an ENGINE that implements RSA by performing private key
2982 exponentiations with the GMP library. The conversions to and from
2983 GMP's mpz_t format aren't optimised nor are any montgomery forms
2984 cached, and on x86 it appears OpenSSL's own performance has caught up.
2985 However there are likely to be other architectures where GMP could
2986 provide a boost. This ENGINE is not built in by default, but it can be
2987 specified at Configure time and should be accompanied by the necessary
2988 linker additions, eg;
2989 ./config -DOPENSSL_USE_GMP -lgmp
2992 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2993 testing availability of engines with "-t" - the old behaviour is
2994 produced by increasing the feature's verbosity with "-tt".
2997 *) ECDSA routines: under certain error conditions uninitialized BN objects
2998 could be freed. Solution: make sure initialization is performed early
2999 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3003 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3004 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3005 software implementations. For DSA and DH, parameter generation can
3006 also be overriden by providing the appropriate method callbacks.
3009 *) Change the "progress" mechanism used in key-generation and
3010 primality testing to functions that take a new BN_GENCB pointer in
3011 place of callback/argument pairs. The new API functions have "_ex"
3012 postfixes and the older functions are reimplemented as wrappers for
3013 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3014 declarations of the old functions to help (graceful) attempts to
3015 migrate to the new functions. Also, the new key-generation API
3016 functions operate on a caller-supplied key-structure and return
3017 success/failure rather than returning a key or NULL - this is to
3018 help make "keygen" another member function of RSA_METHOD etc.
3020 Example for using the new callback interface:
3022 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3026 BN_GENCB_set(&my_cb, my_callback, my_arg);
3028 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3029 /* For the meaning of a, b in calls to my_callback(), see the
3030 * documentation of the function that calls the callback.
3031 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3032 * my_callback should return 1 if it wants BN_is_prime_ex()
3033 * to continue, or 0 to stop.
3038 *) Change the ZLIB compression method to be stateful, and make it
3039 available to TLS with the number defined in
3040 draft-ietf-tls-compression-04.txt.
3043 *) Add the ASN.1 structures and functions for CertificatePair, which
3044 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3046 CertificatePair ::= SEQUENCE {
3047 forward [0] Certificate OPTIONAL,
3048 reverse [1] Certificate OPTIONAL,
3049 -- at least one of the pair shall be present -- }
3051 Also implement the PEM functions to read and write certificate
3052 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3054 This needed to be defined, mostly for the sake of the LDAP
3055 attribute crossCertificatePair, but may prove useful elsewhere as
3059 *) Make it possible to inhibit symlinking of shared libraries in
3060 Makefile.shared, for Cygwin's sake.
3063 *) Extend the BIGNUM API by creating a function
3064 void BN_set_negative(BIGNUM *a, int neg);
3065 and a macro that behave like
3066 int BN_is_negative(const BIGNUM *a);
3068 to avoid the need to access 'a->neg' directly in applications.
3071 *) Implement fast modular reduction for pseudo-Mersenne primes
3072 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3073 EC_GROUP_new_curve_GFp() will now automatically use this
3075 [Nils Larsch <nla@trustcenter.de>]
3077 *) Add new lock type (CRYPTO_LOCK_BN).
3080 *) Change the ENGINE framework to automatically load engines
3081 dynamically from specific directories unless they could be
3082 found to already be built in or loaded. Move all the
3083 current engines except for the cryptodev one to a new
3085 The engines in engines/ are built as shared libraries if
3086 the "shared" options was given to ./Configure or ./config.
3087 Otherwise, they are inserted in libcrypto.a.
3088 /usr/local/ssl/engines is the default directory for dynamic
3089 engines, but that can be overriden at configure time through
3090 the usual use of --prefix and/or --openssldir, and at run
3091 time with the environment variable OPENSSL_ENGINES.
3092 [Geoff Thorpe and Richard Levitte]
3094 *) Add Makefile.shared, a helper makefile to build shared
3095 libraries. Addapt Makefile.org.
3098 *) Add version info to Win32 DLLs.
3099 [Peter 'Luna' Runestig" <peter@runestig.com>]
3101 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3102 can be added using this API to created arbitrary PKCS#12
3103 files while avoiding the low level API.
3105 New options to PKCS12_create(), key or cert can be NULL and
3106 will then be omitted from the output file. The encryption
3107 algorithm NIDs can be set to -1 for no encryption, the mac
3108 iteration count can be set to 0 to omit the mac.
3110 Enhance pkcs12 utility by making the -nokeys and -nocerts
3111 options work when creating a PKCS#12 file. New option -nomac
3112 to omit the mac, NONE can be set for an encryption algorithm.
3113 New code is modified to use the enhanced PKCS12_create()
3114 instead of the low level API.
3117 *) Extend ASN1 encoder to support indefinite length constructed
3118 encoding. This can output sequences tags and octet strings in
3119 this form. Modify pk7_asn1.c to support indefinite length
3120 encoding. This is experimental and needs additional code to
3121 be useful, such as an ASN1 bio and some enhanced streaming
3124 Extend template encode functionality so that tagging is passed
3125 down to the template encoder.
3128 *) Let 'openssl req' fail if an argument to '-newkey' is not
3129 recognized instead of using RSA as a default.
3132 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3133 As these are not official, they are not included in "ALL";
3134 the "ECCdraft" ciphersuite group alias can be used to select them.
3135 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3137 *) Add ECDH engine support.
3138 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3140 *) Add ECDH in new directory crypto/ecdh/.
3141 [Douglas Stebila (Sun Microsystems Laboratories)]
3143 *) Let BN_rand_range() abort with an error after 100 iterations
3144 without success (which indicates a broken PRNG).
3147 *) Change BN_mod_sqrt() so that it verifies that the input value
3148 is really the square of the return value. (Previously,
3149 BN_mod_sqrt would show GIGO behaviour.)
3152 *) Add named elliptic curves over binary fields from X9.62, SECG,
3153 and WAP/WTLS; add OIDs that were still missing.
3155 [Sheueling Chang Shantz and Douglas Stebila
3156 (Sun Microsystems Laboratories)]
3158 *) Extend the EC library for elliptic curves over binary fields
3159 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3162 EC_GF2m_simple_method
3166 EC_GROUP_new_curve_GF2m
3167 EC_GROUP_set_curve_GF2m
3168 EC_GROUP_get_curve_GF2m
3169 EC_POINT_set_affine_coordinates_GF2m
3170 EC_POINT_get_affine_coordinates_GF2m
3171 EC_POINT_set_compressed_coordinates_GF2m
3173 Point compression for binary fields is disabled by default for
3174 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3177 As binary polynomials are represented as BIGNUMs, various members
3178 of the EC_GROUP and EC_POINT data structures can be shared
3179 between the implementations for prime fields and binary fields;
3180 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3181 are essentially identical to their ..._GFp counterparts.
3182 (For simplicity, the '..._GFp' prefix has been dropped from
3183 various internal method names.)
3185 An internal 'field_div' method (similar to 'field_mul' and
3186 'field_sqr') has been added; this is used only for binary fields.
3188 [Sheueling Chang Shantz and Douglas Stebila
3189 (Sun Microsystems Laboratories)]
3191 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3192 through methods ('mul', 'precompute_mult').
3194 The generic implementations (now internally called 'ec_wNAF_mul'
3195 and 'ec_wNAF_precomputed_mult') remain the default if these
3196 methods are undefined.
3198 [Sheueling Chang Shantz and Douglas Stebila
3199 (Sun Microsystems Laboratories)]
3201 *) New function EC_GROUP_get_degree, which is defined through
3202 EC_METHOD. For curves over prime fields, this returns the bit
3203 length of the modulus.
3205 [Sheueling Chang Shantz and Douglas Stebila
3206 (Sun Microsystems Laboratories)]
3208 *) New functions EC_GROUP_dup, EC_POINT_dup.
3209 (These simply call ..._new and ..._copy).
3211 [Sheueling Chang Shantz and Douglas Stebila
3212 (Sun Microsystems Laboratories)]
3214 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3215 Polynomials are represented as BIGNUMs (where the sign bit is not
3216 used) in the following functions [macros]:
3219 BN_GF2m_sub [= BN_GF2m_add]
3220 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3221 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3222 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3224 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3225 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3226 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3227 BN_GF2m_cmp [= BN_ucmp]
3229 (Note that only the 'mod' functions are actually for fields GF(2^m).
3230 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3232 For some functions, an the irreducible polynomial defining a
3233 field can be given as an 'unsigned int[]' with strictly
3234 decreasing elements giving the indices of those bits that are set;
3235 i.e., p[] represents the polynomial
3236 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3238 p[0] > p[1] > ... > p[k] = 0.
3239 This applies to the following functions:
3244 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3245 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3247 BN_GF2m_mod_sqrt_arr
3248 BN_GF2m_mod_solve_quad_arr
3252 Conversion can be performed by the following functions:
3257 bntest.c has additional tests for binary polynomial arithmetic.
3259 Two implementations for BN_GF2m_mod_div() are available.
3260 The default algorithm simply uses BN_GF2m_mod_inv() and
3261 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3262 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3263 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3265 [Sheueling Chang Shantz and Douglas Stebila
3266 (Sun Microsystems Laboratories)]
3268 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3269 functionality is disabled at compile-time.
3270 [Douglas Stebila <douglas.stebila@sun.com>]
3272 *) Change default behaviour of 'openssl asn1parse' so that more
3273 information is visible when viewing, e.g., a certificate:
3275 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3276 mode the content of non-printable OCTET STRINGs is output in a
3277 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3278 avoid the appearance of a printable string.
3279 [Nils Larsch <nla@trustcenter.de>]
3281 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3283 EC_GROUP_set_asn1_flag()
3284 EC_GROUP_get_asn1_flag()
3285 EC_GROUP_set_point_conversion_form()
3286 EC_GROUP_get_point_conversion_form()
3287 These control ASN1 encoding details:
3288 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3289 has been set to OPENSSL_EC_NAMED_CURVE.
3290 - Points are encoded in uncompressed form by default; options for
3291 asn1_for are as for point2oct, namely
3292 POINT_CONVERSION_COMPRESSED
3293 POINT_CONVERSION_UNCOMPRESSED
3294 POINT_CONVERSION_HYBRID
3296 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3299 EC_GROUP_get0_seed()
3300 EC_GROUP_get_seed_len()
3301 This is used only for ASN1 purposes (so far).
3302 [Nils Larsch <nla@trustcenter.de>]
3304 *) Add 'field_type' member to EC_METHOD, which holds the NID
3305 of the appropriate field type OID. The new function
3306 EC_METHOD_get_field_type() returns this value.
3307 [Nils Larsch <nla@trustcenter.de>]
3312 EC_POINT_point2hex()
3313 EC_POINT_hex2point()
3314 providing useful interfaces to EC_POINT_point2oct() and
3315 EC_POINT_oct2point().
3316 [Nils Larsch <nla@trustcenter.de>]
3318 *) Change internals of the EC library so that the functions
3319 EC_GROUP_set_generator()
3320 EC_GROUP_get_generator()
3321 EC_GROUP_get_order()
3322 EC_GROUP_get_cofactor()
3323 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3324 to methods, which would lead to unnecessary code duplication when
3325 adding different types of curves.
3326 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3328 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3329 arithmetic, and such that modified wNAFs are generated
3330 (which avoid length expansion in many cases).
3333 *) Add a function EC_GROUP_check_discriminant() (defined via
3334 EC_METHOD) that verifies that the curve discriminant is non-zero.
3336 Add a function EC_GROUP_check() that makes some sanity tests
3337 on a EC_GROUP, its generator and order. This includes
3338 EC_GROUP_check_discriminant().
3339 [Nils Larsch <nla@trustcenter.de>]
3341 *) Add ECDSA in new directory crypto/ecdsa/.
3343 Add applications 'openssl ecparam' and 'openssl ecdsa'
3344 (these are based on 'openssl dsaparam' and 'openssl dsa').
3346 ECDSA support is also included in various other files across the
3347 library. Most notably,
3348 - 'openssl req' now has a '-newkey ecdsa:file' option;
3349 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3350 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3351 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3352 them suitable for ECDSA where domain parameters must be
3353 extracted before the specific public key;
3354 - ECDSA engine support has been added.
3355 [Nils Larsch <nla@trustcenter.de>]
3357 *) Include some named elliptic curves, and add OIDs from X9.62,
3358 SECG, and WAP/WTLS. Each curve can be obtained from the new
3360 EC_GROUP_new_by_curve_name(),
3361 and the list of available named curves can be obtained with
3362 EC_get_builtin_curves().
3363 Also add a 'curve_name' member to EC_GROUP objects, which can be
3365 EC_GROUP_set_curve_name()
3366 EC_GROUP_get_curve_name()
3367 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3369 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3370 was actually never needed) and in BN_mul(). The removal in BN_mul()
3371 required a small change in bn_mul_part_recursive() and the addition
3372 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3373 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3374 bn_sub_words() and bn_add_words() except they take arrays with
3378 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3380 *) Cleanse PEM buffers before freeing them since they may contain
3382 [Benjamin Bennett <ben@psc.edu>]
3384 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3385 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3386 authentication-only ciphersuites.
3389 *) Since AES128 and AES256 share a single mask bit in the logic of
3390 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3391 kludge to work properly if AES128 is available and AES256 isn't.
3394 *) Expand security boundary to match 1.1.1 module.
3397 *) Remove redundant features: hash file source, editing of test vectors
3398 modify fipsld to use external fips_premain.c signature.
3401 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3402 run algorithm test programs.
3405 *) Make algorithm test programs more tolerant of whitespace.
3408 *) Have SSL/TLS server implementation tolerate "mismatched" record
3409 protocol version while receiving ClientHello even if the
3410 ClientHello is fragmented. (The server can't insist on the
3411 particular protocol version it has chosen before the ServerHello
3412 message has informed the client about his choice.)
3415 *) Load error codes if they are not already present instead of using a
3416 static variable. This allows them to be cleanly unloaded and reloaded.
3419 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3421 *) Introduce limits to prevent malicious keys being able to
3422 cause a denial of service. (CVE-2006-2940)
3423 [Steve Henson, Bodo Moeller]
3425 *) Fix ASN.1 parsing of certain invalid structures that can result
3426 in a denial of service. (CVE-2006-2937) [Steve Henson]
3428 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3429 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3431 *) Fix SSL client code which could crash if connecting to a
3432 malicious SSLv2 server. (CVE-2006-4343)
3433 [Tavis Ormandy and Will Drewry, Google Security Team]
3435 *) Change ciphersuite string processing so that an explicit
3436 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3437 will no longer include "AES128-SHA"), and any other similar
3438 ciphersuite (same bitmap) from *other* protocol versions (so that
3439 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3440 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3441 changes from 0.9.8b and 0.9.8d.
3444 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3446 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3447 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3449 *) Change the Unix randomness entropy gathering to use poll() when
3450 possible instead of select(), since the latter has some
3451 undesirable limitations.
3452 [Darryl Miles via Richard Levitte and Bodo Moeller]
3454 *) Disable rogue ciphersuites:
3456 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3457 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3458 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3460 The latter two were purportedly from
3461 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3464 Also deactive the remaining ciphersuites from
3465 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3466 unofficial, and the ID has long expired.
3469 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3470 dual-core machines) and other potential thread-safety issues.
3473 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3475 *) Adapt fipsld and the build system to link against the validated FIPS
3476 module in FIPS mode.
3479 *) Fixes for VC++ 2005 build under Windows.
3482 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3483 from a Windows bash shell such as MSYS. It is autodetected from the
3484 "config" script when run from a VC++ environment. Modify standard VC++
3485 build to use fipscanister.o from the GNU make build.
3488 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3490 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3491 The value now differs depending on if you build for FIPS or not.
3492 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3493 safely run with a non-FIPSed libcrypto, as it may crash because of
3494 the difference induced by this change.
3497 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3499 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3500 (part of SSL_OP_ALL). This option used to disable the
3501 countermeasure against man-in-the-middle protocol-version
3502 rollback in the SSL 2.0 server implementation, which is a bad
3503 idea. (CVE-2005-2969)
3505 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3506 for Information Security, National Institute of Advanced Industrial
3507 Science and Technology [AIST], Japan)]
3509 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3510 mainly for FIPS compliance and not fully integrated at this stage.
3513 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3514 the exponentiation using a fixed-length exponent. (Otherwise,
3515 the information leaked through timing could expose the secret key
3516 after many signatures; cf. Bleichenbacher's attack on DSA with
3520 *) Make a new fixed-window mod_exp implementation the default for
3521 RSA, DSA, and DH private-key operations so that the sequence of
3522 squares and multiplies and the memory access pattern are
3523 independent of the particular secret key. This will mitigate
3524 cache-timing and potential related attacks.
3526 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3527 and this is automatically used by BN_mod_exp_mont() if the new flag
3528 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3529 will use this BN flag for private exponents unless the flag
3530 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3531 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3533 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3535 *) Change the client implementation for SSLv23_method() and
3536 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3537 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3538 (Previously, the SSL 2.0 backwards compatible Client Hello
3539 message format would be used even with SSL_OP_NO_SSLv2.)
3542 *) Add support for smime-type MIME parameter in S/MIME messages which some
3546 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3547 a threadsafe manner. Modify rsa code to use new function and add calls
3548 to dsa and dh code (which had race conditions before).
3551 *) Include the fixed error library code in the C error file definitions
3552 instead of fixing them up at runtime. This keeps the error code
3553 structures constant.
3556 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3558 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3561 *) Fixes for newer kerberos headers. NB: the casts are needed because
3562 the 'length' field is signed on one version and unsigned on another
3563 with no (?) obvious way to tell the difference, without these VC++
3564 complains. Also the "definition" of FAR (blank) is no longer included
3565 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3566 some needed definitions.
3569 *) Undo Cygwin change.
3572 *) Added support for proxy certificates according to RFC 3820.
3573 Because they may be a security thread to unaware applications,
3574 they must be explicitely allowed in run-time. See
3575 docs/HOWTO/proxy_certificates.txt for further information.
3578 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3580 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3581 server and client random values. Previously
3582 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3583 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3585 This change has negligible security impact because:
3587 1. Server and client random values still have 24 bytes of pseudo random
3590 2. Server and client random values are sent in the clear in the initial
3593 3. The master secret is derived using the premaster secret (48 bytes in
3594 size for static RSA ciphersuites) as well as client server and random
3597 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3600 [Stephen Henson, reported by UK NISCC]
3602 *) Use Windows randomness collection on Cygwin.
3605 *) Fix hang in EGD/PRNGD query when communication socket is closed
3606 prematurely by EGD/PRNGD.
3607 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3609 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3612 *) Back-port of selected performance improvements from development
3613 branch, as well as improved support for PowerPC platforms.
3616 *) Add lots of checks for memory allocation failure, error codes to indicate
3617 failure and freeing up memory if a failure occurs.
3618 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3620 *) Add new -passin argument to dgst.
3623 *) Perform some character comparisons of different types in X509_NAME_cmp:
3624 this is needed for some certificates that reencode DNs into UTF8Strings
3625 (in violation of RFC3280) and can't or wont issue name rollover
3629 *) Make an explicit check during certificate validation to see that
3630 the CA setting in each certificate on the chain is correct. As a
3631 side effect always do the following basic checks on extensions,
3632 not just when there's an associated purpose to the check:
3634 - if there is an unhandled critical extension (unless the user
3635 has chosen to ignore this fault)
3636 - if the path length has been exceeded (if one is set at all)
3637 - that certain extensions fit the associated purpose (if one has
3641 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3643 *) Avoid a race condition when CRLs are checked in a multi threaded
3644 environment. This would happen due to the reordering of the revoked
3645 entries during signature checking and serial number lookup. Now the
3646 encoding is cached and the serial number sort performed under a lock.
3647 Add new STACK function sk_is_sorted().
3650 *) Add Delta CRL to the extension code.
3653 *) Various fixes to s3_pkt.c so alerts are sent properly.
3654 [David Holmes <d.holmes@f5.com>]
3656 *) Reduce the chances of duplicate issuer name and serial numbers (in
3657 violation of RFC3280) using the OpenSSL certificate creation utilities.
3658 This is done by creating a random 64 bit value for the initial serial
3659 number when a serial number file is created or when a self signed
3660 certificate is created using 'openssl req -x509'. The initial serial
3661 number file is created using 'openssl x509 -next_serial' in CA.pl
3662 rather than being initialized to 1.
3665 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3667 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3668 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3669 [Joe Orton, Steve Henson]
3671 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3673 [Joe Orton, Steve Henson]
3675 *) Make it possible to have multiple active certificates with the same
3676 subject in the CA index file. This is done only if the keyword
3677 'unique_subject' is set to 'no' in the main CA section (default
3678 if 'CA_default') of the configuration file. The value is saved
3679 with the database itself in a separate index attribute file,
3680 named like the index file with '.attr' appended to the name.
3683 *) X509 verify fixes. Disable broken certificate workarounds when
3684 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3685 keyUsage extension present. Don't accept CRLs with unhandled critical
3686 extensions: since verify currently doesn't process CRL extensions this
3687 rejects a CRL with *any* critical extensions. Add new verify error codes
3691 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3692 A clarification of RFC2560 will require the use of OCTET STRINGs and
3693 some implementations cannot handle the current raw format. Since OpenSSL
3694 copies and compares OCSP nonces as opaque blobs without any attempt at
3695 parsing them this should not create any compatibility issues.
3698 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3699 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3700 this HMAC (and other) operations are several times slower than OpenSSL
3704 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3705 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3707 *) Use the correct content when signing type "other".
3710 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3712 *) Fix various bugs revealed by running the NISCC test suite:
3714 Stop out of bounds reads in the ASN1 code when presented with
3715 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3717 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3719 If verify callback ignores invalid public key errors don't try to check
3720 certificate signature with the NULL public key.
3724 *) New -ignore_err option in ocsp application to stop the server
3725 exiting on the first error in a request.
3728 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3729 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3733 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3734 extra data after the compression methods not only for TLS 1.0
3735 but also for SSL 3.0 (as required by the specification).
3736 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3738 *) Change X509_certificate_type() to mark the key as exported/exportable
3739 when it's 512 *bits* long, not 512 bytes.
3742 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3743 blocks during encryption.
3746 *) Various fixes to base64 BIO and non blocking I/O. On write
3747 flushes were not handled properly if the BIO retried. On read
3748 data was not being buffered properly and had various logic bugs.
3749 This also affects blocking I/O when the data being decoded is a
3753 *) Various S/MIME bugfixes and compatibility changes:
3754 output correct application/pkcs7 MIME type if
3755 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3756 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3757 of files as .eml work). Correctly handle very long lines in MIME
3761 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3763 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3764 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3765 a protocol version number mismatch like a decryption error
3766 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3769 *) Turn on RSA blinding by default in the default implementation
3770 to avoid a timing attack. Applications that don't want it can call
3771 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3772 They would be ill-advised to do so in most cases.
3773 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3775 *) Change RSA blinding code so that it works when the PRNG is not
3776 seeded (in this case, the secret RSA exponent is abused as
3777 an unpredictable seed -- if it is not unpredictable, there
3778 is no point in blinding anyway). Make RSA blinding thread-safe
3779 by remembering the creator's thread ID in rsa->blinding and
3780 having all other threads use local one-time blinding factors
3781 (this requires more computation than sharing rsa->blinding, but
3782 avoids excessive locking; and if an RSA object is not shared
3783 between threads, blinding will still be very fast).
3786 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3787 ENGINE as defaults for all supported algorithms irrespective of
3788 the 'flags' parameter. 'flags' is now honoured, so applications
3789 should make sure they are passing it correctly.
3792 *) Target "mingw" now allows native Windows code to be generated in
3793 the Cygwin environment as well as with the MinGW compiler.
3796 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3798 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3799 via timing by performing a MAC computation even if incorrrect
3800 block cipher padding has been found. This is a countermeasure
3801 against active attacks where the attacker has to distinguish
3802 between bad padding and a MAC verification error. (CVE-2003-0078)
3804 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3805 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3806 Martin Vuagnoux (EPFL, Ilion)]
3808 *) Make the no-err option work as intended. The intention with no-err
3809 is not to have the whole error stack handling routines removed from
3810 libcrypto, it's only intended to remove all the function name and
3811 reason texts, thereby removing some of the footprint that may not
3812 be interesting if those errors aren't displayed anyway.
3814 NOTE: it's still possible for any application or module to have it's
3815 own set of error texts inserted. The routines are there, just not
3816 used by default when no-err is given.
3819 *) Add support for FreeBSD on IA64.
3820 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3822 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3823 Kerberos function mit_des_cbc_cksum(). Before this change,
3824 the value returned by DES_cbc_cksum() was like the one from
3825 mit_des_cbc_cksum(), except the bytes were swapped.
3826 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3828 *) Allow an application to disable the automatic SSL chain building.
3829 Before this a rather primitive chain build was always performed in
3830 ssl3_output_cert_chain(): an application had no way to send the
3831 correct chain if the automatic operation produced an incorrect result.
3833 Now the chain builder is disabled if either:
3835 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3837 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3839 The reasoning behind this is that an application would not want the
3840 auto chain building to take place if extra chain certificates are
3841 present and it might also want a means of sending no additional
3842 certificates (for example the chain has two certificates and the
3846 *) Add the possibility to build without the ENGINE framework.
3847 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3849 *) Under Win32 gmtime() can return NULL: check return value in
3850 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3853 *) DSA routines: under certain error conditions uninitialized BN objects
3854 could be freed. Solution: make sure initialization is performed early
3855 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3856 Nils Larsch <nla@trustcenter.de> via PR#459)
3859 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3860 checked on reconnect on the client side, therefore session resumption
3861 could still fail with a "ssl session id is different" error. This
3862 behaviour is masked when SSL_OP_ALL is used due to
3863 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3864 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3865 followup to PR #377.
