5 Changes between 1.0.1e and 1.0.1f [xx XXX xxxx]
9 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
13 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
15 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
17 This addresses the flaw in CBC record processing discovered by
18 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
19 at: http://www.isg.rhul.ac.uk/tls/
21 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
22 Security Group at Royal Holloway, University of London
23 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
24 Emilia Käsper for the initial patch.
26 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
28 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
29 ciphersuites which can be exploited in a denial of service attack.
30 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
31 and detecting this bug and to Wolfgang Ettlinger
32 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
36 *) Return an error when checking OCSP signatures when key is NULL.
37 This fixes a DoS attack. (CVE-2013-0166)
40 *) Make openssl verify return errors.
41 [Chris Palmer <palmer@google.com> and Ben Laurie]
43 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
44 the right response is stapled. Also change SSL_get_certificate()
45 so it returns the certificate actually sent.
46 See http://rt.openssl.org/Ticket/Display.html?id=2836.
47 [Rob Stradling <rob.stradling@comodo.com>]
49 *) Fix possible deadlock when decoding public keys.
52 *) Don't use TLS 1.0 record version number in initial client hello
56 Changes between 1.0.1b and 1.0.1c [10 May 2012]
58 *) Sanity check record length before skipping explicit IV in TLS
59 1.2, 1.1 and DTLS to fix DoS attack.
61 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
62 fuzzing as a service testing platform.
66 *) Initialise tkeylen properly when encrypting CMS messages.
67 Thanks to Solar Designer of Openwall for reporting this issue.
70 *) In FIPS mode don't try to use composite ciphers as they are not
74 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
76 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
77 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
78 mean any application compiled against OpenSSL 1.0.0 headers setting
79 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
80 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
81 0x10000000L Any application which was previously compiled against
82 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
83 will need to be recompiled as a result. Letting be results in
84 inability to disable specifically TLS 1.1 and in client context,
85 in unlike event, limit maximum offered version to TLS 1.0 [see below].
88 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
89 disable just protocol X, but all protocols above X *if* there are
90 protocols *below* X still enabled. In more practical terms it means
91 that if application wants to disable TLS1.0 in favor of TLS1.1 and
92 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
93 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
97 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
99 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
100 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
101 in CRYPTO_realloc_clean.
103 Thanks to Tavis Ormandy, Google Security Team, for discovering this
104 issue and to Adam Langley <agl@chromium.org> for fixing it.
106 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
108 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
111 *) Workarounds for some broken servers that "hang" if a client hello
112 record length exceeds 255 bytes.
114 1. Do not use record version number > TLS 1.0 in initial client
115 hello: some (but not all) hanging servers will now work.
116 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
117 the number of ciphers sent in the client hello. This should be
118 set to an even number, such as 50, for example by passing:
119 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
120 Most broken servers should now work.
121 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
122 TLS 1.2 client support entirely.
125 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
128 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
130 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
131 STRING form instead of a DigestInfo.
134 *) The format used for MDC2 RSA signatures is inconsistent between EVP
135 and the RSA_sign/RSA_verify functions. This was made more apparent when
136 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
137 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
138 the correct format in RSA_verify so both forms transparently work.
141 *) Some servers which support TLS 1.0 can choke if we initially indicate
142 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
143 encrypted premaster secret. As a workaround use the maximum pemitted
144 client version in client hello, this should keep such servers happy
145 and still work with previous versions of OpenSSL.
148 *) Add support for TLS/DTLS heartbeats.
149 [Robin Seggelmann <seggelmann@fh-muenster.de>]
151 *) Add support for SCTP.
152 [Robin Seggelmann <seggelmann@fh-muenster.de>]
154 *) Improved PRNG seeding for VOS.
155 [Paul Green <Paul.Green@stratus.com>]
157 *) Extensive assembler packs updates, most notably:
159 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
160 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
161 - x86_64: bit-sliced AES implementation;
162 - ARM: NEON support, contemporary platforms optimizations;
163 - s390x: z196 support;
164 - *: GHASH and GF(2^m) multiplication implementations;
168 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
169 (removal of unnecessary code)
170 [Peter Sylvester <peter.sylvester@edelweb.fr>]
172 *) Add TLS key material exporter from RFC 5705.
175 *) Add DTLS-SRTP negotiation from RFC 5764.
178 *) Add Next Protocol Negotiation,
179 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
180 disabled with a no-npn flag to config or Configure. Code donated
182 [Adam Langley <agl@google.com> and Ben Laurie]
184 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
185 NIST-P256, NIST-P521, with constant-time single point multiplication on
186 typical inputs. Compiler support for the nonstandard type __uint128_t is
187 required to use this (present in gcc 4.4 and later, for 64-bit builds).
188 Code made available under Apache License version 2.0.
190 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
191 line to include this in your build of OpenSSL, and run "make depend" (or
192 "make update"). This enables the following EC_METHODs:
194 EC_GFp_nistp224_method()
195 EC_GFp_nistp256_method()
196 EC_GFp_nistp521_method()
198 EC_GROUP_new_by_curve_name() will automatically use these (while
199 EC_GROUP_new_curve_GFp() currently prefers the more flexible
201 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
203 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
204 all platforms. Move ssize_t definition from e_os.h to the public
205 header file e_os2.h as it now appears in public header file cms.h
208 *) New -sigopt option to the ca, req and x509 utilities. Additional
209 signature parameters can be passed using this option and in
213 *) Add RSA PSS signing function. This will generate and set the
214 appropriate AlgorithmIdentifiers for PSS based on those in the
215 corresponding EVP_MD_CTX structure. No application support yet.
218 *) Support for companion algorithm specific ASN1 signing routines.
219 New function ASN1_item_sign_ctx() signs a pre-initialised
220 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
221 the appropriate parameters.
224 *) Add new algorithm specific ASN1 verification initialisation function
225 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
226 handling will be the same no matter what EVP_PKEY_METHOD is used.
227 Add a PSS handler to support verification of PSS signatures: checked
228 against a number of sample certificates.
231 *) Add signature printing for PSS. Add PSS OIDs.
232 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
234 *) Add algorithm specific signature printing. An individual ASN1 method
235 can now print out signatures instead of the standard hex dump.
237 More complex signatures (e.g. PSS) can print out more meaningful
238 information. Include DSA version that prints out the signature
242 *) Password based recipient info support for CMS library: implementing
246 *) Split password based encryption into PBES2 and PBKDF2 functions. This
247 neatly separates the code into cipher and PBE sections and is required
248 for some algorithms that split PBES2 into separate pieces (such as
252 *) Session-handling fixes:
253 - Fix handling of connections that are resuming with a session ID,
254 but also support Session Tickets.
255 - Fix a bug that suppressed issuing of a new ticket if the client
256 presented a ticket with an expired session.
257 - Try to set the ticket lifetime hint to something reasonable.
258 - Make tickets shorter by excluding irrelevant information.
259 - On the client side, don't ignore renewed tickets.
260 [Adam Langley, Bodo Moeller (Google)]
262 *) Fix PSK session representation.
265 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
267 This work was sponsored by Intel.
270 *) Add GCM support to TLS library. Some custom code is needed to split
271 the IV between the fixed (from PRF) and explicit (from TLS record)
272 portions. This adds all GCM ciphersuites supported by RFC5288 and
273 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
274 add a special AESGCM string for GCM only.
277 *) Expand range of ctrls for AES GCM. Permit setting invocation
278 field on decrypt and retrieval of invocation field only on encrypt.
281 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
282 As required by RFC5289 these ciphersuites cannot be used if for
283 versions of TLS earlier than 1.2.
286 *) For FIPS capable OpenSSL interpret a NULL default public key method
287 as unset and return the appopriate default but do *not* set the default.
288 This means we can return the appopriate method in applications that
289 swicth between FIPS and non-FIPS modes.
292 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
293 ENGINE is used then we cannot handle that in the FIPS module so we
294 keep original code iff non-FIPS operations are allowed.
297 *) Add -attime option to openssl utilities.
298 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
300 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
303 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
304 FIPS EC methods unconditionally for now.
307 *) New build option no-ec2m to disable characteristic 2 code.
310 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
311 all cases can be covered as some introduce binary incompatibilities.
314 *) Redirect RSA operations to FIPS module including keygen,
315 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
318 *) Add similar low level API blocking to ciphers.
321 *) Low level digest APIs are not approved in FIPS mode: any attempt
322 to use these will cause a fatal error. Applications that *really* want
323 to use them can use the private_* version instead.
326 *) Redirect cipher operations to FIPS module for FIPS builds.
329 *) Redirect digest operations to FIPS module for FIPS builds.
332 *) Update build system to add "fips" flag which will link in fipscanister.o
333 for static and shared library builds embedding a signature if needed.
336 *) Output TLS supported curves in preference order instead of numerical
337 order. This is currently hardcoded for the highest order curves first.
338 This should be configurable so applications can judge speed vs strength.
341 *) Add TLS v1.2 server support for client authentication.
344 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
348 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
349 FIPS modules versions.
352 *) Add TLS v1.2 client side support for client authentication. Keep cache
353 of handshake records longer as we don't know the hash algorithm to use
354 until after the certificate request message is received.
357 *) Initial TLS v1.2 client support. Add a default signature algorithms
358 extension including all the algorithms we support. Parse new signature
359 format in client key exchange. Relax some ECC signing restrictions for
360 TLS v1.2 as indicated in RFC5246.
363 *) Add server support for TLS v1.2 signature algorithms extension. Switch
364 to new signature format when needed using client digest preference.
365 All server ciphersuites should now work correctly in TLS v1.2. No client
366 support yet and no support for client certificates.
369 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
370 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
371 ciphersuites. At present only RSA key exchange ciphersuites work with
372 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
373 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
374 and version checking.
377 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
378 with this defined it will not be affected by any changes to ssl internal
379 structures. Add several utility functions to allow openssl application
380 to work with OPENSSL_NO_SSL_INTERN defined.
384 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
386 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
389 *) Permit abbreviated handshakes when renegotiating using the function
390 SSL_renegotiate_abbreviated().
391 [Robin Seggelmann <seggelmann@fh-muenster.de>]
393 *) Add call to ENGINE_register_all_complete() to
394 ENGINE_load_builtin_engines(), so some implementations get used
395 automatically instead of needing explicit application support.
398 *) Add support for TLS key exporter as described in RFC5705.
399 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
401 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
402 a few changes are required:
404 Add SSL_OP_NO_TLSv1_1 flag.
406 Update version checking logic to handle version 1.1.
407 Add explicit IV handling (ported from DTLS code).
408 Add command line options to s_client/s_server.
411 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
413 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
414 in CMS and PKCS7 code. When RSA decryption fails use a random key for
415 content decryption and always return the same error. Note: this attack
416 needs on average 2^20 messages so it only affects automated senders. The
417 old behaviour can be reenabled in the CMS code by setting the
418 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
419 an MMA defence is not necessary.
420 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
421 this issue. (CVE-2012-0884)
424 *) Fix CVE-2011-4619: make sure we really are receiving a
425 client hello before rejecting multiple SGC restarts. Thanks to
426 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
429 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
431 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
432 Thanks to Antonio Martin, Enterprise Secure Access Research and
433 Development, Cisco Systems, Inc. for discovering this bug and
434 preparing a fix. (CVE-2012-0050)
437 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
439 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
440 of the Vaudenay padding oracle attack on CBC mode encryption
441 which enables an efficient plaintext recovery attack against
442 the OpenSSL implementation of DTLS. Their attack exploits timing
443 differences arising during decryption processing. A research
444 paper describing this attack can be found at:
445 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
446 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
447 Security Group at Royal Holloway, University of London
448 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
449 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
450 for preparing the fix. (CVE-2011-4108)
451 [Robin Seggelmann, Michael Tuexen]
453 *) Clear bytes used for block padding of SSL 3.0 records.
455 [Adam Langley (Google)]
457 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
458 Kadianakis <desnacked@gmail.com> for discovering this issue and
459 Adam Langley for preparing the fix. (CVE-2011-4619)
460 [Adam Langley (Google)]
462 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
463 [Andrey Kulikov <amdeich@gmail.com>]
465 *) Prevent malformed RFC3779 data triggering an assertion failure.
466 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
467 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
468 [Rob Austein <sra@hactrn.net>]
470 *) Improved PRNG seeding for VOS.
471 [Paul Green <Paul.Green@stratus.com>]
473 *) Fix ssl_ciph.c set-up race.
474 [Adam Langley (Google)]
476 *) Fix spurious failures in ecdsatest.c.
477 [Emilia Käsper (Google)]
479 *) Fix the BIO_f_buffer() implementation (which was mixing different
480 interpretations of the '..._len' fields).
481 [Adam Langley (Google)]
483 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
484 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
485 threads won't reuse the same blinding coefficients.
487 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
488 lock to call BN_BLINDING_invert_ex, and avoids one use of
489 BN_BLINDING_update for each BN_BLINDING structure (previously,
490 the last update always remained unused).
491 [Emilia Käsper (Google)]
493 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
494 [Bob Buckholz (Google)]
496 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
498 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
499 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
500 [Kaspar Brand <ossl@velox.ch>]
502 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
503 for multi-threaded use of ECDH. (CVE-2011-3210)
504 [Adam Langley (Google)]
506 *) Fix x509_name_ex_d2i memory leak on bad inputs.
509 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
510 signature public key algorithm by using OID xref utilities instead.
511 Before this you could only use some ECC ciphersuites with SHA1 only.
514 *) Add protection against ECDSA timing attacks as mentioned in the paper
515 by Billy Bob Brumley and Nicola Tuveri, see:
517 http://eprint.iacr.org/2011/232.pdf
519 [Billy Bob Brumley and Nicola Tuveri]
521 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
523 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
524 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
526 *) Fix bug in string printing code: if *any* escaping is enabled we must
527 escape the escape character (backslash) or the resulting string is
531 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
533 *) Disable code workaround for ancient and obsolete Netscape browsers
534 and servers: an attacker can use it in a ciphersuite downgrade attack.
535 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
538 *) Fixed J-PAKE implementation error, originally discovered by
539 Sebastien Martini, further info and confirmation from Stefan
540 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
543 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
545 *) Fix extension code to avoid race conditions which can result in a buffer
546 overrun vulnerability: resumed sessions must not be modified as they can
547 be shared by multiple threads. CVE-2010-3864
550 *) Fix WIN32 build system to correctly link an ENGINE directory into
554 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
556 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
558 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
560 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
562 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
563 context. The operation can be customised via the ctrl mechanism in
564 case ENGINEs want to include additional functionality.
567 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
570 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
571 output hashes compatible with older versions of OpenSSL.
572 [Willy Weisz <weisz@vcpc.univie.ac.at>]
574 *) Fix compression algorithm handling: if resuming a session use the
575 compression algorithm of the resumed session instead of determining
576 it from client hello again. Don't allow server to change algorithm.
579 *) Add load_crls() function to apps tidying load_certs() too. Add option
580 to verify utility to allow additional CRLs to be included.
583 *) Update OCSP request code to permit adding custom headers to the request:
584 some responders need this.
587 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
589 [Julia Lawall <julia@diku.dk>]
591 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
592 needlessly dereferenced structures, used obsolete functions and
593 didn't handle all updated verify codes correctly.
