5 Changes between 1.0.1f and 1.0.1g [xx XXX xxxx]
7 *) A missing bounds check in the handling of the TLS heartbeat extension
8 can be used to reveal up to 64k of memory to a connected client or
11 Thanks for Neel Mehta of Google Security for discovering this bug and to
12 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
13 preparing the fix (CVE-2014-0160)
14 [Adam Langley, Bodo Moeller]
16 *) Fix for the attack described in the paper "Recovering OpenSSL
17 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
18 by Yuval Yarom and Naomi Benger. Details can be obtained from:
19 http://eprint.iacr.org/2014/140
21 Thanks to Yuval Yarom and Naomi Benger for discovering this
22 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
23 [Yuval Yarom and Naomi Benger]
25 *) TLS pad extension: draft-agl-tls-padding-03
27 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
28 TLS client Hello record length value would otherwise be > 255 and
29 less that 512 pad with a dummy extension containing zeroes so it
30 is at least 512 bytes long.
32 [Adam Langley, Steve Henson]
34 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
36 *) Fix for TLS record tampering bug. A carefully crafted invalid
37 handshake could crash OpenSSL with a NULL pointer exception.
38 Thanks to Anton Johansson for reporting this issues.
41 *) Keep original DTLS digest and encryption contexts in retransmission
42 structures so we can use the previous session parameters if they need
43 to be resent. (CVE-2013-6450)
46 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
47 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
48 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
49 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
50 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
51 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
52 [Rob Stradling, Adam Langley]
54 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
56 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
57 supporting platforms or when small records were transferred.
58 [Andy Polyakov, Steve Henson]
60 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
62 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
64 This addresses the flaw in CBC record processing discovered by
65 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
66 at: http://www.isg.rhul.ac.uk/tls/
68 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
69 Security Group at Royal Holloway, University of London
70 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
71 Emilia Käsper for the initial patch.
73 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
75 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
76 ciphersuites which can be exploited in a denial of service attack.
77 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
78 and detecting this bug and to Wolfgang Ettlinger
79 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
83 *) Return an error when checking OCSP signatures when key is NULL.
84 This fixes a DoS attack. (CVE-2013-0166)
87 *) Make openssl verify return errors.
88 [Chris Palmer <palmer@google.com> and Ben Laurie]
90 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
91 the right response is stapled. Also change SSL_get_certificate()
92 so it returns the certificate actually sent.
93 See http://rt.openssl.org/Ticket/Display.html?id=2836.
94 [Rob Stradling <rob.stradling@comodo.com>]
96 *) Fix possible deadlock when decoding public keys.
99 *) Don't use TLS 1.0 record version number in initial client hello
103 Changes between 1.0.1b and 1.0.1c [10 May 2012]
105 *) Sanity check record length before skipping explicit IV in TLS
106 1.2, 1.1 and DTLS to fix DoS attack.
108 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
109 fuzzing as a service testing platform.
113 *) Initialise tkeylen properly when encrypting CMS messages.
114 Thanks to Solar Designer of Openwall for reporting this issue.
117 *) In FIPS mode don't try to use composite ciphers as they are not
121 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
123 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
124 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
125 mean any application compiled against OpenSSL 1.0.0 headers setting
126 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
127 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
128 0x10000000L Any application which was previously compiled against
129 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
130 will need to be recompiled as a result. Letting be results in
131 inability to disable specifically TLS 1.1 and in client context,
132 in unlike event, limit maximum offered version to TLS 1.0 [see below].
135 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
136 disable just protocol X, but all protocols above X *if* there are
137 protocols *below* X still enabled. In more practical terms it means
138 that if application wants to disable TLS1.0 in favor of TLS1.1 and
139 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
140 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
144 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
146 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
147 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
148 in CRYPTO_realloc_clean.
150 Thanks to Tavis Ormandy, Google Security Team, for discovering this
151 issue and to Adam Langley <agl@chromium.org> for fixing it.
153 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
155 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
158 *) Workarounds for some broken servers that "hang" if a client hello
159 record length exceeds 255 bytes.
161 1. Do not use record version number > TLS 1.0 in initial client
162 hello: some (but not all) hanging servers will now work.
163 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
164 the number of ciphers sent in the client hello. This should be
165 set to an even number, such as 50, for example by passing:
166 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
167 Most broken servers should now work.
168 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
169 TLS 1.2 client support entirely.
172 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
175 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
177 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
178 STRING form instead of a DigestInfo.
181 *) The format used for MDC2 RSA signatures is inconsistent between EVP
182 and the RSA_sign/RSA_verify functions. This was made more apparent when
183 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
184 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
185 the correct format in RSA_verify so both forms transparently work.
188 *) Some servers which support TLS 1.0 can choke if we initially indicate
189 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
190 encrypted premaster secret. As a workaround use the maximum pemitted
191 client version in client hello, this should keep such servers happy
192 and still work with previous versions of OpenSSL.
195 *) Add support for TLS/DTLS heartbeats.
196 [Robin Seggelmann <seggelmann@fh-muenster.de>]
198 *) Add support for SCTP.
199 [Robin Seggelmann <seggelmann@fh-muenster.de>]
201 *) Improved PRNG seeding for VOS.
202 [Paul Green <Paul.Green@stratus.com>]
204 *) Extensive assembler packs updates, most notably:
206 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
207 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
208 - x86_64: bit-sliced AES implementation;
209 - ARM: NEON support, contemporary platforms optimizations;
210 - s390x: z196 support;
211 - *: GHASH and GF(2^m) multiplication implementations;
215 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
216 (removal of unnecessary code)
217 [Peter Sylvester <peter.sylvester@edelweb.fr>]
219 *) Add TLS key material exporter from RFC 5705.
222 *) Add DTLS-SRTP negotiation from RFC 5764.
225 *) Add Next Protocol Negotiation,
226 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
227 disabled with a no-npn flag to config or Configure. Code donated
229 [Adam Langley <agl@google.com> and Ben Laurie]
231 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
232 NIST-P256, NIST-P521, with constant-time single point multiplication on
233 typical inputs. Compiler support for the nonstandard type __uint128_t is
234 required to use this (present in gcc 4.4 and later, for 64-bit builds).
235 Code made available under Apache License version 2.0.
237 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
238 line to include this in your build of OpenSSL, and run "make depend" (or
239 "make update"). This enables the following EC_METHODs:
241 EC_GFp_nistp224_method()
242 EC_GFp_nistp256_method()
243 EC_GFp_nistp521_method()
245 EC_GROUP_new_by_curve_name() will automatically use these (while
246 EC_GROUP_new_curve_GFp() currently prefers the more flexible
248 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
250 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
251 all platforms. Move ssize_t definition from e_os.h to the public
252 header file e_os2.h as it now appears in public header file cms.h
255 *) New -sigopt option to the ca, req and x509 utilities. Additional
256 signature parameters can be passed using this option and in
260 *) Add RSA PSS signing function. This will generate and set the
261 appropriate AlgorithmIdentifiers for PSS based on those in the
262 corresponding EVP_MD_CTX structure. No application support yet.
265 *) Support for companion algorithm specific ASN1 signing routines.
266 New function ASN1_item_sign_ctx() signs a pre-initialised
267 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
268 the appropriate parameters.
271 *) Add new algorithm specific ASN1 verification initialisation function
272 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
273 handling will be the same no matter what EVP_PKEY_METHOD is used.
274 Add a PSS handler to support verification of PSS signatures: checked
275 against a number of sample certificates.
278 *) Add signature printing for PSS. Add PSS OIDs.
279 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
281 *) Add algorithm specific signature printing. An individual ASN1 method
282 can now print out signatures instead of the standard hex dump.
284 More complex signatures (e.g. PSS) can print out more meaningful
285 information. Include DSA version that prints out the signature
289 *) Password based recipient info support for CMS library: implementing
293 *) Split password based encryption into PBES2 and PBKDF2 functions. This
294 neatly separates the code into cipher and PBE sections and is required
295 for some algorithms that split PBES2 into separate pieces (such as
299 *) Session-handling fixes:
300 - Fix handling of connections that are resuming with a session ID,
301 but also support Session Tickets.
302 - Fix a bug that suppressed issuing of a new ticket if the client
303 presented a ticket with an expired session.
304 - Try to set the ticket lifetime hint to something reasonable.
305 - Make tickets shorter by excluding irrelevant information.
306 - On the client side, don't ignore renewed tickets.
307 [Adam Langley, Bodo Moeller (Google)]
309 *) Fix PSK session representation.
312 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
314 This work was sponsored by Intel.
317 *) Add GCM support to TLS library. Some custom code is needed to split
318 the IV between the fixed (from PRF) and explicit (from TLS record)
319 portions. This adds all GCM ciphersuites supported by RFC5288 and
320 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
321 add a special AESGCM string for GCM only.
324 *) Expand range of ctrls for AES GCM. Permit setting invocation
325 field on decrypt and retrieval of invocation field only on encrypt.
328 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
329 As required by RFC5289 these ciphersuites cannot be used if for
330 versions of TLS earlier than 1.2.
333 *) For FIPS capable OpenSSL interpret a NULL default public key method
334 as unset and return the appopriate default but do *not* set the default.
335 This means we can return the appopriate method in applications that
336 swicth between FIPS and non-FIPS modes.
339 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
340 ENGINE is used then we cannot handle that in the FIPS module so we
341 keep original code iff non-FIPS operations are allowed.
344 *) Add -attime option to openssl utilities.
345 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
347 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
350 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
351 FIPS EC methods unconditionally for now.
354 *) New build option no-ec2m to disable characteristic 2 code.
357 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
358 all cases can be covered as some introduce binary incompatibilities.
361 *) Redirect RSA operations to FIPS module including keygen,
362 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
365 *) Add similar low level API blocking to ciphers.
368 *) Low level digest APIs are not approved in FIPS mode: any attempt
369 to use these will cause a fatal error. Applications that *really* want
370 to use them can use the private_* version instead.
373 *) Redirect cipher operations to FIPS module for FIPS builds.
376 *) Redirect digest operations to FIPS module for FIPS builds.
379 *) Update build system to add "fips" flag which will link in fipscanister.o
380 for static and shared library builds embedding a signature if needed.
383 *) Output TLS supported curves in preference order instead of numerical
384 order. This is currently hardcoded for the highest order curves first.
385 This should be configurable so applications can judge speed vs strength.
388 *) Add TLS v1.2 server support for client authentication.
391 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
395 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
396 FIPS modules versions.
399 *) Add TLS v1.2 client side support for client authentication. Keep cache
400 of handshake records longer as we don't know the hash algorithm to use
401 until after the certificate request message is received.
404 *) Initial TLS v1.2 client support. Add a default signature algorithms
405 extension including all the algorithms we support. Parse new signature
406 format in client key exchange. Relax some ECC signing restrictions for
407 TLS v1.2 as indicated in RFC5246.
410 *) Add server support for TLS v1.2 signature algorithms extension. Switch
411 to new signature format when needed using client digest preference.
412 All server ciphersuites should now work correctly in TLS v1.2. No client
413 support yet and no support for client certificates.
416 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
417 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
418 ciphersuites. At present only RSA key exchange ciphersuites work with
419 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
420 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
421 and version checking.
424 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
425 with this defined it will not be affected by any changes to ssl internal
426 structures. Add several utility functions to allow openssl application
427 to work with OPENSSL_NO_SSL_INTERN defined.
431 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
433 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
436 *) Permit abbreviated handshakes when renegotiating using the function
437 SSL_renegotiate_abbreviated().
438 [Robin Seggelmann <seggelmann@fh-muenster.de>]
440 *) Add call to ENGINE_register_all_complete() to
441 ENGINE_load_builtin_engines(), so some implementations get used
442 automatically instead of needing explicit application support.
445 *) Add support for TLS key exporter as described in RFC5705.
446 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
448 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
449 a few changes are required:
451 Add SSL_OP_NO_TLSv1_1 flag.
453 Update version checking logic to handle version 1.1.
454 Add explicit IV handling (ported from DTLS code).
455 Add command line options to s_client/s_server.
458 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
460 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
462 This addresses the flaw in CBC record processing discovered by
463 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
464 at: http://www.isg.rhul.ac.uk/tls/
466 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
467 Security Group at Royal Holloway, University of London
468 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
469 Emilia Käsper for the initial patch.
471 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
473 *) Return an error when checking OCSP signatures when key is NULL.
474 This fixes a DoS attack. (CVE-2013-0166)
477 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
478 the right response is stapled. Also change SSL_get_certificate()
479 so it returns the certificate actually sent.
480 See http://rt.openssl.org/Ticket/Display.html?id=2836.
482 [Rob Stradling <rob.stradling@comodo.com>]
484 *) Fix possible deadlock when decoding public keys.
