5 Changes between 1.0.1e and 1.0.1f [xx XXX xxxx]
7 *) Fix for TLS record tampering bug. A carefully crafted invalid
8 handshake could crash OpenSSL with a NULL pointer exception.
9 Thanks to Anton Johansson for reporting this issues.
12 *) Keep original DTLS digest and encryption contexts in retransmission
13 structures so we can use the previous session parameters if they need
14 to be resent. (CVE-2013-6450)
17 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
18 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
19 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
20 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
21 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
22 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
23 [Rob Stradling, Adam Langley]
25 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
27 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
28 supporting platforms or when small records were transferred.
29 [Andy Polyakov, Steve Henson]
31 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
33 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
35 This addresses the flaw in CBC record processing discovered by
36 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
37 at: http://www.isg.rhul.ac.uk/tls/
39 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
40 Security Group at Royal Holloway, University of London
41 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
42 Emilia Käsper for the initial patch.
44 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
46 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
47 ciphersuites which can be exploited in a denial of service attack.
48 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
49 and detecting this bug and to Wolfgang Ettlinger
50 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
54 *) Return an error when checking OCSP signatures when key is NULL.
55 This fixes a DoS attack. (CVE-2013-0166)
58 *) Make openssl verify return errors.
59 [Chris Palmer <palmer@google.com> and Ben Laurie]
61 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
62 the right response is stapled. Also change SSL_get_certificate()
63 so it returns the certificate actually sent.
64 See http://rt.openssl.org/Ticket/Display.html?id=2836.
65 [Rob Stradling <rob.stradling@comodo.com>]
67 *) Fix possible deadlock when decoding public keys.
70 *) Don't use TLS 1.0 record version number in initial client hello
74 Changes between 1.0.1b and 1.0.1c [10 May 2012]
76 *) Sanity check record length before skipping explicit IV in TLS
77 1.2, 1.1 and DTLS to fix DoS attack.
79 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
80 fuzzing as a service testing platform.
84 *) Initialise tkeylen properly when encrypting CMS messages.
85 Thanks to Solar Designer of Openwall for reporting this issue.
88 *) In FIPS mode don't try to use composite ciphers as they are not
92 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
94 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
95 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
96 mean any application compiled against OpenSSL 1.0.0 headers setting
97 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
98 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
99 0x10000000L Any application which was previously compiled against
100 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
101 will need to be recompiled as a result. Letting be results in
102 inability to disable specifically TLS 1.1 and in client context,
103 in unlike event, limit maximum offered version to TLS 1.0 [see below].
106 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
107 disable just protocol X, but all protocols above X *if* there are
108 protocols *below* X still enabled. In more practical terms it means
109 that if application wants to disable TLS1.0 in favor of TLS1.1 and
110 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
111 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
115 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
117 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
118 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
119 in CRYPTO_realloc_clean.
121 Thanks to Tavis Ormandy, Google Security Team, for discovering this
122 issue and to Adam Langley <agl@chromium.org> for fixing it.
124 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
126 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
129 *) Workarounds for some broken servers that "hang" if a client hello
130 record length exceeds 255 bytes.
132 1. Do not use record version number > TLS 1.0 in initial client
133 hello: some (but not all) hanging servers will now work.
134 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
135 the number of ciphers sent in the client hello. This should be
136 set to an even number, such as 50, for example by passing:
137 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
138 Most broken servers should now work.
139 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
140 TLS 1.2 client support entirely.
143 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
146 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
148 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
149 STRING form instead of a DigestInfo.
152 *) The format used for MDC2 RSA signatures is inconsistent between EVP
153 and the RSA_sign/RSA_verify functions. This was made more apparent when
154 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
155 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
156 the correct format in RSA_verify so both forms transparently work.
159 *) Some servers which support TLS 1.0 can choke if we initially indicate
160 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
161 encrypted premaster secret. As a workaround use the maximum pemitted
162 client version in client hello, this should keep such servers happy
163 and still work with previous versions of OpenSSL.
166 *) Add support for TLS/DTLS heartbeats.
167 [Robin Seggelmann <seggelmann@fh-muenster.de>]
169 *) Add support for SCTP.
170 [Robin Seggelmann <seggelmann@fh-muenster.de>]
172 *) Improved PRNG seeding for VOS.
173 [Paul Green <Paul.Green@stratus.com>]
175 *) Extensive assembler packs updates, most notably:
177 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
178 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
179 - x86_64: bit-sliced AES implementation;
180 - ARM: NEON support, contemporary platforms optimizations;
181 - s390x: z196 support;
182 - *: GHASH and GF(2^m) multiplication implementations;
186 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
187 (removal of unnecessary code)
188 [Peter Sylvester <peter.sylvester@edelweb.fr>]
190 *) Add TLS key material exporter from RFC 5705.
193 *) Add DTLS-SRTP negotiation from RFC 5764.
196 *) Add Next Protocol Negotiation,
197 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
198 disabled with a no-npn flag to config or Configure. Code donated
200 [Adam Langley <agl@google.com> and Ben Laurie]
202 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
203 NIST-P256, NIST-P521, with constant-time single point multiplication on
204 typical inputs. Compiler support for the nonstandard type __uint128_t is
205 required to use this (present in gcc 4.4 and later, for 64-bit builds).
206 Code made available under Apache License version 2.0.
208 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
209 line to include this in your build of OpenSSL, and run "make depend" (or
210 "make update"). This enables the following EC_METHODs:
212 EC_GFp_nistp224_method()
213 EC_GFp_nistp256_method()
214 EC_GFp_nistp521_method()
216 EC_GROUP_new_by_curve_name() will automatically use these (while
217 EC_GROUP_new_curve_GFp() currently prefers the more flexible
219 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
221 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
222 all platforms. Move ssize_t definition from e_os.h to the public
223 header file e_os2.h as it now appears in public header file cms.h
226 *) New -sigopt option to the ca, req and x509 utilities. Additional
227 signature parameters can be passed using this option and in
231 *) Add RSA PSS signing function. This will generate and set the
232 appropriate AlgorithmIdentifiers for PSS based on those in the
233 corresponding EVP_MD_CTX structure. No application support yet.
236 *) Support for companion algorithm specific ASN1 signing routines.
237 New function ASN1_item_sign_ctx() signs a pre-initialised
238 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
239 the appropriate parameters.
242 *) Add new algorithm specific ASN1 verification initialisation function
243 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
244 handling will be the same no matter what EVP_PKEY_METHOD is used.
245 Add a PSS handler to support verification of PSS signatures: checked
246 against a number of sample certificates.
249 *) Add signature printing for PSS. Add PSS OIDs.
250 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
252 *) Add algorithm specific signature printing. An individual ASN1 method
253 can now print out signatures instead of the standard hex dump.
255 More complex signatures (e.g. PSS) can print out more meaningful
256 information. Include DSA version that prints out the signature
260 *) Password based recipient info support for CMS library: implementing
264 *) Split password based encryption into PBES2 and PBKDF2 functions. This
265 neatly separates the code into cipher and PBE sections and is required
266 for some algorithms that split PBES2 into separate pieces (such as
270 *) Session-handling fixes:
271 - Fix handling of connections that are resuming with a session ID,
272 but also support Session Tickets.
273 - Fix a bug that suppressed issuing of a new ticket if the client
274 presented a ticket with an expired session.
275 - Try to set the ticket lifetime hint to something reasonable.
276 - Make tickets shorter by excluding irrelevant information.
277 - On the client side, don't ignore renewed tickets.
278 [Adam Langley, Bodo Moeller (Google)]
280 *) Fix PSK session representation.
283 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
285 This work was sponsored by Intel.
288 *) Add GCM support to TLS library. Some custom code is needed to split
289 the IV between the fixed (from PRF) and explicit (from TLS record)
290 portions. This adds all GCM ciphersuites supported by RFC5288 and
291 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
292 add a special AESGCM string for GCM only.
295 *) Expand range of ctrls for AES GCM. Permit setting invocation
296 field on decrypt and retrieval of invocation field only on encrypt.
299 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
300 As required by RFC5289 these ciphersuites cannot be used if for
301 versions of TLS earlier than 1.2.
304 *) For FIPS capable OpenSSL interpret a NULL default public key method
305 as unset and return the appopriate default but do *not* set the default.
306 This means we can return the appopriate method in applications that
307 swicth between FIPS and non-FIPS modes.
310 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
311 ENGINE is used then we cannot handle that in the FIPS module so we
312 keep original code iff non-FIPS operations are allowed.
315 *) Add -attime option to openssl utilities.
316 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
318 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
321 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
322 FIPS EC methods unconditionally for now.
325 *) New build option no-ec2m to disable characteristic 2 code.
328 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
329 all cases can be covered as some introduce binary incompatibilities.
332 *) Redirect RSA operations to FIPS module including keygen,
333 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
336 *) Add similar low level API blocking to ciphers.
339 *) Low level digest APIs are not approved in FIPS mode: any attempt
340 to use these will cause a fatal error. Applications that *really* want
341 to use them can use the private_* version instead.
344 *) Redirect cipher operations to FIPS module for FIPS builds.
347 *) Redirect digest operations to FIPS module for FIPS builds.
350 *) Update build system to add "fips" flag which will link in fipscanister.o
351 for static and shared library builds embedding a signature if needed.
354 *) Output TLS supported curves in preference order instead of numerical
355 order. This is currently hardcoded for the highest order curves first.
356 This should be configurable so applications can judge speed vs strength.
359 *) Add TLS v1.2 server support for client authentication.
362 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
366 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
367 FIPS modules versions.
370 *) Add TLS v1.2 client side support for client authentication. Keep cache
371 of handshake records longer as we don't know the hash algorithm to use
372 until after the certificate request message is received.
375 *) Initial TLS v1.2 client support. Add a default signature algorithms
376 extension including all the algorithms we support. Parse new signature
377 format in client key exchange. Relax some ECC signing restrictions for
378 TLS v1.2 as indicated in RFC5246.
381 *) Add server support for TLS v1.2 signature algorithms extension. Switch
382 to new signature format when needed using client digest preference.
383 All server ciphersuites should now work correctly in TLS v1.2. No client
384 support yet and no support for client certificates.
387 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
388 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
389 ciphersuites. At present only RSA key exchange ciphersuites work with
390 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
391 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
392 and version checking.
395 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
396 with this defined it will not be affected by any changes to ssl internal
397 structures. Add several utility functions to allow openssl application
398 to work with OPENSSL_NO_SSL_INTERN defined.
402 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
404 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
407 *) Permit abbreviated handshakes when renegotiating using the function
408 SSL_renegotiate_abbreviated().
409 [Robin Seggelmann <seggelmann@fh-muenster.de>]
411 *) Add call to ENGINE_register_all_complete() to
412 ENGINE_load_builtin_engines(), so some implementations get used
413 automatically instead of needing explicit application support.
416 *) Add support for TLS key exporter as described in RFC5705.
417 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
419 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
420 a few changes are required:
422 Add SSL_OP_NO_TLSv1_1 flag.
424 Update version checking logic to handle version 1.1.
425 Add explicit IV handling (ported from DTLS code).
426 Add command line options to s_client/s_server.
429 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
431 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
433 This addresses the flaw in CBC record processing discovered by
434 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
435 at: http://www.isg.rhul.ac.uk/tls/
437 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
438 Security Group at Royal Holloway, University of London
439 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
440 Emilia Käsper for the initial patch.
442 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
444 *) Return an error when checking OCSP signatures when key is NULL.
445 This fixes a DoS attack. (CVE-2013-0166)
448 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
449 the right response is stapled. Also change SSL_get_certificate()
450 so it returns the certificate actually sent.
451 See http://rt.openssl.org/Ticket/Display.html?id=2836.
453 [Rob Stradling <rob.stradling@comodo.com>]
455 *) Fix possible deadlock when decoding public keys.
458 Changes between 1.0.0i and 1.0.0j [10 May 2012]
460 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
463 *) Sanity check record length before skipping explicit IV in DTLS
466 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
467 fuzzing as a service testing platform.
471 *) Initialise tkeylen properly when encrypting CMS messages.
472 Thanks to Solar Designer of Openwall for reporting this issue.
475 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
477 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
478 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
479 in CRYPTO_realloc_clean.
481 Thanks to Tavis Ormandy, Google Security Team, for discovering this
482 issue and to Adam Langley <agl@chromium.org> for fixing it.
484 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
486 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
488 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
489 in CMS and PKCS7 code. When RSA decryption fails use a random key for
490 content decryption and always return the same error. Note: this attack
491 needs on average 2^20 messages so it only affects automated senders. The
492 old behaviour can be reenabled in the CMS code by setting the
493 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
494 an MMA defence is not necessary.
495 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
496 this issue. (CVE-2012-0884)
499 *) Fix CVE-2011-4619: make sure we really are receiving a
500 client hello before rejecting multiple SGC restarts. Thanks to
501 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
504 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
506 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
507 Thanks to Antonio Martin, Enterprise Secure Access Research and
508 Development, Cisco Systems, Inc. for discovering this bug and
509 preparing a fix. (CVE-2012-0050)
512 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
514 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
515 of the Vaudenay padding oracle attack on CBC mode encryption
516 which enables an efficient plaintext recovery attack against
517 the OpenSSL implementation of DTLS. Their attack exploits timing
518 differences arising during decryption processing. A research
519 paper describing this attack can be found at:
520 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
521 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
522 Security Group at Royal Holloway, University of London
523 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
524 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
525 for preparing the fix. (CVE-2011-4108)
526 [Robin Seggelmann, Michael Tuexen]
528 *) Clear bytes used for block padding of SSL 3.0 records.
530 [Adam Langley (Google)]
532 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
533 Kadianakis <desnacked@gmail.com> for discovering this issue and
534 Adam Langley for preparing the fix. (CVE-2011-4619)
535 [Adam Langley (Google)]
537 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
538 [Andrey Kulikov <amdeich@gmail.com>]
540 *) Prevent malformed RFC3779 data triggering an assertion failure.
541 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
542 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
543 [Rob Austein <sra@hactrn.net>]
545 *) Improved PRNG seeding for VOS.
546 [Paul Green <Paul.Green@stratus.com>]
548 *) Fix ssl_ciph.c set-up race.
549 [Adam Langley (Google)]
551 *) Fix spurious failures in ecdsatest.c.
552 [Emilia Käsper (Google)]
554 *) Fix the BIO_f_buffer() implementation (which was mixing different
555 interpretations of the '..._len' fields).
556 [Adam Langley (Google)]
558 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
559 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
560 threads won't reuse the same blinding coefficients.
562 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
563 lock to call BN_BLINDING_invert_ex, and avoids one use of
564 BN_BLINDING_update for each BN_BLINDING structure (previously,
565 the last update always remained unused).
566 [Emilia Käsper (Google)]
568 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
569 [Bob Buckholz (Google)]
571 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
573 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
574 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
575 [Kaspar Brand <ossl@velox.ch>]
577 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
578 for multi-threaded use of ECDH. (CVE-2011-3210)
579 [Adam Langley (Google)]
581 *) Fix x509_name_ex_d2i memory leak on bad inputs.
584 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
585 signature public key algorithm by using OID xref utilities instead.
586 Before this you could only use some ECC ciphersuites with SHA1 only.
589 *) Add protection against ECDSA timing attacks as mentioned in the paper
590 by Billy Bob Brumley and Nicola Tuveri, see:
592 http://eprint.iacr.org/2011/232.pdf
594 [Billy Bob Brumley and Nicola Tuveri]
596 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
598 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
599 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
601 *) Fix bug in string printing code: if *any* escaping is enabled we must
602 escape the escape character (backslash) or the resulting string is
606 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
608 *) Disable code workaround for ancient and obsolete Netscape browsers
609 and servers: an attacker can use it in a ciphersuite downgrade attack.
610 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
613 *) Fixed J-PAKE implementation error, originally discovered by
614 Sebastien Martini, further info and confirmation from Stefan
615 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
618 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
620 *) Fix extension code to avoid race conditions which can result in a buffer
621 overrun vulnerability: resumed sessions must not be modified as they can
622 be shared by multiple threads. CVE-2010-3864
625 *) Fix WIN32 build system to correctly link an ENGINE directory into
629 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
631 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
633 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
635 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
637 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
638 context. The operation can be customised via the ctrl mechanism in
639 case ENGINEs want to include additional functionality.
642 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
645 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
646 output hashes compatible with older versions of OpenSSL.
647 [Willy Weisz <weisz@vcpc.univie.ac.at>]
649 *) Fix compression algorithm handling: if resuming a session use the
650 compression algorithm of the resumed session instead of determining
651 it from client hello again. Don't allow server to change algorithm.
654 *) Add load_crls() function to apps tidying load_certs() too. Add option
655 to verify utility to allow additional CRLs to be included.
658 *) Update OCSP request code to permit adding custom headers to the request:
659 some responders need this.
662 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
664 [Julia Lawall <julia@diku.dk>]
666 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
667 needlessly dereferenced structures, used obsolete functions and
668 didn't handle all updated verify codes correctly.
671 *) Disable MD2 in the default configuration.
674 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
675 indicate the initial BIO being pushed or popped. This makes it possible
676 to determine whether the BIO is the one explicitly called or as a result
677 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
678 it handles reference counts correctly and doesn't zero out the I/O bio
679 when it is not being explicitly popped. WARNING: applications which
680 included workarounds for the old buggy behaviour will need to be modified
681 or they could free up already freed BIOs.
684 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
685 renaming to all platforms (within the 0.9.8 branch, this was
686 done conditionally on Netware platforms to avoid a name clash).
687 [Guenter <lists@gknw.net>]
689 *) Add ECDHE and PSK support to DTLS.
690 [Michael Tuexen <tuexen@fh-muenster.de>]
692 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
696 *) Add "missing" function EVP_MD_flags() (without this the only way to
697 retrieve a digest flags is by accessing the structure directly. Update
698 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
699 or cipher is registered as in the "from" argument. Print out all
700 registered digests in the dgst usage message instead of manually
701 attempting to work them out.
704 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
705 this allows the use of compression and extensions. Change default cipher
706 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
707 by default unless an application cipher string requests it.
710 *) Alter match criteria in PKCS12_parse(). It used to try to use local
711 key ids to find matching certificates and keys but some PKCS#12 files
712 don't follow the (somewhat unwritten) rules and this strategy fails.
713 Now just gather all certificates together and the first private key
714 then look for the first certificate that matches the key.
717 *) Support use of registered digest and cipher names for dgst and cipher
718 commands instead of having to add each one as a special case. So now
725 openssl dgst -sha256 foo
727 and this works for ENGINE based algorithms too.
731 *) Update Gost ENGINE to support parameter files.
732 [Victor B. Wagner <vitus@cryptocom.ru>]
734 *) Support GeneralizedTime in ca utility.
735 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
737 *) Enhance the hash format used for certificate directory links. The new
738 form uses the canonical encoding (meaning equivalent names will work
739 even if they aren't identical) and uses SHA1 instead of MD5. This form
740 is incompatible with the older format and as a result c_rehash should
741 be used to rebuild symbolic links.
744 *) Make PKCS#8 the default write format for private keys, replacing the
745 traditional format. This form is standardised, more secure and doesn't
746 include an implicit MD5 dependency.
749 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
750 committed to OpenSSL should pass this lot as a minimum.
753 *) Add session ticket override functionality for use by EAP-FAST.
754 [Jouni Malinen <j@w1.fi>]
756 *) Modify HMAC functions to return a value. Since these can be implemented
757 in an ENGINE errors can occur.
760 *) Type-checked OBJ_bsearch_ex.
763 *) Type-checked OBJ_bsearch. Also some constification necessitated
764 by type-checking. Still to come: TXT_DB, bsearch(?),
765 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
769 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
770 seconds to a tm structure directly, instead of going through OS
771 specific date routines. This avoids any issues with OS routines such
772 as the year 2038 bug. New *_adj() functions for ASN1 time structures
773 and X509_time_adj_ex() to cover the extended range. The existing
774 X509_time_adj() is still usable and will no longer have any date issues.
777 *) Delta CRL support. New use deltas option which will attempt to locate
778 and search any appropriate delta CRLs available.
780 This work was sponsored by Google.
783 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
784 code and add additional score elements. Validate alternate CRL paths
785 as part of the CRL checking and indicate a new error "CRL path validation
786 error" in this case. Applications wanting additional details can use
787 the verify callback and check the new "parent" field. If this is not
788 NULL CRL path validation is taking place. Existing applications wont
789 see this because it requires extended CRL support which is off by
792 This work was sponsored by Google.
795 *) Support for freshest CRL extension.
797 This work was sponsored by Google.
800 *) Initial indirect CRL support. Currently only supported in the CRLs
801 passed directly and not via lookup. Process certificate issuer
802 CRL entry extension and lookup CRL entries by bother issuer name
803 and serial number. Check and process CRL issuer entry in IDP extension.
805 This work was sponsored by Google.
808 *) Add support for distinct certificate and CRL paths. The CRL issuer
809 certificate is validated separately in this case. Only enabled if
810 an extended CRL support flag is set: this flag will enable additional
811 CRL functionality in future.
813 This work was sponsored by Google.
816 *) Add support for policy mappings extension.
818 This work was sponsored by Google.
821 *) Fixes to pathlength constraint, self issued certificate handling,
822 policy processing to align with RFC3280 and PKITS tests.
824 This work was sponsored by Google.
827 *) Support for name constraints certificate extension. DN, email, DNS
828 and URI types are currently supported.
830 This work was sponsored by Google.
833 *) To cater for systems that provide a pointer-based thread ID rather
834 than numeric, deprecate the current numeric thread ID mechanism and
835 replace it with a structure and associated callback type. This
836 mechanism allows a numeric "hash" to be extracted from a thread ID in
837 either case, and on platforms where pointers are larger than 'long',
838 mixing is done to help ensure the numeric 'hash' is usable even if it
839 can't be guaranteed unique. The default mechanism is to use "&errno"
840 as a pointer-based thread ID to distinguish between threads.
842 Applications that want to provide their own thread IDs should now use
843 CRYPTO_THREADID_set_callback() to register a callback that will call
844 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
846 Note that ERR_remove_state() is now deprecated, because it is tied
847 to the assumption that thread IDs are numeric. ERR_remove_state(0)
848 to free the current thread's error state should be replaced by
849 ERR_remove_thread_state(NULL).
851 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
852 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
853 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
854 application was previously providing a numeric thread callback that
855 was inappropriate for distinguishing threads, then uniqueness might
856 have been obtained with &errno that happened immediately in the
857 intermediate development versions of OpenSSL; this is no longer the
858 case, the numeric thread callback will now override the automatic use
860 [Geoff Thorpe, with help from Bodo Moeller]
862 *) Initial support for different CRL issuing certificates. This covers a
863 simple case where the self issued certificates in the chain exist and
864 the real CRL issuer is higher in the existing chain.
866 This work was sponsored by Google.
869 *) Removed effectively defunct crypto/store from the build.
872 *) Revamp of STACK to provide stronger type-checking. Still to come:
873 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
874 ASN1_STRING, CONF_VALUE.
877 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
878 RAM on SSL connections. This option can save about 34k per idle SSL.
881 *) Revamp of LHASH to provide stronger type-checking. Still to come:
882 STACK, TXT_DB, bsearch, qsort.
885 *) Initial support for Cryptographic Message Syntax (aka CMS) based
886 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
887 support for data, signedData, compressedData, digestedData and
888 encryptedData, envelopedData types included. Scripts to check against
889 RFC4134 examples draft and interop and consistency checks of many
890 content types and variants.
893 *) Add options to enc utility to support use of zlib compression BIO.
896 *) Extend mk1mf to support importing of options and assembly language
897 files from Configure script, currently only included in VC-WIN32.
898 The assembly language rules can now optionally generate the source
899 files from the associated perl scripts.
902 *) Implement remaining functionality needed to support GOST ciphersuites.
903 Interop testing has been performed using CryptoPro implementations.
904 [Victor B. Wagner <vitus@cryptocom.ru>]
906 *) s390x assembler pack.
909 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
913 *) Implement Opaque PRF Input TLS extension as specified in
914 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
915 official specification yet and no extension type assignment by
916 IANA exists, this extension (for now) will have to be explicitly
917 enabled when building OpenSSL by providing the extension number
918 to use. For example, specify an option
920 -DTLSEXT_TYPE_opaque_prf_input=0x9527
922 to the "config" or "Configure" script to enable the extension,
923 assuming extension number 0x9527 (which is a completely arbitrary
924 and unofficial assignment based on the MD5 hash of the Internet
925 Draft). Note that by doing so, you potentially lose
926 interoperability with other TLS implementations since these might
927 be using the same extension number for other purposes.
929 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
930 opaque PRF input value to use in the handshake. This will create
931 an interal copy of the length-'len' string at 'src', and will
932 return non-zero for success.
934 To get more control and flexibility, provide a callback function
937 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
938 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
942 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
945 Callback function 'cb' will be called in handshakes, and is
946 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
947 Argument 'arg' is for application purposes (the value as given to
948 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
949 be provided to the callback function). The callback function
950 has to return non-zero to report success: usually 1 to use opaque
951 PRF input just if possible, or 2 to enforce use of the opaque PRF
952 input. In the latter case, the library will abort the handshake
953 if opaque PRF input is not successfully negotiated.
955 Arguments 'peerinput' and 'len' given to the callback function
956 will always be NULL and 0 in the case of a client. A server will
957 see the client's opaque PRF input through these variables if
958 available (NULL and 0 otherwise). Note that if the server
959 provides an opaque PRF input, the length must be the same as the
960 length of the client's opaque PRF input.
962 Note that the callback function will only be called when creating
963 a new session (session resumption can resume whatever was
964 previously negotiated), and will not be called in SSL 2.0
965 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
966 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
967 for applications that need to enforce opaque PRF input.
971 *) Update ssl code to support digests other than SHA1+MD5 for handshake
974 [Victor B. Wagner <vitus@cryptocom.ru>]
976 *) Add RFC4507 support to OpenSSL. This includes the corrections in
977 RFC4507bis. The encrypted ticket format is an encrypted encoded
978 SSL_SESSION structure, that way new session features are automatically
981 If a client application caches session in an SSL_SESSION structure
982 support is transparent because tickets are now stored in the encoded
985 The SSL_CTX structure automatically generates keys for ticket
986 protection in servers so again support should be possible
987 with no application modification.
989 If a client or server wishes to disable RFC4507 support then the option
990 SSL_OP_NO_TICKET can be set.
992 Add a TLS extension debugging callback to allow the contents of any client
993 or server extensions to be examined.
995 This work was sponsored by Google.
998 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
999 OpenSSL should now compile cleanly on gcc 4.2
1000 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1002 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1003 support including streaming MAC support: this is required for GOST
1004 ciphersuite support.
1005 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1007 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1008 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1009 to output in BER and PEM format.
1012 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1013 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1014 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1015 ENGINE support for HMAC keys which are unextractable. New -mac and
1016 -macopt options to dgst utility.
1019 *) New option -sigopt to dgst utility. Update dgst to use
1020 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1021 alternative signing paramaters such as X9.31 or PSS in the dgst
1025 *) Change ssl_cipher_apply_rule(), the internal function that does
1026 the work each time a ciphersuite string requests enabling
1027 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1028 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1029 the order of disabled ciphersuites such that those ciphersuites
1030 that most recently went from enabled to disabled not only stay
1031 in order with respect to each other, but also have higher priority
1032 than other disabled ciphersuites the next time ciphersuites are
1035 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1036 the same ciphersuites as with "HIGH" alone, but in a specific
1037 order where the PSK ciphersuites come first (since they are the
1038 most recently disabled ciphersuites when "HIGH" is parsed).
1040 Also, change ssl_create_cipher_list() (using this new
1041 funcionality) such that between otherwise identical
1042 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1046 *) Change ssl_create_cipher_list() so that it automatically
1047 arranges the ciphersuites in reasonable order before starting
1048 to process the rule string. Thus, the definition for "DEFAULT"
1049 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1050 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1051 This makes it much easier to arrive at a reasonable default order
1052 in applications for which anonymous ciphers are OK (meaning
1053 that you can't actually use DEFAULT).
1054 [Bodo Moeller; suggested by Victor Duchovni]
1056 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1057 processing) into multiple integers instead of setting
1058 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1059 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1060 (These masks as well as the individual bit definitions are hidden
1061 away into the non-exported interface ssl/ssl_locl.h, so this
1062 change to the definition of the SSL_CIPHER structure shouldn't
1063 affect applications.) This give us more bits for each of these
1064 categories, so there is no longer a need to coagulate AES128 and
1065 AES256 into a single algorithm bit, and to coagulate Camellia128
1066 and Camellia256 into a single algorithm bit, which has led to all
1069 Thus, among other things, the kludge introduced in 0.9.7m and
1070 0.9.8e for masking out AES256 independently of AES128 or masking
1071 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1073 With the change, we also introduce new ciphersuite aliases that
1074 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1078 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1079 Use the leftmost N bytes of the signature input if the input is
1080 larger than the prime q (with N being the size in bytes of q).
1083 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1084 it yet and it is largely untested.
1087 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1090 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1091 some compilers (gcc 4.2 and later) reject their use. Safestack is
1092 reimplemented. Update ASN1 to avoid use of legacy functions.
1095 *) Win32/64 targets are linked with Winsock2.
1098 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1099 to external functions. This can be used to increase CRL handling
1100 efficiency especially when CRLs are very large by (for example) storing
1101 the CRL revoked certificates in a database.
1104 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1105 new CRLs added to a directory can be used. New command line option
1106 -verify_return_error to s_client and s_server. This causes real errors
1107 to be returned by the verify callback instead of carrying on no matter
1108 what. This reflects the way a "real world" verify callback would behave.
1111 *) GOST engine, supporting several GOST algorithms and public key formats.
1112 Kindly donated by Cryptocom.
1115 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1116 partitioned by DP are handled but no indirect CRL or reason partitioning
1117 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1118 selected via a scoring technique which handles IDP and AKID in CRLs.
1121 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1122 will ultimately be used for all verify operations: this will remove the
1123 X509_STORE dependency on certificate verification and allow alternative
1124 lookup methods. X509_STORE based implementations of these two callbacks.
1127 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1128 Modify get_crl() to find a valid (unexpired) CRL if possible.
1131 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1132 this would be called X509_CRL_cmp() but that name is already used by
1133 a function that just compares CRL issuer names. Cache several CRL
1134 extensions in X509_CRL structure and cache CRLDP in X509.
1137 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1138 this maps equivalent X509_NAME structures into a consistent structure.
1139 Name comparison can then be performed rapidly using memcmp().
1142 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1146 *) Allow digests to supply their own micalg string for S/MIME type using
1147 the ctrl EVP_MD_CTRL_MICALG.
1150 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1151 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1152 ctrl. It can then customise the structure before and/or after signing
1156 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1157 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1158 to free up any added signature OIDs.
1161 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1162 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1163 digest and cipher tables. New options added to openssl utility:
1164 list-message-digest-algorithms and list-cipher-algorithms.
1167 *) Change the array representation of binary polynomials: the list
1168 of degrees of non-zero coefficients is now terminated with -1.
1169 Previously it was terminated with 0, which was also part of the
1170 value; thus, the array representation was not applicable to
1171 polynomials where t^0 has coefficient zero. This change makes
1172 the array representation useful in a more general context.
1175 *) Various modifications and fixes to SSL/TLS cipher string
1176 handling. For ECC, the code now distinguishes between fixed ECDH
1177 with RSA certificates on the one hand and with ECDSA certificates
1178 on the other hand, since these are separate ciphersuites. The
1179 unused code for Fortezza ciphersuites has been removed.
1181 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1182 (not "ECDHE"). For consistency with the code for DH
1183 certificates, use of ECDH certificates is now considered ECDH
1184 authentication, not RSA or ECDSA authentication (the latter is
1185 merely the CA's signing algorithm and not actively used in the
1188 The temporary ciphersuite alias "ECCdraft" is no longer
1189 available, and ECC ciphersuites are no longer excluded from "ALL"
1190 and "DEFAULT". The following aliases now exist for RFC 4492
1191 ciphersuites, most of these by analogy with the DH case:
1193 kECDHr - ECDH cert, signed with RSA
1194 kECDHe - ECDH cert, signed with ECDSA
1195 kECDH - ECDH cert (signed with either RSA or ECDSA)
1196 kEECDH - ephemeral ECDH
1197 ECDH - ECDH cert or ephemeral ECDH
1203 AECDH - anonymous ECDH
1204 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1208 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1209 Use correct micalg parameters depending on digest(s) in signed message.
1212 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1213 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1216 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1217 an engine to register a method. Add ENGINE lookups for methods and
1218 functional reference processing.
1221 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1222 EVP_{Sign,Verify}* which allow an application to customise the signature
1226 *) New -resign option to smime utility. This adds one or more signers
1227 to an existing PKCS#7 signedData structure. Also -md option to use an
1228 alternative message digest algorithm for signing.
1231 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1232 create PKCS7 structures containing multiple signers. Update smime
1233 application to support multiple signers.
1236 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1240 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1241 Reorganize PBE internals to lookup from a static table using NIDs,
1242 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1243 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1244 PRF which will be automatically used with PBES2.
1247 *) Replace the algorithm specific calls to generate keys in "req" with the
1251 *) Update PKCS#7 enveloped data routines to use new API. This is now
1252 supported by any public key method supporting the encrypt operation. A
1253 ctrl is added to allow the public key algorithm to examine or modify
1254 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1258 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1259 a default digest type to use. In most cases this will be SHA1 but some
1260 algorithms (such as GOST) need to specify an alternative digest. The
1261 return value indicates how strong the prefernce is 1 means optional and
1262 2 is mandatory (that is it is the only supported type). Modify
1263 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1264 use the default md. Update openssl utilities to use the default digest
1265 type for signing if it is not explicitly indicated.
1268 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1269 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1270 signing method from the key type. This effectively removes the link
1271 between digests and public key types.
1274 *) Add an OID cross reference table and utility functions. Its purpose is to
1275 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1276 rsaEncryption. This will allow some of the algorithm specific hackery
1277 needed to use the correct OID to be removed.
1280 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1281 structures for PKCS7_sign(). They are now set up by the relevant public
1285 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1288 *) Add support for key derivation (agreement) in the API, DH method and
1292 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1293 public and private key formats. As a side effect these add additional
1294 command line functionality not previously available: DSA signatures can be
1295 generated and verified using pkeyutl and DH key support and generation in
1300 [Oliver Tappe <zooey@hirschkaefer.de>]
1302 *) New make target "install_html_docs" installs HTML renditions of the
1304 [Oliver Tappe <zooey@hirschkaefer.de>]
1306 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1307 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1308 support key and parameter generation and add initial key generation
1309 functionality for RSA.
1312 *) Add functions for main EVP_PKEY_method operations. The undocumented
1313 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1314 EVP_PKEY_{encrypt,decrypt}_old.
1317 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1318 key API, doesn't do much yet.
1321 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1322 public key algorithms. New option to openssl utility:
1323 "list-public-key-algorithms" to print out info.
1326 *) Implement the Supported Elliptic Curves Extension for
1327 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1330 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1331 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1334 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1335 utilities such as rsa, dsa, dsaparam etc except they process any key
1339 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1340 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1341 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1345 *) Initial support for pluggable public key ASN1.
1346 De-spaghettify the public key ASN1 handling. Move public and private
1347 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1348 algorithm specific handling to a single module within the relevant
1349 algorithm directory. Add functions to allow (near) opaque processing
1350 of public and private key structures.