596 *) Disable MD2 in the default configuration.
599 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
600 indicate the initial BIO being pushed or popped. This makes it possible
601 to determine whether the BIO is the one explicitly called or as a result
602 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
603 it handles reference counts correctly and doesn't zero out the I/O bio
604 when it is not being explicitly popped. WARNING: applications which
605 included workarounds for the old buggy behaviour will need to be modified
606 or they could free up already freed BIOs.
609 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
610 renaming to all platforms (within the 0.9.8 branch, this was
611 done conditionally on Netware platforms to avoid a name clash).
612 [Guenter <lists@gknw.net>]
614 *) Add ECDHE and PSK support to DTLS.
615 [Michael Tuexen <tuexen@fh-muenster.de>]
617 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
621 *) Add "missing" function EVP_MD_flags() (without this the only way to
622 retrieve a digest flags is by accessing the structure directly. Update
623 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
624 or cipher is registered as in the "from" argument. Print out all
625 registered digests in the dgst usage message instead of manually
626 attempting to work them out.
629 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
630 this allows the use of compression and extensions. Change default cipher
631 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
632 by default unless an application cipher string requests it.
635 *) Alter match criteria in PKCS12_parse(). It used to try to use local
636 key ids to find matching certificates and keys but some PKCS#12 files
637 don't follow the (somewhat unwritten) rules and this strategy fails.
638 Now just gather all certificates together and the first private key
639 then look for the first certificate that matches the key.
642 *) Support use of registered digest and cipher names for dgst and cipher
643 commands instead of having to add each one as a special case. So now
650 openssl dgst -sha256 foo
652 and this works for ENGINE based algorithms too.
656 *) Update Gost ENGINE to support parameter files.
657 [Victor B. Wagner <vitus@cryptocom.ru>]
659 *) Support GeneralizedTime in ca utility.
660 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
662 *) Enhance the hash format used for certificate directory links. The new
663 form uses the canonical encoding (meaning equivalent names will work
664 even if they aren't identical) and uses SHA1 instead of MD5. This form
665 is incompatible with the older format and as a result c_rehash should
666 be used to rebuild symbolic links.
669 *) Make PKCS#8 the default write format for private keys, replacing the
670 traditional format. This form is standardised, more secure and doesn't
671 include an implicit MD5 dependency.
674 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
675 committed to OpenSSL should pass this lot as a minimum.
678 *) Add session ticket override functionality for use by EAP-FAST.
679 [Jouni Malinen <j@w1.fi>]
681 *) Modify HMAC functions to return a value. Since these can be implemented
682 in an ENGINE errors can occur.
685 *) Type-checked OBJ_bsearch_ex.
688 *) Type-checked OBJ_bsearch. Also some constification necessitated
689 by type-checking. Still to come: TXT_DB, bsearch(?),
690 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
694 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
695 seconds to a tm structure directly, instead of going through OS
696 specific date routines. This avoids any issues with OS routines such
697 as the year 2038 bug. New *_adj() functions for ASN1 time structures
698 and X509_time_adj_ex() to cover the extended range. The existing
699 X509_time_adj() is still usable and will no longer have any date issues.
702 *) Delta CRL support. New use deltas option which will attempt to locate
703 and search any appropriate delta CRLs available.
705 This work was sponsored by Google.
708 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
709 code and add additional score elements. Validate alternate CRL paths
710 as part of the CRL checking and indicate a new error "CRL path validation
711 error" in this case. Applications wanting additional details can use
712 the verify callback and check the new "parent" field. If this is not
713 NULL CRL path validation is taking place. Existing applications wont
714 see this because it requires extended CRL support which is off by
717 This work was sponsored by Google.
720 *) Support for freshest CRL extension.
722 This work was sponsored by Google.
725 *) Initial indirect CRL support. Currently only supported in the CRLs
726 passed directly and not via lookup. Process certificate issuer
727 CRL entry extension and lookup CRL entries by bother issuer name
728 and serial number. Check and process CRL issuer entry in IDP extension.
730 This work was sponsored by Google.
733 *) Add support for distinct certificate and CRL paths. The CRL issuer
734 certificate is validated separately in this case. Only enabled if
735 an extended CRL support flag is set: this flag will enable additional
736 CRL functionality in future.
738 This work was sponsored by Google.
741 *) Add support for policy mappings extension.
743 This work was sponsored by Google.
746 *) Fixes to pathlength constraint, self issued certificate handling,
747 policy processing to align with RFC3280 and PKITS tests.
749 This work was sponsored by Google.
752 *) Support for name constraints certificate extension. DN, email, DNS
753 and URI types are currently supported.
755 This work was sponsored by Google.
758 *) To cater for systems that provide a pointer-based thread ID rather
759 than numeric, deprecate the current numeric thread ID mechanism and
760 replace it with a structure and associated callback type. This
761 mechanism allows a numeric "hash" to be extracted from a thread ID in
762 either case, and on platforms where pointers are larger than 'long',
763 mixing is done to help ensure the numeric 'hash' is usable even if it
764 can't be guaranteed unique. The default mechanism is to use "&errno"
765 as a pointer-based thread ID to distinguish between threads.
767 Applications that want to provide their own thread IDs should now use
768 CRYPTO_THREADID_set_callback() to register a callback that will call
769 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
771 Note that ERR_remove_state() is now deprecated, because it is tied
772 to the assumption that thread IDs are numeric. ERR_remove_state(0)
773 to free the current thread's error state should be replaced by
774 ERR_remove_thread_state(NULL).
776 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
777 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
778 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
779 application was previously providing a numeric thread callback that
780 was inappropriate for distinguishing threads, then uniqueness might
781 have been obtained with &errno that happened immediately in the
782 intermediate development versions of OpenSSL; this is no longer the
783 case, the numeric thread callback will now override the automatic use
785 [Geoff Thorpe, with help from Bodo Moeller]
787 *) Initial support for different CRL issuing certificates. This covers a
788 simple case where the self issued certificates in the chain exist and
789 the real CRL issuer is higher in the existing chain.
791 This work was sponsored by Google.
794 *) Removed effectively defunct crypto/store from the build.
797 *) Revamp of STACK to provide stronger type-checking. Still to come:
798 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
799 ASN1_STRING, CONF_VALUE.
802 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
803 RAM on SSL connections. This option can save about 34k per idle SSL.
806 *) Revamp of LHASH to provide stronger type-checking. Still to come:
807 STACK, TXT_DB, bsearch, qsort.
810 *) Initial support for Cryptographic Message Syntax (aka CMS) based
811 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
812 support for data, signedData, compressedData, digestedData and
813 encryptedData, envelopedData types included. Scripts to check against
814 RFC4134 examples draft and interop and consistency checks of many
815 content types and variants.
818 *) Add options to enc utility to support use of zlib compression BIO.
821 *) Extend mk1mf to support importing of options and assembly language
822 files from Configure script, currently only included in VC-WIN32.
823 The assembly language rules can now optionally generate the source
824 files from the associated perl scripts.
827 *) Implement remaining functionality needed to support GOST ciphersuites.
828 Interop testing has been performed using CryptoPro implementations.
829 [Victor B. Wagner <vitus@cryptocom.ru>]
831 *) s390x assembler pack.
834 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
838 *) Implement Opaque PRF Input TLS extension as specified in
839 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
840 official specification yet and no extension type assignment by
841 IANA exists, this extension (for now) will have to be explicitly
842 enabled when building OpenSSL by providing the extension number
843 to use. For example, specify an option
845 -DTLSEXT_TYPE_opaque_prf_input=0x9527
847 to the "config" or "Configure" script to enable the extension,
848 assuming extension number 0x9527 (which is a completely arbitrary
849 and unofficial assignment based on the MD5 hash of the Internet
850 Draft). Note that by doing so, you potentially lose
851 interoperability with other TLS implementations since these might
852 be using the same extension number for other purposes.
854 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
855 opaque PRF input value to use in the handshake. This will create
856 an interal copy of the length-'len' string at 'src', and will
857 return non-zero for success.
859 To get more control and flexibility, provide a callback function
862 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
863 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
867 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
870 Callback function 'cb' will be called in handshakes, and is
871 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
872 Argument 'arg' is for application purposes (the value as given to
873 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
874 be provided to the callback function). The callback function
875 has to return non-zero to report success: usually 1 to use opaque
876 PRF input just if possible, or 2 to enforce use of the opaque PRF
877 input. In the latter case, the library will abort the handshake
878 if opaque PRF input is not successfully negotiated.
880 Arguments 'peerinput' and 'len' given to the callback function
881 will always be NULL and 0 in the case of a client. A server will
882 see the client's opaque PRF input through these variables if
883 available (NULL and 0 otherwise). Note that if the server
884 provides an opaque PRF input, the length must be the same as the
885 length of the client's opaque PRF input.
887 Note that the callback function will only be called when creating
888 a new session (session resumption can resume whatever was
889 previously negotiated), and will not be called in SSL 2.0
890 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
891 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
892 for applications that need to enforce opaque PRF input.
896 *) Update ssl code to support digests other than SHA1+MD5 for handshake
899 [Victor B. Wagner <vitus@cryptocom.ru>]
901 *) Add RFC4507 support to OpenSSL. This includes the corrections in
902 RFC4507bis. The encrypted ticket format is an encrypted encoded
903 SSL_SESSION structure, that way new session features are automatically
906 If a client application caches session in an SSL_SESSION structure
907 support is transparent because tickets are now stored in the encoded
910 The SSL_CTX structure automatically generates keys for ticket
911 protection in servers so again support should be possible
912 with no application modification.
914 If a client or server wishes to disable RFC4507 support then the option
915 SSL_OP_NO_TICKET can be set.
917 Add a TLS extension debugging callback to allow the contents of any client
918 or server extensions to be examined.
920 This work was sponsored by Google.
923 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
924 OpenSSL should now compile cleanly on gcc 4.2
925 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
927 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
928 support including streaming MAC support: this is required for GOST
930 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
932 *) Add option -stream to use PKCS#7 streaming in smime utility. New
933 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
934 to output in BER and PEM format.
937 *) Experimental support for use of HMAC via EVP_PKEY interface. This
938 allows HMAC to be handled via the EVP_DigestSign*() interface. The
939 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
940 ENGINE support for HMAC keys which are unextractable. New -mac and
941 -macopt options to dgst utility.
944 *) New option -sigopt to dgst utility. Update dgst to use
945 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
946 alternative signing paramaters such as X9.31 or PSS in the dgst
950 *) Change ssl_cipher_apply_rule(), the internal function that does
951 the work each time a ciphersuite string requests enabling
952 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
953 removing ("!foo+bar") a class of ciphersuites: Now it maintains
954 the order of disabled ciphersuites such that those ciphersuites
955 that most recently went from enabled to disabled not only stay
956 in order with respect to each other, but also have higher priority
957 than other disabled ciphersuites the next time ciphersuites are
960 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
961 the same ciphersuites as with "HIGH" alone, but in a specific
962 order where the PSK ciphersuites come first (since they are the
963 most recently disabled ciphersuites when "HIGH" is parsed).
965 Also, change ssl_create_cipher_list() (using this new
966 funcionality) such that between otherwise identical
967 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
971 *) Change ssl_create_cipher_list() so that it automatically
972 arranges the ciphersuites in reasonable order before starting
973 to process the rule string. Thus, the definition for "DEFAULT"
974 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
975 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
976 This makes it much easier to arrive at a reasonable default order
977 in applications for which anonymous ciphers are OK (meaning
978 that you can't actually use DEFAULT).
979 [Bodo Moeller; suggested by Victor Duchovni]
981 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
982 processing) into multiple integers instead of setting
983 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
984 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
985 (These masks as well as the individual bit definitions are hidden
986 away into the non-exported interface ssl/ssl_locl.h, so this
987 change to the definition of the SSL_CIPHER structure shouldn't
988 affect applications.) This give us more bits for each of these
989 categories, so there is no longer a need to coagulate AES128 and
990 AES256 into a single algorithm bit, and to coagulate Camellia128
991 and Camellia256 into a single algorithm bit, which has led to all
994 Thus, among other things, the kludge introduced in 0.9.7m and
995 0.9.8e for masking out AES256 independently of AES128 or masking
996 out Camellia256 independently of AES256 is not needed here in 0.9.9.
998 With the change, we also introduce new ciphersuite aliases that
999 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1003 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1004 Use the leftmost N bytes of the signature input if the input is
1005 larger than the prime q (with N being the size in bytes of q).
1008 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1009 it yet and it is largely untested.
1012 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1015 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1016 some compilers (gcc 4.2 and later) reject their use. Safestack is
1017 reimplemented. Update ASN1 to avoid use of legacy functions.
1020 *) Win32/64 targets are linked with Winsock2.
1023 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1024 to external functions. This can be used to increase CRL handling
1025 efficiency especially when CRLs are very large by (for example) storing
1026 the CRL revoked certificates in a database.
1029 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1030 new CRLs added to a directory can be used. New command line option
1031 -verify_return_error to s_client and s_server. This causes real errors
1032 to be returned by the verify callback instead of carrying on no matter
1033 what. This reflects the way a "real world" verify callback would behave.
1036 *) GOST engine, supporting several GOST algorithms and public key formats.
1037 Kindly donated by Cryptocom.
1040 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1041 partitioned by DP are handled but no indirect CRL or reason partitioning
1042 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1043 selected via a scoring technique which handles IDP and AKID in CRLs.
1046 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1047 will ultimately be used for all verify operations: this will remove the
1048 X509_STORE dependency on certificate verification and allow alternative
1049 lookup methods. X509_STORE based implementations of these two callbacks.
1052 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1053 Modify get_crl() to find a valid (unexpired) CRL if possible.
1056 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1057 this would be called X509_CRL_cmp() but that name is already used by
1058 a function that just compares CRL issuer names. Cache several CRL
1059 extensions in X509_CRL structure and cache CRLDP in X509.
1062 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1063 this maps equivalent X509_NAME structures into a consistent structure.
1064 Name comparison can then be performed rapidly using memcmp().
1067 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1071 *) Allow digests to supply their own micalg string for S/MIME type using
1072 the ctrl EVP_MD_CTRL_MICALG.
1075 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1076 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1077 ctrl. It can then customise the structure before and/or after signing
1081 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1082 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1083 to free up any added signature OIDs.
1086 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1087 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1088 digest and cipher tables. New options added to openssl utility:
1089 list-message-digest-algorithms and list-cipher-algorithms.
1092 *) Change the array representation of binary polynomials: the list
1093 of degrees of non-zero coefficients is now terminated with -1.
1094 Previously it was terminated with 0, which was also part of the
1095 value; thus, the array representation was not applicable to
1096 polynomials where t^0 has coefficient zero. This change makes
1097 the array representation useful in a more general context.
1100 *) Various modifications and fixes to SSL/TLS cipher string
1101 handling. For ECC, the code now distinguishes between fixed ECDH
1102 with RSA certificates on the one hand and with ECDSA certificates
1103 on the other hand, since these are separate ciphersuites. The
1104 unused code for Fortezza ciphersuites has been removed.