487 Changes between 1.0.0i and 1.0.0j [10 May 2012]
489 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
492 *) Sanity check record length before skipping explicit IV in DTLS
495 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
496 fuzzing as a service testing platform.
500 *) Initialise tkeylen properly when encrypting CMS messages.
501 Thanks to Solar Designer of Openwall for reporting this issue.
504 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
506 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
507 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
508 in CRYPTO_realloc_clean.
510 Thanks to Tavis Ormandy, Google Security Team, for discovering this
511 issue and to Adam Langley <agl@chromium.org> for fixing it.
513 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
515 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
517 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
518 in CMS and PKCS7 code. When RSA decryption fails use a random key for
519 content decryption and always return the same error. Note: this attack
520 needs on average 2^20 messages so it only affects automated senders. The
521 old behaviour can be reenabled in the CMS code by setting the
522 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
523 an MMA defence is not necessary.
524 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
525 this issue. (CVE-2012-0884)
528 *) Fix CVE-2011-4619: make sure we really are receiving a
529 client hello before rejecting multiple SGC restarts. Thanks to
530 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
533 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
535 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
536 Thanks to Antonio Martin, Enterprise Secure Access Research and
537 Development, Cisco Systems, Inc. for discovering this bug and
538 preparing a fix. (CVE-2012-0050)
541 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
543 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
544 of the Vaudenay padding oracle attack on CBC mode encryption
545 which enables an efficient plaintext recovery attack against
546 the OpenSSL implementation of DTLS. Their attack exploits timing
547 differences arising during decryption processing. A research
548 paper describing this attack can be found at:
549 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
550 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
551 Security Group at Royal Holloway, University of London
552 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
553 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
554 for preparing the fix. (CVE-2011-4108)
555 [Robin Seggelmann, Michael Tuexen]
557 *) Clear bytes used for block padding of SSL 3.0 records.
559 [Adam Langley (Google)]
561 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
562 Kadianakis <desnacked@gmail.com> for discovering this issue and
563 Adam Langley for preparing the fix. (CVE-2011-4619)
564 [Adam Langley (Google)]
566 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
567 [Andrey Kulikov <amdeich@gmail.com>]
569 *) Prevent malformed RFC3779 data triggering an assertion failure.
570 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
571 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
572 [Rob Austein <sra@hactrn.net>]
574 *) Improved PRNG seeding for VOS.
575 [Paul Green <Paul.Green@stratus.com>]
577 *) Fix ssl_ciph.c set-up race.
578 [Adam Langley (Google)]
580 *) Fix spurious failures in ecdsatest.c.
581 [Emilia Käsper (Google)]
583 *) Fix the BIO_f_buffer() implementation (which was mixing different
584 interpretations of the '..._len' fields).
585 [Adam Langley (Google)]
587 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
588 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
589 threads won't reuse the same blinding coefficients.
591 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
592 lock to call BN_BLINDING_invert_ex, and avoids one use of
593 BN_BLINDING_update for each BN_BLINDING structure (previously,
594 the last update always remained unused).
595 [Emilia Käsper (Google)]
597 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
598 [Bob Buckholz (Google)]
600 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
602 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
603 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
604 [Kaspar Brand <ossl@velox.ch>]
606 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
607 for multi-threaded use of ECDH. (CVE-2011-3210)
608 [Adam Langley (Google)]
610 *) Fix x509_name_ex_d2i memory leak on bad inputs.
613 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
614 signature public key algorithm by using OID xref utilities instead.
615 Before this you could only use some ECC ciphersuites with SHA1 only.
618 *) Add protection against ECDSA timing attacks as mentioned in the paper
619 by Billy Bob Brumley and Nicola Tuveri, see:
621 http://eprint.iacr.org/2011/232.pdf
623 [Billy Bob Brumley and Nicola Tuveri]
625 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
627 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
628 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
630 *) Fix bug in string printing code: if *any* escaping is enabled we must
631 escape the escape character (backslash) or the resulting string is
635 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
637 *) Disable code workaround for ancient and obsolete Netscape browsers
638 and servers: an attacker can use it in a ciphersuite downgrade attack.
639 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
642 *) Fixed J-PAKE implementation error, originally discovered by
643 Sebastien Martini, further info and confirmation from Stefan
644 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
647 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
649 *) Fix extension code to avoid race conditions which can result in a buffer
650 overrun vulnerability: resumed sessions must not be modified as they can
651 be shared by multiple threads. CVE-2010-3864
654 *) Fix WIN32 build system to correctly link an ENGINE directory into
658 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
660 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
662 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
664 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
666 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
667 context. The operation can be customised via the ctrl mechanism in
668 case ENGINEs want to include additional functionality.
671 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
674 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
675 output hashes compatible with older versions of OpenSSL.
676 [Willy Weisz <weisz@vcpc.univie.ac.at>]
678 *) Fix compression algorithm handling: if resuming a session use the
679 compression algorithm of the resumed session instead of determining
680 it from client hello again. Don't allow server to change algorithm.
683 *) Add load_crls() function to apps tidying load_certs() too. Add option
684 to verify utility to allow additional CRLs to be included.
687 *) Update OCSP request code to permit adding custom headers to the request:
688 some responders need this.
691 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
693 [Julia Lawall <julia@diku.dk>]
695 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
696 needlessly dereferenced structures, used obsolete functions and
697 didn't handle all updated verify codes correctly.
700 *) Disable MD2 in the default configuration.
703 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
704 indicate the initial BIO being pushed or popped. This makes it possible
705 to determine whether the BIO is the one explicitly called or as a result
706 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
707 it handles reference counts correctly and doesn't zero out the I/O bio
708 when it is not being explicitly popped. WARNING: applications which
709 included workarounds for the old buggy behaviour will need to be modified
710 or they could free up already freed BIOs.
713 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
714 renaming to all platforms (within the 0.9.8 branch, this was
715 done conditionally on Netware platforms to avoid a name clash).
716 [Guenter <lists@gknw.net>]
718 *) Add ECDHE and PSK support to DTLS.
719 [Michael Tuexen <tuexen@fh-muenster.de>]
721 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
725 *) Add "missing" function EVP_MD_flags() (without this the only way to
726 retrieve a digest flags is by accessing the structure directly. Update
727 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
728 or cipher is registered as in the "from" argument. Print out all
729 registered digests in the dgst usage message instead of manually
730 attempting to work them out.
733 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
734 this allows the use of compression and extensions. Change default cipher
735 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
736 by default unless an application cipher string requests it.
739 *) Alter match criteria in PKCS12_parse(). It used to try to use local
740 key ids to find matching certificates and keys but some PKCS#12 files
741 don't follow the (somewhat unwritten) rules and this strategy fails.
742 Now just gather all certificates together and the first private key
743 then look for the first certificate that matches the key.
746 *) Support use of registered digest and cipher names for dgst and cipher
747 commands instead of having to add each one as a special case. So now
754 openssl dgst -sha256 foo
756 and this works for ENGINE based algorithms too.
760 *) Update Gost ENGINE to support parameter files.
761 [Victor B. Wagner <vitus@cryptocom.ru>]
763 *) Support GeneralizedTime in ca utility.
764 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
766 *) Enhance the hash format used for certificate directory links. The new
767 form uses the canonical encoding (meaning equivalent names will work
768 even if they aren't identical) and uses SHA1 instead of MD5. This form
769 is incompatible with the older format and as a result c_rehash should
770 be used to rebuild symbolic links.
773 *) Make PKCS#8 the default write format for private keys, replacing the
774 traditional format. This form is standardised, more secure and doesn't
775 include an implicit MD5 dependency.
778 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
779 committed to OpenSSL should pass this lot as a minimum.
782 *) Add session ticket override functionality for use by EAP-FAST.
783 [Jouni Malinen <j@w1.fi>]
785 *) Modify HMAC functions to return a value. Since these can be implemented
786 in an ENGINE errors can occur.
789 *) Type-checked OBJ_bsearch_ex.
792 *) Type-checked OBJ_bsearch. Also some constification necessitated
793 by type-checking. Still to come: TXT_DB, bsearch(?),
794 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
798 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
799 seconds to a tm structure directly, instead of going through OS
800 specific date routines. This avoids any issues with OS routines such
801 as the year 2038 bug. New *_adj() functions for ASN1 time structures
802 and X509_time_adj_ex() to cover the extended range. The existing
803 X509_time_adj() is still usable and will no longer have any date issues.
806 *) Delta CRL support. New use deltas option which will attempt to locate
807 and search any appropriate delta CRLs available.
809 This work was sponsored by Google.
812 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
813 code and add additional score elements. Validate alternate CRL paths
814 as part of the CRL checking and indicate a new error "CRL path validation
815 error" in this case. Applications wanting additional details can use
816 the verify callback and check the new "parent" field. If this is not
817 NULL CRL path validation is taking place. Existing applications wont
818 see this because it requires extended CRL support which is off by
821 This work was sponsored by Google.
824 *) Support for freshest CRL extension.
826 This work was sponsored by Google.
829 *) Initial indirect CRL support. Currently only supported in the CRLs
830 passed directly and not via lookup. Process certificate issuer
831 CRL entry extension and lookup CRL entries by bother issuer name
832 and serial number. Check and process CRL issuer entry in IDP extension.
834 This work was sponsored by Google.
837 *) Add support for distinct certificate and CRL paths. The CRL issuer
838 certificate is validated separately in this case. Only enabled if
839 an extended CRL support flag is set: this flag will enable additional
840 CRL functionality in future.
842 This work was sponsored by Google.
845 *) Add support for policy mappings extension.
847 This work was sponsored by Google.
850 *) Fixes to pathlength constraint, self issued certificate handling,
851 policy processing to align with RFC3280 and PKITS tests.
853 This work was sponsored by Google.
856 *) Support for name constraints certificate extension. DN, email, DNS
857 and URI types are currently supported.
859 This work was sponsored by Google.
862 *) To cater for systems that provide a pointer-based thread ID rather
863 than numeric, deprecate the current numeric thread ID mechanism and
864 replace it with a structure and associated callback type. This
865 mechanism allows a numeric "hash" to be extracted from a thread ID in
866 either case, and on platforms where pointers are larger than 'long',
867 mixing is done to help ensure the numeric 'hash' is usable even if it
868 can't be guaranteed unique. The default mechanism is to use "&errno"
869 as a pointer-based thread ID to distinguish between threads.
871 Applications that want to provide their own thread IDs should now use
872 CRYPTO_THREADID_set_callback() to register a callback that will call
873 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
875 Note that ERR_remove_state() is now deprecated, because it is tied
876 to the assumption that thread IDs are numeric. ERR_remove_state(0)
877 to free the current thread's error state should be replaced by
878 ERR_remove_thread_state(NULL).
880 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
881 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
882 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
883 application was previously providing a numeric thread callback that
884 was inappropriate for distinguishing threads, then uniqueness might
885 have been obtained with &errno that happened immediately in the
886 intermediate development versions of OpenSSL; this is no longer the
887 case, the numeric thread callback will now override the automatic use
889 [Geoff Thorpe, with help from Bodo Moeller]
891 *) Initial support for different CRL issuing certificates. This covers a
892 simple case where the self issued certificates in the chain exist and
893 the real CRL issuer is higher in the existing chain.
895 This work was sponsored by Google.
898 *) Removed effectively defunct crypto/store from the build.
901 *) Revamp of STACK to provide stronger type-checking. Still to come:
902 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
903 ASN1_STRING, CONF_VALUE.
906 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
907 RAM on SSL connections. This option can save about 34k per idle SSL.
910 *) Revamp of LHASH to provide stronger type-checking. Still to come:
911 STACK, TXT_DB, bsearch, qsort.
914 *) Initial support for Cryptographic Message Syntax (aka CMS) based
915 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
916 support for data, signedData, compressedData, digestedData and
917 encryptedData, envelopedData types included. Scripts to check against
918 RFC4134 examples draft and interop and consistency checks of many
919 content types and variants.
922 *) Add options to enc utility to support use of zlib compression BIO.
925 *) Extend mk1mf to support importing of options and assembly language
926 files from Configure script, currently only included in VC-WIN32.
927 The assembly language rules can now optionally generate the source
928 files from the associated perl scripts.
931 *) Implement remaining functionality needed to support GOST ciphersuites.
932 Interop testing has been performed using CryptoPro implementations.
933 [Victor B. Wagner <vitus@cryptocom.ru>]
935 *) s390x assembler pack.