1353 *) Implement the Supported Point Formats Extension for
1354 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1357 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1358 for the psk identity [hint] and the psk callback functions to the
1359 SSL_SESSION, SSL and SSL_CTX structure.
1362 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1366 SSL_CTX_use_psk_identity_hint
1367 SSL_get_psk_identity_hint
1368 SSL_get_psk_identity
1369 SSL_use_psk_identity_hint
1371 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1373 *) Add RFC 3161 compliant time stamp request creation, response generation
1374 and response verification functionality.
1375 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1377 *) Add initial support for TLS extensions, specifically for the server_name
1378 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1379 have new members for a host name. The SSL data structure has an
1380 additional member SSL_CTX *initial_ctx so that new sessions can be
1381 stored in that context to allow for session resumption, even after the
1382 SSL has been switched to a new SSL_CTX in reaction to a client's
1383 server_name extension.
1385 New functions (subject to change):
1387 SSL_get_servername()
1388 SSL_get_servername_type()
1391 New CTRL codes and macros (subject to change):
1393 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1394 - SSL_CTX_set_tlsext_servername_callback()
1395 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1396 - SSL_CTX_set_tlsext_servername_arg()
1397 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1399 openssl s_client has a new '-servername ...' option.
1401 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1402 '-key2 ...', '-servername_fatal' (subject to change). This allows
1403 testing the HostName extension for a specific single host name ('-cert'
1404 and '-key' remain fallbacks for handshakes without HostName
1405 negotiation). If the unrecogninzed_name alert has to be sent, this by
1406 default is a warning; it becomes fatal with the '-servername_fatal'
1409 [Peter Sylvester, Remy Allais, Christophe Renou]
1411 *) Whirlpool hash implementation is added.
1414 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1415 bn(64,32). Because of instruction set limitations it doesn't have
1416 any negative impact on performance. This was done mostly in order
1417 to make it possible to share assembler modules, such as bn_mul_mont
1418 implementations, between 32- and 64-bit builds without hassle.
1421 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1422 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1426 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1427 dedicated Montgomery multiplication procedure, is introduced.
1428 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1429 "64-bit" performance on certain 32-bit targets.
1432 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1433 in SSL structures. New SSL ctrl to set maximum send fragment size.
1434 Save memory by seeting the I/O buffer sizes dynamically instead of
1435 using the maximum available value.
1438 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1439 in addition to the text details.
1442 *) Very, very preliminary EXPERIMENTAL support for printing of general
1443 ASN1 structures. This currently produces rather ugly output and doesn't
1444 handle several customised structures at all.
1447 *) Integrated support for PVK file format and some related formats such
1448 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1449 these in the 'rsa' and 'dsa' utilities.
1452 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1455 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1456 place for the (very old) "NETSCAPE" format certificates which are now
1457 handled using new ASN1 code equivalents.
1460 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1461 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1462 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1465 *) Modify CRL distribution points extension code to print out previously
1466 unsupported fields. Enhance extension setting code to allow setting of
1470 *) Add print and set support for Issuing Distribution Point CRL extension.
1473 *) Change 'Configure' script to enable Camellia by default.
1476 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1478 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1480 This addresses the flaw in CBC record processing discovered by
1481 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1482 at: http://www.isg.rhul.ac.uk/tls/
1484 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1485 Security Group at Royal Holloway, University of London
1486 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1487 Emilia Käsper for the initial patch.
1489 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1491 *) Return an error when checking OCSP signatures when key is NULL.
1492 This fixes a DoS attack. (CVE-2013-0166)
1495 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1496 the right response is stapled. Also change SSL_get_certificate()
1497 so it returns the certificate actually sent.
1498 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1499 (This is a backport)
1500 [Rob Stradling <rob.stradling@comodo.com>]
1502 *) Fix possible deadlock when decoding public keys.
1505 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1507 *) Sanity check record length before skipping explicit IV in DTLS
1510 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1511 fuzzing as a service testing platform.
1515 *) Initialise tkeylen properly when encrypting CMS messages.
1516 Thanks to Solar Designer of Openwall for reporting this issue.
1519 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1521 *) The fix for CVE-2012-2110 did not take into account that the
1522 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1523 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1524 rejecting negative len parameter. (CVE-2012-2131)
1525 [Tomas Hoger <thoger@redhat.com>]
1527 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1529 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1530 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1531 in CRYPTO_realloc_clean.
1533 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1534 issue and to Adam Langley <agl@chromium.org> for fixing it.
1536 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1538 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1540 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1541 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1542 content decryption and always return the same error. Note: this attack
1543 needs on average 2^20 messages so it only affects automated senders. The
1544 old behaviour can be reenabled in the CMS code by setting the
1545 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1546 an MMA defence is not necessary.
1547 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1548 this issue. (CVE-2012-0884)
1551 *) Fix CVE-2011-4619: make sure we really are receiving a
1552 client hello before rejecting multiple SGC restarts. Thanks to
1553 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1556 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1558 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1559 Thanks to Antonio Martin, Enterprise Secure Access Research and
1560 Development, Cisco Systems, Inc. for discovering this bug and
1561 preparing a fix. (CVE-2012-0050)
1564 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1566 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1567 of the Vaudenay padding oracle attack on CBC mode encryption
1568 which enables an efficient plaintext recovery attack against
1569 the OpenSSL implementation of DTLS. Their attack exploits timing
1570 differences arising during decryption processing. A research
1571 paper describing this attack can be found at:
1572 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1573 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1574 Security Group at Royal Holloway, University of London
1575 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1576 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1577 for preparing the fix. (CVE-2011-4108)
1578 [Robin Seggelmann, Michael Tuexen]
1580 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1581 [Ben Laurie, Kasper <ekasper@google.com>]
1583 *) Clear bytes used for block padding of SSL 3.0 records.
1585 [Adam Langley (Google)]
1587 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1588 Kadianakis <desnacked@gmail.com> for discovering this issue and
1589 Adam Langley for preparing the fix. (CVE-2011-4619)
1590 [Adam Langley (Google)]
1592 *) Prevent malformed RFC3779 data triggering an assertion failure.
1593 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1594 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1595 [Rob Austein <sra@hactrn.net>]
1597 *) Fix ssl_ciph.c set-up race.
1598 [Adam Langley (Google)]
1600 *) Fix spurious failures in ecdsatest.c.
1601 [Emilia Käsper (Google)]
1603 *) Fix the BIO_f_buffer() implementation (which was mixing different
1604 interpretations of the '..._len' fields).
1605 [Adam Langley (Google)]
1607 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1608 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1609 threads won't reuse the same blinding coefficients.
1611 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1612 lock to call BN_BLINDING_invert_ex, and avoids one use of
1613 BN_BLINDING_update for each BN_BLINDING structure (previously,
1614 the last update always remained unused).
1615 [Emilia Käsper (Google)]
1617 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1618 for multi-threaded use of ECDH.
1619 [Adam Langley (Google)]
1621 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1624 *) Add protection against ECDSA timing attacks as mentioned in the paper
1625 by Billy Bob Brumley and Nicola Tuveri, see:
1627 http://eprint.iacr.org/2011/232.pdf
1629 [Billy Bob Brumley and Nicola Tuveri]
1631 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1633 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1634 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1636 *) Fix bug in string printing code: if *any* escaping is enabled we must
1637 escape the escape character (backslash) or the resulting string is
1641 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1643 *) Disable code workaround for ancient and obsolete Netscape browsers
1644 and servers: an attacker can use it in a ciphersuite downgrade attack.
1645 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1648 *) Fixed J-PAKE implementation error, originally discovered by
1649 Sebastien Martini, further info and confirmation from Stefan
1650 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1653 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1655 *) Fix extension code to avoid race conditions which can result in a buffer
1656 overrun vulnerability: resumed sessions must not be modified as they can
1657 be shared by multiple threads. CVE-2010-3864
1660 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1663 *) Don't reencode certificate when calculating signature: cache and use
1664 the original encoding instead. This makes signature verification of
1665 some broken encodings work correctly.
1668 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1669 is also one of the inputs.
1670 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1672 *) Don't repeatedly append PBE algorithms to table if they already exist.
1673 Sort table on each new add. This effectively makes the table read only
1674 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1678 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1680 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1683 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1684 access or freeing data twice (CVE-2010-0742)
1685 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1687 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1688 common in certificates and some applications which only call
1689 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1693 Reduce copying into .apps and .test in makevms.com
1694 Don't try to use blank CA certificate in CA.com
1695 Allow use of C files from original directories in maketests.com
1696 [Steven M. Schweda" <sms@antinode.info>]
1698 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1700 *) When rejecting SSL/TLS records due to an incorrect version number, never
1701 update s->server with a new major version number. As of
1702 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1703 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1704 the previous behavior could result in a read attempt at NULL when
1705 receiving specific incorrect SSL/TLS records once record payload
1706 protection is active. (CVE-2010-0740)
1707 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1709 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1710 could be crashed if the relevant tables were not present (e.g. chrooted).
1711 [Tomas Hoger <thoger@redhat.com>]
1713 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1715 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1716 [Martin Olsson, Neel Mehta]
1718 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1719 accommodate for stack sorting, always a write lock!).
1722 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1723 excessive delays in the RAND_poll(): over a minute. As a workaround
1724 include a time check in the inner Heap32Next loop too.
1727 *) The code that handled flushing of data in SSL/TLS originally used the
1728 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1729 the problem outlined in PR#1949. The fix suggested there however can
1730 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1731 of Apache). So instead simplify the code to flush unconditionally.
1732 This should be fine since flushing with no data to flush is a no op.
1735 *) Handle TLS versions 2.0 and later properly and correctly use the
1736 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1737 off ancient servers have a habit of sticking around for a while...
1740 *) Modify compression code so it frees up structures without using the
1741 ex_data callbacks. This works around a problem where some applications
1742 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1743 restarting) then use compression (e.g. SSL with compression) later.
1744 This results in significant per-connection memory leaks and
1745 has caused some security issues including CVE-2008-1678 and
1749 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1750 change when encrypting or decrypting.
1753 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1754 connect and renegotiate with servers which do not support RI.
1755 Until RI is more widely deployed this option is enabled by default.
1758 *) Add "missing" ssl ctrls to clear options and mode.
1761 *) If client attempts to renegotiate and doesn't support RI respond with
1762 a no_renegotiation alert as required by RFC5746. Some renegotiating
1763 TLS clients will continue a connection gracefully when they receive
1764 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1765 waiting for a server hello which it will never receive. Now we treat a
1766 received no_renegotiation alert as a fatal error. This is because
1767 applications requesting a renegotiation might well expect it to succeed
1768 and would have no code in place to handle the server denying it so the
1769 only safe thing to do is to terminate the connection.
1772 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1773 peer supports secure renegotiation and 0 otherwise. Print out peer
1774 renegotiation support in s_client/s_server.
1777 *) Replace the highly broken and deprecated SPKAC certification method with
1778 the updated NID creation version. This should correctly handle UTF8.
1781 *) Implement RFC5746. Re-enable renegotiation but require the extension
1782 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1783 turns out to be a bad idea. It has been replaced by
1784 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1785 SSL_CTX_set_options(). This is really not recommended unless you
1786 know what you are doing.
1787 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1789 *) Fixes to stateless session resumption handling. Use initial_ctx when
1790 issuing and attempting to decrypt tickets in case it has changed during
1791 servername handling. Use a non-zero length session ID when attempting
1792 stateless session resumption: this makes it possible to determine if
1793 a resumption has occurred immediately after receiving server hello
1794 (several places in OpenSSL subtly assume this) instead of later in
1798 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1799 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1800 fixes for a few places where the return code is not checked
1802 [Julia Lawall <julia@diku.dk>]
1804 *) Add --strict-warnings option to Configure script to include devteam
1805 warnings in other configurations.
1808 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1809 makes it possible to install openssl libraries in locations which
1810 have names other than "lib", for example "/usr/lib64" which some
1812 [Steve Henson, based on patch from Jeremy Utley]
1814 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1815 X690 8.9.12 and can produce some misleading textual output of OIDs.
1816 [Steve Henson, reported by Dan Kaminsky]
1818 *) Delete MD2 from algorithm tables. This follows the recommendation in
1819 several standards that it is not used in new applications due to
1820 several cryptographic weaknesses. For binary compatibility reasons
1821 the MD2 API is still compiled in by default.
1824 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1828 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1829 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1831 [Guenter <lists@gknw.net>]
1833 *) Fix the server certificate chain building code to use X509_verify_cert(),
1834 it used to have an ad-hoc builder which was unable to cope with anything
1835 other than a simple chain.
1836 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1838 *) Don't check self signed certificate signatures in X509_verify_cert()
1839 by default (a flag can override this): it just wastes time without
1840 adding any security. As a useful side effect self signed root CAs
1841 with non-FIPS digests are now usable in FIPS mode.
1844 *) In dtls1_process_out_of_seq_message() the check if the current message
1845 is already buffered was missing. For every new message was memory
1846 allocated, allowing an attacker to perform an denial of service attack
1847 with sending out of seq handshake messages until there is no memory
1848 left. Additionally every future messege was buffered, even if the
1849 sequence number made no sense and would be part of another handshake.
1850 So only messages with sequence numbers less than 10 in advance will be
1851 buffered. (CVE-2009-1378)
1852 [Robin Seggelmann, discovered by Daniel Mentz]
1854 *) Records are buffered if they arrive with a future epoch to be
1855 processed after finishing the corresponding handshake. There is
1856 currently no limitation to this buffer allowing an attacker to perform
1857 a DOS attack with sending records with future epochs until there is no
1858 memory left. This patch adds the pqueue_size() function to detemine
1859 the size of a buffer and limits the record buffer to 100 entries.
1861 [Robin Seggelmann, discovered by Daniel Mentz]
1863 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1864 parent structure is freed. (CVE-2009-1379)
1867 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1868 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1871 [Ilya O. <vrghost@gmail.com>]
1873 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1875 *) Disable renegotiation completely - this fixes a severe security
1876 problem (CVE-2009-3555) at the cost of breaking all
1877 renegotiation. Renegotiation can be re-enabled by setting
1878 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1879 run-time. This is really not recommended unless you know what
1883 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1885 *) Don't set val to NULL when freeing up structures, it is freed up by
1886 underlying code. If sizeof(void *) > sizeof(long) this can result in
1887 zeroing past the valid field. (CVE-2009-0789)
1888 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1890 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1891 checked correctly. This would allow some invalid signed attributes to
1892 appear to verify correctly. (CVE-2009-0591)
1893 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1895 *) Reject UniversalString and BMPString types with invalid lengths. This
1896 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1897 a legal length. (CVE-2009-0590)
1900 *) Set S/MIME signing as the default purpose rather than setting it
1901 unconditionally. This allows applications to override it at the store
1905 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1906 to handle some structures.
1909 *) Improve efficiency of mem_gets: don't search whole buffer each time
1911 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1913 *) New -hex option for openssl rand.
1916 *) Print out UTF8String and NumericString when parsing ASN1.
1919 *) Support NumericString type for name components.
1922 *) Allow CC in the environment to override the automatically chosen
1923 compiler. Note that nothing is done to ensure flags work with the
1927 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1929 *) Properly check EVP_VerifyFinal() and similar return values
1931 [Ben Laurie, Bodo Moeller, Google Security Team]
1933 *) Enable TLS extensions by default.
1936 *) Allow the CHIL engine to be loaded, whether the application is
1937 multithreaded or not. (This does not release the developer from the
1938 obligation to set up the dynamic locking callbacks.)
1939 [Sander Temme <sander@temme.net>]
1941 *) Use correct exit code if there is an error in dgst command.
1942 [Steve Henson; problem pointed out by Roland Dirlewanger]
1944 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1945 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1948 *) Add experimental JPAKE support, including demo authentication in
1949 s_client and s_server.
1952 *) Set the comparison function in v3_addr_canonize().
1953 [Rob Austein <sra@hactrn.net>]
1955 *) Add support for XMPP STARTTLS in s_client.
1956 [Philip Paeps <philip@freebsd.org>]
1958 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1959 to ensure that even with this option, only ciphersuites in the
1960 server's preference list will be accepted. (Note that the option
1961 applies only when resuming a session, so the earlier behavior was
1962 just about the algorithm choice for symmetric cryptography.)
1965 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1967 *) Fix NULL pointer dereference if a DTLS server received
1968 ChangeCipherSpec as first record (CVE-2009-1386).
1971 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1972 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1975 *) The fix in 0.9.8c that supposedly got rid of unsafe
1976 double-checked locking was incomplete for RSA blinding,
1977 addressing just one layer of what turns out to have been
1978 doubly unsafe triple-checked locking.
1980 So now fix this for real by retiring the MONT_HELPER macro
1981 in crypto/rsa/rsa_eay.c.
1983 [Bodo Moeller; problem pointed out by Marius Schilder]
1985 *) Various precautionary measures:
1987 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1989 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1990 (NB: This would require knowledge of the secret session ticket key
1991 to exploit, in which case you'd be SOL either way.)
1993 - Change bn_nist.c so that it will properly handle input BIGNUMs
1994 outside the expected range.
1996 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1999 [Neel Mehta, Bodo Moeller]
2001 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2002 the load fails. Useful for distros.
2003 [Ben Laurie and the FreeBSD team]
2005 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2008 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2011 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2013 This work was sponsored by Logica.
2016 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2017 keystores. Support for SSL/TLS client authentication too.
2018 Not compiled unless enable-capieng specified to Configure.
2020 This work was sponsored by Logica.
2023 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2024 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2025 attribute creation routines such as certifcate requests and PKCS#12
2029 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2031 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2032 handshake which could lead to a cilent crash as found using the
2033 Codenomicon TLS test suite (CVE-2008-1672)
2034 [Steve Henson, Mark Cox]
2036 *) Fix double free in TLS server name extensions which could lead to
2037 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2040 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2042 Clear the error queue to ensure that error entries left from
2043 older function calls do not interfere with the correct operation.
2044 [Lutz Jaenicke, Erik de Castro Lopo]
2046 *) Remove root CA certificates of commercial CAs:
2048 The OpenSSL project does not recommend any specific CA and does not
2049 have any policy with respect to including or excluding any CA.
2050 Therefore it does not make any sense to ship an arbitrary selection
2051 of root CA certificates with the OpenSSL software.
2054 *) RSA OAEP patches to fix two separate invalid memory reads.
2055 The first one involves inputs when 'lzero' is greater than
2056 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2057 before the beginning of from). The second one involves inputs where
2058 the 'db' section contains nothing but zeroes (there is a one-byte
2059 invalid read after the end of 'db').
2060 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2062 *) Partial backport from 0.9.9-dev:
2064 Introduce bn_mul_mont (dedicated Montgomery multiplication
2065 procedure) as a candidate for BIGNUM assembler implementation.
2066 While 0.9.9-dev uses assembler for various architectures, only
2067 x86_64 is available by default here in the 0.9.8 branch, and
2068 32-bit x86 is available through a compile-time setting.
2070 To try the 32-bit x86 assembler implementation, use Configure
2071 option "enable-montasm" (which exists only for this backport).
2073 As "enable-montasm" for 32-bit x86 disclaims code stability
2074 anyway, in this constellation we activate additional code
2075 backported from 0.9.9-dev for further performance improvements,
2076 namely BN_from_montgomery_word. (To enable this otherwise,
2077 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2079 [Andy Polyakov (backport partially by Bodo Moeller)]
2081 *) Add TLS session ticket callback. This allows an application to set
2082 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2083 values. This is useful for key rollover for example where several key
2084 sets may exist with different names.
2087 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2088 This was broken until now in 0.9.8 releases, such that the only way
2089 a registered ENGINE could be used (assuming it initialises
2090 successfully on the host) was to explicitly set it as the default
2091 for the relevant algorithms. This is in contradiction with 0.9.7
2092 behaviour and the documentation. With this fix, when an ENGINE is
2093 registered into a given algorithm's table of implementations, the
2094 'uptodate' flag is reset so that auto-discovery will be used next
2095 time a new context for that algorithm attempts to select an
2097 [Ian Lister (tweaked by Geoff Thorpe)]
2099 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2100 implemention in the following ways:
2102 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2105 Lack of BER streaming support means one pass streaming processing is
2106 only supported if data is detached: setting the streaming flag is
2107 ignored for embedded content.
2109 CMS support is disabled by default and must be explicitly enabled
2110 with the enable-cms configuration option.
2113 *) Update the GMP engine glue to do direct copies between BIGNUM and
2114 mpz_t when openssl and GMP use the same limb size. Otherwise the
2115 existing "conversion via a text string export" trick is still used.
2116 [Paul Sheer <paulsheer@gmail.com>]
2118 *) Zlib compression BIO. This is a filter BIO which compressed and
2119 uncompresses any data passed through it.
2122 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2123 RFC3394 compatible AES key wrapping.
2126 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2127 sets string data without copying. X509_ALGOR_set0() and
2128 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2129 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2130 from an X509_ATTRIBUTE structure optionally checking it occurs only
2131 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2135 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2136 to get the expected BN_FLG_CONSTTIME behavior.
2137 [Bodo Moeller (Google)]
2141 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2142 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2143 - added some more tests to do_tests.pl
2144 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2145 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2146 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2147 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2148 - various changes to netware.pl to enable gcc-cross builds on Win32
2150 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2151 - various changes to fix missing prototype warnings
2152 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2153 - added AES, WHIRLPOOL and CPUID assembler code to build files
2154 - added missing AES assembler make rules to mk1mf.pl
2155 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2156 [Guenter Knauf <eflash@gmx.net>]
2158 *) Implement certificate status request TLS extension defined in RFC3546.
2159 A client can set the appropriate parameters and receive the encoded
2160 OCSP response via a callback. A server can query the supplied parameters
2161 and set the encoded OCSP response in the callback. Add simplified examples
2162 to s_client and s_server.
2165 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2167 *) Fix various bugs:
2168 + Binary incompatibility of ssl_ctx_st structure
2169 + DTLS interoperation with non-compliant servers
2170 + Don't call get_session_cb() without proposed session
2171 + Fix ia64 assembler code
2172 [Andy Polyakov, Steve Henson]
2174 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2176 *) DTLS Handshake overhaul. There were longstanding issues with
2177 OpenSSL DTLS implementation, which were making it impossible for
2178 RFC 4347 compliant client to communicate with OpenSSL server.
2179 Unfortunately just fixing these incompatibilities would "cut off"
2180 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2181 server keeps tolerating non RFC compliant syntax. The opposite is
2182 not true, 0.9.8f client can not communicate with earlier server.
2183 This update even addresses CVE-2007-4995.
2186 *) Changes to avoid need for function casts in OpenSSL: some compilers
2187 (gcc 4.2 and later) reject their use.
2188 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2191 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2192 RFC4507bis. The encrypted ticket format is an encrypted encoded
2193 SSL_SESSION structure, that way new session features are automatically
2196 If a client application caches session in an SSL_SESSION structure
2197 support is transparent because tickets are now stored in the encoded
2200 The SSL_CTX structure automatically generates keys for ticket
2201 protection in servers so again support should be possible
2202 with no application modification.
2204 If a client or server wishes to disable RFC4507 support then the option
2205 SSL_OP_NO_TICKET can be set.
2207 Add a TLS extension debugging callback to allow the contents of any client
2208 or server extensions to be examined.
2210 This work was sponsored by Google.
2213 *) Add initial support for TLS extensions, specifically for the server_name
2214 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2215 have new members for a host name. The SSL data structure has an
2216 additional member SSL_CTX *initial_ctx so that new sessions can be
2217 stored in that context to allow for session resumption, even after the
2218 SSL has been switched to a new SSL_CTX in reaction to a client's
2219 server_name extension.
2221 New functions (subject to change):
2223 SSL_get_servername()
2224 SSL_get_servername_type()
2227 New CTRL codes and macros (subject to change):
2229 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2230 - SSL_CTX_set_tlsext_servername_callback()
2231 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2232 - SSL_CTX_set_tlsext_servername_arg()
2233 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2235 openssl s_client has a new '-servername ...' option.
2237 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2238 '-key2 ...', '-servername_fatal' (subject to change). This allows
2239 testing the HostName extension for a specific single host name ('-cert'
2240 and '-key' remain fallbacks for handshakes without HostName
2241 negotiation). If the unrecogninzed_name alert has to be sent, this by
2242 default is a warning; it becomes fatal with the '-servername_fatal'
2245 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2247 *) Add AES and SSE2 assembly language support to VC++ build.
2250 *) Mitigate attack on final subtraction in Montgomery reduction.
2253 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2254 (which previously caused an internal error).
2257 *) Squeeze another 10% out of IGE mode when in != out.
2260 *) AES IGE mode speedup.
2261 [Dean Gaudet (Google)]
2263 *) Add the Korean symmetric 128-bit cipher SEED (see
2264 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2265 add SEED ciphersuites from RFC 4162:
2267 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2268 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2269 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2270 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2272 To minimize changes between patchlevels in the OpenSSL 0.9.8
2273 series, SEED remains excluded from compilation unless OpenSSL
2274 is configured with 'enable-seed'.
2275 [KISA, Bodo Moeller]
2277 *) Mitigate branch prediction attacks, which can be practical if a
2278 single processor is shared, allowing a spy process to extract
2279 information. For detailed background information, see
2280 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2281 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2282 and Necessary Software Countermeasures"). The core of the change
2283 are new versions BN_div_no_branch() and
2284 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2285 respectively, which are slower, but avoid the security-relevant
2286 conditional branches. These are automatically called by BN_div()
2287 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2288 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2289 remove a conditional branch.
2291 BN_FLG_CONSTTIME is the new name for the previous
2292 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2293 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2294 in the exponent causes BN_mod_exp_mont() to use the alternative
2295 implementation in BN_mod_exp_mont_consttime().) The old name
2296 remains as a deprecated alias.
2298 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2299 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2300 constant-time implementations for more than just exponentiation.
2301 Here too the old name is kept as a deprecated alias.
2303 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2304 the BN_BLINDING structure gets an independent copy of the
2305 modulus. This means that the previous "BIGNUM *m" argument to
2306 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2307 essentially becomes "const BIGNUM *m", although we can't actually
2308 change this in the header file before 0.9.9. It allows
2309 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2310 enable BN_FLG_CONSTTIME.
2312 [Matthew D Wood (Intel Corp)]
2314 *) In the SSL/TLS server implementation, be strict about session ID
2315 context matching (which matters if an application uses a single
2316 external cache for different purposes). Previously,
2317 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2318 set. This did ensure strict client verification, but meant that,
2319 with applications using a single external cache for quite
2320 different requirements, clients could circumvent ciphersuite
2321 restrictions for a given session ID context by starting a session
2322 in a different context.
2325 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2326 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2327 authentication-only ciphersuites.
2330 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2331 not complete and could lead to a possible single byte overflow
2332 (CVE-2007-5135) [Ben Laurie]
2334 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2336 *) Since AES128 and AES256 (and similarly Camellia128 and
2337 Camellia256) share a single mask bit in the logic of
2338 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2339 kludge to work properly if AES128 is available and AES256 isn't
2340 (or if Camellia128 is available and Camellia256 isn't).
2343 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2344 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2345 When a point or a seed is encoded in a BIT STRING, we need to
2346 prevent the removal of trailing zero bits to get the proper DER
2347 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2348 of a NamedBitList, for which trailing 0 bits need to be removed.)
2351 *) Have SSL/TLS server implementation tolerate "mismatched" record
2352 protocol version while receiving ClientHello even if the
2353 ClientHello is fragmented. (The server can't insist on the
2354 particular protocol version it has chosen before the ServerHello
2355 message has informed the client about his choice.)
2358 *) Add RFC 3779 support.
2359 [Rob Austein for ARIN, Ben Laurie]
2361 *) Load error codes if they are not already present instead of using a
2362 static variable. This allows them to be cleanly unloaded and reloaded.
2363 Improve header file function name parsing.
2366 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2367 or CAPABILITY handshake as required by RFCs.
2370 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2372 *) Introduce limits to prevent malicious keys being able to
2373 cause a denial of service. (CVE-2006-2940)
2374 [Steve Henson, Bodo Moeller]
2376 *) Fix ASN.1 parsing of certain invalid structures that can result
2377 in a denial of service. (CVE-2006-2937) [Steve Henson]
2379 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2380 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2382 *) Fix SSL client code which could crash if connecting to a
2383 malicious SSLv2 server. (CVE-2006-4343)
2384 [Tavis Ormandy and Will Drewry, Google Security Team]
2386 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2387 match only those. Before that, "AES256-SHA" would be interpreted
2388 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2389 the same strength classification in 0.9.7h) as we currently only
2390 have a single AES bit in the ciphersuite description bitmap.
2391 That change, however, also applied to ciphersuite strings such as
2392 "RC4-MD5" that intentionally matched multiple ciphersuites --
2393 namely, SSL 2.0 ciphersuites in addition to the more common ones
2394 from SSL 3.0/TLS 1.0.
2396 So we change the selection algorithm again: Naming an explicit
2397 ciphersuite selects this one ciphersuite, and any other similar
2398 ciphersuite (same bitmap) from *other* protocol versions.
2399 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2400 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2402 Since SSL 2.0 does not have any ciphersuites for which the
2403 128/256 bit distinction would be relevant, this works for now.
2404 The proper fix will be to use different bits for AES128 and
2405 AES256, which would have avoided the problems from the beginning;
2406 however, bits are scarce, so we can only do this in a new release
2407 (not just a patchlevel) when we can change the SSL_CIPHER
2408 definition to split the single 'unsigned long mask' bitmap into
2409 multiple values to extend the available space.
2413 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2415 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2416 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2418 *) Add AES IGE and biIGE modes.
2421 *) Change the Unix randomness entropy gathering to use poll() when
2422 possible instead of select(), since the latter has some
2423 undesirable limitations.
2424 [Darryl Miles via Richard Levitte and Bodo Moeller]
2426 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2427 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2428 cannot be implicitly activated as part of, e.g., the "AES" alias.
2429 However, please upgrade to OpenSSL 0.9.9[-dev] for
2430 non-experimental use of the ECC ciphersuites to get TLS extension
2431 support, which is required for curve and point format negotiation
2432 to avoid potential handshake problems.
2435 *) Disable rogue ciphersuites:
2437 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2438 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2439 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2441 The latter two were purportedly from
2442 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2445 Also deactivate the remaining ciphersuites from
2446 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2447 unofficial, and the ID has long expired.
2450 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2451 dual-core machines) and other potential thread-safety issues.
2454 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2455 versions), which is now available for royalty-free use
2456 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2457 Also, add Camellia TLS ciphersuites from RFC 4132.
2459 To minimize changes between patchlevels in the OpenSSL 0.9.8
2460 series, Camellia remains excluded from compilation unless OpenSSL
2461 is configured with 'enable-camellia'.
2464 *) Disable the padding bug check when compression is in use. The padding
2465 bug check assumes the first packet is of even length, this is not
2466 necessarily true if compresssion is enabled and can result in false
2467 positives causing handshake failure. The actual bug test is ancient
2468 code so it is hoped that implementations will either have fixed it by
2469 now or any which still have the bug do not support compression.
2472 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2474 *) When applying a cipher rule check to see if string match is an explicit
2475 cipher suite and only match that one cipher suite if it is.
2478 *) Link in manifests for VC++ if needed.
2479 [Austin Ziegler <halostatue@gmail.com>]
2481 *) Update support for ECC-based TLS ciphersuites according to
2482 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2483 TLS extensions, which are supported starting with the 0.9.9
2484 branch, not in the OpenSSL 0.9.8 branch).
2487 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2488 opaque EVP_CIPHER_CTX handling.
2491 *) Fixes and enhancements to zlib compression code. We now only use
2492 "zlib1.dll" and use the default __cdecl calling convention on Win32
2493 to conform with the standards mentioned here:
2494 http://www.zlib.net/DLL_FAQ.txt
2495 Static zlib linking now works on Windows and the new --with-zlib-include
2496 --with-zlib-lib options to Configure can be used to supply the location
2497 of the headers and library. Gracefully handle case where zlib library
2501 *) Several fixes and enhancements to the OID generation code. The old code
2502 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2503 handle numbers larger than ULONG_MAX, truncated printing and had a
2504 non standard OBJ_obj2txt() behaviour.
2507 *) Add support for building of engines under engine/ as shared libraries
2508 under VC++ build system.
2511 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2512 Hopefully, we will not see any false combination of paths any more.
2515 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2517 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2518 (part of SSL_OP_ALL). This option used to disable the
2519 countermeasure against man-in-the-middle protocol-version
2520 rollback in the SSL 2.0 server implementation, which is a bad
2521 idea. (CVE-2005-2969)
2523 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2524 for Information Security, National Institute of Advanced Industrial
2525 Science and Technology [AIST], Japan)]
2527 *) Add two function to clear and return the verify parameter flags.
2530 *) Keep cipherlists sorted in the source instead of sorting them at
2531 runtime, thus removing the need for a lock.
2534 *) Avoid some small subgroup attacks in Diffie-Hellman.
2535 [Nick Mathewson and Ben Laurie]
2537 *) Add functions for well-known primes.
2540 *) Extended Windows CE support.
2541 [Satoshi Nakamura and Andy Polyakov]
2543 *) Initialize SSL_METHOD structures at compile time instead of during
2544 runtime, thus removing the need for a lock.
2547 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2548 attempting to decrypt each encrypted key in turn. Add support to
2552 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2554 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2557 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2560 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2561 key into the same file any more.
2564 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2567 *) Add -utf8 command line and config file option to 'ca'.
2568 [Stefan <stf@udoma.org]
2570 *) Removed the macro des_crypt(), as it seems to conflict with some
2571 libraries. Use DES_crypt().
2574 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2575 involves renaming the source and generated shared-libs for
2576 both. The engines will accept the corrected or legacy ids
2577 ('ncipher' and '4758_cca' respectively) when binding. NB,
2578 this only applies when building 'shared'.
2579 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2581 *) Add attribute functions to EVP_PKEY structure. Modify
2582 PKCS12_create() to recognize a CSP name attribute and
2583 use it. Make -CSP option work again in pkcs12 utility.
2586 *) Add new functionality to the bn blinding code:
2587 - automatic re-creation of the BN_BLINDING parameters after
2588 a fixed number of uses (currently 32)
2589 - add new function for parameter creation
2590 - introduce flags to control the update behaviour of the
2591 BN_BLINDING parameters
2592 - hide BN_BLINDING structure
2593 Add a second BN_BLINDING slot to the RSA structure to improve
2594 performance when a single RSA object is shared among several
2598 *) Add support for DTLS.
2599 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2601 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2602 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2605 *) Remove buggy and incompletet DH cert support from
2606 ssl/ssl_rsa.c and ssl/s3_both.c
2609 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2610 the apps/openssl applications.
2613 *) Compile clean with "-Wall -Wmissing-prototypes
2614 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2615 DEBUG_SAFESTACK must also be set.
2618 *) Change ./Configure so that certain algorithms can be disabled by default.
2619 The new counterpiece to "no-xxx" is "enable-xxx".
2621 The patented RC5 and MDC2 algorithms will now be disabled unless
2622 "enable-rc5" and "enable-mdc2", respectively, are specified.
2624 (IDEA remains enabled despite being patented. This is because IDEA
2625 is frequently required for interoperability, and there is no license
2626 fee for non-commercial use. As before, "no-idea" can be used to
2627 avoid this algorithm.)
2631 *) Add processing of proxy certificates (see RFC 3820). This work was
2632 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2633 EGEE (Enabling Grids for E-science in Europe).
2636 *) RC4 performance overhaul on modern architectures/implementations, such
2637 as Intel P4, IA-64 and AMD64.
2640 *) New utility extract-section.pl. This can be used specify an alternative
2641 section number in a pod file instead of having to treat each file as
2642 a separate case in Makefile. This can be done by adding two lines to the
2645 =for comment openssl_section:XXX
2647 The blank line is mandatory.
2651 *) New arguments -certform, -keyform and -pass for s_client and s_server
2652 to allow alternative format key and certificate files and passphrase
2656 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2657 update associated structures and add various utility functions.