1106 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1107 (not "ECDHE"). For consistency with the code for DH
1108 certificates, use of ECDH certificates is now considered ECDH
1109 authentication, not RSA or ECDSA authentication (the latter is
1110 merely the CA's signing algorithm and not actively used in the
1113 The temporary ciphersuite alias "ECCdraft" is no longer
1114 available, and ECC ciphersuites are no longer excluded from "ALL"
1115 and "DEFAULT". The following aliases now exist for RFC 4492
1116 ciphersuites, most of these by analogy with the DH case:
1118 kECDHr - ECDH cert, signed with RSA
1119 kECDHe - ECDH cert, signed with ECDSA
1120 kECDH - ECDH cert (signed with either RSA or ECDSA)
1121 kEECDH - ephemeral ECDH
1122 ECDH - ECDH cert or ephemeral ECDH
1128 AECDH - anonymous ECDH
1129 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1133 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1134 Use correct micalg parameters depending on digest(s) in signed message.
1137 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1138 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1141 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1142 an engine to register a method. Add ENGINE lookups for methods and
1143 functional reference processing.
1146 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1147 EVP_{Sign,Verify}* which allow an application to customise the signature
1151 *) New -resign option to smime utility. This adds one or more signers
1152 to an existing PKCS#7 signedData structure. Also -md option to use an
1153 alternative message digest algorithm for signing.
1156 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1157 create PKCS7 structures containing multiple signers. Update smime
1158 application to support multiple signers.
1161 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1165 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1166 Reorganize PBE internals to lookup from a static table using NIDs,
1167 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1168 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1169 PRF which will be automatically used with PBES2.
1172 *) Replace the algorithm specific calls to generate keys in "req" with the
1176 *) Update PKCS#7 enveloped data routines to use new API. This is now
1177 supported by any public key method supporting the encrypt operation. A
1178 ctrl is added to allow the public key algorithm to examine or modify
1179 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1183 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1184 a default digest type to use. In most cases this will be SHA1 but some
1185 algorithms (such as GOST) need to specify an alternative digest. The
1186 return value indicates how strong the prefernce is 1 means optional and
1187 2 is mandatory (that is it is the only supported type). Modify
1188 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1189 use the default md. Update openssl utilities to use the default digest
1190 type for signing if it is not explicitly indicated.
1193 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1194 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1195 signing method from the key type. This effectively removes the link
1196 between digests and public key types.
1199 *) Add an OID cross reference table and utility functions. Its purpose is to
1200 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1201 rsaEncryption. This will allow some of the algorithm specific hackery
1202 needed to use the correct OID to be removed.
1205 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1206 structures for PKCS7_sign(). They are now set up by the relevant public
1210 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1213 *) Add support for key derivation (agreement) in the API, DH method and
1217 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1218 public and private key formats. As a side effect these add additional
1219 command line functionality not previously available: DSA signatures can be
1220 generated and verified using pkeyutl and DH key support and generation in
1225 [Oliver Tappe <zooey@hirschkaefer.de>]
1227 *) New make target "install_html_docs" installs HTML renditions of the
1229 [Oliver Tappe <zooey@hirschkaefer.de>]
1231 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1232 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1233 support key and parameter generation and add initial key generation
1234 functionality for RSA.
1237 *) Add functions for main EVP_PKEY_method operations. The undocumented
1238 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1239 EVP_PKEY_{encrypt,decrypt}_old.
1242 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1243 key API, doesn't do much yet.
1246 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1247 public key algorithms. New option to openssl utility:
1248 "list-public-key-algorithms" to print out info.
1251 *) Implement the Supported Elliptic Curves Extension for
1252 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1255 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1256 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1259 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1260 utilities such as rsa, dsa, dsaparam etc except they process any key
1264 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1265 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1266 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1270 *) Initial support for pluggable public key ASN1.
1271 De-spaghettify the public key ASN1 handling. Move public and private
1272 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1273 algorithm specific handling to a single module within the relevant
1274 algorithm directory. Add functions to allow (near) opaque processing
1275 of public and private key structures.
1278 *) Implement the Supported Point Formats Extension for
1279 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1282 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1283 for the psk identity [hint] and the psk callback functions to the
1284 SSL_SESSION, SSL and SSL_CTX structure.
1287 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1291 SSL_CTX_use_psk_identity_hint
1292 SSL_get_psk_identity_hint
1293 SSL_get_psk_identity
1294 SSL_use_psk_identity_hint
1296 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1298 *) Add RFC 3161 compliant time stamp request creation, response generation
1299 and response verification functionality.
1300 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1302 *) Add initial support for TLS extensions, specifically for the server_name
1303 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1304 have new members for a host name. The SSL data structure has an
1305 additional member SSL_CTX *initial_ctx so that new sessions can be
1306 stored in that context to allow for session resumption, even after the
1307 SSL has been switched to a new SSL_CTX in reaction to a client's
1308 server_name extension.
1310 New functions (subject to change):
1312 SSL_get_servername()
1313 SSL_get_servername_type()
1316 New CTRL codes and macros (subject to change):
1318 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1319 - SSL_CTX_set_tlsext_servername_callback()
1320 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1321 - SSL_CTX_set_tlsext_servername_arg()
1322 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1324 openssl s_client has a new '-servername ...' option.
1326 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1327 '-key2 ...', '-servername_fatal' (subject to change). This allows
1328 testing the HostName extension for a specific single host name ('-cert'
1329 and '-key' remain fallbacks for handshakes without HostName
1330 negotiation). If the unrecogninzed_name alert has to be sent, this by
1331 default is a warning; it becomes fatal with the '-servername_fatal'
1334 [Peter Sylvester, Remy Allais, Christophe Renou]
1336 *) Whirlpool hash implementation is added.
1339 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1340 bn(64,32). Because of instruction set limitations it doesn't have
1341 any negative impact on performance. This was done mostly in order
1342 to make it possible to share assembler modules, such as bn_mul_mont
1343 implementations, between 32- and 64-bit builds without hassle.
1346 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1347 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1351 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1352 dedicated Montgomery multiplication procedure, is introduced.
1353 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1354 "64-bit" performance on certain 32-bit targets.
1357 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1358 in SSL structures. New SSL ctrl to set maximum send fragment size.
1359 Save memory by seeting the I/O buffer sizes dynamically instead of
1360 using the maximum available value.
1363 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1364 in addition to the text details.
1367 *) Very, very preliminary EXPERIMENTAL support for printing of general
1368 ASN1 structures. This currently produces rather ugly output and doesn't
1369 handle several customised structures at all.
1372 *) Integrated support for PVK file format and some related formats such
1373 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1374 these in the 'rsa' and 'dsa' utilities.
1377 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1380 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1381 place for the (very old) "NETSCAPE" format certificates which are now
1382 handled using new ASN1 code equivalents.
1385 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1386 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1387 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1390 *) Modify CRL distribution points extension code to print out previously
1391 unsupported fields. Enhance extension setting code to allow setting of
1395 *) Add print and set support for Issuing Distribution Point CRL extension.
1398 *) Change 'Configure' script to enable Camellia by default.
1401 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1403 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1404 Thanks to Antonio Martin, Enterprise Secure Access Research and
1405 Development, Cisco Systems, Inc. for discovering this bug and
1406 preparing a fix. (CVE-2012-0050)
1409 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1411 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1412 of the Vaudenay padding oracle attack on CBC mode encryption
1413 which enables an efficient plaintext recovery attack against
1414 the OpenSSL implementation of DTLS. Their attack exploits timing
1415 differences arising during decryption processing. A research
1416 paper describing this attack can be found at:
1417 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1418 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1419 Security Group at Royal Holloway, University of London
1420 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1421 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1422 for preparing the fix. (CVE-2011-4108)
1423 [Robin Seggelmann, Michael Tuexen]
1425 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1426 [Ben Laurie, Kasper <ekasper@google.com>]
1428 *) Clear bytes used for block padding of SSL 3.0 records.
1430 [Adam Langley (Google)]
1432 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1433 Kadianakis <desnacked@gmail.com> for discovering this issue and
1434 Adam Langley for preparing the fix. (CVE-2011-4619)
1435 [Adam Langley (Google)]
1437 *) Prevent malformed RFC3779 data triggering an assertion failure.
1438 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1439 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1440 [Rob Austein <sra@hactrn.net>]
1442 *) Fix ssl_ciph.c set-up race.
1443 [Adam Langley (Google)]
1445 *) Fix spurious failures in ecdsatest.c.
1446 [Emilia Käsper (Google)]
1448 *) Fix the BIO_f_buffer() implementation (which was mixing different
1449 interpretations of the '..._len' fields).
1450 [Adam Langley (Google)]
1452 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1453 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1454 threads won't reuse the same blinding coefficients.
1456 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1457 lock to call BN_BLINDING_invert_ex, and avoids one use of
1458 BN_BLINDING_update for each BN_BLINDING structure (previously,
1459 the last update always remained unused).
1460 [Emilia Käsper (Google)]
1462 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1463 for multi-threaded use of ECDH.
1464 [Adam Langley (Google)]
1466 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1469 *) Add protection against ECDSA timing attacks as mentioned in the paper
1470 by Billy Bob Brumley and Nicola Tuveri, see:
1472 http://eprint.iacr.org/2011/232.pdf
1474 [Billy Bob Brumley and Nicola Tuveri]
1476 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1478 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1479 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1481 *) Fix bug in string printing code: if *any* escaping is enabled we must
1482 escape the escape character (backslash) or the resulting string is
1486 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1488 *) Disable code workaround for ancient and obsolete Netscape browsers
1489 and servers: an attacker can use it in a ciphersuite downgrade attack.
1490 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1493 *) Fixed J-PAKE implementation error, originally discovered by
1494 Sebastien Martini, further info and confirmation from Stefan
1495 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1498 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1500 *) Fix extension code to avoid race conditions which can result in a buffer
1501 overrun vulnerability: resumed sessions must not be modified as they can
1502 be shared by multiple threads. CVE-2010-3864
1505 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1508 *) Don't reencode certificate when calculating signature: cache and use
1509 the original encoding instead. This makes signature verification of
1510 some broken encodings work correctly.
1513 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1514 is also one of the inputs.
1515 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1517 *) Don't repeatedly append PBE algorithms to table if they already exist.
1518 Sort table on each new add. This effectively makes the table read only
1519 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1523 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1525 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1528 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1529 access or freeing data twice (CVE-2010-0742)
1530 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1532 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1533 common in certificates and some applications which only call
1534 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1538 Reduce copying into .apps and .test in makevms.com
1539 Don't try to use blank CA certificate in CA.com
1540 Allow use of C files from original directories in maketests.com
1541 [Steven M. Schweda" <sms@antinode.info>]
1543 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1545 *) When rejecting SSL/TLS records due to an incorrect version number, never
1546 update s->server with a new major version number. As of
1547 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1548 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1549 the previous behavior could result in a read attempt at NULL when
1550 receiving specific incorrect SSL/TLS records once record payload
1551 protection is active. (CVE-2010-0740)
1552 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1554 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1555 could be crashed if the relevant tables were not present (e.g. chrooted).
1556 [Tomas Hoger <thoger@redhat.com>]
1558 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1560 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1561 [Martin Olsson, Neel Mehta]
1563 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1564 accommodate for stack sorting, always a write lock!).
1567 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1568 excessive delays in the RAND_poll(): over a minute. As a workaround
1569 include a time check in the inner Heap32Next loop too.
1572 *) The code that handled flushing of data in SSL/TLS originally used the
1573 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1574 the problem outlined in PR#1949. The fix suggested there however can
1575 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1576 of Apache). So instead simplify the code to flush unconditionally.
1577 This should be fine since flushing with no data to flush is a no op.
1580 *) Handle TLS versions 2.0 and later properly and correctly use the
1581 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1582 off ancient servers have a habit of sticking around for a while...
1585 *) Modify compression code so it frees up structures without using the
1586 ex_data callbacks. This works around a problem where some applications
1587 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1588 restarting) then use compression (e.g. SSL with compression) later.
1589 This results in significant per-connection memory leaks and
1590 has caused some security issues including CVE-2008-1678 and
1594 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1595 change when encrypting or decrypting.
1598 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1599 connect and renegotiate with servers which do not support RI.
1600 Until RI is more widely deployed this option is enabled by default.
1603 *) Add "missing" ssl ctrls to clear options and mode.
1606 *) If client attempts to renegotiate and doesn't support RI respond with
1607 a no_renegotiation alert as required by RFC5746. Some renegotiating
1608 TLS clients will continue a connection gracefully when they receive
1609 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1610 waiting for a server hello which it will never receive. Now we treat a
1611 received no_renegotiation alert as a fatal error. This is because
1612 applications requesting a renegotiation might well expect it to succeed
1613 and would have no code in place to handle the server denying it so the
1614 only safe thing to do is to terminate the connection.
1617 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1618 peer supports secure renegotiation and 0 otherwise. Print out peer
1619 renegotiation support in s_client/s_server.
1622 *) Replace the highly broken and deprecated SPKAC certification method with
1623 the updated NID creation version. This should correctly handle UTF8.
1626 *) Implement RFC5746. Re-enable renegotiation but require the extension
1627 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1628 turns out to be a bad idea. It has been replaced by
1629 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1630 SSL_CTX_set_options(). This is really not recommended unless you
1631 know what you are doing.
1632 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1634 *) Fixes to stateless session resumption handling. Use initial_ctx when
1635 issuing and attempting to decrypt tickets in case it has changed during
1636 servername handling. Use a non-zero length session ID when attempting
1637 stateless session resumption: this makes it possible to determine if
1638 a resumption has occurred immediately after receiving server hello
1639 (several places in OpenSSL subtly assume this) instead of later in
1643 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1644 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1645 fixes for a few places where the return code is not checked
1647 [Julia Lawall <julia@diku.dk>]
1649 *) Add --strict-warnings option to Configure script to include devteam
1650 warnings in other configurations.
1653 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1654 makes it possible to install openssl libraries in locations which
1655 have names other than "lib", for example "/usr/lib64" which some
1657 [Steve Henson, based on patch from Jeremy Utley]
1659 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1660 X690 8.9.12 and can produce some misleading textual output of OIDs.
1661 [Steve Henson, reported by Dan Kaminsky]
1663 *) Delete MD2 from algorithm tables. This follows the recommendation in
1664 several standards that it is not used in new applications due to
1665 several cryptographic weaknesses. For binary compatibility reasons
1666 the MD2 API is still compiled in by default.
1669 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1673 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1674 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1676 [Guenter <lists@gknw.net>]
1678 *) Fix the server certificate chain building code to use X509_verify_cert(),
1679 it used to have an ad-hoc builder which was unable to cope with anything
1680 other than a simple chain.
1681 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1683 *) Don't check self signed certificate signatures in X509_verify_cert()
1684 by default (a flag can override this): it just wastes time without
1685 adding any security. As a useful side effect self signed root CAs
1686 with non-FIPS digests are now usable in FIPS mode.
1689 *) In dtls1_process_out_of_seq_message() the check if the current message
1690 is already buffered was missing. For every new message was memory
1691 allocated, allowing an attacker to perform an denial of service attack
1692 with sending out of seq handshake messages until there is no memory
1693 left. Additionally every future messege was buffered, even if the
1694 sequence number made no sense and would be part of another handshake.