3868 *) IA-32 assembler support enhancements: unified ELF targets, support
3869 for SCO/Caldera platforms, fix for Cygwin shared build.
3872 *) Add support for FreeBSD on sparc64. As a consequence, support for
3873 FreeBSD on non-x86 processors is separate from x86 processors on
3874 the config script, much like the NetBSD support.
3875 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3877 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3879 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3882 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3883 code (06) was taken as the first octet of the session ID and the last
3884 octet was ignored consequently. As a result SSLv2 client side session
3885 caching could not have worked due to the session ID mismatch between
3887 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3891 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3892 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3896 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3897 seems that in spite of existing for more than a year, many application
3898 author have done nothing to provide the necessary callbacks, which
3899 means that this particular engine will not work properly anywhere.
3900 This is a very unfortunate situation which forces us, in the name
3901 of usability, to give the hw_ncipher.c a static lock, which is part
3903 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3904 appear in 0.9.8 or later. We EXPECT application authors to have
3905 dealt properly with this when 0.9.8 is released (unless we actually
3906 make such changes in the libcrypto locking code that changes will
3907 have to be made anyway).
3910 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3911 octets have been read, EOF or an error occurs. Without this change
3912 some truncated ASN1 structures will not produce an error.
3915 *) Disable Heimdal support, since it hasn't been fully implemented.
3916 Still give the possibility to force the use of Heimdal, but with
3917 warnings and a request that patches get sent to openssl-dev.
3920 *) Add the VC-CE target, introduce the WINCE sysname, and add
3921 INSTALL.WCE and appropriate conditionals to make it build.
3922 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3924 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3925 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3926 edit numbers of the version.
3927 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3929 *) Introduce safe string copy and catenation functions
3930 (BUF_strlcpy() and BUF_strlcat()).
3931 [Ben Laurie (CHATS) and Richard Levitte]
3933 *) Avoid using fixed-size buffers for one-line DNs.
3934 [Ben Laurie (CHATS)]
3936 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3937 resizing buffers containing secrets, and use where appropriate.
3938 [Ben Laurie (CHATS)]
3940 *) Avoid using fixed size buffers for configuration file location.
3941 [Ben Laurie (CHATS)]
3943 *) Avoid filename truncation for various CA files.
3944 [Ben Laurie (CHATS)]
3946 *) Use sizeof in preference to magic numbers.
3947 [Ben Laurie (CHATS)]
3949 *) Avoid filename truncation in cert requests.
3950 [Ben Laurie (CHATS)]
3952 *) Add assertions to check for (supposedly impossible) buffer
3954 [Ben Laurie (CHATS)]
3956 *) Don't cache truncated DNS entries in the local cache (this could
3957 potentially lead to a spoofing attack).
3958 [Ben Laurie (CHATS)]
3960 *) Fix various buffers to be large enough for hex/decimal
3961 representations in a platform independent manner.
3962 [Ben Laurie (CHATS)]
3964 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3965 resizing buffers containing secrets, and use where appropriate.
3966 [Ben Laurie (CHATS)]
3968 *) Add BIO_indent() to avoid much slightly worrying code to do
3970 [Ben Laurie (CHATS)]
3972 *) Convert sprintf()/BIO_puts() to BIO_printf().
3973 [Ben Laurie (CHATS)]
3975 *) buffer_gets() could terminate with the buffer only half
3977 [Ben Laurie (CHATS)]
3979 *) Add assertions to prevent user-supplied crypto functions from
3980 overflowing internal buffers by having large block sizes, etc.
3981 [Ben Laurie (CHATS)]
3983 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3985 [Ben Laurie (CHATS)]
3987 *) Eliminate unused copy of key in RC4.
3988 [Ben Laurie (CHATS)]
3990 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3991 [Ben Laurie (CHATS)]
3993 *) Fix off-by-one error in EGD path.
3994 [Ben Laurie (CHATS)]
3996 *) If RANDFILE path is too long, ignore instead of truncating.
3997 [Ben Laurie (CHATS)]
3999 *) Eliminate unused and incorrectly sized X.509 structure
4001 [Ben Laurie (CHATS)]
4003 *) Eliminate unused and dangerous function knumber().
4004 [Ben Laurie (CHATS)]
4006 *) Eliminate unused and dangerous structure, KSSL_ERR.
4007 [Ben Laurie (CHATS)]
4009 *) Protect against overlong session ID context length in an encoded
4010 session object. Since these are local, this does not appear to be
4012 [Ben Laurie (CHATS)]
4014 *) Change from security patch (see 0.9.6e below) that did not affect
4015 the 0.9.6 release series:
4017 Remote buffer overflow in SSL3 protocol - an attacker could
4018 supply an oversized master key in Kerberos-enabled versions.
4020 [Ben Laurie (CHATS)]
4022 *) Change the SSL kerb5 codes to match RFC 2712.
4025 *) Make -nameopt work fully for req and add -reqopt switch.
4026 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4028 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4029 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4031 *) Make sure tests can be performed even if the corresponding algorithms
4032 have been removed entirely. This was also the last step to make
4033 OpenSSL compilable with DJGPP under all reasonable conditions.
4034 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4036 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4037 to allow version independent disabling of normally unselected ciphers,
4038 which may be activated as a side-effect of selecting a single cipher.
4040 (E.g., cipher list string "RSA" enables ciphersuites that are left
4041 out of "ALL" because they do not provide symmetric encryption.
4042 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4043 [Lutz Jaenicke, Bodo Moeller]
4045 *) Add appropriate support for separate platform-dependent build
4046 directories. The recommended way to make a platform-dependent
4047 build directory is the following (tested on Linux), maybe with
4050 # Place yourself outside of the OpenSSL source tree. In
4051 # this example, the environment variable OPENSSL_SOURCE
4052 # is assumed to contain the absolute OpenSSL source directory.
4053 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4054 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4055 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4056 mkdir -p `dirname $F`
4057 ln -s $OPENSSL_SOURCE/$F $F
4060 To be absolutely sure not to disturb the source tree, a "make clean"
4061 is a good thing. If it isn't successfull, don't worry about it,
4062 it probably means the source directory is very clean.
4065 *) Make sure any ENGINE control commands make local copies of string
4066 pointers passed to them whenever necessary. Otherwise it is possible
4067 the caller may have overwritten (or deallocated) the original string
4068 data when a later ENGINE operation tries to use the stored values.
4069 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4071 *) Improve diagnostics in file reading and command-line digests.
4072 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4074 *) Add AES modes CFB and OFB to the object database. Correct an
4075 error in AES-CFB decryption.
4078 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4079 allows existing EVP_CIPHER_CTX structures to be reused after
4080 calling EVP_*Final(). This behaviour is used by encryption
4081 BIOs and some applications. This has the side effect that
4082 applications must explicitly clean up cipher contexts with
4083 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4086 *) Check the values of dna and dnb in bn_mul_recursive before calling
4087 bn_mul_comba (a non zero value means the a or b arrays do not contain
4088 n2 elements) and fallback to bn_mul_normal if either is not zero.
4091 *) Fix escaping of non-ASCII characters when using the -subj option
4092 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4095 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4096 form for "surname", serialNumber has no short form.
4097 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4098 therefore remove "mail" short name for "internet 7".
4099 The OID for unique identifiers in X509 certificates is
4100 x500UniqueIdentifier, not uniqueIdentifier.
4101 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4104 *) Add an "init" command to the ENGINE config module and auto initialize
4105 ENGINEs. Without any "init" command the ENGINE will be initialized
4106 after all ctrl commands have been executed on it. If init=1 the
4107 ENGINE is initailized at that point (ctrls before that point are run
4108 on the uninitialized ENGINE and after on the initialized one). If
4109 init=0 then the ENGINE will not be iniatialized at all.
4112 *) Fix the 'app_verify_callback' interface so that the user-defined
4113 argument is actually passed to the callback: In the
4114 SSL_CTX_set_cert_verify_callback() prototype, the callback
4115 declaration has been changed from
4118 int (*cb)(X509_STORE_CTX *,void *);
4119 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4120 i=s->ctx->app_verify_callback(&ctx)
4121 has been changed into
4122 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4124 To update applications using SSL_CTX_set_cert_verify_callback(),
4125 a dummy argument can be added to their callback functions.
4126 [D. K. Smetters <smetters@parc.xerox.com>]
4128 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4129 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4131 *) Add and OPENSSL_LOAD_CONF define which will cause
4132 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4133 This allows older applications to transparently support certain
4134 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4135 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4136 load the config file and OPENSSL_add_all_algorithms_conf() which will
4137 always load it have also been added.
4140 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4141 Adjust NIDs and EVP layer.
4142 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4144 *) Config modules support in openssl utility.
4146 Most commands now load modules from the config file,
4147 though in a few (such as version) this isn't done
4148 because it couldn't be used for anything.
4150 In the case of ca and req the config file used is
4151 the same as the utility itself: that is the -config
4152 command line option can be used to specify an
4156 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4157 use "openssl_conf" if filename is NULL use default openssl config file.
4160 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4161 config section name. Add a new flag to tolerate a missing config file
4162 and move code to CONF_modules_load_file().
4165 *) Support for crypto accelerator cards from Accelerated Encryption
4166 Processing, www.aep.ie. (Use engine 'aep')
4167 The support was copied from 0.9.6c [engine] and adapted/corrected
4168 to work with the new engine framework.
4169 [AEP Inc. and Richard Levitte]
4171 *) Support for SureWare crypto accelerator cards from Baltimore
4172 Technologies. (Use engine 'sureware')
4173 The support was copied from 0.9.6c [engine] and adapted
4174 to work with the new engine framework.
4177 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4178 make the newer ENGINE framework commands for the CHIL engine work.
4179 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4181 *) Make it possible to produce shared libraries on ReliantUNIX.
4182 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4184 *) Add the configuration target debug-linux-ppro.
4185 Make 'openssl rsa' use the general key loading routines
4186 implemented in apps.c, and make those routines able to
4187 handle the key format FORMAT_NETSCAPE and the variant
4189 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4191 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4192 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4194 *) Add -keyform to rsautl, and document -engine.
4195 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4197 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4198 BIO_R_NO_SUCH_FILE error code rather than the generic
4199 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4202 *) Add new functions
4204 ERR_peek_last_error_line
4205 ERR_peek_last_error_line_data.
4206 These are similar to
4209 ERR_peek_error_line_data,
4210 but report on the latest error recorded rather than the first one
4211 still in the error queue.
4212 [Ben Laurie, Bodo Moeller]
4214 *) default_algorithms option in ENGINE config module. This allows things
4216 default_algorithms = ALL
4217 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4220 *) Prelminary ENGINE config module.
4223 *) New experimental application configuration code.
4226 *) Change the AES code to follow the same name structure as all other
4227 symmetric ciphers, and behave the same way. Move everything to
4228 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4229 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4231 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4232 [Ben Laurie and Theo de Raadt]
4234 *) Add option to output public keys in req command.
4235 [Massimiliano Pala madwolf@openca.org]
4237 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4238 (up to about 10% better than before for P-192 and P-224).
4241 *) New functions/macros
4243 SSL_CTX_set_msg_callback(ctx, cb)
4244 SSL_CTX_set_msg_callback_arg(ctx, arg)
4245 SSL_set_msg_callback(ssl, cb)
4246 SSL_set_msg_callback_arg(ssl, arg)
4248 to request calling a callback function
4250 void cb(int write_p, int version, int content_type,
4251 const void *buf, size_t len, SSL *ssl, void *arg)
4253 whenever a protocol message has been completely received
4254 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4255 protocol version according to which the SSL library interprets
4256 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4257 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4258 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4259 specification (change_cipher_spec(20), alert(21), handshake(22)).
4260 'buf' and 'len' point to the actual message, 'ssl' to the
4261 SSL object, and 'arg' is the application-defined value set by
4262 SSL[_CTX]_set_msg_callback_arg().
4264 'openssl s_client' and 'openssl s_server' have new '-msg' options
4265 to enable a callback that displays all protocol messages.
4268 *) Change the shared library support so shared libraries are built as
4269 soon as the corresponding static library is finished, and thereby get
4270 openssl and the test programs linked against the shared library.
4271 This still only happens when the keyword "shard" has been given to
4272 the configuration scripts.
4274 NOTE: shared library support is still an experimental thing, and
4275 backward binary compatibility is still not guaranteed.
4276 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4278 *) Add support for Subject Information Access extension.
4279 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4281 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4282 additional bytes when new memory had to be allocated, not just
4283 when reusing an existing buffer.
4286 *) New command line and configuration option 'utf8' for the req command.
4287 This allows field values to be specified as UTF8 strings.
4290 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4291 runs for the former and machine-readable output for the latter.
4294 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4295 of the e-mail address in the DN (i.e., it will go into a certificate
4296 extension only). The new configuration file option 'email_in_dn = no'
4297 has the same effect.
4298 [Massimiliano Pala madwolf@openca.org]
4300 *) Change all functions with names starting with des_ to be starting
4301 with DES_ instead. Add wrappers that are compatible with libdes,
4302 but are named _ossl_old_des_*. Finally, add macros that map the
4303 des_* symbols to the corresponding _ossl_old_des_* if libdes
4304 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4305 desired, the des_* symbols will be mapped to DES_*, with one
4308 Since we provide two compatibility mappings, the user needs to
4309 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4310 compatibility is desired. The default (i.e., when that macro
4311 isn't defined) is OpenSSL 0.9.6c compatibility.
4313 There are also macros that enable and disable the support of old
4314 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4315 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4316 are defined, the default will apply: to support the old des routines.
4318 In either case, one must include openssl/des.h to get the correct
4319 definitions. Do not try to just include openssl/des_old.h, that
4322 NOTE: This is a major break of an old API into a new one. Software
4323 authors are encouraged to switch to the DES_ style functions. Some
4324 time in the future, des_old.h and the libdes compatibility functions
4325 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4326 default), and then completely removed.
4329 *) Test for certificates which contain unsupported critical extensions.
4330 If such a certificate is found during a verify operation it is
4331 rejected by default: this behaviour can be overridden by either
4332 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4333 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4334 X509_supported_extension() has also been added which returns 1 if a
4335 particular extension is supported.
4338 *) Modify the behaviour of EVP cipher functions in similar way to digests
4339 to retain compatibility with existing code.
4342 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4343 compatibility with existing code. In particular the 'ctx' parameter does
4344 not have to be to be initialized before the call to EVP_DigestInit() and
4345 it is tidied up after a call to EVP_DigestFinal(). New function
4346 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4347 EVP_MD_CTX_copy() changed to not require the destination to be
4348 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4349 requires the destination to be valid.
4351 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4352 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4355 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4356 so that complete 'Handshake' protocol structures are kept in memory
4357 instead of overwriting 'msg_type' and 'length' with 'body' data.
4360 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4361 [Massimo Santin via Richard Levitte]
4363 *) Major restructuring to the underlying ENGINE code. This includes
4364 reduction of linker bloat, separation of pure "ENGINE" manipulation
4365 (initialisation, etc) from functionality dealing with implementations
4366 of specific crypto iterfaces. This change also introduces integrated
4367 support for symmetric ciphers and digest implementations - so ENGINEs
4368 can now accelerate these by providing EVP_CIPHER and EVP_MD
4369 implementations of their own. This is detailed in crypto/engine/README
4370 as it couldn't be adequately described here. However, there are a few
4371 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4372 were changed in the original introduction of ENGINE code have now
4373 reverted back - the hooking from this code to ENGINE is now a good
4374 deal more passive and at run-time, operations deal directly with
4375 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4376 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4377 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4378 they were not being used by the framework as there is no concept of a
4379 BIGNUM_METHOD and they could not be generalised to the new
4380 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4381 ENGINE_cpy() has been removed as it cannot be consistently defined in
4385 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4388 *) Change mkdef.pl to sort symbols that get the same entry number,
4389 and make sure the automatically generated functions ERR_load_*
4390 become part of libeay.num as well.
4393 *) New function SSL_renegotiate_pending(). This returns true once
4394 renegotiation has been requested (either SSL_renegotiate() call
4395 or HelloRequest/ClientHello receveived from the peer) and becomes
4396 false once a handshake has been completed.
4397 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4398 sends a HelloRequest, but does not ensure that a handshake takes
4399 place. SSL_renegotiate_pending() is useful for checking if the
4400 client has followed the request.)
4403 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4404 By default, clients may request session resumption even during
4405 renegotiation (if session ID contexts permit); with this option,
4406 session resumption is possible only in the first handshake.
4408 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4409 more bits available for options that should not be part of
4410 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4413 *) Add some demos for certificate and certificate request creation.
4416 *) Make maximum certificate chain size accepted from the peer application
4417 settable (SSL*_get/set_max_cert_list()), as proposed by
4418 "Douglas E. Engert" <deengert@anl.gov>.
4421 *) Add support for shared libraries for Unixware-7
4422 (Boyd Lynn Gerber <gerberb@zenez.com>).
4425 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4426 be done prior to destruction. Use this to unload error strings from
4427 ENGINEs that load their own error strings. NB: This adds two new API
4428 functions to "get" and "set" this destroy handler in an ENGINE.
4431 *) Alter all existing ENGINE implementations (except "openssl" and
4432 "openbsd") to dynamically instantiate their own error strings. This
4433 makes them more flexible to be built both as statically-linked ENGINEs
4434 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4435 Also, add stub code to each that makes building them as self-contained
4436 shared-libraries easier (see README.ENGINE).
4439 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4440 implementations into applications that are completely implemented in
4441 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4442 commands that can be used to configure what shared-library to load and
4443 to control aspects of the way it is handled. Also, made an update to
4444 the README.ENGINE file that brings its information up-to-date and
4445 provides some information and instructions on the "dynamic" ENGINE
4446 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4449 *) Make it possible to unload ranges of ERR strings with a new
4450 "ERR_unload_strings" function.
4453 *) Add a copy() function to EVP_MD.
4456 *) Make EVP_MD routines take a context pointer instead of just the
4457 md_data void pointer.
4460 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4461 that the digest can only process a single chunk of data
4462 (typically because it is provided by a piece of
4463 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4464 is only going to provide a single chunk of data, and hence the
4465 framework needn't accumulate the data for oneshot drivers.
4468 *) As with "ERR", make it possible to replace the underlying "ex_data"
4469 functions. This change also alters the storage and management of global
4470 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4471 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4472 index counters. The API functions that use this state have been changed
4473 to take a "class_index" rather than pointers to the class's local STACK
4474 and counter, and there is now an API function to dynamically create new
4475 classes. This centralisation allows us to (a) plug a lot of the
4476 thread-safety problems that existed, and (b) makes it possible to clean
4477 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4478 such data would previously have always leaked in application code and
4479 workarounds were in place to make the memory debugging turn a blind eye
4480 to it. Application code that doesn't use this new function will still
4481 leak as before, but their memory debugging output will announce it now
4482 rather than letting it slide.
4484 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4485 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4486 has a return value to indicate success or failure.
4489 *) Make it possible to replace the underlying "ERR" functions such that the
4490 global state (2 LHASH tables and 2 locks) is only used by the "default"
4491 implementation. This change also adds two functions to "get" and "set"
4492 the implementation prior to it being automatically set the first time
4493 any other ERR function takes place. Ie. an application can call "get",
4494 pass the return value to a module it has just loaded, and that module
4495 can call its own "set" function using that value. This means the
4496 module's "ERR" operations will use (and modify) the error state in the
4497 application and not in its own statically linked copy of OpenSSL code.
4500 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4501 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4502 the operation, and provides a more encapsulated way for external code
4503 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4504 to use these functions rather than manually incrementing the counts.
4506 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4509 *) Add EVP test program.
4512 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4515 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4516 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4517 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4518 These allow a CRL to be built without having to access X509_CRL fields
4519 directly. Modify 'ca' application to use new functions.
4522 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4523 bug workarounds. Rollback attack detection is a security feature.
4524 The problem will only arise on OpenSSL servers when TLSv1 is not
4525 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4526 Software authors not wanting to support TLSv1 will have special reasons
4527 for their choice and can explicitly enable this option.
4528 [Bodo Moeller, Lutz Jaenicke]
4530 *) Rationalise EVP so it can be extended: don't include a union of
4531 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4532 (similar to those existing for EVP_CIPHER_CTX).
4537 EVP_MD_CTX_init(&md); /* new function call */
4538 EVP_DigestInit(&md, EVP_sha1());
4539 EVP_DigestUpdate(&md, in, len);
4540 EVP_DigestFinal(&md, out, NULL);
4541 EVP_MD_CTX_cleanup(&md); /* new function call */
4545 *) Make DES key schedule conform to the usual scheme, as well as
4546 correcting its structure. This means that calls to DES functions
4547 now have to pass a pointer to a des_key_schedule instead of a
4548 plain des_key_schedule (which was actually always a pointer
4551 des_key_schedule ks;
4553 des_set_key_checked(..., &ks);
4554 des_ncbc_encrypt(..., &ks, ...);
4556 (Note that a later change renames 'des_...' into 'DES_...'.)
4559 *) Initial reduction of linker bloat: the use of some functions, such as
4560 PEM causes large amounts of unused functions to be linked in due to
4561 poor organisation. For example pem_all.c contains every PEM function
4562 which has a knock on effect of linking in large amounts of (unused)
4563 ASN1 code. Grouping together similar functions and splitting unrelated
4564 functions prevents this.
4567 *) Cleanup of EVP macros.
4570 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4571 correct _ecb suffix.
4574 *) Add initial OCSP responder support to ocsp application. The
4575 revocation information is handled using the text based index
4576 use by the ca application. The responder can either handle
4577 requests generated internally, supplied in files (for example
4578 via a CGI script) or using an internal minimal server.
4581 *) Add configuration choices to get zlib compression for TLS.
4584 *) Changes to Kerberos SSL for RFC 2712 compliance:
4585 1. Implemented real KerberosWrapper, instead of just using
4586 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4587 2. Implemented optional authenticator field of KerberosWrapper.
4589 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4590 and authenticator structs; see crypto/krb5/.
4592 Generalized Kerberos calls to support multiple Kerberos libraries.
4593 [Vern Staats <staatsvr@asc.hpc.mil>,
4594 Jeffrey Altman <jaltman@columbia.edu>
4595 via Richard Levitte]
4597 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4598 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4599 values for each of the key sizes rather than having just
4600 parameters (and 'speed' generating keys each time).
4603 *) Speed up EVP routines.
4606 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4607 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4608 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4609 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4611 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4612 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4613 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4616 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4618 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4621 *) Added the OS2-EMX target.
4622 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4624 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4625 to support NCONF routines in extension code. New function CONF_set_nconf()
4626 to allow functions which take an NCONF to also handle the old LHASH
4627 structure: this means that the old CONF compatible routines can be
4628 retained (in particular wrt extensions) without having to duplicate the
4629 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4632 *) Enhance the general user interface with mechanisms for inner control
4633 and with possibilities to have yes/no kind of prompts.
4636 *) Change all calls to low level digest routines in the library and
4637 applications to use EVP. Add missing calls to HMAC_cleanup() and
4638 don't assume HMAC_CTX can be copied using memcpy().
4639 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4641 *) Add the possibility to control engines through control names but with
4642 arbitrary arguments instead of just a string.
4643 Change the key loaders to take a UI_METHOD instead of a callback
4644 function pointer. NOTE: this breaks binary compatibility with earlier
4645 versions of OpenSSL [engine].
4646 Adapt the nCipher code for these new conditions and add a card insertion
4650 *) Enhance the general user interface with mechanisms to better support
4651 dialog box interfaces, application-defined prompts, the possibility
4652 to use defaults (for example default passwords from somewhere else)
4653 and interrupts/cancellations.
4656 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4657 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4660 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4661 tidy up some unnecessarily weird code in 'sk_new()').
4662 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4664 *) Change the key loading routines for ENGINEs to use the same kind
4665 callback (pem_password_cb) as all other routines that need this
4669 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4670 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4671 than this minimum value is recommended.
4674 *) New random seeder for OpenVMS, using the system process statistics
4675 that are easily reachable.
4678 *) Windows apparently can't transparently handle global
4679 variables defined in DLLs. Initialisations such as:
4681 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4683 wont compile. This is used by the any applications that need to
4684 declare their own ASN1 modules. This was fixed by adding the option
4685 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4686 needed for static libraries under Win32.
4689 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4690 setting of purpose and trust fields. New X509_STORE trust and
4691 purpose functions and tidy up setting in other SSL functions.
4694 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4695 structure. These are inherited by X509_STORE_CTX when it is
4696 initialised. This allows various defaults to be set in the
4697 X509_STORE structure (such as flags for CRL checking and custom
4698 purpose or trust settings) for functions which only use X509_STORE_CTX
4699 internally such as S/MIME.
4701 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4702 trust settings if they are not set in X509_STORE. This allows X509_STORE
4703 purposes and trust (in S/MIME for example) to override any set by default.
4705 Add command line options for CRL checking to smime, s_client and s_server
4709 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4710 are set then the CRL is looked up in the X509_STORE structure and
4711 its validity and signature checked, then if the certificate is found
4712 in the CRL the verify fails with a revoked error.
4714 Various new CRL related callbacks added to X509_STORE_CTX structure.
4716 Command line options added to 'verify' application to support this.
4718 This needs some additional work, such as being able to handle multiple
4719 CRLs with different times, extension based lookup (rather than just
4720 by subject name) and ultimately more complete V2 CRL extension
4724 *) Add a general user interface API (crypto/ui/). This is designed
4725 to replace things like des_read_password and friends (backward
4726 compatibility functions using this new API are provided).