2659 Add new policy related verify parameters, include policy checking in
2660 standard verify code. Enhance 'smime' application with extra parameters
2661 to support policy checking and print out.
2664 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2665 Nehemiah processors. These extensions support AES encryption in hardware
2666 as well as RNG (though RNG support is currently disabled).
2667 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2669 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2672 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2673 [Andy Polyakov and a number of other people]
2675 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2676 implementation contributed by IBM.
2677 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2679 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2680 exponent rather than 'unsigned long'. There is a corresponding change to
2681 the new 'rsa_keygen' element of the RSA_METHOD structure.
2682 [Jelte Jansen, Geoff Thorpe]
2684 *) Functionality for creating the initial serial number file is now
2685 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2687 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2688 number file to 1, which is bound to cause problems. To avoid
2689 the problems while respecting compatibility between different 0.9.7
2690 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2691 CA.pl for serial number initialization. With the new release 0.9.8,
2692 we can fix the problem directly in the 'ca' utility.)
2695 *) Reduced header interdepencies by declaring more opaque objects in
2696 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2697 give fewer recursive includes, which could break lazy source code - so
2698 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2699 developers should define this symbol when building and using openssl to
2700 ensure they track the recommended behaviour, interfaces, [etc], but
2701 backwards-compatible behaviour prevails when this isn't defined.
2704 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2707 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2708 This will generate a random key of the appropriate length based on the
2709 cipher context. The EVP_CIPHER can provide its own random key generation
2710 routine to support keys of a specific form. This is used in the des and
2711 3des routines to generate a key of the correct parity. Update S/MIME
2712 code to use new functions and hence generate correct parity DES keys.
2713 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2714 valid (weak or incorrect parity).
2717 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2718 as looking them up. This is useful when the verified structure may contain
2719 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2720 present unless the new PKCS7_NO_CRL flag is asserted.
2723 *) Extend ASN1 oid configuration module. It now additionally accepts the
2726 shortName = some long name, 1.2.3.4
2729 *) Reimplemented the BN_CTX implementation. There is now no more static
2730 limitation on the number of variables it can handle nor the depth of the
2731 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2732 information can now expand as required, and rather than having a single
2733 static array of bignums, BN_CTX now uses a linked-list of such arrays
2734 allowing it to expand on demand whilst maintaining the usefulness of
2735 BN_CTX's "bundling".
2738 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2739 to allow all RSA operations to function using a single BN_CTX.
2742 *) Preliminary support for certificate policy evaluation and checking. This
2743 is initially intended to pass the tests outlined in "Conformance Testing
2744 of Relying Party Client Certificate Path Processing Logic" v1.07.
2747 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2748 remained unused and not that useful. A variety of other little bignum
2749 tweaks and fixes have also been made continuing on from the audit (see
2753 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2754 associated ASN1, EVP and SSL functions and old ASN1 macros.
2757 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2758 and this should never fail. So the return value from the use of
2759 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2760 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2763 *) BN_CTX_get() should return zero-valued bignums, providing the same
2764 initialised value as BN_new().
2765 [Geoff Thorpe, suggested by Ulf Möller]
2767 *) Support for inhibitAnyPolicy certificate extension.
2770 *) An audit of the BIGNUM code is underway, for which debugging code is
2771 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2772 is considered valid when processing BIGNUMs, and causes execution to
2773 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2774 further steps are taken to deliberately pollute unused data in BIGNUM
2775 structures to try and expose faulty code further on. For now, openssl will
2776 (in its default mode of operation) continue to tolerate the inconsistent
2777 forms that it has tolerated in the past, but authors and packagers should
2778 consider trying openssl and their own applications when compiled with
2779 these debugging symbols defined. It will help highlight potential bugs in
2780 their own code, and will improve the test coverage for OpenSSL itself. At
2781 some point, these tighter rules will become openssl's default to improve
2782 maintainability, though the assert()s and other overheads will remain only
2783 in debugging configurations. See bn.h for more details.
2784 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2786 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2787 that can only be obtained through BN_CTX_new() (which implicitly
2788 initialises it). The presence of this function only made it possible
2789 to overwrite an existing structure (and cause memory leaks).
2792 *) Because of the callback-based approach for implementing LHASH as a
2793 template type, lh_insert() adds opaque objects to hash-tables and
2794 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2795 to clean up those corresponding objects before destroying the hash table
2796 (and losing the object pointers). So some over-zealous constifications in
2797 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2798 objects as "const" and the lh_doall[_arg] callback wrappers are not
2799 prototyped to have "const" restrictions on the object pointers they are
2800 given (and so aren't required to cast them away any more).
2803 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2804 (speed) prefers to use its own implementation. The two implementations
2805 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2806 its object type properly exposed (MS_TM) instead of casting to/from "char
2807 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2808 aren't necessarily the greatest nomenclatures - but this is what was used
2809 internally to the implementation so I've used that for now.
2812 *) Ensure that deprecated functions do not get compiled when
2813 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2814 the self-tests were still using deprecated key-generation functions so
2815 these have been updated also.
2818 *) Reorganise PKCS#7 code to separate the digest location functionality
2819 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2820 New function PKCS7_set_digest() to set the digest type for PKCS#7
2821 digestedData type. Add additional code to correctly generate the
2822 digestedData type and add support for this type in PKCS7 initialization
2826 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2827 structure of type "other".
2830 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2831 sure the loop does correctly stop and breaking ("division by zero")
2832 modulus operations are not performed. The (pre-generated) prime
2833 table crypto/bn/bn_prime.h was already correct, but it could not be
2834 re-generated on some platforms because of the "division by zero"
2835 situation in the script.
2836 [Ralf S. Engelschall]
2838 *) Update support for ECC-based TLS ciphersuites according to
2839 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2840 SHA-1 now is only used for "small" curves (where the
2841 representation of a field element takes up to 24 bytes); for
2842 larger curves, the field element resulting from ECDH is directly
2843 used as premaster secret.
2844 [Douglas Stebila (Sun Microsystems Laboratories)]
2846 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2847 curve secp160r1 to the tests.
2848 [Douglas Stebila (Sun Microsystems Laboratories)]
2850 *) Add the possibility to load symbols globally with DSO.
2851 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2853 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2854 control of the error stack.
2857 *) Add support for STORE in ENGINE.
2860 *) Add the STORE type. The intention is to provide a common interface
2861 to certificate and key stores, be they simple file-based stores, or
2862 HSM-type store, or LDAP stores, or...
2863 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2866 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2867 pass a list of arguments to any function as well as provide a way
2868 for a function to pass data back to the caller.
2871 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2872 works like BUF_strdup() but can be used to duplicate a portion of
2873 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2877 *) Add the function sk_find_ex() which works like sk_find(), but will
2878 return an index to an element even if an exact match couldn't be
2879 found. The index is guaranteed to point at the element where the
2880 searched-for key would be inserted to preserve sorting order.
2883 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2884 takes an extra flags argument for optional functionality. Currently,
2885 the following flags are defined:
2887 OBJ_BSEARCH_VALUE_ON_NOMATCH
2888 This one gets OBJ_bsearch_ex() to return a pointer to the first
2889 element where the comparing function returns a negative or zero
2892 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2893 This one gets OBJ_bsearch_ex() to return a pointer to the first
2894 element where the comparing function returns zero. This is useful
2895 if there are more than one element where the comparing function
2899 *) Make it possible to create self-signed certificates with 'openssl ca'
2900 in such a way that the self-signed certificate becomes part of the
2901 CA database and uses the same mechanisms for serial number generation
2902 as all other certificate signing. The new flag '-selfsign' enables
2903 this functionality. Adapt CA.sh and CA.pl.in.
2906 *) Add functionality to check the public key of a certificate request
2907 against a given private. This is useful to check that a certificate
2908 request can be signed by that key (self-signing).
2911 *) Make it possible to have multiple active certificates with the same
2912 subject in the CA index file. This is done only if the keyword
2913 'unique_subject' is set to 'no' in the main CA section (default
2914 if 'CA_default') of the configuration file. The value is saved
2915 with the database itself in a separate index attribute file,
2916 named like the index file with '.attr' appended to the name.
2919 *) Generate muti valued AVAs using '+' notation in config files for
2923 *) Support for nameConstraints certificate extension.
2926 *) Support for policyConstraints certificate extension.
2929 *) Support for policyMappings certificate extension.
2932 *) Make sure the default DSA_METHOD implementation only uses its
2933 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2934 and change its own handlers to be NULL so as to remove unnecessary
2935 indirection. This lets alternative implementations fallback to the
2936 default implementation more easily.
2939 *) Support for directoryName in GeneralName related extensions
2943 *) Make it possible to link applications using Makefile.shared.
2944 Make that possible even when linking against static libraries!
2947 *) Support for single pass processing for S/MIME signing. This now
2948 means that S/MIME signing can be done from a pipe, in addition
2949 cleartext signing (multipart/signed type) is effectively streaming
2950 and the signed data does not need to be all held in memory.
2952 This is done with a new flag PKCS7_STREAM. When this flag is set
2953 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2954 is done after the data is output (and digests calculated) in
2955 SMIME_write_PKCS7().
2958 *) Add full support for -rpath/-R, both in shared libraries and
2959 applications, at least on the platforms where it's known how
2963 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2964 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2965 will now compute a table of multiples of the generator that
2966 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2967 faster (notably in the case of a single point multiplication,
2968 scalar * generator).
2969 [Nils Larsch, Bodo Moeller]
2971 *) IPv6 support for certificate extensions. The various extensions
2972 which use the IP:a.b.c.d can now take IPv6 addresses using the
2973 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2977 *) Added an ENGINE that implements RSA by performing private key
2978 exponentiations with the GMP library. The conversions to and from
2979 GMP's mpz_t format aren't optimised nor are any montgomery forms
2980 cached, and on x86 it appears OpenSSL's own performance has caught up.
2981 However there are likely to be other architectures where GMP could
2982 provide a boost. This ENGINE is not built in by default, but it can be
2983 specified at Configure time and should be accompanied by the necessary
2984 linker additions, eg;
2985 ./config -DOPENSSL_USE_GMP -lgmp
2988 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2989 testing availability of engines with "-t" - the old behaviour is
2990 produced by increasing the feature's verbosity with "-tt".
2993 *) ECDSA routines: under certain error conditions uninitialized BN objects
2994 could be freed. Solution: make sure initialization is performed early
2995 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2999 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3000 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3001 software implementations. For DSA and DH, parameter generation can
3002 also be overriden by providing the appropriate method callbacks.
3005 *) Change the "progress" mechanism used in key-generation and
3006 primality testing to functions that take a new BN_GENCB pointer in
3007 place of callback/argument pairs. The new API functions have "_ex"
3008 postfixes and the older functions are reimplemented as wrappers for
3009 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3010 declarations of the old functions to help (graceful) attempts to
3011 migrate to the new functions. Also, the new key-generation API
3012 functions operate on a caller-supplied key-structure and return
3013 success/failure rather than returning a key or NULL - this is to
3014 help make "keygen" another member function of RSA_METHOD etc.
3016 Example for using the new callback interface:
3018 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3022 BN_GENCB_set(&my_cb, my_callback, my_arg);
3024 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3025 /* For the meaning of a, b in calls to my_callback(), see the
3026 * documentation of the function that calls the callback.
3027 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3028 * my_callback should return 1 if it wants BN_is_prime_ex()
3029 * to continue, or 0 to stop.
3034 *) Change the ZLIB compression method to be stateful, and make it
3035 available to TLS with the number defined in
3036 draft-ietf-tls-compression-04.txt.
3039 *) Add the ASN.1 structures and functions for CertificatePair, which
3040 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3042 CertificatePair ::= SEQUENCE {
3043 forward [0] Certificate OPTIONAL,
3044 reverse [1] Certificate OPTIONAL,
3045 -- at least one of the pair shall be present -- }
3047 Also implement the PEM functions to read and write certificate
3048 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3050 This needed to be defined, mostly for the sake of the LDAP
3051 attribute crossCertificatePair, but may prove useful elsewhere as
3055 *) Make it possible to inhibit symlinking of shared libraries in
3056 Makefile.shared, for Cygwin's sake.
3059 *) Extend the BIGNUM API by creating a function
3060 void BN_set_negative(BIGNUM *a, int neg);
3061 and a macro that behave like
3062 int BN_is_negative(const BIGNUM *a);
3064 to avoid the need to access 'a->neg' directly in applications.
3067 *) Implement fast modular reduction for pseudo-Mersenne primes
3068 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3069 EC_GROUP_new_curve_GFp() will now automatically use this
3071 [Nils Larsch <nla@trustcenter.de>]
3073 *) Add new lock type (CRYPTO_LOCK_BN).
3076 *) Change the ENGINE framework to automatically load engines
3077 dynamically from specific directories unless they could be
3078 found to already be built in or loaded. Move all the
3079 current engines except for the cryptodev one to a new
3081 The engines in engines/ are built as shared libraries if
3082 the "shared" options was given to ./Configure or ./config.
3083 Otherwise, they are inserted in libcrypto.a.
3084 /usr/local/ssl/engines is the default directory for dynamic
3085 engines, but that can be overriden at configure time through
3086 the usual use of --prefix and/or --openssldir, and at run
3087 time with the environment variable OPENSSL_ENGINES.
3088 [Geoff Thorpe and Richard Levitte]
3090 *) Add Makefile.shared, a helper makefile to build shared
3091 libraries. Addapt Makefile.org.
3094 *) Add version info to Win32 DLLs.
3095 [Peter 'Luna' Runestig" <peter@runestig.com>]
3097 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3098 can be added using this API to created arbitrary PKCS#12
3099 files while avoiding the low level API.
3101 New options to PKCS12_create(), key or cert can be NULL and
3102 will then be omitted from the output file. The encryption
3103 algorithm NIDs can be set to -1 for no encryption, the mac
3104 iteration count can be set to 0 to omit the mac.
3106 Enhance pkcs12 utility by making the -nokeys and -nocerts
3107 options work when creating a PKCS#12 file. New option -nomac
3108 to omit the mac, NONE can be set for an encryption algorithm.
3109 New code is modified to use the enhanced PKCS12_create()
3110 instead of the low level API.
3113 *) Extend ASN1 encoder to support indefinite length constructed
3114 encoding. This can output sequences tags and octet strings in
3115 this form. Modify pk7_asn1.c to support indefinite length
3116 encoding. This is experimental and needs additional code to
3117 be useful, such as an ASN1 bio and some enhanced streaming
3120 Extend template encode functionality so that tagging is passed
3121 down to the template encoder.
3124 *) Let 'openssl req' fail if an argument to '-newkey' is not
3125 recognized instead of using RSA as a default.
3128 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3129 As these are not official, they are not included in "ALL";
3130 the "ECCdraft" ciphersuite group alias can be used to select them.
3131 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3133 *) Add ECDH engine support.
3134 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3136 *) Add ECDH in new directory crypto/ecdh/.
3137 [Douglas Stebila (Sun Microsystems Laboratories)]
3139 *) Let BN_rand_range() abort with an error after 100 iterations
3140 without success (which indicates a broken PRNG).
3143 *) Change BN_mod_sqrt() so that it verifies that the input value
3144 is really the square of the return value. (Previously,
3145 BN_mod_sqrt would show GIGO behaviour.)
3148 *) Add named elliptic curves over binary fields from X9.62, SECG,
3149 and WAP/WTLS; add OIDs that were still missing.
3151 [Sheueling Chang Shantz and Douglas Stebila
3152 (Sun Microsystems Laboratories)]
3154 *) Extend the EC library for elliptic curves over binary fields
3155 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3158 EC_GF2m_simple_method
3162 EC_GROUP_new_curve_GF2m
3163 EC_GROUP_set_curve_GF2m
3164 EC_GROUP_get_curve_GF2m
3165 EC_POINT_set_affine_coordinates_GF2m
3166 EC_POINT_get_affine_coordinates_GF2m
3167 EC_POINT_set_compressed_coordinates_GF2m
3169 Point compression for binary fields is disabled by default for
3170 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3173 As binary polynomials are represented as BIGNUMs, various members
3174 of the EC_GROUP and EC_POINT data structures can be shared
3175 between the implementations for prime fields and binary fields;
3176 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3177 are essentially identical to their ..._GFp counterparts.
3178 (For simplicity, the '..._GFp' prefix has been dropped from
3179 various internal method names.)
3181 An internal 'field_div' method (similar to 'field_mul' and
3182 'field_sqr') has been added; this is used only for binary fields.
3184 [Sheueling Chang Shantz and Douglas Stebila
3185 (Sun Microsystems Laboratories)]
3187 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3188 through methods ('mul', 'precompute_mult').
3190 The generic implementations (now internally called 'ec_wNAF_mul'
3191 and 'ec_wNAF_precomputed_mult') remain the default if these
3192 methods are undefined.
3194 [Sheueling Chang Shantz and Douglas Stebila
3195 (Sun Microsystems Laboratories)]
3197 *) New function EC_GROUP_get_degree, which is defined through
3198 EC_METHOD. For curves over prime fields, this returns the bit
3199 length of the modulus.
3201 [Sheueling Chang Shantz and Douglas Stebila
3202 (Sun Microsystems Laboratories)]
3204 *) New functions EC_GROUP_dup, EC_POINT_dup.
3205 (These simply call ..._new and ..._copy).
3207 [Sheueling Chang Shantz and Douglas Stebila
3208 (Sun Microsystems Laboratories)]
3210 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3211 Polynomials are represented as BIGNUMs (where the sign bit is not
3212 used) in the following functions [macros]:
3215 BN_GF2m_sub [= BN_GF2m_add]
3216 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3217 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3218 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3220 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3221 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3222 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3223 BN_GF2m_cmp [= BN_ucmp]
3225 (Note that only the 'mod' functions are actually for fields GF(2^m).
3226 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3228 For some functions, an the irreducible polynomial defining a
3229 field can be given as an 'unsigned int[]' with strictly
3230 decreasing elements giving the indices of those bits that are set;
3231 i.e., p[] represents the polynomial
3232 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3234 p[0] > p[1] > ... > p[k] = 0.
3235 This applies to the following functions:
3240 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3241 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3243 BN_GF2m_mod_sqrt_arr
3244 BN_GF2m_mod_solve_quad_arr
3248 Conversion can be performed by the following functions:
3253 bntest.c has additional tests for binary polynomial arithmetic.
3255 Two implementations for BN_GF2m_mod_div() are available.
3256 The default algorithm simply uses BN_GF2m_mod_inv() and
3257 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3258 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3259 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3261 [Sheueling Chang Shantz and Douglas Stebila
3262 (Sun Microsystems Laboratories)]
3264 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3265 functionality is disabled at compile-time.
3266 [Douglas Stebila <douglas.stebila@sun.com>]
3268 *) Change default behaviour of 'openssl asn1parse' so that more
3269 information is visible when viewing, e.g., a certificate:
3271 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3272 mode the content of non-printable OCTET STRINGs is output in a
3273 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3274 avoid the appearance of a printable string.
3275 [Nils Larsch <nla@trustcenter.de>]
3277 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3279 EC_GROUP_set_asn1_flag()
3280 EC_GROUP_get_asn1_flag()
3281 EC_GROUP_set_point_conversion_form()
3282 EC_GROUP_get_point_conversion_form()
3283 These control ASN1 encoding details:
3284 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3285 has been set to OPENSSL_EC_NAMED_CURVE.
3286 - Points are encoded in uncompressed form by default; options for
3287 asn1_for are as for point2oct, namely
3288 POINT_CONVERSION_COMPRESSED
3289 POINT_CONVERSION_UNCOMPRESSED
3290 POINT_CONVERSION_HYBRID
3292 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3295 EC_GROUP_get0_seed()
3296 EC_GROUP_get_seed_len()
3297 This is used only for ASN1 purposes (so far).
3298 [Nils Larsch <nla@trustcenter.de>]
3300 *) Add 'field_type' member to EC_METHOD, which holds the NID
3301 of the appropriate field type OID. The new function
3302 EC_METHOD_get_field_type() returns this value.
3303 [Nils Larsch <nla@trustcenter.de>]
3308 EC_POINT_point2hex()
3309 EC_POINT_hex2point()
3310 providing useful interfaces to EC_POINT_point2oct() and
3311 EC_POINT_oct2point().
3312 [Nils Larsch <nla@trustcenter.de>]
3314 *) Change internals of the EC library so that the functions
3315 EC_GROUP_set_generator()
3316 EC_GROUP_get_generator()
3317 EC_GROUP_get_order()
3318 EC_GROUP_get_cofactor()
3319 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3320 to methods, which would lead to unnecessary code duplication when
3321 adding different types of curves.
3322 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3324 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3325 arithmetic, and such that modified wNAFs are generated
3326 (which avoid length expansion in many cases).
3329 *) Add a function EC_GROUP_check_discriminant() (defined via
3330 EC_METHOD) that verifies that the curve discriminant is non-zero.
3332 Add a function EC_GROUP_check() that makes some sanity tests
3333 on a EC_GROUP, its generator and order. This includes
3334 EC_GROUP_check_discriminant().
3335 [Nils Larsch <nla@trustcenter.de>]
3337 *) Add ECDSA in new directory crypto/ecdsa/.
3339 Add applications 'openssl ecparam' and 'openssl ecdsa'
3340 (these are based on 'openssl dsaparam' and 'openssl dsa').
3342 ECDSA support is also included in various other files across the
3343 library. Most notably,
3344 - 'openssl req' now has a '-newkey ecdsa:file' option;
3345 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3346 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3347 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3348 them suitable for ECDSA where domain parameters must be
3349 extracted before the specific public key;
3350 - ECDSA engine support has been added.
3351 [Nils Larsch <nla@trustcenter.de>]
3353 *) Include some named elliptic curves, and add OIDs from X9.62,
3354 SECG, and WAP/WTLS. Each curve can be obtained from the new
3356 EC_GROUP_new_by_curve_name(),
3357 and the list of available named curves can be obtained with
3358 EC_get_builtin_curves().
3359 Also add a 'curve_name' member to EC_GROUP objects, which can be
3361 EC_GROUP_set_curve_name()
3362 EC_GROUP_get_curve_name()
3363 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3365 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3366 was actually never needed) and in BN_mul(). The removal in BN_mul()
3367 required a small change in bn_mul_part_recursive() and the addition
3368 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3369 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3370 bn_sub_words() and bn_add_words() except they take arrays with
3374 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3376 *) Cleanse PEM buffers before freeing them since they may contain
3378 [Benjamin Bennett <ben@psc.edu>]
3380 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3381 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3382 authentication-only ciphersuites.
3385 *) Since AES128 and AES256 share a single mask bit in the logic of
3386 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3387 kludge to work properly if AES128 is available and AES256 isn't.
3390 *) Expand security boundary to match 1.1.1 module.
3393 *) Remove redundant features: hash file source, editing of test vectors
3394 modify fipsld to use external fips_premain.c signature.
3397 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3398 run algorithm test programs.
3401 *) Make algorithm test programs more tolerant of whitespace.
3404 *) Have SSL/TLS server implementation tolerate "mismatched" record
3405 protocol version while receiving ClientHello even if the
3406 ClientHello is fragmented. (The server can't insist on the
3407 particular protocol version it has chosen before the ServerHello
3408 message has informed the client about his choice.)
3411 *) Load error codes if they are not already present instead of using a
3412 static variable. This allows them to be cleanly unloaded and reloaded.
3415 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3417 *) Introduce limits to prevent malicious keys being able to
3418 cause a denial of service. (CVE-2006-2940)
3419 [Steve Henson, Bodo Moeller]
3421 *) Fix ASN.1 parsing of certain invalid structures that can result
3422 in a denial of service. (CVE-2006-2937) [Steve Henson]
3424 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3425 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3427 *) Fix SSL client code which could crash if connecting to a
3428 malicious SSLv2 server. (CVE-2006-4343)
3429 [Tavis Ormandy and Will Drewry, Google Security Team]
3431 *) Change ciphersuite string processing so that an explicit
3432 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3433 will no longer include "AES128-SHA"), and any other similar
3434 ciphersuite (same bitmap) from *other* protocol versions (so that
3435 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3436 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3437 changes from 0.9.8b and 0.9.8d.
3440 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3442 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3443 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3445 *) Change the Unix randomness entropy gathering to use poll() when
3446 possible instead of select(), since the latter has some
3447 undesirable limitations.
3448 [Darryl Miles via Richard Levitte and Bodo Moeller]
3450 *) Disable rogue ciphersuites:
3452 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3453 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3454 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3456 The latter two were purportedly from
3457 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3460 Also deactive the remaining ciphersuites from
3461 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3462 unofficial, and the ID has long expired.
3465 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3466 dual-core machines) and other potential thread-safety issues.
3469 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3471 *) Adapt fipsld and the build system to link against the validated FIPS
3472 module in FIPS mode.
3475 *) Fixes for VC++ 2005 build under Windows.
3478 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3479 from a Windows bash shell such as MSYS. It is autodetected from the
3480 "config" script when run from a VC++ environment. Modify standard VC++
3481 build to use fipscanister.o from the GNU make build.
3484 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3486 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3487 The value now differs depending on if you build for FIPS or not.
3488 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3489 safely run with a non-FIPSed libcrypto, as it may crash because of
3490 the difference induced by this change.
3493 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3495 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3496 (part of SSL_OP_ALL). This option used to disable the
3497 countermeasure against man-in-the-middle protocol-version
3498 rollback in the SSL 2.0 server implementation, which is a bad
3499 idea. (CVE-2005-2969)
3501 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3502 for Information Security, National Institute of Advanced Industrial
3503 Science and Technology [AIST], Japan)]
3505 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3506 mainly for FIPS compliance and not fully integrated at this stage.
3509 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3510 the exponentiation using a fixed-length exponent. (Otherwise,
3511 the information leaked through timing could expose the secret key
3512 after many signatures; cf. Bleichenbacher's attack on DSA with
3516 *) Make a new fixed-window mod_exp implementation the default for
3517 RSA, DSA, and DH private-key operations so that the sequence of
3518 squares and multiplies and the memory access pattern are
3519 independent of the particular secret key. This will mitigate
3520 cache-timing and potential related attacks.
3522 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3523 and this is automatically used by BN_mod_exp_mont() if the new flag
3524 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3525 will use this BN flag for private exponents unless the flag
3526 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3527 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3529 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3531 *) Change the client implementation for SSLv23_method() and
3532 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3533 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3534 (Previously, the SSL 2.0 backwards compatible Client Hello
3535 message format would be used even with SSL_OP_NO_SSLv2.)
3538 *) Add support for smime-type MIME parameter in S/MIME messages which some
3542 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3543 a threadsafe manner. Modify rsa code to use new function and add calls
3544 to dsa and dh code (which had race conditions before).
3547 *) Include the fixed error library code in the C error file definitions
3548 instead of fixing them up at runtime. This keeps the error code
3549 structures constant.
3552 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3554 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3557 *) Fixes for newer kerberos headers. NB: the casts are needed because
3558 the 'length' field is signed on one version and unsigned on another
3559 with no (?) obvious way to tell the difference, without these VC++
3560 complains. Also the "definition" of FAR (blank) is no longer included
3561 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3562 some needed definitions.
3565 *) Undo Cygwin change.
3568 *) Added support for proxy certificates according to RFC 3820.
3569 Because they may be a security thread to unaware applications,
3570 they must be explicitely allowed in run-time. See
3571 docs/HOWTO/proxy_certificates.txt for further information.
3574 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3576 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3577 server and client random values. Previously
3578 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3579 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3581 This change has negligible security impact because:
3583 1. Server and client random values still have 24 bytes of pseudo random
3586 2. Server and client random values are sent in the clear in the initial
3589 3. The master secret is derived using the premaster secret (48 bytes in
3590 size for static RSA ciphersuites) as well as client server and random
3593 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3596 [Stephen Henson, reported by UK NISCC]
3598 *) Use Windows randomness collection on Cygwin.
3601 *) Fix hang in EGD/PRNGD query when communication socket is closed
3602 prematurely by EGD/PRNGD.
3603 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3605 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3608 *) Back-port of selected performance improvements from development
3609 branch, as well as improved support for PowerPC platforms.
3612 *) Add lots of checks for memory allocation failure, error codes to indicate
3613 failure and freeing up memory if a failure occurs.
3614 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3616 *) Add new -passin argument to dgst.
3619 *) Perform some character comparisons of different types in X509_NAME_cmp:
3620 this is needed for some certificates that reencode DNs into UTF8Strings
3621 (in violation of RFC3280) and can't or wont issue name rollover
3625 *) Make an explicit check during certificate validation to see that
3626 the CA setting in each certificate on the chain is correct. As a
3627 side effect always do the following basic checks on extensions,
3628 not just when there's an associated purpose to the check:
3630 - if there is an unhandled critical extension (unless the user
3631 has chosen to ignore this fault)
3632 - if the path length has been exceeded (if one is set at all)
3633 - that certain extensions fit the associated purpose (if one has
3637 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3639 *) Avoid a race condition when CRLs are checked in a multi threaded
3640 environment. This would happen due to the reordering of the revoked
3641 entries during signature checking and serial number lookup. Now the
3642 encoding is cached and the serial number sort performed under a lock.
3643 Add new STACK function sk_is_sorted().
3646 *) Add Delta CRL to the extension code.
3649 *) Various fixes to s3_pkt.c so alerts are sent properly.
3650 [David Holmes <d.holmes@f5.com>]
3652 *) Reduce the chances of duplicate issuer name and serial numbers (in
3653 violation of RFC3280) using the OpenSSL certificate creation utilities.
3654 This is done by creating a random 64 bit value for the initial serial
3655 number when a serial number file is created or when a self signed
3656 certificate is created using 'openssl req -x509'. The initial serial
3657 number file is created using 'openssl x509 -next_serial' in CA.pl
3658 rather than being initialized to 1.
3661 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3663 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3664 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3665 [Joe Orton, Steve Henson]
3667 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3669 [Joe Orton, Steve Henson]
3671 *) Make it possible to have multiple active certificates with the same
3672 subject in the CA index file. This is done only if the keyword
3673 'unique_subject' is set to 'no' in the main CA section (default
3674 if 'CA_default') of the configuration file. The value is saved
3675 with the database itself in a separate index attribute file,
3676 named like the index file with '.attr' appended to the name.
3679 *) X509 verify fixes. Disable broken certificate workarounds when
3680 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3681 keyUsage extension present. Don't accept CRLs with unhandled critical
3682 extensions: since verify currently doesn't process CRL extensions this
3683 rejects a CRL with *any* critical extensions. Add new verify error codes
3687 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3688 A clarification of RFC2560 will require the use of OCTET STRINGs and
3689 some implementations cannot handle the current raw format. Since OpenSSL
3690 copies and compares OCSP nonces as opaque blobs without any attempt at
3691 parsing them this should not create any compatibility issues.
3694 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3695 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3696 this HMAC (and other) operations are several times slower than OpenSSL
3700 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3701 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3703 *) Use the correct content when signing type "other".
3706 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3708 *) Fix various bugs revealed by running the NISCC test suite:
3710 Stop out of bounds reads in the ASN1 code when presented with
3711 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3713 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3715 If verify callback ignores invalid public key errors don't try to check
3716 certificate signature with the NULL public key.
3720 *) New -ignore_err option in ocsp application to stop the server
3721 exiting on the first error in a request.
3724 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3725 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3729 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3730 extra data after the compression methods not only for TLS 1.0
3731 but also for SSL 3.0 (as required by the specification).
3732 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3734 *) Change X509_certificate_type() to mark the key as exported/exportable
3735 when it's 512 *bits* long, not 512 bytes.
3738 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3739 blocks during encryption.
3742 *) Various fixes to base64 BIO and non blocking I/O. On write
3743 flushes were not handled properly if the BIO retried. On read
3744 data was not being buffered properly and had various logic bugs.
3745 This also affects blocking I/O when the data being decoded is a
3749 *) Various S/MIME bugfixes and compatibility changes:
3750 output correct application/pkcs7 MIME type if
3751 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3752 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3753 of files as .eml work). Correctly handle very long lines in MIME
3757 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3759 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3760 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3761 a protocol version number mismatch like a decryption error
3762 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3765 *) Turn on RSA blinding by default in the default implementation
3766 to avoid a timing attack. Applications that don't want it can call
3767 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3768 They would be ill-advised to do so in most cases.
3769 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3771 *) Change RSA blinding code so that it works when the PRNG is not
3772 seeded (in this case, the secret RSA exponent is abused as
3773 an unpredictable seed -- if it is not unpredictable, there
3774 is no point in blinding anyway). Make RSA blinding thread-safe
3775 by remembering the creator's thread ID in rsa->blinding and
3776 having all other threads use local one-time blinding factors
3777 (this requires more computation than sharing rsa->blinding, but
3778 avoids excessive locking; and if an RSA object is not shared
3779 between threads, blinding will still be very fast).
3782 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3783 ENGINE as defaults for all supported algorithms irrespective of
3784 the 'flags' parameter. 'flags' is now honoured, so applications
3785 should make sure they are passing it correctly.
3788 *) Target "mingw" now allows native Windows code to be generated in
3789 the Cygwin environment as well as with the MinGW compiler.
3792 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3794 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3795 via timing by performing a MAC computation even if incorrrect
3796 block cipher padding has been found. This is a countermeasure
3797 against active attacks where the attacker has to distinguish
3798 between bad padding and a MAC verification error. (CVE-2003-0078)
3800 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3801 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3802 Martin Vuagnoux (EPFL, Ilion)]
3804 *) Make the no-err option work as intended. The intention with no-err
3805 is not to have the whole error stack handling routines removed from
3806 libcrypto, it's only intended to remove all the function name and
3807 reason texts, thereby removing some of the footprint that may not
3808 be interesting if those errors aren't displayed anyway.
3810 NOTE: it's still possible for any application or module to have it's
3811 own set of error texts inserted. The routines are there, just not
3812 used by default when no-err is given.
3815 *) Add support for FreeBSD on IA64.
3816 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3818 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3819 Kerberos function mit_des_cbc_cksum(). Before this change,
3820 the value returned by DES_cbc_cksum() was like the one from
3821 mit_des_cbc_cksum(), except the bytes were swapped.
3822 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3824 *) Allow an application to disable the automatic SSL chain building.
3825 Before this a rather primitive chain build was always performed in
3826 ssl3_output_cert_chain(): an application had no way to send the
3827 correct chain if the automatic operation produced an incorrect result.
3829 Now the chain builder is disabled if either:
3831 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3833 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3835 The reasoning behind this is that an application would not want the
3836 auto chain building to take place if extra chain certificates are
3837 present and it might also want a means of sending no additional
3838 certificates (for example the chain has two certificates and the
3842 *) Add the possibility to build without the ENGINE framework.
3843 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3845 *) Under Win32 gmtime() can return NULL: check return value in
3846 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3849 *) DSA routines: under certain error conditions uninitialized BN objects
3850 could be freed. Solution: make sure initialization is performed early
3851 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3852 Nils Larsch <nla@trustcenter.de> via PR#459)
3855 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3856 checked on reconnect on the client side, therefore session resumption
3857 could still fail with a "ssl session id is different" error. This
3858 behaviour is masked when SSL_OP_ALL is used due to
3859 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3860 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3861 followup to PR #377.
3864 *) IA-32 assembler support enhancements: unified ELF targets, support
3865 for SCO/Caldera platforms, fix for Cygwin shared build.
3868 *) Add support for FreeBSD on sparc64. As a consequence, support for
3869 FreeBSD on non-x86 processors is separate from x86 processors on
3870 the config script, much like the NetBSD support.