1695 So only messages with sequence numbers less than 10 in advance will be
1696 buffered. (CVE-2009-1378)
1697 [Robin Seggelmann, discovered by Daniel Mentz]
1699 *) Records are buffered if they arrive with a future epoch to be
1700 processed after finishing the corresponding handshake. There is
1701 currently no limitation to this buffer allowing an attacker to perform
1702 a DOS attack with sending records with future epochs until there is no
1703 memory left. This patch adds the pqueue_size() function to detemine
1704 the size of a buffer and limits the record buffer to 100 entries.
1706 [Robin Seggelmann, discovered by Daniel Mentz]
1708 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1709 parent structure is freed. (CVE-2009-1379)
1712 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1713 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1716 [Ilya O. <vrghost@gmail.com>]
1718 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1720 *) Disable renegotiation completely - this fixes a severe security
1721 problem (CVE-2009-3555) at the cost of breaking all
1722 renegotiation. Renegotiation can be re-enabled by setting
1723 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1724 run-time. This is really not recommended unless you know what
1728 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1730 *) Don't set val to NULL when freeing up structures, it is freed up by
1731 underlying code. If sizeof(void *) > sizeof(long) this can result in
1732 zeroing past the valid field. (CVE-2009-0789)
1733 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1735 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1736 checked correctly. This would allow some invalid signed attributes to
1737 appear to verify correctly. (CVE-2009-0591)
1738 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1740 *) Reject UniversalString and BMPString types with invalid lengths. This
1741 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1742 a legal length. (CVE-2009-0590)
1745 *) Set S/MIME signing as the default purpose rather than setting it
1746 unconditionally. This allows applications to override it at the store
1750 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1751 to handle some structures.
1754 *) Improve efficiency of mem_gets: don't search whole buffer each time
1756 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1758 *) New -hex option for openssl rand.
1761 *) Print out UTF8String and NumericString when parsing ASN1.
1764 *) Support NumericString type for name components.
1767 *) Allow CC in the environment to override the automatically chosen
1768 compiler. Note that nothing is done to ensure flags work with the
1772 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1774 *) Properly check EVP_VerifyFinal() and similar return values
1776 [Ben Laurie, Bodo Moeller, Google Security Team]
1778 *) Enable TLS extensions by default.
1781 *) Allow the CHIL engine to be loaded, whether the application is
1782 multithreaded or not. (This does not release the developer from the
1783 obligation to set up the dynamic locking callbacks.)
1784 [Sander Temme <sander@temme.net>]
1786 *) Use correct exit code if there is an error in dgst command.
1787 [Steve Henson; problem pointed out by Roland Dirlewanger]
1789 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1790 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1793 *) Add experimental JPAKE support, including demo authentication in
1794 s_client and s_server.
1797 *) Set the comparison function in v3_addr_canonize().
1798 [Rob Austein <sra@hactrn.net>]
1800 *) Add support for XMPP STARTTLS in s_client.
1801 [Philip Paeps <philip@freebsd.org>]
1803 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1804 to ensure that even with this option, only ciphersuites in the
1805 server's preference list will be accepted. (Note that the option
1806 applies only when resuming a session, so the earlier behavior was
1807 just about the algorithm choice for symmetric cryptography.)
1810 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1812 *) Fix NULL pointer dereference if a DTLS server received
1813 ChangeCipherSpec as first record (CVE-2009-1386).
1816 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1817 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1820 *) The fix in 0.9.8c that supposedly got rid of unsafe
1821 double-checked locking was incomplete for RSA blinding,
1822 addressing just one layer of what turns out to have been
1823 doubly unsafe triple-checked locking.
1825 So now fix this for real by retiring the MONT_HELPER macro
1826 in crypto/rsa/rsa_eay.c.
1828 [Bodo Moeller; problem pointed out by Marius Schilder]
1830 *) Various precautionary measures:
1832 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1834 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1835 (NB: This would require knowledge of the secret session ticket key
1836 to exploit, in which case you'd be SOL either way.)
1838 - Change bn_nist.c so that it will properly handle input BIGNUMs
1839 outside the expected range.
1841 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1844 [Neel Mehta, Bodo Moeller]
1846 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1847 the load fails. Useful for distros.
1848 [Ben Laurie and the FreeBSD team]
1850 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1853 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1856 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1858 This work was sponsored by Logica.
1861 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1862 keystores. Support for SSL/TLS client authentication too.
1863 Not compiled unless enable-capieng specified to Configure.
1865 This work was sponsored by Logica.
1868 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1869 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1870 attribute creation routines such as certifcate requests and PKCS#12
1874 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1876 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1877 handshake which could lead to a cilent crash as found using the
1878 Codenomicon TLS test suite (CVE-2008-1672)
1879 [Steve Henson, Mark Cox]
1881 *) Fix double free in TLS server name extensions which could lead to
1882 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1885 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1887 Clear the error queue to ensure that error entries left from
1888 older function calls do not interfere with the correct operation.
1889 [Lutz Jaenicke, Erik de Castro Lopo]
1891 *) Remove root CA certificates of commercial CAs:
1893 The OpenSSL project does not recommend any specific CA and does not
1894 have any policy with respect to including or excluding any CA.
1895 Therefore it does not make any sense to ship an arbitrary selection
1896 of root CA certificates with the OpenSSL software.
1899 *) RSA OAEP patches to fix two separate invalid memory reads.
1900 The first one involves inputs when 'lzero' is greater than
1901 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1902 before the beginning of from). The second one involves inputs where
1903 the 'db' section contains nothing but zeroes (there is a one-byte
1904 invalid read after the end of 'db').
1905 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1907 *) Partial backport from 0.9.9-dev:
1909 Introduce bn_mul_mont (dedicated Montgomery multiplication
1910 procedure) as a candidate for BIGNUM assembler implementation.
1911 While 0.9.9-dev uses assembler for various architectures, only
1912 x86_64 is available by default here in the 0.9.8 branch, and
1913 32-bit x86 is available through a compile-time setting.
1915 To try the 32-bit x86 assembler implementation, use Configure
1916 option "enable-montasm" (which exists only for this backport).
1918 As "enable-montasm" for 32-bit x86 disclaims code stability
1919 anyway, in this constellation we activate additional code
1920 backported from 0.9.9-dev for further performance improvements,
1921 namely BN_from_montgomery_word. (To enable this otherwise,
1922 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1924 [Andy Polyakov (backport partially by Bodo Moeller)]
1926 *) Add TLS session ticket callback. This allows an application to set
1927 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1928 values. This is useful for key rollover for example where several key
1929 sets may exist with different names.
1932 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1933 This was broken until now in 0.9.8 releases, such that the only way
1934 a registered ENGINE could be used (assuming it initialises
1935 successfully on the host) was to explicitly set it as the default
1936 for the relevant algorithms. This is in contradiction with 0.9.7
1937 behaviour and the documentation. With this fix, when an ENGINE is
1938 registered into a given algorithm's table of implementations, the
1939 'uptodate' flag is reset so that auto-discovery will be used next
1940 time a new context for that algorithm attempts to select an
1942 [Ian Lister (tweaked by Geoff Thorpe)]
1944 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1945 implemention in the following ways:
1947 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1950 Lack of BER streaming support means one pass streaming processing is
1951 only supported if data is detached: setting the streaming flag is
1952 ignored for embedded content.
1954 CMS support is disabled by default and must be explicitly enabled
1955 with the enable-cms configuration option.
1958 *) Update the GMP engine glue to do direct copies between BIGNUM and
1959 mpz_t when openssl and GMP use the same limb size. Otherwise the
1960 existing "conversion via a text string export" trick is still used.
1961 [Paul Sheer <paulsheer@gmail.com>]
1963 *) Zlib compression BIO. This is a filter BIO which compressed and
1964 uncompresses any data passed through it.
1967 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1968 RFC3394 compatible AES key wrapping.
1971 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1972 sets string data without copying. X509_ALGOR_set0() and
1973 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1974 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1975 from an X509_ATTRIBUTE structure optionally checking it occurs only
1976 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1980 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1981 to get the expected BN_FLG_CONSTTIME behavior.
1982 [Bodo Moeller (Google)]
1986 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1987 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1988 - added some more tests to do_tests.pl
1989 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1990 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1991 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1992 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1993 - various changes to netware.pl to enable gcc-cross builds on Win32
1995 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1996 - various changes to fix missing prototype warnings
1997 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1998 - added AES, WHIRLPOOL and CPUID assembler code to build files
1999 - added missing AES assembler make rules to mk1mf.pl
2000 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2001 [Guenter Knauf <eflash@gmx.net>]
2003 *) Implement certificate status request TLS extension defined in RFC3546.
2004 A client can set the appropriate parameters and receive the encoded
2005 OCSP response via a callback. A server can query the supplied parameters
2006 and set the encoded OCSP response in the callback. Add simplified examples
2007 to s_client and s_server.
2010 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2012 *) Fix various bugs:
2013 + Binary incompatibility of ssl_ctx_st structure
2014 + DTLS interoperation with non-compliant servers
2015 + Don't call get_session_cb() without proposed session
2016 + Fix ia64 assembler code
2017 [Andy Polyakov, Steve Henson]
2019 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2021 *) DTLS Handshake overhaul. There were longstanding issues with
2022 OpenSSL DTLS implementation, which were making it impossible for
2023 RFC 4347 compliant client to communicate with OpenSSL server.
2024 Unfortunately just fixing these incompatibilities would "cut off"
2025 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2026 server keeps tolerating non RFC compliant syntax. The opposite is
2027 not true, 0.9.8f client can not communicate with earlier server.
2028 This update even addresses CVE-2007-4995.
2031 *) Changes to avoid need for function casts in OpenSSL: some compilers
2032 (gcc 4.2 and later) reject their use.
2033 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2036 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2037 RFC4507bis. The encrypted ticket format is an encrypted encoded
2038 SSL_SESSION structure, that way new session features are automatically
2041 If a client application caches session in an SSL_SESSION structure
2042 support is transparent because tickets are now stored in the encoded
2045 The SSL_CTX structure automatically generates keys for ticket
2046 protection in servers so again support should be possible
2047 with no application modification.
2049 If a client or server wishes to disable RFC4507 support then the option
2050 SSL_OP_NO_TICKET can be set.
2052 Add a TLS extension debugging callback to allow the contents of any client
2053 or server extensions to be examined.
2055 This work was sponsored by Google.
2058 *) Add initial support for TLS extensions, specifically for the server_name
2059 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2060 have new members for a host name. The SSL data structure has an
2061 additional member SSL_CTX *initial_ctx so that new sessions can be
2062 stored in that context to allow for session resumption, even after the
2063 SSL has been switched to a new SSL_CTX in reaction to a client's
2064 server_name extension.
2066 New functions (subject to change):
2068 SSL_get_servername()
2069 SSL_get_servername_type()
2072 New CTRL codes and macros (subject to change):
2074 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2075 - SSL_CTX_set_tlsext_servername_callback()
2076 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2077 - SSL_CTX_set_tlsext_servername_arg()
2078 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2080 openssl s_client has a new '-servername ...' option.
2082 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2083 '-key2 ...', '-servername_fatal' (subject to change). This allows
2084 testing the HostName extension for a specific single host name ('-cert'
2085 and '-key' remain fallbacks for handshakes without HostName
2086 negotiation). If the unrecogninzed_name alert has to be sent, this by
2087 default is a warning; it becomes fatal with the '-servername_fatal'
2090 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2092 *) Add AES and SSE2 assembly language support to VC++ build.
2095 *) Mitigate attack on final subtraction in Montgomery reduction.
2098 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2099 (which previously caused an internal error).
2102 *) Squeeze another 10% out of IGE mode when in != out.
2105 *) AES IGE mode speedup.
2106 [Dean Gaudet (Google)]
2108 *) Add the Korean symmetric 128-bit cipher SEED (see
2109 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2110 add SEED ciphersuites from RFC 4162:
2112 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2113 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2114 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2115 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2117 To minimize changes between patchlevels in the OpenSSL 0.9.8
2118 series, SEED remains excluded from compilation unless OpenSSL
2119 is configured with 'enable-seed'.
2120 [KISA, Bodo Moeller]
2122 *) Mitigate branch prediction attacks, which can be practical if a
2123 single processor is shared, allowing a spy process to extract
2124 information. For detailed background information, see
2125 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2126 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2127 and Necessary Software Countermeasures"). The core of the change
2128 are new versions BN_div_no_branch() and
2129 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2130 respectively, which are slower, but avoid the security-relevant
2131 conditional branches. These are automatically called by BN_div()
2132 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2133 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2134 remove a conditional branch.
2136 BN_FLG_CONSTTIME is the new name for the previous
2137 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2138 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2139 in the exponent causes BN_mod_exp_mont() to use the alternative
2140 implementation in BN_mod_exp_mont_consttime().) The old name
2141 remains as a deprecated alias.
2143 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2144 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2145 constant-time implementations for more than just exponentiation.
2146 Here too the old name is kept as a deprecated alias.
2148 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2149 the BN_BLINDING structure gets an independent copy of the
2150 modulus. This means that the previous "BIGNUM *m" argument to
2151 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2152 essentially becomes "const BIGNUM *m", although we can't actually
2153 change this in the header file before 0.9.9. It allows
2154 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2155 enable BN_FLG_CONSTTIME.
2157 [Matthew D Wood (Intel Corp)]
2159 *) In the SSL/TLS server implementation, be strict about session ID
2160 context matching (which matters if an application uses a single
2161 external cache for different purposes). Previously,
2162 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2163 set. This did ensure strict client verification, but meant that,
2164 with applications using a single external cache for quite
2165 different requirements, clients could circumvent ciphersuite
2166 restrictions for a given session ID context by starting a session
2167 in a different context.
2170 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2171 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2172 authentication-only ciphersuites.
2175 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2176 not complete and could lead to a possible single byte overflow
2177 (CVE-2007-5135) [Ben Laurie]
2179 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2181 *) Since AES128 and AES256 (and similarly Camellia128 and
2182 Camellia256) share a single mask bit in the logic of
2183 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2184 kludge to work properly if AES128 is available and AES256 isn't
2185 (or if Camellia128 is available and Camellia256 isn't).
2188 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2189 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2190 When a point or a seed is encoded in a BIT STRING, we need to
2191 prevent the removal of trailing zero bits to get the proper DER
2192 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2193 of a NamedBitList, for which trailing 0 bits need to be removed.)
2196 *) Have SSL/TLS server implementation tolerate "mismatched" record
2197 protocol version while receiving ClientHello even if the
2198 ClientHello is fragmented. (The server can't insist on the
2199 particular protocol version it has chosen before the ServerHello
2200 message has informed the client about his choice.)
2203 *) Add RFC 3779 support.
2204 [Rob Austein for ARIN, Ben Laurie]
2206 *) Load error codes if they are not already present instead of using a
2207 static variable. This allows them to be cleanly unloaded and reloaded.
2208 Improve header file function name parsing.
2211 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2212 or CAPABILITY handshake as required by RFCs.