938 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
942 *) Implement Opaque PRF Input TLS extension as specified in
943 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
944 official specification yet and no extension type assignment by
945 IANA exists, this extension (for now) will have to be explicitly
946 enabled when building OpenSSL by providing the extension number
947 to use. For example, specify an option
949 -DTLSEXT_TYPE_opaque_prf_input=0x9527
951 to the "config" or "Configure" script to enable the extension,
952 assuming extension number 0x9527 (which is a completely arbitrary
953 and unofficial assignment based on the MD5 hash of the Internet
954 Draft). Note that by doing so, you potentially lose
955 interoperability with other TLS implementations since these might
956 be using the same extension number for other purposes.
958 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
959 opaque PRF input value to use in the handshake. This will create
960 an interal copy of the length-'len' string at 'src', and will
961 return non-zero for success.
963 To get more control and flexibility, provide a callback function
966 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
967 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
971 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
974 Callback function 'cb' will be called in handshakes, and is
975 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
976 Argument 'arg' is for application purposes (the value as given to
977 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
978 be provided to the callback function). The callback function
979 has to return non-zero to report success: usually 1 to use opaque
980 PRF input just if possible, or 2 to enforce use of the opaque PRF
981 input. In the latter case, the library will abort the handshake
982 if opaque PRF input is not successfully negotiated.
984 Arguments 'peerinput' and 'len' given to the callback function
985 will always be NULL and 0 in the case of a client. A server will
986 see the client's opaque PRF input through these variables if
987 available (NULL and 0 otherwise). Note that if the server
988 provides an opaque PRF input, the length must be the same as the
989 length of the client's opaque PRF input.
991 Note that the callback function will only be called when creating
992 a new session (session resumption can resume whatever was
993 previously negotiated), and will not be called in SSL 2.0
994 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
995 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
996 for applications that need to enforce opaque PRF input.
1000 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1003 [Victor B. Wagner <vitus@cryptocom.ru>]
1005 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1006 RFC4507bis. The encrypted ticket format is an encrypted encoded
1007 SSL_SESSION structure, that way new session features are automatically
1010 If a client application caches session in an SSL_SESSION structure
1011 support is transparent because tickets are now stored in the encoded
1014 The SSL_CTX structure automatically generates keys for ticket
1015 protection in servers so again support should be possible
1016 with no application modification.
1018 If a client or server wishes to disable RFC4507 support then the option
1019 SSL_OP_NO_TICKET can be set.
1021 Add a TLS extension debugging callback to allow the contents of any client
1022 or server extensions to be examined.
1024 This work was sponsored by Google.
1027 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1028 OpenSSL should now compile cleanly on gcc 4.2
1029 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1031 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1032 support including streaming MAC support: this is required for GOST
1033 ciphersuite support.
1034 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1036 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1037 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1038 to output in BER and PEM format.
1041 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1042 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1043 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1044 ENGINE support for HMAC keys which are unextractable. New -mac and
1045 -macopt options to dgst utility.
1048 *) New option -sigopt to dgst utility. Update dgst to use
1049 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1050 alternative signing paramaters such as X9.31 or PSS in the dgst
1054 *) Change ssl_cipher_apply_rule(), the internal function that does
1055 the work each time a ciphersuite string requests enabling
1056 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1057 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1058 the order of disabled ciphersuites such that those ciphersuites
1059 that most recently went from enabled to disabled not only stay
1060 in order with respect to each other, but also have higher priority
1061 than other disabled ciphersuites the next time ciphersuites are
1064 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1065 the same ciphersuites as with "HIGH" alone, but in a specific
1066 order where the PSK ciphersuites come first (since they are the
1067 most recently disabled ciphersuites when "HIGH" is parsed).
1069 Also, change ssl_create_cipher_list() (using this new
1070 funcionality) such that between otherwise identical
1071 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1075 *) Change ssl_create_cipher_list() so that it automatically
1076 arranges the ciphersuites in reasonable order before starting
1077 to process the rule string. Thus, the definition for "DEFAULT"
1078 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1079 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1080 This makes it much easier to arrive at a reasonable default order
1081 in applications for which anonymous ciphers are OK (meaning
1082 that you can't actually use DEFAULT).
1083 [Bodo Moeller; suggested by Victor Duchovni]
1085 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1086 processing) into multiple integers instead of setting
1087 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1088 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1089 (These masks as well as the individual bit definitions are hidden
1090 away into the non-exported interface ssl/ssl_locl.h, so this
1091 change to the definition of the SSL_CIPHER structure shouldn't
1092 affect applications.) This give us more bits for each of these
1093 categories, so there is no longer a need to coagulate AES128 and
1094 AES256 into a single algorithm bit, and to coagulate Camellia128
1095 and Camellia256 into a single algorithm bit, which has led to all
1098 Thus, among other things, the kludge introduced in 0.9.7m and
1099 0.9.8e for masking out AES256 independently of AES128 or masking
1100 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1102 With the change, we also introduce new ciphersuite aliases that
1103 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1107 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1108 Use the leftmost N bytes of the signature input if the input is
1109 larger than the prime q (with N being the size in bytes of q).
1112 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1113 it yet and it is largely untested.
1116 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1119 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1120 some compilers (gcc 4.2 and later) reject their use. Safestack is
1121 reimplemented. Update ASN1 to avoid use of legacy functions.
1124 *) Win32/64 targets are linked with Winsock2.
1127 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1128 to external functions. This can be used to increase CRL handling
1129 efficiency especially when CRLs are very large by (for example) storing
1130 the CRL revoked certificates in a database.
1133 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1134 new CRLs added to a directory can be used. New command line option
1135 -verify_return_error to s_client and s_server. This causes real errors
1136 to be returned by the verify callback instead of carrying on no matter
1137 what. This reflects the way a "real world" verify callback would behave.
1140 *) GOST engine, supporting several GOST algorithms and public key formats.
1141 Kindly donated by Cryptocom.
1144 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1145 partitioned by DP are handled but no indirect CRL or reason partitioning
1146 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1147 selected via a scoring technique which handles IDP and AKID in CRLs.
1150 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1151 will ultimately be used for all verify operations: this will remove the
1152 X509_STORE dependency on certificate verification and allow alternative
1153 lookup methods. X509_STORE based implementations of these two callbacks.
1156 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1157 Modify get_crl() to find a valid (unexpired) CRL if possible.
1160 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1161 this would be called X509_CRL_cmp() but that name is already used by
1162 a function that just compares CRL issuer names. Cache several CRL
1163 extensions in X509_CRL structure and cache CRLDP in X509.
1166 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1167 this maps equivalent X509_NAME structures into a consistent structure.
1168 Name comparison can then be performed rapidly using memcmp().
1171 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1175 *) Allow digests to supply their own micalg string for S/MIME type using
1176 the ctrl EVP_MD_CTRL_MICALG.
1179 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1180 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1181 ctrl. It can then customise the structure before and/or after signing
1185 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1186 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1187 to free up any added signature OIDs.
1190 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1191 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1192 digest and cipher tables. New options added to openssl utility:
1193 list-message-digest-algorithms and list-cipher-algorithms.
1196 *) Change the array representation of binary polynomials: the list
1197 of degrees of non-zero coefficients is now terminated with -1.
1198 Previously it was terminated with 0, which was also part of the
1199 value; thus, the array representation was not applicable to
1200 polynomials where t^0 has coefficient zero. This change makes
1201 the array representation useful in a more general context.
1204 *) Various modifications and fixes to SSL/TLS cipher string
1205 handling. For ECC, the code now distinguishes between fixed ECDH
1206 with RSA certificates on the one hand and with ECDSA certificates
1207 on the other hand, since these are separate ciphersuites. The
1208 unused code for Fortezza ciphersuites has been removed.
1210 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1211 (not "ECDHE"). For consistency with the code for DH
1212 certificates, use of ECDH certificates is now considered ECDH
1213 authentication, not RSA or ECDSA authentication (the latter is
1214 merely the CA's signing algorithm and not actively used in the
1217 The temporary ciphersuite alias "ECCdraft" is no longer
1218 available, and ECC ciphersuites are no longer excluded from "ALL"
1219 and "DEFAULT". The following aliases now exist for RFC 4492
1220 ciphersuites, most of these by analogy with the DH case:
1222 kECDHr - ECDH cert, signed with RSA
1223 kECDHe - ECDH cert, signed with ECDSA
1224 kECDH - ECDH cert (signed with either RSA or ECDSA)
1225 kEECDH - ephemeral ECDH
1226 ECDH - ECDH cert or ephemeral ECDH
1232 AECDH - anonymous ECDH
1233 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1237 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1238 Use correct micalg parameters depending on digest(s) in signed message.
1241 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1242 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1245 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1246 an engine to register a method. Add ENGINE lookups for methods and
1247 functional reference processing.
1250 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1251 EVP_{Sign,Verify}* which allow an application to customise the signature
1255 *) New -resign option to smime utility. This adds one or more signers
1256 to an existing PKCS#7 signedData structure. Also -md option to use an
1257 alternative message digest algorithm for signing.
1260 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1261 create PKCS7 structures containing multiple signers. Update smime
1262 application to support multiple signers.
1265 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1269 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1270 Reorganize PBE internals to lookup from a static table using NIDs,
1271 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1272 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1273 PRF which will be automatically used with PBES2.
1276 *) Replace the algorithm specific calls to generate keys in "req" with the
1280 *) Update PKCS#7 enveloped data routines to use new API. This is now
1281 supported by any public key method supporting the encrypt operation. A
1282 ctrl is added to allow the public key algorithm to examine or modify
1283 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1287 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1288 a default digest type to use. In most cases this will be SHA1 but some
1289 algorithms (such as GOST) need to specify an alternative digest. The
1290 return value indicates how strong the prefernce is 1 means optional and
1291 2 is mandatory (that is it is the only supported type). Modify
1292 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1293 use the default md. Update openssl utilities to use the default digest
1294 type for signing if it is not explicitly indicated.
1297 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1298 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1299 signing method from the key type. This effectively removes the link
1300 between digests and public key types.
1303 *) Add an OID cross reference table and utility functions. Its purpose is to
1304 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1305 rsaEncryption. This will allow some of the algorithm specific hackery
1306 needed to use the correct OID to be removed.
1309 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1310 structures for PKCS7_sign(). They are now set up by the relevant public
1314 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1317 *) Add support for key derivation (agreement) in the API, DH method and
1321 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1322 public and private key formats. As a side effect these add additional
1323 command line functionality not previously available: DSA signatures can be
1324 generated and verified using pkeyutl and DH key support and generation in
1329 [Oliver Tappe <zooey@hirschkaefer.de>]
1331 *) New make target "install_html_docs" installs HTML renditions of the
1333 [Oliver Tappe <zooey@hirschkaefer.de>]
1335 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1336 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1337 support key and parameter generation and add initial key generation
1338 functionality for RSA.
1341 *) Add functions for main EVP_PKEY_method operations. The undocumented
1342 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1343 EVP_PKEY_{encrypt,decrypt}_old.
1346 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1347 key API, doesn't do much yet.
1350 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1351 public key algorithms. New option to openssl utility:
1352 "list-public-key-algorithms" to print out info.
1355 *) Implement the Supported Elliptic Curves Extension for
1356 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1359 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1360 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1363 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1364 utilities such as rsa, dsa, dsaparam etc except they process any key
1368 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1369 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1370 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1374 *) Initial support for pluggable public key ASN1.
1375 De-spaghettify the public key ASN1 handling. Move public and private
1376 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1377 algorithm specific handling to a single module within the relevant
1378 algorithm directory. Add functions to allow (near) opaque processing
1379 of public and private key structures.
1382 *) Implement the Supported Point Formats Extension for
1383 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1386 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1387 for the psk identity [hint] and the psk callback functions to the
1388 SSL_SESSION, SSL and SSL_CTX structure.
1391 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1395 SSL_CTX_use_psk_identity_hint
1396 SSL_get_psk_identity_hint
1397 SSL_get_psk_identity
1398 SSL_use_psk_identity_hint
1400 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1402 *) Add RFC 3161 compliant time stamp request creation, response generation
1403 and response verification functionality.
1404 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1406 *) Add initial support for TLS extensions, specifically for the server_name
1407 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1408 have new members for a host name. The SSL data structure has an
1409 additional member SSL_CTX *initial_ctx so that new sessions can be
1410 stored in that context to allow for session resumption, even after the
1411 SSL has been switched to a new SSL_CTX in reaction to a client's
1412 server_name extension.
1414 New functions (subject to change):
1416 SSL_get_servername()
1417 SSL_get_servername_type()
1420 New CTRL codes and macros (subject to change):
1422 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1423 - SSL_CTX_set_tlsext_servername_callback()
1424 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1425 - SSL_CTX_set_tlsext_servername_arg()
1426 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1428 openssl s_client has a new '-servername ...' option.