3871 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3873 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3875 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3878 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3879 code (06) was taken as the first octet of the session ID and the last
3880 octet was ignored consequently. As a result SSLv2 client side session
3881 caching could not have worked due to the session ID mismatch between
3883 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3887 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3888 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3892 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3893 seems that in spite of existing for more than a year, many application
3894 author have done nothing to provide the necessary callbacks, which
3895 means that this particular engine will not work properly anywhere.
3896 This is a very unfortunate situation which forces us, in the name
3897 of usability, to give the hw_ncipher.c a static lock, which is part
3899 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3900 appear in 0.9.8 or later. We EXPECT application authors to have
3901 dealt properly with this when 0.9.8 is released (unless we actually
3902 make such changes in the libcrypto locking code that changes will
3903 have to be made anyway).
3906 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3907 octets have been read, EOF or an error occurs. Without this change
3908 some truncated ASN1 structures will not produce an error.
3911 *) Disable Heimdal support, since it hasn't been fully implemented.
3912 Still give the possibility to force the use of Heimdal, but with
3913 warnings and a request that patches get sent to openssl-dev.
3916 *) Add the VC-CE target, introduce the WINCE sysname, and add
3917 INSTALL.WCE and appropriate conditionals to make it build.
3918 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3920 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3921 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3922 edit numbers of the version.
3923 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3925 *) Introduce safe string copy and catenation functions
3926 (BUF_strlcpy() and BUF_strlcat()).
3927 [Ben Laurie (CHATS) and Richard Levitte]
3929 *) Avoid using fixed-size buffers for one-line DNs.
3930 [Ben Laurie (CHATS)]
3932 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3933 resizing buffers containing secrets, and use where appropriate.
3934 [Ben Laurie (CHATS)]
3936 *) Avoid using fixed size buffers for configuration file location.
3937 [Ben Laurie (CHATS)]
3939 *) Avoid filename truncation for various CA files.
3940 [Ben Laurie (CHATS)]
3942 *) Use sizeof in preference to magic numbers.
3943 [Ben Laurie (CHATS)]
3945 *) Avoid filename truncation in cert requests.
3946 [Ben Laurie (CHATS)]
3948 *) Add assertions to check for (supposedly impossible) buffer
3950 [Ben Laurie (CHATS)]
3952 *) Don't cache truncated DNS entries in the local cache (this could
3953 potentially lead to a spoofing attack).
3954 [Ben Laurie (CHATS)]
3956 *) Fix various buffers to be large enough for hex/decimal
3957 representations in a platform independent manner.
3958 [Ben Laurie (CHATS)]
3960 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3961 resizing buffers containing secrets, and use where appropriate.
3962 [Ben Laurie (CHATS)]
3964 *) Add BIO_indent() to avoid much slightly worrying code to do
3966 [Ben Laurie (CHATS)]
3968 *) Convert sprintf()/BIO_puts() to BIO_printf().
3969 [Ben Laurie (CHATS)]
3971 *) buffer_gets() could terminate with the buffer only half
3973 [Ben Laurie (CHATS)]
3975 *) Add assertions to prevent user-supplied crypto functions from
3976 overflowing internal buffers by having large block sizes, etc.
3977 [Ben Laurie (CHATS)]
3979 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3981 [Ben Laurie (CHATS)]
3983 *) Eliminate unused copy of key in RC4.
3984 [Ben Laurie (CHATS)]
3986 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3987 [Ben Laurie (CHATS)]
3989 *) Fix off-by-one error in EGD path.
3990 [Ben Laurie (CHATS)]
3992 *) If RANDFILE path is too long, ignore instead of truncating.
3993 [Ben Laurie (CHATS)]
3995 *) Eliminate unused and incorrectly sized X.509 structure
3997 [Ben Laurie (CHATS)]
3999 *) Eliminate unused and dangerous function knumber().
4000 [Ben Laurie (CHATS)]
4002 *) Eliminate unused and dangerous structure, KSSL_ERR.
4003 [Ben Laurie (CHATS)]
4005 *) Protect against overlong session ID context length in an encoded
4006 session object. Since these are local, this does not appear to be
4008 [Ben Laurie (CHATS)]
4010 *) Change from security patch (see 0.9.6e below) that did not affect
4011 the 0.9.6 release series:
4013 Remote buffer overflow in SSL3 protocol - an attacker could
4014 supply an oversized master key in Kerberos-enabled versions.
4016 [Ben Laurie (CHATS)]
4018 *) Change the SSL kerb5 codes to match RFC 2712.
4021 *) Make -nameopt work fully for req and add -reqopt switch.
4022 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4024 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4025 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4027 *) Make sure tests can be performed even if the corresponding algorithms
4028 have been removed entirely. This was also the last step to make
4029 OpenSSL compilable with DJGPP under all reasonable conditions.
4030 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4032 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4033 to allow version independent disabling of normally unselected ciphers,
4034 which may be activated as a side-effect of selecting a single cipher.
4036 (E.g., cipher list string "RSA" enables ciphersuites that are left
4037 out of "ALL" because they do not provide symmetric encryption.
4038 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4039 [Lutz Jaenicke, Bodo Moeller]
4041 *) Add appropriate support for separate platform-dependent build
4042 directories. The recommended way to make a platform-dependent
4043 build directory is the following (tested on Linux), maybe with
4046 # Place yourself outside of the OpenSSL source tree. In
4047 # this example, the environment variable OPENSSL_SOURCE
4048 # is assumed to contain the absolute OpenSSL source directory.
4049 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4050 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4051 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4052 mkdir -p `dirname $F`
4053 ln -s $OPENSSL_SOURCE/$F $F
4056 To be absolutely sure not to disturb the source tree, a "make clean"
4057 is a good thing. If it isn't successfull, don't worry about it,
4058 it probably means the source directory is very clean.
4061 *) Make sure any ENGINE control commands make local copies of string
4062 pointers passed to them whenever necessary. Otherwise it is possible
4063 the caller may have overwritten (or deallocated) the original string
4064 data when a later ENGINE operation tries to use the stored values.
4065 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4067 *) Improve diagnostics in file reading and command-line digests.
4068 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4070 *) Add AES modes CFB and OFB to the object database. Correct an
4071 error in AES-CFB decryption.
4074 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4075 allows existing EVP_CIPHER_CTX structures to be reused after
4076 calling EVP_*Final(). This behaviour is used by encryption
4077 BIOs and some applications. This has the side effect that
4078 applications must explicitly clean up cipher contexts with
4079 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4082 *) Check the values of dna and dnb in bn_mul_recursive before calling
4083 bn_mul_comba (a non zero value means the a or b arrays do not contain
4084 n2 elements) and fallback to bn_mul_normal if either is not zero.
4087 *) Fix escaping of non-ASCII characters when using the -subj option
4088 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4091 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4092 form for "surname", serialNumber has no short form.
4093 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4094 therefore remove "mail" short name for "internet 7".
4095 The OID for unique identifiers in X509 certificates is
4096 x500UniqueIdentifier, not uniqueIdentifier.
4097 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4100 *) Add an "init" command to the ENGINE config module and auto initialize
4101 ENGINEs. Without any "init" command the ENGINE will be initialized
4102 after all ctrl commands have been executed on it. If init=1 the
4103 ENGINE is initailized at that point (ctrls before that point are run
4104 on the uninitialized ENGINE and after on the initialized one). If
4105 init=0 then the ENGINE will not be iniatialized at all.
4108 *) Fix the 'app_verify_callback' interface so that the user-defined
4109 argument is actually passed to the callback: In the
4110 SSL_CTX_set_cert_verify_callback() prototype, the callback
4111 declaration has been changed from
4114 int (*cb)(X509_STORE_CTX *,void *);
4115 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4116 i=s->ctx->app_verify_callback(&ctx)
4117 has been changed into
4118 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4120 To update applications using SSL_CTX_set_cert_verify_callback(),
4121 a dummy argument can be added to their callback functions.
4122 [D. K. Smetters <smetters@parc.xerox.com>]
4124 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4125 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4127 *) Add and OPENSSL_LOAD_CONF define which will cause
4128 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4129 This allows older applications to transparently support certain
4130 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4131 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4132 load the config file and OPENSSL_add_all_algorithms_conf() which will
4133 always load it have also been added.
4136 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4137 Adjust NIDs and EVP layer.
4138 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4140 *) Config modules support in openssl utility.
4142 Most commands now load modules from the config file,
4143 though in a few (such as version) this isn't done
4144 because it couldn't be used for anything.
4146 In the case of ca and req the config file used is
4147 the same as the utility itself: that is the -config
4148 command line option can be used to specify an
4152 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4153 use "openssl_conf" if filename is NULL use default openssl config file.
4156 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4157 config section name. Add a new flag to tolerate a missing config file
4158 and move code to CONF_modules_load_file().
4161 *) Support for crypto accelerator cards from Accelerated Encryption
4162 Processing, www.aep.ie. (Use engine 'aep')
4163 The support was copied from 0.9.6c [engine] and adapted/corrected
4164 to work with the new engine framework.
4165 [AEP Inc. and Richard Levitte]
4167 *) Support for SureWare crypto accelerator cards from Baltimore
4168 Technologies. (Use engine 'sureware')
4169 The support was copied from 0.9.6c [engine] and adapted
4170 to work with the new engine framework.
4173 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4174 make the newer ENGINE framework commands for the CHIL engine work.
4175 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4177 *) Make it possible to produce shared libraries on ReliantUNIX.
4178 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4180 *) Add the configuration target debug-linux-ppro.
4181 Make 'openssl rsa' use the general key loading routines
4182 implemented in apps.c, and make those routines able to
4183 handle the key format FORMAT_NETSCAPE and the variant
4185 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4187 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4188 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4190 *) Add -keyform to rsautl, and document -engine.
4191 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4193 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4194 BIO_R_NO_SUCH_FILE error code rather than the generic
4195 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4198 *) Add new functions
4200 ERR_peek_last_error_line
4201 ERR_peek_last_error_line_data.
4202 These are similar to
4205 ERR_peek_error_line_data,
4206 but report on the latest error recorded rather than the first one
4207 still in the error queue.
4208 [Ben Laurie, Bodo Moeller]
4210 *) default_algorithms option in ENGINE config module. This allows things
4212 default_algorithms = ALL
4213 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4216 *) Prelminary ENGINE config module.
4219 *) New experimental application configuration code.
4222 *) Change the AES code to follow the same name structure as all other
4223 symmetric ciphers, and behave the same way. Move everything to
4224 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4225 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4227 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4228 [Ben Laurie and Theo de Raadt]
4230 *) Add option to output public keys in req command.
4231 [Massimiliano Pala madwolf@openca.org]
4233 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4234 (up to about 10% better than before for P-192 and P-224).
4237 *) New functions/macros
4239 SSL_CTX_set_msg_callback(ctx, cb)
4240 SSL_CTX_set_msg_callback_arg(ctx, arg)
4241 SSL_set_msg_callback(ssl, cb)
4242 SSL_set_msg_callback_arg(ssl, arg)
4244 to request calling a callback function
4246 void cb(int write_p, int version, int content_type,
4247 const void *buf, size_t len, SSL *ssl, void *arg)
4249 whenever a protocol message has been completely received
4250 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4251 protocol version according to which the SSL library interprets
4252 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4253 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4254 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4255 specification (change_cipher_spec(20), alert(21), handshake(22)).
4256 'buf' and 'len' point to the actual message, 'ssl' to the
4257 SSL object, and 'arg' is the application-defined value set by
4258 SSL[_CTX]_set_msg_callback_arg().
4260 'openssl s_client' and 'openssl s_server' have new '-msg' options
4261 to enable a callback that displays all protocol messages.
4264 *) Change the shared library support so shared libraries are built as
4265 soon as the corresponding static library is finished, and thereby get
4266 openssl and the test programs linked against the shared library.
4267 This still only happens when the keyword "shard" has been given to
4268 the configuration scripts.
4270 NOTE: shared library support is still an experimental thing, and
4271 backward binary compatibility is still not guaranteed.
4272 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4274 *) Add support for Subject Information Access extension.
4275 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4277 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4278 additional bytes when new memory had to be allocated, not just
4279 when reusing an existing buffer.
4282 *) New command line and configuration option 'utf8' for the req command.
4283 This allows field values to be specified as UTF8 strings.
4286 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4287 runs for the former and machine-readable output for the latter.
4290 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4291 of the e-mail address in the DN (i.e., it will go into a certificate
4292 extension only). The new configuration file option 'email_in_dn = no'
4293 has the same effect.
4294 [Massimiliano Pala madwolf@openca.org]
4296 *) Change all functions with names starting with des_ to be starting
4297 with DES_ instead. Add wrappers that are compatible with libdes,
4298 but are named _ossl_old_des_*. Finally, add macros that map the
4299 des_* symbols to the corresponding _ossl_old_des_* if libdes
4300 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4301 desired, the des_* symbols will be mapped to DES_*, with one
4304 Since we provide two compatibility mappings, the user needs to
4305 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4306 compatibility is desired. The default (i.e., when that macro
4307 isn't defined) is OpenSSL 0.9.6c compatibility.
4309 There are also macros that enable and disable the support of old
4310 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4311 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4312 are defined, the default will apply: to support the old des routines.
4314 In either case, one must include openssl/des.h to get the correct
4315 definitions. Do not try to just include openssl/des_old.h, that
4318 NOTE: This is a major break of an old API into a new one. Software
4319 authors are encouraged to switch to the DES_ style functions. Some
4320 time in the future, des_old.h and the libdes compatibility functions
4321 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4322 default), and then completely removed.
4325 *) Test for certificates which contain unsupported critical extensions.
4326 If such a certificate is found during a verify operation it is
4327 rejected by default: this behaviour can be overridden by either
4328 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4329 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4330 X509_supported_extension() has also been added which returns 1 if a
4331 particular extension is supported.
4334 *) Modify the behaviour of EVP cipher functions in similar way to digests
4335 to retain compatibility with existing code.
4338 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4339 compatibility with existing code. In particular the 'ctx' parameter does
4340 not have to be to be initialized before the call to EVP_DigestInit() and
4341 it is tidied up after a call to EVP_DigestFinal(). New function
4342 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4343 EVP_MD_CTX_copy() changed to not require the destination to be
4344 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4345 requires the destination to be valid.
4347 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4348 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4351 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4352 so that complete 'Handshake' protocol structures are kept in memory
4353 instead of overwriting 'msg_type' and 'length' with 'body' data.
4356 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4357 [Massimo Santin via Richard Levitte]
4359 *) Major restructuring to the underlying ENGINE code. This includes
4360 reduction of linker bloat, separation of pure "ENGINE" manipulation
4361 (initialisation, etc) from functionality dealing with implementations
4362 of specific crypto iterfaces. This change also introduces integrated
4363 support for symmetric ciphers and digest implementations - so ENGINEs
4364 can now accelerate these by providing EVP_CIPHER and EVP_MD
4365 implementations of their own. This is detailed in crypto/engine/README
4366 as it couldn't be adequately described here. However, there are a few
4367 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4368 were changed in the original introduction of ENGINE code have now
4369 reverted back - the hooking from this code to ENGINE is now a good
4370 deal more passive and at run-time, operations deal directly with
4371 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4372 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4373 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4374 they were not being used by the framework as there is no concept of a
4375 BIGNUM_METHOD and they could not be generalised to the new
4376 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4377 ENGINE_cpy() has been removed as it cannot be consistently defined in
4381 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4384 *) Change mkdef.pl to sort symbols that get the same entry number,
4385 and make sure the automatically generated functions ERR_load_*
4386 become part of libeay.num as well.
4389 *) New function SSL_renegotiate_pending(). This returns true once
4390 renegotiation has been requested (either SSL_renegotiate() call
4391 or HelloRequest/ClientHello receveived from the peer) and becomes
4392 false once a handshake has been completed.
4393 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4394 sends a HelloRequest, but does not ensure that a handshake takes
4395 place. SSL_renegotiate_pending() is useful for checking if the
4396 client has followed the request.)
4399 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4400 By default, clients may request session resumption even during
4401 renegotiation (if session ID contexts permit); with this option,
4402 session resumption is possible only in the first handshake.
4404 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4405 more bits available for options that should not be part of
4406 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4409 *) Add some demos for certificate and certificate request creation.
4412 *) Make maximum certificate chain size accepted from the peer application
4413 settable (SSL*_get/set_max_cert_list()), as proposed by
4414 "Douglas E. Engert" <deengert@anl.gov>.
4417 *) Add support for shared libraries for Unixware-7
4418 (Boyd Lynn Gerber <gerberb@zenez.com>).
4421 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4422 be done prior to destruction. Use this to unload error strings from
4423 ENGINEs that load their own error strings. NB: This adds two new API
4424 functions to "get" and "set" this destroy handler in an ENGINE.
4427 *) Alter all existing ENGINE implementations (except "openssl" and
4428 "openbsd") to dynamically instantiate their own error strings. This
4429 makes them more flexible to be built both as statically-linked ENGINEs
4430 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4431 Also, add stub code to each that makes building them as self-contained
4432 shared-libraries easier (see README.ENGINE).
4435 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4436 implementations into applications that are completely implemented in
4437 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4438 commands that can be used to configure what shared-library to load and
4439 to control aspects of the way it is handled. Also, made an update to
4440 the README.ENGINE file that brings its information up-to-date and
4441 provides some information and instructions on the "dynamic" ENGINE
4442 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4445 *) Make it possible to unload ranges of ERR strings with a new
4446 "ERR_unload_strings" function.
4449 *) Add a copy() function to EVP_MD.
4452 *) Make EVP_MD routines take a context pointer instead of just the
4453 md_data void pointer.
4456 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4457 that the digest can only process a single chunk of data
4458 (typically because it is provided by a piece of
4459 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4460 is only going to provide a single chunk of data, and hence the
4461 framework needn't accumulate the data for oneshot drivers.
4464 *) As with "ERR", make it possible to replace the underlying "ex_data"
4465 functions. This change also alters the storage and management of global
4466 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4467 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4468 index counters. The API functions that use this state have been changed
4469 to take a "class_index" rather than pointers to the class's local STACK
4470 and counter, and there is now an API function to dynamically create new
4471 classes. This centralisation allows us to (a) plug a lot of the
4472 thread-safety problems that existed, and (b) makes it possible to clean
4473 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4474 such data would previously have always leaked in application code and
4475 workarounds were in place to make the memory debugging turn a blind eye
4476 to it. Application code that doesn't use this new function will still
4477 leak as before, but their memory debugging output will announce it now
4478 rather than letting it slide.
4480 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4481 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4482 has a return value to indicate success or failure.
4485 *) Make it possible to replace the underlying "ERR" functions such that the
4486 global state (2 LHASH tables and 2 locks) is only used by the "default"
4487 implementation. This change also adds two functions to "get" and "set"
4488 the implementation prior to it being automatically set the first time
4489 any other ERR function takes place. Ie. an application can call "get",
4490 pass the return value to a module it has just loaded, and that module
4491 can call its own "set" function using that value. This means the
4492 module's "ERR" operations will use (and modify) the error state in the
4493 application and not in its own statically linked copy of OpenSSL code.
4496 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4497 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4498 the operation, and provides a more encapsulated way for external code
4499 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4500 to use these functions rather than manually incrementing the counts.
4502 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4505 *) Add EVP test program.
4508 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4511 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4512 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4513 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4514 These allow a CRL to be built without having to access X509_CRL fields
4515 directly. Modify 'ca' application to use new functions.
4518 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4519 bug workarounds. Rollback attack detection is a security feature.
4520 The problem will only arise on OpenSSL servers when TLSv1 is not
4521 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4522 Software authors not wanting to support TLSv1 will have special reasons
4523 for their choice and can explicitly enable this option.
4524 [Bodo Moeller, Lutz Jaenicke]
4526 *) Rationalise EVP so it can be extended: don't include a union of
4527 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4528 (similar to those existing for EVP_CIPHER_CTX).
4533 EVP_MD_CTX_init(&md); /* new function call */
4534 EVP_DigestInit(&md, EVP_sha1());
4535 EVP_DigestUpdate(&md, in, len);
4536 EVP_DigestFinal(&md, out, NULL);
4537 EVP_MD_CTX_cleanup(&md); /* new function call */
4541 *) Make DES key schedule conform to the usual scheme, as well as
4542 correcting its structure. This means that calls to DES functions
4543 now have to pass a pointer to a des_key_schedule instead of a
4544 plain des_key_schedule (which was actually always a pointer
4547 des_key_schedule ks;
4549 des_set_key_checked(..., &ks);
4550 des_ncbc_encrypt(..., &ks, ...);
4552 (Note that a later change renames 'des_...' into 'DES_...'.)
4555 *) Initial reduction of linker bloat: the use of some functions, such as
4556 PEM causes large amounts of unused functions to be linked in due to
4557 poor organisation. For example pem_all.c contains every PEM function
4558 which has a knock on effect of linking in large amounts of (unused)
4559 ASN1 code. Grouping together similar functions and splitting unrelated
4560 functions prevents this.
4563 *) Cleanup of EVP macros.
4566 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4567 correct _ecb suffix.
4570 *) Add initial OCSP responder support to ocsp application. The
4571 revocation information is handled using the text based index
4572 use by the ca application. The responder can either handle
4573 requests generated internally, supplied in files (for example
4574 via a CGI script) or using an internal minimal server.
4577 *) Add configuration choices to get zlib compression for TLS.
4580 *) Changes to Kerberos SSL for RFC 2712 compliance:
4581 1. Implemented real KerberosWrapper, instead of just using
4582 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4583 2. Implemented optional authenticator field of KerberosWrapper.
4585 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4586 and authenticator structs; see crypto/krb5/.
4588 Generalized Kerberos calls to support multiple Kerberos libraries.
4589 [Vern Staats <staatsvr@asc.hpc.mil>,
4590 Jeffrey Altman <jaltman@columbia.edu>
4591 via Richard Levitte]
4593 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4594 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4595 values for each of the key sizes rather than having just
4596 parameters (and 'speed' generating keys each time).
4599 *) Speed up EVP routines.
4602 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4603 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4604 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4605 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4607 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4608 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4609 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4612 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4614 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4617 *) Added the OS2-EMX target.
4618 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4620 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4621 to support NCONF routines in extension code. New function CONF_set_nconf()
4622 to allow functions which take an NCONF to also handle the old LHASH
4623 structure: this means that the old CONF compatible routines can be
4624 retained (in particular wrt extensions) without having to duplicate the
4625 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4628 *) Enhance the general user interface with mechanisms for inner control
4629 and with possibilities to have yes/no kind of prompts.
4632 *) Change all calls to low level digest routines in the library and
4633 applications to use EVP. Add missing calls to HMAC_cleanup() and
4634 don't assume HMAC_CTX can be copied using memcpy().
4635 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4637 *) Add the possibility to control engines through control names but with
4638 arbitrary arguments instead of just a string.
4639 Change the key loaders to take a UI_METHOD instead of a callback
4640 function pointer. NOTE: this breaks binary compatibility with earlier
4641 versions of OpenSSL [engine].
4642 Adapt the nCipher code for these new conditions and add a card insertion
4646 *) Enhance the general user interface with mechanisms to better support
4647 dialog box interfaces, application-defined prompts, the possibility
4648 to use defaults (for example default passwords from somewhere else)
4649 and interrupts/cancellations.
4652 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4653 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4656 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4657 tidy up some unnecessarily weird code in 'sk_new()').
4658 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4660 *) Change the key loading routines for ENGINEs to use the same kind
4661 callback (pem_password_cb) as all other routines that need this
4665 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4666 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4667 than this minimum value is recommended.
4670 *) New random seeder for OpenVMS, using the system process statistics
4671 that are easily reachable.
4674 *) Windows apparently can't transparently handle global
4675 variables defined in DLLs. Initialisations such as:
4677 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4679 wont compile. This is used by the any applications that need to
4680 declare their own ASN1 modules. This was fixed by adding the option
4681 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4682 needed for static libraries under Win32.
4685 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4686 setting of purpose and trust fields. New X509_STORE trust and
4687 purpose functions and tidy up setting in other SSL functions.
4690 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4691 structure. These are inherited by X509_STORE_CTX when it is
4692 initialised. This allows various defaults to be set in the
4693 X509_STORE structure (such as flags for CRL checking and custom
4694 purpose or trust settings) for functions which only use X509_STORE_CTX
4695 internally such as S/MIME.
4697 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4698 trust settings if they are not set in X509_STORE. This allows X509_STORE
4699 purposes and trust (in S/MIME for example) to override any set by default.
4701 Add command line options for CRL checking to smime, s_client and s_server
4705 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4706 are set then the CRL is looked up in the X509_STORE structure and
4707 its validity and signature checked, then if the certificate is found
4708 in the CRL the verify fails with a revoked error.
4710 Various new CRL related callbacks added to X509_STORE_CTX structure.
4712 Command line options added to 'verify' application to support this.
4714 This needs some additional work, such as being able to handle multiple
4715 CRLs with different times, extension based lookup (rather than just
4716 by subject name) and ultimately more complete V2 CRL extension
4720 *) Add a general user interface API (crypto/ui/). This is designed
4721 to replace things like des_read_password and friends (backward
4722 compatibility functions using this new API are provided).
4723 The purpose is to remove prompting functions from the DES code
4724 section as well as provide for prompting through dialog boxes in
4725 a window system and the like.
4728 *) Add "ex_data" support to ENGINE so implementations can add state at a
4729 per-structure level rather than having to store it globally.
4732 *) Make it possible for ENGINE structures to be copied when retrieved by
4733 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4734 This causes the "original" ENGINE structure to act like a template,
4735 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4736 operational state can be localised to each ENGINE structure, despite the
4737 fact they all share the same "methods". New ENGINE structures returned in
4738 this case have no functional references and the return value is the single
4739 structural reference. This matches the single structural reference returned
4740 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4744 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4745 needs to match any other type at all we need to manually clear the
4749 *) Changes to the "openssl engine" utility to include;
4750 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4751 about an ENGINE's available control commands.
4752 - executing control commands from command line arguments using the
4753 '-pre' and '-post' switches. '-post' is only used if '-t' is
4754 specified and the ENGINE is successfully initialised. The syntax for
4755 the individual commands are colon-separated, for example;
4756 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4759 *) New dynamic control command support for ENGINEs. ENGINEs can now
4760 declare their own commands (numbers), names (strings), descriptions,
4761 and input types for run-time discovery by calling applications. A
4762 subset of these commands are implicitly classed as "executable"
4763 depending on their input type, and only these can be invoked through
4764 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4765 can be based on user input, config files, etc). The distinction is
4766 that "executable" commands cannot return anything other than a boolean
4767 result and can only support numeric or string input, whereas some
4768 discoverable commands may only be for direct use through
4769 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4770 pointers, or other custom uses. The "executable" commands are to
4771 support parameterisations of ENGINE behaviour that can be
4772 unambiguously defined by ENGINEs and used consistently across any
4773 OpenSSL-based application. Commands have been added to all the
4774 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4775 control over shared-library paths without source code alterations.
4778 *) Changed all ENGINE implementations to dynamically allocate their
4779 ENGINEs rather than declaring them statically. Apart from this being
4780 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4781 this also allows the implementations to compile without using the
4782 internal engine_int.h header.
4785 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4786 'const' value. Any code that should be able to modify a RAND_METHOD
4787 should already have non-const pointers to it (ie. they should only
4788 modify their own ones).
4791 *) Made a variety of little tweaks to the ENGINE code.
4792 - "atalla" and "ubsec" string definitions were moved from header files
4793 to C code. "nuron" string definitions were placed in variables
4794 rather than hard-coded - allowing parameterisation of these values
4795 later on via ctrl() commands.
4796 - Removed unused "#if 0"'d code.
4797 - Fixed engine list iteration code so it uses ENGINE_free() to release
4798 structural references.
4799 - Constified the RAND_METHOD element of ENGINE structures.
4800 - Constified various get/set functions as appropriate and added
4801 missing functions (including a catch-all ENGINE_cpy that duplicates
4802 all ENGINE values onto a new ENGINE except reference counts/state).
4803 - Removed NULL parameter checks in get/set functions. Setting a method
4804 or function to NULL is a way of cancelling out a previously set
4805 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4806 and doesn't justify the extra error symbols and code.
4807 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4808 flags from engine_int.h to engine.h.
4809 - Changed prototypes for ENGINE handler functions (init(), finish(),
4810 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4813 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4814 to the algorithm using long division. The binary algorithm can be
4815 used only if the modulus is odd. On 32-bit systems, it is faster
4816 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4817 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4818 up to 450 bits. In 64-bit environments, the binary algorithm
4819 appears to be advantageous for much longer moduli; here we use it
4820 for moduli up to 2048 bits.
4823 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4824 could not support the combine flag in choice fields.
4827 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4828 extensions from a certificate request to the certificate.
4831 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4832 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4833 file: this allows the display of the certificate about to be
4834 signed to be customised, to allow certain fields to be included
4835 or excluded and extension details. The old system didn't display
4836 multicharacter strings properly, omitted fields not in the policy
4837 and couldn't display additional details such as extensions.
4840 *) Function EC_POINTs_mul for multiple scalar multiplication
4841 of an arbitrary number of elliptic curve points
4842 \sum scalars[i]*points[i],
4843 optionally including the generator defined for the EC_GROUP:
4844 scalar*generator + \sum scalars[i]*points[i].
4846 EC_POINT_mul is a simple wrapper function for the typical case
4847 that the point list has just one item (besides the optional
4851 *) First EC_METHODs for curves over GF(p):
4853 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4854 operations and provides various method functions that can also
4855 operate with faster implementations of modular arithmetic.
4857 EC_GFp_mont_method() reuses most functions that are part of
4858 EC_GFp_simple_method, but uses Montgomery arithmetic.
4860 [Bodo Moeller; point addition and point doubling
4861 implementation directly derived from source code provided by
4862 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4864 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4865 crypto/ec/ec_lib.c):
4867 Curves are EC_GROUP objects (with an optional group generator)
4868 based on EC_METHODs that are built into the library.
4870 Points are EC_POINT objects based on EC_GROUP objects.
4872 Most of the framework would be able to handle curves over arbitrary
4873 finite fields, but as there are no obvious types for fields other
4874 than GF(p), some functions are limited to that for now.
4877 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4878 that the file contains a complete HTTP response.
4881 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4882 change the def and num file printf format specifier from "%-40sXXX"
4883 to "%-39s XXX". The latter will always guarantee a space after the
4884 field while the former will cause them to run together if the field
4885 is 40 of more characters long.
4888 *) Constify the cipher and digest 'method' functions and structures
4889 and modify related functions to take constant EVP_MD and EVP_CIPHER
4893 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4894 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4897 *) Modify EVP_Digest*() routines so they now return values. Although the
4898 internal software routines can never fail additional hardware versions
4902 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4904 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4905 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4908 ERR_R_NESTED_ASN1_ERROR
4910 ERR_R_MISSING_ASN1_EOS
4911 were 4 .. 9, conflicting with
4912 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4914 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4915 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4917 Add new error code 'ERR_R_INTERNAL_ERROR'.
4920 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4924 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4925 sets the subject name for a new request or supersedes the
4926 subject name in a given request. Formats that can be parsed are
4927 'CN=Some Name, OU=myOU, C=IT'
4929 'CN=Some Name/OU=myOU/C=IT'.
4931 Add options '-batch' and '-verbose' to 'openssl req'.
4932 [Massimiliano Pala <madwolf@hackmasters.net>]
4934 *) Introduce the possibility to access global variables through
4935 functions on platform were that's the best way to handle exporting
4936 global variables in shared libraries. To enable this functionality,
4937 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4938 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4939 is normally done by Configure or something similar).
4941 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4942 in the source file (foo.c) like this:
4944 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4945 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4947 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4948 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4950 OPENSSL_DECLARE_GLOBAL(int,foo);
4951 #define foo OPENSSL_GLOBAL_REF(foo)
4952 OPENSSL_DECLARE_GLOBAL(double,bar);
4953 #define bar OPENSSL_GLOBAL_REF(bar)
4955 The #defines are very important, and therefore so is including the
4956 header file everywhere where the defined globals are used.
4958 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4959 of ASN.1 items, but that structure is a bit different.
4961 The largest change is in util/mkdef.pl which has been enhanced with
4962 better and easier to understand logic to choose which symbols should
4963 go into the Windows .def files as well as a number of fixes and code
4964 cleanup (among others, algorithm keywords are now sorted
4965 lexicographically to avoid constant rewrites).
4968 *) In BN_div() keep a copy of the sign of 'num' before writing the
4969 result to 'rm' because if rm==num the value will be overwritten
4970 and produce the wrong result if 'num' is negative: this caused
4971 problems with BN_mod() and BN_nnmod().
4974 *) Function OCSP_request_verify(). This checks the signature on an
4975 OCSP request and verifies the signer certificate. The signer
4976 certificate is just checked for a generic purpose and OCSP request
4980 *) Add OCSP_check_validity() function to check the validity of OCSP
4981 responses. OCSP responses are prepared in real time and may only
4982 be a few seconds old. Simply checking that the current time lies
4983 between thisUpdate and nextUpdate max reject otherwise valid responses
4984 caused by either OCSP responder or client clock inaccuracy. Instead
4985 we allow thisUpdate and nextUpdate to fall within a certain period of
4986 the current time. The age of the response can also optionally be
4987 checked. Two new options -validity_period and -status_age added to
4991 *) If signature or public key algorithm is unrecognized print out its
4992 OID rather that just UNKNOWN.
4995 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4996 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4997 ID to be generated from the issuer certificate alone which can then be
4998 passed to OCSP_id_issuer_cmp().
5001 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5002 ASN1 modules to export functions returning ASN1_ITEM pointers
5003 instead of the ASN1_ITEM structures themselves. This adds several
5004 new macros which allow the underlying ASN1 function/structure to
5005 be accessed transparently. As a result code should not use ASN1_ITEM
5006 references directly (such as &X509_it) but instead use the relevant
5007 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5008 use of the new ASN1 code on platforms where exporting structures
5009 is problematical (for example in shared libraries) but exporting
5010 functions returning pointers to structures is not.
5013 *) Add support for overriding the generation of SSL/TLS session IDs.
5014 These callbacks can be registered either in an SSL_CTX or per SSL.
5015 The purpose of this is to allow applications to control, if they wish,
5016 the arbitrary values chosen for use as session IDs, particularly as it
5017 can be useful for session caching in multiple-server environments. A
5018 command-line switch for testing this (and any client code that wishes
5019 to use such a feature) has been added to "s_server".
5020 [Geoff Thorpe, Lutz Jaenicke]
5022 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5023 of the form '#if defined(...) || defined(...) || ...' and
5024 '#if !defined(...) && !defined(...) && ...'. This also avoids
5025 the growing number of special cases it was previously handling.
5028 *) Make all configuration macros available for application by making
5029 sure they are available in opensslconf.h, by giving them names starting
5030 with "OPENSSL_" to avoid conflicts with other packages and by making
5031 sure e_os2.h will cover all platform-specific cases together with
5033 Additionally, it is now possible to define configuration/platform-
5034 specific names (called "system identities"). In the C code, these
5035 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5036 macro with the name beginning with "OPENSSL_SYS_", which is determined
5037 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5041 *) New option -set_serial to 'req' and 'x509' this allows the serial
5042 number to use to be specified on the command line. Previously self
5043 signed certificates were hard coded with serial number 0 and the
5044 CA options of 'x509' had to use a serial number in a file which was
5048 *) New options to 'ca' utility to support V2 CRL entry extensions.
5049 Currently CRL reason, invalidity date and hold instruction are
5050 supported. Add new CRL extensions to V3 code and some new objects.
5053 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5054 disable standard block padding (aka PKCS#5 padding) in the EVP
5055 API, which was previously mandatory. This means that the data is
5056 not padded in any way and so the total length much be a multiple
5057 of the block size, otherwise an error occurs.