2215 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2217 *) Introduce limits to prevent malicious keys being able to
2218 cause a denial of service. (CVE-2006-2940)
2219 [Steve Henson, Bodo Moeller]
2221 *) Fix ASN.1 parsing of certain invalid structures that can result
2222 in a denial of service. (CVE-2006-2937) [Steve Henson]
2224 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2225 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2227 *) Fix SSL client code which could crash if connecting to a
2228 malicious SSLv2 server. (CVE-2006-4343)
2229 [Tavis Ormandy and Will Drewry, Google Security Team]
2231 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2232 match only those. Before that, "AES256-SHA" would be interpreted
2233 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2234 the same strength classification in 0.9.7h) as we currently only
2235 have a single AES bit in the ciphersuite description bitmap.
2236 That change, however, also applied to ciphersuite strings such as
2237 "RC4-MD5" that intentionally matched multiple ciphersuites --
2238 namely, SSL 2.0 ciphersuites in addition to the more common ones
2239 from SSL 3.0/TLS 1.0.
2241 So we change the selection algorithm again: Naming an explicit
2242 ciphersuite selects this one ciphersuite, and any other similar
2243 ciphersuite (same bitmap) from *other* protocol versions.
2244 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2245 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2247 Since SSL 2.0 does not have any ciphersuites for which the
2248 128/256 bit distinction would be relevant, this works for now.
2249 The proper fix will be to use different bits for AES128 and
2250 AES256, which would have avoided the problems from the beginning;
2251 however, bits are scarce, so we can only do this in a new release
2252 (not just a patchlevel) when we can change the SSL_CIPHER
2253 definition to split the single 'unsigned long mask' bitmap into
2254 multiple values to extend the available space.
2258 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2260 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2261 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2263 *) Add AES IGE and biIGE modes.
2266 *) Change the Unix randomness entropy gathering to use poll() when
2267 possible instead of select(), since the latter has some
2268 undesirable limitations.
2269 [Darryl Miles via Richard Levitte and Bodo Moeller]
2271 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2272 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2273 cannot be implicitly activated as part of, e.g., the "AES" alias.
2274 However, please upgrade to OpenSSL 0.9.9[-dev] for
2275 non-experimental use of the ECC ciphersuites to get TLS extension
2276 support, which is required for curve and point format negotiation
2277 to avoid potential handshake problems.
2280 *) Disable rogue ciphersuites:
2282 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2283 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2284 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2286 The latter two were purportedly from
2287 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2290 Also deactivate the remaining ciphersuites from
2291 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2292 unofficial, and the ID has long expired.
2295 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2296 dual-core machines) and other potential thread-safety issues.
2299 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2300 versions), which is now available for royalty-free use
2301 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2302 Also, add Camellia TLS ciphersuites from RFC 4132.
2304 To minimize changes between patchlevels in the OpenSSL 0.9.8
2305 series, Camellia remains excluded from compilation unless OpenSSL
2306 is configured with 'enable-camellia'.
2309 *) Disable the padding bug check when compression is in use. The padding
2310 bug check assumes the first packet is of even length, this is not
2311 necessarily true if compresssion is enabled and can result in false
2312 positives causing handshake failure. The actual bug test is ancient
2313 code so it is hoped that implementations will either have fixed it by
2314 now or any which still have the bug do not support compression.
2317 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2319 *) When applying a cipher rule check to see if string match is an explicit
2320 cipher suite and only match that one cipher suite if it is.
2323 *) Link in manifests for VC++ if needed.
2324 [Austin Ziegler <halostatue@gmail.com>]
2326 *) Update support for ECC-based TLS ciphersuites according to
2327 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2328 TLS extensions, which are supported starting with the 0.9.9
2329 branch, not in the OpenSSL 0.9.8 branch).
2332 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2333 opaque EVP_CIPHER_CTX handling.
2336 *) Fixes and enhancements to zlib compression code. We now only use
2337 "zlib1.dll" and use the default __cdecl calling convention on Win32
2338 to conform with the standards mentioned here:
2339 http://www.zlib.net/DLL_FAQ.txt
2340 Static zlib linking now works on Windows and the new --with-zlib-include
2341 --with-zlib-lib options to Configure can be used to supply the location
2342 of the headers and library. Gracefully handle case where zlib library
2346 *) Several fixes and enhancements to the OID generation code. The old code
2347 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2348 handle numbers larger than ULONG_MAX, truncated printing and had a
2349 non standard OBJ_obj2txt() behaviour.
2352 *) Add support for building of engines under engine/ as shared libraries
2353 under VC++ build system.
2356 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2357 Hopefully, we will not see any false combination of paths any more.
2360 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2362 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2363 (part of SSL_OP_ALL). This option used to disable the
2364 countermeasure against man-in-the-middle protocol-version
2365 rollback in the SSL 2.0 server implementation, which is a bad
2366 idea. (CVE-2005-2969)
2368 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2369 for Information Security, National Institute of Advanced Industrial
2370 Science and Technology [AIST], Japan)]
2372 *) Add two function to clear and return the verify parameter flags.
2375 *) Keep cipherlists sorted in the source instead of sorting them at
2376 runtime, thus removing the need for a lock.
2379 *) Avoid some small subgroup attacks in Diffie-Hellman.
2380 [Nick Mathewson and Ben Laurie]
2382 *) Add functions for well-known primes.
2385 *) Extended Windows CE support.
2386 [Satoshi Nakamura and Andy Polyakov]
2388 *) Initialize SSL_METHOD structures at compile time instead of during
2389 runtime, thus removing the need for a lock.
2392 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2393 attempting to decrypt each encrypted key in turn. Add support to
2397 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2399 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2402 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2405 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2406 key into the same file any more.
2409 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2412 *) Add -utf8 command line and config file option to 'ca'.
2413 [Stefan <stf@udoma.org]
2415 *) Removed the macro des_crypt(), as it seems to conflict with some
2416 libraries. Use DES_crypt().
2419 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2420 involves renaming the source and generated shared-libs for
2421 both. The engines will accept the corrected or legacy ids
2422 ('ncipher' and '4758_cca' respectively) when binding. NB,
2423 this only applies when building 'shared'.
2424 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2426 *) Add attribute functions to EVP_PKEY structure. Modify
2427 PKCS12_create() to recognize a CSP name attribute and
2428 use it. Make -CSP option work again in pkcs12 utility.
2431 *) Add new functionality to the bn blinding code:
2432 - automatic re-creation of the BN_BLINDING parameters after
2433 a fixed number of uses (currently 32)
2434 - add new function for parameter creation
2435 - introduce flags to control the update behaviour of the
2436 BN_BLINDING parameters
2437 - hide BN_BLINDING structure
2438 Add a second BN_BLINDING slot to the RSA structure to improve
2439 performance when a single RSA object is shared among several
2443 *) Add support for DTLS.
2444 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2446 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2447 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2450 *) Remove buggy and incompletet DH cert support from
2451 ssl/ssl_rsa.c and ssl/s3_both.c
2454 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2455 the apps/openssl applications.
2458 *) Compile clean with "-Wall -Wmissing-prototypes
2459 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2460 DEBUG_SAFESTACK must also be set.
2463 *) Change ./Configure so that certain algorithms can be disabled by default.
2464 The new counterpiece to "no-xxx" is "enable-xxx".
2466 The patented RC5 and MDC2 algorithms will now be disabled unless
2467 "enable-rc5" and "enable-mdc2", respectively, are specified.
2469 (IDEA remains enabled despite being patented. This is because IDEA
2470 is frequently required for interoperability, and there is no license
2471 fee for non-commercial use. As before, "no-idea" can be used to
2472 avoid this algorithm.)
2476 *) Add processing of proxy certificates (see RFC 3820). This work was
2477 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2478 EGEE (Enabling Grids for E-science in Europe).
2481 *) RC4 performance overhaul on modern architectures/implementations, such
2482 as Intel P4, IA-64 and AMD64.
2485 *) New utility extract-section.pl. This can be used specify an alternative
2486 section number in a pod file instead of having to treat each file as
2487 a separate case in Makefile. This can be done by adding two lines to the
2490 =for comment openssl_section:XXX
2492 The blank line is mandatory.
2496 *) New arguments -certform, -keyform and -pass for s_client and s_server
2497 to allow alternative format key and certificate files and passphrase
2501 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2502 update associated structures and add various utility functions.
2504 Add new policy related verify parameters, include policy checking in
2505 standard verify code. Enhance 'smime' application with extra parameters
2506 to support policy checking and print out.
2509 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2510 Nehemiah processors. These extensions support AES encryption in hardware
2511 as well as RNG (though RNG support is currently disabled).
2512 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2514 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2517 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2518 [Andy Polyakov and a number of other people]
2520 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2521 implementation contributed by IBM.
2522 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2524 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2525 exponent rather than 'unsigned long'. There is a corresponding change to
2526 the new 'rsa_keygen' element of the RSA_METHOD structure.
2527 [Jelte Jansen, Geoff Thorpe]
2529 *) Functionality for creating the initial serial number file is now
2530 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2532 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2533 number file to 1, which is bound to cause problems. To avoid
2534 the problems while respecting compatibility between different 0.9.7
2535 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2536 CA.pl for serial number initialization. With the new release 0.9.8,
2537 we can fix the problem directly in the 'ca' utility.)
2540 *) Reduced header interdepencies by declaring more opaque objects in
2541 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2542 give fewer recursive includes, which could break lazy source code - so
2543 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2544 developers should define this symbol when building and using openssl to
2545 ensure they track the recommended behaviour, interfaces, [etc], but
2546 backwards-compatible behaviour prevails when this isn't defined.
2549 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2552 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2553 This will generate a random key of the appropriate length based on the
2554 cipher context. The EVP_CIPHER can provide its own random key generation
2555 routine to support keys of a specific form. This is used in the des and
2556 3des routines to generate a key of the correct parity. Update S/MIME
2557 code to use new functions and hence generate correct parity DES keys.
2558 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2559 valid (weak or incorrect parity).
2562 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2563 as looking them up. This is useful when the verified structure may contain
2564 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2565 present unless the new PKCS7_NO_CRL flag is asserted.
2568 *) Extend ASN1 oid configuration module. It now additionally accepts the
2571 shortName = some long name, 1.2.3.4
2574 *) Reimplemented the BN_CTX implementation. There is now no more static
2575 limitation on the number of variables it can handle nor the depth of the
2576 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2577 information can now expand as required, and rather than having a single
2578 static array of bignums, BN_CTX now uses a linked-list of such arrays
2579 allowing it to expand on demand whilst maintaining the usefulness of
2580 BN_CTX's "bundling".
2583 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2584 to allow all RSA operations to function using a single BN_CTX.
2587 *) Preliminary support for certificate policy evaluation and checking. This
2588 is initially intended to pass the tests outlined in "Conformance Testing
2589 of Relying Party Client Certificate Path Processing Logic" v1.07.
2592 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2593 remained unused and not that useful. A variety of other little bignum
2594 tweaks and fixes have also been made continuing on from the audit (see
2598 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2599 associated ASN1, EVP and SSL functions and old ASN1 macros.
2602 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2603 and this should never fail. So the return value from the use of
2604 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2605 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2608 *) BN_CTX_get() should return zero-valued bignums, providing the same
2609 initialised value as BN_new().
2610 [Geoff Thorpe, suggested by Ulf Möller]
2612 *) Support for inhibitAnyPolicy certificate extension.
2615 *) An audit of the BIGNUM code is underway, for which debugging code is
2616 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2617 is considered valid when processing BIGNUMs, and causes execution to
2618 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2619 further steps are taken to deliberately pollute unused data in BIGNUM
2620 structures to try and expose faulty code further on. For now, openssl will
2621 (in its default mode of operation) continue to tolerate the inconsistent
2622 forms that it has tolerated in the past, but authors and packagers should
2623 consider trying openssl and their own applications when compiled with
2624 these debugging symbols defined. It will help highlight potential bugs in
2625 their own code, and will improve the test coverage for OpenSSL itself. At
2626 some point, these tighter rules will become openssl's default to improve
2627 maintainability, though the assert()s and other overheads will remain only
2628 in debugging configurations. See bn.h for more details.
2629 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2631 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2632 that can only be obtained through BN_CTX_new() (which implicitly
2633 initialises it). The presence of this function only made it possible
2634 to overwrite an existing structure (and cause memory leaks).
2637 *) Because of the callback-based approach for implementing LHASH as a
2638 template type, lh_insert() adds opaque objects to hash-tables and
2639 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2640 to clean up those corresponding objects before destroying the hash table
2641 (and losing the object pointers). So some over-zealous constifications in
2642 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2643 objects as "const" and the lh_doall[_arg] callback wrappers are not
2644 prototyped to have "const" restrictions on the object pointers they are
2645 given (and so aren't required to cast them away any more).
2648 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2649 (speed) prefers to use its own implementation. The two implementations
2650 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2651 its object type properly exposed (MS_TM) instead of casting to/from "char
2652 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2653 aren't necessarily the greatest nomenclatures - but this is what was used
2654 internally to the implementation so I've used that for now.
2657 *) Ensure that deprecated functions do not get compiled when
2658 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2659 the self-tests were still using deprecated key-generation functions so
2660 these have been updated also.
2663 *) Reorganise PKCS#7 code to separate the digest location functionality
2664 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2665 New function PKCS7_set_digest() to set the digest type for PKCS#7
2666 digestedData type. Add additional code to correctly generate the
2667 digestedData type and add support for this type in PKCS7 initialization
2671 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2672 structure of type "other".
2675 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2676 sure the loop does correctly stop and breaking ("division by zero")
2677 modulus operations are not performed. The (pre-generated) prime
2678 table crypto/bn/bn_prime.h was already correct, but it could not be
2679 re-generated on some platforms because of the "division by zero"
2680 situation in the script.
2681 [Ralf S. Engelschall]
2683 *) Update support for ECC-based TLS ciphersuites according to
2684 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2685 SHA-1 now is only used for "small" curves (where the
2686 representation of a field element takes up to 24 bytes); for
2687 larger curves, the field element resulting from ECDH is directly
2688 used as premaster secret.
2689 [Douglas Stebila (Sun Microsystems Laboratories)]
2691 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2692 curve secp160r1 to the tests.
2693 [Douglas Stebila (Sun Microsystems Laboratories)]
2695 *) Add the possibility to load symbols globally with DSO.
2696 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2698 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2699 control of the error stack.
2702 *) Add support for STORE in ENGINE.
2705 *) Add the STORE type. The intention is to provide a common interface
2706 to certificate and key stores, be they simple file-based stores, or
2707 HSM-type store, or LDAP stores, or...
2708 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2711 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2712 pass a list of arguments to any function as well as provide a way
2713 for a function to pass data back to the caller.
2716 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2717 works like BUF_strdup() but can be used to duplicate a portion of
2718 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2722 *) Add the function sk_find_ex() which works like sk_find(), but will
2723 return an index to an element even if an exact match couldn't be
2724 found. The index is guaranteed to point at the element where the
2725 searched-for key would be inserted to preserve sorting order.
2728 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2729 takes an extra flags argument for optional functionality. Currently,
2730 the following flags are defined:
2732 OBJ_BSEARCH_VALUE_ON_NOMATCH
2733 This one gets OBJ_bsearch_ex() to return a pointer to the first
2734 element where the comparing function returns a negative or zero
2737 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2738 This one gets OBJ_bsearch_ex() to return a pointer to the first
2739 element where the comparing function returns zero. This is useful
2740 if there are more than one element where the comparing function
2744 *) Make it possible to create self-signed certificates with 'openssl ca'
2745 in such a way that the self-signed certificate becomes part of the
2746 CA database and uses the same mechanisms for serial number generation
2747 as all other certificate signing. The new flag '-selfsign' enables
2748 this functionality. Adapt CA.sh and CA.pl.in.