4727 The purpose is to remove prompting functions from the DES code
4728 section as well as provide for prompting through dialog boxes in
4729 a window system and the like.
4732 *) Add "ex_data" support to ENGINE so implementations can add state at a
4733 per-structure level rather than having to store it globally.
4736 *) Make it possible for ENGINE structures to be copied when retrieved by
4737 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4738 This causes the "original" ENGINE structure to act like a template,
4739 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4740 operational state can be localised to each ENGINE structure, despite the
4741 fact they all share the same "methods". New ENGINE structures returned in
4742 this case have no functional references and the return value is the single
4743 structural reference. This matches the single structural reference returned
4744 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4748 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4749 needs to match any other type at all we need to manually clear the
4753 *) Changes to the "openssl engine" utility to include;
4754 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4755 about an ENGINE's available control commands.
4756 - executing control commands from command line arguments using the
4757 '-pre' and '-post' switches. '-post' is only used if '-t' is
4758 specified and the ENGINE is successfully initialised. The syntax for
4759 the individual commands are colon-separated, for example;
4760 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4763 *) New dynamic control command support for ENGINEs. ENGINEs can now
4764 declare their own commands (numbers), names (strings), descriptions,
4765 and input types for run-time discovery by calling applications. A
4766 subset of these commands are implicitly classed as "executable"
4767 depending on their input type, and only these can be invoked through
4768 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4769 can be based on user input, config files, etc). The distinction is
4770 that "executable" commands cannot return anything other than a boolean
4771 result and can only support numeric or string input, whereas some
4772 discoverable commands may only be for direct use through
4773 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4774 pointers, or other custom uses. The "executable" commands are to
4775 support parameterisations of ENGINE behaviour that can be
4776 unambiguously defined by ENGINEs and used consistently across any
4777 OpenSSL-based application. Commands have been added to all the
4778 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4779 control over shared-library paths without source code alterations.
4782 *) Changed all ENGINE implementations to dynamically allocate their
4783 ENGINEs rather than declaring them statically. Apart from this being
4784 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4785 this also allows the implementations to compile without using the
4786 internal engine_int.h header.
4789 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4790 'const' value. Any code that should be able to modify a RAND_METHOD
4791 should already have non-const pointers to it (ie. they should only
4792 modify their own ones).
4795 *) Made a variety of little tweaks to the ENGINE code.
4796 - "atalla" and "ubsec" string definitions were moved from header files
4797 to C code. "nuron" string definitions were placed in variables
4798 rather than hard-coded - allowing parameterisation of these values
4799 later on via ctrl() commands.
4800 - Removed unused "#if 0"'d code.
4801 - Fixed engine list iteration code so it uses ENGINE_free() to release
4802 structural references.
4803 - Constified the RAND_METHOD element of ENGINE structures.
4804 - Constified various get/set functions as appropriate and added
4805 missing functions (including a catch-all ENGINE_cpy that duplicates
4806 all ENGINE values onto a new ENGINE except reference counts/state).
4807 - Removed NULL parameter checks in get/set functions. Setting a method
4808 or function to NULL is a way of cancelling out a previously set
4809 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4810 and doesn't justify the extra error symbols and code.
4811 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4812 flags from engine_int.h to engine.h.
4813 - Changed prototypes for ENGINE handler functions (init(), finish(),
4814 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4817 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4818 to the algorithm using long division. The binary algorithm can be
4819 used only if the modulus is odd. On 32-bit systems, it is faster
4820 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4821 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4822 up to 450 bits. In 64-bit environments, the binary algorithm
4823 appears to be advantageous for much longer moduli; here we use it
4824 for moduli up to 2048 bits.
4827 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4828 could not support the combine flag in choice fields.
4831 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4832 extensions from a certificate request to the certificate.
4835 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4836 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4837 file: this allows the display of the certificate about to be
4838 signed to be customised, to allow certain fields to be included
4839 or excluded and extension details. The old system didn't display
4840 multicharacter strings properly, omitted fields not in the policy
4841 and couldn't display additional details such as extensions.
4844 *) Function EC_POINTs_mul for multiple scalar multiplication
4845 of an arbitrary number of elliptic curve points
4846 \sum scalars[i]*points[i],
4847 optionally including the generator defined for the EC_GROUP:
4848 scalar*generator + \sum scalars[i]*points[i].
4850 EC_POINT_mul is a simple wrapper function for the typical case
4851 that the point list has just one item (besides the optional
4855 *) First EC_METHODs for curves over GF(p):
4857 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4858 operations and provides various method functions that can also
4859 operate with faster implementations of modular arithmetic.
4861 EC_GFp_mont_method() reuses most functions that are part of
4862 EC_GFp_simple_method, but uses Montgomery arithmetic.
4864 [Bodo Moeller; point addition and point doubling
4865 implementation directly derived from source code provided by
4866 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4868 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4869 crypto/ec/ec_lib.c):
4871 Curves are EC_GROUP objects (with an optional group generator)
4872 based on EC_METHODs that are built into the library.
4874 Points are EC_POINT objects based on EC_GROUP objects.
4876 Most of the framework would be able to handle curves over arbitrary
4877 finite fields, but as there are no obvious types for fields other
4878 than GF(p), some functions are limited to that for now.
4881 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4882 that the file contains a complete HTTP response.
4885 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4886 change the def and num file printf format specifier from "%-40sXXX"
4887 to "%-39s XXX". The latter will always guarantee a space after the
4888 field while the former will cause them to run together if the field
4889 is 40 of more characters long.
4892 *) Constify the cipher and digest 'method' functions and structures
4893 and modify related functions to take constant EVP_MD and EVP_CIPHER
4897 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4898 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4901 *) Modify EVP_Digest*() routines so they now return values. Although the
4902 internal software routines can never fail additional hardware versions
4906 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4908 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4909 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4912 ERR_R_NESTED_ASN1_ERROR
4914 ERR_R_MISSING_ASN1_EOS
4915 were 4 .. 9, conflicting with
4916 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4918 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4919 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4921 Add new error code 'ERR_R_INTERNAL_ERROR'.
4924 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4928 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4929 sets the subject name for a new request or supersedes the
4930 subject name in a given request. Formats that can be parsed are
4931 'CN=Some Name, OU=myOU, C=IT'
4933 'CN=Some Name/OU=myOU/C=IT'.
4935 Add options '-batch' and '-verbose' to 'openssl req'.
4936 [Massimiliano Pala <madwolf@hackmasters.net>]
4938 *) Introduce the possibility to access global variables through
4939 functions on platform were that's the best way to handle exporting
4940 global variables in shared libraries. To enable this functionality,
4941 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4942 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4943 is normally done by Configure or something similar).
4945 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4946 in the source file (foo.c) like this:
4948 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4949 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4951 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4952 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4954 OPENSSL_DECLARE_GLOBAL(int,foo);
4955 #define foo OPENSSL_GLOBAL_REF(foo)
4956 OPENSSL_DECLARE_GLOBAL(double,bar);
4957 #define bar OPENSSL_GLOBAL_REF(bar)
4959 The #defines are very important, and therefore so is including the
4960 header file everywhere where the defined globals are used.
4962 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4963 of ASN.1 items, but that structure is a bit different.
4965 The largest change is in util/mkdef.pl which has been enhanced with
4966 better and easier to understand logic to choose which symbols should
4967 go into the Windows .def files as well as a number of fixes and code
4968 cleanup (among others, algorithm keywords are now sorted
4969 lexicographically to avoid constant rewrites).
4972 *) In BN_div() keep a copy of the sign of 'num' before writing the
4973 result to 'rm' because if rm==num the value will be overwritten
4974 and produce the wrong result if 'num' is negative: this caused
4975 problems with BN_mod() and BN_nnmod().
4978 *) Function OCSP_request_verify(). This checks the signature on an
4979 OCSP request and verifies the signer certificate. The signer
4980 certificate is just checked for a generic purpose and OCSP request
4984 *) Add OCSP_check_validity() function to check the validity of OCSP
4985 responses. OCSP responses are prepared in real time and may only
4986 be a few seconds old. Simply checking that the current time lies
4987 between thisUpdate and nextUpdate max reject otherwise valid responses
4988 caused by either OCSP responder or client clock inaccuracy. Instead
4989 we allow thisUpdate and nextUpdate to fall within a certain period of
4990 the current time. The age of the response can also optionally be
4991 checked. Two new options -validity_period and -status_age added to
4995 *) If signature or public key algorithm is unrecognized print out its
4996 OID rather that just UNKNOWN.
4999 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5000 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5001 ID to be generated from the issuer certificate alone which can then be
5002 passed to OCSP_id_issuer_cmp().
5005 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5006 ASN1 modules to export functions returning ASN1_ITEM pointers
5007 instead of the ASN1_ITEM structures themselves. This adds several
5008 new macros which allow the underlying ASN1 function/structure to
5009 be accessed transparently. As a result code should not use ASN1_ITEM
5010 references directly (such as &X509_it) but instead use the relevant
5011 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5012 use of the new ASN1 code on platforms where exporting structures
5013 is problematical (for example in shared libraries) but exporting
5014 functions returning pointers to structures is not.
5017 *) Add support for overriding the generation of SSL/TLS session IDs.
5018 These callbacks can be registered either in an SSL_CTX or per SSL.
5019 The purpose of this is to allow applications to control, if they wish,
5020 the arbitrary values chosen for use as session IDs, particularly as it
5021 can be useful for session caching in multiple-server environments. A
5022 command-line switch for testing this (and any client code that wishes
5023 to use such a feature) has been added to "s_server".
5024 [Geoff Thorpe, Lutz Jaenicke]
5026 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5027 of the form '#if defined(...) || defined(...) || ...' and
5028 '#if !defined(...) && !defined(...) && ...'. This also avoids
5029 the growing number of special cases it was previously handling.
5032 *) Make all configuration macros available for application by making
5033 sure they are available in opensslconf.h, by giving them names starting
5034 with "OPENSSL_" to avoid conflicts with other packages and by making
5035 sure e_os2.h will cover all platform-specific cases together with
5037 Additionally, it is now possible to define configuration/platform-
5038 specific names (called "system identities"). In the C code, these
5039 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5040 macro with the name beginning with "OPENSSL_SYS_", which is determined
5041 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5045 *) New option -set_serial to 'req' and 'x509' this allows the serial
5046 number to use to be specified on the command line. Previously self
5047 signed certificates were hard coded with serial number 0 and the
5048 CA options of 'x509' had to use a serial number in a file which was
5052 *) New options to 'ca' utility to support V2 CRL entry extensions.
5053 Currently CRL reason, invalidity date and hold instruction are
5054 supported. Add new CRL extensions to V3 code and some new objects.
5057 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5058 disable standard block padding (aka PKCS#5 padding) in the EVP
5059 API, which was previously mandatory. This means that the data is
5060 not padded in any way and so the total length much be a multiple
5061 of the block size, otherwise an error occurs.
5064 *) Initial (incomplete) OCSP SSL support.
5067 *) New function OCSP_parse_url(). This splits up a URL into its host,
5068 port and path components: primarily to parse OCSP URLs. New -url
5069 option to ocsp utility.
5072 *) New nonce behavior. The return value of OCSP_check_nonce() now
5073 reflects the various checks performed. Applications can decide
5074 whether to tolerate certain situations such as an absent nonce
5075 in a response when one was present in a request: the ocsp application
5076 just prints out a warning. New function OCSP_add1_basic_nonce()
5077 this is to allow responders to include a nonce in a response even if
5078 the request is nonce-less.
5081 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5082 skipped when using openssl x509 multiple times on a single input file,
5083 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5086 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5087 set string type: to handle setting ASN1_TIME structures. Fix ca
5088 utility to correctly initialize revocation date of CRLs.
5091 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5092 the clients preferred ciphersuites and rather use its own preferences.
5093 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5094 Internet Explorer by ensuring unchanged hash method during stepup.
5095 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5098 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5099 to aes and add a new 'exist' option to print out symbols that don't
5103 *) Additional options to ocsp utility to allow flags to be set and
5104 additional certificates supplied.
5107 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5108 OCSP client a number of certificate to only verify the response
5112 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5113 handle the new API. Currently only ECB, CBC modes supported. Add new
5116 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5117 Encryption Standard (AES) Ciphersuites for Transport Layer
5118 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5119 not enabled by default and were not part of the "ALL" ciphersuite
5120 alias because they were not yet official; they could be
5121 explicitly requested by specifying the "AESdraft" ciphersuite
5122 group alias. In the final release of OpenSSL 0.9.7, the group
5123 alias is called "AES" and is part of "ALL".)
5124 [Ben Laurie, Steve Henson, Bodo Moeller]
5126 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5127 request to response.
5130 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5131 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5132 extract information from a certificate request. OCSP_response_create()
5133 creates a response and optionally adds a basic response structure.
5134 OCSP_basic_add1_status() adds a complete single response to a basic
5135 response and returns the OCSP_SINGLERESP structure just added (to allow
5136 extensions to be included for example). OCSP_basic_add1_cert() adds a
5137 certificate to a basic response and OCSP_basic_sign() signs a basic
5138 response with various flags. New helper functions ASN1_TIME_check()
5139 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5140 (converts ASN1_TIME to GeneralizedTime).
5143 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5144 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5145 structure from a certificate. X509_pubkey_digest() digests the public_key
5146 contents: this is used in various key identifiers.
5149 *) Make sk_sort() tolerate a NULL argument.
5150 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5152 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5153 passed by the function are trusted implicitly. If any of them signed the
5154 response then it is assumed to be valid and is not verified.
5157 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5158 to data. This was previously part of the PKCS7 ASN1 code. This
5159 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5160 [Steve Henson, reported by Kenneth R. Robinette
5161 <support@securenetterm.com>]
5163 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5164 routines: without these tracing memory leaks is very painful.
5165 Fix leaks in PKCS12 and PKCS7 routines.
5168 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5169 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5170 effectively meant GeneralizedTime would never be used. Now it
5171 is initialised to -1 but X509_time_adj() now has to check the value
5172 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5173 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5174 [Steve Henson, reported by Kenneth R. Robinette
5175 <support@securenetterm.com>]
5177 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5178 result in a zero length in the ASN1_INTEGER structure which was
5179 not consistent with the structure when d2i_ASN1_INTEGER() was used
5180 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5181 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5182 where it did not print out a minus for negative ASN1_INTEGER.
5185 *) Add summary printout to ocsp utility. The various functions which
5186 convert status values to strings have been renamed to:
5187 OCSP_response_status_str(), OCSP_cert_status_str() and
5188 OCSP_crl_reason_str() and are no longer static. New options
5189 to verify nonce values and to disable verification. OCSP response
5190 printout format cleaned up.
5193 *) Add additional OCSP certificate checks. These are those specified
5194 in RFC2560. This consists of two separate checks: the CA of the
5195 certificate being checked must either be the OCSP signer certificate
5196 or the issuer of the OCSP signer certificate. In the latter case the
5197 OCSP signer certificate must contain the OCSP signing extended key
5198 usage. This check is performed by attempting to match the OCSP
5199 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5200 in the OCSP_CERTID structures of the response.
5203 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5204 and related routines. This uses the standard OpenSSL certificate
5205 verify routines to perform initial checks (just CA validity) and
5206 to obtain the certificate chain. Then additional checks will be
5207 performed on the chain. Currently the root CA is checked to see
5208 if it is explicitly trusted for OCSP signing. This is used to set
5209 a root CA as a global signing root: that is any certificate that
5210 chains to that CA is an acceptable OCSP signing certificate.
5213 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5214 extensions from a separate configuration file.
5215 As when reading extensions from the main configuration file,
5216 the '-extensions ...' option may be used for specifying the
5218 [Massimiliano Pala <madwolf@comune.modena.it>]
5220 *) New OCSP utility. Allows OCSP requests to be generated or
5221 read. The request can be sent to a responder and the output
5222 parsed, outputed or printed in text form. Not complete yet:
5223 still needs to check the OCSP response validity.
5226 *) New subcommands for 'openssl ca':
5227 'openssl ca -status <serial>' prints the status of the cert with
5228 the given serial number (according to the index file).
5229 'openssl ca -updatedb' updates the expiry status of certificates
5231 [Massimiliano Pala <madwolf@comune.modena.it>]
5233 *) New '-newreq-nodes' command option to CA.pl. This is like
5234 '-newreq', but calls 'openssl req' with the '-nodes' option
5235 so that the resulting key is not encrypted.
5236 [Damien Miller <djm@mindrot.org>]
5238 *) New configuration for the GNU Hurd.
5239 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5241 *) Initial code to implement OCSP basic response verify. This
5242 is currently incomplete. Currently just finds the signer's
5243 certificate and verifies the signature on the response.
5246 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5247 value of OPENSSLDIR. This is available via the new '-d' option
5248 to 'openssl version', and is also included in 'openssl version -a'.
5251 *) Allowing defining memory allocation callbacks that will be given
5252 file name and line number information in additional arguments
5253 (a const char* and an int). The basic functionality remains, as
5254 well as the original possibility to just replace malloc(),
5255 realloc() and free() by functions that do not know about these
5256 additional arguments. To register and find out the current
5257 settings for extended allocation functions, the following
5258 functions are provided:
5260 CRYPTO_set_mem_ex_functions
5261 CRYPTO_set_locked_mem_ex_functions
5262 CRYPTO_get_mem_ex_functions
5263 CRYPTO_get_locked_mem_ex_functions
5265 These work the same way as CRYPTO_set_mem_functions and friends.
5266 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5267 extended allocation function is enabled.
5268 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5269 a conventional allocation function is enabled.
5270 [Richard Levitte, Bodo Moeller]
5272 *) Finish off removing the remaining LHASH function pointer casts.
5273 There should no longer be any prototype-casting required when using
5274 the LHASH abstraction, and any casts that remain are "bugs". See
5275 the callback types and macros at the head of lhash.h for details
5276 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5279 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5280 If /dev/[u]random devices are not available or do not return enough
5281 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5283 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5284 /etc/entropy will be queried once each in this sequence, quering stops
5285 when enough entropy was collected without querying more sockets.
5288 *) Change the Unix RAND_poll() variant to be able to poll several
5289 random devices, as specified by DEVRANDOM, until a sufficient amount
5290 of data has been collected. We spend at most 10 ms on each file
5291 (select timeout) and read in non-blocking mode. DEVRANDOM now
5292 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5293 (previously it was just the string "/dev/urandom"), so on typical
5294 platforms the 10 ms delay will never occur.
5295 Also separate out the Unix variant to its own file, rand_unix.c.
5296 For VMS, there's a currently-empty rand_vms.c.
5299 *) Move OCSP client related routines to ocsp_cl.c. These
5300 provide utility functions which an application needing
5301 to issue a request to an OCSP responder and analyse the
5302 response will typically need: as opposed to those which an
5303 OCSP responder itself would need which will be added later.
5305 OCSP_request_sign() signs an OCSP request with an API similar
5306 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5307 response. OCSP_response_get1_basic() extracts basic response
5308 from response. OCSP_resp_find_status(): finds and extracts status
5309 information from an OCSP_CERTID structure (which will be created
5310 when the request structure is built). These are built from lower
5311 level functions which work on OCSP_SINGLERESP structures but
5312 wont normally be used unless the application wishes to examine
5313 extensions in the OCSP response for example.
5315 Replace nonce routines with a pair of functions.
5316 OCSP_request_add1_nonce() adds a nonce value and optionally
5317 generates a random value. OCSP_check_nonce() checks the
5318 validity of the nonce in an OCSP response.
5321 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5322 This doesn't copy the supplied OCSP_CERTID and avoids the
5323 need to free up the newly created id. Change return type
5324 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5325 This can then be used to add extensions to the request.
5326 Deleted OCSP_request_new(), since most of its functionality
5327 is now in OCSP_REQUEST_new() (and the case insensitive name
5328 clash) apart from the ability to set the request name which
5329 will be added elsewhere.
5332 *) Update OCSP API. Remove obsolete extensions argument from
5333 various functions. Extensions are now handled using the new
5334 OCSP extension code. New simple OCSP HTTP function which
5335 can be used to send requests and parse the response.
5338 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5339 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5340 uses the special reorder version of SET OF to sort the attributes
5341 and reorder them to match the encoded order. This resolves a long
5342 standing problem: a verify on a PKCS7 structure just after signing
5343 it used to fail because the attribute order did not match the
5344 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5345 it uses the received order. This is necessary to tolerate some broken
5346 software that does not order SET OF. This is handled by encoding
5347 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5348 to produce the required SET OF.
5351 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5352 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5353 files to get correct declarations of the ASN.1 item variables.
5356 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5357 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5358 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5359 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5360 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5361 ASN1_ITEM and no wrapper functions.
5364 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5365 replace the old function pointer based I/O routines. Change most of
5366 the *_d2i_bio() and *_d2i_fp() functions to use these.
5369 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5370 lines, recognice more "algorithms" that can be deselected, and make
5371 it complain about algorithm deselection that isn't recognised.
5374 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5375 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5376 to use new functions. Add NO_ASN1_OLD which can be set to remove
5377 some old style ASN1 functions: this can be used to determine if old
5378 code will still work when these eventually go away.
5381 *) New extension functions for OCSP structures, these follow the
5382 same conventions as certificates and CRLs.
5385 *) New function X509V3_add1_i2d(). This automatically encodes and
5386 adds an extension. Its behaviour can be customised with various
5387 flags to append, replace or delete. Various wrappers added for
5388 certifcates and CRLs.
5391 *) Fix to avoid calling the underlying ASN1 print routine when
5392 an extension cannot be parsed. Correct a typo in the
5393 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5396 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5397 entries for variables.
5400 *) Add functionality to apps/openssl.c for detecting locking
5401 problems: As the program is single-threaded, all we have
5402 to do is register a locking callback using an array for
5403 storing which locks are currently held by the program.
5406 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5407 SSL_get_ex_data_X509_STORE_idx(), which is used in
5408 ssl_verify_cert_chain() and thus can be called at any time
5409 during TLS/SSL handshakes so that thread-safety is essential.
5410 Unfortunately, the ex_data design is not at all suited
5411 for multi-threaded use, so it probably should be abolished.
5414 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5415 [Broadcom, tweaked and integrated by Geoff Thorpe]
5417 *) Move common extension printing code to new function
5418 X509V3_print_extensions(). Reorganise OCSP print routines and
5419 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5422 *) New function X509_signature_print() to remove duplication in some
5426 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5427 set (this was treated exactly the same as SET OF previously). This
5428 is used to reorder the STACK representing the structure to match the
5429 encoding. This will be used to get round a problem where a PKCS7
5430 structure which was signed could not be verified because the STACK
5431 order did not reflect the encoded order.
5434 *) Reimplement the OCSP ASN1 module using the new code.
5437 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5438 for its ASN1 operations. The old style function pointers still exist
5439 for now but they will eventually go away.
5442 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5443 completely replaces the old ASN1 functionality with a table driven
5444 encoder and decoder which interprets an ASN1_ITEM structure describing
5445 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5446 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5447 has also been converted to the new form.
5450 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5451 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5452 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5453 for negative moduli.
5456 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5457 of not touching the result's sign bit.
5460 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5464 *) Changed the LHASH code to use prototypes for callbacks, and created
5465 macros to declare and implement thin (optionally static) functions
5466 that provide type-safety and avoid function pointer casting for the
5467 type-specific callbacks.
5470 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5472 [Veers Staats <staatsvr@asc.hpc.mil>,
5473 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5475 *) Reformat the FAQ so the different questions and answers can be divided
5476 in sections depending on the subject.
5479 *) Have the zlib compression code load ZLIB.DLL dynamically under
5483 *) New function BN_mod_sqrt for computing square roots modulo a prime
5484 (using the probabilistic Tonelli-Shanks algorithm unless
5485 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5486 be handled deterministically).
5487 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5489 *) Make BN_mod_inverse faster by explicitly handling small quotients
5490 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5491 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5494 *) New function BN_kronecker.
5497 *) Fix BN_gcd so that it works on negative inputs; the result is
5498 positive unless both parameters are zero.
5499 Previously something reasonably close to an infinite loop was
5500 possible because numbers could be growing instead of shrinking
5501 in the implementation of Euclid's algorithm.
5504 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5505 sign of the number in question.
5507 Fix BN_is_word(a,w) to work correctly for w == 0.
5509 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5510 because its test if the absolute value of 'a' equals 'w'.
5511 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5512 it exists mostly for use in the implementations of BN_is_zero(),
5513 BN_is_one(), and BN_is_word().
5516 *) New function BN_swap.
5519 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5520 the exponentiation functions are more likely to produce reasonable
5521 results on negative inputs.
5524 *) Change BN_mod_mul so that the result is always non-negative.
5525 Previously, it could be negative if one of the factors was negative;
5526 I don't think anyone really wanted that behaviour.
5529 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5530 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5531 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5532 and add new functions:
5541 BN_mod_lshift1_quick
5545 These functions always generate non-negative results.
5547 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5548 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5550 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5551 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5552 be reduced modulo m.
5553 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5556 The following entry accidentily appeared in the CHANGES file
5557 distributed with OpenSSL 0.9.7. The modifications described in
5558 it do *not* apply to OpenSSL 0.9.7.
5560 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5561 was actually never needed) and in BN_mul(). The removal in BN_mul()
5562 required a small change in bn_mul_part_recursive() and the addition
5563 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5564 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5565 bn_sub_words() and bn_add_words() except they take arrays with
5570 *) In 'openssl passwd', verify passwords read from the terminal
5571 unless the '-salt' option is used (which usually means that
5572 verification would just waste user's time since the resulting
5573 hash is going to be compared with some given password hash)
5574 or the new '-noverify' option is used.