1430 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1431 '-key2 ...', '-servername_fatal' (subject to change). This allows
1432 testing the HostName extension for a specific single host name ('-cert'
1433 and '-key' remain fallbacks for handshakes without HostName
1434 negotiation). If the unrecogninzed_name alert has to be sent, this by
1435 default is a warning; it becomes fatal with the '-servername_fatal'
1438 [Peter Sylvester, Remy Allais, Christophe Renou]
1440 *) Whirlpool hash implementation is added.
1443 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1444 bn(64,32). Because of instruction set limitations it doesn't have
1445 any negative impact on performance. This was done mostly in order
1446 to make it possible to share assembler modules, such as bn_mul_mont
1447 implementations, between 32- and 64-bit builds without hassle.
1450 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1451 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1455 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1456 dedicated Montgomery multiplication procedure, is introduced.
1457 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1458 "64-bit" performance on certain 32-bit targets.
1461 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1462 in SSL structures. New SSL ctrl to set maximum send fragment size.
1463 Save memory by seeting the I/O buffer sizes dynamically instead of
1464 using the maximum available value.
1467 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1468 in addition to the text details.
1471 *) Very, very preliminary EXPERIMENTAL support for printing of general
1472 ASN1 structures. This currently produces rather ugly output and doesn't
1473 handle several customised structures at all.
1476 *) Integrated support for PVK file format and some related formats such
1477 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1478 these in the 'rsa' and 'dsa' utilities.
1481 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1484 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1485 place for the (very old) "NETSCAPE" format certificates which are now
1486 handled using new ASN1 code equivalents.
1489 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1490 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1491 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1494 *) Modify CRL distribution points extension code to print out previously
1495 unsupported fields. Enhance extension setting code to allow setting of
1499 *) Add print and set support for Issuing Distribution Point CRL extension.
1502 *) Change 'Configure' script to enable Camellia by default.
1505 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1507 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1509 This addresses the flaw in CBC record processing discovered by
1510 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1511 at: http://www.isg.rhul.ac.uk/tls/
1513 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1514 Security Group at Royal Holloway, University of London
1515 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1516 Emilia Käsper for the initial patch.
1518 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1520 *) Return an error when checking OCSP signatures when key is NULL.
1521 This fixes a DoS attack. (CVE-2013-0166)
1524 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1525 the right response is stapled. Also change SSL_get_certificate()
1526 so it returns the certificate actually sent.
1527 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1528 (This is a backport)
1529 [Rob Stradling <rob.stradling@comodo.com>]
1531 *) Fix possible deadlock when decoding public keys.
1534 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1536 *) Sanity check record length before skipping explicit IV in DTLS
1539 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1540 fuzzing as a service testing platform.
1544 *) Initialise tkeylen properly when encrypting CMS messages.
1545 Thanks to Solar Designer of Openwall for reporting this issue.
1548 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1550 *) The fix for CVE-2012-2110 did not take into account that the
1551 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1552 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1553 rejecting negative len parameter. (CVE-2012-2131)
1554 [Tomas Hoger <thoger@redhat.com>]
1556 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1558 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1559 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1560 in CRYPTO_realloc_clean.
1562 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1563 issue and to Adam Langley <agl@chromium.org> for fixing it.
1565 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1567 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1569 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1570 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1571 content decryption and always return the same error. Note: this attack
1572 needs on average 2^20 messages so it only affects automated senders. The
1573 old behaviour can be reenabled in the CMS code by setting the
1574 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1575 an MMA defence is not necessary.
1576 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1577 this issue. (CVE-2012-0884)
1580 *) Fix CVE-2011-4619: make sure we really are receiving a
1581 client hello before rejecting multiple SGC restarts. Thanks to
1582 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1585 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1587 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1588 Thanks to Antonio Martin, Enterprise Secure Access Research and
1589 Development, Cisco Systems, Inc. for discovering this bug and
1590 preparing a fix. (CVE-2012-0050)
1593 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1595 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1596 of the Vaudenay padding oracle attack on CBC mode encryption
1597 which enables an efficient plaintext recovery attack against
1598 the OpenSSL implementation of DTLS. Their attack exploits timing
1599 differences arising during decryption processing. A research
1600 paper describing this attack can be found at:
1601 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1602 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1603 Security Group at Royal Holloway, University of London
1604 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1605 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1606 for preparing the fix. (CVE-2011-4108)
1607 [Robin Seggelmann, Michael Tuexen]
1609 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1610 [Ben Laurie, Kasper <ekasper@google.com>]
1612 *) Clear bytes used for block padding of SSL 3.0 records.
1614 [Adam Langley (Google)]
1616 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1617 Kadianakis <desnacked@gmail.com> for discovering this issue and
1618 Adam Langley for preparing the fix. (CVE-2011-4619)
1619 [Adam Langley (Google)]
1621 *) Prevent malformed RFC3779 data triggering an assertion failure.
1622 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1623 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1624 [Rob Austein <sra@hactrn.net>]
1626 *) Fix ssl_ciph.c set-up race.
1627 [Adam Langley (Google)]
1629 *) Fix spurious failures in ecdsatest.c.
1630 [Emilia Käsper (Google)]
1632 *) Fix the BIO_f_buffer() implementation (which was mixing different
1633 interpretations of the '..._len' fields).
1634 [Adam Langley (Google)]
1636 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1637 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1638 threads won't reuse the same blinding coefficients.
1640 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1641 lock to call BN_BLINDING_invert_ex, and avoids one use of
1642 BN_BLINDING_update for each BN_BLINDING structure (previously,
1643 the last update always remained unused).
1644 [Emilia Käsper (Google)]
1646 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1647 for multi-threaded use of ECDH.
1648 [Adam Langley (Google)]
1650 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1653 *) Add protection against ECDSA timing attacks as mentioned in the paper
1654 by Billy Bob Brumley and Nicola Tuveri, see:
1656 http://eprint.iacr.org/2011/232.pdf
1658 [Billy Bob Brumley and Nicola Tuveri]
1660 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1662 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1663 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1665 *) Fix bug in string printing code: if *any* escaping is enabled we must
1666 escape the escape character (backslash) or the resulting string is
1670 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1672 *) Disable code workaround for ancient and obsolete Netscape browsers
1673 and servers: an attacker can use it in a ciphersuite downgrade attack.
1674 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1677 *) Fixed J-PAKE implementation error, originally discovered by
1678 Sebastien Martini, further info and confirmation from Stefan
1679 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1682 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1684 *) Fix extension code to avoid race conditions which can result in a buffer
1685 overrun vulnerability: resumed sessions must not be modified as they can
1686 be shared by multiple threads. CVE-2010-3864
1689 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1692 *) Don't reencode certificate when calculating signature: cache and use
1693 the original encoding instead. This makes signature verification of
1694 some broken encodings work correctly.
1697 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1698 is also one of the inputs.
1699 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1701 *) Don't repeatedly append PBE algorithms to table if they already exist.
1702 Sort table on each new add. This effectively makes the table read only
1703 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1707 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1709 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1712 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1713 access or freeing data twice (CVE-2010-0742)
1714 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1716 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1717 common in certificates and some applications which only call
1718 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1722 Reduce copying into .apps and .test in makevms.com
1723 Don't try to use blank CA certificate in CA.com
1724 Allow use of C files from original directories in maketests.com
1725 [Steven M. Schweda" <sms@antinode.info>]
1727 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1729 *) When rejecting SSL/TLS records due to an incorrect version number, never
1730 update s->server with a new major version number. As of
1731 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1732 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1733 the previous behavior could result in a read attempt at NULL when
1734 receiving specific incorrect SSL/TLS records once record payload
1735 protection is active. (CVE-2010-0740)
1736 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1738 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1739 could be crashed if the relevant tables were not present (e.g. chrooted).
1740 [Tomas Hoger <thoger@redhat.com>]
1742 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1744 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1745 [Martin Olsson, Neel Mehta]
1747 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1748 accommodate for stack sorting, always a write lock!).
1751 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1752 excessive delays in the RAND_poll(): over a minute. As a workaround
1753 include a time check in the inner Heap32Next loop too.
1756 *) The code that handled flushing of data in SSL/TLS originally used the
1757 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1758 the problem outlined in PR#1949. The fix suggested there however can
1759 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1760 of Apache). So instead simplify the code to flush unconditionally.
1761 This should be fine since flushing with no data to flush is a no op.
1764 *) Handle TLS versions 2.0 and later properly and correctly use the
1765 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1766 off ancient servers have a habit of sticking around for a while...
1769 *) Modify compression code so it frees up structures without using the
1770 ex_data callbacks. This works around a problem where some applications
1771 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1772 restarting) then use compression (e.g. SSL with compression) later.
1773 This results in significant per-connection memory leaks and
1774 has caused some security issues including CVE-2008-1678 and
1778 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1779 change when encrypting or decrypting.
1782 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1783 connect and renegotiate with servers which do not support RI.
1784 Until RI is more widely deployed this option is enabled by default.
1787 *) Add "missing" ssl ctrls to clear options and mode.
1790 *) If client attempts to renegotiate and doesn't support RI respond with
1791 a no_renegotiation alert as required by RFC5746. Some renegotiating
1792 TLS clients will continue a connection gracefully when they receive
1793 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1794 waiting for a server hello which it will never receive. Now we treat a
1795 received no_renegotiation alert as a fatal error. This is because
1796 applications requesting a renegotiation might well expect it to succeed
1797 and would have no code in place to handle the server denying it so the
1798 only safe thing to do is to terminate the connection.
1801 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1802 peer supports secure renegotiation and 0 otherwise. Print out peer
1803 renegotiation support in s_client/s_server.
1806 *) Replace the highly broken and deprecated SPKAC certification method with
1807 the updated NID creation version. This should correctly handle UTF8.
1810 *) Implement RFC5746. Re-enable renegotiation but require the extension
1811 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1812 turns out to be a bad idea. It has been replaced by
1813 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1814 SSL_CTX_set_options(). This is really not recommended unless you
1815 know what you are doing.
1816 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1818 *) Fixes to stateless session resumption handling. Use initial_ctx when
1819 issuing and attempting to decrypt tickets in case it has changed during
1820 servername handling. Use a non-zero length session ID when attempting
1821 stateless session resumption: this makes it possible to determine if
1822 a resumption has occurred immediately after receiving server hello
1823 (several places in OpenSSL subtly assume this) instead of later in
1827 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1828 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1829 fixes for a few places where the return code is not checked
1831 [Julia Lawall <julia@diku.dk>]
1833 *) Add --strict-warnings option to Configure script to include devteam
1834 warnings in other configurations.
1837 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1838 makes it possible to install openssl libraries in locations which
1839 have names other than "lib", for example "/usr/lib64" which some
1841 [Steve Henson, based on patch from Jeremy Utley]
1843 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1844 X690 8.9.12 and can produce some misleading textual output of OIDs.
1845 [Steve Henson, reported by Dan Kaminsky]
1847 *) Delete MD2 from algorithm tables. This follows the recommendation in
1848 several standards that it is not used in new applications due to
1849 several cryptographic weaknesses. For binary compatibility reasons
1850 the MD2 API is still compiled in by default.
1853 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1857 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1858 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1860 [Guenter <lists@gknw.net>]
1862 *) Fix the server certificate chain building code to use X509_verify_cert(),
1863 it used to have an ad-hoc builder which was unable to cope with anything
1864 other than a simple chain.
1865 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1867 *) Don't check self signed certificate signatures in X509_verify_cert()
1868 by default (a flag can override this): it just wastes time without
1869 adding any security. As a useful side effect self signed root CAs
1870 with non-FIPS digests are now usable in FIPS mode.
1873 *) In dtls1_process_out_of_seq_message() the check if the current message
1874 is already buffered was missing. For every new message was memory
1875 allocated, allowing an attacker to perform an denial of service attack
1876 with sending out of seq handshake messages until there is no memory
1877 left. Additionally every future messege was buffered, even if the
1878 sequence number made no sense and would be part of another handshake.
1879 So only messages with sequence numbers less than 10 in advance will be
1880 buffered. (CVE-2009-1378)
1881 [Robin Seggelmann, discovered by Daniel Mentz]
1883 *) Records are buffered if they arrive with a future epoch to be
1884 processed after finishing the corresponding handshake. There is
1885 currently no limitation to this buffer allowing an attacker to perform
1886 a DOS attack with sending records with future epochs until there is no
1887 memory left. This patch adds the pqueue_size() function to detemine
1888 the size of a buffer and limits the record buffer to 100 entries.