2751 *) Add functionality to check the public key of a certificate request
2752 against a given private. This is useful to check that a certificate
2753 request can be signed by that key (self-signing).
2756 *) Make it possible to have multiple active certificates with the same
2757 subject in the CA index file. This is done only if the keyword
2758 'unique_subject' is set to 'no' in the main CA section (default
2759 if 'CA_default') of the configuration file. The value is saved
2760 with the database itself in a separate index attribute file,
2761 named like the index file with '.attr' appended to the name.
2764 *) Generate muti valued AVAs using '+' notation in config files for
2768 *) Support for nameConstraints certificate extension.
2771 *) Support for policyConstraints certificate extension.
2774 *) Support for policyMappings certificate extension.
2777 *) Make sure the default DSA_METHOD implementation only uses its
2778 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2779 and change its own handlers to be NULL so as to remove unnecessary
2780 indirection. This lets alternative implementations fallback to the
2781 default implementation more easily.
2784 *) Support for directoryName in GeneralName related extensions
2788 *) Make it possible to link applications using Makefile.shared.
2789 Make that possible even when linking against static libraries!
2792 *) Support for single pass processing for S/MIME signing. This now
2793 means that S/MIME signing can be done from a pipe, in addition
2794 cleartext signing (multipart/signed type) is effectively streaming
2795 and the signed data does not need to be all held in memory.
2797 This is done with a new flag PKCS7_STREAM. When this flag is set
2798 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2799 is done after the data is output (and digests calculated) in
2800 SMIME_write_PKCS7().
2803 *) Add full support for -rpath/-R, both in shared libraries and
2804 applications, at least on the platforms where it's known how
2808 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2809 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2810 will now compute a table of multiples of the generator that
2811 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2812 faster (notably in the case of a single point multiplication,
2813 scalar * generator).
2814 [Nils Larsch, Bodo Moeller]
2816 *) IPv6 support for certificate extensions. The various extensions
2817 which use the IP:a.b.c.d can now take IPv6 addresses using the
2818 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2822 *) Added an ENGINE that implements RSA by performing private key
2823 exponentiations with the GMP library. The conversions to and from
2824 GMP's mpz_t format aren't optimised nor are any montgomery forms
2825 cached, and on x86 it appears OpenSSL's own performance has caught up.
2826 However there are likely to be other architectures where GMP could
2827 provide a boost. This ENGINE is not built in by default, but it can be
2828 specified at Configure time and should be accompanied by the necessary
2829 linker additions, eg;
2830 ./config -DOPENSSL_USE_GMP -lgmp
2833 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2834 testing availability of engines with "-t" - the old behaviour is
2835 produced by increasing the feature's verbosity with "-tt".
2838 *) ECDSA routines: under certain error conditions uninitialized BN objects
2839 could be freed. Solution: make sure initialization is performed early
2840 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2844 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2845 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2846 software implementations. For DSA and DH, parameter generation can
2847 also be overriden by providing the appropriate method callbacks.
2850 *) Change the "progress" mechanism used in key-generation and
2851 primality testing to functions that take a new BN_GENCB pointer in
2852 place of callback/argument pairs. The new API functions have "_ex"
2853 postfixes and the older functions are reimplemented as wrappers for
2854 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2855 declarations of the old functions to help (graceful) attempts to
2856 migrate to the new functions. Also, the new key-generation API
2857 functions operate on a caller-supplied key-structure and return
2858 success/failure rather than returning a key or NULL - this is to
2859 help make "keygen" another member function of RSA_METHOD etc.
2861 Example for using the new callback interface:
2863 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2867 BN_GENCB_set(&my_cb, my_callback, my_arg);
2869 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2870 /* For the meaning of a, b in calls to my_callback(), see the
2871 * documentation of the function that calls the callback.
2872 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2873 * my_callback should return 1 if it wants BN_is_prime_ex()
2874 * to continue, or 0 to stop.
2879 *) Change the ZLIB compression method to be stateful, and make it
2880 available to TLS with the number defined in
2881 draft-ietf-tls-compression-04.txt.
2884 *) Add the ASN.1 structures and functions for CertificatePair, which
2885 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2887 CertificatePair ::= SEQUENCE {
2888 forward [0] Certificate OPTIONAL,
2889 reverse [1] Certificate OPTIONAL,
2890 -- at least one of the pair shall be present -- }
2892 Also implement the PEM functions to read and write certificate
2893 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2895 This needed to be defined, mostly for the sake of the LDAP
2896 attribute crossCertificatePair, but may prove useful elsewhere as
2900 *) Make it possible to inhibit symlinking of shared libraries in
2901 Makefile.shared, for Cygwin's sake.
2904 *) Extend the BIGNUM API by creating a function
2905 void BN_set_negative(BIGNUM *a, int neg);
2906 and a macro that behave like
2907 int BN_is_negative(const BIGNUM *a);
2909 to avoid the need to access 'a->neg' directly in applications.
2912 *) Implement fast modular reduction for pseudo-Mersenne primes
2913 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2914 EC_GROUP_new_curve_GFp() will now automatically use this
2916 [Nils Larsch <nla@trustcenter.de>]
2918 *) Add new lock type (CRYPTO_LOCK_BN).
2921 *) Change the ENGINE framework to automatically load engines
2922 dynamically from specific directories unless they could be
2923 found to already be built in or loaded. Move all the
2924 current engines except for the cryptodev one to a new
2926 The engines in engines/ are built as shared libraries if
2927 the "shared" options was given to ./Configure or ./config.
2928 Otherwise, they are inserted in libcrypto.a.
2929 /usr/local/ssl/engines is the default directory for dynamic
2930 engines, but that can be overriden at configure time through
2931 the usual use of --prefix and/or --openssldir, and at run
2932 time with the environment variable OPENSSL_ENGINES.
2933 [Geoff Thorpe and Richard Levitte]
2935 *) Add Makefile.shared, a helper makefile to build shared
2936 libraries. Addapt Makefile.org.
2939 *) Add version info to Win32 DLLs.
2940 [Peter 'Luna' Runestig" <peter@runestig.com>]
2942 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2943 can be added using this API to created arbitrary PKCS#12
2944 files while avoiding the low level API.
2946 New options to PKCS12_create(), key or cert can be NULL and
2947 will then be omitted from the output file. The encryption
2948 algorithm NIDs can be set to -1 for no encryption, the mac
2949 iteration count can be set to 0 to omit the mac.
2951 Enhance pkcs12 utility by making the -nokeys and -nocerts
2952 options work when creating a PKCS#12 file. New option -nomac
2953 to omit the mac, NONE can be set for an encryption algorithm.
2954 New code is modified to use the enhanced PKCS12_create()
2955 instead of the low level API.
2958 *) Extend ASN1 encoder to support indefinite length constructed
2959 encoding. This can output sequences tags and octet strings in
2960 this form. Modify pk7_asn1.c to support indefinite length
2961 encoding. This is experimental and needs additional code to
2962 be useful, such as an ASN1 bio and some enhanced streaming
2965 Extend template encode functionality so that tagging is passed
2966 down to the template encoder.
2969 *) Let 'openssl req' fail if an argument to '-newkey' is not
2970 recognized instead of using RSA as a default.
2973 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2974 As these are not official, they are not included in "ALL";
2975 the "ECCdraft" ciphersuite group alias can be used to select them.
2976 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2978 *) Add ECDH engine support.
2979 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2981 *) Add ECDH in new directory crypto/ecdh/.
2982 [Douglas Stebila (Sun Microsystems Laboratories)]
2984 *) Let BN_rand_range() abort with an error after 100 iterations
2985 without success (which indicates a broken PRNG).
2988 *) Change BN_mod_sqrt() so that it verifies that the input value
2989 is really the square of the return value. (Previously,
2990 BN_mod_sqrt would show GIGO behaviour.)
2993 *) Add named elliptic curves over binary fields from X9.62, SECG,
2994 and WAP/WTLS; add OIDs that were still missing.
2996 [Sheueling Chang Shantz and Douglas Stebila
2997 (Sun Microsystems Laboratories)]
2999 *) Extend the EC library for elliptic curves over binary fields
3000 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3003 EC_GF2m_simple_method
3007 EC_GROUP_new_curve_GF2m
3008 EC_GROUP_set_curve_GF2m
3009 EC_GROUP_get_curve_GF2m
3010 EC_POINT_set_affine_coordinates_GF2m
3011 EC_POINT_get_affine_coordinates_GF2m
3012 EC_POINT_set_compressed_coordinates_GF2m
3014 Point compression for binary fields is disabled by default for
3015 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3018 As binary polynomials are represented as BIGNUMs, various members
3019 of the EC_GROUP and EC_POINT data structures can be shared
3020 between the implementations for prime fields and binary fields;
3021 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3022 are essentially identical to their ..._GFp counterparts.
3023 (For simplicity, the '..._GFp' prefix has been dropped from
3024 various internal method names.)
3026 An internal 'field_div' method (similar to 'field_mul' and
3027 'field_sqr') has been added; this is used only for binary fields.
3029 [Sheueling Chang Shantz and Douglas Stebila
3030 (Sun Microsystems Laboratories)]
3032 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3033 through methods ('mul', 'precompute_mult').
3035 The generic implementations (now internally called 'ec_wNAF_mul'
3036 and 'ec_wNAF_precomputed_mult') remain the default if these
3037 methods are undefined.
3039 [Sheueling Chang Shantz and Douglas Stebila
3040 (Sun Microsystems Laboratories)]
3042 *) New function EC_GROUP_get_degree, which is defined through
3043 EC_METHOD. For curves over prime fields, this returns the bit
3044 length of the modulus.
3046 [Sheueling Chang Shantz and Douglas Stebila
3047 (Sun Microsystems Laboratories)]
3049 *) New functions EC_GROUP_dup, EC_POINT_dup.
3050 (These simply call ..._new and ..._copy).
3052 [Sheueling Chang Shantz and Douglas Stebila
3053 (Sun Microsystems Laboratories)]
3055 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3056 Polynomials are represented as BIGNUMs (where the sign bit is not
3057 used) in the following functions [macros]:
3060 BN_GF2m_sub [= BN_GF2m_add]
3061 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3062 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3063 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3065 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3066 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3067 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3068 BN_GF2m_cmp [= BN_ucmp]
3070 (Note that only the 'mod' functions are actually for fields GF(2^m).
3071 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3073 For some functions, an the irreducible polynomial defining a
3074 field can be given as an 'unsigned int[]' with strictly
3075 decreasing elements giving the indices of those bits that are set;
3076 i.e., p[] represents the polynomial
3077 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3079 p[0] > p[1] > ... > p[k] = 0.
3080 This applies to the following functions:
3085 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3086 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3088 BN_GF2m_mod_sqrt_arr
3089 BN_GF2m_mod_solve_quad_arr
3093 Conversion can be performed by the following functions:
3098 bntest.c has additional tests for binary polynomial arithmetic.
3100 Two implementations for BN_GF2m_mod_div() are available.
3101 The default algorithm simply uses BN_GF2m_mod_inv() and
3102 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3103 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3104 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3106 [Sheueling Chang Shantz and Douglas Stebila
3107 (Sun Microsystems Laboratories)]
3109 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3110 functionality is disabled at compile-time.
3111 [Douglas Stebila <douglas.stebila@sun.com>]
3113 *) Change default behaviour of 'openssl asn1parse' so that more
3114 information is visible when viewing, e.g., a certificate:
3116 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3117 mode the content of non-printable OCTET STRINGs is output in a
3118 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3119 avoid the appearance of a printable string.
3120 [Nils Larsch <nla@trustcenter.de>]
3122 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3124 EC_GROUP_set_asn1_flag()
3125 EC_GROUP_get_asn1_flag()
3126 EC_GROUP_set_point_conversion_form()
3127 EC_GROUP_get_point_conversion_form()
3128 These control ASN1 encoding details:
3129 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3130 has been set to OPENSSL_EC_NAMED_CURVE.
3131 - Points are encoded in uncompressed form by default; options for
3132 asn1_for are as for point2oct, namely
3133 POINT_CONVERSION_COMPRESSED
3134 POINT_CONVERSION_UNCOMPRESSED
3135 POINT_CONVERSION_HYBRID
3137 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3140 EC_GROUP_get0_seed()
3141 EC_GROUP_get_seed_len()
3142 This is used only for ASN1 purposes (so far).
3143 [Nils Larsch <nla@trustcenter.de>]
3145 *) Add 'field_type' member to EC_METHOD, which holds the NID
3146 of the appropriate field type OID. The new function
3147 EC_METHOD_get_field_type() returns this value.
3148 [Nils Larsch <nla@trustcenter.de>]
3153 EC_POINT_point2hex()
3154 EC_POINT_hex2point()
3155 providing useful interfaces to EC_POINT_point2oct() and
3156 EC_POINT_oct2point().
3157 [Nils Larsch <nla@trustcenter.de>]
3159 *) Change internals of the EC library so that the functions
3160 EC_GROUP_set_generator()
3161 EC_GROUP_get_generator()
3162 EC_GROUP_get_order()
3163 EC_GROUP_get_cofactor()
3164 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3165 to methods, which would lead to unnecessary code duplication when
3166 adding different types of curves.
3167 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3169 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3170 arithmetic, and such that modified wNAFs are generated
3171 (which avoid length expansion in many cases).
3174 *) Add a function EC_GROUP_check_discriminant() (defined via
3175 EC_METHOD) that verifies that the curve discriminant is non-zero.
3177 Add a function EC_GROUP_check() that makes some sanity tests
3178 on a EC_GROUP, its generator and order. This includes
3179 EC_GROUP_check_discriminant().
3180 [Nils Larsch <nla@trustcenter.de>]
3182 *) Add ECDSA in new directory crypto/ecdsa/.
3184 Add applications 'openssl ecparam' and 'openssl ecdsa'
3185 (these are based on 'openssl dsaparam' and 'openssl dsa').
3187 ECDSA support is also included in various other files across the
3188 library. Most notably,
3189 - 'openssl req' now has a '-newkey ecdsa:file' option;
3190 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3191 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3192 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3193 them suitable for ECDSA where domain parameters must be
3194 extracted before the specific public key;
3195 - ECDSA engine support has been added.
3196 [Nils Larsch <nla@trustcenter.de>]
3198 *) Include some named elliptic curves, and add OIDs from X9.62,
3199 SECG, and WAP/WTLS. Each curve can be obtained from the new
3201 EC_GROUP_new_by_curve_name(),
3202 and the list of available named curves can be obtained with
3203 EC_get_builtin_curves().
3204 Also add a 'curve_name' member to EC_GROUP objects, which can be
3206 EC_GROUP_set_curve_name()
3207 EC_GROUP_get_curve_name()
3208 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3210 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3211 was actually never needed) and in BN_mul(). The removal in BN_mul()
3212 required a small change in bn_mul_part_recursive() and the addition
3213 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3214 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3215 bn_sub_words() and bn_add_words() except they take arrays with
3219 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3221 *) Cleanse PEM buffers before freeing them since they may contain
3223 [Benjamin Bennett <ben@psc.edu>]
3225 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3226 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3227 authentication-only ciphersuites.