5060 *) Initial (incomplete) OCSP SSL support.
5063 *) New function OCSP_parse_url(). This splits up a URL into its host,
5064 port and path components: primarily to parse OCSP URLs. New -url
5065 option to ocsp utility.
5068 *) New nonce behavior. The return value of OCSP_check_nonce() now
5069 reflects the various checks performed. Applications can decide
5070 whether to tolerate certain situations such as an absent nonce
5071 in a response when one was present in a request: the ocsp application
5072 just prints out a warning. New function OCSP_add1_basic_nonce()
5073 this is to allow responders to include a nonce in a response even if
5074 the request is nonce-less.
5077 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5078 skipped when using openssl x509 multiple times on a single input file,
5079 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5082 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5083 set string type: to handle setting ASN1_TIME structures. Fix ca
5084 utility to correctly initialize revocation date of CRLs.
5087 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5088 the clients preferred ciphersuites and rather use its own preferences.
5089 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5090 Internet Explorer by ensuring unchanged hash method during stepup.
5091 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5094 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5095 to aes and add a new 'exist' option to print out symbols that don't
5099 *) Additional options to ocsp utility to allow flags to be set and
5100 additional certificates supplied.
5103 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5104 OCSP client a number of certificate to only verify the response
5108 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5109 handle the new API. Currently only ECB, CBC modes supported. Add new
5112 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5113 Encryption Standard (AES) Ciphersuites for Transport Layer
5114 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5115 not enabled by default and were not part of the "ALL" ciphersuite
5116 alias because they were not yet official; they could be
5117 explicitly requested by specifying the "AESdraft" ciphersuite
5118 group alias. In the final release of OpenSSL 0.9.7, the group
5119 alias is called "AES" and is part of "ALL".)
5120 [Ben Laurie, Steve Henson, Bodo Moeller]
5122 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5123 request to response.
5126 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5127 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5128 extract information from a certificate request. OCSP_response_create()
5129 creates a response and optionally adds a basic response structure.
5130 OCSP_basic_add1_status() adds a complete single response to a basic
5131 response and returns the OCSP_SINGLERESP structure just added (to allow
5132 extensions to be included for example). OCSP_basic_add1_cert() adds a
5133 certificate to a basic response and OCSP_basic_sign() signs a basic
5134 response with various flags. New helper functions ASN1_TIME_check()
5135 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5136 (converts ASN1_TIME to GeneralizedTime).
5139 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5140 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5141 structure from a certificate. X509_pubkey_digest() digests the public_key
5142 contents: this is used in various key identifiers.
5145 *) Make sk_sort() tolerate a NULL argument.
5146 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5148 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5149 passed by the function are trusted implicitly. If any of them signed the
5150 response then it is assumed to be valid and is not verified.
5153 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5154 to data. This was previously part of the PKCS7 ASN1 code. This
5155 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5156 [Steve Henson, reported by Kenneth R. Robinette
5157 <support@securenetterm.com>]
5159 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5160 routines: without these tracing memory leaks is very painful.
5161 Fix leaks in PKCS12 and PKCS7 routines.
5164 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5165 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5166 effectively meant GeneralizedTime would never be used. Now it
5167 is initialised to -1 but X509_time_adj() now has to check the value
5168 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5169 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5170 [Steve Henson, reported by Kenneth R. Robinette
5171 <support@securenetterm.com>]
5173 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5174 result in a zero length in the ASN1_INTEGER structure which was
5175 not consistent with the structure when d2i_ASN1_INTEGER() was used
5176 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5177 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5178 where it did not print out a minus for negative ASN1_INTEGER.
5181 *) Add summary printout to ocsp utility. The various functions which
5182 convert status values to strings have been renamed to:
5183 OCSP_response_status_str(), OCSP_cert_status_str() and
5184 OCSP_crl_reason_str() and are no longer static. New options
5185 to verify nonce values and to disable verification. OCSP response
5186 printout format cleaned up.
5189 *) Add additional OCSP certificate checks. These are those specified
5190 in RFC2560. This consists of two separate checks: the CA of the
5191 certificate being checked must either be the OCSP signer certificate
5192 or the issuer of the OCSP signer certificate. In the latter case the
5193 OCSP signer certificate must contain the OCSP signing extended key
5194 usage. This check is performed by attempting to match the OCSP
5195 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5196 in the OCSP_CERTID structures of the response.
5199 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5200 and related routines. This uses the standard OpenSSL certificate
5201 verify routines to perform initial checks (just CA validity) and
5202 to obtain the certificate chain. Then additional checks will be
5203 performed on the chain. Currently the root CA is checked to see
5204 if it is explicitly trusted for OCSP signing. This is used to set
5205 a root CA as a global signing root: that is any certificate that
5206 chains to that CA is an acceptable OCSP signing certificate.
5209 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5210 extensions from a separate configuration file.
5211 As when reading extensions from the main configuration file,
5212 the '-extensions ...' option may be used for specifying the
5214 [Massimiliano Pala <madwolf@comune.modena.it>]
5216 *) New OCSP utility. Allows OCSP requests to be generated or
5217 read. The request can be sent to a responder and the output
5218 parsed, outputed or printed in text form. Not complete yet:
5219 still needs to check the OCSP response validity.
5222 *) New subcommands for 'openssl ca':
5223 'openssl ca -status <serial>' prints the status of the cert with
5224 the given serial number (according to the index file).
5225 'openssl ca -updatedb' updates the expiry status of certificates
5227 [Massimiliano Pala <madwolf@comune.modena.it>]
5229 *) New '-newreq-nodes' command option to CA.pl. This is like
5230 '-newreq', but calls 'openssl req' with the '-nodes' option
5231 so that the resulting key is not encrypted.
5232 [Damien Miller <djm@mindrot.org>]
5234 *) New configuration for the GNU Hurd.
5235 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5237 *) Initial code to implement OCSP basic response verify. This
5238 is currently incomplete. Currently just finds the signer's
5239 certificate and verifies the signature on the response.
5242 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5243 value of OPENSSLDIR. This is available via the new '-d' option
5244 to 'openssl version', and is also included in 'openssl version -a'.
5247 *) Allowing defining memory allocation callbacks that will be given
5248 file name and line number information in additional arguments
5249 (a const char* and an int). The basic functionality remains, as
5250 well as the original possibility to just replace malloc(),
5251 realloc() and free() by functions that do not know about these
5252 additional arguments. To register and find out the current
5253 settings for extended allocation functions, the following
5254 functions are provided:
5256 CRYPTO_set_mem_ex_functions
5257 CRYPTO_set_locked_mem_ex_functions
5258 CRYPTO_get_mem_ex_functions
5259 CRYPTO_get_locked_mem_ex_functions
5261 These work the same way as CRYPTO_set_mem_functions and friends.
5262 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5263 extended allocation function is enabled.
5264 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5265 a conventional allocation function is enabled.
5266 [Richard Levitte, Bodo Moeller]
5268 *) Finish off removing the remaining LHASH function pointer casts.
5269 There should no longer be any prototype-casting required when using
5270 the LHASH abstraction, and any casts that remain are "bugs". See
5271 the callback types and macros at the head of lhash.h for details
5272 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5275 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5276 If /dev/[u]random devices are not available or do not return enough
5277 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5279 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5280 /etc/entropy will be queried once each in this sequence, quering stops
5281 when enough entropy was collected without querying more sockets.
5284 *) Change the Unix RAND_poll() variant to be able to poll several
5285 random devices, as specified by DEVRANDOM, until a sufficient amount
5286 of data has been collected. We spend at most 10 ms on each file
5287 (select timeout) and read in non-blocking mode. DEVRANDOM now
5288 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5289 (previously it was just the string "/dev/urandom"), so on typical
5290 platforms the 10 ms delay will never occur.
5291 Also separate out the Unix variant to its own file, rand_unix.c.
5292 For VMS, there's a currently-empty rand_vms.c.
5295 *) Move OCSP client related routines to ocsp_cl.c. These
5296 provide utility functions which an application needing
5297 to issue a request to an OCSP responder and analyse the
5298 response will typically need: as opposed to those which an
5299 OCSP responder itself would need which will be added later.
5301 OCSP_request_sign() signs an OCSP request with an API similar
5302 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5303 response. OCSP_response_get1_basic() extracts basic response
5304 from response. OCSP_resp_find_status(): finds and extracts status
5305 information from an OCSP_CERTID structure (which will be created
5306 when the request structure is built). These are built from lower
5307 level functions which work on OCSP_SINGLERESP structures but
5308 wont normally be used unless the application wishes to examine
5309 extensions in the OCSP response for example.
5311 Replace nonce routines with a pair of functions.
5312 OCSP_request_add1_nonce() adds a nonce value and optionally
5313 generates a random value. OCSP_check_nonce() checks the
5314 validity of the nonce in an OCSP response.
5317 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5318 This doesn't copy the supplied OCSP_CERTID and avoids the
5319 need to free up the newly created id. Change return type
5320 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5321 This can then be used to add extensions to the request.
5322 Deleted OCSP_request_new(), since most of its functionality
5323 is now in OCSP_REQUEST_new() (and the case insensitive name
5324 clash) apart from the ability to set the request name which
5325 will be added elsewhere.
5328 *) Update OCSP API. Remove obsolete extensions argument from
5329 various functions. Extensions are now handled using the new
5330 OCSP extension code. New simple OCSP HTTP function which
5331 can be used to send requests and parse the response.
5334 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5335 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5336 uses the special reorder version of SET OF to sort the attributes
5337 and reorder them to match the encoded order. This resolves a long
5338 standing problem: a verify on a PKCS7 structure just after signing
5339 it used to fail because the attribute order did not match the
5340 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5341 it uses the received order. This is necessary to tolerate some broken
5342 software that does not order SET OF. This is handled by encoding
5343 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5344 to produce the required SET OF.
5347 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5348 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5349 files to get correct declarations of the ASN.1 item variables.
5352 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5353 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5354 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5355 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5356 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5357 ASN1_ITEM and no wrapper functions.
5360 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5361 replace the old function pointer based I/O routines. Change most of
5362 the *_d2i_bio() and *_d2i_fp() functions to use these.
5365 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5366 lines, recognice more "algorithms" that can be deselected, and make
5367 it complain about algorithm deselection that isn't recognised.
5370 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5371 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5372 to use new functions. Add NO_ASN1_OLD which can be set to remove
5373 some old style ASN1 functions: this can be used to determine if old
5374 code will still work when these eventually go away.
5377 *) New extension functions for OCSP structures, these follow the
5378 same conventions as certificates and CRLs.
5381 *) New function X509V3_add1_i2d(). This automatically encodes and
5382 adds an extension. Its behaviour can be customised with various
5383 flags to append, replace or delete. Various wrappers added for
5384 certifcates and CRLs.
5387 *) Fix to avoid calling the underlying ASN1 print routine when
5388 an extension cannot be parsed. Correct a typo in the
5389 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5392 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5393 entries for variables.
5396 *) Add functionality to apps/openssl.c for detecting locking
5397 problems: As the program is single-threaded, all we have
5398 to do is register a locking callback using an array for
5399 storing which locks are currently held by the program.
5402 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5403 SSL_get_ex_data_X509_STORE_idx(), which is used in
5404 ssl_verify_cert_chain() and thus can be called at any time
5405 during TLS/SSL handshakes so that thread-safety is essential.
5406 Unfortunately, the ex_data design is not at all suited
5407 for multi-threaded use, so it probably should be abolished.
5410 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5411 [Broadcom, tweaked and integrated by Geoff Thorpe]
5413 *) Move common extension printing code to new function
5414 X509V3_print_extensions(). Reorganise OCSP print routines and
5415 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5418 *) New function X509_signature_print() to remove duplication in some
5422 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5423 set (this was treated exactly the same as SET OF previously). This
5424 is used to reorder the STACK representing the structure to match the
5425 encoding. This will be used to get round a problem where a PKCS7
5426 structure which was signed could not be verified because the STACK
5427 order did not reflect the encoded order.
5430 *) Reimplement the OCSP ASN1 module using the new code.
5433 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5434 for its ASN1 operations. The old style function pointers still exist
5435 for now but they will eventually go away.
5438 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5439 completely replaces the old ASN1 functionality with a table driven
5440 encoder and decoder which interprets an ASN1_ITEM structure describing
5441 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5442 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5443 has also been converted to the new form.
5446 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5447 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5448 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5449 for negative moduli.
5452 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5453 of not touching the result's sign bit.
5456 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5460 *) Changed the LHASH code to use prototypes for callbacks, and created
5461 macros to declare and implement thin (optionally static) functions
5462 that provide type-safety and avoid function pointer casting for the
5463 type-specific callbacks.
5466 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5468 [Veers Staats <staatsvr@asc.hpc.mil>,
5469 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5471 *) Reformat the FAQ so the different questions and answers can be divided
5472 in sections depending on the subject.
5475 *) Have the zlib compression code load ZLIB.DLL dynamically under
5479 *) New function BN_mod_sqrt for computing square roots modulo a prime
5480 (using the probabilistic Tonelli-Shanks algorithm unless
5481 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5482 be handled deterministically).
5483 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5485 *) Make BN_mod_inverse faster by explicitly handling small quotients
5486 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5487 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5490 *) New function BN_kronecker.
5493 *) Fix BN_gcd so that it works on negative inputs; the result is
5494 positive unless both parameters are zero.
5495 Previously something reasonably close to an infinite loop was
5496 possible because numbers could be growing instead of shrinking
5497 in the implementation of Euclid's algorithm.
5500 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5501 sign of the number in question.
5503 Fix BN_is_word(a,w) to work correctly for w == 0.
5505 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5506 because its test if the absolute value of 'a' equals 'w'.
5507 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5508 it exists mostly for use in the implementations of BN_is_zero(),
5509 BN_is_one(), and BN_is_word().
5512 *) New function BN_swap.
5515 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5516 the exponentiation functions are more likely to produce reasonable
5517 results on negative inputs.
5520 *) Change BN_mod_mul so that the result is always non-negative.
5521 Previously, it could be negative if one of the factors was negative;
5522 I don't think anyone really wanted that behaviour.
5525 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5526 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5527 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5528 and add new functions:
5537 BN_mod_lshift1_quick
5541 These functions always generate non-negative results.
5543 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5544 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5546 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5547 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5548 be reduced modulo m.
5549 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5552 The following entry accidentily appeared in the CHANGES file
5553 distributed with OpenSSL 0.9.7. The modifications described in
5554 it do *not* apply to OpenSSL 0.9.7.
5556 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5557 was actually never needed) and in BN_mul(). The removal in BN_mul()
5558 required a small change in bn_mul_part_recursive() and the addition
5559 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5560 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5561 bn_sub_words() and bn_add_words() except they take arrays with
5566 *) In 'openssl passwd', verify passwords read from the terminal
5567 unless the '-salt' option is used (which usually means that
5568 verification would just waste user's time since the resulting
5569 hash is going to be compared with some given password hash)
5570 or the new '-noverify' option is used.
5572 This is an incompatible change, but it does not affect
5573 non-interactive use of 'openssl passwd' (passwords on the command
5574 line, '-stdin' option, '-in ...' option) and thus should not
5578 *) Remove all references to RSAref, since there's no more need for it.
5581 *) Make DSO load along a path given through an environment variable
5582 (SHLIB_PATH) with shl_load().
5585 *) Constify the ENGINE code as a result of BIGNUM constification.
5586 Also constify the RSA code and most things related to it. In a
5587 few places, most notable in the depth of the ASN.1 code, ugly
5588 casts back to non-const were required (to be solved at a later
5592 *) Make it so the openssl application has all engines loaded by default.
5595 *) Constify the BIGNUM routines a little more.
5598 *) Add the following functions:
5600 ENGINE_load_cswift()
5602 ENGINE_load_atalla()
5604 ENGINE_load_builtin_engines()
5606 That way, an application can itself choose if external engines that
5607 are built-in in OpenSSL shall ever be used or not. The benefit is
5608 that applications won't have to be linked with libdl or other dso
5609 libraries unless it's really needed.
5611 Changed 'openssl engine' to load all engines on demand.
5612 Changed the engine header files to avoid the duplication of some
5613 declarations (they differed!).
5616 *) 'openssl engine' can now list capabilities.
5619 *) Better error reporting in 'openssl engine'.
5622 *) Never call load_dh_param(NULL) in s_server.
5625 *) Add engine application. It can currently list engines by name and
5626 identity, and test if they are actually available.
5629 *) Improve RPM specification file by forcing symbolic linking and making
5630 sure the installed documentation is also owned by root.root.
5631 [Damien Miller <djm@mindrot.org>]
5633 *) Give the OpenSSL applications more possibilities to make use of
5634 keys (public as well as private) handled by engines.
5637 *) Add OCSP code that comes from CertCo.
5640 *) Add VMS support for the Rijndael code.
5643 *) Added untested support for Nuron crypto accelerator.
5646 *) Add support for external cryptographic devices. This code was
5647 previously distributed separately as the "engine" branch.
5648 [Geoff Thorpe, Richard Levitte]
5650 *) Rework the filename-translation in the DSO code. It is now possible to
5651 have far greater control over how a "name" is turned into a filename
5652 depending on the operating environment and any oddities about the
5653 different shared library filenames on each system.
5656 *) Support threads on FreeBSD-elf in Configure.
5659 *) Fix for SHA1 assembly problem with MASM: it produces
5660 warnings about corrupt line number information when assembling
5661 with debugging information. This is caused by the overlapping
5663 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5666 NCONF_get_number() has no error checking at all. As a replacement,
5667 NCONF_get_number_e() is defined (_e for "error checking") and is
5668 promoted strongly. The old NCONF_get_number is kept around for
5669 binary backward compatibility.
5670 Make it possible for methods to load from something other than a BIO,
5671 by providing a function pointer that is given a name instead of a BIO.
5672 For example, this could be used to load configuration data from an
5676 *) Fix for non blocking accept BIOs. Added new I/O special reason
5677 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5678 with non blocking I/O was not possible because no retry code was
5679 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5683 *) Added the beginnings of Rijndael support.
5686 *) Fix for bug in DirectoryString mask setting. Add support for
5687 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5688 to allow certificate printing to more controllable, additional
5689 'certopt' option to 'x509' to allow new printing options to be
5693 *) Clean old EAY MD5 hack from e_os.h.
5696 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5698 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5699 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5700 [Joe Orton, Steve Henson]
5702 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5704 *) Fix additional bug revealed by the NISCC test suite:
5706 Stop bug triggering large recursion when presented with
5707 certain ASN.1 tags (CVE-2003-0851)
5710 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5712 *) Fix various bugs revealed by running the NISCC test suite:
5714 Stop out of bounds reads in the ASN1 code when presented with
5715 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5717 If verify callback ignores invalid public key errors don't try to check
5718 certificate signature with the NULL public key.
5722 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5723 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5727 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5728 extra data after the compression methods not only for TLS 1.0
5729 but also for SSL 3.0 (as required by the specification).
5730 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5732 *) Change X509_certificate_type() to mark the key as exported/exportable
5733 when it's 512 *bits* long, not 512 bytes.
5736 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5738 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5739 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5740 a protocol version number mismatch like a decryption error
5741 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5744 *) Turn on RSA blinding by default in the default implementation
5745 to avoid a timing attack. Applications that don't want it can call
5746 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5747 They would be ill-advised to do so in most cases.
5748 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5750 *) Change RSA blinding code so that it works when the PRNG is not
5751 seeded (in this case, the secret RSA exponent is abused as
5752 an unpredictable seed -- if it is not unpredictable, there
5753 is no point in blinding anyway). Make RSA blinding thread-safe
5754 by remembering the creator's thread ID in rsa->blinding and
5755 having all other threads use local one-time blinding factors
5756 (this requires more computation than sharing rsa->blinding, but
5757 avoids excessive locking; and if an RSA object is not shared
5758 between threads, blinding will still be very fast).
5761 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5763 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5764 via timing by performing a MAC computation even if incorrrect
5765 block cipher padding has been found. This is a countermeasure
5766 against active attacks where the attacker has to distinguish
5767 between bad padding and a MAC verification error. (CVE-2003-0078)
5769 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5770 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5771 Martin Vuagnoux (EPFL, Ilion)]
5773 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5775 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5776 memory from it's contents. This is done with a counter that will
5777 place alternating values in each byte. This can be used to solve
5778 two issues: 1) the removal of calls to memset() by highly optimizing
5779 compilers, and 2) cleansing with other values than 0, since those can
5780 be read through on certain media, for example a swap space on disk.
5783 *) Bugfix: client side session caching did not work with external caching,
5784 because the session->cipher setting was not restored when reloading
5785 from the external cache. This problem was masked, when
5786 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5787 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5790 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5791 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5792 [Zeev Lieber <zeev-l@yahoo.com>]
5794 *) Undo an undocumented change introduced in 0.9.6e which caused
5795 repeated calls to OpenSSL_add_all_ciphers() and
5796 OpenSSL_add_all_digests() to be ignored, even after calling
5800 *) Change the default configuration reader to deal with last line not
5801 being properly terminated.
5804 *) Change X509_NAME_cmp() so it applies the special rules on handling
5805 DN values that are of type PrintableString, as well as RDNs of type
5806 emailAddress where the value has the type ia5String.
5807 [stefank@valicert.com via Richard Levitte]
5809 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5810 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5811 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5812 the bitwise-OR of the two for use by the majority of applications
5813 wanting this behaviour, and update the docs. The documented
5814 behaviour and actual behaviour were inconsistent and had been
5815 changing anyway, so this is more a bug-fix than a behavioural
5817 [Geoff Thorpe, diagnosed by Nadav Har'El]
5819 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5820 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5823 *) Fix initialization code race conditions in
5824 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5825 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5826 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5827 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5828 ssl2_get_cipher_by_char(),
5829 ssl3_get_cipher_by_char().
5830 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5832 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5833 the cached sessions are flushed, as the remove_cb() might use ex_data
5834 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5835 (see [openssl.org #212]).
5836 [Geoff Thorpe, Lutz Jaenicke]
5838 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5839 length, instead of the encoding length to d2i_ASN1_OBJECT.
5842 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5844 *) [In 0.9.6g-engine release:]
5845 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5846 [Lynn Gazis <lgazis@rainbow.com>]
5848 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5850 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5851 and get fix the header length calculation.
5852 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5853 Alon Kantor <alonk@checkpoint.com> (and others),
5856 *) Use proper error handling instead of 'assertions' in buffer
5857 overflow checks added in 0.9.6e. This prevents DoS (the
5858 assertions could call abort()).
5859 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5861 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5863 *) Add various sanity checks to asn1_get_length() to reject
5864 the ASN1 length bytes if they exceed sizeof(long), will appear
5865 negative or the content length exceeds the length of the
5867 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5869 *) Fix cipher selection routines: ciphers without encryption had no flags
5870 for the cipher strength set and where therefore not handled correctly
5871 by the selection routines (PR #130).
5874 *) Fix EVP_dsa_sha macro.
5878 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5879 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5880 that was added in OpenSSL 0.9.6d.
5882 As the countermeasure turned out to be incompatible with some
5883 broken SSL implementations, the new option is part of SSL_OP_ALL.
5884 SSL_OP_ALL is usually employed when compatibility with weird SSL
5885 implementations is desired (e.g. '-bugs' option to 's_client' and
5886 's_server'), so the new option is automatically set in many
5890 *) Changes in security patch:
5892 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5893 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5894 Air Force Materiel Command, USAF, under agreement number
5897 *) Add various sanity checks to asn1_get_length() to reject
5898 the ASN1 length bytes if they exceed sizeof(long), will appear
5899 negative or the content length exceeds the length of the
5900 supplied buffer. (CVE-2002-0659)
5901 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5903 *) Assertions for various potential buffer overflows, not known to
5905 [Ben Laurie (CHATS)]
5907 *) Various temporary buffers to hold ASCII versions of integers were
5908 too small for 64 bit platforms. (CVE-2002-0655)
5909 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5911 *) Remote buffer overflow in SSL3 protocol - an attacker could
5912 supply an oversized session ID to a client. (CVE-2002-0656)
5913 [Ben Laurie (CHATS)]
5915 *) Remote buffer overflow in SSL2 protocol - an attacker could
5916 supply an oversized client master key. (CVE-2002-0656)
5917 [Ben Laurie (CHATS)]
5919 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5921 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5922 encoded as NULL) with id-dsa-with-sha1.
5923 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5925 *) Check various X509_...() return values in apps/req.c.
5926 [Nils Larsch <nla@trustcenter.de>]
5928 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5929 an end-of-file condition would erronously be flagged, when the CRLF
5930 was just at the end of a processed block. The bug was discovered when
5931 processing data through a buffering memory BIO handing the data to a
5932 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5933 <ptsekov@syntrex.com> and Nedelcho Stanev.
5936 *) Implement a countermeasure against a vulnerability recently found
5937 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5938 before application data chunks to avoid the use of known IVs
5939 with data potentially chosen by the attacker.
5942 *) Fix length checks in ssl3_get_client_hello().
5945 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5946 to prevent ssl3_read_internal() from incorrectly assuming that
5947 ssl3_read_bytes() found application data while handshake
5948 processing was enabled when in fact s->s3->in_read_app_data was
5949 merely automatically cleared during the initial handshake.
5950 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5952 *) Fix object definitions for Private and Enterprise: they were not
5953 recognized in their shortname (=lowercase) representation. Extend
5954 obj_dat.pl to issue an error when using undefined keywords instead
5955 of silently ignoring the problem (Svenning Sorensen
5956 <sss@sss.dnsalias.net>).
5959 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5960 generators, i.e. generators other than 2 and 5. (Previously, the
5961 code did not properly initialise the 'add' and 'rem' values to
5962 BN_generate_prime().)
5964 In the new general case, we do not insist that 'generator' is
5965 actually a primitive root: This requirement is rather pointless;
5966 a generator of the order-q subgroup is just as good, if not
5970 *) Map new X509 verification errors to alerts. Discovered and submitted by
5971 Tom Wu <tom@arcot.com>.
5974 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5975 returning non-zero before the data has been completely received
5976 when using non-blocking I/O.
5977 [Bodo Moeller; problem pointed out by John Hughes]
5979 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5980 [Ben Laurie, Lutz Jaenicke]
5982 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5983 Yoram Zahavi <YoramZ@gilian.com>).
5986 *) Add information about CygWin 1.3 and on, and preserve proper
5987 configuration for the versions before that.
5988 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5990 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5991 check whether we deal with a copy of a session and do not delete from
5992 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5993 <izhar@checkpoint.com>.
5996 *) Do not store session data into the internal session cache, if it
5997 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5998 flag is set). Proposed by Aslam <aslam@funk.com>.
6001 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6005 *) [In 0.9.6d-engine release:]
6006 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6007 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6009 *) Add the configuration target linux-s390x.
6010 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6012 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6013 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6014 variable as an indication that a ClientHello message has been
6015 received. As the flag value will be lost between multiple
6016 invocations of ssl3_accept when using non-blocking I/O, the
6017 function may not be aware that a handshake has actually taken
6018 place, thus preventing a new session from being added to the
6021 To avoid this problem, we now set s->new_session to 2 instead of
6022 using a local variable.
6023 [Lutz Jaenicke, Bodo Moeller]
6025 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6026 if the SSL_R_LENGTH_MISMATCH error is detected.
6027 [Geoff Thorpe, Bodo Moeller]
6029 *) New 'shared_ldflag' column in Configure platform table.
6032 *) Fix EVP_CIPHER_mode macro.
6033 ["Dan S. Camper" <dan@bti.net>]
6035 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6036 type, we must throw them away by setting rr->length to 0.
6037 [D P Chang <dpc@qualys.com>]
6039 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6041 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6042 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6043 worked incorrectly for those cases where range = 10..._2 and
6044 3*range is two bits longer than range.)
6047 *) Only add signing time to PKCS7 structures if it is not already
6051 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6052 OBJ_ld_ce should be OBJ_id_ce.
6053 Also some ip-pda OIDs in crypto/objects/objects.txt were
6054 incorrect (cf. RFC 3039).
6055 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6057 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6058 returns early because it has nothing to do.
6059 [Andy Schneider <andy.schneider@bjss.co.uk>]
6061 *) [In 0.9.6c-engine release:]
6062 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6063 [Andy Schneider <andy.schneider@bjss.co.uk>]
6065 *) [In 0.9.6c-engine release:]
6066 Add support for Cryptographic Appliance's keyserver technology.
6067 (Use engine 'keyclient')
6068 [Cryptographic Appliances and Geoff Thorpe]
6070 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6071 is called via tools/c89.sh because arguments have to be
6072 rearranged (all '-L' options must appear before the first object
6074 [Richard Shapiro <rshapiro@abinitio.com>]
6076 *) [In 0.9.6c-engine release:]
6077 Add support for Broadcom crypto accelerator cards, backported
6079 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6081 *) [In 0.9.6c-engine release:]
6082 Add support for SureWare crypto accelerator cards from
6083 Baltimore Technologies. (Use engine 'sureware')
6084 [Baltimore Technologies and Mark Cox]
6086 *) [In 0.9.6c-engine release:]
6087 Add support for crypto accelerator cards from Accelerated
6088 Encryption Processing, www.aep.ie. (Use engine 'aep')
6089 [AEP Inc. and Mark Cox]
6091 *) Add a configuration entry for gcc on UnixWare.
6092 [Gary Benson <gbenson@redhat.com>]
6094 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6095 messages are stored in a single piece (fixed-length part and
6096 variable-length part combined) and fix various bugs found on the way.
6099 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6100 instead. BIO_gethostbyname() does not know what timeouts are
6101 appropriate, so entries would stay in cache even when they have
6103 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6105 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6106 faced with a pathologically small ClientHello fragment that does
6107 not contain client_version: Instead of aborting with an error,
6108 simply choose the highest available protocol version (i.e.,
6109 TLS 1.0 unless it is disabled). In practice, ClientHello
6110 messages are never sent like this, but this change gives us
6111 strictly correct behaviour at least for TLS.
6114 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6115 never resets s->method to s->ctx->method when called from within
6116 one of the SSL handshake functions.
6117 [Bodo Moeller; problem pointed out by Niko Baric]
6119 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6120 (sent using the client's version number) if client_version is
6121 smaller than the protocol version in use. Also change
6122 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6123 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6124 the client will at least see that alert.
6127 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6131 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6132 client receives HelloRequest while in a handshake.
6133 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6135 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6136 should end in 'break', not 'goto end' which circuments various
6137 cleanups done in state SSL_ST_OK. But session related stuff
6138 must be disabled for SSL_ST_OK in the case that we just sent a
6141 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6142 before just sending a HelloRequest.
6143 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6145 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6146 reveal whether illegal block cipher padding was found or a MAC
6147 verification error occured. (Neither SSLerr() codes nor alerts
6148 are directly visible to potential attackers, but the information
6149 may leak via logfiles.)
6151 Similar changes are not required for the SSL 2.0 implementation
6152 because the number of padding bytes is sent in clear for SSL 2.0,
6153 and the extra bytes are just ignored. However ssl/s2_pkt.c
6154 failed to verify that the purported number of padding bytes is in
6158 *) Add OpenUNIX-8 support including shared libraries
6159 (Boyd Lynn Gerber <gerberb@zenez.com>).
6162 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6163 'wristwatch attack' using huge encoding parameters (cf.
6164 James H. Manger's CRYPTO 2001 paper). Note that the
6165 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6166 encoding parameters and hence was not vulnerable.
6169 *) BN_sqr() bug fix.
6170 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6172 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6173 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6174 followed by modular reduction.
6175 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6177 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6178 equivalent based on BN_pseudo_rand() instead of BN_rand().
6181 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6182 This function was broken, as the check for a new client hello message
6183 to handle SGC did not allow these large messages.
6184 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6187 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6190 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6191 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6194 *) Rework the configuration and shared library support for Tru64 Unix.
6195 The configuration part makes use of modern compiler features and
6196 still retains old compiler behavior for those that run older versions
6197 of the OS. The shared library support part includes a variant that
6198 uses the RPATH feature, and is available through the special
6199 configuration target "alpha-cc-rpath", which will never be selected
6201 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6203 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6204 with the same message size as in ssl3_get_certificate_request().
6205 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6206 messages might inadvertently be reject as too long.
6207 [Petr Lampa <lampa@fee.vutbr.cz>]
6209 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6212 *) Modified SSL library such that the verify_callback that has been set
6213 specificly for an SSL object with SSL_set_verify() is actually being
6214 used. Before the change, a verify_callback set with this function was
6215 ignored and the verify_callback() set in the SSL_CTX at the time of
6216 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6217 to allow the necessary settings.
6220 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6221 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6222 done automatically (in contradiction to the requirements of the C
6223 standard). This made problems when used from OpenSSH.
6226 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6227 dh->length and always used
6229 BN_rand_range(priv_key, dh->p).
6231 BN_rand_range() is not necessary for Diffie-Hellman, and this
6232 specific range makes Diffie-Hellman unnecessarily inefficient if
6233 dh->length (recommended exponent length) is much smaller than the
6234 length of dh->p. We could use BN_rand_range() if the order of
6235 the subgroup was stored in the DH structure, but we only have
6240 BN_rand(priv_key, l, ...)
6242 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6248 RSA_eay_public_encrypt
6249 RSA_eay_private_decrypt
6250 RSA_eay_private_encrypt (signing)
6251 RSA_eay_public_decrypt (signature verification)
6253 (default implementations for RSA_public_encrypt,
6254 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6255 always reject numbers >= n.
6258 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6259 to synchronize access to 'locking_thread'. This is necessary on
6260 systems where access to 'locking_thread' (an 'unsigned long'
6261 variable) is not atomic.
6264 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6265 *before* setting the 'crypto_lock_rand' flag. The previous code had
6266 a race condition if 0 is a valid thread ID.
6267 [Travis Vitek <vitek@roguewave.com>]
6269 *) Add support for shared libraries under Irix.
6270 [Albert Chin-A-Young <china@thewrittenword.com>]
6272 *) Add configuration option to build on Linux on both big-endian and
6274 [Ralf Baechle <ralf@uni-koblenz.de>]
6276 *) Add the possibility to create shared libraries on HP-UX.
6279 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6281 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6282 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6283 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6284 PRNG state recovery was possible based on the output of
6285 one PRNG request appropriately sized to gain knowledge on
6286 'md' followed by enough consecutive 1-byte PRNG requests
6287 to traverse all of 'state'.
6289 1. When updating 'md_local' (the current thread's copy of 'md')
6290 during PRNG output generation, hash all of the previous
6291 'md_local' value, not just the half used for PRNG output.
6293 2. Make the number of bytes from 'state' included into the hash
6294 independent from the number of PRNG bytes requested.
6296 The first measure alone would be sufficient to avoid
6297 Markku-Juhani's attack. (Actually it had never occurred
6298 to me that the half of 'md_local' used for chaining was the
6299 half from which PRNG output bytes were taken -- I had always
6300 assumed that the secret half would be used.) The second
6301 measure makes sure that additional data from 'state' is never
6302 mixed into 'md_local' in small portions; this heuristically
6303 further strengthens the PRNG.
6306 *) Fix crypto/bn/asm/mips3.s.
6309 *) When only the key is given to "enc", the IV is undefined. Print out
6310 an error message in this case.
6313 *) Handle special case when X509_NAME is empty in X509 printing routines.
6316 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6317 positive and less than q.
6320 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6321 used: it isn't thread safe and the add_lock_callback should handle
6323 [Paul Rose <Paul.Rose@bridge.com>]
6325 *) Verify that incoming data obeys the block size in
6326 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6330 [Ulf Möller, Bodo Möller]
6332 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6333 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6334 when fixing the server behaviour for backwards-compatible 'client
6335 hello' messages. (Note that the attack is impractical against
6336 SSL 3.0 and TLS 1.0 anyway because length and version checking
6337 means that the probability of guessing a valid ciphertext is
6338 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6341 Before 0.9.5, the countermeasure (hide the error by generating a
6342 random 'decryption result') did not work properly because
6343 ERR_clear_error() was missing, meaning that SSL_get_error() would
6344 detect the supposedly ignored error.