1890 [Robin Seggelmann, discovered by Daniel Mentz]
1892 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1893 parent structure is freed. (CVE-2009-1379)
1896 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1897 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1900 [Ilya O. <vrghost@gmail.com>]
1902 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1904 *) Disable renegotiation completely - this fixes a severe security
1905 problem (CVE-2009-3555) at the cost of breaking all
1906 renegotiation. Renegotiation can be re-enabled by setting
1907 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1908 run-time. This is really not recommended unless you know what
1912 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1914 *) Don't set val to NULL when freeing up structures, it is freed up by
1915 underlying code. If sizeof(void *) > sizeof(long) this can result in
1916 zeroing past the valid field. (CVE-2009-0789)
1917 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1919 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1920 checked correctly. This would allow some invalid signed attributes to
1921 appear to verify correctly. (CVE-2009-0591)
1922 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1924 *) Reject UniversalString and BMPString types with invalid lengths. This
1925 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1926 a legal length. (CVE-2009-0590)
1929 *) Set S/MIME signing as the default purpose rather than setting it
1930 unconditionally. This allows applications to override it at the store
1934 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1935 to handle some structures.
1938 *) Improve efficiency of mem_gets: don't search whole buffer each time
1940 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1942 *) New -hex option for openssl rand.
1945 *) Print out UTF8String and NumericString when parsing ASN1.
1948 *) Support NumericString type for name components.
1951 *) Allow CC in the environment to override the automatically chosen
1952 compiler. Note that nothing is done to ensure flags work with the
1956 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1958 *) Properly check EVP_VerifyFinal() and similar return values
1960 [Ben Laurie, Bodo Moeller, Google Security Team]
1962 *) Enable TLS extensions by default.
1965 *) Allow the CHIL engine to be loaded, whether the application is
1966 multithreaded or not. (This does not release the developer from the
1967 obligation to set up the dynamic locking callbacks.)
1968 [Sander Temme <sander@temme.net>]
1970 *) Use correct exit code if there is an error in dgst command.
1971 [Steve Henson; problem pointed out by Roland Dirlewanger]
1973 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1974 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1977 *) Add experimental JPAKE support, including demo authentication in
1978 s_client and s_server.
1981 *) Set the comparison function in v3_addr_canonize().
1982 [Rob Austein <sra@hactrn.net>]
1984 *) Add support for XMPP STARTTLS in s_client.
1985 [Philip Paeps <philip@freebsd.org>]
1987 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1988 to ensure that even with this option, only ciphersuites in the
1989 server's preference list will be accepted. (Note that the option
1990 applies only when resuming a session, so the earlier behavior was
1991 just about the algorithm choice for symmetric cryptography.)
1994 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1996 *) Fix NULL pointer dereference if a DTLS server received
1997 ChangeCipherSpec as first record (CVE-2009-1386).
2000 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2001 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2004 *) The fix in 0.9.8c that supposedly got rid of unsafe
2005 double-checked locking was incomplete for RSA blinding,
2006 addressing just one layer of what turns out to have been
2007 doubly unsafe triple-checked locking.
2009 So now fix this for real by retiring the MONT_HELPER macro
2010 in crypto/rsa/rsa_eay.c.
2012 [Bodo Moeller; problem pointed out by Marius Schilder]
2014 *) Various precautionary measures:
2016 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2018 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2019 (NB: This would require knowledge of the secret session ticket key
2020 to exploit, in which case you'd be SOL either way.)
2022 - Change bn_nist.c so that it will properly handle input BIGNUMs
2023 outside the expected range.
2025 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2028 [Neel Mehta, Bodo Moeller]
2030 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2031 the load fails. Useful for distros.
2032 [Ben Laurie and the FreeBSD team]
2034 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2037 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2040 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2042 This work was sponsored by Logica.
2045 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2046 keystores. Support for SSL/TLS client authentication too.
2047 Not compiled unless enable-capieng specified to Configure.
2049 This work was sponsored by Logica.
2052 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2053 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2054 attribute creation routines such as certifcate requests and PKCS#12
2058 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2060 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2061 handshake which could lead to a cilent crash as found using the
2062 Codenomicon TLS test suite (CVE-2008-1672)
2063 [Steve Henson, Mark Cox]
2065 *) Fix double free in TLS server name extensions which could lead to
2066 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2069 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2071 Clear the error queue to ensure that error entries left from
2072 older function calls do not interfere with the correct operation.
2073 [Lutz Jaenicke, Erik de Castro Lopo]
2075 *) Remove root CA certificates of commercial CAs:
2077 The OpenSSL project does not recommend any specific CA and does not
2078 have any policy with respect to including or excluding any CA.
2079 Therefore it does not make any sense to ship an arbitrary selection
2080 of root CA certificates with the OpenSSL software.
2083 *) RSA OAEP patches to fix two separate invalid memory reads.
2084 The first one involves inputs when 'lzero' is greater than
2085 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2086 before the beginning of from). The second one involves inputs where
2087 the 'db' section contains nothing but zeroes (there is a one-byte
2088 invalid read after the end of 'db').
2089 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2091 *) Partial backport from 0.9.9-dev:
2093 Introduce bn_mul_mont (dedicated Montgomery multiplication
2094 procedure) as a candidate for BIGNUM assembler implementation.
2095 While 0.9.9-dev uses assembler for various architectures, only
2096 x86_64 is available by default here in the 0.9.8 branch, and
2097 32-bit x86 is available through a compile-time setting.
2099 To try the 32-bit x86 assembler implementation, use Configure
2100 option "enable-montasm" (which exists only for this backport).
2102 As "enable-montasm" for 32-bit x86 disclaims code stability
2103 anyway, in this constellation we activate additional code
2104 backported from 0.9.9-dev for further performance improvements,
2105 namely BN_from_montgomery_word. (To enable this otherwise,
2106 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2108 [Andy Polyakov (backport partially by Bodo Moeller)]
2110 *) Add TLS session ticket callback. This allows an application to set
2111 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2112 values. This is useful for key rollover for example where several key
2113 sets may exist with different names.
2116 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2117 This was broken until now in 0.9.8 releases, such that the only way
2118 a registered ENGINE could be used (assuming it initialises
2119 successfully on the host) was to explicitly set it as the default
2120 for the relevant algorithms. This is in contradiction with 0.9.7
2121 behaviour and the documentation. With this fix, when an ENGINE is
2122 registered into a given algorithm's table of implementations, the
2123 'uptodate' flag is reset so that auto-discovery will be used next
2124 time a new context for that algorithm attempts to select an
2126 [Ian Lister (tweaked by Geoff Thorpe)]
2128 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2129 implemention in the following ways:
2131 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2134 Lack of BER streaming support means one pass streaming processing is
2135 only supported if data is detached: setting the streaming flag is
2136 ignored for embedded content.
2138 CMS support is disabled by default and must be explicitly enabled
2139 with the enable-cms configuration option.
2142 *) Update the GMP engine glue to do direct copies between BIGNUM and
2143 mpz_t when openssl and GMP use the same limb size. Otherwise the
2144 existing "conversion via a text string export" trick is still used.
2145 [Paul Sheer <paulsheer@gmail.com>]
2147 *) Zlib compression BIO. This is a filter BIO which compressed and
2148 uncompresses any data passed through it.
2151 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2152 RFC3394 compatible AES key wrapping.
2155 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2156 sets string data without copying. X509_ALGOR_set0() and
2157 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2158 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2159 from an X509_ATTRIBUTE structure optionally checking it occurs only
2160 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2164 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2165 to get the expected BN_FLG_CONSTTIME behavior.
2166 [Bodo Moeller (Google)]
2170 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2171 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2172 - added some more tests to do_tests.pl
2173 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2174 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2175 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2176 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2177 - various changes to netware.pl to enable gcc-cross builds on Win32
2179 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2180 - various changes to fix missing prototype warnings
2181 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2182 - added AES, WHIRLPOOL and CPUID assembler code to build files
2183 - added missing AES assembler make rules to mk1mf.pl
2184 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2185 [Guenter Knauf <eflash@gmx.net>]
2187 *) Implement certificate status request TLS extension defined in RFC3546.
2188 A client can set the appropriate parameters and receive the encoded
2189 OCSP response via a callback. A server can query the supplied parameters
2190 and set the encoded OCSP response in the callback. Add simplified examples
2191 to s_client and s_server.
2194 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2196 *) Fix various bugs:
2197 + Binary incompatibility of ssl_ctx_st structure
2198 + DTLS interoperation with non-compliant servers
2199 + Don't call get_session_cb() without proposed session
2200 + Fix ia64 assembler code
2201 [Andy Polyakov, Steve Henson]
2203 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2205 *) DTLS Handshake overhaul. There were longstanding issues with
2206 OpenSSL DTLS implementation, which were making it impossible for
2207 RFC 4347 compliant client to communicate with OpenSSL server.
2208 Unfortunately just fixing these incompatibilities would "cut off"
2209 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2210 server keeps tolerating non RFC compliant syntax. The opposite is
2211 not true, 0.9.8f client can not communicate with earlier server.
2212 This update even addresses CVE-2007-4995.
2215 *) Changes to avoid need for function casts in OpenSSL: some compilers
2216 (gcc 4.2 and later) reject their use.
2217 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2220 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2221 RFC4507bis. The encrypted ticket format is an encrypted encoded
2222 SSL_SESSION structure, that way new session features are automatically
2225 If a client application caches session in an SSL_SESSION structure
2226 support is transparent because tickets are now stored in the encoded
2229 The SSL_CTX structure automatically generates keys for ticket
2230 protection in servers so again support should be possible
2231 with no application modification.
2233 If a client or server wishes to disable RFC4507 support then the option
2234 SSL_OP_NO_TICKET can be set.
2236 Add a TLS extension debugging callback to allow the contents of any client
2237 or server extensions to be examined.
2239 This work was sponsored by Google.
2242 *) Add initial support for TLS extensions, specifically for the server_name
2243 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2244 have new members for a host name. The SSL data structure has an
2245 additional member SSL_CTX *initial_ctx so that new sessions can be
2246 stored in that context to allow for session resumption, even after the
2247 SSL has been switched to a new SSL_CTX in reaction to a client's
2248 server_name extension.
2250 New functions (subject to change):
2252 SSL_get_servername()
2253 SSL_get_servername_type()
2256 New CTRL codes and macros (subject to change):
2258 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2259 - SSL_CTX_set_tlsext_servername_callback()
2260 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2261 - SSL_CTX_set_tlsext_servername_arg()
2262 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2264 openssl s_client has a new '-servername ...' option.
2266 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2267 '-key2 ...', '-servername_fatal' (subject to change). This allows
2268 testing the HostName extension for a specific single host name ('-cert'
2269 and '-key' remain fallbacks for handshakes without HostName
2270 negotiation). If the unrecogninzed_name alert has to be sent, this by
2271 default is a warning; it becomes fatal with the '-servername_fatal'
2274 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2276 *) Add AES and SSE2 assembly language support to VC++ build.
2279 *) Mitigate attack on final subtraction in Montgomery reduction.
2282 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2283 (which previously caused an internal error).
2286 *) Squeeze another 10% out of IGE mode when in != out.
2289 *) AES IGE mode speedup.
2290 [Dean Gaudet (Google)]
2292 *) Add the Korean symmetric 128-bit cipher SEED (see
2293 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2294 add SEED ciphersuites from RFC 4162:
2296 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2297 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2298 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2299 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2301 To minimize changes between patchlevels in the OpenSSL 0.9.8
2302 series, SEED remains excluded from compilation unless OpenSSL
2303 is configured with 'enable-seed'.
2304 [KISA, Bodo Moeller]
2306 *) Mitigate branch prediction attacks, which can be practical if a
2307 single processor is shared, allowing a spy process to extract
2308 information. For detailed background information, see
2309 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2310 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2311 and Necessary Software Countermeasures"). The core of the change
2312 are new versions BN_div_no_branch() and
2313 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2314 respectively, which are slower, but avoid the security-relevant
2315 conditional branches. These are automatically called by BN_div()
2316 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2317 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2318 remove a conditional branch.
2320 BN_FLG_CONSTTIME is the new name for the previous
2321 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2322 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2323 in the exponent causes BN_mod_exp_mont() to use the alternative
2324 implementation in BN_mod_exp_mont_consttime().) The old name
2325 remains as a deprecated alias.
2327 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2328 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2329 constant-time implementations for more than just exponentiation.
2330 Here too the old name is kept as a deprecated alias.
2332 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2333 the BN_BLINDING structure gets an independent copy of the
2334 modulus. This means that the previous "BIGNUM *m" argument to
2335 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2336 essentially becomes "const BIGNUM *m", although we can't actually
2337 change this in the header file before 0.9.9. It allows
2338 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2339 enable BN_FLG_CONSTTIME.
2341 [Matthew D Wood (Intel Corp)]
2343 *) In the SSL/TLS server implementation, be strict about session ID
2344 context matching (which matters if an application uses a single
2345 external cache for different purposes). Previously,
2346 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2347 set. This did ensure strict client verification, but meant that,
2348 with applications using a single external cache for quite
2349 different requirements, clients could circumvent ciphersuite
2350 restrictions for a given session ID context by starting a session
2351 in a different context.