3230 *) Since AES128 and AES256 share a single mask bit in the logic of
3231 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3232 kludge to work properly if AES128 is available and AES256 isn't.
3235 *) Expand security boundary to match 1.1.1 module.
3238 *) Remove redundant features: hash file source, editing of test vectors
3239 modify fipsld to use external fips_premain.c signature.
3242 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3243 run algorithm test programs.
3246 *) Make algorithm test programs more tolerant of whitespace.
3249 *) Have SSL/TLS server implementation tolerate "mismatched" record
3250 protocol version while receiving ClientHello even if the
3251 ClientHello is fragmented. (The server can't insist on the
3252 particular protocol version it has chosen before the ServerHello
3253 message has informed the client about his choice.)
3256 *) Load error codes if they are not already present instead of using a
3257 static variable. This allows them to be cleanly unloaded and reloaded.
3260 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3262 *) Introduce limits to prevent malicious keys being able to
3263 cause a denial of service. (CVE-2006-2940)
3264 [Steve Henson, Bodo Moeller]
3266 *) Fix ASN.1 parsing of certain invalid structures that can result
3267 in a denial of service. (CVE-2006-2937) [Steve Henson]
3269 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3270 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3272 *) Fix SSL client code which could crash if connecting to a
3273 malicious SSLv2 server. (CVE-2006-4343)
3274 [Tavis Ormandy and Will Drewry, Google Security Team]
3276 *) Change ciphersuite string processing so that an explicit
3277 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3278 will no longer include "AES128-SHA"), and any other similar
3279 ciphersuite (same bitmap) from *other* protocol versions (so that
3280 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3281 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3282 changes from 0.9.8b and 0.9.8d.
3285 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3287 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3288 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3290 *) Change the Unix randomness entropy gathering to use poll() when
3291 possible instead of select(), since the latter has some
3292 undesirable limitations.
3293 [Darryl Miles via Richard Levitte and Bodo Moeller]
3295 *) Disable rogue ciphersuites:
3297 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3298 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3299 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3301 The latter two were purportedly from
3302 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3305 Also deactive the remaining ciphersuites from
3306 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3307 unofficial, and the ID has long expired.
3310 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3311 dual-core machines) and other potential thread-safety issues.
3314 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3316 *) Adapt fipsld and the build system to link against the validated FIPS
3317 module in FIPS mode.
3320 *) Fixes for VC++ 2005 build under Windows.
3323 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3324 from a Windows bash shell such as MSYS. It is autodetected from the
3325 "config" script when run from a VC++ environment. Modify standard VC++
3326 build to use fipscanister.o from the GNU make build.
3329 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3331 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3332 The value now differs depending on if you build for FIPS or not.
3333 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3334 safely run with a non-FIPSed libcrypto, as it may crash because of
3335 the difference induced by this change.
3338 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3340 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3341 (part of SSL_OP_ALL). This option used to disable the
3342 countermeasure against man-in-the-middle protocol-version
3343 rollback in the SSL 2.0 server implementation, which is a bad
3344 idea. (CVE-2005-2969)
3346 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3347 for Information Security, National Institute of Advanced Industrial
3348 Science and Technology [AIST], Japan)]
3350 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3351 mainly for FIPS compliance and not fully integrated at this stage.
3354 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3355 the exponentiation using a fixed-length exponent. (Otherwise,
3356 the information leaked through timing could expose the secret key
3357 after many signatures; cf. Bleichenbacher's attack on DSA with
3361 *) Make a new fixed-window mod_exp implementation the default for
3362 RSA, DSA, and DH private-key operations so that the sequence of
3363 squares and multiplies and the memory access pattern are
3364 independent of the particular secret key. This will mitigate
3365 cache-timing and potential related attacks.
3367 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3368 and this is automatically used by BN_mod_exp_mont() if the new flag
3369 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3370 will use this BN flag for private exponents unless the flag
3371 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3372 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3374 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3376 *) Change the client implementation for SSLv23_method() and
3377 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3378 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3379 (Previously, the SSL 2.0 backwards compatible Client Hello
3380 message format would be used even with SSL_OP_NO_SSLv2.)
3383 *) Add support for smime-type MIME parameter in S/MIME messages which some
3387 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3388 a threadsafe manner. Modify rsa code to use new function and add calls
3389 to dsa and dh code (which had race conditions before).
3392 *) Include the fixed error library code in the C error file definitions
3393 instead of fixing them up at runtime. This keeps the error code
3394 structures constant.
3397 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3399 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3402 *) Fixes for newer kerberos headers. NB: the casts are needed because
3403 the 'length' field is signed on one version and unsigned on another
3404 with no (?) obvious way to tell the difference, without these VC++
3405 complains. Also the "definition" of FAR (blank) is no longer included
3406 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3407 some needed definitions.
3410 *) Undo Cygwin change.
3413 *) Added support for proxy certificates according to RFC 3820.
3414 Because they may be a security thread to unaware applications,
3415 they must be explicitely allowed in run-time. See
3416 docs/HOWTO/proxy_certificates.txt for further information.
3419 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3421 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3422 server and client random values. Previously
3423 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3424 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3426 This change has negligible security impact because:
3428 1. Server and client random values still have 24 bytes of pseudo random
3431 2. Server and client random values are sent in the clear in the initial
3434 3. The master secret is derived using the premaster secret (48 bytes in
3435 size for static RSA ciphersuites) as well as client server and random
3438 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3441 [Stephen Henson, reported by UK NISCC]
3443 *) Use Windows randomness collection on Cygwin.
3446 *) Fix hang in EGD/PRNGD query when communication socket is closed
3447 prematurely by EGD/PRNGD.
3448 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3450 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3453 *) Back-port of selected performance improvements from development
3454 branch, as well as improved support for PowerPC platforms.
3457 *) Add lots of checks for memory allocation failure, error codes to indicate
3458 failure and freeing up memory if a failure occurs.
3459 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3461 *) Add new -passin argument to dgst.
3464 *) Perform some character comparisons of different types in X509_NAME_cmp:
3465 this is needed for some certificates that reencode DNs into UTF8Strings
3466 (in violation of RFC3280) and can't or wont issue name rollover
3470 *) Make an explicit check during certificate validation to see that
3471 the CA setting in each certificate on the chain is correct. As a
3472 side effect always do the following basic checks on extensions,
3473 not just when there's an associated purpose to the check:
3475 - if there is an unhandled critical extension (unless the user
3476 has chosen to ignore this fault)
3477 - if the path length has been exceeded (if one is set at all)
3478 - that certain extensions fit the associated purpose (if one has
3482 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3484 *) Avoid a race condition when CRLs are checked in a multi threaded
3485 environment. This would happen due to the reordering of the revoked
3486 entries during signature checking and serial number lookup. Now the
3487 encoding is cached and the serial number sort performed under a lock.
3488 Add new STACK function sk_is_sorted().
3491 *) Add Delta CRL to the extension code.
3494 *) Various fixes to s3_pkt.c so alerts are sent properly.
3495 [David Holmes <d.holmes@f5.com>]
3497 *) Reduce the chances of duplicate issuer name and serial numbers (in
3498 violation of RFC3280) using the OpenSSL certificate creation utilities.
3499 This is done by creating a random 64 bit value for the initial serial
3500 number when a serial number file is created or when a self signed
3501 certificate is created using 'openssl req -x509'. The initial serial
3502 number file is created using 'openssl x509 -next_serial' in CA.pl
3503 rather than being initialized to 1.
3506 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3508 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3509 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3510 [Joe Orton, Steve Henson]
3512 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3514 [Joe Orton, Steve Henson]
3516 *) Make it possible to have multiple active certificates with the same
3517 subject in the CA index file. This is done only if the keyword
3518 'unique_subject' is set to 'no' in the main CA section (default
3519 if 'CA_default') of the configuration file. The value is saved
3520 with the database itself in a separate index attribute file,
3521 named like the index file with '.attr' appended to the name.
3524 *) X509 verify fixes. Disable broken certificate workarounds when
3525 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3526 keyUsage extension present. Don't accept CRLs with unhandled critical
3527 extensions: since verify currently doesn't process CRL extensions this
3528 rejects a CRL with *any* critical extensions. Add new verify error codes
3532 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3533 A clarification of RFC2560 will require the use of OCTET STRINGs and
3534 some implementations cannot handle the current raw format. Since OpenSSL
3535 copies and compares OCSP nonces as opaque blobs without any attempt at
3536 parsing them this should not create any compatibility issues.
3539 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3540 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3541 this HMAC (and other) operations are several times slower than OpenSSL
3545 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3546 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3548 *) Use the correct content when signing type "other".
3551 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3553 *) Fix various bugs revealed by running the NISCC test suite:
3555 Stop out of bounds reads in the ASN1 code when presented with
3556 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3558 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3560 If verify callback ignores invalid public key errors don't try to check
3561 certificate signature with the NULL public key.
3565 *) New -ignore_err option in ocsp application to stop the server
3566 exiting on the first error in a request.
3569 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3570 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3574 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3575 extra data after the compression methods not only for TLS 1.0
3576 but also for SSL 3.0 (as required by the specification).
3577 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3579 *) Change X509_certificate_type() to mark the key as exported/exportable
3580 when it's 512 *bits* long, not 512 bytes.
3583 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3584 blocks during encryption.
3587 *) Various fixes to base64 BIO and non blocking I/O. On write
3588 flushes were not handled properly if the BIO retried. On read
3589 data was not being buffered properly and had various logic bugs.
3590 This also affects blocking I/O when the data being decoded is a
3594 *) Various S/MIME bugfixes and compatibility changes:
3595 output correct application/pkcs7 MIME type if
3596 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3597 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3598 of files as .eml work). Correctly handle very long lines in MIME
3602 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3604 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3605 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3606 a protocol version number mismatch like a decryption error
3607 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3610 *) Turn on RSA blinding by default in the default implementation
3611 to avoid a timing attack. Applications that don't want it can call
3612 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3613 They would be ill-advised to do so in most cases.
3614 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3616 *) Change RSA blinding code so that it works when the PRNG is not
3617 seeded (in this case, the secret RSA exponent is abused as
3618 an unpredictable seed -- if it is not unpredictable, there
3619 is no point in blinding anyway). Make RSA blinding thread-safe
3620 by remembering the creator's thread ID in rsa->blinding and
3621 having all other threads use local one-time blinding factors
3622 (this requires more computation than sharing rsa->blinding, but
3623 avoids excessive locking; and if an RSA object is not shared
3624 between threads, blinding will still be very fast).
3627 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3628 ENGINE as defaults for all supported algorithms irrespective of
3629 the 'flags' parameter. 'flags' is now honoured, so applications
3630 should make sure they are passing it correctly.
3633 *) Target "mingw" now allows native Windows code to be generated in
3634 the Cygwin environment as well as with the MinGW compiler.
3637 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3639 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3640 via timing by performing a MAC computation even if incorrrect
3641 block cipher padding has been found. This is a countermeasure
3642 against active attacks where the attacker has to distinguish
3643 between bad padding and a MAC verification error. (CVE-2003-0078)
3645 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3646 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3647 Martin Vuagnoux (EPFL, Ilion)]
3649 *) Make the no-err option work as intended. The intention with no-err
3650 is not to have the whole error stack handling routines removed from
3651 libcrypto, it's only intended to remove all the function name and
3652 reason texts, thereby removing some of the footprint that may not
3653 be interesting if those errors aren't displayed anyway.
3655 NOTE: it's still possible for any application or module to have it's
3656 own set of error texts inserted. The routines are there, just not
3657 used by default when no-err is given.
3660 *) Add support for FreeBSD on IA64.
3661 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3663 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3664 Kerberos function mit_des_cbc_cksum(). Before this change,
3665 the value returned by DES_cbc_cksum() was like the one from
3666 mit_des_cbc_cksum(), except the bytes were swapped.
3667 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3669 *) Allow an application to disable the automatic SSL chain building.
3670 Before this a rather primitive chain build was always performed in
3671 ssl3_output_cert_chain(): an application had no way to send the
3672 correct chain if the automatic operation produced an incorrect result.
3674 Now the chain builder is disabled if either:
3676 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3678 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3680 The reasoning behind this is that an application would not want the
3681 auto chain building to take place if extra chain certificates are
3682 present and it might also want a means of sending no additional
3683 certificates (for example the chain has two certificates and the
3687 *) Add the possibility to build without the ENGINE framework.
3688 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3690 *) Under Win32 gmtime() can return NULL: check return value in
3691 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3694 *) DSA routines: under certain error conditions uninitialized BN objects
3695 could be freed. Solution: make sure initialization is performed early
3696 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3697 Nils Larsch <nla@trustcenter.de> via PR#459)
3700 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3701 checked on reconnect on the client side, therefore session resumption
3702 could still fail with a "ssl session id is different" error. This
3703 behaviour is masked when SSL_OP_ALL is used due to
3704 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3705 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3706 followup to PR #377.
3709 *) IA-32 assembler support enhancements: unified ELF targets, support
3710 for SCO/Caldera platforms, fix for Cygwin shared build.
3713 *) Add support for FreeBSD on sparc64. As a consequence, support for
3714 FreeBSD on non-x86 processors is separate from x86 processors on
3715 the config script, much like the NetBSD support.
3716 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3718 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3720 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3723 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3724 code (06) was taken as the first octet of the session ID and the last
3725 octet was ignored consequently. As a result SSLv2 client side session
3726 caching could not have worked due to the session ID mismatch between
3728 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3732 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3733 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3737 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3738 seems that in spite of existing for more than a year, many application
3739 author have done nothing to provide the necessary callbacks, which
3740 means that this particular engine will not work properly anywhere.
3741 This is a very unfortunate situation which forces us, in the name
3742 of usability, to give the hw_ncipher.c a static lock, which is part
3744 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3745 appear in 0.9.8 or later. We EXPECT application authors to have
3746 dealt properly with this when 0.9.8 is released (unless we actually
3747 make such changes in the libcrypto locking code that changes will
3748 have to be made anyway).
3751 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3752 octets have been read, EOF or an error occurs. Without this change
3753 some truncated ASN1 structures will not produce an error.
3756 *) Disable Heimdal support, since it hasn't been fully implemented.
3757 Still give the possibility to force the use of Heimdal, but with
3758 warnings and a request that patches get sent to openssl-dev.
3761 *) Add the VC-CE target, introduce the WINCE sysname, and add
3762 INSTALL.WCE and appropriate conditionals to make it build.
3763 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3765 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3766 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3767 edit numbers of the version.
3768 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3770 *) Introduce safe string copy and catenation functions
3771 (BUF_strlcpy() and BUF_strlcat()).
3772 [Ben Laurie (CHATS) and Richard Levitte]
3774 *) Avoid using fixed-size buffers for one-line DNs.
3775 [Ben Laurie (CHATS)]
3777 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3778 resizing buffers containing secrets, and use where appropriate.
3779 [Ben Laurie (CHATS)]
3781 *) Avoid using fixed size buffers for configuration file location.