5576 This is an incompatible change, but it does not affect
5577 non-interactive use of 'openssl passwd' (passwords on the command
5578 line, '-stdin' option, '-in ...' option) and thus should not
5582 *) Remove all references to RSAref, since there's no more need for it.
5585 *) Make DSO load along a path given through an environment variable
5586 (SHLIB_PATH) with shl_load().
5589 *) Constify the ENGINE code as a result of BIGNUM constification.
5590 Also constify the RSA code and most things related to it. In a
5591 few places, most notable in the depth of the ASN.1 code, ugly
5592 casts back to non-const were required (to be solved at a later
5596 *) Make it so the openssl application has all engines loaded by default.
5599 *) Constify the BIGNUM routines a little more.
5602 *) Add the following functions:
5604 ENGINE_load_cswift()
5606 ENGINE_load_atalla()
5608 ENGINE_load_builtin_engines()
5610 That way, an application can itself choose if external engines that
5611 are built-in in OpenSSL shall ever be used or not. The benefit is
5612 that applications won't have to be linked with libdl or other dso
5613 libraries unless it's really needed.
5615 Changed 'openssl engine' to load all engines on demand.
5616 Changed the engine header files to avoid the duplication of some
5617 declarations (they differed!).
5620 *) 'openssl engine' can now list capabilities.
5623 *) Better error reporting in 'openssl engine'.
5626 *) Never call load_dh_param(NULL) in s_server.
5629 *) Add engine application. It can currently list engines by name and
5630 identity, and test if they are actually available.
5633 *) Improve RPM specification file by forcing symbolic linking and making
5634 sure the installed documentation is also owned by root.root.
5635 [Damien Miller <djm@mindrot.org>]
5637 *) Give the OpenSSL applications more possibilities to make use of
5638 keys (public as well as private) handled by engines.
5641 *) Add OCSP code that comes from CertCo.
5644 *) Add VMS support for the Rijndael code.
5647 *) Added untested support for Nuron crypto accelerator.
5650 *) Add support for external cryptographic devices. This code was
5651 previously distributed separately as the "engine" branch.
5652 [Geoff Thorpe, Richard Levitte]
5654 *) Rework the filename-translation in the DSO code. It is now possible to
5655 have far greater control over how a "name" is turned into a filename
5656 depending on the operating environment and any oddities about the
5657 different shared library filenames on each system.
5660 *) Support threads on FreeBSD-elf in Configure.
5663 *) Fix for SHA1 assembly problem with MASM: it produces
5664 warnings about corrupt line number information when assembling
5665 with debugging information. This is caused by the overlapping
5667 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5670 NCONF_get_number() has no error checking at all. As a replacement,
5671 NCONF_get_number_e() is defined (_e for "error checking") and is
5672 promoted strongly. The old NCONF_get_number is kept around for
5673 binary backward compatibility.
5674 Make it possible for methods to load from something other than a BIO,
5675 by providing a function pointer that is given a name instead of a BIO.
5676 For example, this could be used to load configuration data from an
5680 *) Fix for non blocking accept BIOs. Added new I/O special reason
5681 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5682 with non blocking I/O was not possible because no retry code was
5683 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5687 *) Added the beginnings of Rijndael support.
5690 *) Fix for bug in DirectoryString mask setting. Add support for
5691 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5692 to allow certificate printing to more controllable, additional
5693 'certopt' option to 'x509' to allow new printing options to be
5697 *) Clean old EAY MD5 hack from e_os.h.
5700 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5702 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5703 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5704 [Joe Orton, Steve Henson]
5706 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5708 *) Fix additional bug revealed by the NISCC test suite:
5710 Stop bug triggering large recursion when presented with
5711 certain ASN.1 tags (CVE-2003-0851)
5714 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5716 *) Fix various bugs revealed by running the NISCC test suite:
5718 Stop out of bounds reads in the ASN1 code when presented with
5719 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5721 If verify callback ignores invalid public key errors don't try to check
5722 certificate signature with the NULL public key.
5726 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5727 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5731 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5732 extra data after the compression methods not only for TLS 1.0
5733 but also for SSL 3.0 (as required by the specification).
5734 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5736 *) Change X509_certificate_type() to mark the key as exported/exportable
5737 when it's 512 *bits* long, not 512 bytes.
5740 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5742 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5743 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5744 a protocol version number mismatch like a decryption error
5745 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5748 *) Turn on RSA blinding by default in the default implementation
5749 to avoid a timing attack. Applications that don't want it can call
5750 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5751 They would be ill-advised to do so in most cases.
5752 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5754 *) Change RSA blinding code so that it works when the PRNG is not
5755 seeded (in this case, the secret RSA exponent is abused as
5756 an unpredictable seed -- if it is not unpredictable, there
5757 is no point in blinding anyway). Make RSA blinding thread-safe
5758 by remembering the creator's thread ID in rsa->blinding and
5759 having all other threads use local one-time blinding factors
5760 (this requires more computation than sharing rsa->blinding, but
5761 avoids excessive locking; and if an RSA object is not shared
5762 between threads, blinding will still be very fast).
5765 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5767 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5768 via timing by performing a MAC computation even if incorrrect
5769 block cipher padding has been found. This is a countermeasure
5770 against active attacks where the attacker has to distinguish
5771 between bad padding and a MAC verification error. (CVE-2003-0078)
5773 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5774 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5775 Martin Vuagnoux (EPFL, Ilion)]
5777 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5779 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5780 memory from it's contents. This is done with a counter that will
5781 place alternating values in each byte. This can be used to solve
5782 two issues: 1) the removal of calls to memset() by highly optimizing
5783 compilers, and 2) cleansing with other values than 0, since those can
5784 be read through on certain media, for example a swap space on disk.
5787 *) Bugfix: client side session caching did not work with external caching,
5788 because the session->cipher setting was not restored when reloading
5789 from the external cache. This problem was masked, when
5790 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5791 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5794 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5795 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5796 [Zeev Lieber <zeev-l@yahoo.com>]
5798 *) Undo an undocumented change introduced in 0.9.6e which caused
5799 repeated calls to OpenSSL_add_all_ciphers() and
5800 OpenSSL_add_all_digests() to be ignored, even after calling
5804 *) Change the default configuration reader to deal with last line not
5805 being properly terminated.
5808 *) Change X509_NAME_cmp() so it applies the special rules on handling
5809 DN values that are of type PrintableString, as well as RDNs of type
5810 emailAddress where the value has the type ia5String.
5811 [stefank@valicert.com via Richard Levitte]
5813 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5814 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5815 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5816 the bitwise-OR of the two for use by the majority of applications
5817 wanting this behaviour, and update the docs. The documented
5818 behaviour and actual behaviour were inconsistent and had been
5819 changing anyway, so this is more a bug-fix than a behavioural
5821 [Geoff Thorpe, diagnosed by Nadav Har'El]
5823 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5824 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5827 *) Fix initialization code race conditions in
5828 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5829 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5830 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5831 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5832 ssl2_get_cipher_by_char(),
5833 ssl3_get_cipher_by_char().
5834 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5836 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5837 the cached sessions are flushed, as the remove_cb() might use ex_data
5838 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5839 (see [openssl.org #212]).
5840 [Geoff Thorpe, Lutz Jaenicke]
5842 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5843 length, instead of the encoding length to d2i_ASN1_OBJECT.
5846 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5848 *) [In 0.9.6g-engine release:]
5849 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5850 [Lynn Gazis <lgazis@rainbow.com>]
5852 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5854 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5855 and get fix the header length calculation.
5856 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5857 Alon Kantor <alonk@checkpoint.com> (and others),
5860 *) Use proper error handling instead of 'assertions' in buffer
5861 overflow checks added in 0.9.6e. This prevents DoS (the
5862 assertions could call abort()).
5863 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5865 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5867 *) Add various sanity checks to asn1_get_length() to reject
5868 the ASN1 length bytes if they exceed sizeof(long), will appear
5869 negative or the content length exceeds the length of the
5871 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5873 *) Fix cipher selection routines: ciphers without encryption had no flags
5874 for the cipher strength set and where therefore not handled correctly
5875 by the selection routines (PR #130).
5878 *) Fix EVP_dsa_sha macro.
5882 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5883 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5884 that was added in OpenSSL 0.9.6d.
5886 As the countermeasure turned out to be incompatible with some
5887 broken SSL implementations, the new option is part of SSL_OP_ALL.
5888 SSL_OP_ALL is usually employed when compatibility with weird SSL
5889 implementations is desired (e.g. '-bugs' option to 's_client' and
5890 's_server'), so the new option is automatically set in many
5894 *) Changes in security patch:
5896 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5897 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5898 Air Force Materiel Command, USAF, under agreement number
5901 *) Add various sanity checks to asn1_get_length() to reject
5902 the ASN1 length bytes if they exceed sizeof(long), will appear
5903 negative or the content length exceeds the length of the
5904 supplied buffer. (CVE-2002-0659)
5905 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5907 *) Assertions for various potential buffer overflows, not known to
5909 [Ben Laurie (CHATS)]
5911 *) Various temporary buffers to hold ASCII versions of integers were
5912 too small for 64 bit platforms. (CVE-2002-0655)
5913 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5915 *) Remote buffer overflow in SSL3 protocol - an attacker could
5916 supply an oversized session ID to a client. (CVE-2002-0656)
5917 [Ben Laurie (CHATS)]
5919 *) Remote buffer overflow in SSL2 protocol - an attacker could
5920 supply an oversized client master key. (CVE-2002-0656)
5921 [Ben Laurie (CHATS)]
5923 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5925 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5926 encoded as NULL) with id-dsa-with-sha1.
5927 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5929 *) Check various X509_...() return values in apps/req.c.
5930 [Nils Larsch <nla@trustcenter.de>]
5932 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5933 an end-of-file condition would erronously be flagged, when the CRLF
5934 was just at the end of a processed block. The bug was discovered when
5935 processing data through a buffering memory BIO handing the data to a
5936 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5937 <ptsekov@syntrex.com> and Nedelcho Stanev.
5940 *) Implement a countermeasure against a vulnerability recently found
5941 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5942 before application data chunks to avoid the use of known IVs
5943 with data potentially chosen by the attacker.
5946 *) Fix length checks in ssl3_get_client_hello().
5949 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5950 to prevent ssl3_read_internal() from incorrectly assuming that
5951 ssl3_read_bytes() found application data while handshake
5952 processing was enabled when in fact s->s3->in_read_app_data was
5953 merely automatically cleared during the initial handshake.
5954 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5956 *) Fix object definitions for Private and Enterprise: they were not
5957 recognized in their shortname (=lowercase) representation. Extend
5958 obj_dat.pl to issue an error when using undefined keywords instead
5959 of silently ignoring the problem (Svenning Sorensen
5960 <sss@sss.dnsalias.net>).
5963 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5964 generators, i.e. generators other than 2 and 5. (Previously, the
5965 code did not properly initialise the 'add' and 'rem' values to
5966 BN_generate_prime().)
5968 In the new general case, we do not insist that 'generator' is
5969 actually a primitive root: This requirement is rather pointless;
5970 a generator of the order-q subgroup is just as good, if not
5974 *) Map new X509 verification errors to alerts. Discovered and submitted by
5975 Tom Wu <tom@arcot.com>.
5978 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5979 returning non-zero before the data has been completely received
5980 when using non-blocking I/O.
5981 [Bodo Moeller; problem pointed out by John Hughes]
5983 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5984 [Ben Laurie, Lutz Jaenicke]
5986 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5987 Yoram Zahavi <YoramZ@gilian.com>).
5990 *) Add information about CygWin 1.3 and on, and preserve proper
5991 configuration for the versions before that.
5992 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5994 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5995 check whether we deal with a copy of a session and do not delete from
5996 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5997 <izhar@checkpoint.com>.
6000 *) Do not store session data into the internal session cache, if it
6001 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6002 flag is set). Proposed by Aslam <aslam@funk.com>.
6005 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6009 *) [In 0.9.6d-engine release:]
6010 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6011 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6013 *) Add the configuration target linux-s390x.
6014 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6016 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6017 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6018 variable as an indication that a ClientHello message has been
6019 received. As the flag value will be lost between multiple
6020 invocations of ssl3_accept when using non-blocking I/O, the
6021 function may not be aware that a handshake has actually taken
6022 place, thus preventing a new session from being added to the
6025 To avoid this problem, we now set s->new_session to 2 instead of
6026 using a local variable.
6027 [Lutz Jaenicke, Bodo Moeller]
6029 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6030 if the SSL_R_LENGTH_MISMATCH error is detected.
6031 [Geoff Thorpe, Bodo Moeller]
6033 *) New 'shared_ldflag' column in Configure platform table.
6036 *) Fix EVP_CIPHER_mode macro.
6037 ["Dan S. Camper" <dan@bti.net>]
6039 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6040 type, we must throw them away by setting rr->length to 0.
6041 [D P Chang <dpc@qualys.com>]
6043 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6045 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6046 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6047 worked incorrectly for those cases where range = 10..._2 and
6048 3*range is two bits longer than range.)
6051 *) Only add signing time to PKCS7 structures if it is not already
6055 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6056 OBJ_ld_ce should be OBJ_id_ce.
6057 Also some ip-pda OIDs in crypto/objects/objects.txt were
6058 incorrect (cf. RFC 3039).
6059 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6061 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6062 returns early because it has nothing to do.
6063 [Andy Schneider <andy.schneider@bjss.co.uk>]
6065 *) [In 0.9.6c-engine release:]
6066 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6067 [Andy Schneider <andy.schneider@bjss.co.uk>]
6069 *) [In 0.9.6c-engine release:]
6070 Add support for Cryptographic Appliance's keyserver technology.
6071 (Use engine 'keyclient')
6072 [Cryptographic Appliances and Geoff Thorpe]
6074 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6075 is called via tools/c89.sh because arguments have to be
6076 rearranged (all '-L' options must appear before the first object
6078 [Richard Shapiro <rshapiro@abinitio.com>]
6080 *) [In 0.9.6c-engine release:]
6081 Add support for Broadcom crypto accelerator cards, backported
6083 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6085 *) [In 0.9.6c-engine release:]
6086 Add support for SureWare crypto accelerator cards from
6087 Baltimore Technologies. (Use engine 'sureware')
6088 [Baltimore Technologies and Mark Cox]
6090 *) [In 0.9.6c-engine release:]
6091 Add support for crypto accelerator cards from Accelerated
6092 Encryption Processing, www.aep.ie. (Use engine 'aep')
6093 [AEP Inc. and Mark Cox]
6095 *) Add a configuration entry for gcc on UnixWare.
6096 [Gary Benson <gbenson@redhat.com>]
6098 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6099 messages are stored in a single piece (fixed-length part and
6100 variable-length part combined) and fix various bugs found on the way.
6103 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6104 instead. BIO_gethostbyname() does not know what timeouts are
6105 appropriate, so entries would stay in cache even when they have
6107 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6109 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6110 faced with a pathologically small ClientHello fragment that does
6111 not contain client_version: Instead of aborting with an error,
6112 simply choose the highest available protocol version (i.e.,
6113 TLS 1.0 unless it is disabled). In practice, ClientHello
6114 messages are never sent like this, but this change gives us
6115 strictly correct behaviour at least for TLS.
6118 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6119 never resets s->method to s->ctx->method when called from within
6120 one of the SSL handshake functions.
6121 [Bodo Moeller; problem pointed out by Niko Baric]
6123 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6124 (sent using the client's version number) if client_version is
6125 smaller than the protocol version in use. Also change
6126 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6127 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6128 the client will at least see that alert.
6131 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6135 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6136 client receives HelloRequest while in a handshake.
6137 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6139 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6140 should end in 'break', not 'goto end' which circuments various
6141 cleanups done in state SSL_ST_OK. But session related stuff
6142 must be disabled for SSL_ST_OK in the case that we just sent a
6145 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6146 before just sending a HelloRequest.
6147 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6149 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6150 reveal whether illegal block cipher padding was found or a MAC
6151 verification error occured. (Neither SSLerr() codes nor alerts
6152 are directly visible to potential attackers, but the information
6153 may leak via logfiles.)
6155 Similar changes are not required for the SSL 2.0 implementation
6156 because the number of padding bytes is sent in clear for SSL 2.0,
6157 and the extra bytes are just ignored. However ssl/s2_pkt.c
6158 failed to verify that the purported number of padding bytes is in
6162 *) Add OpenUNIX-8 support including shared libraries
6163 (Boyd Lynn Gerber <gerberb@zenez.com>).
6166 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6167 'wristwatch attack' using huge encoding parameters (cf.
6168 James H. Manger's CRYPTO 2001 paper). Note that the
6169 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6170 encoding parameters and hence was not vulnerable.
6173 *) BN_sqr() bug fix.
6174 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6176 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6177 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6178 followed by modular reduction.
6179 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6181 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6182 equivalent based on BN_pseudo_rand() instead of BN_rand().
6185 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6186 This function was broken, as the check for a new client hello message
6187 to handle SGC did not allow these large messages.
6188 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6191 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6194 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6195 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6198 *) Rework the configuration and shared library support for Tru64 Unix.
6199 The configuration part makes use of modern compiler features and
6200 still retains old compiler behavior for those that run older versions
6201 of the OS. The shared library support part includes a variant that
6202 uses the RPATH feature, and is available through the special
6203 configuration target "alpha-cc-rpath", which will never be selected
6205 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6207 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6208 with the same message size as in ssl3_get_certificate_request().
6209 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6210 messages might inadvertently be reject as too long.
6211 [Petr Lampa <lampa@fee.vutbr.cz>]
6213 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6216 *) Modified SSL library such that the verify_callback that has been set
6217 specificly for an SSL object with SSL_set_verify() is actually being
6218 used. Before the change, a verify_callback set with this function was
6219 ignored and the verify_callback() set in the SSL_CTX at the time of
6220 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6221 to allow the necessary settings.
6224 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6225 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6226 done automatically (in contradiction to the requirements of the C
6227 standard). This made problems when used from OpenSSH.
6230 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6231 dh->length and always used
6233 BN_rand_range(priv_key, dh->p).
6235 BN_rand_range() is not necessary for Diffie-Hellman, and this
6236 specific range makes Diffie-Hellman unnecessarily inefficient if
6237 dh->length (recommended exponent length) is much smaller than the
6238 length of dh->p. We could use BN_rand_range() if the order of
6239 the subgroup was stored in the DH structure, but we only have
6244 BN_rand(priv_key, l, ...)
6246 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6252 RSA_eay_public_encrypt
6253 RSA_eay_private_decrypt
6254 RSA_eay_private_encrypt (signing)
6255 RSA_eay_public_decrypt (signature verification)
6257 (default implementations for RSA_public_encrypt,
6258 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6259 always reject numbers >= n.
6262 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6263 to synchronize access to 'locking_thread'. This is necessary on
6264 systems where access to 'locking_thread' (an 'unsigned long'
6265 variable) is not atomic.
6268 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6269 *before* setting the 'crypto_lock_rand' flag. The previous code had
6270 a race condition if 0 is a valid thread ID.
6271 [Travis Vitek <vitek@roguewave.com>]
6273 *) Add support for shared libraries under Irix.
6274 [Albert Chin-A-Young <china@thewrittenword.com>]
6276 *) Add configuration option to build on Linux on both big-endian and
6278 [Ralf Baechle <ralf@uni-koblenz.de>]
6280 *) Add the possibility to create shared libraries on HP-UX.
6283 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6285 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6286 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6287 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6288 PRNG state recovery was possible based on the output of
6289 one PRNG request appropriately sized to gain knowledge on
6290 'md' followed by enough consecutive 1-byte PRNG requests
6291 to traverse all of 'state'.
6293 1. When updating 'md_local' (the current thread's copy of 'md')
6294 during PRNG output generation, hash all of the previous
6295 'md_local' value, not just the half used for PRNG output.
6297 2. Make the number of bytes from 'state' included into the hash
6298 independent from the number of PRNG bytes requested.
6300 The first measure alone would be sufficient to avoid
6301 Markku-Juhani's attack. (Actually it had never occurred
6302 to me that the half of 'md_local' used for chaining was the
6303 half from which PRNG output bytes were taken -- I had always
6304 assumed that the secret half would be used.) The second
6305 measure makes sure that additional data from 'state' is never
6306 mixed into 'md_local' in small portions; this heuristically
6307 further strengthens the PRNG.
6310 *) Fix crypto/bn/asm/mips3.s.
6313 *) When only the key is given to "enc", the IV is undefined. Print out
6314 an error message in this case.
6317 *) Handle special case when X509_NAME is empty in X509 printing routines.
6320 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6321 positive and less than q.
6324 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6325 used: it isn't thread safe and the add_lock_callback should handle
6327 [Paul Rose <Paul.Rose@bridge.com>]
6329 *) Verify that incoming data obeys the block size in
6330 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6334 [Ulf Möller, Bodo Möller]
6336 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6337 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6338 when fixing the server behaviour for backwards-compatible 'client
6339 hello' messages. (Note that the attack is impractical against
6340 SSL 3.0 and TLS 1.0 anyway because length and version checking
6341 means that the probability of guessing a valid ciphertext is
6342 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6345 Before 0.9.5, the countermeasure (hide the error by generating a
6346 random 'decryption result') did not work properly because
6347 ERR_clear_error() was missing, meaning that SSL_get_error() would
6348 detect the supposedly ignored error.
6350 Both problems are now fixed.
6353 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6354 (previously it was 1024).
6357 *) Fix for compatibility mode trust settings: ignore trust settings
6358 unless some valid trust or reject settings are present.
6361 *) Fix for blowfish EVP: its a variable length cipher.
6364 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6365 parameters in DSA public key structures and return an error in the
6366 DSA routines if parameters are absent.
6369 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6370 in the current directory if neither $RANDFILE nor $HOME was set.
6371 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6372 caused some confusion to Windows users who haven't defined $HOME.
6373 Thus RAND_file_name() is changed again: e_os.h can define a
6374 DEFAULT_HOME, which will be used if $HOME is not set.
6375 For Windows, we use "C:"; on other platforms, we still require
6376 environment variables.
6378 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6379 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6380 having multiple threads call RAND_poll() concurrently.
6383 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6384 combination of a flag and a thread ID variable.
6385 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6386 flag), *other* threads can enter ssleay_add_bytes without obeying
6387 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6388 that they do not hold after the first thread unsets add_do_not_lock).
6391 *) Change bctest again: '-x' expressions are not available in all
6395 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6397 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6398 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6400 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6401 the default extension for executables, if any. Also, make the perl
6402 scripts that use symlink() to test if it really exists and use "cp"
6403 if it doesn't. All this made OpenSSL compilable and installable in
6407 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6408 If SEQUENCE is length is indefinite just set c->slen to the total
6409 amount of data available.
6410 [Steve Henson, reported by shige@FreeBSD.org]
6411 [This change does not apply to 0.9.7.]
6413 *) Change bctest to avoid here-documents inside command substitution
6414 (workaround for FreeBSD /bin/sh bug).
6415 For compatibility with Ultrix, avoid shell functions (introduced
6416 in the bctest version that searches along $PATH).
6419 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6420 with des_encrypt() defined on some operating systems, like Solaris
6424 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6425 On the Importance of Eliminating Errors in Cryptographic
6426 Computations, J. Cryptology 14 (2001) 2, 101-119,
6427 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6430 *) MIPS assembler BIGNUM division bug fix.
6433 *) Disabled incorrect Alpha assembler code.
6436 *) Fix PKCS#7 decode routines so they correctly update the length
6437 after reading an EOC for the EXPLICIT tag.
6439 [This change does not apply to 0.9.7.]
6441 *) Fix bug in PKCS#12 key generation routines. This was triggered
6442 if a 3DES key was generated with a 0 initial byte. Include
6443 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6444 (but broken) behaviour.
6447 *) Enhance bctest to search for a working bc along $PATH and print
6449 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6451 *) Fix memory leaks in err.c: free err_data string if necessary;
6452 don't write to the wrong index in ERR_set_error_data.
6455 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6459 *) Replace rdtsc with _emit statements for VC++ version 5.
6460 [Jeremy Cooper <jeremy@baymoo.org>]
6462 *) Make it possible to reuse SSLv2 sessions.
6465 *) In copy_email() check for >= 0 as a return value for
6466 X509_NAME_get_index_by_NID() since 0 is a valid index.
6467 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6469 *) Avoid coredump with unsupported or invalid public keys by checking if
6470 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6471 PKCS7_verify() fails with non detached data.
6474 *) Don't use getenv in library functions when run as setuid/setgid.
6475 New function OPENSSL_issetugid().
6478 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6479 due to incorrect handling of multi-threading:
6481 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6483 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6485 3. Count how many times MemCheck_off() has been called so that
6486 nested use can be treated correctly. This also avoids
6487 inband-signalling in the previous code (which relied on the
6488 assumption that thread ID 0 is impossible).
6491 *) Add "-rand" option also to s_client and s_server.
6494 *) Fix CPU detection on Irix 6.x.
6495 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6496 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6498 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6501 [This change does not apply to 0.9.7.]
6503 *) Use the cached encoding of an X509_NAME structure rather than
6504 copying it. This is apparently the reason for the libsafe "errors"
6505 but the code is actually correct.
6508 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6509 Bleichenbacher's DSA attack.
6510 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6511 to be set and top=0 forces the highest bit to be set; top=-1 is new
6512 and leaves the highest bit random.
6513 [Ulf Moeller, Bodo Moeller]
6515 *) In the NCONF_...-based implementations for CONF_... queries
6516 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6517 a temporary CONF structure with the data component set to NULL
6518 (which gives segmentation faults in lh_retrieve).
6519 Instead, use NULL for the CONF pointer in CONF_get_string and
6520 CONF_get_number (which may use environment variables) and directly
6521 return NULL from CONF_get_section.