2354 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2355 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2356 authentication-only ciphersuites.
2359 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2360 not complete and could lead to a possible single byte overflow
2361 (CVE-2007-5135) [Ben Laurie]
2363 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2365 *) Since AES128 and AES256 (and similarly Camellia128 and
2366 Camellia256) share a single mask bit in the logic of
2367 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2368 kludge to work properly if AES128 is available and AES256 isn't
2369 (or if Camellia128 is available and Camellia256 isn't).
2372 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2373 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2374 When a point or a seed is encoded in a BIT STRING, we need to
2375 prevent the removal of trailing zero bits to get the proper DER
2376 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2377 of a NamedBitList, for which trailing 0 bits need to be removed.)
2380 *) Have SSL/TLS server implementation tolerate "mismatched" record
2381 protocol version while receiving ClientHello even if the
2382 ClientHello is fragmented. (The server can't insist on the
2383 particular protocol version it has chosen before the ServerHello
2384 message has informed the client about his choice.)
2387 *) Add RFC 3779 support.
2388 [Rob Austein for ARIN, Ben Laurie]
2390 *) Load error codes if they are not already present instead of using a
2391 static variable. This allows them to be cleanly unloaded and reloaded.
2392 Improve header file function name parsing.
2395 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2396 or CAPABILITY handshake as required by RFCs.
2399 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2401 *) Introduce limits to prevent malicious keys being able to
2402 cause a denial of service. (CVE-2006-2940)
2403 [Steve Henson, Bodo Moeller]
2405 *) Fix ASN.1 parsing of certain invalid structures that can result
2406 in a denial of service. (CVE-2006-2937) [Steve Henson]
2408 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2409 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2411 *) Fix SSL client code which could crash if connecting to a
2412 malicious SSLv2 server. (CVE-2006-4343)
2413 [Tavis Ormandy and Will Drewry, Google Security Team]
2415 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2416 match only those. Before that, "AES256-SHA" would be interpreted
2417 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2418 the same strength classification in 0.9.7h) as we currently only
2419 have a single AES bit in the ciphersuite description bitmap.
2420 That change, however, also applied to ciphersuite strings such as
2421 "RC4-MD5" that intentionally matched multiple ciphersuites --
2422 namely, SSL 2.0 ciphersuites in addition to the more common ones
2423 from SSL 3.0/TLS 1.0.
2425 So we change the selection algorithm again: Naming an explicit
2426 ciphersuite selects this one ciphersuite, and any other similar
2427 ciphersuite (same bitmap) from *other* protocol versions.
2428 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2429 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2431 Since SSL 2.0 does not have any ciphersuites for which the
2432 128/256 bit distinction would be relevant, this works for now.
2433 The proper fix will be to use different bits for AES128 and
2434 AES256, which would have avoided the problems from the beginning;
2435 however, bits are scarce, so we can only do this in a new release
2436 (not just a patchlevel) when we can change the SSL_CIPHER
2437 definition to split the single 'unsigned long mask' bitmap into
2438 multiple values to extend the available space.
2442 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2444 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2445 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2447 *) Add AES IGE and biIGE modes.
2450 *) Change the Unix randomness entropy gathering to use poll() when
2451 possible instead of select(), since the latter has some
2452 undesirable limitations.
2453 [Darryl Miles via Richard Levitte and Bodo Moeller]
2455 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2456 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2457 cannot be implicitly activated as part of, e.g., the "AES" alias.
2458 However, please upgrade to OpenSSL 0.9.9[-dev] for
2459 non-experimental use of the ECC ciphersuites to get TLS extension
2460 support, which is required for curve and point format negotiation
2461 to avoid potential handshake problems.
2464 *) Disable rogue ciphersuites:
2466 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2467 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2468 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2470 The latter two were purportedly from
2471 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2474 Also deactivate the remaining ciphersuites from
2475 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2476 unofficial, and the ID has long expired.
2479 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2480 dual-core machines) and other potential thread-safety issues.
2483 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2484 versions), which is now available for royalty-free use
2485 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2486 Also, add Camellia TLS ciphersuites from RFC 4132.
2488 To minimize changes between patchlevels in the OpenSSL 0.9.8
2489 series, Camellia remains excluded from compilation unless OpenSSL
2490 is configured with 'enable-camellia'.
2493 *) Disable the padding bug check when compression is in use. The padding
2494 bug check assumes the first packet is of even length, this is not
2495 necessarily true if compresssion is enabled and can result in false
2496 positives causing handshake failure. The actual bug test is ancient
2497 code so it is hoped that implementations will either have fixed it by
2498 now or any which still have the bug do not support compression.
2501 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2503 *) When applying a cipher rule check to see if string match is an explicit
2504 cipher suite and only match that one cipher suite if it is.
2507 *) Link in manifests for VC++ if needed.
2508 [Austin Ziegler <halostatue@gmail.com>]
2510 *) Update support for ECC-based TLS ciphersuites according to
2511 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2512 TLS extensions, which are supported starting with the 0.9.9
2513 branch, not in the OpenSSL 0.9.8 branch).
2516 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2517 opaque EVP_CIPHER_CTX handling.
2520 *) Fixes and enhancements to zlib compression code. We now only use
2521 "zlib1.dll" and use the default __cdecl calling convention on Win32
2522 to conform with the standards mentioned here:
2523 http://www.zlib.net/DLL_FAQ.txt
2524 Static zlib linking now works on Windows and the new --with-zlib-include
2525 --with-zlib-lib options to Configure can be used to supply the location
2526 of the headers and library. Gracefully handle case where zlib library
2530 *) Several fixes and enhancements to the OID generation code. The old code
2531 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2532 handle numbers larger than ULONG_MAX, truncated printing and had a
2533 non standard OBJ_obj2txt() behaviour.
2536 *) Add support for building of engines under engine/ as shared libraries
2537 under VC++ build system.
2540 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2541 Hopefully, we will not see any false combination of paths any more.
2544 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2546 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2547 (part of SSL_OP_ALL). This option used to disable the
2548 countermeasure against man-in-the-middle protocol-version
2549 rollback in the SSL 2.0 server implementation, which is a bad
2550 idea. (CVE-2005-2969)
2552 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2553 for Information Security, National Institute of Advanced Industrial
2554 Science and Technology [AIST], Japan)]
2556 *) Add two function to clear and return the verify parameter flags.
2559 *) Keep cipherlists sorted in the source instead of sorting them at
2560 runtime, thus removing the need for a lock.
2563 *) Avoid some small subgroup attacks in Diffie-Hellman.
2564 [Nick Mathewson and Ben Laurie]
2566 *) Add functions for well-known primes.
2569 *) Extended Windows CE support.
2570 [Satoshi Nakamura and Andy Polyakov]
2572 *) Initialize SSL_METHOD structures at compile time instead of during
2573 runtime, thus removing the need for a lock.
2576 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2577 attempting to decrypt each encrypted key in turn. Add support to
2581 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2583 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2586 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2589 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2590 key into the same file any more.
2593 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2596 *) Add -utf8 command line and config file option to 'ca'.
2597 [Stefan <stf@udoma.org]
2599 *) Removed the macro des_crypt(), as it seems to conflict with some
2600 libraries. Use DES_crypt().
2603 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2604 involves renaming the source and generated shared-libs for
2605 both. The engines will accept the corrected or legacy ids
2606 ('ncipher' and '4758_cca' respectively) when binding. NB,
2607 this only applies when building 'shared'.
2608 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2610 *) Add attribute functions to EVP_PKEY structure. Modify
2611 PKCS12_create() to recognize a CSP name attribute and
2612 use it. Make -CSP option work again in pkcs12 utility.
2615 *) Add new functionality to the bn blinding code:
2616 - automatic re-creation of the BN_BLINDING parameters after
2617 a fixed number of uses (currently 32)
2618 - add new function for parameter creation
2619 - introduce flags to control the update behaviour of the
2620 BN_BLINDING parameters
2621 - hide BN_BLINDING structure
2622 Add a second BN_BLINDING slot to the RSA structure to improve
2623 performance when a single RSA object is shared among several
2627 *) Add support for DTLS.
2628 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2630 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2631 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2634 *) Remove buggy and incompletet DH cert support from
2635 ssl/ssl_rsa.c and ssl/s3_both.c
2638 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2639 the apps/openssl applications.
2642 *) Compile clean with "-Wall -Wmissing-prototypes
2643 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2644 DEBUG_SAFESTACK must also be set.
2647 *) Change ./Configure so that certain algorithms can be disabled by default.
2648 The new counterpiece to "no-xxx" is "enable-xxx".
2650 The patented RC5 and MDC2 algorithms will now be disabled unless
2651 "enable-rc5" and "enable-mdc2", respectively, are specified.
2653 (IDEA remains enabled despite being patented. This is because IDEA
2654 is frequently required for interoperability, and there is no license
2655 fee for non-commercial use. As before, "no-idea" can be used to
2656 avoid this algorithm.)
2660 *) Add processing of proxy certificates (see RFC 3820). This work was
2661 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2662 EGEE (Enabling Grids for E-science in Europe).
2665 *) RC4 performance overhaul on modern architectures/implementations, such
2666 as Intel P4, IA-64 and AMD64.
2669 *) New utility extract-section.pl. This can be used specify an alternative
2670 section number in a pod file instead of having to treat each file as
2671 a separate case in Makefile. This can be done by adding two lines to the
2674 =for comment openssl_section:XXX
2676 The blank line is mandatory.
2680 *) New arguments -certform, -keyform and -pass for s_client and s_server
2681 to allow alternative format key and certificate files and passphrase
2685 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2686 update associated structures and add various utility functions.
2688 Add new policy related verify parameters, include policy checking in
2689 standard verify code. Enhance 'smime' application with extra parameters
2690 to support policy checking and print out.
2693 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2694 Nehemiah processors. These extensions support AES encryption in hardware
2695 as well as RNG (though RNG support is currently disabled).
2696 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2698 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2701 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2702 [Andy Polyakov and a number of other people]
2704 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2705 implementation contributed by IBM.
2706 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2708 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2709 exponent rather than 'unsigned long'. There is a corresponding change to
2710 the new 'rsa_keygen' element of the RSA_METHOD structure.
2711 [Jelte Jansen, Geoff Thorpe]
2713 *) Functionality for creating the initial serial number file is now
2714 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2716 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2717 number file to 1, which is bound to cause problems. To avoid
2718 the problems while respecting compatibility between different 0.9.7
2719 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2720 CA.pl for serial number initialization. With the new release 0.9.8,
2721 we can fix the problem directly in the 'ca' utility.)
2724 *) Reduced header interdepencies by declaring more opaque objects in
2725 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2726 give fewer recursive includes, which could break lazy source code - so
2727 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2728 developers should define this symbol when building and using openssl to
2729 ensure they track the recommended behaviour, interfaces, [etc], but
2730 backwards-compatible behaviour prevails when this isn't defined.
2733 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2736 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2737 This will generate a random key of the appropriate length based on the
2738 cipher context. The EVP_CIPHER can provide its own random key generation
2739 routine to support keys of a specific form. This is used in the des and
2740 3des routines to generate a key of the correct parity. Update S/MIME
2741 code to use new functions and hence generate correct parity DES keys.
2742 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2743 valid (weak or incorrect parity).
2746 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2747 as looking them up. This is useful when the verified structure may contain
2748 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2749 present unless the new PKCS7_NO_CRL flag is asserted.
2752 *) Extend ASN1 oid configuration module. It now additionally accepts the
2755 shortName = some long name, 1.2.3.4
2758 *) Reimplemented the BN_CTX implementation. There is now no more static
2759 limitation on the number of variables it can handle nor the depth of the
2760 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2761 information can now expand as required, and rather than having a single
2762 static array of bignums, BN_CTX now uses a linked-list of such arrays
2763 allowing it to expand on demand whilst maintaining the usefulness of
2764 BN_CTX's "bundling".
2767 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2768 to allow all RSA operations to function using a single BN_CTX.
2771 *) Preliminary support for certificate policy evaluation and checking. This
2772 is initially intended to pass the tests outlined in "Conformance Testing
2773 of Relying Party Client Certificate Path Processing Logic" v1.07.
2776 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2777 remained unused and not that useful. A variety of other little bignum
2778 tweaks and fixes have also been made continuing on from the audit (see
2782 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2783 associated ASN1, EVP and SSL functions and old ASN1 macros.
2786 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2787 and this should never fail. So the return value from the use of
2788 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2789 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2792 *) BN_CTX_get() should return zero-valued bignums, providing the same
2793 initialised value as BN_new().