6346 Both problems are now fixed.
6349 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6350 (previously it was 1024).
6353 *) Fix for compatibility mode trust settings: ignore trust settings
6354 unless some valid trust or reject settings are present.
6357 *) Fix for blowfish EVP: its a variable length cipher.
6360 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6361 parameters in DSA public key structures and return an error in the
6362 DSA routines if parameters are absent.
6365 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6366 in the current directory if neither $RANDFILE nor $HOME was set.
6367 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6368 caused some confusion to Windows users who haven't defined $HOME.
6369 Thus RAND_file_name() is changed again: e_os.h can define a
6370 DEFAULT_HOME, which will be used if $HOME is not set.
6371 For Windows, we use "C:"; on other platforms, we still require
6372 environment variables.
6374 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6375 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6376 having multiple threads call RAND_poll() concurrently.
6379 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6380 combination of a flag and a thread ID variable.
6381 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6382 flag), *other* threads can enter ssleay_add_bytes without obeying
6383 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6384 that they do not hold after the first thread unsets add_do_not_lock).
6387 *) Change bctest again: '-x' expressions are not available in all
6391 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6393 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6394 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6396 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6397 the default extension for executables, if any. Also, make the perl
6398 scripts that use symlink() to test if it really exists and use "cp"
6399 if it doesn't. All this made OpenSSL compilable and installable in
6403 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6404 If SEQUENCE is length is indefinite just set c->slen to the total
6405 amount of data available.
6406 [Steve Henson, reported by shige@FreeBSD.org]
6407 [This change does not apply to 0.9.7.]
6409 *) Change bctest to avoid here-documents inside command substitution
6410 (workaround for FreeBSD /bin/sh bug).
6411 For compatibility with Ultrix, avoid shell functions (introduced
6412 in the bctest version that searches along $PATH).
6415 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6416 with des_encrypt() defined on some operating systems, like Solaris
6420 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6421 On the Importance of Eliminating Errors in Cryptographic
6422 Computations, J. Cryptology 14 (2001) 2, 101-119,
6423 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6426 *) MIPS assembler BIGNUM division bug fix.
6429 *) Disabled incorrect Alpha assembler code.
6432 *) Fix PKCS#7 decode routines so they correctly update the length
6433 after reading an EOC for the EXPLICIT tag.
6435 [This change does not apply to 0.9.7.]
6437 *) Fix bug in PKCS#12 key generation routines. This was triggered
6438 if a 3DES key was generated with a 0 initial byte. Include
6439 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6440 (but broken) behaviour.
6443 *) Enhance bctest to search for a working bc along $PATH and print
6445 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6447 *) Fix memory leaks in err.c: free err_data string if necessary;
6448 don't write to the wrong index in ERR_set_error_data.
6451 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6455 *) Replace rdtsc with _emit statements for VC++ version 5.
6456 [Jeremy Cooper <jeremy@baymoo.org>]
6458 *) Make it possible to reuse SSLv2 sessions.
6461 *) In copy_email() check for >= 0 as a return value for
6462 X509_NAME_get_index_by_NID() since 0 is a valid index.
6463 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6465 *) Avoid coredump with unsupported or invalid public keys by checking if
6466 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6467 PKCS7_verify() fails with non detached data.
6470 *) Don't use getenv in library functions when run as setuid/setgid.
6471 New function OPENSSL_issetugid().
6474 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6475 due to incorrect handling of multi-threading:
6477 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6479 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6481 3. Count how many times MemCheck_off() has been called so that
6482 nested use can be treated correctly. This also avoids
6483 inband-signalling in the previous code (which relied on the
6484 assumption that thread ID 0 is impossible).
6487 *) Add "-rand" option also to s_client and s_server.
6490 *) Fix CPU detection on Irix 6.x.
6491 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6492 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6494 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6497 [This change does not apply to 0.9.7.]
6499 *) Use the cached encoding of an X509_NAME structure rather than
6500 copying it. This is apparently the reason for the libsafe "errors"
6501 but the code is actually correct.
6504 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6505 Bleichenbacher's DSA attack.
6506 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6507 to be set and top=0 forces the highest bit to be set; top=-1 is new
6508 and leaves the highest bit random.
6509 [Ulf Moeller, Bodo Moeller]
6511 *) In the NCONF_...-based implementations for CONF_... queries
6512 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6513 a temporary CONF structure with the data component set to NULL
6514 (which gives segmentation faults in lh_retrieve).
6515 Instead, use NULL for the CONF pointer in CONF_get_string and
6516 CONF_get_number (which may use environment variables) and directly
6517 return NULL from CONF_get_section.
6520 *) Fix potential buffer overrun for EBCDIC.
6523 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6524 keyUsage if basicConstraints absent for a CA.
6527 *) Make SMIME_write_PKCS7() write mail header values with a format that
6528 is more generally accepted (no spaces before the semicolon), since
6529 some programs can't parse those values properly otherwise. Also make
6530 sure BIO's that break lines after each write do not create invalid
6534 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6535 macros previously used would not encode an empty SEQUENCE OF
6536 and break the signature.
6538 [This change does not apply to 0.9.7.]
6540 *) Zero the premaster secret after deriving the master secret in
6544 *) Add some EVP_add_digest_alias registrations (as found in
6545 OpenSSL_add_all_digests()) to SSL_library_init()
6546 aka OpenSSL_add_ssl_algorithms(). This provides improved
6547 compatibility with peers using X.509 certificates
6548 with unconventional AlgorithmIdentifier OIDs.
6551 *) Fix for Irix with NO_ASM.
6552 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6554 *) ./config script fixes.
6555 [Ulf Moeller, Richard Levitte]
6557 *) Fix 'openssl passwd -1'.
6560 *) Change PKCS12_key_gen_asc() so it can cope with non null
6561 terminated strings whose length is passed in the passlen
6562 parameter, for example from PEM callbacks. This was done
6563 by adding an extra length parameter to asc2uni().
6564 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6566 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6567 call failed, free the DSA structure.
6570 *) Fix to uni2asc() to cope with zero length Unicode strings.
6571 These are present in some PKCS#12 files.
6574 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6575 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6576 when writing a 32767 byte record.
6577 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6579 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6580 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6582 (RSA objects have a reference count access to which is protected
6583 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6584 so they are meant to be shared between threads.)
6585 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6586 "Reddie, Steven" <Steven.Reddie@ca.com>]
6588 *) Fix a deadlock in CRYPTO_mem_leaks().
6591 *) Use better test patterns in bntest.
6594 *) rand_win.c fix for Borland C.
6597 *) BN_rshift bugfix for n == 0.
6600 *) Add a 'bctest' script that checks for some known 'bc' bugs
6601 so that 'make test' does not abort just because 'bc' is broken.
6604 *) Store verify_result within SSL_SESSION also for client side to
6605 avoid potential security hole. (Re-used sessions on the client side
6606 always resulted in verify_result==X509_V_OK, not using the original
6607 result of the server certificate verification.)
6610 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6611 SSL3_RT_APPLICATION_DATA, return 0.
6612 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6616 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6617 releases, have been re-implemented by renaming the previous
6618 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6619 and ssl3_read_internal, respectively, and adding 'peek' parameters
6620 to them. The new ssl[23]_{read,peek} functions are calls to
6621 ssl[23]_read_internal with the 'peek' flag set appropriately.
6622 A 'peek' parameter has also been added to ssl3_read_bytes, which
6623 does the actual work for ssl3_read_internal.
6626 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6627 the method-specific "init()" handler. Also clean up ex_data after
6628 calling the method-specific "finish()" handler. Previously, this was
6629 happening the other way round.
6632 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6633 The previous value, 12, was not always sufficient for BN_mod_exp().
6636 *) Make sure that shared libraries get the internal name engine with
6637 the full version number and not just 0. This should mark the
6638 shared libraries as not backward compatible. Of course, this should
6639 be changed again when we can guarantee backward binary compatibility.
6642 *) Fix typo in get_cert_by_subject() in by_dir.c
6643 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6645 *) Rework the system to generate shared libraries:
6647 - Make note of the expected extension for the shared libraries and
6648 if there is a need for symbolic links from for example libcrypto.so.0
6649 to libcrypto.so.0.9.7. There is extended info in Configure for
6652 - Make as few rebuilds of the shared libraries as possible.
6654 - Still avoid linking the OpenSSL programs with the shared libraries.
6656 - When installing, install the shared libraries separately from the
6660 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6662 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6663 and not in SSL_clear because the latter is also used by the
6664 accept/connect functions; previously, the settings made by
6665 SSL_set_read_ahead would be lost during the handshake.
6666 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6668 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6669 Previously, it would create entries for disableed algorithms no
6673 *) Added several new manual pages for SSL_* function.
6676 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6678 *) In ssl23_get_client_hello, generate an error message when faced
6679 with an initial SSL 3.0/TLS record that is too small to contain the
6680 first two bytes of the ClientHello message, i.e. client_version.
6681 (Note that this is a pathologic case that probably has never happened
6682 in real life.) The previous approach was to use the version number
6683 from the record header as a substitute; but our protocol choice
6684 should not depend on that one because it is not authenticated
6685 by the Finished messages.
6688 *) More robust randomness gathering functions for Windows.
6689 [Jeffrey Altman <jaltman@columbia.edu>]
6691 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6692 not set then we don't setup the error code for issuer check errors
6693 to avoid possibly overwriting other errors which the callback does
6694 handle. If an application does set the flag then we assume it knows
6695 what it is doing and can handle the new informational codes
6699 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6700 a general "ANY" type, as such it should be able to decode anything
6701 including tagged types. However it didn't check the class so it would
6702 wrongly interpret tagged types in the same way as their universal
6703 counterpart and unknown types were just rejected. Changed so that the
6704 tagged and unknown types are handled in the same way as a SEQUENCE:
6705 that is the encoding is stored intact. There is also a new type
6706 "V_ASN1_OTHER" which is used when the class is not universal, in this
6707 case we have no idea what the actual type is so we just lump them all
6711 *) On VMS, stdout may very well lead to a file that is written to
6712 in a record-oriented fashion. That means that every write() will
6713 write a separate record, which will be read separately by the
6714 programs trying to read from it. This can be very confusing.
6716 The solution is to put a BIO filter in the way that will buffer
6717 text until a linefeed is reached, and then write everything a
6718 line at a time, so every record written will be an actual line,
6719 not chunks of lines and not (usually doesn't happen, but I've
6720 seen it once) several lines in one record. BIO_f_linebuffer() is
6723 Currently, it's a VMS-only method, because that's where it has
6724 been tested well enough.
6727 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6728 it can return incorrect results.
6729 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6730 but it was in 0.9.6-beta[12].)
6733 *) Disable the check for content being present when verifying detached
6734 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6735 include zero length content when signing messages.
6738 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6739 BIO_ctrl (for BIO pairs).
6742 *) Add DSO method for VMS.
6745 *) Bug fix: Montgomery multiplication could produce results with the
6749 *) Add RPM specification openssl.spec and modify it to build three
6750 packages. The default package contains applications, application
6751 documentation and run-time libraries. The devel package contains
6752 include files, static libraries and function documentation. The
6753 doc package contains the contents of the doc directory. The original
6754 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6757 *) Add a large number of documentation files for many SSL routines.
6758 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6760 *) Add a configuration entry for Sony News 4.
6761 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6763 *) Don't set the two most significant bits to one when generating a
6764 random number < q in the DSA library.
6767 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6768 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6769 the underlying transport is blocking) if a handshake took place.
6770 (The default behaviour is needed by applications such as s_client
6771 and s_server that use select() to determine when to use SSL_read;
6772 but for applications that know in advance when to expect data, it
6773 just makes things more complicated.)
6776 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6780 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6781 work better on such systems.
6782 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6784 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6785 Update PKCS12_parse() so it copies the friendlyName and the
6786 keyid to the certificates aux info.
6789 *) Fix bug in PKCS7_verify() which caused an infinite loop
6790 if there was more than one signature.
6791 [Sven Uszpelkat <su@celocom.de>]
6793 *) Major change in util/mkdef.pl to include extra information
6794 about each symbol, as well as presentig variables as well
6795 as functions. This change means that there's n more need
6796 to rebuild the .num files when some algorithms are excluded.
6799 *) Allow the verify time to be set by an application,
6800 rather than always using the current time.
6803 *) Phase 2 verify code reorganisation. The certificate
6804 verify code now looks up an issuer certificate by a
6805 number of criteria: subject name, authority key id
6806 and key usage. It also verifies self signed certificates
6807 by the same criteria. The main comparison function is
6808 X509_check_issued() which performs these checks.
6810 Lot of changes were necessary in order to support this
6811 without completely rewriting the lookup code.
6813 Authority and subject key identifier are now cached.
6815 The LHASH 'certs' is X509_STORE has now been replaced
6816 by a STACK_OF(X509_OBJECT). This is mainly because an
6817 LHASH can't store or retrieve multiple objects with
6818 the same hash value.
6820 As a result various functions (which were all internal
6821 use only) have changed to handle the new X509_STORE
6822 structure. This will break anything that messed round
6823 with X509_STORE internally.
6825 The functions X509_STORE_add_cert() now checks for an
6826 exact match, rather than just subject name.
6828 The X509_STORE API doesn't directly support the retrieval
6829 of multiple certificates matching a given criteria, however
6830 this can be worked round by performing a lookup first
6831 (which will fill the cache with candidate certificates)
6832 and then examining the cache for matches. This is probably
6833 the best we can do without throwing out X509_LOOKUP
6834 entirely (maybe later...).
6836 The X509_VERIFY_CTX structure has been enhanced considerably.
6838 All certificate lookup operations now go via a get_issuer()
6839 callback. Although this currently uses an X509_STORE it
6840 can be replaced by custom lookups. This is a simple way
6841 to bypass the X509_STORE hackery necessary to make this
6842 work and makes it possible to use more efficient techniques
6843 in future. A very simple version which uses a simple
6844 STACK for its trusted certificate store is also provided
6845 using X509_STORE_CTX_trusted_stack().
6847 The verify_cb() and verify() callbacks now have equivalents
6848 in the X509_STORE_CTX structure.
6850 X509_STORE_CTX also has a 'flags' field which can be used
6851 to customise the verify behaviour.
6854 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6855 excludes S/MIME capabilities.
6858 *) When a certificate request is read in keep a copy of the
6859 original encoding of the signed data and use it when outputing
6860 again. Signatures then use the original encoding rather than
6861 a decoded, encoded version which may cause problems if the
6862 request is improperly encoded.
6865 *) For consistency with other BIO_puts implementations, call
6866 buffer_write(b, ...) directly in buffer_puts instead of calling
6869 In BIO_puts, increment b->num_write as in BIO_write.
6870 [Peter.Sylvester@EdelWeb.fr]
6872 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6873 BN_zero, we may not return a BIGNUM with an array consisting of
6877 *) Avoid calling abort() from within the library when problems are
6878 detected, except if preprocessor symbols have been defined
6879 (such as REF_CHECK, BN_DEBUG etc.).
6882 *) New openssl application 'rsautl'. This utility can be
6883 used for low level RSA operations. DER public key
6884 BIO/fp routines also added.
6887 *) New Configure entry and patches for compiling on QNX 4.
6888 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6890 *) A demo state-machine implementation was sponsored by
6891 Nuron (http://www.nuron.com/) and is now available in
6892 demos/state_machine.
6895 *) New options added to the 'dgst' utility for signature
6896 generation and verification.
6899 *) Unrecognized PKCS#7 content types are now handled via a
6900 catch all ASN1_TYPE structure. This allows unsupported
6901 types to be stored as a "blob" and an application can
6902 encode and decode it manually.
6905 *) Fix various signed/unsigned issues to make a_strex.c
6907 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6909 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6910 length if passed a buffer. ASN1_INTEGER_to_BN failed
6911 if passed a NULL BN and its argument was negative.
6912 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6914 *) Modification to PKCS#7 encoding routines to output definite
6915 length encoding. Since currently the whole structures are in
6916 memory there's not real point in using indefinite length
6917 constructed encoding. However if OpenSSL is compiled with
6918 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6921 *) Added BIO_vprintf() and BIO_vsnprintf().
6924 *) Added more prefixes to parse for in the the strings written
6925 through a logging bio, to cover all the levels that are available
6926 through syslog. The prefixes are now:
6928 PANIC, EMERG, EMR => LOG_EMERG
6929 ALERT, ALR => LOG_ALERT
6930 CRIT, CRI => LOG_CRIT
6931 ERROR, ERR => LOG_ERR
6932 WARNING, WARN, WAR => LOG_WARNING
6933 NOTICE, NOTE, NOT => LOG_NOTICE
6934 INFO, INF => LOG_INFO
6935 DEBUG, DBG => LOG_DEBUG
6937 and as before, if none of those prefixes are present at the
6938 beginning of the string, LOG_ERR is chosen.
6940 On Win32, the LOG_* levels are mapped according to this:
6942 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6943 LOG_WARNING => EVENTLOG_WARNING_TYPE
6944 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6948 *) Made it possible to reconfigure with just the configuration
6949 argument "reconf" or "reconfigure". The command line arguments
6950 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6951 and are retrieved from there when reconfiguring.
6955 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6957 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6960 *) The obj_dat.pl script was messing up the sorting of object
6961 names. The reason was that it compared the quoted version
6962 of strings as a result "OCSP" > "OCSP Signing" because
6963 " > SPACE. Changed script to store unquoted versions of
6964 names and add quotes on output. It was also omitting some
6965 names from the lookup table if they were given a default
6966 value (that is if SN is missing it is given the same
6967 value as LN and vice versa), these are now added on the
6968 grounds that if an object has a name we should be able to
6969 look it up. Finally added warning output when duplicate
6970 short or long names are found.
6973 *) Changes needed for Tandem NSK.
6974 [Scott Uroff <scott@xypro.com>]
6976 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6977 RSA_padding_check_SSLv23(), special padding was never detected
6978 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6979 version rollback attacks was not effective.
6981 In s23_clnt.c, don't use special rollback-attack detection padding
6982 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6983 client; similarly, in s23_srvr.c, don't do the rollback check if
6984 SSL 2.0 is the only protocol enabled in the server.
6987 *) Make it possible to get hexdumps of unprintable data with 'openssl
6988 asn1parse'. By implication, the functions ASN1_parse_dump() and
6989 BIO_dump_indent() are added.
6992 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6993 these print out strings and name structures based on various
6994 flags including RFC2253 support and proper handling of
6995 multibyte characters. Added options to the 'x509' utility
6996 to allow the various flags to be set.
6999 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7000 Also change the functions X509_cmp_current_time() and
7001 X509_gmtime_adj() work with an ASN1_TIME structure,
7002 this will enable certificates using GeneralizedTime in validity
7003 dates to be checked.
7006 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7007 negative public key encodings) on by default,
7008 NO_NEG_PUBKEY_BUG can be set to disable it.
7011 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7012 content octets. An i2c_ASN1_OBJECT is unnecessary because
7013 the encoding can be trivially obtained from the structure.
7016 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7017 not read locks (CRYPTO_r_[un]lock).
7020 *) A first attempt at creating official support for shared
7021 libraries through configuration. I've kept it so the
7022 default is static libraries only, and the OpenSSL programs
7023 are always statically linked for now, but there are
7024 preparations for dynamic linking in place.
7025 This has been tested on Linux and Tru64.
7028 *) Randomness polling function for Win9x, as described in:
7029 Peter Gutmann, Software Generation of Practically Strong
7033 *) Fix so PRNG is seeded in req if using an already existing
7037 *) New options to smime application. -inform and -outform
7038 allow alternative formats for the S/MIME message including
7039 PEM and DER. The -content option allows the content to be
7040 specified separately. This should allow things like Netscape
7041 form signing output easier to verify.
7044 *) Fix the ASN1 encoding of tags using the 'long form'.
7047 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7048 STRING types. These convert content octets to and from the
7049 underlying type. The actual tag and length octets are
7050 already assumed to have been read in and checked. These
7051 are needed because all other string types have virtually
7052 identical handling apart from the tag. By having versions
7053 of the ASN1 functions that just operate on content octets
7054 IMPLICIT tagging can be handled properly. It also allows
7055 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7056 and ASN1_INTEGER are identical apart from the tag.
7059 *) Change the handling of OID objects as follows:
7061 - New object identifiers are inserted in objects.txt, following
7062 the syntax given in objects.README.
7063 - objects.pl is used to process obj_mac.num and create a new
7065 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7068 This is currently kind of a hack, and the perl code in objects.pl
7069 isn't very elegant, but it works as I intended. The simplest way
7070 to check that it worked correctly is to look in obj_dat.h and
7071 check the array nid_objs and make sure the objects haven't moved
7072 around (this is important!). Additions are OK, as well as
7073 consistent name changes.
7076 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7079 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7080 The given file adds to whatever has already been seeded into the
7081 random pool through the RANDFILE configuration file option or
7082 environment variable, or the default random state file.
7085 *) mkstack.pl now sorts each macro group into lexical order.
7086 Previously the output order depended on the order the files
7087 appeared in the directory, resulting in needless rewriting
7091 *) Patches to make OpenSSL compile under Win32 again. Mostly
7092 work arounds for the VC++ problem that it treats func() as
7093 func(void). Also stripped out the parts of mkdef.pl that
7094 added extra typesafe functions: these no longer exist.
7097 *) Reorganisation of the stack code. The macros are now all
7098 collected in safestack.h . Each macro is defined in terms of
7099 a "stack macro" of the form SKM_<name>(type, a, b). The
7100 DEBUG_SAFESTACK is now handled in terms of function casts,
7101 this has the advantage of retaining type safety without the
7102 use of additional functions. If DEBUG_SAFESTACK is not defined
7103 then the non typesafe macros are used instead. Also modified the
7104 mkstack.pl script to handle the new form. Needs testing to see
7105 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7106 the default if no major problems. Similar behaviour for ASN1_SET_OF
7107 and PKCS12_STACK_OF.
7110 *) When some versions of IIS use the 'NET' form of private key the
7111 key derivation algorithm is different. Normally MD5(password) is
7112 used as a 128 bit RC4 key. In the modified case
7113 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7114 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7115 as the old Netscape_RSA functions except they have an additional
7116 'sgckey' parameter which uses the modified algorithm. Also added
7117 an -sgckey command line option to the rsa utility. Thanks to
7118 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7119 algorithm to openssl-dev.
7122 *) The evp_local.h macros were using 'c.##kname' which resulted in
7123 invalid expansion on some systems (SCO 5.0.5 for example).
7124 Corrected to 'c.kname'.
7125 [Phillip Porch <root@theporch.com>]
7127 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7128 a STACK of email addresses from a certificate or request, these look
7129 in the subject name and the subject alternative name extensions and
7130 omit any duplicate addresses.
7133 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7134 This makes DSA verification about 2 % faster.
7137 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7138 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7139 plus overhead for 1024 bit moduli).
7140 This makes exponentiations about 0.5 % faster for 1024 bit
7141 exponents (as measured by "openssl speed rsa2048").
7144 *) Rename memory handling macros to avoid conflicts with other
7146 Malloc => OPENSSL_malloc
7147 Malloc_locked => OPENSSL_malloc_locked
7148 Realloc => OPENSSL_realloc
7149 Free => OPENSSL_free
7152 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7153 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7156 *) CygWin32 support.
7157 [John Jarvie <jjarvie@newsguy.com>]
7159 *) The type-safe stack code has been rejigged. It is now only compiled
7160 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7161 by default all type-specific stack functions are "#define"d back to
7162 standard stack functions. This results in more streamlined output
7163 but retains the type-safety checking possibilities of the original
7167 *) The STACK code has been cleaned up, and certain type declarations
7168 that didn't make a lot of sense have been brought in line. This has
7169 also involved a cleanup of sorts in safestack.h to more correctly
7170 map type-safe stack functions onto their plain stack counterparts.
7171 This work has also resulted in a variety of "const"ifications of
7172 lots of the code, especially "_cmp" operations which should normally
7173 be prototyped with "const" parameters anyway.
7176 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7177 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7178 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7179 where all of 'md' is used each time the PRNG is used, but 'state'
7180 is used only indexed by a cyclic counter. As entropy may not be
7181 well distributed from the beginning, 'md' is important as a
7182 chaining variable. However, the output function chains only half
7183 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7184 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7185 in all of 'state' being rewritten, with the new values depending
7186 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7189 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7190 the handshake is continued after ssl_verify_cert_chain();
7191 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7192 can lead to 'unexplainable' connection aborts later.
7193 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7195 *) Major EVP API cipher revision.
7196 Add hooks for extra EVP features. This allows various cipher
7197 parameters to be set in the EVP interface. Support added for variable
7198 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7199 setting of RC2 and RC5 parameters.
7201 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7204 Remove lots of duplicated code from the EVP library. For example *every*
7205 cipher init() function handles the 'iv' in the same way according to the
7206 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7207 for CFB and OFB modes they zero ctx->num.
7209 New functionality allows removal of S/MIME code RC2 hack.
7211 Most of the routines have the same form and so can be declared in terms
7214 By shifting this to the top level EVP_CipherInit() it can be removed from
7215 all individual ciphers. If the cipher wants to handle IVs or keys
7216 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7219 Change lots of functions like EVP_EncryptUpdate() to now return a
7220 value: although software versions of the algorithms cannot fail
7221 any installed hardware versions can.
7224 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7225 this option is set, tolerate broken clients that send the negotiated
7226 protocol version number instead of the requested protocol version
7230 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7231 i.e. non-zero for export ciphersuites, zero otherwise.
7232 Previous versions had this flag inverted, inconsistent with
7233 rsa_tmp_cb (..._TMP_RSA_CB).
7234 [Bodo Moeller; problem reported by Amit Chopra]
7236 *) Add missing DSA library text string. Work around for some IIS
7237 key files with invalid SEQUENCE encoding.
7240 *) Add a document (doc/standards.txt) that list all kinds of standards
7241 and so on that are implemented in OpenSSL.
7244 *) Enhance c_rehash script. Old version would mishandle certificates
7245 with the same subject name hash and wouldn't handle CRLs at all.
7246 Added -fingerprint option to crl utility, to support new c_rehash
7250 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7253 *) Fix for SSL server purpose checking. Server checking was
7254 rejecting certificates which had extended key usage present
7255 but no ssl client purpose.
7256 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7258 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7259 is a little unclear about how a blank password is handled.
7260 Since the password in encoded as a BMPString with terminating
7261 double NULL a zero length password would end up as just the
7262 double NULL. However no password at all is different and is
7263 handled differently in the PKCS#12 key generation code. NS
7264 treats a blank password as zero length. MSIE treats it as no
7265 password on export: but it will try both on import. We now do
7266 the same: PKCS12_parse() tries zero length and no password if
7267 the password is set to "" or NULL (NULL is now a valid password:
7268 it wasn't before) as does the pkcs12 application.
7271 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7272 perror when PEM_read_bio_X509_REQ fails, the error message must
7273 be obtained from the error queue.
7276 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7277 it in ERR_remove_state if appropriate, and change ERR_get_state
7278 accordingly to avoid race conditions (this is necessary because
7279 thread_hash is no longer constant once set).
7282 *) Bugfix for linux-elf makefile.one.
7285 *) RSA_get_default_method() will now cause a default
7286 RSA_METHOD to be chosen if one doesn't exist already.
7287 Previously this was only set during a call to RSA_new()
7288 or RSA_new_method(NULL) meaning it was possible for
7289 RSA_get_default_method() to return NULL.
7292 *) Added native name translation to the existing DSO code
7293 that will convert (if the flag to do so is set) filenames
7294 that are sufficiently small and have no path information
7295 into a canonical native form. Eg. "blah" converted to
7296 "libblah.so" or "blah.dll" etc.
7299 *) New function ERR_error_string_n(e, buf, len) which is like
7300 ERR_error_string(e, buf), but writes at most 'len' bytes
7301 including the 0 terminator. For ERR_error_string_n, 'buf'
7303 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7305 *) CONF library reworked to become more general. A new CONF
7306 configuration file reader "class" is implemented as well as a
7307 new functions (NCONF_*, for "New CONF") to handle it. The now
7308 old CONF_* functions are still there, but are reimplemented to
7309 work in terms of the new functions. Also, a set of functions
7310 to handle the internal storage of the configuration data is
7311 provided to make it easier to write new configuration file
7312 reader "classes" (I can definitely see something reading a
7313 configuration file in XML format, for example), called _CONF_*,
7314 or "the configuration storage API"...
7316 The new configuration file reading functions are:
7318 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7319 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7321 NCONF_default, NCONF_WIN32
7323 NCONF_dump_fp, NCONF_dump_bio
7325 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7326 NCONF_new creates a new CONF object. This works in the same way
7327 as other interfaces in OpenSSL, like the BIO interface.
7328 NCONF_dump_* dump the internal storage of the configuration file,
7329 which is useful for debugging. All other functions take the same
7330 arguments as the old CONF_* functions wth the exception of the
7331 first that must be a `CONF *' instead of a `LHASH *'.
7333 To make it easer to use the new classes with the old CONF_* functions,
7334 the function CONF_set_default_method is provided.
7337 *) Add '-tls1' option to 'openssl ciphers', which was already
7338 mentioned in the documentation but had not been implemented.
7339 (This option is not yet really useful because even the additional
7340 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7343 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7344 OpenSSL-based applications) load shared libraries and bind to
7345 them in a portable way.
7346 [Geoff Thorpe, with contributions from Richard Levitte]
7348 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7350 *) Make sure _lrotl and _lrotr are only used with MSVC.
7352 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7353 (the default implementation of RAND_status).
7355 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7356 to '-clrext' (= clear extensions), as intended and documented.
7357 [Bodo Moeller; inconsistency pointed out by Michael Attili
7360 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7361 was larger than the MD block size.
7362 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7364 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7365 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7366 using the passed key: if the passed key was a private key the result
7367 of X509_print(), for example, would be to print out all the private key
7371 *) des_quad_cksum() byte order bug fix.
7372 [Ulf Möller, using the problem description in krb4-0.9.7, where
7373 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7375 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7377 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7379 *) For easily testing in shell scripts whether some command
7380 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7381 returns with exit code 0 iff no command of the given name is available.
7382 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7383 the output goes to stdout and nothing is printed to stderr.
7384 Additional arguments are always ignored.
7386 Since for each cipher there is a command of the same name,
7387 the 'no-cipher' compilation switches can be tested this way.
7389 ('openssl no-XXX' is not able to detect pseudo-commands such
7390 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7393 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7396 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7397 is set; it will be thrown away anyway because each handshake creates
7399 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7400 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7401 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7402 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7405 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7406 'Q' and 'R' lose their special meanings (quit/renegotiate).
7407 This is part of what -quiet does; unlike -quiet, -ign_eof
7408 does not suppress any output.
7411 *) Add compatibility options to the purpose and trust code. The
7412 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7413 accepts a certificate or CA, this was the previous behaviour,
7414 with all the associated security issues.
7416 X509_TRUST_COMPAT is the old trust behaviour: only and
7417 automatically trust self signed roots in certificate store. A
7418 new trust setting X509_TRUST_DEFAULT is used to specify that
7419 a purpose has no associated trust setting and it should instead
7420 use the value in the default purpose.
7423 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7424 and fix a memory leak.
7427 *) In util/mkerr.pl (which implements 'make errors'), preserve
7428 reason strings from the previous version of the .c file, as
7429 the default to have only downcase letters (and digits) in
7430 automatically generated reasons codes is not always appropriate.
7433 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7434 using strerror. Previously, ERR_reason_error_string() returned
7435 library names as reason strings for SYSerr; but SYSerr is a special
7436 case where small numbers are errno values, not library numbers.
7439 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7440 converts DSA parameters into DH parameters. (When creating parameters,
7441 DSA_generate_parameters is used.)
7444 *) Include 'length' (recommended exponent length) in C code generated
7445 by 'openssl dhparam -C'.
7448 *) The second argument to set_label in perlasm was already being used
7449 so couldn't be used as a "file scope" flag. Moved to third argument
7453 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7454 instead of RAND_bytes for encryption IVs and salts.
7457 *) Include RAND_status() into RAND_METHOD instead of implementing
7458 it only for md_rand.c Otherwise replacing the PRNG by calling
7459 RAND_set_rand_method would be impossible.
7462 *) Don't let DSA_generate_key() enter an infinite loop if the random
7463 number generation fails.
7466 *) New 'rand' application for creating pseudo-random output.
7469 *) Added configuration support for Linux/IA64
7470 [Rolf Haberrecker <rolf@suse.de>]
7472 *) Assembler module support for Mingw32.
7475 *) Shared library support for HPUX (in shlib/).
7476 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7478 *) Shared library support for Solaris gcc.
7479 [Lutz Behnke <behnke@trustcenter.de>]
7481 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7483 *) PKCS7_encrypt() was adding text MIME headers twice because they
7484 were added manually and by SMIME_crlf_copy().
7487 *) In bntest.c don't call BN_rand with zero bits argument.
7488 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7490 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7491 case was implemented. This caused BN_div_recp() to fail occasionally.
7494 *) Add an optional second argument to the set_label() in the perl
7495 assembly language builder. If this argument exists and is set
7496 to 1 it signals that the assembler should use a symbol whose
7497 scope is the entire file, not just the current function. This
7498 is needed with MASM which uses the format label:: for this scope.
7499 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7501 *) Change the ASN1 types so they are typedefs by default. Before
7502 almost all types were #define'd to ASN1_STRING which was causing
7503 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7507 *) Change names of new functions to the new get1/get0 naming
7508 convention: After 'get1', the caller owns a reference count
7509 and has to call ..._free; 'get0' returns a pointer to some
7510 data structure without incrementing reference counters.
7511 (Some of the existing 'get' functions increment a reference
7512 counter, some don't.)
7513 Similarly, 'set1' and 'add1' functions increase reference
7514 counters or duplicate objects.
7517 *) Allow for the possibility of temp RSA key generation failure:
7518 the code used to assume it always worked and crashed on failure.
7521 *) Fix potential buffer overrun problem in BIO_printf().
7522 [Ulf Möller, using public domain code by Patrick Powell; problem
7523 pointed out by David Sacerdote <das33@cornell.edu>]
7525 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7526 RAND_egd() and RAND_status(). In the command line application,
7527 the EGD socket can be specified like a seed file using RANDFILE
7531 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7532 Some CAs (e.g. Verisign) distribute certificates in this form.
7535 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7536 list to exclude them. This means that no special compilation option
7537 is needed to use anonymous DH: it just needs to be included in the
7541 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7542 EVP_MD_type. The old functionality is available in a new macro called
7543 EVP_MD_md(). Change code that uses it and update docs.
7546 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7547 where the 'void *' argument is replaced by a function pointer argument.
7548 Previously 'void *' was abused to point to functions, which works on
7549 many platforms, but is not correct. As these functions are usually
7550 called by macros defined in OpenSSL header files, most source code
7551 should work without changes.
7554 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7555 sections with information on -D... compiler switches used for
7556 compiling the library so that applications can see them. To enable
7557 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7558 must be defined. E.g.,
7559 #define OPENSSL_ALGORITHM_DEFINES
7560 #include <openssl/opensslconf.h>
7561 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7562 [Richard Levitte, Ulf and Bodo Möller]
7564 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7568 *) Change the 'other' type in certificate aux info to a STACK_OF
7569 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7570 the required ASN1 format: arbitrary types determined by an OID.