6524 *) Fix potential buffer overrun for EBCDIC.
6527 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6528 keyUsage if basicConstraints absent for a CA.
6531 *) Make SMIME_write_PKCS7() write mail header values with a format that
6532 is more generally accepted (no spaces before the semicolon), since
6533 some programs can't parse those values properly otherwise. Also make
6534 sure BIO's that break lines after each write do not create invalid
6538 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6539 macros previously used would not encode an empty SEQUENCE OF
6540 and break the signature.
6542 [This change does not apply to 0.9.7.]
6544 *) Zero the premaster secret after deriving the master secret in
6548 *) Add some EVP_add_digest_alias registrations (as found in
6549 OpenSSL_add_all_digests()) to SSL_library_init()
6550 aka OpenSSL_add_ssl_algorithms(). This provides improved
6551 compatibility with peers using X.509 certificates
6552 with unconventional AlgorithmIdentifier OIDs.
6555 *) Fix for Irix with NO_ASM.
6556 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6558 *) ./config script fixes.
6559 [Ulf Moeller, Richard Levitte]
6561 *) Fix 'openssl passwd -1'.
6564 *) Change PKCS12_key_gen_asc() so it can cope with non null
6565 terminated strings whose length is passed in the passlen
6566 parameter, for example from PEM callbacks. This was done
6567 by adding an extra length parameter to asc2uni().
6568 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6570 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6571 call failed, free the DSA structure.
6574 *) Fix to uni2asc() to cope with zero length Unicode strings.
6575 These are present in some PKCS#12 files.
6578 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6579 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6580 when writing a 32767 byte record.
6581 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6583 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6584 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6586 (RSA objects have a reference count access to which is protected
6587 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6588 so they are meant to be shared between threads.)
6589 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6590 "Reddie, Steven" <Steven.Reddie@ca.com>]
6592 *) Fix a deadlock in CRYPTO_mem_leaks().
6595 *) Use better test patterns in bntest.
6598 *) rand_win.c fix for Borland C.
6601 *) BN_rshift bugfix for n == 0.
6604 *) Add a 'bctest' script that checks for some known 'bc' bugs
6605 so that 'make test' does not abort just because 'bc' is broken.
6608 *) Store verify_result within SSL_SESSION also for client side to
6609 avoid potential security hole. (Re-used sessions on the client side
6610 always resulted in verify_result==X509_V_OK, not using the original
6611 result of the server certificate verification.)
6614 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6615 SSL3_RT_APPLICATION_DATA, return 0.
6616 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6620 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6621 releases, have been re-implemented by renaming the previous
6622 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6623 and ssl3_read_internal, respectively, and adding 'peek' parameters
6624 to them. The new ssl[23]_{read,peek} functions are calls to
6625 ssl[23]_read_internal with the 'peek' flag set appropriately.
6626 A 'peek' parameter has also been added to ssl3_read_bytes, which
6627 does the actual work for ssl3_read_internal.
6630 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6631 the method-specific "init()" handler. Also clean up ex_data after
6632 calling the method-specific "finish()" handler. Previously, this was
6633 happening the other way round.
6636 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6637 The previous value, 12, was not always sufficient for BN_mod_exp().
6640 *) Make sure that shared libraries get the internal name engine with
6641 the full version number and not just 0. This should mark the
6642 shared libraries as not backward compatible. Of course, this should
6643 be changed again when we can guarantee backward binary compatibility.
6646 *) Fix typo in get_cert_by_subject() in by_dir.c
6647 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6649 *) Rework the system to generate shared libraries:
6651 - Make note of the expected extension for the shared libraries and
6652 if there is a need for symbolic links from for example libcrypto.so.0
6653 to libcrypto.so.0.9.7. There is extended info in Configure for
6656 - Make as few rebuilds of the shared libraries as possible.
6658 - Still avoid linking the OpenSSL programs with the shared libraries.
6660 - When installing, install the shared libraries separately from the
6664 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6666 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6667 and not in SSL_clear because the latter is also used by the
6668 accept/connect functions; previously, the settings made by
6669 SSL_set_read_ahead would be lost during the handshake.
6670 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6672 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6673 Previously, it would create entries for disableed algorithms no
6677 *) Added several new manual pages for SSL_* function.
6680 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6682 *) In ssl23_get_client_hello, generate an error message when faced
6683 with an initial SSL 3.0/TLS record that is too small to contain the
6684 first two bytes of the ClientHello message, i.e. client_version.
6685 (Note that this is a pathologic case that probably has never happened
6686 in real life.) The previous approach was to use the version number
6687 from the record header as a substitute; but our protocol choice
6688 should not depend on that one because it is not authenticated
6689 by the Finished messages.
6692 *) More robust randomness gathering functions for Windows.
6693 [Jeffrey Altman <jaltman@columbia.edu>]
6695 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6696 not set then we don't setup the error code for issuer check errors
6697 to avoid possibly overwriting other errors which the callback does
6698 handle. If an application does set the flag then we assume it knows
6699 what it is doing and can handle the new informational codes
6703 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6704 a general "ANY" type, as such it should be able to decode anything
6705 including tagged types. However it didn't check the class so it would
6706 wrongly interpret tagged types in the same way as their universal
6707 counterpart and unknown types were just rejected. Changed so that the
6708 tagged and unknown types are handled in the same way as a SEQUENCE:
6709 that is the encoding is stored intact. There is also a new type
6710 "V_ASN1_OTHER" which is used when the class is not universal, in this
6711 case we have no idea what the actual type is so we just lump them all
6715 *) On VMS, stdout may very well lead to a file that is written to
6716 in a record-oriented fashion. That means that every write() will
6717 write a separate record, which will be read separately by the
6718 programs trying to read from it. This can be very confusing.
6720 The solution is to put a BIO filter in the way that will buffer
6721 text until a linefeed is reached, and then write everything a
6722 line at a time, so every record written will be an actual line,
6723 not chunks of lines and not (usually doesn't happen, but I've
6724 seen it once) several lines in one record. BIO_f_linebuffer() is
6727 Currently, it's a VMS-only method, because that's where it has
6728 been tested well enough.
6731 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6732 it can return incorrect results.
6733 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6734 but it was in 0.9.6-beta[12].)
6737 *) Disable the check for content being present when verifying detached
6738 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6739 include zero length content when signing messages.
6742 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6743 BIO_ctrl (for BIO pairs).
6746 *) Add DSO method for VMS.
6749 *) Bug fix: Montgomery multiplication could produce results with the
6753 *) Add RPM specification openssl.spec and modify it to build three
6754 packages. The default package contains applications, application
6755 documentation and run-time libraries. The devel package contains
6756 include files, static libraries and function documentation. The
6757 doc package contains the contents of the doc directory. The original
6758 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6761 *) Add a large number of documentation files for many SSL routines.
6762 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6764 *) Add a configuration entry for Sony News 4.
6765 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6767 *) Don't set the two most significant bits to one when generating a
6768 random number < q in the DSA library.
6771 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6772 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6773 the underlying transport is blocking) if a handshake took place.
6774 (The default behaviour is needed by applications such as s_client
6775 and s_server that use select() to determine when to use SSL_read;
6776 but for applications that know in advance when to expect data, it
6777 just makes things more complicated.)
6780 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6784 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6785 work better on such systems.
6786 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6788 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6789 Update PKCS12_parse() so it copies the friendlyName and the
6790 keyid to the certificates aux info.
6793 *) Fix bug in PKCS7_verify() which caused an infinite loop
6794 if there was more than one signature.
6795 [Sven Uszpelkat <su@celocom.de>]
6797 *) Major change in util/mkdef.pl to include extra information
6798 about each symbol, as well as presentig variables as well
6799 as functions. This change means that there's n more need
6800 to rebuild the .num files when some algorithms are excluded.
6803 *) Allow the verify time to be set by an application,
6804 rather than always using the current time.
6807 *) Phase 2 verify code reorganisation. The certificate
6808 verify code now looks up an issuer certificate by a
6809 number of criteria: subject name, authority key id
6810 and key usage. It also verifies self signed certificates
6811 by the same criteria. The main comparison function is
6812 X509_check_issued() which performs these checks.
6814 Lot of changes were necessary in order to support this
6815 without completely rewriting the lookup code.
6817 Authority and subject key identifier are now cached.
6819 The LHASH 'certs' is X509_STORE has now been replaced
6820 by a STACK_OF(X509_OBJECT). This is mainly because an
6821 LHASH can't store or retrieve multiple objects with
6822 the same hash value.
6824 As a result various functions (which were all internal
6825 use only) have changed to handle the new X509_STORE
6826 structure. This will break anything that messed round
6827 with X509_STORE internally.
6829 The functions X509_STORE_add_cert() now checks for an
6830 exact match, rather than just subject name.
6832 The X509_STORE API doesn't directly support the retrieval
6833 of multiple certificates matching a given criteria, however
6834 this can be worked round by performing a lookup first
6835 (which will fill the cache with candidate certificates)
6836 and then examining the cache for matches. This is probably
6837 the best we can do without throwing out X509_LOOKUP
6838 entirely (maybe later...).
6840 The X509_VERIFY_CTX structure has been enhanced considerably.
6842 All certificate lookup operations now go via a get_issuer()
6843 callback. Although this currently uses an X509_STORE it
6844 can be replaced by custom lookups. This is a simple way
6845 to bypass the X509_STORE hackery necessary to make this
6846 work and makes it possible to use more efficient techniques
6847 in future. A very simple version which uses a simple
6848 STACK for its trusted certificate store is also provided
6849 using X509_STORE_CTX_trusted_stack().
6851 The verify_cb() and verify() callbacks now have equivalents
6852 in the X509_STORE_CTX structure.
6854 X509_STORE_CTX also has a 'flags' field which can be used
6855 to customise the verify behaviour.
6858 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6859 excludes S/MIME capabilities.
6862 *) When a certificate request is read in keep a copy of the
6863 original encoding of the signed data and use it when outputing
6864 again. Signatures then use the original encoding rather than
6865 a decoded, encoded version which may cause problems if the
6866 request is improperly encoded.
6869 *) For consistency with other BIO_puts implementations, call
6870 buffer_write(b, ...) directly in buffer_puts instead of calling
6873 In BIO_puts, increment b->num_write as in BIO_write.
6874 [Peter.Sylvester@EdelWeb.fr]
6876 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6877 BN_zero, we may not return a BIGNUM with an array consisting of
6881 *) Avoid calling abort() from within the library when problems are
6882 detected, except if preprocessor symbols have been defined
6883 (such as REF_CHECK, BN_DEBUG etc.).
6886 *) New openssl application 'rsautl'. This utility can be
6887 used for low level RSA operations. DER public key
6888 BIO/fp routines also added.
6891 *) New Configure entry and patches for compiling on QNX 4.
6892 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6894 *) A demo state-machine implementation was sponsored by
6895 Nuron (http://www.nuron.com/) and is now available in
6896 demos/state_machine.
6899 *) New options added to the 'dgst' utility for signature
6900 generation and verification.
6903 *) Unrecognized PKCS#7 content types are now handled via a
6904 catch all ASN1_TYPE structure. This allows unsupported
6905 types to be stored as a "blob" and an application can
6906 encode and decode it manually.
6909 *) Fix various signed/unsigned issues to make a_strex.c
6911 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6913 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6914 length if passed a buffer. ASN1_INTEGER_to_BN failed
6915 if passed a NULL BN and its argument was negative.
6916 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6918 *) Modification to PKCS#7 encoding routines to output definite
6919 length encoding. Since currently the whole structures are in
6920 memory there's not real point in using indefinite length
6921 constructed encoding. However if OpenSSL is compiled with
6922 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6925 *) Added BIO_vprintf() and BIO_vsnprintf().
6928 *) Added more prefixes to parse for in the the strings written
6929 through a logging bio, to cover all the levels that are available
6930 through syslog. The prefixes are now:
6932 PANIC, EMERG, EMR => LOG_EMERG
6933 ALERT, ALR => LOG_ALERT
6934 CRIT, CRI => LOG_CRIT
6935 ERROR, ERR => LOG_ERR
6936 WARNING, WARN, WAR => LOG_WARNING
6937 NOTICE, NOTE, NOT => LOG_NOTICE
6938 INFO, INF => LOG_INFO
6939 DEBUG, DBG => LOG_DEBUG
6941 and as before, if none of those prefixes are present at the
6942 beginning of the string, LOG_ERR is chosen.
6944 On Win32, the LOG_* levels are mapped according to this:
6946 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6947 LOG_WARNING => EVENTLOG_WARNING_TYPE
6948 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6952 *) Made it possible to reconfigure with just the configuration
6953 argument "reconf" or "reconfigure". The command line arguments
6954 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6955 and are retrieved from there when reconfiguring.
6959 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6961 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6964 *) The obj_dat.pl script was messing up the sorting of object
6965 names. The reason was that it compared the quoted version
6966 of strings as a result "OCSP" > "OCSP Signing" because
6967 " > SPACE. Changed script to store unquoted versions of
6968 names and add quotes on output. It was also omitting some
6969 names from the lookup table if they were given a default
6970 value (that is if SN is missing it is given the same
6971 value as LN and vice versa), these are now added on the
6972 grounds that if an object has a name we should be able to
6973 look it up. Finally added warning output when duplicate
6974 short or long names are found.
6977 *) Changes needed for Tandem NSK.
6978 [Scott Uroff <scott@xypro.com>]
6980 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6981 RSA_padding_check_SSLv23(), special padding was never detected
6982 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6983 version rollback attacks was not effective.
6985 In s23_clnt.c, don't use special rollback-attack detection padding
6986 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6987 client; similarly, in s23_srvr.c, don't do the rollback check if
6988 SSL 2.0 is the only protocol enabled in the server.
6991 *) Make it possible to get hexdumps of unprintable data with 'openssl
6992 asn1parse'. By implication, the functions ASN1_parse_dump() and
6993 BIO_dump_indent() are added.
6996 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6997 these print out strings and name structures based on various
6998 flags including RFC2253 support and proper handling of
6999 multibyte characters. Added options to the 'x509' utility
7000 to allow the various flags to be set.
7003 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7004 Also change the functions X509_cmp_current_time() and
7005 X509_gmtime_adj() work with an ASN1_TIME structure,
7006 this will enable certificates using GeneralizedTime in validity
7007 dates to be checked.
7010 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7011 negative public key encodings) on by default,
7012 NO_NEG_PUBKEY_BUG can be set to disable it.
7015 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7016 content octets. An i2c_ASN1_OBJECT is unnecessary because
7017 the encoding can be trivially obtained from the structure.
7020 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7021 not read locks (CRYPTO_r_[un]lock).
7024 *) A first attempt at creating official support for shared
7025 libraries through configuration. I've kept it so the
7026 default is static libraries only, and the OpenSSL programs
7027 are always statically linked for now, but there are
7028 preparations for dynamic linking in place.
7029 This has been tested on Linux and Tru64.
7032 *) Randomness polling function for Win9x, as described in:
7033 Peter Gutmann, Software Generation of Practically Strong
7037 *) Fix so PRNG is seeded in req if using an already existing
7041 *) New options to smime application. -inform and -outform
7042 allow alternative formats for the S/MIME message including
7043 PEM and DER. The -content option allows the content to be
7044 specified separately. This should allow things like Netscape
7045 form signing output easier to verify.
7048 *) Fix the ASN1 encoding of tags using the 'long form'.
7051 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7052 STRING types. These convert content octets to and from the
7053 underlying type. The actual tag and length octets are
7054 already assumed to have been read in and checked. These
7055 are needed because all other string types have virtually
7056 identical handling apart from the tag. By having versions
7057 of the ASN1 functions that just operate on content octets
7058 IMPLICIT tagging can be handled properly. It also allows
7059 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7060 and ASN1_INTEGER are identical apart from the tag.
7063 *) Change the handling of OID objects as follows:
7065 - New object identifiers are inserted in objects.txt, following
7066 the syntax given in objects.README.
7067 - objects.pl is used to process obj_mac.num and create a new
7069 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7072 This is currently kind of a hack, and the perl code in objects.pl
7073 isn't very elegant, but it works as I intended. The simplest way
7074 to check that it worked correctly is to look in obj_dat.h and
7075 check the array nid_objs and make sure the objects haven't moved
7076 around (this is important!). Additions are OK, as well as
7077 consistent name changes.
7080 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7083 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7084 The given file adds to whatever has already been seeded into the
7085 random pool through the RANDFILE configuration file option or
7086 environment variable, or the default random state file.
7089 *) mkstack.pl now sorts each macro group into lexical order.
7090 Previously the output order depended on the order the files
7091 appeared in the directory, resulting in needless rewriting
7095 *) Patches to make OpenSSL compile under Win32 again. Mostly
7096 work arounds for the VC++ problem that it treats func() as
7097 func(void). Also stripped out the parts of mkdef.pl that
7098 added extra typesafe functions: these no longer exist.
7101 *) Reorganisation of the stack code. The macros are now all
7102 collected in safestack.h . Each macro is defined in terms of
7103 a "stack macro" of the form SKM_<name>(type, a, b). The
7104 DEBUG_SAFESTACK is now handled in terms of function casts,
7105 this has the advantage of retaining type safety without the
7106 use of additional functions. If DEBUG_SAFESTACK is not defined
7107 then the non typesafe macros are used instead. Also modified the
7108 mkstack.pl script to handle the new form. Needs testing to see
7109 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7110 the default if no major problems. Similar behaviour for ASN1_SET_OF
7111 and PKCS12_STACK_OF.
7114 *) When some versions of IIS use the 'NET' form of private key the
7115 key derivation algorithm is different. Normally MD5(password) is
7116 used as a 128 bit RC4 key. In the modified case
7117 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7118 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7119 as the old Netscape_RSA functions except they have an additional
7120 'sgckey' parameter which uses the modified algorithm. Also added
7121 an -sgckey command line option to the rsa utility. Thanks to
7122 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7123 algorithm to openssl-dev.
7126 *) The evp_local.h macros were using 'c.##kname' which resulted in
7127 invalid expansion on some systems (SCO 5.0.5 for example).
7128 Corrected to 'c.kname'.
7129 [Phillip Porch <root@theporch.com>]
7131 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7132 a STACK of email addresses from a certificate or request, these look
7133 in the subject name and the subject alternative name extensions and
7134 omit any duplicate addresses.
7137 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7138 This makes DSA verification about 2 % faster.
7141 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7142 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7143 plus overhead for 1024 bit moduli).
7144 This makes exponentiations about 0.5 % faster for 1024 bit
7145 exponents (as measured by "openssl speed rsa2048").
7148 *) Rename memory handling macros to avoid conflicts with other
7150 Malloc => OPENSSL_malloc
7151 Malloc_locked => OPENSSL_malloc_locked
7152 Realloc => OPENSSL_realloc
7153 Free => OPENSSL_free
7156 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7157 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7160 *) CygWin32 support.
7161 [John Jarvie <jjarvie@newsguy.com>]
7163 *) The type-safe stack code has been rejigged. It is now only compiled
7164 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7165 by default all type-specific stack functions are "#define"d back to
7166 standard stack functions. This results in more streamlined output
7167 but retains the type-safety checking possibilities of the original
7171 *) The STACK code has been cleaned up, and certain type declarations
7172 that didn't make a lot of sense have been brought in line. This has
7173 also involved a cleanup of sorts in safestack.h to more correctly
7174 map type-safe stack functions onto their plain stack counterparts.
7175 This work has also resulted in a variety of "const"ifications of
7176 lots of the code, especially "_cmp" operations which should normally
7177 be prototyped with "const" parameters anyway.
7180 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7181 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7182 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7183 where all of 'md' is used each time the PRNG is used, but 'state'
7184 is used only indexed by a cyclic counter. As entropy may not be
7185 well distributed from the beginning, 'md' is important as a
7186 chaining variable. However, the output function chains only half
7187 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7188 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7189 in all of 'state' being rewritten, with the new values depending
7190 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7193 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7194 the handshake is continued after ssl_verify_cert_chain();
7195 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7196 can lead to 'unexplainable' connection aborts later.
7197 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7199 *) Major EVP API cipher revision.
7200 Add hooks for extra EVP features. This allows various cipher
7201 parameters to be set in the EVP interface. Support added for variable
7202 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7203 setting of RC2 and RC5 parameters.
7205 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7208 Remove lots of duplicated code from the EVP library. For example *every*
7209 cipher init() function handles the 'iv' in the same way according to the
7210 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7211 for CFB and OFB modes they zero ctx->num.
7213 New functionality allows removal of S/MIME code RC2 hack.
7215 Most of the routines have the same form and so can be declared in terms
7218 By shifting this to the top level EVP_CipherInit() it can be removed from
7219 all individual ciphers. If the cipher wants to handle IVs or keys
7220 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7223 Change lots of functions like EVP_EncryptUpdate() to now return a
7224 value: although software versions of the algorithms cannot fail
7225 any installed hardware versions can.
7228 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7229 this option is set, tolerate broken clients that send the negotiated
7230 protocol version number instead of the requested protocol version
7234 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7235 i.e. non-zero for export ciphersuites, zero otherwise.
7236 Previous versions had this flag inverted, inconsistent with
7237 rsa_tmp_cb (..._TMP_RSA_CB).
7238 [Bodo Moeller; problem reported by Amit Chopra]
7240 *) Add missing DSA library text string. Work around for some IIS
7241 key files with invalid SEQUENCE encoding.
7244 *) Add a document (doc/standards.txt) that list all kinds of standards
7245 and so on that are implemented in OpenSSL.
7248 *) Enhance c_rehash script. Old version would mishandle certificates
7249 with the same subject name hash and wouldn't handle CRLs at all.
7250 Added -fingerprint option to crl utility, to support new c_rehash
7254 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7257 *) Fix for SSL server purpose checking. Server checking was
7258 rejecting certificates which had extended key usage present
7259 but no ssl client purpose.
7260 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7262 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7263 is a little unclear about how a blank password is handled.
7264 Since the password in encoded as a BMPString with terminating
7265 double NULL a zero length password would end up as just the
7266 double NULL. However no password at all is different and is
7267 handled differently in the PKCS#12 key generation code. NS
7268 treats a blank password as zero length. MSIE treats it as no
7269 password on export: but it will try both on import. We now do
7270 the same: PKCS12_parse() tries zero length and no password if
7271 the password is set to "" or NULL (NULL is now a valid password:
7272 it wasn't before) as does the pkcs12 application.
7275 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7276 perror when PEM_read_bio_X509_REQ fails, the error message must
7277 be obtained from the error queue.
7280 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7281 it in ERR_remove_state if appropriate, and change ERR_get_state
7282 accordingly to avoid race conditions (this is necessary because
7283 thread_hash is no longer constant once set).
7286 *) Bugfix for linux-elf makefile.one.
7289 *) RSA_get_default_method() will now cause a default
7290 RSA_METHOD to be chosen if one doesn't exist already.
7291 Previously this was only set during a call to RSA_new()
7292 or RSA_new_method(NULL) meaning it was possible for
7293 RSA_get_default_method() to return NULL.
7296 *) Added native name translation to the existing DSO code
7297 that will convert (if the flag to do so is set) filenames
7298 that are sufficiently small and have no path information
7299 into a canonical native form. Eg. "blah" converted to
7300 "libblah.so" or "blah.dll" etc.
7303 *) New function ERR_error_string_n(e, buf, len) which is like
7304 ERR_error_string(e, buf), but writes at most 'len' bytes
7305 including the 0 terminator. For ERR_error_string_n, 'buf'
7307 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7309 *) CONF library reworked to become more general. A new CONF
7310 configuration file reader "class" is implemented as well as a
7311 new functions (NCONF_*, for "New CONF") to handle it. The now
7312 old CONF_* functions are still there, but are reimplemented to
7313 work in terms of the new functions. Also, a set of functions
7314 to handle the internal storage of the configuration data is
7315 provided to make it easier to write new configuration file
7316 reader "classes" (I can definitely see something reading a
7317 configuration file in XML format, for example), called _CONF_*,
7318 or "the configuration storage API"...
7320 The new configuration file reading functions are:
7322 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7323 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7325 NCONF_default, NCONF_WIN32
7327 NCONF_dump_fp, NCONF_dump_bio
7329 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7330 NCONF_new creates a new CONF object. This works in the same way
7331 as other interfaces in OpenSSL, like the BIO interface.
7332 NCONF_dump_* dump the internal storage of the configuration file,
7333 which is useful for debugging. All other functions take the same
7334 arguments as the old CONF_* functions wth the exception of the
7335 first that must be a `CONF *' instead of a `LHASH *'.
7337 To make it easer to use the new classes with the old CONF_* functions,
7338 the function CONF_set_default_method is provided.
7341 *) Add '-tls1' option to 'openssl ciphers', which was already
7342 mentioned in the documentation but had not been implemented.
7343 (This option is not yet really useful because even the additional
7344 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7347 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7348 OpenSSL-based applications) load shared libraries and bind to
7349 them in a portable way.
7350 [Geoff Thorpe, with contributions from Richard Levitte]
7352 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7354 *) Make sure _lrotl and _lrotr are only used with MSVC.
7356 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7357 (the default implementation of RAND_status).
7359 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7360 to '-clrext' (= clear extensions), as intended and documented.
7361 [Bodo Moeller; inconsistency pointed out by Michael Attili
7364 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7365 was larger than the MD block size.
7366 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7368 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7369 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7370 using the passed key: if the passed key was a private key the result
7371 of X509_print(), for example, would be to print out all the private key
7375 *) des_quad_cksum() byte order bug fix.
7376 [Ulf Möller, using the problem description in krb4-0.9.7, where
7377 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7379 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7381 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7383 *) For easily testing in shell scripts whether some command
7384 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7385 returns with exit code 0 iff no command of the given name is available.