2794 [Geoff Thorpe, suggested by Ulf Möller]
2796 *) Support for inhibitAnyPolicy certificate extension.
2799 *) An audit of the BIGNUM code is underway, for which debugging code is
2800 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2801 is considered valid when processing BIGNUMs, and causes execution to
2802 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2803 further steps are taken to deliberately pollute unused data in BIGNUM
2804 structures to try and expose faulty code further on. For now, openssl will
2805 (in its default mode of operation) continue to tolerate the inconsistent
2806 forms that it has tolerated in the past, but authors and packagers should
2807 consider trying openssl and their own applications when compiled with
2808 these debugging symbols defined. It will help highlight potential bugs in
2809 their own code, and will improve the test coverage for OpenSSL itself. At
2810 some point, these tighter rules will become openssl's default to improve
2811 maintainability, though the assert()s and other overheads will remain only
2812 in debugging configurations. See bn.h for more details.
2813 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2815 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2816 that can only be obtained through BN_CTX_new() (which implicitly
2817 initialises it). The presence of this function only made it possible
2818 to overwrite an existing structure (and cause memory leaks).
2821 *) Because of the callback-based approach for implementing LHASH as a
2822 template type, lh_insert() adds opaque objects to hash-tables and
2823 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2824 to clean up those corresponding objects before destroying the hash table
2825 (and losing the object pointers). So some over-zealous constifications in
2826 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2827 objects as "const" and the lh_doall[_arg] callback wrappers are not
2828 prototyped to have "const" restrictions on the object pointers they are
2829 given (and so aren't required to cast them away any more).
2832 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2833 (speed) prefers to use its own implementation. The two implementations
2834 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2835 its object type properly exposed (MS_TM) instead of casting to/from "char
2836 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2837 aren't necessarily the greatest nomenclatures - but this is what was used
2838 internally to the implementation so I've used that for now.
2841 *) Ensure that deprecated functions do not get compiled when
2842 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2843 the self-tests were still using deprecated key-generation functions so
2844 these have been updated also.
2847 *) Reorganise PKCS#7 code to separate the digest location functionality
2848 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2849 New function PKCS7_set_digest() to set the digest type for PKCS#7
2850 digestedData type. Add additional code to correctly generate the
2851 digestedData type and add support for this type in PKCS7 initialization
2855 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2856 structure of type "other".
2859 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2860 sure the loop does correctly stop and breaking ("division by zero")
2861 modulus operations are not performed. The (pre-generated) prime
2862 table crypto/bn/bn_prime.h was already correct, but it could not be
2863 re-generated on some platforms because of the "division by zero"
2864 situation in the script.
2865 [Ralf S. Engelschall]
2867 *) Update support for ECC-based TLS ciphersuites according to
2868 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2869 SHA-1 now is only used for "small" curves (where the
2870 representation of a field element takes up to 24 bytes); for
2871 larger curves, the field element resulting from ECDH is directly
2872 used as premaster secret.
2873 [Douglas Stebila (Sun Microsystems Laboratories)]
2875 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2876 curve secp160r1 to the tests.
2877 [Douglas Stebila (Sun Microsystems Laboratories)]
2879 *) Add the possibility to load symbols globally with DSO.
2880 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2882 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2883 control of the error stack.
2886 *) Add support for STORE in ENGINE.
2889 *) Add the STORE type. The intention is to provide a common interface
2890 to certificate and key stores, be they simple file-based stores, or
2891 HSM-type store, or LDAP stores, or...
2892 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2895 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2896 pass a list of arguments to any function as well as provide a way
2897 for a function to pass data back to the caller.
2900 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2901 works like BUF_strdup() but can be used to duplicate a portion of
2902 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2906 *) Add the function sk_find_ex() which works like sk_find(), but will
2907 return an index to an element even if an exact match couldn't be
2908 found. The index is guaranteed to point at the element where the
2909 searched-for key would be inserted to preserve sorting order.
2912 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2913 takes an extra flags argument for optional functionality. Currently,
2914 the following flags are defined:
2916 OBJ_BSEARCH_VALUE_ON_NOMATCH
2917 This one gets OBJ_bsearch_ex() to return a pointer to the first
2918 element where the comparing function returns a negative or zero
2921 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2922 This one gets OBJ_bsearch_ex() to return a pointer to the first
2923 element where the comparing function returns zero. This is useful
2924 if there are more than one element where the comparing function
2928 *) Make it possible to create self-signed certificates with 'openssl ca'
2929 in such a way that the self-signed certificate becomes part of the
2930 CA database and uses the same mechanisms for serial number generation
2931 as all other certificate signing. The new flag '-selfsign' enables
2932 this functionality. Adapt CA.sh and CA.pl.in.
2935 *) Add functionality to check the public key of a certificate request
2936 against a given private. This is useful to check that a certificate
2937 request can be signed by that key (self-signing).
2940 *) Make it possible to have multiple active certificates with the same
2941 subject in the CA index file. This is done only if the keyword
2942 'unique_subject' is set to 'no' in the main CA section (default
2943 if 'CA_default') of the configuration file. The value is saved
2944 with the database itself in a separate index attribute file,
2945 named like the index file with '.attr' appended to the name.
2948 *) Generate muti valued AVAs using '+' notation in config files for
2952 *) Support for nameConstraints certificate extension.
2955 *) Support for policyConstraints certificate extension.
2958 *) Support for policyMappings certificate extension.
2961 *) Make sure the default DSA_METHOD implementation only uses its
2962 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2963 and change its own handlers to be NULL so as to remove unnecessary
2964 indirection. This lets alternative implementations fallback to the
2965 default implementation more easily.
2968 *) Support for directoryName in GeneralName related extensions
2972 *) Make it possible to link applications using Makefile.shared.
2973 Make that possible even when linking against static libraries!
2976 *) Support for single pass processing for S/MIME signing. This now
2977 means that S/MIME signing can be done from a pipe, in addition
2978 cleartext signing (multipart/signed type) is effectively streaming
2979 and the signed data does not need to be all held in memory.
2981 This is done with a new flag PKCS7_STREAM. When this flag is set
2982 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2983 is done after the data is output (and digests calculated) in
2984 SMIME_write_PKCS7().
2987 *) Add full support for -rpath/-R, both in shared libraries and
2988 applications, at least on the platforms where it's known how
2992 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2993 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2994 will now compute a table of multiples of the generator that
2995 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2996 faster (notably in the case of a single point multiplication,
2997 scalar * generator).
2998 [Nils Larsch, Bodo Moeller]
3000 *) IPv6 support for certificate extensions. The various extensions
3001 which use the IP:a.b.c.d can now take IPv6 addresses using the
3002 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3006 *) Added an ENGINE that implements RSA by performing private key
3007 exponentiations with the GMP library. The conversions to and from
3008 GMP's mpz_t format aren't optimised nor are any montgomery forms
3009 cached, and on x86 it appears OpenSSL's own performance has caught up.
3010 However there are likely to be other architectures where GMP could
3011 provide a boost. This ENGINE is not built in by default, but it can be
3012 specified at Configure time and should be accompanied by the necessary
3013 linker additions, eg;
3014 ./config -DOPENSSL_USE_GMP -lgmp
3017 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3018 testing availability of engines with "-t" - the old behaviour is
3019 produced by increasing the feature's verbosity with "-tt".
3022 *) ECDSA routines: under certain error conditions uninitialized BN objects
3023 could be freed. Solution: make sure initialization is performed early
3024 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3028 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3029 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3030 software implementations. For DSA and DH, parameter generation can
3031 also be overriden by providing the appropriate method callbacks.
3034 *) Change the "progress" mechanism used in key-generation and
3035 primality testing to functions that take a new BN_GENCB pointer in
3036 place of callback/argument pairs. The new API functions have "_ex"
3037 postfixes and the older functions are reimplemented as wrappers for
3038 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3039 declarations of the old functions to help (graceful) attempts to
3040 migrate to the new functions. Also, the new key-generation API
3041 functions operate on a caller-supplied key-structure and return
3042 success/failure rather than returning a key or NULL - this is to
3043 help make "keygen" another member function of RSA_METHOD etc.
3045 Example for using the new callback interface:
3047 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3051 BN_GENCB_set(&my_cb, my_callback, my_arg);
3053 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3054 /* For the meaning of a, b in calls to my_callback(), see the
3055 * documentation of the function that calls the callback.
3056 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3057 * my_callback should return 1 if it wants BN_is_prime_ex()
3058 * to continue, or 0 to stop.
3063 *) Change the ZLIB compression method to be stateful, and make it
3064 available to TLS with the number defined in
3065 draft-ietf-tls-compression-04.txt.
3068 *) Add the ASN.1 structures and functions for CertificatePair, which
3069 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3071 CertificatePair ::= SEQUENCE {
3072 forward [0] Certificate OPTIONAL,
3073 reverse [1] Certificate OPTIONAL,
3074 -- at least one of the pair shall be present -- }
3076 Also implement the PEM functions to read and write certificate
3077 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3079 This needed to be defined, mostly for the sake of the LDAP
3080 attribute crossCertificatePair, but may prove useful elsewhere as
3084 *) Make it possible to inhibit symlinking of shared libraries in
3085 Makefile.shared, for Cygwin's sake.
3088 *) Extend the BIGNUM API by creating a function
3089 void BN_set_negative(BIGNUM *a, int neg);
3090 and a macro that behave like
3091 int BN_is_negative(const BIGNUM *a);
3093 to avoid the need to access 'a->neg' directly in applications.
3096 *) Implement fast modular reduction for pseudo-Mersenne primes
3097 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3098 EC_GROUP_new_curve_GFp() will now automatically use this
3100 [Nils Larsch <nla@trustcenter.de>]
3102 *) Add new lock type (CRYPTO_LOCK_BN).
3105 *) Change the ENGINE framework to automatically load engines
3106 dynamically from specific directories unless they could be
3107 found to already be built in or loaded. Move all the
3108 current engines except for the cryptodev one to a new
3110 The engines in engines/ are built as shared libraries if
3111 the "shared" options was given to ./Configure or ./config.
3112 Otherwise, they are inserted in libcrypto.a.
3113 /usr/local/ssl/engines is the default directory for dynamic
3114 engines, but that can be overriden at configure time through
3115 the usual use of --prefix and/or --openssldir, and at run
3116 time with the environment variable OPENSSL_ENGINES.
3117 [Geoff Thorpe and Richard Levitte]
3119 *) Add Makefile.shared, a helper makefile to build shared
3120 libraries. Addapt Makefile.org.
3123 *) Add version info to Win32 DLLs.
3124 [Peter 'Luna' Runestig" <peter@runestig.com>]
3126 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3127 can be added using this API to created arbitrary PKCS#12
3128 files while avoiding the low level API.
3130 New options to PKCS12_create(), key or cert can be NULL and
3131 will then be omitted from the output file. The encryption
3132 algorithm NIDs can be set to -1 for no encryption, the mac
3133 iteration count can be set to 0 to omit the mac.
3135 Enhance pkcs12 utility by making the -nokeys and -nocerts
3136 options work when creating a PKCS#12 file. New option -nomac
3137 to omit the mac, NONE can be set for an encryption algorithm.
3138 New code is modified to use the enhanced PKCS12_create()
3139 instead of the low level API.
3142 *) Extend ASN1 encoder to support indefinite length constructed
3143 encoding. This can output sequences tags and octet strings in
3144 this form. Modify pk7_asn1.c to support indefinite length
3145 encoding. This is experimental and needs additional code to
3146 be useful, such as an ASN1 bio and some enhanced streaming
3149 Extend template encode functionality so that tagging is passed
3150 down to the template encoder.
3153 *) Let 'openssl req' fail if an argument to '-newkey' is not
3154 recognized instead of using RSA as a default.
3157 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3158 As these are not official, they are not included in "ALL";
3159 the "ECCdraft" ciphersuite group alias can be used to select them.
3160 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3162 *) Add ECDH engine support.
3163 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3165 *) Add ECDH in new directory crypto/ecdh/.
3166 [Douglas Stebila (Sun Microsystems Laboratories)]
3168 *) Let BN_rand_range() abort with an error after 100 iterations
3169 without success (which indicates a broken PRNG).
3172 *) Change BN_mod_sqrt() so that it verifies that the input value
3173 is really the square of the return value. (Previously,
3174 BN_mod_sqrt would show GIGO behaviour.)
3177 *) Add named elliptic curves over binary fields from X9.62, SECG,
3178 and WAP/WTLS; add OIDs that were still missing.