7573 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7574 argument to 'req'. This is not because the function is newer or
7575 better than others it just uses the work 'NEW' in the certificate
7576 request header lines. Some software needs this.
7579 *) Reorganise password command line arguments: now passwords can be
7580 obtained from various sources. Delete the PEM_cb function and make
7581 it the default behaviour: i.e. if the callback is NULL and the
7582 usrdata argument is not NULL interpret it as a null terminated pass
7583 phrase. If usrdata and the callback are NULL then the pass phrase
7584 is prompted for as usual.
7587 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7588 the support is automatically enabled. The resulting binaries will
7589 autodetect the card and use it if present.
7590 [Ben Laurie and Compaq Inc.]
7592 *) Work around for Netscape hang bug. This sends certificate request
7593 and server done in one record. Since this is perfectly legal in the
7594 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7595 the bugs/SSLv3 entry for more info.
7598 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7601 *) Add -rand argument to smime and pkcs12 applications and read/write
7605 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7608 *) Add command line password options to the remaining applications.
7611 *) Bug fix for BN_div_recp() for numerators with an even number of
7615 *) More tests in bntest.c, and changed test_bn output.
7618 *) ./config recognizes MacOS X now.
7621 *) Bug fix for BN_div() when the first words of num and divsor are
7622 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7625 *) Add support for various broken PKCS#8 formats, and command line
7626 options to produce them.
7629 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7630 get temporary BIGNUMs from a BN_CTX.
7633 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7637 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7638 include a #define from the old name to the new. The original intent
7639 was that statically linked binaries could for example just call
7640 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7641 link with digests. This never worked becayse SSLeay_add_all_digests()
7642 and SSLeay_add_all_ciphers() were in the same source file so calling
7643 one would link with the other. They are now in separate source files.
7646 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7649 *) Use a less unusual form of the Miller-Rabin primality test (it used
7650 a binary algorithm for exponentiation integrated into the Miller-Rabin
7651 loop, our standard modexp algorithms are faster).
7654 *) Support for the EBCDIC character set completed.
7655 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7657 *) Source code cleanups: use const where appropriate, eliminate casts,
7658 use void * instead of char * in lhash.
7661 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7662 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7663 this the server could overwrite ephemeral keys that the client
7667 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7668 using 50 iterations of the Rabin-Miller test.
7670 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7671 iterations of the Rabin-Miller test as required by the appendix
7672 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7673 As BN_is_prime_fasttest includes trial division, DSA parameter
7674 generation becomes much faster.
7676 This implies a change for the callback functions in DSA_is_prime
7677 and DSA_generate_parameters: The callback function is called once
7678 for each positive witness in the Rabin-Miller test, not just
7679 occasionally in the inner loop; and the parameters to the
7680 callback function now provide an iteration count for the outer
7681 loop rather than for the current invocation of the inner loop.
7682 DSA_generate_parameters additionally can call the callback
7683 function with an 'iteration count' of -1, meaning that a
7684 candidate has passed the trial division test (when q is generated
7685 from an application-provided seed, trial division is skipped).
7688 *) New function BN_is_prime_fasttest that optionally does trial
7689 division before starting the Rabin-Miller test and has
7690 an additional BN_CTX * argument (whereas BN_is_prime always
7691 has to allocate at least one BN_CTX).
7692 'callback(1, -1, cb_arg)' is called when a number has passed the
7693 trial division stage.
7696 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7700 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7703 *) New function BN_pseudo_rand().
7706 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7707 bignum version of BN_from_montgomery() with the working code from
7708 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7712 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7713 made it impossible to use the same SSL_SESSION data structure in
7714 SSL2 clients in multiple threads.
7717 *) The return value of RAND_load_file() no longer counts bytes obtained
7718 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7719 to seed the PRNG (previously an explicit byte count was required).
7720 [Ulf Möller, Bodo Möller]
7722 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7723 used (char *) instead of (void *) and had casts all over the place.
7726 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7729 *) Retain source code compatibility for BN_prime_checks macro:
7730 BN_is_prime(..., BN_prime_checks, ...) now uses
7731 BN_prime_checks_for_size to determine the appropriate number of
7732 Rabin-Miller iterations.
7735 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7736 DH_CHECK_P_NOT_SAFE_PRIME.
7737 (Check if this is true? OpenPGP calls them "strong".)
7740 *) Merge the functionality of "dh" and "gendh" programs into a new program
7741 "dhparam". The old programs are retained for now but will handle DH keys
7742 (instead of parameters) in future.
7745 *) Make the ciphers, s_server and s_client programs check the return values
7746 when a new cipher list is set.
7749 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7750 ciphers. Before when the 56bit ciphers were enabled the sorting was
7753 The syntax for the cipher sorting has been extended to support sorting by
7754 cipher-strength (using the strength_bits hard coded in the tables).
7755 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7757 Fix a bug in the cipher-command parser: when supplying a cipher command
7758 string with an "undefined" symbol (neither command nor alphanumeric
7759 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7760 an error is flagged.
7762 Due to the strength-sorting extension, the code of the
7763 ssl_create_cipher_list() function was completely rearranged. I hope that
7764 the readability was also increased :-)
7765 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7767 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7768 for the first serial number and places 2 in the serial number file. This
7769 avoids problems when the root CA is created with serial number zero and
7770 the first user certificate has the same issuer name and serial number
7774 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7775 the new code. Add documentation for this stuff.
7778 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7779 X509_*() to X509at_*() on the grounds that they don't handle X509
7780 structures and behave in an analagous way to the X509v3 functions:
7781 they shouldn't be called directly but wrapper functions should be used
7784 So we also now have some wrapper functions that call the X509at functions
7785 when passed certificate requests. (TO DO: similar things can be done with
7786 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7787 things. Some of these need some d2i or i2d and print functionality
7788 because they handle more complex structures.)
7791 *) Add missing #ifndefs that caused missing symbols when building libssl
7792 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7793 NO_RSA in ssl/s2*.c.
7794 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7796 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7797 has a return value which indicates the quality of the random data
7798 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7799 error queue. New function RAND_pseudo_bytes() generates output that is
7800 guaranteed to be unique but not unpredictable. RAND_add is like
7801 RAND_seed, but takes an extra argument for an entropy estimate
7802 (RAND_seed always assumes full entropy).
7805 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7806 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7807 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7808 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7809 false-positive rate of at most 2^-80 for random input.
7812 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7815 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7816 in the 0.9.5 release), this returns the chain
7817 from an X509_CTX structure with a dup of the stack and all
7818 the X509 reference counts upped: so the stack will exist
7819 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7822 Also make SSL_SESSION_print() print out the verify return
7826 *) Add manpage for the pkcs12 command. Also change the default
7827 behaviour so MAC iteration counts are used unless the new
7828 -nomaciter option is used. This improves file security and
7829 only older versions of MSIE (4.0 for example) need it.
7832 *) Honor the no-xxx Configure options when creating .DEF files.
7835 *) Add PKCS#10 attributes to field table: challengePassword,
7836 unstructuredName and unstructuredAddress. These are taken from
7837 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7838 international characters are used.
7840 More changes to X509_ATTRIBUTE code: allow the setting of types
7841 based on strings. Remove the 'loc' parameter when adding
7842 attributes because these will be a SET OF encoding which is sorted
7846 *) Initial changes to the 'req' utility to allow request generation
7847 automation. This will allow an application to just generate a template
7848 file containing all the field values and have req construct the
7851 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7852 used all over the place including certificate requests and PKCS#7
7853 structures. They are currently handled manually where necessary with
7854 some primitive wrappers for PKCS#7. The new functions behave in a
7855 manner analogous to the X509 extension functions: they allow
7856 attributes to be looked up by NID and added.
7858 Later something similar to the X509V3 code would be desirable to
7859 automatically handle the encoding, decoding and printing of the
7860 more complex types. The string types like challengePassword can
7861 be handled by the string table functions.
7863 Also modified the multi byte string table handling. Now there is
7864 a 'global mask' which masks out certain types. The table itself
7865 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7866 is useful when for example there is only one permissible type
7867 (as in countryName) and using the mask might result in no valid
7871 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7872 SSL_get_peer_finished to allow applications to obtain the latest
7873 Finished messages sent to the peer or expected from the peer,
7874 respectively. (SSL_get_peer_finished is usually the Finished message
7875 actually received from the peer, otherwise the protocol will be aborted.)
7877 As the Finished message are message digests of the complete handshake
7878 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7879 be used for external authentication procedures when the authentication
7880 provided by SSL/TLS is not desired or is not enough.
7883 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7884 the host supports BWX extension and if Compaq C is present on the
7885 $PATH. Just exploiting of the BWX extension results in 20-30%
7886 performance kick for some algorithms, e.g. DES and RC4 to mention
7887 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7891 *) Add support for MS "fast SGC". This is arguably a violation of the
7892 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7893 weak crypto and after checking the certificate is SGC a second one
7894 with strong crypto. MS SGC stops the first handshake after receiving
7895 the server certificate message and sends a second client hello. Since
7896 a server will typically do all the time consuming operations before
7897 expecting any further messages from the client (server key exchange
7898 is the most expensive) there is little difference between the two.
7900 To get OpenSSL to support MS SGC we have to permit a second client
7901 hello message after we have sent server done. In addition we have to
7902 reset the MAC if we do get this second client hello.
7905 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7906 if a DER encoded private key is RSA or DSA traditional format. Changed
7907 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7908 format DER encoded private key. Newer code should use PKCS#8 format which
7909 has the key type encoded in the ASN1 structure. Added DER private key
7910 support to pkcs8 application.
7913 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7914 ciphersuites has been selected (as required by the SSL 3/TLS 1
7915 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7916 is set, we interpret this as a request to violate the specification
7917 (the worst that can happen is a handshake failure, and 'correct'
7918 behaviour would result in a handshake failure anyway).
7921 *) In SSL_CTX_add_session, take into account that there might be multiple
7922 SSL_SESSION structures with the same session ID (e.g. when two threads
7923 concurrently obtain them from an external cache).
7924 The internal cache can handle only one SSL_SESSION with a given ID,
7925 so if there's a conflict, we now throw out the old one to achieve
7929 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7930 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7931 some routines that use cipher OIDs: some ciphers do not have OIDs
7932 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7936 *) Simplify the trust setting structure and code. Now we just have
7937 two sequences of OIDs for trusted and rejected settings. These will
7938 typically have values the same as the extended key usage extension
7939 and any application specific purposes.
7941 The trust checking code now has a default behaviour: it will just
7942 check for an object with the same NID as the passed id. Functions can
7943 be provided to override either the default behaviour or the behaviour
7944 for a given id. SSL client, server and email already have functions
7945 in place for compatibility: they check the NID and also return "trusted"
7946 if the certificate is self signed.
7949 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7950 traditional format into an EVP_PKEY structure.
7953 *) Add a password callback function PEM_cb() which either prompts for
7954 a password if usr_data is NULL or otherwise assumes it is a null
7955 terminated password. Allow passwords to be passed on command line
7956 environment or config files in a few more utilities.
7959 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7960 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7961 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7962 Update documentation.
7965 *) Support for ASN1 "NULL" type. This could be handled before by using
7966 ASN1_TYPE but there wasn't any function that would try to read a NULL
7967 and produce an error if it couldn't. For compatibility we also have
7968 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7969 don't allocate anything because they don't need to.
7972 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7974 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7976 *) Rebuild of the memory allocation routines used by OpenSSL code and
7977 possibly others as well. The purpose is to make an interface that
7978 provide hooks so anyone can build a separate set of allocation and
7979 deallocation routines to be used by OpenSSL, for example memory
7980 pool implementations, or something else, which was previously hard
7981 since Malloc(), Realloc() and Free() were defined as macros having
7982 the values malloc, realloc and free, respectively (except for Win32
7983 compilations). The same is provided for memory debugging code.
7984 OpenSSL already comes with functionality to find memory leaks, but
7985 this gives people a chance to debug other memory problems.
7987 With these changes, a new set of functions and macros have appeared:
7989 CRYPTO_set_mem_debug_functions() [F]
7990 CRYPTO_get_mem_debug_functions() [F]
7991 CRYPTO_dbg_set_options() [F]
7992 CRYPTO_dbg_get_options() [F]
7993 CRYPTO_malloc_debug_init() [M]
7995 The memory debug functions are NULL by default, unless the library
7996 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7997 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7998 gives the standard debugging functions that come with OpenSSL) or
7999 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8000 provided by the library user) must be used. When the standard
8001 debugging functions are used, CRYPTO_dbg_set_options can be used to
8002 request additional information:
8003 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8004 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8006 Also, things like CRYPTO_set_mem_functions will always give the
8007 expected result (the new set of functions is used for allocation
8008 and deallocation) at all times, regardless of platform and compiler
8011 To finish it up, some functions that were never use in any other
8012 way than through macros have a new API and new semantic:
8015 CRYPTO_dbg_realloc()
8018 All macros of value have retained their old syntax.
8019 [Richard Levitte and Bodo Moeller]
8021 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8022 ordering of SMIMECapabilities wasn't in "strength order" and there
8023 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8027 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8028 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8029 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8031 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8032 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8033 functionality to handle multipart/signed properly) and a utility
8034 called 'smime' to call all this stuff. This is based on code I
8035 originally wrote for Celo who have kindly allowed it to be
8036 included in OpenSSL.
8039 *) Add variants des_set_key_checked and des_set_key_unchecked of
8040 des_set_key (aka des_key_sched). Global variable des_check_key
8041 decides which of these is called by des_set_key; this way
8042 des_check_key behaves as it always did, but applications and
8043 the library itself, which was buggy for des_check_key == 1,
8044 have a cleaner way to pick the version they need.
8047 *) New function PKCS12_newpass() which changes the password of a
8051 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8052 dynamic mix. In both cases the ids can be used as an index into the
8053 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8054 functions so they accept a list of the field values and the
8055 application doesn't need to directly manipulate the X509_TRUST
8059 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8063 *) Modify the way the V3 extension code looks up extensions. This now
8064 works in a similar way to the object code: we have some "standard"
8065 extensions in a static table which is searched with OBJ_bsearch()
8066 and the application can add dynamic ones if needed. The file
8067 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8068 updated whenever a new extension is added to the core code and kept
8069 in ext_nid order. There is a simple program 'tabtest.c' which checks
8070 this. New extensions are not added too often so this file can readily
8071 be maintained manually.
8073 There are two big advantages in doing things this way. The extensions
8074 can be looked up immediately and no longer need to be "added" using
8075 X509V3_add_standard_extensions(): this function now does nothing.
8076 [Side note: I get *lots* of email saying the extension code doesn't
8077 work because people forget to call this function]
8078 Also no dynamic allocation is done unless new extensions are added:
8079 so if we don't add custom extensions there is no need to call
8080 X509V3_EXT_cleanup().
8083 *) Modify enc utility's salting as follows: make salting the default. Add a
8084 magic header, so unsalted files fail gracefully instead of just decrypting
8085 to garbage. This is because not salting is a big security hole, so people
8086 should be discouraged from doing it.
8089 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8090 digest to be passed on the command line but it only used this
8091 parameter when signing a certificate. Modified so all relevant
8092 operations are affected by the digest parameter including the
8093 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8094 DSA key was used because it didn't fix the digest.
8097 *) Initial certificate chain verify code. Currently tests the untrusted
8098 certificates for consistency with the verify purpose (which is set
8099 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8101 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8102 this is because it will reject chains with invalid extensions whereas
8103 every previous version of OpenSSL and SSLeay made no checks at all.
8105 Trust code: checks the root CA for the relevant trust settings. Trust
8106 settings have an initial value consistent with the verify purpose: e.g.
8107 if the verify purpose is for SSL client use it expects the CA to be
8108 trusted for SSL client use. However the default value can be changed to
8109 permit custom trust settings: one example of this would be to only trust
8110 certificates from a specific "secure" set of CAs.
8112 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8113 which should be used for version portability: especially since the
8114 verify structure is likely to change more often now.
8116 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8117 to set them. If not set then assume SSL clients will verify SSL servers
8120 Two new options to the verify program: -untrusted allows a set of
8121 untrusted certificates to be passed in and -purpose which sets the
8122 intended purpose of the certificate. If a purpose is set then the
8123 new chain verify code is used to check extension consistency.
8126 *) Support for the authority information access extension.
8129 *) Modify RSA and DSA PEM read routines to transparently handle
8130 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8131 public keys in a format compatible with certificate
8132 SubjectPublicKeyInfo structures. Unfortunately there were already
8133 functions called *_PublicKey_* which used various odd formats so
8134 these are retained for compatibility: however the DSA variants were
8135 never in a public release so they have been deleted. Changed dsa/rsa
8136 utilities to handle the new format: note no releases ever handled public
8137 keys so we should be OK.
8139 The primary motivation for this change is to avoid the same fiasco
8140 that dogs private keys: there are several incompatible private key
8141 formats some of which are standard and some OpenSSL specific and
8142 require various evil hacks to allow partial transparent handling and
8143 even then it doesn't work with DER formats. Given the option anything
8144 other than PKCS#8 should be dumped: but the other formats have to
8145 stay in the name of compatibility.
8147 With public keys and the benefit of hindsight one standard format
8148 is used which works with EVP_PKEY, RSA or DSA structures: though
8149 it clearly returns an error if you try to read the wrong kind of key.
8151 Added a -pubkey option to the 'x509' utility to output the public key.
8152 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8153 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8154 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8155 that do the same as the EVP_PKEY_assign_*() except they up the
8156 reference count of the added key (they don't "swallow" the
8160 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8161 CRLs would fail if the file contained no certificates or no CRLs:
8162 added a new function to read in both types and return the number
8163 read: this means that if none are read it will be an error. The
8164 DER versions of the certificate and CRL reader would always fail
8165 because it isn't possible to mix certificates and CRLs in DER format
8166 without choking one or the other routine. Changed this to just read
8167 a certificate: this is the best we can do. Also modified the code
8168 in apps/verify.c to take notice of return codes: it was previously
8169 attempting to read in certificates from NULL pointers and ignoring
8170 any errors: this is one reason why the cert and CRL reader seemed
8171 to work. It doesn't check return codes from the default certificate
8172 routines: these may well fail if the certificates aren't installed.
8175 *) Code to support otherName option in GeneralName.
8178 *) First update to verify code. Change the verify utility
8179 so it warns if it is passed a self signed certificate:
8180 for consistency with the normal behaviour. X509_verify
8181 has been modified to it will now verify a self signed
8182 certificate if *exactly* the same certificate appears
8183 in the store: it was previously impossible to trust a
8184 single self signed certificate. This means that:
8185 openssl verify ss.pem
8186 now gives a warning about a self signed certificate but
8187 openssl verify -CAfile ss.pem ss.pem
8191 *) For servers, store verify_result in SSL_SESSION data structure
8192 (and add it to external session representation).
8193 This is needed when client certificate verifications fails,
8194 but an application-provided verification callback (set by
8195 SSL_CTX_set_cert_verify_callback) allows accepting the session
8196 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8197 but returns 1): When the session is reused, we have to set
8198 ssl->verify_result to the appropriate error code to avoid
8200 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8202 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8203 case in PKCS7_dataInit() where the signed PKCS7 structure
8204 didn't contain any existing data because it was being created.
8205 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8207 *) Add a salt to the key derivation routines in enc.c. This
8208 forms the first 8 bytes of the encrypted file. Also add a
8209 -S option to allow a salt to be input on the command line.
8212 *) New function X509_cmp(). Oddly enough there wasn't a function
8213 to compare two certificates. We do this by working out the SHA1
8214 hash and comparing that. X509_cmp() will be needed by the trust
8218 *) SSL_get1_session() is like SSL_get_session(), but increments
8219 the reference count in the SSL_SESSION returned.
8220 [Geoff Thorpe <geoff@eu.c2.net>]
8222 *) Fix for 'req': it was adding a null to request attributes.
8223 Also change the X509_LOOKUP and X509_INFO code to handle
8224 certificate auxiliary information.
8227 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8231 *) Add the possibility to add extra information to the memory leak
8232 detecting output, to form tracebacks, showing from where each
8233 allocation was originated: CRYPTO_push_info("constant string") adds
8234 the string plus current file name and line number to a per-thread
8235 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8236 is like calling CYRPTO_pop_info() until the stack is empty.
8237 Also updated memory leak detection code to be multi-thread-safe.
8240 *) Add options -text and -noout to pkcs7 utility and delete the
8241 encryption options which never did anything. Update docs.
8244 *) Add options to some of the utilities to allow the pass phrase
8245 to be included on either the command line (not recommended on
8246 OSes like Unix) or read from the environment. Update the
8247 manpages and fix a few bugs.
8250 *) Add a few manpages for some of the openssl commands.
8253 *) Fix the -revoke option in ca. It was freeing up memory twice,
8254 leaking and not finding already revoked certificates.
8257 *) Extensive changes to support certificate auxiliary information.
8258 This involves the use of X509_CERT_AUX structure and X509_AUX
8259 functions. An X509_AUX function such as PEM_read_X509_AUX()
8260 can still read in a certificate file in the usual way but it
8261 will also read in any additional "auxiliary information". By
8262 doing things this way a fair degree of compatibility can be
8263 retained: existing certificates can have this information added
8264 using the new 'x509' options.
8266 Current auxiliary information includes an "alias" and some trust
8267 settings. The trust settings will ultimately be used in enhanced
8268 certificate chain verification routines: currently a certificate
8269 can only be trusted if it is self signed and then it is trusted
8273 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8274 The problem was that one of the replacement routines had not been working
8275 since SSLeay releases. For now the offending routine has been replaced
8276 with non-optimised assembler. Even so, this now gives around 95%
8277 performance improvement for 1024 bit RSA signs.
8280 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8281 handling. Most clients have the effective key size in bits equal to
8282 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8283 A few however don't do this and instead use the size of the decrypted key
8284 to determine the RC2 key length and the AlgorithmIdentifier to determine
8285 the effective key length. In this case the effective key length can still
8286 be 40 bits but the key length can be 168 bits for example. This is fixed
8287 by manually forcing an RC2 key into the EVP_PKEY structure because the
8288 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8289 the key length and effective key length are equal.
8292 *) Add a bunch of functions that should simplify the creation of
8293 X509_NAME structures. Now you should be able to do:
8294 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8295 and have it automatically work out the correct field type and fill in
8296 the structures. The more adventurous can try:
8297 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8298 and it will (hopefully) work out the correct multibyte encoding.
8301 *) Change the 'req' utility to use the new field handling and multibyte
8302 copy routines. Before the DN field creation was handled in an ad hoc
8303 way in req, ca, and x509 which was rather broken and didn't support
8304 BMPStrings or UTF8Strings. Since some software doesn't implement
8305 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8306 using the dirstring_type option. See the new comment in the default
8307 openssl.cnf for more info.
8310 *) Make crypto/rand/md_rand.c more robust:
8311 - Assure unique random numbers after fork().
8312 - Make sure that concurrent threads access the global counter and
8313 md serializably so that we never lose entropy in them
8314 or use exactly the same state in multiple threads.
8315 Access to the large state is not always serializable because
8316 the additional locking could be a performance killer, and
8317 md should be large enough anyway.
8320 *) New file apps/app_rand.c with commonly needed functionality
8321 for handling the random seed file.
8323 Use the random seed file in some applications that previously did not:
8325 dsaparam -genkey (which also ignored its '-rand' option),
8328 x509 (when signing).
8329 Except on systems with /dev/urandom, it is crucial to have a random
8330 seed file at least for key creation, DSA signing, and for DH exchanges;
8331 for RSA signatures we could do without one.
8333 gendh and gendsa (unlike genrsa) used to read only the first byte
8334 of each file listed in the '-rand' option. The function as previously
8335 found in genrsa is now in app_rand.c and is used by all programs
8336 that support '-rand'.
8339 *) In RAND_write_file, use mode 0600 for creating files;
8340 don't just chmod when it may be too late.
8343 *) Report an error from X509_STORE_load_locations
8344 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8347 *) New function ASN1_mbstring_copy() this copies a string in either
8348 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8349 into an ASN1_STRING type. A mask of permissible types is passed
8350 and it chooses the "minimal" type to use or an error if not type
8354 *) Add function equivalents to the various macros in asn1.h. The old
8355 macros are retained with an M_ prefix. Code inside the library can
8356 use the M_ macros. External code (including the openssl utility)
8357 should *NOT* in order to be "shared library friendly".
8360 *) Add various functions that can check a certificate's extensions
8361 to see if it usable for various purposes such as SSL client,
8362 server or S/MIME and CAs of these types. This is currently
8363 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8364 verification. Also added a -purpose flag to x509 utility to
8365 print out all the purposes.
8368 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8372 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8373 for, obtain and decode and extension and obtain its critical flag.
8374 This allows all the necessary extension code to be handled in a
8375 single function call.
8378 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8379 platforms. See crypto/rc4/rc4_enc.c for further details.
8382 *) New -noout option to asn1parse. This causes no output to be produced
8383 its main use is when combined with -strparse and -out to extract data
8384 from a file (which may not be in ASN.1 format).
8387 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8388 when producing the local key id.
8389 [Richard Levitte <levitte@stacken.kth.se>]
8391 *) New option -dhparam in s_server. This allows a DH parameter file to be
8392 stated explicitly. If it is not stated then it tries the first server
8393 certificate file. The previous behaviour hard coded the filename
8397 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8398 a public key to be input or output. For example:
8399 openssl rsa -in key.pem -pubout -out pubkey.pem
8400 Also added necessary DSA public key functions to handle this.
8403 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8404 in the message. This was handled by allowing
8405 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8406 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8408 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8409 to the end of the strings whereas this didn't. This would cause problems
8410 if strings read with d2i_ASN1_bytes() were later modified.
8411 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8413 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8414 data and it contains EOF it will end up returning an error. This is
8415 caused by input 46 bytes long. The cause is due to the way base64
8416 BIOs find the start of base64 encoded data. They do this by trying a
8417 trial decode on each line until they find one that works. When they
8418 do a flag is set and it starts again knowing it can pass all the
8419 data directly through the decoder. Unfortunately it doesn't reset
8420 the context it uses. This means that if EOF is reached an attempt
8421 is made to pass two EOFs through the context and this causes the
8422 resulting error. This can also cause other problems as well. As is
8423 usual with these problems it takes *ages* to find and the fix is
8424 trivial: move one line.
8425 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8427 *) Ugly workaround to get s_client and s_server working under Windows. The
8428 old code wouldn't work because it needed to select() on sockets and the
8429 tty (for keypresses and to see if data could be written). Win32 only
8430 supports select() on sockets so we select() with a 1s timeout on the
8431 sockets and then see if any characters are waiting to be read, if none
8432 are present then we retry, we also assume we can always write data to
8433 the tty. This isn't nice because the code then blocks until we've
8434 received a complete line of data and it is effectively polling the
8435 keyboard at 1s intervals: however it's quite a bit better than not
8436 working at all :-) A dedicated Windows application might handle this
8437 with an event loop for example.
8440 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8441 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8442 will be called when RSA_sign() and RSA_verify() are used. This is useful
8443 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8444 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8445 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8446 This necessitated the support of an extra signature type NID_md5_sha1
8447 for SSL signatures and modifications to the SSL library to use it instead
8448 of calling RSA_public_decrypt() and RSA_private_encrypt().
8451 *) Add new -verify -CAfile and -CApath options to the crl program, these
8452 will lookup a CRL issuers certificate and verify the signature in a
8453 similar way to the verify program. Tidy up the crl program so it
8454 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8455 less strict. It will now permit CRL extensions even if it is not
8456 a V2 CRL: this will allow it to tolerate some broken CRLs.
8459 *) Initialize all non-automatic variables each time one of the openssl
8460 sub-programs is started (this is necessary as they may be started
8461 multiple times from the "OpenSSL>" prompt).
8462 [Lennart Bang, Bodo Moeller]
8464 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8465 removing all other RSA functionality (this is what NO_RSA does). This
8466 is so (for example) those in the US can disable those operations covered
8467 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8471 *) Non-copying interface to BIO pairs.
8472 (still largely untested)
8475 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8476 ASCII string. This was handled independently in various places before.
8479 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8480 UTF8 strings a character at a time.
8483 *) Use client_version from client hello to select the protocol
8484 (s23_srvr.c) and for RSA client key exchange verification
8485 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8488 *) Add various utility functions to handle SPKACs, these were previously
8489 handled by poking round in the structure internals. Added new function
8490 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8491 print, verify and generate SPKACs. Based on an original idea from
8492 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8495 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8498 *) Allow the config file extension section to be overwritten on the
8499 command line. Based on an original idea from Massimiliano Pala
8500 <madwolf@comune.modena.it>. The new option is called -extensions
8501 and can be applied to ca, req and x509. Also -reqexts to override
8502 the request extensions in req and -crlexts to override the crl extensions
8506 *) Add new feature to the SPKAC handling in ca. Now you can include
8507 the same field multiple times by preceding it by "XXXX." for example:
8510 this is the same syntax as used in the req config file.
8513 *) Allow certificate extensions to be added to certificate requests. These
8514 are specified in a 'req_extensions' option of the req section of the
8515 config file. They can be printed out with the -text option to req but
8516 are otherwise ignored at present.
8519 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8520 data read consists of only the final block it would not decrypted because
8521 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8522 A misplaced 'break' also meant the decrypted final block might not be
8523 copied until the next read.
8526 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8527 a few extra parameters to the DH structure: these will be useful if
8528 for example we want the value of 'q' or implement X9.42 DH.
8531 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8532 provides hooks that allow the default DSA functions or functions on a
8533 "per key" basis to be replaced. This allows hardware acceleration and
8534 hardware key storage to be handled without major modification to the
8535 library. Also added low level modexp hooks and CRYPTO_EX structure and
8536 associated functions.
8539 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8540 as "read only": it can't be written to and the buffer it points to will
8541 not be freed. Reading from a read only BIO is much more efficient than
8542 a normal memory BIO. This was added because there are several times when
8543 an area of memory needs to be read from a BIO. The previous method was
8544 to create a memory BIO and write the data to it, this results in two
8545 copies of the data and an O(n^2) reading algorithm. There is a new
8546 function BIO_new_mem_buf() which creates a read only memory BIO from
8547 an area of memory. Also modified the PKCS#7 routines to use read only
8551 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8552 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8553 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8554 but a retry condition occured while trying to read the rest.
8557 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8558 NID_pkcs7_encrypted by default: this was wrong since this should almost
8559 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8560 the encrypted data type: this is a more sensible place to put it and it
8561 allows the PKCS#12 code to be tidied up that duplicated this
8565 *) Changed obj_dat.pl script so it takes its input and output files on
8566 the command line. This should avoid shell escape redirection problems
8570 *) Initial support for certificate extension requests, these are included
8571 in things like Xenroll certificate requests. Included functions to allow
8572 extensions to be obtained and added.
8575 *) -crlf option to s_client and s_server for sending newlines as
8576 CRLF (as required by many protocols).
8579 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8581 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8582 [Ralf S. Engelschall]
8584 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8585 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8587 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8591 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8592 DH parameters/keys (q is lost during that conversion, but the resulting
8593 DH parameters contain its length).
8595 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8596 much faster than DH_generate_parameters (which creates parameters
8597 where p = 2*q + 1), and also the smaller q makes DH computations
8598 much more efficient (160-bit exponentiation instead of 1024-bit
8599 exponentiation); so this provides a convenient way to support DHE
8600 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8601 utter importance to use
8602 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8604 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8605 when such DH parameters are used, because otherwise small subgroup
8606 attacks may become possible!
8609 *) Avoid memory leak in i2d_DHparams.
8612 *) Allow the -k option to be used more than once in the enc program:
8613 this allows the same encrypted message to be read by multiple recipients.
8616 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8617 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8618 it will always use the numerical form of the OID, even if it has a short
8622 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8623 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8624 otherwise bn_mod_exp was called. In the case of hardware keys for example
8625 no private key components need be present and it might store extra data
8626 in the RSA structure, which cannot be accessed from bn_mod_exp.
8627 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8628 private key operations.
8631 *) Added support for SPARC Linux.
8634 *) pem_password_cb function type incompatibly changed from
8635 typedef int pem_password_cb(char *buf, int size, int rwflag);
8637 ....(char *buf, int size, int rwflag, void *userdata);
8638 so that applications can pass data to their callbacks:
8639 The PEM[_ASN1]_{read,write}... functions and macros now take an
8640 additional void * argument, which is just handed through whenever
8641 the password callback is called.
8642 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8644 New function SSL_CTX_set_default_passwd_cb_userdata.
8646 Compatibility note: As many C implementations push function arguments
8647 onto the stack in reverse order, the new library version is likely to
8648 interoperate with programs that have been compiled with the old
8649 pem_password_cb definition (PEM_whatever takes some data that
8650 happens to be on the stack as its last argument, and the callback
8651 just ignores this garbage); but there is no guarantee whatsoever that
8654 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8655 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8656 problems not only on Windows, but also on some Unix platforms.
8657 To avoid problematic command lines, these definitions are now in an
8658 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8659 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8662 *) MIPS III/IV assembler module is reimplemented.
8665 *) More DES library cleanups: remove references to srand/rand and
8666 delete an unused file.
8669 *) Add support for the the free Netwide assembler (NASM) under Win32,
8670 since not many people have MASM (ml) and it can be hard to obtain.
8671 This is currently experimental but it seems to work OK and pass all
8672 the tests. Check out INSTALL.W32 for info.
8675 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8676 without temporary keys kept an extra copy of the server key,
8677 and connections with temporary keys did not free everything in case
8681 *) New function RSA_check_key and new openssl rsa option -check
8682 for verifying the consistency of RSA keys.
8683 [Ulf Moeller, Bodo Moeller]
8685 *) Various changes to make Win32 compile work:
8686 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8687 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8688 comparison" warnings.
8689 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8692 *) Add a debugging option to PKCS#5 v2 key generation function: when
8693 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8694 derived keys are printed to stderr.
8697 *) Copy the flags in ASN1_STRING_dup().
8698 [Roman E. Pavlov <pre@mo.msk.ru>]
8700 *) The x509 application mishandled signing requests containing DSA
8701 keys when the signing key was also DSA and the parameters didn't match.
8703 It was supposed to omit the parameters when they matched the signing key:
8704 the verifying software was then supposed to automatically use the CA's
8705 parameters if they were absent from the end user certificate.
8707 Omitting parameters is no longer recommended. The test was also
8708 the wrong way round! This was probably due to unusual behaviour in
8709 EVP_cmp_parameters() which returns 1 if the parameters match.
8710 This meant that parameters were omitted when they *didn't* match and
8711 the certificate was useless. Certificates signed with 'ca' didn't have
8713 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8715 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8716 The interface is as follows:
8717 Applications can use
8718 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8719 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8720 "off" is now the default.
8721 The library internally uses
8722 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8723 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8724 to disable memory-checking temporarily.
8726 Some inconsistent states that previously were possible (and were
8727 even the default) are now avoided.
8729 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8730 with each memory chunk allocated; this is occasionally more helpful
8731 than just having a counter.
8733 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8735 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8739 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8740 which largely parallels "options", but is for changing API behaviour,
8741 whereas "options" are about protocol behaviour.
8742 Initial "mode" flags are:
8744 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8745 a single record has been written.
8746 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8747 retries use the same buffer location.
8748 (But all of the contents must be
8752 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8755 *) Fix problems with no-hmac etc.
8756 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8758 *) New functions RSA_get_default_method(), RSA_set_method() and
8759 RSA_get_method(). These allows replacement of RSA_METHODs without having
8760 to mess around with the internals of an RSA structure.