7386 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7387 the output goes to stdout and nothing is printed to stderr.
7388 Additional arguments are always ignored.
7390 Since for each cipher there is a command of the same name,
7391 the 'no-cipher' compilation switches can be tested this way.
7393 ('openssl no-XXX' is not able to detect pseudo-commands such
7394 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7397 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7400 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7401 is set; it will be thrown away anyway because each handshake creates
7403 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7404 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7405 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7406 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7409 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7410 'Q' and 'R' lose their special meanings (quit/renegotiate).
7411 This is part of what -quiet does; unlike -quiet, -ign_eof
7412 does not suppress any output.
7415 *) Add compatibility options to the purpose and trust code. The
7416 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7417 accepts a certificate or CA, this was the previous behaviour,
7418 with all the associated security issues.
7420 X509_TRUST_COMPAT is the old trust behaviour: only and
7421 automatically trust self signed roots in certificate store. A
7422 new trust setting X509_TRUST_DEFAULT is used to specify that
7423 a purpose has no associated trust setting and it should instead
7424 use the value in the default purpose.
7427 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7428 and fix a memory leak.
7431 *) In util/mkerr.pl (which implements 'make errors'), preserve
7432 reason strings from the previous version of the .c file, as
7433 the default to have only downcase letters (and digits) in
7434 automatically generated reasons codes is not always appropriate.
7437 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7438 using strerror. Previously, ERR_reason_error_string() returned
7439 library names as reason strings for SYSerr; but SYSerr is a special
7440 case where small numbers are errno values, not library numbers.
7443 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7444 converts DSA parameters into DH parameters. (When creating parameters,
7445 DSA_generate_parameters is used.)
7448 *) Include 'length' (recommended exponent length) in C code generated
7449 by 'openssl dhparam -C'.
7452 *) The second argument to set_label in perlasm was already being used
7453 so couldn't be used as a "file scope" flag. Moved to third argument
7457 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7458 instead of RAND_bytes for encryption IVs and salts.
7461 *) Include RAND_status() into RAND_METHOD instead of implementing
7462 it only for md_rand.c Otherwise replacing the PRNG by calling
7463 RAND_set_rand_method would be impossible.
7466 *) Don't let DSA_generate_key() enter an infinite loop if the random
7467 number generation fails.
7470 *) New 'rand' application for creating pseudo-random output.
7473 *) Added configuration support for Linux/IA64
7474 [Rolf Haberrecker <rolf@suse.de>]
7476 *) Assembler module support for Mingw32.
7479 *) Shared library support for HPUX (in shlib/).
7480 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7482 *) Shared library support for Solaris gcc.
7483 [Lutz Behnke <behnke@trustcenter.de>]
7485 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7487 *) PKCS7_encrypt() was adding text MIME headers twice because they
7488 were added manually and by SMIME_crlf_copy().
7491 *) In bntest.c don't call BN_rand with zero bits argument.
7492 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7494 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7495 case was implemented. This caused BN_div_recp() to fail occasionally.
7498 *) Add an optional second argument to the set_label() in the perl
7499 assembly language builder. If this argument exists and is set
7500 to 1 it signals that the assembler should use a symbol whose
7501 scope is the entire file, not just the current function. This
7502 is needed with MASM which uses the format label:: for this scope.
7503 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7505 *) Change the ASN1 types so they are typedefs by default. Before
7506 almost all types were #define'd to ASN1_STRING which was causing
7507 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7511 *) Change names of new functions to the new get1/get0 naming
7512 convention: After 'get1', the caller owns a reference count
7513 and has to call ..._free; 'get0' returns a pointer to some
7514 data structure without incrementing reference counters.
7515 (Some of the existing 'get' functions increment a reference
7516 counter, some don't.)
7517 Similarly, 'set1' and 'add1' functions increase reference
7518 counters or duplicate objects.
7521 *) Allow for the possibility of temp RSA key generation failure:
7522 the code used to assume it always worked and crashed on failure.
7525 *) Fix potential buffer overrun problem in BIO_printf().
7526 [Ulf Möller, using public domain code by Patrick Powell; problem
7527 pointed out by David Sacerdote <das33@cornell.edu>]
7529 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7530 RAND_egd() and RAND_status(). In the command line application,
7531 the EGD socket can be specified like a seed file using RANDFILE
7535 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7536 Some CAs (e.g. Verisign) distribute certificates in this form.
7539 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7540 list to exclude them. This means that no special compilation option
7541 is needed to use anonymous DH: it just needs to be included in the
7545 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7546 EVP_MD_type. The old functionality is available in a new macro called
7547 EVP_MD_md(). Change code that uses it and update docs.
7550 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7551 where the 'void *' argument is replaced by a function pointer argument.
7552 Previously 'void *' was abused to point to functions, which works on
7553 many platforms, but is not correct. As these functions are usually
7554 called by macros defined in OpenSSL header files, most source code
7555 should work without changes.
7558 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7559 sections with information on -D... compiler switches used for
7560 compiling the library so that applications can see them. To enable
7561 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7562 must be defined. E.g.,
7563 #define OPENSSL_ALGORITHM_DEFINES
7564 #include <openssl/opensslconf.h>
7565 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7566 [Richard Levitte, Ulf and Bodo Möller]
7568 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7572 *) Change the 'other' type in certificate aux info to a STACK_OF
7573 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7574 the required ASN1 format: arbitrary types determined by an OID.
7577 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7578 argument to 'req'. This is not because the function is newer or
7579 better than others it just uses the work 'NEW' in the certificate
7580 request header lines. Some software needs this.
7583 *) Reorganise password command line arguments: now passwords can be
7584 obtained from various sources. Delete the PEM_cb function and make
7585 it the default behaviour: i.e. if the callback is NULL and the
7586 usrdata argument is not NULL interpret it as a null terminated pass
7587 phrase. If usrdata and the callback are NULL then the pass phrase
7588 is prompted for as usual.
7591 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7592 the support is automatically enabled. The resulting binaries will
7593 autodetect the card and use it if present.
7594 [Ben Laurie and Compaq Inc.]
7596 *) Work around for Netscape hang bug. This sends certificate request
7597 and server done in one record. Since this is perfectly legal in the
7598 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7599 the bugs/SSLv3 entry for more info.
7602 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7605 *) Add -rand argument to smime and pkcs12 applications and read/write
7609 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7612 *) Add command line password options to the remaining applications.
7615 *) Bug fix for BN_div_recp() for numerators with an even number of
7619 *) More tests in bntest.c, and changed test_bn output.
7622 *) ./config recognizes MacOS X now.
7625 *) Bug fix for BN_div() when the first words of num and divsor are
7626 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7629 *) Add support for various broken PKCS#8 formats, and command line
7630 options to produce them.
7633 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7634 get temporary BIGNUMs from a BN_CTX.
7637 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7641 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7642 include a #define from the old name to the new. The original intent
7643 was that statically linked binaries could for example just call
7644 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7645 link with digests. This never worked becayse SSLeay_add_all_digests()
7646 and SSLeay_add_all_ciphers() were in the same source file so calling
7647 one would link with the other. They are now in separate source files.
7650 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7653 *) Use a less unusual form of the Miller-Rabin primality test (it used
7654 a binary algorithm for exponentiation integrated into the Miller-Rabin
7655 loop, our standard modexp algorithms are faster).
7658 *) Support for the EBCDIC character set completed.
7659 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7661 *) Source code cleanups: use const where appropriate, eliminate casts,
7662 use void * instead of char * in lhash.
7665 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7666 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7667 this the server could overwrite ephemeral keys that the client
7671 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7672 using 50 iterations of the Rabin-Miller test.
7674 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7675 iterations of the Rabin-Miller test as required by the appendix
7676 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7677 As BN_is_prime_fasttest includes trial division, DSA parameter
7678 generation becomes much faster.
7680 This implies a change for the callback functions in DSA_is_prime
7681 and DSA_generate_parameters: The callback function is called once
7682 for each positive witness in the Rabin-Miller test, not just
7683 occasionally in the inner loop; and the parameters to the
7684 callback function now provide an iteration count for the outer
7685 loop rather than for the current invocation of the inner loop.
7686 DSA_generate_parameters additionally can call the callback
7687 function with an 'iteration count' of -1, meaning that a
7688 candidate has passed the trial division test (when q is generated
7689 from an application-provided seed, trial division is skipped).
7692 *) New function BN_is_prime_fasttest that optionally does trial
7693 division before starting the Rabin-Miller test and has
7694 an additional BN_CTX * argument (whereas BN_is_prime always
7695 has to allocate at least one BN_CTX).
7696 'callback(1, -1, cb_arg)' is called when a number has passed the
7697 trial division stage.
7700 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7704 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7707 *) New function BN_pseudo_rand().
7710 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7711 bignum version of BN_from_montgomery() with the working code from
7712 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7716 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7717 made it impossible to use the same SSL_SESSION data structure in
7718 SSL2 clients in multiple threads.
7721 *) The return value of RAND_load_file() no longer counts bytes obtained
7722 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7723 to seed the PRNG (previously an explicit byte count was required).
7724 [Ulf Möller, Bodo Möller]
7726 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7727 used (char *) instead of (void *) and had casts all over the place.
7730 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7733 *) Retain source code compatibility for BN_prime_checks macro:
7734 BN_is_prime(..., BN_prime_checks, ...) now uses
7735 BN_prime_checks_for_size to determine the appropriate number of
7736 Rabin-Miller iterations.
7739 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7740 DH_CHECK_P_NOT_SAFE_PRIME.
7741 (Check if this is true? OpenPGP calls them "strong".)
7744 *) Merge the functionality of "dh" and "gendh" programs into a new program
7745 "dhparam". The old programs are retained for now but will handle DH keys
7746 (instead of parameters) in future.
7749 *) Make the ciphers, s_server and s_client programs check the return values
7750 when a new cipher list is set.
7753 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7754 ciphers. Before when the 56bit ciphers were enabled the sorting was
7757 The syntax for the cipher sorting has been extended to support sorting by
7758 cipher-strength (using the strength_bits hard coded in the tables).
7759 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7761 Fix a bug in the cipher-command parser: when supplying a cipher command
7762 string with an "undefined" symbol (neither command nor alphanumeric
7763 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7764 an error is flagged.
7766 Due to the strength-sorting extension, the code of the
7767 ssl_create_cipher_list() function was completely rearranged. I hope that
7768 the readability was also increased :-)
7769 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7771 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7772 for the first serial number and places 2 in the serial number file. This
7773 avoids problems when the root CA is created with serial number zero and
7774 the first user certificate has the same issuer name and serial number
7778 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7779 the new code. Add documentation for this stuff.
7782 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7783 X509_*() to X509at_*() on the grounds that they don't handle X509
7784 structures and behave in an analagous way to the X509v3 functions:
7785 they shouldn't be called directly but wrapper functions should be used
7788 So we also now have some wrapper functions that call the X509at functions
7789 when passed certificate requests. (TO DO: similar things can be done with
7790 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7791 things. Some of these need some d2i or i2d and print functionality
7792 because they handle more complex structures.)
7795 *) Add missing #ifndefs that caused missing symbols when building libssl
7796 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7797 NO_RSA in ssl/s2*.c.
7798 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7800 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7801 has a return value which indicates the quality of the random data
7802 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7803 error queue. New function RAND_pseudo_bytes() generates output that is
7804 guaranteed to be unique but not unpredictable. RAND_add is like
7805 RAND_seed, but takes an extra argument for an entropy estimate
7806 (RAND_seed always assumes full entropy).
7809 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7810 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7811 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7812 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7813 false-positive rate of at most 2^-80 for random input.
7816 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7819 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7820 in the 0.9.5 release), this returns the chain
7821 from an X509_CTX structure with a dup of the stack and all
7822 the X509 reference counts upped: so the stack will exist
7823 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7826 Also make SSL_SESSION_print() print out the verify return
7830 *) Add manpage for the pkcs12 command. Also change the default
7831 behaviour so MAC iteration counts are used unless the new
7832 -nomaciter option is used. This improves file security and
7833 only older versions of MSIE (4.0 for example) need it.
7836 *) Honor the no-xxx Configure options when creating .DEF files.
7839 *) Add PKCS#10 attributes to field table: challengePassword,
7840 unstructuredName and unstructuredAddress. These are taken from
7841 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7842 international characters are used.
7844 More changes to X509_ATTRIBUTE code: allow the setting of types
7845 based on strings. Remove the 'loc' parameter when adding
7846 attributes because these will be a SET OF encoding which is sorted
7850 *) Initial changes to the 'req' utility to allow request generation
7851 automation. This will allow an application to just generate a template
7852 file containing all the field values and have req construct the
7855 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7856 used all over the place including certificate requests and PKCS#7
7857 structures. They are currently handled manually where necessary with
7858 some primitive wrappers for PKCS#7. The new functions behave in a
7859 manner analogous to the X509 extension functions: they allow
7860 attributes to be looked up by NID and added.
7862 Later something similar to the X509V3 code would be desirable to
7863 automatically handle the encoding, decoding and printing of the
7864 more complex types. The string types like challengePassword can
7865 be handled by the string table functions.
7867 Also modified the multi byte string table handling. Now there is
7868 a 'global mask' which masks out certain types. The table itself
7869 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7870 is useful when for example there is only one permissible type
7871 (as in countryName) and using the mask might result in no valid
7875 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7876 SSL_get_peer_finished to allow applications to obtain the latest
7877 Finished messages sent to the peer or expected from the peer,
7878 respectively. (SSL_get_peer_finished is usually the Finished message
7879 actually received from the peer, otherwise the protocol will be aborted.)
7881 As the Finished message are message digests of the complete handshake
7882 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7883 be used for external authentication procedures when the authentication
7884 provided by SSL/TLS is not desired or is not enough.
7887 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7888 the host supports BWX extension and if Compaq C is present on the
7889 $PATH. Just exploiting of the BWX extension results in 20-30%
7890 performance kick for some algorithms, e.g. DES and RC4 to mention
7891 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7895 *) Add support for MS "fast SGC". This is arguably a violation of the
7896 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7897 weak crypto and after checking the certificate is SGC a second one
7898 with strong crypto. MS SGC stops the first handshake after receiving
7899 the server certificate message and sends a second client hello. Since
7900 a server will typically do all the time consuming operations before
7901 expecting any further messages from the client (server key exchange
7902 is the most expensive) there is little difference between the two.
7904 To get OpenSSL to support MS SGC we have to permit a second client
7905 hello message after we have sent server done. In addition we have to
7906 reset the MAC if we do get this second client hello.
7909 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7910 if a DER encoded private key is RSA or DSA traditional format. Changed
7911 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7912 format DER encoded private key. Newer code should use PKCS#8 format which
7913 has the key type encoded in the ASN1 structure. Added DER private key
7914 support to pkcs8 application.
7917 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7918 ciphersuites has been selected (as required by the SSL 3/TLS 1
7919 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7920 is set, we interpret this as a request to violate the specification
7921 (the worst that can happen is a handshake failure, and 'correct'
7922 behaviour would result in a handshake failure anyway).
7925 *) In SSL_CTX_add_session, take into account that there might be multiple
7926 SSL_SESSION structures with the same session ID (e.g. when two threads
7927 concurrently obtain them from an external cache).
7928 The internal cache can handle only one SSL_SESSION with a given ID,
7929 so if there's a conflict, we now throw out the old one to achieve
7933 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7934 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7935 some routines that use cipher OIDs: some ciphers do not have OIDs
7936 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7940 *) Simplify the trust setting structure and code. Now we just have
7941 two sequences of OIDs for trusted and rejected settings. These will
7942 typically have values the same as the extended key usage extension
7943 and any application specific purposes.
7945 The trust checking code now has a default behaviour: it will just
7946 check for an object with the same NID as the passed id. Functions can
7947 be provided to override either the default behaviour or the behaviour
7948 for a given id. SSL client, server and email already have functions
7949 in place for compatibility: they check the NID and also return "trusted"
7950 if the certificate is self signed.
7953 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7954 traditional format into an EVP_PKEY structure.
7957 *) Add a password callback function PEM_cb() which either prompts for
7958 a password if usr_data is NULL or otherwise assumes it is a null
7959 terminated password. Allow passwords to be passed on command line
7960 environment or config files in a few more utilities.
7963 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7964 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7965 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7966 Update documentation.
7969 *) Support for ASN1 "NULL" type. This could be handled before by using
7970 ASN1_TYPE but there wasn't any function that would try to read a NULL
7971 and produce an error if it couldn't. For compatibility we also have
7972 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7973 don't allocate anything because they don't need to.
7976 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7978 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7980 *) Rebuild of the memory allocation routines used by OpenSSL code and
7981 possibly others as well. The purpose is to make an interface that
7982 provide hooks so anyone can build a separate set of allocation and
7983 deallocation routines to be used by OpenSSL, for example memory
7984 pool implementations, or something else, which was previously hard
7985 since Malloc(), Realloc() and Free() were defined as macros having
7986 the values malloc, realloc and free, respectively (except for Win32
7987 compilations). The same is provided for memory debugging code.
7988 OpenSSL already comes with functionality to find memory leaks, but
7989 this gives people a chance to debug other memory problems.
7991 With these changes, a new set of functions and macros have appeared:
7993 CRYPTO_set_mem_debug_functions() [F]
7994 CRYPTO_get_mem_debug_functions() [F]
7995 CRYPTO_dbg_set_options() [F]
7996 CRYPTO_dbg_get_options() [F]
7997 CRYPTO_malloc_debug_init() [M]
7999 The memory debug functions are NULL by default, unless the library
8000 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8001 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8002 gives the standard debugging functions that come with OpenSSL) or
8003 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8004 provided by the library user) must be used. When the standard
8005 debugging functions are used, CRYPTO_dbg_set_options can be used to
8006 request additional information:
8007 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8008 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8010 Also, things like CRYPTO_set_mem_functions will always give the
8011 expected result (the new set of functions is used for allocation
8012 and deallocation) at all times, regardless of platform and compiler
8015 To finish it up, some functions that were never use in any other
8016 way than through macros have a new API and new semantic:
8019 CRYPTO_dbg_realloc()
8022 All macros of value have retained their old syntax.
8023 [Richard Levitte and Bodo Moeller]
8025 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8026 ordering of SMIMECapabilities wasn't in "strength order" and there
8027 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8031 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8032 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8033 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8035 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8036 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8037 functionality to handle multipart/signed properly) and a utility
8038 called 'smime' to call all this stuff. This is based on code I
8039 originally wrote for Celo who have kindly allowed it to be
8040 included in OpenSSL.
8043 *) Add variants des_set_key_checked and des_set_key_unchecked of
8044 des_set_key (aka des_key_sched). Global variable des_check_key
8045 decides which of these is called by des_set_key; this way
8046 des_check_key behaves as it always did, but applications and
8047 the library itself, which was buggy for des_check_key == 1,
8048 have a cleaner way to pick the version they need.
8051 *) New function PKCS12_newpass() which changes the password of a
8055 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8056 dynamic mix. In both cases the ids can be used as an index into the
8057 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8058 functions so they accept a list of the field values and the
8059 application doesn't need to directly manipulate the X509_TRUST
8063 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8067 *) Modify the way the V3 extension code looks up extensions. This now
8068 works in a similar way to the object code: we have some "standard"
8069 extensions in a static table which is searched with OBJ_bsearch()
8070 and the application can add dynamic ones if needed. The file
8071 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8072 updated whenever a new extension is added to the core code and kept
8073 in ext_nid order. There is a simple program 'tabtest.c' which checks
8074 this. New extensions are not added too often so this file can readily
8075 be maintained manually.
8077 There are two big advantages in doing things this way. The extensions
8078 can be looked up immediately and no longer need to be "added" using
8079 X509V3_add_standard_extensions(): this function now does nothing.
8080 [Side note: I get *lots* of email saying the extension code doesn't
8081 work because people forget to call this function]
8082 Also no dynamic allocation is done unless new extensions are added:
8083 so if we don't add custom extensions there is no need to call
8084 X509V3_EXT_cleanup().
8087 *) Modify enc utility's salting as follows: make salting the default. Add a
8088 magic header, so unsalted files fail gracefully instead of just decrypting
8089 to garbage. This is because not salting is a big security hole, so people
8090 should be discouraged from doing it.
8093 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8094 digest to be passed on the command line but it only used this
8095 parameter when signing a certificate. Modified so all relevant
8096 operations are affected by the digest parameter including the
8097 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8098 DSA key was used because it didn't fix the digest.
8101 *) Initial certificate chain verify code. Currently tests the untrusted
8102 certificates for consistency with the verify purpose (which is set
8103 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8105 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8106 this is because it will reject chains with invalid extensions whereas
8107 every previous version of OpenSSL and SSLeay made no checks at all.
8109 Trust code: checks the root CA for the relevant trust settings. Trust
8110 settings have an initial value consistent with the verify purpose: e.g.
8111 if the verify purpose is for SSL client use it expects the CA to be
8112 trusted for SSL client use. However the default value can be changed to
8113 permit custom trust settings: one example of this would be to only trust
8114 certificates from a specific "secure" set of CAs.
8116 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8117 which should be used for version portability: especially since the
8118 verify structure is likely to change more often now.
8120 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8121 to set them. If not set then assume SSL clients will verify SSL servers
8124 Two new options to the verify program: -untrusted allows a set of
8125 untrusted certificates to be passed in and -purpose which sets the
8126 intended purpose of the certificate. If a purpose is set then the
8127 new chain verify code is used to check extension consistency.
8130 *) Support for the authority information access extension.
8133 *) Modify RSA and DSA PEM read routines to transparently handle
8134 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8135 public keys in a format compatible with certificate
8136 SubjectPublicKeyInfo structures. Unfortunately there were already
8137 functions called *_PublicKey_* which used various odd formats so
8138 these are retained for compatibility: however the DSA variants were
8139 never in a public release so they have been deleted. Changed dsa/rsa
8140 utilities to handle the new format: note no releases ever handled public
8141 keys so we should be OK.
8143 The primary motivation for this change is to avoid the same fiasco
8144 that dogs private keys: there are several incompatible private key
8145 formats some of which are standard and some OpenSSL specific and
8146 require various evil hacks to allow partial transparent handling and
8147 even then it doesn't work with DER formats. Given the option anything
8148 other than PKCS#8 should be dumped: but the other formats have to
8149 stay in the name of compatibility.
8151 With public keys and the benefit of hindsight one standard format
8152 is used which works with EVP_PKEY, RSA or DSA structures: though
8153 it clearly returns an error if you try to read the wrong kind of key.
8155 Added a -pubkey option to the 'x509' utility to output the public key.
8156 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8157 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8158 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8159 that do the same as the EVP_PKEY_assign_*() except they up the
8160 reference count of the added key (they don't "swallow" the
8164 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8165 CRLs would fail if the file contained no certificates or no CRLs:
8166 added a new function to read in both types and return the number
8167 read: this means that if none are read it will be an error. The
8168 DER versions of the certificate and CRL reader would always fail
8169 because it isn't possible to mix certificates and CRLs in DER format
8170 without choking one or the other routine. Changed this to just read
8171 a certificate: this is the best we can do. Also modified the code
8172 in apps/verify.c to take notice of return codes: it was previously
8173 attempting to read in certificates from NULL pointers and ignoring
8174 any errors: this is one reason why the cert and CRL reader seemed
8175 to work. It doesn't check return codes from the default certificate
8176 routines: these may well fail if the certificates aren't installed.
8179 *) Code to support otherName option in GeneralName.
8182 *) First update to verify code. Change the verify utility
8183 so it warns if it is passed a self signed certificate:
8184 for consistency with the normal behaviour. X509_verify
8185 has been modified to it will now verify a self signed
8186 certificate if *exactly* the same certificate appears
8187 in the store: it was previously impossible to trust a
8188 single self signed certificate. This means that:
8189 openssl verify ss.pem
8190 now gives a warning about a self signed certificate but
8191 openssl verify -CAfile ss.pem ss.pem
8195 *) For servers, store verify_result in SSL_SESSION data structure
8196 (and add it to external session representation).
8197 This is needed when client certificate verifications fails,
8198 but an application-provided verification callback (set by
8199 SSL_CTX_set_cert_verify_callback) allows accepting the session
8200 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8201 but returns 1): When the session is reused, we have to set
8202 ssl->verify_result to the appropriate error code to avoid
8204 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8206 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8207 case in PKCS7_dataInit() where the signed PKCS7 structure
8208 didn't contain any existing data because it was being created.
8209 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8211 *) Add a salt to the key derivation routines in enc.c. This
8212 forms the first 8 bytes of the encrypted file. Also add a
8213 -S option to allow a salt to be input on the command line.
8216 *) New function X509_cmp(). Oddly enough there wasn't a function
8217 to compare two certificates. We do this by working out the SHA1
8218 hash and comparing that. X509_cmp() will be needed by the trust
8222 *) SSL_get1_session() is like SSL_get_session(), but increments
8223 the reference count in the SSL_SESSION returned.
8224 [Geoff Thorpe <geoff@eu.c2.net>]
8226 *) Fix for 'req': it was adding a null to request attributes.
8227 Also change the X509_LOOKUP and X509_INFO code to handle
8228 certificate auxiliary information.
8231 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8235 *) Add the possibility to add extra information to the memory leak
8236 detecting output, to form tracebacks, showing from where each
8237 allocation was originated: CRYPTO_push_info("constant string") adds
8238 the string plus current file name and line number to a per-thread
8239 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8240 is like calling CYRPTO_pop_info() until the stack is empty.
8241 Also updated memory leak detection code to be multi-thread-safe.
8244 *) Add options -text and -noout to pkcs7 utility and delete the
8245 encryption options which never did anything. Update docs.