3180 [Sheueling Chang Shantz and Douglas Stebila
3181 (Sun Microsystems Laboratories)]
3183 *) Extend the EC library for elliptic curves over binary fields
3184 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3187 EC_GF2m_simple_method
3191 EC_GROUP_new_curve_GF2m
3192 EC_GROUP_set_curve_GF2m
3193 EC_GROUP_get_curve_GF2m
3194 EC_POINT_set_affine_coordinates_GF2m
3195 EC_POINT_get_affine_coordinates_GF2m
3196 EC_POINT_set_compressed_coordinates_GF2m
3198 Point compression for binary fields is disabled by default for
3199 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3202 As binary polynomials are represented as BIGNUMs, various members
3203 of the EC_GROUP and EC_POINT data structures can be shared
3204 between the implementations for prime fields and binary fields;
3205 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3206 are essentially identical to their ..._GFp counterparts.
3207 (For simplicity, the '..._GFp' prefix has been dropped from
3208 various internal method names.)
3210 An internal 'field_div' method (similar to 'field_mul' and
3211 'field_sqr') has been added; this is used only for binary fields.
3213 [Sheueling Chang Shantz and Douglas Stebila
3214 (Sun Microsystems Laboratories)]
3216 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3217 through methods ('mul', 'precompute_mult').
3219 The generic implementations (now internally called 'ec_wNAF_mul'
3220 and 'ec_wNAF_precomputed_mult') remain the default if these
3221 methods are undefined.
3223 [Sheueling Chang Shantz and Douglas Stebila
3224 (Sun Microsystems Laboratories)]
3226 *) New function EC_GROUP_get_degree, which is defined through
3227 EC_METHOD. For curves over prime fields, this returns the bit
3228 length of the modulus.
3230 [Sheueling Chang Shantz and Douglas Stebila
3231 (Sun Microsystems Laboratories)]
3233 *) New functions EC_GROUP_dup, EC_POINT_dup.
3234 (These simply call ..._new and ..._copy).
3236 [Sheueling Chang Shantz and Douglas Stebila
3237 (Sun Microsystems Laboratories)]
3239 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3240 Polynomials are represented as BIGNUMs (where the sign bit is not
3241 used) in the following functions [macros]:
3244 BN_GF2m_sub [= BN_GF2m_add]
3245 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3246 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3247 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3249 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3250 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3251 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3252 BN_GF2m_cmp [= BN_ucmp]
3254 (Note that only the 'mod' functions are actually for fields GF(2^m).
3255 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3257 For some functions, an the irreducible polynomial defining a
3258 field can be given as an 'unsigned int[]' with strictly
3259 decreasing elements giving the indices of those bits that are set;
3260 i.e., p[] represents the polynomial
3261 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3263 p[0] > p[1] > ... > p[k] = 0.
3264 This applies to the following functions:
3269 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3270 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3272 BN_GF2m_mod_sqrt_arr
3273 BN_GF2m_mod_solve_quad_arr
3277 Conversion can be performed by the following functions:
3282 bntest.c has additional tests for binary polynomial arithmetic.
3284 Two implementations for BN_GF2m_mod_div() are available.
3285 The default algorithm simply uses BN_GF2m_mod_inv() and
3286 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3287 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3288 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3290 [Sheueling Chang Shantz and Douglas Stebila
3291 (Sun Microsystems Laboratories)]
3293 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3294 functionality is disabled at compile-time.
3295 [Douglas Stebila <douglas.stebila@sun.com>]
3297 *) Change default behaviour of 'openssl asn1parse' so that more
3298 information is visible when viewing, e.g., a certificate:
3300 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3301 mode the content of non-printable OCTET STRINGs is output in a
3302 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3303 avoid the appearance of a printable string.
3304 [Nils Larsch <nla@trustcenter.de>]
3306 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3308 EC_GROUP_set_asn1_flag()
3309 EC_GROUP_get_asn1_flag()
3310 EC_GROUP_set_point_conversion_form()
3311 EC_GROUP_get_point_conversion_form()
3312 These control ASN1 encoding details:
3313 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3314 has been set to OPENSSL_EC_NAMED_CURVE.
3315 - Points are encoded in uncompressed form by default; options for
3316 asn1_for are as for point2oct, namely
3317 POINT_CONVERSION_COMPRESSED
3318 POINT_CONVERSION_UNCOMPRESSED
3319 POINT_CONVERSION_HYBRID
3321 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3324 EC_GROUP_get0_seed()
3325 EC_GROUP_get_seed_len()
3326 This is used only for ASN1 purposes (so far).
3327 [Nils Larsch <nla@trustcenter.de>]
3329 *) Add 'field_type' member to EC_METHOD, which holds the NID
3330 of the appropriate field type OID. The new function
3331 EC_METHOD_get_field_type() returns this value.
3332 [Nils Larsch <nla@trustcenter.de>]
3337 EC_POINT_point2hex()
3338 EC_POINT_hex2point()
3339 providing useful interfaces to EC_POINT_point2oct() and
3340 EC_POINT_oct2point().
3341 [Nils Larsch <nla@trustcenter.de>]
3343 *) Change internals of the EC library so that the functions
3344 EC_GROUP_set_generator()
3345 EC_GROUP_get_generator()
3346 EC_GROUP_get_order()
3347 EC_GROUP_get_cofactor()
3348 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3349 to methods, which would lead to unnecessary code duplication when
3350 adding different types of curves.
3351 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3353 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3354 arithmetic, and such that modified wNAFs are generated
3355 (which avoid length expansion in many cases).
3358 *) Add a function EC_GROUP_check_discriminant() (defined via
3359 EC_METHOD) that verifies that the curve discriminant is non-zero.
3361 Add a function EC_GROUP_check() that makes some sanity tests
3362 on a EC_GROUP, its generator and order. This includes
3363 EC_GROUP_check_discriminant().
3364 [Nils Larsch <nla@trustcenter.de>]
3366 *) Add ECDSA in new directory crypto/ecdsa/.
3368 Add applications 'openssl ecparam' and 'openssl ecdsa'
3369 (these are based on 'openssl dsaparam' and 'openssl dsa').
3371 ECDSA support is also included in various other files across the
3372 library. Most notably,
3373 - 'openssl req' now has a '-newkey ecdsa:file' option;
3374 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3375 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3376 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3377 them suitable for ECDSA where domain parameters must be
3378 extracted before the specific public key;
3379 - ECDSA engine support has been added.
3380 [Nils Larsch <nla@trustcenter.de>]
3382 *) Include some named elliptic curves, and add OIDs from X9.62,
3383 SECG, and WAP/WTLS. Each curve can be obtained from the new
3385 EC_GROUP_new_by_curve_name(),
3386 and the list of available named curves can be obtained with
3387 EC_get_builtin_curves().
3388 Also add a 'curve_name' member to EC_GROUP objects, which can be
3390 EC_GROUP_set_curve_name()
3391 EC_GROUP_get_curve_name()
3392 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3394 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3395 was actually never needed) and in BN_mul(). The removal in BN_mul()
3396 required a small change in bn_mul_part_recursive() and the addition
3397 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3398 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3399 bn_sub_words() and bn_add_words() except they take arrays with
3403 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3405 *) Cleanse PEM buffers before freeing them since they may contain
3407 [Benjamin Bennett <ben@psc.edu>]
3409 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3410 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3411 authentication-only ciphersuites.
3414 *) Since AES128 and AES256 share a single mask bit in the logic of
3415 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3416 kludge to work properly if AES128 is available and AES256 isn't.
3419 *) Expand security boundary to match 1.1.1 module.
3422 *) Remove redundant features: hash file source, editing of test vectors
3423 modify fipsld to use external fips_premain.c signature.
3426 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3427 run algorithm test programs.
3430 *) Make algorithm test programs more tolerant of whitespace.
3433 *) Have SSL/TLS server implementation tolerate "mismatched" record
3434 protocol version while receiving ClientHello even if the
3435 ClientHello is fragmented. (The server can't insist on the
3436 particular protocol version it has chosen before the ServerHello
3437 message has informed the client about his choice.)
3440 *) Load error codes if they are not already present instead of using a
3441 static variable. This allows them to be cleanly unloaded and reloaded.
3444 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3446 *) Introduce limits to prevent malicious keys being able to
3447 cause a denial of service. (CVE-2006-2940)
3448 [Steve Henson, Bodo Moeller]
3450 *) Fix ASN.1 parsing of certain invalid structures that can result
3451 in a denial of service. (CVE-2006-2937) [Steve Henson]
3453 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3454 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3456 *) Fix SSL client code which could crash if connecting to a
3457 malicious SSLv2 server. (CVE-2006-4343)
3458 [Tavis Ormandy and Will Drewry, Google Security Team]
3460 *) Change ciphersuite string processing so that an explicit
3461 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3462 will no longer include "AES128-SHA"), and any other similar
3463 ciphersuite (same bitmap) from *other* protocol versions (so that
3464 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3465 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3466 changes from 0.9.8b and 0.9.8d.
3469 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3471 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3472 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3474 *) Change the Unix randomness entropy gathering to use poll() when
3475 possible instead of select(), since the latter has some
3476 undesirable limitations.
3477 [Darryl Miles via Richard Levitte and Bodo Moeller]
3479 *) Disable rogue ciphersuites:
3481 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3482 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3483 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3485 The latter two were purportedly from
3486 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3489 Also deactive the remaining ciphersuites from
3490 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3491 unofficial, and the ID has long expired.
3494 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3495 dual-core machines) and other potential thread-safety issues.
3498 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3500 *) Adapt fipsld and the build system to link against the validated FIPS
3501 module in FIPS mode.
3504 *) Fixes for VC++ 2005 build under Windows.
3507 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3508 from a Windows bash shell such as MSYS. It is autodetected from the
3509 "config" script when run from a VC++ environment. Modify standard VC++
3510 build to use fipscanister.o from the GNU make build.
3513 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3515 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3516 The value now differs depending on if you build for FIPS or not.
3517 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3518 safely run with a non-FIPSed libcrypto, as it may crash because of
3519 the difference induced by this change.
3522 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3524 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3525 (part of SSL_OP_ALL). This option used to disable the
3526 countermeasure against man-in-the-middle protocol-version
3527 rollback in the SSL 2.0 server implementation, which is a bad
3528 idea. (CVE-2005-2969)
3530 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3531 for Information Security, National Institute of Advanced Industrial
3532 Science and Technology [AIST], Japan)]
3534 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3535 mainly for FIPS compliance and not fully integrated at this stage.
3538 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3539 the exponentiation using a fixed-length exponent. (Otherwise,
3540 the information leaked through timing could expose the secret key
3541 after many signatures; cf. Bleichenbacher's attack on DSA with
3545 *) Make a new fixed-window mod_exp implementation the default for
3546 RSA, DSA, and DH private-key operations so that the sequence of
3547 squares and multiplies and the memory access pattern are
3548 independent of the particular secret key. This will mitigate
3549 cache-timing and potential related attacks.
3551 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3552 and this is automatically used by BN_mod_exp_mont() if the new flag
3553 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3554 will use this BN flag for private exponents unless the flag
3555 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3556 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3558 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3560 *) Change the client implementation for SSLv23_method() and
3561 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3562 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3563 (Previously, the SSL 2.0 backwards compatible Client Hello
3564 message format would be used even with SSL_OP_NO_SSLv2.)
3567 *) Add support for smime-type MIME parameter in S/MIME messages which some
3571 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3572 a threadsafe manner. Modify rsa code to use new function and add calls
3573 to dsa and dh code (which had race conditions before).
3576 *) Include the fixed error library code in the C error file definitions
3577 instead of fixing them up at runtime. This keeps the error code
3578 structures constant.
3581 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3583 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3586 *) Fixes for newer kerberos headers. NB: the casts are needed because
3587 the 'length' field is signed on one version and unsigned on another
3588 with no (?) obvious way to tell the difference, without these VC++
3589 complains. Also the "definition" of FAR (blank) is no longer included
3590 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3591 some needed definitions.
3594 *) Undo Cygwin change.
3597 *) Added support for proxy certificates according to RFC 3820.
3598 Because they may be a security thread to unaware applications,
3599 they must be explicitely allowed in run-time. See
3600 docs/HOWTO/proxy_certificates.txt for further information.
3603 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3605 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3606 server and client random values. Previously
3607 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3608 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3610 This change has negligible security impact because:
3612 1. Server and client random values still have 24 bytes of pseudo random
3615 2. Server and client random values are sent in the clear in the initial
3618 3. The master secret is derived using the premaster secret (48 bytes in
3619 size for static RSA ciphersuites) as well as client server and random
3622 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3625 [Stephen Henson, reported by UK NISCC]
3627 *) Use Windows randomness collection on Cygwin.
3630 *) Fix hang in EGD/PRNGD query when communication socket is closed
3631 prematurely by EGD/PRNGD.
3632 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3634 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3637 *) Back-port of selected performance improvements from development
3638 branch, as well as improved support for PowerPC platforms.