8763 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8764 Also really enable memory leak checks in openssl.c and in some
8766 [Chad C. Mulligan, Bodo Moeller]
8768 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8769 up the length of negative integers. This has now been simplified to just
8770 store the length when it is first determined and use it later, rather
8771 than trying to keep track of where data is copied and updating it to
8773 [Steve Henson, reported by Brien Wheeler
8774 <bwheeler@authentica-security.com>]
8776 *) Add a new function PKCS7_signatureVerify. This allows the verification
8777 of a PKCS#7 signature but with the signing certificate passed to the
8778 function itself. This contrasts with PKCS7_dataVerify which assumes the
8779 certificate is present in the PKCS#7 structure. This isn't always the
8780 case: certificates can be omitted from a PKCS#7 structure and be
8781 distributed by "out of band" means (such as a certificate database).
8784 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8785 function prototypes in pem.h, also change util/mkdef.pl to add the
8786 necessary function names.
8789 *) mk1mf.pl (used by Windows builds) did not properly read the
8790 options set by Configure in the top level Makefile, and Configure
8791 was not even able to write more than one option correctly.
8792 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8795 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8796 file to be loaded from a BIO or FILE pointer. The BIO version will
8797 for example allow memory BIOs to contain config info.
8800 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8801 Whoever hopes to achieve shared-library compatibility across versions
8802 must use this, not the compile-time macro.
8803 (Exercise 0.9.4: Which is the minimum library version required by
8805 Note: All this applies only to multi-threaded programs, others don't
8809 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8810 through a BIO pair triggered the default case, i.e.
8811 SSLerr(...,SSL_R_UNKNOWN_STATE).
8814 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8815 can use the SSL library even if none of the specific BIOs is
8819 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8820 for the encoded length.
8821 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8823 *) Add initial documentation of the X509V3 functions.
8826 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8827 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8828 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8829 secure PKCS#8 private key format with a high iteration count.
8832 *) Fix determination of Perl interpreter: A perl or perl5
8833 _directory_ in $PATH was also accepted as the interpreter.
8834 [Ralf S. Engelschall]
8836 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8837 wrong with it but it was very old and did things like calling
8838 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8842 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8843 to use the new extension code.
8846 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8847 with macros. This should make it easier to change their form, add extra
8848 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8852 *) Add to configuration table a new entry that can specify an alternative
8853 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8854 according to Mark Crispin <MRC@Panda.COM>.
8858 *) DES CBC did not update the IV. Weird.
8861 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8862 Changing the behaviour of the former might break existing programs --
8863 where IV updating is needed, des_ncbc_encrypt can be used.
8866 *) When bntest is run from "make test" it drives bc to check its
8867 calculations, as well as internally checking them. If an internal check
8868 fails, it needs to cause bc to give a non-zero result or make test carries
8869 on without noticing the failure. Fixed.
8872 *) DES library cleanups.
8875 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8876 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8877 ciphers. NOTE: although the key derivation function has been verified
8878 against some published test vectors it has not been extensively tested
8879 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8883 *) Instead of "mkdir -p", which is not fully portable, use new
8884 Perl script "util/mkdir-p.pl".
8887 *) Rewrite the way password based encryption (PBE) is handled. It used to
8888 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8889 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8890 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8891 the 'parameter' field of the AlgorithmIdentifier is passed to the
8892 underlying key generation function so it must do its own ASN1 parsing.
8893 This has also changed the EVP_PBE_CipherInit() function which now has a
8894 'parameter' argument instead of literal salt and iteration count values
8895 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8898 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8899 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8900 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8901 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8902 value was just used as a "magic string" and not used directly its
8903 value doesn't matter.
8906 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8910 *) "linux-sparc64" configuration (ultrapenguin).
8911 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8912 "linux-sparc" configuration.
8913 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8915 *) config now generates no-xxx options for missing ciphers.
8918 *) Support the EBCDIC character set (work in progress).
8919 File ebcdic.c not yet included because it has a different license.
8920 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8922 *) Support BS2000/OSD-POSIX.
8923 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8925 *) Make callbacks for key generation use void * instead of char *.
8928 *) Make S/MIME samples compile (not yet tested).
8931 *) Additional typesafe stacks.
8934 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8938 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8940 *) New configuration variant "sco5-gcc".
8942 *) Updated some demos.
8943 [Sean O Riordain, Wade Scholine]
8945 *) Add missing BIO_free at exit of pkcs12 application.
8948 *) Fix memory leak in conf.c.
8951 *) Updates for Win32 to assembler version of MD5.
8954 *) Set #! path to perl in apps/der_chop to where we found it
8955 instead of using a fixed path.
8958 *) SHA library changes for irix64-mips4-cc.
8961 *) Improvements for VMS support.
8965 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8967 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8968 This also avoids the problems with SC4.2 and unpatched SC5.
8969 [Andy Polyakov <appro@fy.chalmers.se>]
8971 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8972 These are required because of the typesafe stack would otherwise break
8973 existing code. If old code used a structure member which used to be STACK
8974 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8975 sk_num or sk_value it would produce an error because the num, data members
8976 are not present in STACK_OF. Now it just produces a warning. sk_set
8977 replaces the old method of assigning a value to sk_value
8978 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8979 that does this will no longer work (and should use sk_set instead) but
8980 this could be regarded as a "questionable" behaviour anyway.
8983 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8984 correctly handle encrypted S/MIME data.
8987 *) Change type of various DES function arguments from des_cblock
8988 (which means, in function argument declarations, pointer to char)
8989 to des_cblock * (meaning pointer to array with 8 char elements),
8990 which allows the compiler to do more typechecking; it was like
8991 that back in SSLeay, but with lots of ugly casts.
8993 Introduce new type const_des_cblock.
8996 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8997 problems: find RecipientInfo structure that matches recipient certificate
8998 and initialise the ASN1 structures properly based on passed cipher.
9001 *) Belatedly make the BN tests actually check the results.
9004 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9005 to and from BNs: it was completely broken. New compilation option
9006 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9007 key elements as negative integers.
9010 *) Reorganize and speed up MD5.
9011 [Andy Polyakov <appro@fy.chalmers.se>]
9014 [Richard Levitte <richard@levitte.org>]
9016 *) New option -out to asn1parse to allow the parsed structure to be
9017 output to a file. This is most useful when combined with the -strparse
9018 option to examine the output of things like OCTET STRINGS.
9021 *) Make SSL library a little more fool-proof by not requiring any longer
9022 that SSL_set_{accept,connect}_state be called before
9023 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9024 in many applications because usually everything *appeared* to work as
9025 intended anyway -- now it really works as intended).
9028 *) Move openssl.cnf out of lib/.
9031 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9032 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9033 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9034 [Ralf S. Engelschall]
9036 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9037 handle PKCS#7 enveloped data properly.
9038 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9040 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9041 copying pointers. The cert_st handling is changed by this in
9042 various ways (and thus what used to be known as ctx->default_cert
9043 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9044 any longer when s->cert does not give us what we need).
9045 ssl_cert_instantiate becomes obsolete by this change.
9046 As soon as we've got the new code right (possibly it already is?),
9047 we have solved a couple of bugs of the earlier code where s->cert
9048 was used as if it could not have been shared with other SSL structures.
9050 Note that using the SSL API in certain dirty ways now will result
9051 in different behaviour than observed with earlier library versions:
9052 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9053 does not influence s as it used to.
9055 In order to clean up things more thoroughly, inside SSL_SESSION
9056 we don't use CERT any longer, but a new structure SESS_CERT
9057 that holds per-session data (if available); currently, this is
9058 the peer's certificate chain and, for clients, the server's certificate
9059 and temporary key. CERT holds only those values that can have
9060 meaningful defaults in an SSL_CTX.
9063 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9064 from the internal representation. Various PKCS#7 fixes: remove some
9065 evil casts and set the enc_dig_alg field properly based on the signing
9069 *) Allow PKCS#12 password to be set from the command line or the
9070 environment. Let 'ca' get its config file name from the environment
9071 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9075 *) Allow certificate policies extension to use an IA5STRING for the
9076 organization field. This is contrary to the PKIX definition but
9077 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9081 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9082 without disallowing inline assembler and the like for non-pedantic builds.
9085 *) Support Borland C++ builder.
9086 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9091 *) SHA-1 cleanups and performance enhancements.
9092 [Andy Polyakov <appro@fy.chalmers.se>]
9094 *) Sparc v8plus assembler for the bignum library.
9095 [Andy Polyakov <appro@fy.chalmers.se>]
9097 *) Accept any -xxx and +xxx compiler options in Configure.
9100 *) Update HPUX configuration.
9103 *) Add missing sk_<type>_unshift() function to safestack.h
9104 [Ralf S. Engelschall]
9106 *) New function SSL_CTX_use_certificate_chain_file that sets the
9107 "extra_cert"s in addition to the certificate. (This makes sense
9108 only for "PEM" format files, as chains as a whole are not
9112 *) Support verify_depth from the SSL API.
9113 x509_vfy.c had what can be considered an off-by-one-error:
9114 Its depth (which was not part of the external interface)
9115 was actually counting the number of certificates in a chain;
9116 now it really counts the depth.
9119 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9120 instead of X509err, which often resulted in confusing error
9121 messages since the error codes are not globally unique
9122 (e.g. an alleged error in ssl3_accept when a certificate
9123 didn't match the private key).
9125 *) New function SSL_CTX_set_session_id_context that allows to set a default
9126 value (so that you don't need SSL_set_session_id_context for each
9127 connection using the SSL_CTX).
9130 *) OAEP decoding bug fix.
9133 *) Support INSTALL_PREFIX for package builders, as proposed by
9137 *) New Configure options "threads" and "no-threads". For systems
9138 where the proper compiler options are known (currently Solaris
9139 and Linux), "threads" is the default.
9142 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9145 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9146 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9147 such as /usr/local/bin.
9150 *) "make linux-shared" to build shared libraries.
9151 [Niels Poppe <niels@netbox.org>]
9153 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9156 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9157 extension adding in x509 utility.
9160 *) Remove NOPROTO sections and error code comments.
9163 *) Partial rewrite of the DEF file generator to now parse the ANSI
9167 *) New Configure options --prefix=DIR and --openssldir=DIR.
9170 *) Complete rewrite of the error code script(s). It is all now handled
9171 by one script at the top level which handles error code gathering,
9172 header rewriting and C source file generation. It should be much better
9173 than the old method: it now uses a modified version of Ulf's parser to
9174 read the ANSI prototypes in all header files (thus the old K&R definitions
9175 aren't needed for error creation any more) and do a better job of
9176 translating function codes into names. The old 'ASN1 error code imbedded
9177 in a comment' is no longer necessary and it doesn't use .err files which
9178 have now been deleted. Also the error code call doesn't have to appear all
9179 on one line (which resulted in some large lines...).
9182 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9185 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9186 0 (which usually indicates a closed connection), but continue reading.
9189 *) Fix some race conditions.
9192 *) Add support for CRL distribution points extension. Add Certificate
9193 Policies and CRL distribution points documentation.
9196 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9199 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9200 8 of keying material. Merlin has also confirmed interop with this fix
9201 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9202 [Merlin Hughes <merlin@baltimore.ie>]
9204 *) Fix lots of warnings.
9205 [Richard Levitte <levitte@stacken.kth.se>]
9207 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9208 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9209 [Richard Levitte <levitte@stacken.kth.se>]
9211 *) Fix problems with sizeof(long) == 8.
9212 [Andy Polyakov <appro@fy.chalmers.se>]
9214 *) Change functions to ANSI C.
9217 *) Fix typos in error codes.
9218 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9220 *) Remove defunct assembler files from Configure.
9223 *) SPARC v8 assembler BIGNUM implementation.
9224 [Andy Polyakov <appro@fy.chalmers.se>]
9226 *) Support for Certificate Policies extension: both print and set.
9227 Various additions to support the r2i method this uses.
9230 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9231 return a const string when you are expecting an allocated buffer.
9234 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9235 types DirectoryString and DisplayText.
9238 *) Add code to allow r2i extensions to access the configuration database,
9239 add an LHASH database driver and add several ctx helper functions.
9242 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9243 fail when they extended the size of a BIGNUM.
9246 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9247 support typesafe stack.
9250 *) Fix typo in SSL_[gs]et_options().
9251 [Nils Frostberg <nils@medcom.se>]
9253 *) Delete various functions and files that belonged to the (now obsolete)
9254 old X509V3 handling code.
9257 *) New Configure option "rsaref".
9260 *) Don't auto-generate pem.h.
9263 *) Introduce type-safe ASN.1 SETs.
9266 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9267 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9269 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9270 that links with OpenSSL (well at least cause lots of warnings), but fear
9271 not: the conversion is trivial, and it eliminates loads of evil casts. A
9272 few STACKed things have been converted already. Feel free to convert more.
9273 In the fullness of time, I'll do away with the STACK type altogether.
9276 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9277 specified in <certfile> by updating the entry in the index.txt file.
9278 This way one no longer has to edit the index.txt file manually for
9279 revoking a certificate. The -revoke option does the gory details now.
9280 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9282 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9283 `-text' option at all and this way the `-noout -text' combination was
9284 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9285 [Ralf S. Engelschall]
9287 *) Make sure a corresponding plain text error message exists for the
9288 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9289 verify callback function determined that a certificate was revoked.
9290 [Ralf S. Engelschall]
9292 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9293 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9294 all available cipers including rc5, which was forgotten until now.
9295 In order to let the testing shell script know which algorithms
9296 are available, a new (up to now undocumented) command
9297 "openssl list-cipher-commands" is used.
9300 *) Bugfix: s_client occasionally would sleep in select() when
9301 it should have checked SSL_pending() first.
9304 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9305 the raw DSA values prior to ASN.1 encoding.
9308 *) Tweaks to Configure
9309 [Niels Poppe <niels@netbox.org>]
9311 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9315 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9318 *) New config option to avoid instructions that are illegal on the 80386.
9319 The default code is faster, but requires at least a 486.
9322 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9323 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9324 same as SSL2_VERSION anyway.
9327 *) New "-showcerts" option for s_client.
9330 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9331 application. Various cleanups and fixes.
9334 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9335 modify error routines to work internally. Add error codes and PBE init
9336 to library startup routines.
9339 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9340 packing functions to asn1 and evp. Changed function names and error
9341 codes along the way.
9344 *) PKCS12 integration: and so it begins... First of several patches to
9345 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9346 objects to objects.h
9349 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9350 and display support for Thawte strong extranet extension.
9353 *) Add LinuxPPC support.
9354 [Jeff Dubrule <igor@pobox.org>]
9356 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9357 bn_div_words in alpha.s.
9358 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9360 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9361 OAEP isn't supported when OpenSSL is built with RSAref.
9362 [Ulf Moeller <ulf@fitug.de>]
9364 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9365 so they no longer are missing under -DNOPROTO.
9366 [Soren S. Jorvang <soren@t.dk>]
9369 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9371 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9372 doesn't work when the session is reused. Coming soon!
9375 *) Fix a security hole, that allows sessions to be reused in the wrong
9376 context thus bypassing client cert protection! All software that uses
9377 client certs and session caches in multiple contexts NEEDS PATCHING to
9378 allow session reuse! A fuller solution is in the works.
9379 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9381 *) Some more source tree cleanups (removed obsolete files
9382 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9383 permission on "config" script to be executable) and a fix for the INSTALL
9385 [Ulf Moeller <ulf@fitug.de>]
9387 *) Remove some legacy and erroneous uses of malloc, free instead of
9389 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9391 *) Make rsa_oaep_test return non-zero on error.
9392 [Ulf Moeller <ulf@fitug.de>]
9394 *) Add support for native Solaris shared libraries. Configure
9395 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9396 if someone would make that last step automatic.
9397 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9399 *) ctx_size was not built with the right compiler during "make links". Fixed.
9402 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9403 except NULL ciphers". This means the default cipher list will no longer
9404 enable NULL ciphers. They need to be specifically enabled e.g. with
9405 the string "DEFAULT:eNULL".
9408 *) Fix to RSA private encryption routines: if p < q then it would
9409 occasionally produce an invalid result. This will only happen with
9410 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9413 *) Be less restrictive and allow also `perl util/perlpath.pl
9414 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9415 because this way one can also use an interpreter named `perl5' (which is
9416 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9417 installed as `perl').
9418 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9420 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9421 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9423 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9424 advapi32.lib to Win32 build and change the pem test comparision
9425 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9426 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9427 and crypto/des/ede_cbcm_enc.c.
9430 *) DES quad checksum was broken on big-endian architectures. Fixed.
9433 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9434 Win32 test batch file so it (might) work again. The Win32 test batch file
9435 is horrible: I feel ill....
9438 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9439 in e_os.h. Audit of header files to check ANSI and non ANSI
9440 sections: 10 functions were absent from non ANSI section and not exported
9441 from Windows DLLs. Fixed up libeay.num for new functions.
9444 *) Make `openssl version' output lines consistent.
9445 [Ralf S. Engelschall]
9447 *) Fix Win32 symbol export lists for BIO functions: Added
9448 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9449 to ms/libeay{16,32}.def.
9450 [Ralf S. Engelschall]
9452 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9453 fine under Unix and passes some trivial tests I've now added. But the
9454 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9455 added to make sure no one expects that this stuff really works in the
9456 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9457 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9459 [Ralf S. Engelschall]
9461 *) Fix the generation of two part addresses in perl.
9462 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9464 *) Add config entry for Linux on MIPS.
9465 [John Tobey <jtobey@channel1.com>]
9467 *) Make links whenever Configure is run, unless we are on Windoze.
9470 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9471 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9475 *) Add a useful kludge to allow package maintainers to specify compiler and
9476 other platforms details on the command line without having to patch the
9477 Configure script everytime: One now can use ``perl Configure
9478 <id>:<details>'', i.e. platform ids are allowed to have details appended
9479 to them (seperated by colons). This is treated as there would be a static
9480 pre-configured entry in Configure's %table under key <id> with value
9481 <details> and ``perl Configure <id>'' is called. So, when you want to
9482 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9483 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9484 now, which overrides the FreeBSD-elf entry on-the-fly.
9485 [Ralf S. Engelschall]
9487 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9490 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9491 on the `perl Configure ...' command line. This way one can compile
9492 OpenSSL libraries with Position Independent Code (PIC) which is needed
9493 for linking it into DSOs.
9494 [Ralf S. Engelschall]
9496 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9500 *) Cleaned up the LICENSE document: The official contact for any license
9501 questions now is the OpenSSL core team under openssl-core@openssl.org.
9502 And add a paragraph about the dual-license situation to make sure people
9503 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9504 to the OpenSSL toolkit.
9505 [Ralf S. Engelschall]
9507 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9508 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9509 Additonally cleaned up the `make links' target: Remove unnecessary
9510 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9511 to speed processing and no longer clutter the display with confusing
9512 stuff. Instead only the actually done links are displayed.
9513 [Ralf S. Engelschall]
9515 *) Permit null encryption ciphersuites, used for authentication only. It used
9516 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9517 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9521 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9522 signed attributes when verifying signatures (this would break them),
9523 the detached data encoding was wrong and public keys obtained using
9524 X509_get_pubkey() weren't freed.
9527 *) Add text documentation for the BUFFER functions. Also added a work around
9528 to a Win95 console bug. This was triggered by the password read stuff: the
9529 last character typed gets carried over to the next fread(). If you were
9530 generating a new cert request using 'req' for example then the last
9531 character of the passphrase would be CR which would then enter the first
9535 *) Added the new `Includes OpenSSL Cryptography Software' button as
9536 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9537 button and can be used by applications based on OpenSSL to show the
9538 relationship to the OpenSSL project.
9539 [Ralf S. Engelschall]
9541 *) Remove confusing variables in function signatures in files
9542 ssl/ssl_lib.c and ssl/ssl.h.
9543 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9545 *) Don't install bss_file.c under PREFIX/include/
9546 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9548 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9549 functions that return function pointers and has support for NT specific
9550 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9551 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9552 unsigned to signed types: this was killing the Win32 compile.
9555 *) Add new certificate file to stack functions,
9556 SSL_add_dir_cert_subjects_to_stack() and
9557 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9558 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9559 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9560 This means that Apache-SSL and similar packages don't have to mess around
9561 to add as many CAs as they want to the preferred list.
9564 *) Experiment with doxygen documentation. Currently only partially applied to
9566 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9567 openssl.doxy as the configuration file.
9570 *) Get rid of remaining C++-style comments which strict C compilers hate.
9571 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9573 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9574 compiled in by default: it has problems with large keys.
9577 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9578 DH private keys and/or callback functions which directly correspond to
9579 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9580 is needed for applications which have to configure certificates on a
9581 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9583 For the RSA certificate situation is makes no difference, but
9584 for the DSA certificate situation this fixes the "no shared cipher"
9585 problem where the OpenSSL cipher selection procedure failed because the
9586 temporary keys were not overtaken from the context and the API provided
9587 no way to reconfigure them.
9588 The new functions now let applications reconfigure the stuff and they
9589 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9590 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9591 non-public-API function ssl_cert_instantiate() is used as a helper
9592 function and also to reduce code redundancy inside ssl_rsa.c.
9593 [Ralf S. Engelschall]
9595 *) Move s_server -dcert and -dkey options out of the undocumented feature
9596 area because they are useful for the DSA situation and should be
9597 recognized by the users.
9598 [Ralf S. Engelschall]
9600 *) Fix the cipher decision scheme for export ciphers: the export bits are
9601 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9602 SSL_EXP_MASK. So, the original variable has to be used instead of the
9603 already masked variable.
9604 [Richard Levitte <levitte@stacken.kth.se>]
9606 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9607 [Richard Levitte <levitte@stacken.kth.se>]
9609 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9610 from `int' to `unsigned int' because it's a length and initialized by
9611 EVP_DigestFinal() which expects an `unsigned int *'.
9612 [Richard Levitte <levitte@stacken.kth.se>]
9614 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9615 script. Instead use the usual Shell->Perl transition trick.
9616 [Ralf S. Engelschall]
9618 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9619 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9620 -noout -modulus' as it's already the case for `openssl rsa -noout
9621 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9622 currently the public key is printed (a decision which was already done by
9623 `openssl dsa -modulus' in the past) which serves a similar purpose.
9624 Additionally the NO_RSA no longer completely removes the whole -modulus
9625 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9627 [Ralf S. Engelschall]
9629 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9630 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9631 [Arne Ansper <arne@ats.cyber.ee>]
9633 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9634 to be added. Now both 'req' and 'ca' can use new objects defined in the
9638 *) Add cool BIO that does syslog (or event log on NT).
9639 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9641 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9642 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9643 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9644 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9647 *) Add preliminary config info for new extension code.
9650 *) Make RSA_NO_PADDING really use no padding.
9651 [Ulf Moeller <ulf@fitug.de>]
9653 *) Generate errors when private/public key check is done.
9656 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9657 for some CRL extensions and new objects added.
9660 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9661 key usage extension and fuller support for authority key id.
9664 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9665 padding method for RSA, which is recommended for new applications in PKCS
9666 #1 v2.0 (RFC 2437, October 1998).
9667 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9668 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9669 against Bleichbacher's attack on RSA.
9670 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9673 *) Updates to the new SSL compression code
9674 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9676 *) Fix so that the version number in the master secret, when passed
9677 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9678 (because the server will not accept higher), that the version number
9679 is 0x03,0x01, not 0x03,0x00
9680 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9682 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9683 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9684 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9687 *) Support for RAW extensions where an arbitrary extension can be
9688 created by including its DER encoding. See apps/openssl.cnf for
9692 *) Make sure latest Perl versions don't interpret some generated C array
9693 code as Perl array code in the crypto/err/err_genc.pl script.
9694 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9696 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9697 not many people have the assembler. Various Win32 compilation fixes and
9698 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9702 *) Modify configure script 'Configure' to automatically create crypto/date.h
9703 file under Win32 and also build pem.h from pem.org. New script
9704 util/mkfiles.pl to create the MINFO file on environments that can't do a
9705 'make files': perl util/mkfiles.pl >MINFO should work.
9708 *) Major rework of DES function declarations, in the pursuit of correctness
9709 and purity. As a result, many evil casts evaporated, and some weirdness,
9710 too. You may find this causes warnings in your code. Zapping your evil
9711 casts will probably fix them. Mostly.
9714 *) Fix for a typo in asn1.h. Bug fix to object creation script
9715 obj_dat.pl. It considered a zero in an object definition to mean
9716 "end of object": none of the objects in objects.h have any zeros
9717 so it wasn't spotted.
9718 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9720 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9721 Masking (CBCM). In the absence of test vectors, the best I have been able
9722 to do is check that the decrypt undoes the encrypt, so far. Send me test
9723 vectors if you have them.
9726 *) Correct calculation of key length for export ciphers (too much space was
9727 allocated for null ciphers). This has not been tested!
9730 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9731 message is now correct (it understands "crypto" and "ssl" on its
9732 command line). There is also now an "update" option. This will update
9733 the util/ssleay.num and util/libeay.num files with any new functions.
9735 perl util/mkdef.pl crypto ssl update
9736 it will update them.
9739 *) Overhauled the Perl interface (perl/*):
9740 - ported BN stuff to OpenSSL's different BN library
9741 - made the perl/ source tree CVS-aware
9742 - renamed the package from SSLeay to OpenSSL (the files still contain
9743 their history because I've copied them in the repository)
9744 - removed obsolete files (the test scripts will be replaced
9745 by better Test::Harness variants in the future)
9746 [Ralf S. Engelschall]
9748 *) First cut for a very conservative source tree cleanup:
9749 1. merge various obsolete readme texts into doc/ssleay.txt
9750 where we collect the old documents and readme texts.
9751 2. remove the first part of files where I'm already sure that we no
9752 longer need them because of three reasons: either they are just temporary
9753 files which were left by Eric or they are preserved original files where
9754 I've verified that the diff is also available in the CVS via "cvs diff
9755 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9756 the crypto/md/ stuff).
9757 [Ralf S. Engelschall]
9759 *) More extension code. Incomplete support for subject and issuer alt
9760 name, issuer and authority key id. Change the i2v function parameters
9761 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9762 what that's for :-) Fix to ASN1 macro which messed up
9763 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9766 *) Preliminary support for ENUMERATED type. This is largely copied from the
9770 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9771 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9773 *) Make sure `make rehash' target really finds the `openssl' program.
9774 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9776 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9777 like to hear about it if this slows down other processors.
9780 *) Add CygWin32 platform information to Configure script.
9781 [Alan Batie <batie@aahz.jf.intel.com>]
9783 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9784 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9786 *) New program nseq to manipulate netscape certificate sequences
9789 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9793 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9794 but the BN code had some problems that would cause failures when
9795 doing certificate verification and some other functions.
9796 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9798 *) Add ASN1 and PEM code to support netscape certificate sequences.
9801 *) Add ASN1 and PEM code to support netscape certificate sequences.
9804 *) Add several PKIX and private extended key usage OIDs.
9807 *) Modify the 'ca' program to handle the new extension code. Modify
9808 openssl.cnf for new extension format, add comments.
9811 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9812 and add a sample to openssl.cnf so req -x509 now adds appropriate
9816 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9817 error code, add initial support to X509_print() and x509 application.
9820 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9821 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9822 stuff is currently isolated and isn't even compiled yet.
9825 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9826 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9827 Removed the versions check from X509 routines when loading extensions:
9828 this allows certain broken certificates that don't set the version
9829 properly to be processed.
9832 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9833 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9834 can still be regenerated with "make depend".
9837 *) Spelling mistake in C version of CAST-128.
9838 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9840 *) Changes to the error generation code. The perl script err-code.pl
9841 now reads in the old error codes and retains the old numbers, only
9842 adding new ones if necessary. It also only changes the .err files if new
9843 codes are added. The makefiles have been modified to only insert errors
9844 when needed (to avoid needlessly modifying header files). This is done
9845 by only inserting errors if the .err file is newer than the auto generated
9846 C file. To rebuild all the error codes from scratch (the old behaviour)
9847 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9848 or delete all the .err files.
9851 *) CAST-128 was incorrectly implemented for short keys. The C version has
9852 been fixed, but is untested. The assembler versions are also fixed, but
9853 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9854 to regenerate it if needed.
9855 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9856 Hagino <itojun@kame.net>]
9858 *) File was opened incorrectly in randfile.c.
9859 [Ulf Möller <ulf@fitug.de>]
9861 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9862 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9863 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9864 al: it's just almost always a UTCTime. Note this patch adds new error
9865 codes so do a "make errors" if there are problems.
9868 *) Correct Linux 1 recognition in config.
9869 [Ulf Möller <ulf@fitug.de>]
9871 *) Remove pointless MD5 hash when using DSA keys in ca.
9872 [Anonymous <nobody@replay.com>]
9874 *) Generate an error if given an empty string as a cert directory. Also
9875 generate an error if handed NULL (previously returned 0 to indicate an
9876 error, but didn't set one).
9877 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9879 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9882 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9883 parameters. This was causing a warning which killed off the Win32 compile.
9886 *) Remove C++ style comments from crypto/bn/bn_local.h.
9887 [Neil Costigan <neil.costigan@celocom.com>]
9889 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9890 based on a text string, looking up short and long names and finally
9891 "dot" format. The "dot" format stuff didn't work. Added new function
9892 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9893 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9894 OID is not part of the table.
9897 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9898 X509_LOOKUP_by_alias().
9901 *) Sort openssl functions by name.
9904 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9905 encryption from sample DSA keys (in case anyone is interested the password
9909 *) Make _all_ *_free functions accept a NULL pointer.
9910 [Frans Heymans <fheymans@isaserver.be>]
9912 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9914 [Anonymous <nobody@replay.com>]
9916 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9917 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9919 *) Don't blow it for numeric -newkey arguments to apps/req.
9920 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9922 *) Temp key "for export" tests were wrong in s3_srvr.c.
9923 [Anonymous <nobody@replay.com>]
9925 *) Add prototype for temp key callback functions
9926 SSL_CTX_set_tmp_{rsa,dh}_callback().
9929 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9930 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9933 *) X509_name_add_entry() freed the wrong thing after an error.
9934 [Arne Ansper <arne@ats.cyber.ee>]
9936 *) rsa_eay.c would attempt to free a NULL context.
9937 [Arne Ansper <arne@ats.cyber.ee>]
9939 *) BIO_s_socket() had a broken should_retry() on Windoze.
9940 [Arne Ansper <arne@ats.cyber.ee>]
9942 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9943 [Arne Ansper <arne@ats.cyber.ee>]
9945 *) Make sure the already existing X509_STORE->depth variable is initialized
9946 in X509_STORE_new(), but document the fact that this variable is still
9947 unused in the certificate verification process.
9948 [Ralf S. Engelschall]
9950 *) Fix the various library and apps files to free up pkeys obtained from
9951 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9954 *) Fix reference counting in X509_PUBKEY_get(). This makes
9955 demos/maurice/example2.c work, amongst others, probably.
9956 [Steve Henson and Ben Laurie]
9958 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9959 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9960 are no longer created. This way we have a single and consistent command
9961 line interface `openssl <command>', similar to `cvs <command>'.
9962 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9964 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9965 BIT STRING wrapper always have zero unused bits.
9968 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9971 *) Make the top-level INSTALL documentation easier to understand.
9974 *) Makefiles updated to exit if an error occurs in a sub-directory
9975 make (including if user presses ^C) [Paul Sutton]
9977 *) Make Montgomery context stuff explicit in RSA data structure.
9980 *) Fix build order of pem and err to allow for generated pem.h.
9983 *) Fix renumbering bug in X509_NAME_delete_entry().
9986 *) Enhanced the err-ins.pl script so it makes the error library number
9987 global and can add a library name. This is needed for external ASN1 and
9988 other error libraries.
9991 *) Fixed sk_insert which never worked properly.
9994 *) Fix ASN1 macros so they can handle indefinite length construted
9995 EXPLICIT tags. Some non standard certificates use these: they can now
9999 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10000 into a single doc/ssleay.txt bundle. This way the information is still
10001 preserved but no longer messes up this directory. Now it's new room for
10002 the new set of documenation files.
10003 [Ralf S. Engelschall]
10005 *) SETs were incorrectly DER encoded. This was a major pain, because they
10006 shared code with SEQUENCEs, which aren't coded the same. This means that
10007 almost everything to do with SETs or SEQUENCEs has either changed name or
10008 number of arguments.
10009 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10011 *) Fix test data to work with the above.
10014 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10015 was already fixed by Eric for 0.9.1 it seems.
10016 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10018 *) Autodetect FreeBSD3.
10021 *) Fix various bugs in Configure. This affects the following platforms:
10025 unixware-2.0-pentium
10029 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10030 before they are needed.
10033 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10037 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10039 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10040 changed SSLeay to OpenSSL in version strings.
10041 [Ralf S. Engelschall]
10043 *) Some fixups to the top-level documents.
10046 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10047 because the symlink to include/ was missing.
10048 [Ralf S. Engelschall]
10050 *) Incorporated the popular no-RSA/DSA-only patches
10051 which allow to compile a RSA-free SSLeay.
10052 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10054 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10055 when "ssleay" is still not found.
10056 [Ralf S. Engelschall]
10058 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10059 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10061 *) Updated the README file.
10062 [Ralf S. Engelschall]
10064 *) Added various .cvsignore files in the CVS repository subdirs
10065 to make a "cvs update" really silent.
10066 [Ralf S. Engelschall]
10068 *) Recompiled the error-definition header files and added
10069 missing symbols to the Win32 linker tables.
10070 [Ralf S. Engelschall]
10072 *) Cleaned up the top-level documents;
10073 o new files: CHANGES and LICENSE
10074 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10075 o merged COPYRIGHT into LICENSE
10076 o removed obsolete TODO file
10077 o renamed MICROSOFT to INSTALL.W32
10078 [Ralf S. Engelschall]
10080 *) Removed dummy files from the 0.9.1b source tree:
10081 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10082 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10083 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10084 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10085 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10086 [Ralf S. Engelschall]
10088 *) Added various platform portability fixes.
10091 *) The Genesis of the OpenSSL rpject:
10092 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10093 Young and Tim J. Hudson created while they were working for C2Net until
10095 [The OpenSSL Project]
10098 Changes between 0.9.0b and 0.9.1b [not released]
10100 *) Updated a few CA certificates under certs/
10103 *) Changed some BIGNUM api stuff.
10106 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10107 DGUX x86, Linux Alpha, etc.
10110 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10111 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10115 *) Add -strparse option to asn1pars program which parses nested
10117 [Dr Stephen Henson <shenson@bigfoot.com>]
10119 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10122 *) DSA fix for "ca" program.
10125 *) Added "-genkey" option to "dsaparam" program.
10128 *) Added RIPE MD160 (rmd160) message digest.
10131 *) Added -a (all) option to "ssleay version" command.
10134 *) Added PLATFORM define which is the id given to Configure.
10137 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10140 *) Extended the ASN.1 parser routines.
10143 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10146 *) Added a BN_CTX to the BN library.
10149 *) Fixed the weak key values in DES library
10152 *) Changed API in EVP library for cipher aliases.
10155 *) Added support for RC2/64bit cipher.
10158 *) Converted the lhash library to the crypto/mem.c functions.
10161 *) Added more recognized ASN.1 object ids.
10164 *) Added more RSA padding checks for SSL/TLS.
10167 *) Added BIO proxy/filter functionality.
10170 *) Added extra_certs to SSL_CTX which can be used
10171 send extra CA certificates to the client in the CA cert chain sending
10172 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10175 *) Now Fortezza is denied in the authentication phase because
10176 this is key exchange mechanism is not supported by SSLeay at all.
10179 *) Additional PKCS1 checks.
10182 *) Support the string "TLSv1" for all TLS v1 ciphers.
10185 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10186 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10189 *) Fixed a few memory leaks.
10192 *) Fixed various code and comment typos.
10195 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10196 bytes sent in the client random.
10197 [Edward Bishop <ebishop@spyglass.com>]