8248 *) Add options to some of the utilities to allow the pass phrase
8249 to be included on either the command line (not recommended on
8250 OSes like Unix) or read from the environment. Update the
8251 manpages and fix a few bugs.
8254 *) Add a few manpages for some of the openssl commands.
8257 *) Fix the -revoke option in ca. It was freeing up memory twice,
8258 leaking and not finding already revoked certificates.
8261 *) Extensive changes to support certificate auxiliary information.
8262 This involves the use of X509_CERT_AUX structure and X509_AUX
8263 functions. An X509_AUX function such as PEM_read_X509_AUX()
8264 can still read in a certificate file in the usual way but it
8265 will also read in any additional "auxiliary information". By
8266 doing things this way a fair degree of compatibility can be
8267 retained: existing certificates can have this information added
8268 using the new 'x509' options.
8270 Current auxiliary information includes an "alias" and some trust
8271 settings. The trust settings will ultimately be used in enhanced
8272 certificate chain verification routines: currently a certificate
8273 can only be trusted if it is self signed and then it is trusted
8277 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8278 The problem was that one of the replacement routines had not been working
8279 since SSLeay releases. For now the offending routine has been replaced
8280 with non-optimised assembler. Even so, this now gives around 95%
8281 performance improvement for 1024 bit RSA signs.
8284 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8285 handling. Most clients have the effective key size in bits equal to
8286 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8287 A few however don't do this and instead use the size of the decrypted key
8288 to determine the RC2 key length and the AlgorithmIdentifier to determine
8289 the effective key length. In this case the effective key length can still
8290 be 40 bits but the key length can be 168 bits for example. This is fixed
8291 by manually forcing an RC2 key into the EVP_PKEY structure because the
8292 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8293 the key length and effective key length are equal.
8296 *) Add a bunch of functions that should simplify the creation of
8297 X509_NAME structures. Now you should be able to do:
8298 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8299 and have it automatically work out the correct field type and fill in
8300 the structures. The more adventurous can try:
8301 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8302 and it will (hopefully) work out the correct multibyte encoding.
8305 *) Change the 'req' utility to use the new field handling and multibyte
8306 copy routines. Before the DN field creation was handled in an ad hoc
8307 way in req, ca, and x509 which was rather broken and didn't support
8308 BMPStrings or UTF8Strings. Since some software doesn't implement
8309 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8310 using the dirstring_type option. See the new comment in the default
8311 openssl.cnf for more info.
8314 *) Make crypto/rand/md_rand.c more robust:
8315 - Assure unique random numbers after fork().
8316 - Make sure that concurrent threads access the global counter and
8317 md serializably so that we never lose entropy in them
8318 or use exactly the same state in multiple threads.
8319 Access to the large state is not always serializable because
8320 the additional locking could be a performance killer, and
8321 md should be large enough anyway.
8324 *) New file apps/app_rand.c with commonly needed functionality
8325 for handling the random seed file.
8327 Use the random seed file in some applications that previously did not:
8329 dsaparam -genkey (which also ignored its '-rand' option),
8332 x509 (when signing).
8333 Except on systems with /dev/urandom, it is crucial to have a random
8334 seed file at least for key creation, DSA signing, and for DH exchanges;
8335 for RSA signatures we could do without one.
8337 gendh and gendsa (unlike genrsa) used to read only the first byte
8338 of each file listed in the '-rand' option. The function as previously
8339 found in genrsa is now in app_rand.c and is used by all programs
8340 that support '-rand'.
8343 *) In RAND_write_file, use mode 0600 for creating files;
8344 don't just chmod when it may be too late.
8347 *) Report an error from X509_STORE_load_locations
8348 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8351 *) New function ASN1_mbstring_copy() this copies a string in either
8352 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8353 into an ASN1_STRING type. A mask of permissible types is passed
8354 and it chooses the "minimal" type to use or an error if not type
8358 *) Add function equivalents to the various macros in asn1.h. The old
8359 macros are retained with an M_ prefix. Code inside the library can
8360 use the M_ macros. External code (including the openssl utility)
8361 should *NOT* in order to be "shared library friendly".
8364 *) Add various functions that can check a certificate's extensions
8365 to see if it usable for various purposes such as SSL client,
8366 server or S/MIME and CAs of these types. This is currently
8367 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8368 verification. Also added a -purpose flag to x509 utility to
8369 print out all the purposes.
8372 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8376 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8377 for, obtain and decode and extension and obtain its critical flag.
8378 This allows all the necessary extension code to be handled in a
8379 single function call.
8382 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8383 platforms. See crypto/rc4/rc4_enc.c for further details.
8386 *) New -noout option to asn1parse. This causes no output to be produced
8387 its main use is when combined with -strparse and -out to extract data
8388 from a file (which may not be in ASN.1 format).
8391 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8392 when producing the local key id.
8393 [Richard Levitte <levitte@stacken.kth.se>]
8395 *) New option -dhparam in s_server. This allows a DH parameter file to be
8396 stated explicitly. If it is not stated then it tries the first server
8397 certificate file. The previous behaviour hard coded the filename
8401 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8402 a public key to be input or output. For example:
8403 openssl rsa -in key.pem -pubout -out pubkey.pem
8404 Also added necessary DSA public key functions to handle this.
8407 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8408 in the message. This was handled by allowing
8409 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8410 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8412 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8413 to the end of the strings whereas this didn't. This would cause problems
8414 if strings read with d2i_ASN1_bytes() were later modified.
8415 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8417 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8418 data and it contains EOF it will end up returning an error. This is
8419 caused by input 46 bytes long. The cause is due to the way base64
8420 BIOs find the start of base64 encoded data. They do this by trying a
8421 trial decode on each line until they find one that works. When they
8422 do a flag is set and it starts again knowing it can pass all the
8423 data directly through the decoder. Unfortunately it doesn't reset
8424 the context it uses. This means that if EOF is reached an attempt
8425 is made to pass two EOFs through the context and this causes the
8426 resulting error. This can also cause other problems as well. As is
8427 usual with these problems it takes *ages* to find and the fix is
8428 trivial: move one line.
8429 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8431 *) Ugly workaround to get s_client and s_server working under Windows. The
8432 old code wouldn't work because it needed to select() on sockets and the
8433 tty (for keypresses and to see if data could be written). Win32 only
8434 supports select() on sockets so we select() with a 1s timeout on the
8435 sockets and then see if any characters are waiting to be read, if none
8436 are present then we retry, we also assume we can always write data to
8437 the tty. This isn't nice because the code then blocks until we've
8438 received a complete line of data and it is effectively polling the
8439 keyboard at 1s intervals: however it's quite a bit better than not
8440 working at all :-) A dedicated Windows application might handle this
8441 with an event loop for example.
8444 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8445 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8446 will be called when RSA_sign() and RSA_verify() are used. This is useful
8447 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8448 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8449 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8450 This necessitated the support of an extra signature type NID_md5_sha1
8451 for SSL signatures and modifications to the SSL library to use it instead
8452 of calling RSA_public_decrypt() and RSA_private_encrypt().
8455 *) Add new -verify -CAfile and -CApath options to the crl program, these
8456 will lookup a CRL issuers certificate and verify the signature in a
8457 similar way to the verify program. Tidy up the crl program so it
8458 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8459 less strict. It will now permit CRL extensions even if it is not
8460 a V2 CRL: this will allow it to tolerate some broken CRLs.
8463 *) Initialize all non-automatic variables each time one of the openssl
8464 sub-programs is started (this is necessary as they may be started
8465 multiple times from the "OpenSSL>" prompt).
8466 [Lennart Bang, Bodo Moeller]
8468 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8469 removing all other RSA functionality (this is what NO_RSA does). This
8470 is so (for example) those in the US can disable those operations covered
8471 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8475 *) Non-copying interface to BIO pairs.
8476 (still largely untested)
8479 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8480 ASCII string. This was handled independently in various places before.
8483 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8484 UTF8 strings a character at a time.
8487 *) Use client_version from client hello to select the protocol
8488 (s23_srvr.c) and for RSA client key exchange verification
8489 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8492 *) Add various utility functions to handle SPKACs, these were previously
8493 handled by poking round in the structure internals. Added new function
8494 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8495 print, verify and generate SPKACs. Based on an original idea from
8496 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8499 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8502 *) Allow the config file extension section to be overwritten on the
8503 command line. Based on an original idea from Massimiliano Pala
8504 <madwolf@comune.modena.it>. The new option is called -extensions
8505 and can be applied to ca, req and x509. Also -reqexts to override
8506 the request extensions in req and -crlexts to override the crl extensions
8510 *) Add new feature to the SPKAC handling in ca. Now you can include
8511 the same field multiple times by preceding it by "XXXX." for example:
8514 this is the same syntax as used in the req config file.
8517 *) Allow certificate extensions to be added to certificate requests. These
8518 are specified in a 'req_extensions' option of the req section of the
8519 config file. They can be printed out with the -text option to req but
8520 are otherwise ignored at present.
8523 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8524 data read consists of only the final block it would not decrypted because
8525 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8526 A misplaced 'break' also meant the decrypted final block might not be
8527 copied until the next read.
8530 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8531 a few extra parameters to the DH structure: these will be useful if
8532 for example we want the value of 'q' or implement X9.42 DH.
8535 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8536 provides hooks that allow the default DSA functions or functions on a
8537 "per key" basis to be replaced. This allows hardware acceleration and
8538 hardware key storage to be handled without major modification to the
8539 library. Also added low level modexp hooks and CRYPTO_EX structure and
8540 associated functions.
8543 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8544 as "read only": it can't be written to and the buffer it points to will
8545 not be freed. Reading from a read only BIO is much more efficient than
8546 a normal memory BIO. This was added because there are several times when
8547 an area of memory needs to be read from a BIO. The previous method was
8548 to create a memory BIO and write the data to it, this results in two
8549 copies of the data and an O(n^2) reading algorithm. There is a new
8550 function BIO_new_mem_buf() which creates a read only memory BIO from
8551 an area of memory. Also modified the PKCS#7 routines to use read only
8555 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8556 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8557 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8558 but a retry condition occured while trying to read the rest.
8561 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8562 NID_pkcs7_encrypted by default: this was wrong since this should almost
8563 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8564 the encrypted data type: this is a more sensible place to put it and it
8565 allows the PKCS#12 code to be tidied up that duplicated this
8569 *) Changed obj_dat.pl script so it takes its input and output files on
8570 the command line. This should avoid shell escape redirection problems
8574 *) Initial support for certificate extension requests, these are included
8575 in things like Xenroll certificate requests. Included functions to allow
8576 extensions to be obtained and added.
8579 *) -crlf option to s_client and s_server for sending newlines as
8580 CRLF (as required by many protocols).
8583 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8585 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8586 [Ralf S. Engelschall]
8588 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8589 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8591 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8595 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8596 DH parameters/keys (q is lost during that conversion, but the resulting
8597 DH parameters contain its length).
8599 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8600 much faster than DH_generate_parameters (which creates parameters
8601 where p = 2*q + 1), and also the smaller q makes DH computations
8602 much more efficient (160-bit exponentiation instead of 1024-bit
8603 exponentiation); so this provides a convenient way to support DHE
8604 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8605 utter importance to use
8606 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8608 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8609 when such DH parameters are used, because otherwise small subgroup
8610 attacks may become possible!
8613 *) Avoid memory leak in i2d_DHparams.
8616 *) Allow the -k option to be used more than once in the enc program:
8617 this allows the same encrypted message to be read by multiple recipients.
8620 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8621 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8622 it will always use the numerical form of the OID, even if it has a short
8626 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8627 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8628 otherwise bn_mod_exp was called. In the case of hardware keys for example
8629 no private key components need be present and it might store extra data
8630 in the RSA structure, which cannot be accessed from bn_mod_exp.
8631 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8632 private key operations.
8635 *) Added support for SPARC Linux.
8638 *) pem_password_cb function type incompatibly changed from
8639 typedef int pem_password_cb(char *buf, int size, int rwflag);
8641 ....(char *buf, int size, int rwflag, void *userdata);
8642 so that applications can pass data to their callbacks:
8643 The PEM[_ASN1]_{read,write}... functions and macros now take an
8644 additional void * argument, which is just handed through whenever
8645 the password callback is called.
8646 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8648 New function SSL_CTX_set_default_passwd_cb_userdata.
8650 Compatibility note: As many C implementations push function arguments
8651 onto the stack in reverse order, the new library version is likely to
8652 interoperate with programs that have been compiled with the old
8653 pem_password_cb definition (PEM_whatever takes some data that
8654 happens to be on the stack as its last argument, and the callback
8655 just ignores this garbage); but there is no guarantee whatsoever that
8658 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8659 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8660 problems not only on Windows, but also on some Unix platforms.
8661 To avoid problematic command lines, these definitions are now in an
8662 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8663 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8666 *) MIPS III/IV assembler module is reimplemented.
8669 *) More DES library cleanups: remove references to srand/rand and
8670 delete an unused file.
8673 *) Add support for the the free Netwide assembler (NASM) under Win32,
8674 since not many people have MASM (ml) and it can be hard to obtain.
8675 This is currently experimental but it seems to work OK and pass all
8676 the tests. Check out INSTALL.W32 for info.
8679 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8680 without temporary keys kept an extra copy of the server key,
8681 and connections with temporary keys did not free everything in case
8685 *) New function RSA_check_key and new openssl rsa option -check
8686 for verifying the consistency of RSA keys.
8687 [Ulf Moeller, Bodo Moeller]
8689 *) Various changes to make Win32 compile work:
8690 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8691 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8692 comparison" warnings.
8693 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8696 *) Add a debugging option to PKCS#5 v2 key generation function: when
8697 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8698 derived keys are printed to stderr.
8701 *) Copy the flags in ASN1_STRING_dup().
8702 [Roman E. Pavlov <pre@mo.msk.ru>]
8704 *) The x509 application mishandled signing requests containing DSA
8705 keys when the signing key was also DSA and the parameters didn't match.
8707 It was supposed to omit the parameters when they matched the signing key:
8708 the verifying software was then supposed to automatically use the CA's
8709 parameters if they were absent from the end user certificate.
8711 Omitting parameters is no longer recommended. The test was also
8712 the wrong way round! This was probably due to unusual behaviour in
8713 EVP_cmp_parameters() which returns 1 if the parameters match.
8714 This meant that parameters were omitted when they *didn't* match and
8715 the certificate was useless. Certificates signed with 'ca' didn't have
8717 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8719 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8720 The interface is as follows:
8721 Applications can use
8722 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8723 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8724 "off" is now the default.
8725 The library internally uses
8726 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8727 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8728 to disable memory-checking temporarily.
8730 Some inconsistent states that previously were possible (and were
8731 even the default) are now avoided.
8733 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8734 with each memory chunk allocated; this is occasionally more helpful
8735 than just having a counter.
8737 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8739 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8743 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8744 which largely parallels "options", but is for changing API behaviour,
8745 whereas "options" are about protocol behaviour.
8746 Initial "mode" flags are:
8748 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8749 a single record has been written.
8750 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8751 retries use the same buffer location.
8752 (But all of the contents must be
8756 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8759 *) Fix problems with no-hmac etc.
8760 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8762 *) New functions RSA_get_default_method(), RSA_set_method() and
8763 RSA_get_method(). These allows replacement of RSA_METHODs without having
8764 to mess around with the internals of an RSA structure.
8767 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8768 Also really enable memory leak checks in openssl.c and in some
8770 [Chad C. Mulligan, Bodo Moeller]
8772 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8773 up the length of negative integers. This has now been simplified to just
8774 store the length when it is first determined and use it later, rather
8775 than trying to keep track of where data is copied and updating it to
8777 [Steve Henson, reported by Brien Wheeler
8778 <bwheeler@authentica-security.com>]
8780 *) Add a new function PKCS7_signatureVerify. This allows the verification
8781 of a PKCS#7 signature but with the signing certificate passed to the
8782 function itself. This contrasts with PKCS7_dataVerify which assumes the
8783 certificate is present in the PKCS#7 structure. This isn't always the
8784 case: certificates can be omitted from a PKCS#7 structure and be
8785 distributed by "out of band" means (such as a certificate database).
8788 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8789 function prototypes in pem.h, also change util/mkdef.pl to add the
8790 necessary function names.
8793 *) mk1mf.pl (used by Windows builds) did not properly read the
8794 options set by Configure in the top level Makefile, and Configure
8795 was not even able to write more than one option correctly.
8796 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8799 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8800 file to be loaded from a BIO or FILE pointer. The BIO version will
8801 for example allow memory BIOs to contain config info.
8804 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8805 Whoever hopes to achieve shared-library compatibility across versions
8806 must use this, not the compile-time macro.
8807 (Exercise 0.9.4: Which is the minimum library version required by
8809 Note: All this applies only to multi-threaded programs, others don't
8813 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8814 through a BIO pair triggered the default case, i.e.
8815 SSLerr(...,SSL_R_UNKNOWN_STATE).
8818 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8819 can use the SSL library even if none of the specific BIOs is
8823 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8824 for the encoded length.
8825 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8827 *) Add initial documentation of the X509V3 functions.
8830 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8831 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8832 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8833 secure PKCS#8 private key format with a high iteration count.
8836 *) Fix determination of Perl interpreter: A perl or perl5
8837 _directory_ in $PATH was also accepted as the interpreter.
8838 [Ralf S. Engelschall]
8840 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8841 wrong with it but it was very old and did things like calling
8842 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8846 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8847 to use the new extension code.
8850 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8851 with macros. This should make it easier to change their form, add extra
8852 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8856 *) Add to configuration table a new entry that can specify an alternative
8857 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8858 according to Mark Crispin <MRC@Panda.COM>.
8862 *) DES CBC did not update the IV. Weird.
8865 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8866 Changing the behaviour of the former might break existing programs --
8867 where IV updating is needed, des_ncbc_encrypt can be used.
8870 *) When bntest is run from "make test" it drives bc to check its
8871 calculations, as well as internally checking them. If an internal check
8872 fails, it needs to cause bc to give a non-zero result or make test carries
8873 on without noticing the failure. Fixed.
8876 *) DES library cleanups.
8879 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8880 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8881 ciphers. NOTE: although the key derivation function has been verified
8882 against some published test vectors it has not been extensively tested
8883 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8887 *) Instead of "mkdir -p", which is not fully portable, use new
8888 Perl script "util/mkdir-p.pl".
8891 *) Rewrite the way password based encryption (PBE) is handled. It used to
8892 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8893 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8894 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8895 the 'parameter' field of the AlgorithmIdentifier is passed to the
8896 underlying key generation function so it must do its own ASN1 parsing.
8897 This has also changed the EVP_PBE_CipherInit() function which now has a
8898 'parameter' argument instead of literal salt and iteration count values
8899 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8902 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8903 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8904 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8905 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8906 value was just used as a "magic string" and not used directly its
8907 value doesn't matter.
8910 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8914 *) "linux-sparc64" configuration (ultrapenguin).
8915 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8916 "linux-sparc" configuration.
8917 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8919 *) config now generates no-xxx options for missing ciphers.
8922 *) Support the EBCDIC character set (work in progress).
8923 File ebcdic.c not yet included because it has a different license.
8924 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8926 *) Support BS2000/OSD-POSIX.
8927 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8929 *) Make callbacks for key generation use void * instead of char *.
8932 *) Make S/MIME samples compile (not yet tested).
8935 *) Additional typesafe stacks.
8938 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8942 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8944 *) New configuration variant "sco5-gcc".
8946 *) Updated some demos.
8947 [Sean O Riordain, Wade Scholine]
8949 *) Add missing BIO_free at exit of pkcs12 application.
8952 *) Fix memory leak in conf.c.
8955 *) Updates for Win32 to assembler version of MD5.
8958 *) Set #! path to perl in apps/der_chop to where we found it
8959 instead of using a fixed path.
8962 *) SHA library changes for irix64-mips4-cc.
8965 *) Improvements for VMS support.
8969 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8971 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8972 This also avoids the problems with SC4.2 and unpatched SC5.
8973 [Andy Polyakov <appro@fy.chalmers.se>]
8975 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8976 These are required because of the typesafe stack would otherwise break
8977 existing code. If old code used a structure member which used to be STACK
8978 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8979 sk_num or sk_value it would produce an error because the num, data members
8980 are not present in STACK_OF. Now it just produces a warning. sk_set
8981 replaces the old method of assigning a value to sk_value
8982 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8983 that does this will no longer work (and should use sk_set instead) but
8984 this could be regarded as a "questionable" behaviour anyway.
8987 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8988 correctly handle encrypted S/MIME data.
8991 *) Change type of various DES function arguments from des_cblock
8992 (which means, in function argument declarations, pointer to char)
8993 to des_cblock * (meaning pointer to array with 8 char elements),
8994 which allows the compiler to do more typechecking; it was like
8995 that back in SSLeay, but with lots of ugly casts.
8997 Introduce new type const_des_cblock.
9000 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9001 problems: find RecipientInfo structure that matches recipient certificate
9002 and initialise the ASN1 structures properly based on passed cipher.
9005 *) Belatedly make the BN tests actually check the results.
9008 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9009 to and from BNs: it was completely broken. New compilation option
9010 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9011 key elements as negative integers.
9014 *) Reorganize and speed up MD5.
9015 [Andy Polyakov <appro@fy.chalmers.se>]
9018 [Richard Levitte <richard@levitte.org>]
9020 *) New option -out to asn1parse to allow the parsed structure to be
9021 output to a file. This is most useful when combined with the -strparse
9022 option to examine the output of things like OCTET STRINGS.
9025 *) Make SSL library a little more fool-proof by not requiring any longer
9026 that SSL_set_{accept,connect}_state be called before
9027 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9028 in many applications because usually everything *appeared* to work as
9029 intended anyway -- now it really works as intended).
9032 *) Move openssl.cnf out of lib/.
9035 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9036 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9037 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9038 [Ralf S. Engelschall]
9040 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9041 handle PKCS#7 enveloped data properly.
9042 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9044 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9045 copying pointers. The cert_st handling is changed by this in
9046 various ways (and thus what used to be known as ctx->default_cert
9047 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9048 any longer when s->cert does not give us what we need).
9049 ssl_cert_instantiate becomes obsolete by this change.
9050 As soon as we've got the new code right (possibly it already is?),
9051 we have solved a couple of bugs of the earlier code where s->cert
9052 was used as if it could not have been shared with other SSL structures.
9054 Note that using the SSL API in certain dirty ways now will result
9055 in different behaviour than observed with earlier library versions:
9056 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9057 does not influence s as it used to.
9059 In order to clean up things more thoroughly, inside SSL_SESSION
9060 we don't use CERT any longer, but a new structure SESS_CERT
9061 that holds per-session data (if available); currently, this is
9062 the peer's certificate chain and, for clients, the server's certificate
9063 and temporary key. CERT holds only those values that can have
9064 meaningful defaults in an SSL_CTX.
9067 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9068 from the internal representation. Various PKCS#7 fixes: remove some
9069 evil casts and set the enc_dig_alg field properly based on the signing
9073 *) Allow PKCS#12 password to be set from the command line or the
9074 environment. Let 'ca' get its config file name from the environment
9075 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9079 *) Allow certificate policies extension to use an IA5STRING for the
9080 organization field. This is contrary to the PKIX definition but
9081 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9085 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9086 without disallowing inline assembler and the like for non-pedantic builds.
9089 *) Support Borland C++ builder.
9090 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9095 *) SHA-1 cleanups and performance enhancements.
9096 [Andy Polyakov <appro@fy.chalmers.se>]
9098 *) Sparc v8plus assembler for the bignum library.
9099 [Andy Polyakov <appro@fy.chalmers.se>]
9101 *) Accept any -xxx and +xxx compiler options in Configure.
9104 *) Update HPUX configuration.
9107 *) Add missing sk_<type>_unshift() function to safestack.h
9108 [Ralf S. Engelschall]
9110 *) New function SSL_CTX_use_certificate_chain_file that sets the
9111 "extra_cert"s in addition to the certificate. (This makes sense
9112 only for "PEM" format files, as chains as a whole are not
9116 *) Support verify_depth from the SSL API.
9117 x509_vfy.c had what can be considered an off-by-one-error:
9118 Its depth (which was not part of the external interface)
9119 was actually counting the number of certificates in a chain;
9120 now it really counts the depth.
9123 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9124 instead of X509err, which often resulted in confusing error
9125 messages since the error codes are not globally unique
9126 (e.g. an alleged error in ssl3_accept when a certificate
9127 didn't match the private key).
9129 *) New function SSL_CTX_set_session_id_context that allows to set a default
9130 value (so that you don't need SSL_set_session_id_context for each
9131 connection using the SSL_CTX).
9134 *) OAEP decoding bug fix.
9137 *) Support INSTALL_PREFIX for package builders, as proposed by
9141 *) New Configure options "threads" and "no-threads". For systems
9142 where the proper compiler options are known (currently Solaris
9143 and Linux), "threads" is the default.
9146 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9149 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9150 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9151 such as /usr/local/bin.
9154 *) "make linux-shared" to build shared libraries.
9155 [Niels Poppe <niels@netbox.org>]
9157 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9160 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9161 extension adding in x509 utility.
9164 *) Remove NOPROTO sections and error code comments.
9167 *) Partial rewrite of the DEF file generator to now parse the ANSI
9171 *) New Configure options --prefix=DIR and --openssldir=DIR.
9174 *) Complete rewrite of the error code script(s). It is all now handled
9175 by one script at the top level which handles error code gathering,
9176 header rewriting and C source file generation. It should be much better
9177 than the old method: it now uses a modified version of Ulf's parser to
9178 read the ANSI prototypes in all header files (thus the old K&R definitions
9179 aren't needed for error creation any more) and do a better job of
9180 translating function codes into names. The old 'ASN1 error code imbedded
9181 in a comment' is no longer necessary and it doesn't use .err files which
9182 have now been deleted. Also the error code call doesn't have to appear all
9183 on one line (which resulted in some large lines...).
9186 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9189 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9190 0 (which usually indicates a closed connection), but continue reading.
9193 *) Fix some race conditions.
9196 *) Add support for CRL distribution points extension. Add Certificate
9197 Policies and CRL distribution points documentation.
9200 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9203 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9204 8 of keying material. Merlin has also confirmed interop with this fix
9205 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9206 [Merlin Hughes <merlin@baltimore.ie>]
9208 *) Fix lots of warnings.
9209 [Richard Levitte <levitte@stacken.kth.se>]
9211 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9212 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9213 [Richard Levitte <levitte@stacken.kth.se>]
9215 *) Fix problems with sizeof(long) == 8.
9216 [Andy Polyakov <appro@fy.chalmers.se>]
9218 *) Change functions to ANSI C.
9221 *) Fix typos in error codes.
9222 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9224 *) Remove defunct assembler files from Configure.
9227 *) SPARC v8 assembler BIGNUM implementation.
9228 [Andy Polyakov <appro@fy.chalmers.se>]
9230 *) Support for Certificate Policies extension: both print and set.
9231 Various additions to support the r2i method this uses.
9234 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9235 return a const string when you are expecting an allocated buffer.
9238 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9239 types DirectoryString and DisplayText.
9242 *) Add code to allow r2i extensions to access the configuration database,
9243 add an LHASH database driver and add several ctx helper functions.
9246 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9247 fail when they extended the size of a BIGNUM.
9250 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9251 support typesafe stack.
9254 *) Fix typo in SSL_[gs]et_options().
9255 [Nils Frostberg <nils@medcom.se>]
9257 *) Delete various functions and files that belonged to the (now obsolete)
9258 old X509V3 handling code.
9261 *) New Configure option "rsaref".
9264 *) Don't auto-generate pem.h.
9267 *) Introduce type-safe ASN.1 SETs.
9270 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9271 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9273 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9274 that links with OpenSSL (well at least cause lots of warnings), but fear
9275 not: the conversion is trivial, and it eliminates loads of evil casts. A
9276 few STACKed things have been converted already. Feel free to convert more.
9277 In the fullness of time, I'll do away with the STACK type altogether.
9280 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9281 specified in <certfile> by updating the entry in the index.txt file.
9282 This way one no longer has to edit the index.txt file manually for
9283 revoking a certificate. The -revoke option does the gory details now.
9284 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9286 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9287 `-text' option at all and this way the `-noout -text' combination was
9288 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9289 [Ralf S. Engelschall]
9291 *) Make sure a corresponding plain text error message exists for the
9292 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9293 verify callback function determined that a certificate was revoked.
9294 [Ralf S. Engelschall]
9296 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9297 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9298 all available cipers including rc5, which was forgotten until now.
9299 In order to let the testing shell script know which algorithms
9300 are available, a new (up to now undocumented) command
9301 "openssl list-cipher-commands" is used.
9304 *) Bugfix: s_client occasionally would sleep in select() when
9305 it should have checked SSL_pending() first.
9308 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9309 the raw DSA values prior to ASN.1 encoding.
9312 *) Tweaks to Configure
9313 [Niels Poppe <niels@netbox.org>]
9315 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9319 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9322 *) New config option to avoid instructions that are illegal on the 80386.
9323 The default code is faster, but requires at least a 486.
9326 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9327 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9328 same as SSL2_VERSION anyway.
9331 *) New "-showcerts" option for s_client.
9334 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9335 application. Various cleanups and fixes.
9338 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9339 modify error routines to work internally. Add error codes and PBE init
9340 to library startup routines.
9343 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9344 packing functions to asn1 and evp. Changed function names and error
9345 codes along the way.
9348 *) PKCS12 integration: and so it begins... First of several patches to
9349 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9350 objects to objects.h
9353 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9354 and display support for Thawte strong extranet extension.
9357 *) Add LinuxPPC support.
9358 [Jeff Dubrule <igor@pobox.org>]
9360 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9361 bn_div_words in alpha.s.
9362 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9364 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9365 OAEP isn't supported when OpenSSL is built with RSAref.
9366 [Ulf Moeller <ulf@fitug.de>]
9368 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9369 so they no longer are missing under -DNOPROTO.
9370 [Soren S. Jorvang <soren@t.dk>]
9373 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9375 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9376 doesn't work when the session is reused. Coming soon!
9379 *) Fix a security hole, that allows sessions to be reused in the wrong
9380 context thus bypassing client cert protection! All software that uses
9381 client certs and session caches in multiple contexts NEEDS PATCHING to
9382 allow session reuse! A fuller solution is in the works.
9383 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9385 *) Some more source tree cleanups (removed obsolete files
9386 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9387 permission on "config" script to be executable) and a fix for the INSTALL
9389 [Ulf Moeller <ulf@fitug.de>]
9391 *) Remove some legacy and erroneous uses of malloc, free instead of
9393 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9395 *) Make rsa_oaep_test return non-zero on error.
9396 [Ulf Moeller <ulf@fitug.de>]
9398 *) Add support for native Solaris shared libraries. Configure
9399 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9400 if someone would make that last step automatic.
9401 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9403 *) ctx_size was not built with the right compiler during "make links". Fixed.
9406 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9407 except NULL ciphers". This means the default cipher list will no longer
9408 enable NULL ciphers. They need to be specifically enabled e.g. with
9409 the string "DEFAULT:eNULL".
9412 *) Fix to RSA private encryption routines: if p < q then it would
9413 occasionally produce an invalid result. This will only happen with
9414 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9417 *) Be less restrictive and allow also `perl util/perlpath.pl
9418 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9419 because this way one can also use an interpreter named `perl5' (which is
9420 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9421 installed as `perl').
9422 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9424 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9425 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9427 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9428 advapi32.lib to Win32 build and change the pem test comparision
9429 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9430 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9431 and crypto/des/ede_cbcm_enc.c.
9434 *) DES quad checksum was broken on big-endian architectures. Fixed.
9437 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9438 Win32 test batch file so it (might) work again. The Win32 test batch file
9439 is horrible: I feel ill....
9442 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9443 in e_os.h. Audit of header files to check ANSI and non ANSI
9444 sections: 10 functions were absent from non ANSI section and not exported
9445 from Windows DLLs. Fixed up libeay.num for new functions.
9448 *) Make `openssl version' output lines consistent.
9449 [Ralf S. Engelschall]
9451 *) Fix Win32 symbol export lists for BIO functions: Added
9452 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9453 to ms/libeay{16,32}.def.
9454 [Ralf S. Engelschall]
9456 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9457 fine under Unix and passes some trivial tests I've now added. But the
9458 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9459 added to make sure no one expects that this stuff really works in the
9460 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9461 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9463 [Ralf S. Engelschall]
9465 *) Fix the generation of two part addresses in perl.
9466 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9468 *) Add config entry for Linux on MIPS.
9469 [John Tobey <jtobey@channel1.com>]
9471 *) Make links whenever Configure is run, unless we are on Windoze.
9474 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9475 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9479 *) Add a useful kludge to allow package maintainers to specify compiler and
9480 other platforms details on the command line without having to patch the
9481 Configure script everytime: One now can use ``perl Configure
9482 <id>:<details>'', i.e. platform ids are allowed to have details appended
9483 to them (seperated by colons). This is treated as there would be a static
9484 pre-configured entry in Configure's %table under key <id> with value
9485 <details> and ``perl Configure <id>'' is called. So, when you want to
9486 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9487 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9488 now, which overrides the FreeBSD-elf entry on-the-fly.
9489 [Ralf S. Engelschall]
9491 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9494 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9495 on the `perl Configure ...' command line. This way one can compile
9496 OpenSSL libraries with Position Independent Code (PIC) which is needed
9497 for linking it into DSOs.
9498 [Ralf S. Engelschall]
9500 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9504 *) Cleaned up the LICENSE document: The official contact for any license
9505 questions now is the OpenSSL core team under openssl-core@openssl.org.
9506 And add a paragraph about the dual-license situation to make sure people
9507 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9508 to the OpenSSL toolkit.
9509 [Ralf S. Engelschall]
9511 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9512 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9513 Additonally cleaned up the `make links' target: Remove unnecessary
9514 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9515 to speed processing and no longer clutter the display with confusing
9516 stuff. Instead only the actually done links are displayed.
9517 [Ralf S. Engelschall]
9519 *) Permit null encryption ciphersuites, used for authentication only. It used
9520 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9521 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9525 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9526 signed attributes when verifying signatures (this would break them),
9527 the detached data encoding was wrong and public keys obtained using
9528 X509_get_pubkey() weren't freed.
9531 *) Add text documentation for the BUFFER functions. Also added a work around
9532 to a Win95 console bug. This was triggered by the password read stuff: the
9533 last character typed gets carried over to the next fread(). If you were
9534 generating a new cert request using 'req' for example then the last
9535 character of the passphrase would be CR which would then enter the first
9539 *) Added the new `Includes OpenSSL Cryptography Software' button as
9540 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9541 button and can be used by applications based on OpenSSL to show the
9542 relationship to the OpenSSL project.
9543 [Ralf S. Engelschall]
9545 *) Remove confusing variables in function signatures in files
9546 ssl/ssl_lib.c and ssl/ssl.h.
9547 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9549 *) Don't install bss_file.c under PREFIX/include/
9550 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9552 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9553 functions that return function pointers and has support for NT specific
9554 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9555 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9556 unsigned to signed types: this was killing the Win32 compile.
9559 *) Add new certificate file to stack functions,
9560 SSL_add_dir_cert_subjects_to_stack() and
9561 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9562 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9563 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9564 This means that Apache-SSL and similar packages don't have to mess around
9565 to add as many CAs as they want to the preferred list.
9568 *) Experiment with doxygen documentation. Currently only partially applied to
9570 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9571 openssl.doxy as the configuration file.
9574 *) Get rid of remaining C++-style comments which strict C compilers hate.
9575 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9577 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9578 compiled in by default: it has problems with large keys.
9581 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9582 DH private keys and/or callback functions which directly correspond to
9583 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9584 is needed for applications which have to configure certificates on a
9585 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9587 For the RSA certificate situation is makes no difference, but
9588 for the DSA certificate situation this fixes the "no shared cipher"
9589 problem where the OpenSSL cipher selection procedure failed because the
9590 temporary keys were not overtaken from the context and the API provided
9591 no way to reconfigure them.
9592 The new functions now let applications reconfigure the stuff and they
9593 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9594 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9595 non-public-API function ssl_cert_instantiate() is used as a helper
9596 function and also to reduce code redundancy inside ssl_rsa.c.
9597 [Ralf S. Engelschall]
9599 *) Move s_server -dcert and -dkey options out of the undocumented feature
9600 area because they are useful for the DSA situation and should be
9601 recognized by the users.
9602 [Ralf S. Engelschall]
9604 *) Fix the cipher decision scheme for export ciphers: the export bits are
9605 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9606 SSL_EXP_MASK. So, the original variable has to be used instead of the
9607 already masked variable.
9608 [Richard Levitte <levitte@stacken.kth.se>]
9610 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9611 [Richard Levitte <levitte@stacken.kth.se>]
9613 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9614 from `int' to `unsigned int' because it's a length and initialized by
9615 EVP_DigestFinal() which expects an `unsigned int *'.
9616 [Richard Levitte <levitte@stacken.kth.se>]
9618 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9619 script. Instead use the usual Shell->Perl transition trick.
9620 [Ralf S. Engelschall]
9622 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9623 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9624 -noout -modulus' as it's already the case for `openssl rsa -noout
9625 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9626 currently the public key is printed (a decision which was already done by
9627 `openssl dsa -modulus' in the past) which serves a similar purpose.
9628 Additionally the NO_RSA no longer completely removes the whole -modulus
9629 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9631 [Ralf S. Engelschall]
9633 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9634 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9635 [Arne Ansper <arne@ats.cyber.ee>]
9637 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9638 to be added. Now both 'req' and 'ca' can use new objects defined in the
9642 *) Add cool BIO that does syslog (or event log on NT).
9643 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9645 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9646 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9647 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9648 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9651 *) Add preliminary config info for new extension code.
9654 *) Make RSA_NO_PADDING really use no padding.
9655 [Ulf Moeller <ulf@fitug.de>]
9657 *) Generate errors when private/public key check is done.
9660 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9661 for some CRL extensions and new objects added.
9664 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9665 key usage extension and fuller support for authority key id.
9668 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9669 padding method for RSA, which is recommended for new applications in PKCS
9670 #1 v2.0 (RFC 2437, October 1998).
9671 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9672 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9673 against Bleichbacher's attack on RSA.
9674 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9677 *) Updates to the new SSL compression code
9678 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9680 *) Fix so that the version number in the master secret, when passed
9681 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9682 (because the server will not accept higher), that the version number
9683 is 0x03,0x01, not 0x03,0x00
9684 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9686 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9687 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9688 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9691 *) Support for RAW extensions where an arbitrary extension can be
9692 created by including its DER encoding. See apps/openssl.cnf for
9696 *) Make sure latest Perl versions don't interpret some generated C array
9697 code as Perl array code in the crypto/err/err_genc.pl script.
9698 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9700 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9701 not many people have the assembler. Various Win32 compilation fixes and
9702 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9706 *) Modify configure script 'Configure' to automatically create crypto/date.h
9707 file under Win32 and also build pem.h from pem.org. New script
9708 util/mkfiles.pl to create the MINFO file on environments that can't do a
9709 'make files': perl util/mkfiles.pl >MINFO should work.
9712 *) Major rework of DES function declarations, in the pursuit of correctness
9713 and purity. As a result, many evil casts evaporated, and some weirdness,
9714 too. You may find this causes warnings in your code. Zapping your evil
9715 casts will probably fix them. Mostly.
9718 *) Fix for a typo in asn1.h. Bug fix to object creation script
9719 obj_dat.pl. It considered a zero in an object definition to mean
9720 "end of object": none of the objects in objects.h have any zeros
9721 so it wasn't spotted.
9722 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9724 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9725 Masking (CBCM). In the absence of test vectors, the best I have been able
9726 to do is check that the decrypt undoes the encrypt, so far. Send me test
9727 vectors if you have them.
9730 *) Correct calculation of key length for export ciphers (too much space was
9731 allocated for null ciphers). This has not been tested!
9734 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9735 message is now correct (it understands "crypto" and "ssl" on its
9736 command line). There is also now an "update" option. This will update
9737 the util/ssleay.num and util/libeay.num files with any new functions.
9739 perl util/mkdef.pl crypto ssl update
9740 it will update them.
9743 *) Overhauled the Perl interface (perl/*):
9744 - ported BN stuff to OpenSSL's different BN library
9745 - made the perl/ source tree CVS-aware
9746 - renamed the package from SSLeay to OpenSSL (the files still contain
9747 their history because I've copied them in the repository)
9748 - removed obsolete files (the test scripts will be replaced
9749 by better Test::Harness variants in the future)
9750 [Ralf S. Engelschall]
9752 *) First cut for a very conservative source tree cleanup:
9753 1. merge various obsolete readme texts into doc/ssleay.txt
9754 where we collect the old documents and readme texts.
9755 2. remove the first part of files where I'm already sure that we no
9756 longer need them because of three reasons: either they are just temporary
9757 files which were left by Eric or they are preserved original files where
9758 I've verified that the diff is also available in the CVS via "cvs diff
9759 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9760 the crypto/md/ stuff).
9761 [Ralf S. Engelschall]
9763 *) More extension code. Incomplete support for subject and issuer alt
9764 name, issuer and authority key id. Change the i2v function parameters
9765 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9766 what that's for :-) Fix to ASN1 macro which messed up
9767 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9770 *) Preliminary support for ENUMERATED type. This is largely copied from the
9774 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9775 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9777 *) Make sure `make rehash' target really finds the `openssl' program.
9778 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9780 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9781 like to hear about it if this slows down other processors.
9784 *) Add CygWin32 platform information to Configure script.
9785 [Alan Batie <batie@aahz.jf.intel.com>]
9787 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9788 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9790 *) New program nseq to manipulate netscape certificate sequences
9793 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9797 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9798 but the BN code had some problems that would cause failures when
9799 doing certificate verification and some other functions.
9800 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9802 *) Add ASN1 and PEM code to support netscape certificate sequences.
9805 *) Add ASN1 and PEM code to support netscape certificate sequences.
9808 *) Add several PKIX and private extended key usage OIDs.
9811 *) Modify the 'ca' program to handle the new extension code. Modify
9812 openssl.cnf for new extension format, add comments.
9815 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9816 and add a sample to openssl.cnf so req -x509 now adds appropriate
9820 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9821 error code, add initial support to X509_print() and x509 application.
9824 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9825 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9826 stuff is currently isolated and isn't even compiled yet.
9829 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9830 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9831 Removed the versions check from X509 routines when loading extensions:
9832 this allows certain broken certificates that don't set the version
9833 properly to be processed.
9836 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9837 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9838 can still be regenerated with "make depend".
9841 *) Spelling mistake in C version of CAST-128.
9842 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9844 *) Changes to the error generation code. The perl script err-code.pl
9845 now reads in the old error codes and retains the old numbers, only
9846 adding new ones if necessary. It also only changes the .err files if new
9847 codes are added. The makefiles have been modified to only insert errors
9848 when needed (to avoid needlessly modifying header files). This is done
9849 by only inserting errors if the .err file is newer than the auto generated
9850 C file. To rebuild all the error codes from scratch (the old behaviour)
9851 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9852 or delete all the .err files.
9855 *) CAST-128 was incorrectly implemented for short keys. The C version has
9856 been fixed, but is untested. The assembler versions are also fixed, but
9857 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9858 to regenerate it if needed.
9859 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9860 Hagino <itojun@kame.net>]
9862 *) File was opened incorrectly in randfile.c.
9863 [Ulf Möller <ulf@fitug.de>]
9865 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9866 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9867 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9868 al: it's just almost always a UTCTime. Note this patch adds new error
9869 codes so do a "make errors" if there are problems.
9872 *) Correct Linux 1 recognition in config.
9873 [Ulf Möller <ulf@fitug.de>]
9875 *) Remove pointless MD5 hash when using DSA keys in ca.
9876 [Anonymous <nobody@replay.com>]
9878 *) Generate an error if given an empty string as a cert directory. Also
9879 generate an error if handed NULL (previously returned 0 to indicate an
9880 error, but didn't set one).
9881 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9883 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9886 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9887 parameters. This was causing a warning which killed off the Win32 compile.
9890 *) Remove C++ style comments from crypto/bn/bn_local.h.
9891 [Neil Costigan <neil.costigan@celocom.com>]
9893 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9894 based on a text string, looking up short and long names and finally
9895 "dot" format. The "dot" format stuff didn't work. Added new function
9896 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9897 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9898 OID is not part of the table.
9901 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9902 X509_LOOKUP_by_alias().
9905 *) Sort openssl functions by name.
9908 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9909 encryption from sample DSA keys (in case anyone is interested the password
9913 *) Make _all_ *_free functions accept a NULL pointer.
9914 [Frans Heymans <fheymans@isaserver.be>]
9916 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9918 [Anonymous <nobody@replay.com>]
9920 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9921 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9923 *) Don't blow it for numeric -newkey arguments to apps/req.
9924 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9926 *) Temp key "for export" tests were wrong in s3_srvr.c.
9927 [Anonymous <nobody@replay.com>]
9929 *) Add prototype for temp key callback functions
9930 SSL_CTX_set_tmp_{rsa,dh}_callback().
9933 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9934 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9937 *) X509_name_add_entry() freed the wrong thing after an error.
9938 [Arne Ansper <arne@ats.cyber.ee>]
9940 *) rsa_eay.c would attempt to free a NULL context.
9941 [Arne Ansper <arne@ats.cyber.ee>]
9943 *) BIO_s_socket() had a broken should_retry() on Windoze.
9944 [Arne Ansper <arne@ats.cyber.ee>]
9946 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9947 [Arne Ansper <arne@ats.cyber.ee>]
9949 *) Make sure the already existing X509_STORE->depth variable is initialized
9950 in X509_STORE_new(), but document the fact that this variable is still
9951 unused in the certificate verification process.
9952 [Ralf S. Engelschall]
9954 *) Fix the various library and apps files to free up pkeys obtained from
9955 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9958 *) Fix reference counting in X509_PUBKEY_get(). This makes
9959 demos/maurice/example2.c work, amongst others, probably.
9960 [Steve Henson and Ben Laurie]
9962 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9963 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9964 are no longer created. This way we have a single and consistent command
9965 line interface `openssl <command>', similar to `cvs <command>'.
9966 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9968 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9969 BIT STRING wrapper always have zero unused bits.
9972 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9975 *) Make the top-level INSTALL documentation easier to understand.
9978 *) Makefiles updated to exit if an error occurs in a sub-directory
9979 make (including if user presses ^C) [Paul Sutton]
9981 *) Make Montgomery context stuff explicit in RSA data structure.
9984 *) Fix build order of pem and err to allow for generated pem.h.
9987 *) Fix renumbering bug in X509_NAME_delete_entry().
9990 *) Enhanced the err-ins.pl script so it makes the error library number
9991 global and can add a library name. This is needed for external ASN1 and
9992 other error libraries.
9995 *) Fixed sk_insert which never worked properly.
9998 *) Fix ASN1 macros so they can handle indefinite length construted
9999 EXPLICIT tags. Some non standard certificates use these: they can now
10003 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10004 into a single doc/ssleay.txt bundle. This way the information is still
10005 preserved but no longer messes up this directory. Now it's new room for
10006 the new set of documenation files.
10007 [Ralf S. Engelschall]
10009 *) SETs were incorrectly DER encoded. This was a major pain, because they
10010 shared code with SEQUENCEs, which aren't coded the same. This means that
10011 almost everything to do with SETs or SEQUENCEs has either changed name or
10012 number of arguments.
10013 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10015 *) Fix test data to work with the above.
10018 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10019 was already fixed by Eric for 0.9.1 it seems.
10020 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10022 *) Autodetect FreeBSD3.
10025 *) Fix various bugs in Configure. This affects the following platforms:
10029 unixware-2.0-pentium
10033 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10034 before they are needed.
10037 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10041 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10043 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10044 changed SSLeay to OpenSSL in version strings.
10045 [Ralf S. Engelschall]
10047 *) Some fixups to the top-level documents.
10050 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10051 because the symlink to include/ was missing.
10052 [Ralf S. Engelschall]
10054 *) Incorporated the popular no-RSA/DSA-only patches
10055 which allow to compile a RSA-free SSLeay.
10056 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10058 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10059 when "ssleay" is still not found.
10060 [Ralf S. Engelschall]
10062 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10063 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10065 *) Updated the README file.
10066 [Ralf S. Engelschall]
10068 *) Added various .cvsignore files in the CVS repository subdirs
10069 to make a "cvs update" really silent.
10070 [Ralf S. Engelschall]
10072 *) Recompiled the error-definition header files and added
10073 missing symbols to the Win32 linker tables.
10074 [Ralf S. Engelschall]
10076 *) Cleaned up the top-level documents;
10077 o new files: CHANGES and LICENSE
10078 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10079 o merged COPYRIGHT into LICENSE
10080 o removed obsolete TODO file
10081 o renamed MICROSOFT to INSTALL.W32
10082 [Ralf S. Engelschall]
10084 *) Removed dummy files from the 0.9.1b source tree:
10085 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10086 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10087 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10088 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10089 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10090 [Ralf S. Engelschall]
10092 *) Added various platform portability fixes.
10095 *) The Genesis of the OpenSSL rpject:
10096 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10097 Young and Tim J. Hudson created while they were working for C2Net until
10099 [The OpenSSL Project]
10102 Changes between 0.9.0b and 0.9.1b [not released]
10104 *) Updated a few CA certificates under certs/
10107 *) Changed some BIGNUM api stuff.
10110 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10111 DGUX x86, Linux Alpha, etc.
10114 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10115 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10119 *) Add -strparse option to asn1pars program which parses nested
10121 [Dr Stephen Henson <shenson@bigfoot.com>]
10123 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10126 *) DSA fix for "ca" program.
10129 *) Added "-genkey" option to "dsaparam" program.
10132 *) Added RIPE MD160 (rmd160) message digest.
10135 *) Added -a (all) option to "ssleay version" command.
10138 *) Added PLATFORM define which is the id given to Configure.
10141 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10144 *) Extended the ASN.1 parser routines.
10147 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10150 *) Added a BN_CTX to the BN library.
10153 *) Fixed the weak key values in DES library
10156 *) Changed API in EVP library for cipher aliases.
10159 *) Added support for RC2/64bit cipher.
10162 *) Converted the lhash library to the crypto/mem.c functions.
10165 *) Added more recognized ASN.1 object ids.
10168 *) Added more RSA padding checks for SSL/TLS.
10171 *) Added BIO proxy/filter functionality.
10174 *) Added extra_certs to SSL_CTX which can be used
10175 send extra CA certificates to the client in the CA cert chain sending
10176 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10179 *) Now Fortezza is denied in the authentication phase because
10180 this is key exchange mechanism is not supported by SSLeay at all.
10183 *) Additional PKCS1 checks.
10186 *) Support the string "TLSv1" for all TLS v1 ciphers.
10189 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10190 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10193 *) Fixed a few memory leaks.
10196 *) Fixed various code and comment typos.
10199 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10200 bytes sent in the client random.
10201 [Edward Bishop <ebishop@spyglass.com>]