3782 [Ben Laurie (CHATS)]
3784 *) Avoid filename truncation for various CA files.
3785 [Ben Laurie (CHATS)]
3787 *) Use sizeof in preference to magic numbers.
3788 [Ben Laurie (CHATS)]
3790 *) Avoid filename truncation in cert requests.
3791 [Ben Laurie (CHATS)]
3793 *) Add assertions to check for (supposedly impossible) buffer
3795 [Ben Laurie (CHATS)]
3797 *) Don't cache truncated DNS entries in the local cache (this could
3798 potentially lead to a spoofing attack).
3799 [Ben Laurie (CHATS)]
3801 *) Fix various buffers to be large enough for hex/decimal
3802 representations in a platform independent manner.
3803 [Ben Laurie (CHATS)]
3805 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3806 resizing buffers containing secrets, and use where appropriate.
3807 [Ben Laurie (CHATS)]
3809 *) Add BIO_indent() to avoid much slightly worrying code to do
3811 [Ben Laurie (CHATS)]
3813 *) Convert sprintf()/BIO_puts() to BIO_printf().
3814 [Ben Laurie (CHATS)]
3816 *) buffer_gets() could terminate with the buffer only half
3818 [Ben Laurie (CHATS)]
3820 *) Add assertions to prevent user-supplied crypto functions from
3821 overflowing internal buffers by having large block sizes, etc.
3822 [Ben Laurie (CHATS)]
3824 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3826 [Ben Laurie (CHATS)]
3828 *) Eliminate unused copy of key in RC4.
3829 [Ben Laurie (CHATS)]
3831 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3832 [Ben Laurie (CHATS)]
3834 *) Fix off-by-one error in EGD path.
3835 [Ben Laurie (CHATS)]
3837 *) If RANDFILE path is too long, ignore instead of truncating.
3838 [Ben Laurie (CHATS)]
3840 *) Eliminate unused and incorrectly sized X.509 structure
3842 [Ben Laurie (CHATS)]
3844 *) Eliminate unused and dangerous function knumber().
3845 [Ben Laurie (CHATS)]
3847 *) Eliminate unused and dangerous structure, KSSL_ERR.
3848 [Ben Laurie (CHATS)]
3850 *) Protect against overlong session ID context length in an encoded
3851 session object. Since these are local, this does not appear to be
3853 [Ben Laurie (CHATS)]
3855 *) Change from security patch (see 0.9.6e below) that did not affect
3856 the 0.9.6 release series:
3858 Remote buffer overflow in SSL3 protocol - an attacker could
3859 supply an oversized master key in Kerberos-enabled versions.
3861 [Ben Laurie (CHATS)]
3863 *) Change the SSL kerb5 codes to match RFC 2712.
3866 *) Make -nameopt work fully for req and add -reqopt switch.
3867 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3869 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3870 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3872 *) Make sure tests can be performed even if the corresponding algorithms
3873 have been removed entirely. This was also the last step to make
3874 OpenSSL compilable with DJGPP under all reasonable conditions.
3875 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3877 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3878 to allow version independent disabling of normally unselected ciphers,
3879 which may be activated as a side-effect of selecting a single cipher.
3881 (E.g., cipher list string "RSA" enables ciphersuites that are left
3882 out of "ALL" because they do not provide symmetric encryption.
3883 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3884 [Lutz Jaenicke, Bodo Moeller]
3886 *) Add appropriate support for separate platform-dependent build
3887 directories. The recommended way to make a platform-dependent
3888 build directory is the following (tested on Linux), maybe with
3891 # Place yourself outside of the OpenSSL source tree. In
3892 # this example, the environment variable OPENSSL_SOURCE
3893 # is assumed to contain the absolute OpenSSL source directory.
3894 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3895 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3896 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3897 mkdir -p `dirname $F`
3898 ln -s $OPENSSL_SOURCE/$F $F
3901 To be absolutely sure not to disturb the source tree, a "make clean"
3902 is a good thing. If it isn't successfull, don't worry about it,
3903 it probably means the source directory is very clean.
3906 *) Make sure any ENGINE control commands make local copies of string
3907 pointers passed to them whenever necessary. Otherwise it is possible
3908 the caller may have overwritten (or deallocated) the original string
3909 data when a later ENGINE operation tries to use the stored values.
3910 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3912 *) Improve diagnostics in file reading and command-line digests.
3913 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3915 *) Add AES modes CFB and OFB to the object database. Correct an
3916 error in AES-CFB decryption.
3919 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3920 allows existing EVP_CIPHER_CTX structures to be reused after
3921 calling EVP_*Final(). This behaviour is used by encryption
3922 BIOs and some applications. This has the side effect that
3923 applications must explicitly clean up cipher contexts with
3924 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3927 *) Check the values of dna and dnb in bn_mul_recursive before calling
3928 bn_mul_comba (a non zero value means the a or b arrays do not contain
3929 n2 elements) and fallback to bn_mul_normal if either is not zero.
3932 *) Fix escaping of non-ASCII characters when using the -subj option
3933 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3936 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3937 form for "surname", serialNumber has no short form.
3938 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3939 therefore remove "mail" short name for "internet 7".
3940 The OID for unique identifiers in X509 certificates is
3941 x500UniqueIdentifier, not uniqueIdentifier.
3942 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3945 *) Add an "init" command to the ENGINE config module and auto initialize
3946 ENGINEs. Without any "init" command the ENGINE will be initialized
3947 after all ctrl commands have been executed on it. If init=1 the
3948 ENGINE is initailized at that point (ctrls before that point are run
3949 on the uninitialized ENGINE and after on the initialized one). If
3950 init=0 then the ENGINE will not be iniatialized at all.
3953 *) Fix the 'app_verify_callback' interface so that the user-defined
3954 argument is actually passed to the callback: In the
3955 SSL_CTX_set_cert_verify_callback() prototype, the callback
3956 declaration has been changed from
3959 int (*cb)(X509_STORE_CTX *,void *);
3960 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3961 i=s->ctx->app_verify_callback(&ctx)
3962 has been changed into
3963 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3965 To update applications using SSL_CTX_set_cert_verify_callback(),
3966 a dummy argument can be added to their callback functions.
3967 [D. K. Smetters <smetters@parc.xerox.com>]
3969 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3970 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3972 *) Add and OPENSSL_LOAD_CONF define which will cause
3973 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3974 This allows older applications to transparently support certain
3975 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3976 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3977 load the config file and OPENSSL_add_all_algorithms_conf() which will
3978 always load it have also been added.
3981 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3982 Adjust NIDs and EVP layer.
3983 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3985 *) Config modules support in openssl utility.
3987 Most commands now load modules from the config file,
3988 though in a few (such as version) this isn't done
3989 because it couldn't be used for anything.
3991 In the case of ca and req the config file used is
3992 the same as the utility itself: that is the -config
3993 command line option can be used to specify an
3997 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3998 use "openssl_conf" if filename is NULL use default openssl config file.
4001 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4002 config section name. Add a new flag to tolerate a missing config file
4003 and move code to CONF_modules_load_file().
4006 *) Support for crypto accelerator cards from Accelerated Encryption
4007 Processing, www.aep.ie. (Use engine 'aep')
4008 The support was copied from 0.9.6c [engine] and adapted/corrected
4009 to work with the new engine framework.
4010 [AEP Inc. and Richard Levitte]
4012 *) Support for SureWare crypto accelerator cards from Baltimore
4013 Technologies. (Use engine 'sureware')
4014 The support was copied from 0.9.6c [engine] and adapted
4015 to work with the new engine framework.
4018 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4019 make the newer ENGINE framework commands for the CHIL engine work.
4020 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4022 *) Make it possible to produce shared libraries on ReliantUNIX.
4023 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4025 *) Add the configuration target debug-linux-ppro.
4026 Make 'openssl rsa' use the general key loading routines
4027 implemented in apps.c, and make those routines able to
4028 handle the key format FORMAT_NETSCAPE and the variant
4030 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4032 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4033 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4035 *) Add -keyform to rsautl, and document -engine.
4036 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4038 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4039 BIO_R_NO_SUCH_FILE error code rather than the generic
4040 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4043 *) Add new functions
4045 ERR_peek_last_error_line
4046 ERR_peek_last_error_line_data.
4047 These are similar to
4050 ERR_peek_error_line_data,
4051 but report on the latest error recorded rather than the first one
4052 still in the error queue.
4053 [Ben Laurie, Bodo Moeller]
4055 *) default_algorithms option in ENGINE config module. This allows things
4057 default_algorithms = ALL
4058 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4061 *) Prelminary ENGINE config module.
4064 *) New experimental application configuration code.
4067 *) Change the AES code to follow the same name structure as all other
4068 symmetric ciphers, and behave the same way. Move everything to
4069 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4070 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4072 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4073 [Ben Laurie and Theo de Raadt]
4075 *) Add option to output public keys in req command.
4076 [Massimiliano Pala madwolf@openca.org]
4078 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4079 (up to about 10% better than before for P-192 and P-224).
4082 *) New functions/macros
4084 SSL_CTX_set_msg_callback(ctx, cb)
4085 SSL_CTX_set_msg_callback_arg(ctx, arg)
4086 SSL_set_msg_callback(ssl, cb)
4087 SSL_set_msg_callback_arg(ssl, arg)
4089 to request calling a callback function
4091 void cb(int write_p, int version, int content_type,
4092 const void *buf, size_t len, SSL *ssl, void *arg)
4094 whenever a protocol message has been completely received
4095 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4096 protocol version according to which the SSL library interprets
4097 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4098 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4099 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4100 specification (change_cipher_spec(20), alert(21), handshake(22)).
4101 'buf' and 'len' point to the actual message, 'ssl' to the
4102 SSL object, and 'arg' is the application-defined value set by
4103 SSL[_CTX]_set_msg_callback_arg().
4105 'openssl s_client' and 'openssl s_server' have new '-msg' options
4106 to enable a callback that displays all protocol messages.
4109 *) Change the shared library support so shared libraries are built as
4110 soon as the corresponding static library is finished, and thereby get
4111 openssl and the test programs linked against the shared library.
4112 This still only happens when the keyword "shard" has been given to
4113 the configuration scripts.
4115 NOTE: shared library support is still an experimental thing, and
4116 backward binary compatibility is still not guaranteed.
4117 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4119 *) Add support for Subject Information Access extension.
4120 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4122 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4123 additional bytes when new memory had to be allocated, not just
4124 when reusing an existing buffer.
4127 *) New command line and configuration option 'utf8' for the req command.
4128 This allows field values to be specified as UTF8 strings.
4131 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4132 runs for the former and machine-readable output for the latter.
4135 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4136 of the e-mail address in the DN (i.e., it will go into a certificate
4137 extension only). The new configuration file option 'email_in_dn = no'
4138 has the same effect.
4139 [Massimiliano Pala madwolf@openca.org]
4141 *) Change all functions with names starting with des_ to be starting
4142 with DES_ instead. Add wrappers that are compatible with libdes,
4143 but are named _ossl_old_des_*. Finally, add macros that map the
4144 des_* symbols to the corresponding _ossl_old_des_* if libdes
4145 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4146 desired, the des_* symbols will be mapped to DES_*, with one
4149 Since we provide two compatibility mappings, the user needs to
4150 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4151 compatibility is desired. The default (i.e., when that macro
4152 isn't defined) is OpenSSL 0.9.6c compatibility.
4154 There are also macros that enable and disable the support of old
4155 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4156 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4157 are defined, the default will apply: to support the old des routines.
4159 In either case, one must include openssl/des.h to get the correct
4160 definitions. Do not try to just include openssl/des_old.h, that
4163 NOTE: This is a major break of an old API into a new one. Software
4164 authors are encouraged to switch to the DES_ style functions. Some
4165 time in the future, des_old.h and the libdes compatibility functions
4166 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4167 default), and then completely removed.
4170 *) Test for certificates which contain unsupported critical extensions.
4171 If such a certificate is found during a verify operation it is
4172 rejected by default: this behaviour can be overridden by either
4173 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4174 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4175 X509_supported_extension() has also been added which returns 1 if a
4176 particular extension is supported.
4179 *) Modify the behaviour of EVP cipher functions in similar way to digests
4180 to retain compatibility with existing code.
4183 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4184 compatibility with existing code. In particular the 'ctx' parameter does
4185 not have to be to be initialized before the call to EVP_DigestInit() and
4186 it is tidied up after a call to EVP_DigestFinal(). New function
4187 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4188 EVP_MD_CTX_copy() changed to not require the destination to be
4189 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4190 requires the destination to be valid.
4192 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4193 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4196 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4197 so that complete 'Handshake' protocol structures are kept in memory
4198 instead of overwriting 'msg_type' and 'length' with 'body' data.
4201 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4202 [Massimo Santin via Richard Levitte]
4204 *) Major restructuring to the underlying ENGINE code. This includes
4205 reduction of linker bloat, separation of pure "ENGINE" manipulation
4206 (initialisation, etc) from functionality dealing with implementations
4207 of specific crypto iterfaces. This change also introduces integrated
4208 support for symmetric ciphers and digest implementations - so ENGINEs
4209 can now accelerate these by providing EVP_CIPHER and EVP_MD
4210 implementations of their own. This is detailed in crypto/engine/README
4211 as it couldn't be adequately described here. However, there are a few
4212 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4213 were changed in the original introduction of ENGINE code have now
4214 reverted back - the hooking from this code to ENGINE is now a good
4215 deal more passive and at run-time, operations deal directly with
4216 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4217 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4218 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4219 they were not being used by the framework as there is no concept of a
4220 BIGNUM_METHOD and they could not be generalised to the new
4221 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4222 ENGINE_cpy() has been removed as it cannot be consistently defined in
4226 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4229 *) Change mkdef.pl to sort symbols that get the same entry number,
4230 and make sure the automatically generated functions ERR_load_*
4231 become part of libeay.num as well.
4234 *) New function SSL_renegotiate_pending(). This returns true once
4235 renegotiation has been requested (either SSL_renegotiate() call
4236 or HelloRequest/ClientHello receveived from the peer) and becomes
4237 false once a handshake has been completed.
4238 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4239 sends a HelloRequest, but does not ensure that a handshake takes
4240 place. SSL_renegotiate_pending() is useful for checking if the
4241 client has followed the request.)
4244 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4245 By default, clients may request session resumption even during
4246 renegotiation (if session ID contexts permit); with this option,
4247 session resumption is possible only in the first handshake.
4249 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4250 more bits available for options that should not be part of
4251 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4254 *) Add some demos for certificate and certificate request creation.
4257 *) Make maximum certificate chain size accepted from the peer application
4258 settable (SSL*_get/set_max_cert_list()), as proposed by
4259 "Douglas E. Engert" <deengert@anl.gov>.
4262 *) Add support for shared libraries for Unixware-7
4263 (Boyd Lynn Gerber <gerberb@zenez.com>).
4266 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4267 be done prior to destruction. Use this to unload error strings from
4268 ENGINEs that load their own error strings. NB: This adds two new API
4269 functions to "get" and "set" this destroy handler in an ENGINE.
4272 *) Alter all existing ENGINE implementations (except "openssl" and
4273 "openbsd") to dynamically instantiate their own error strings. This
4274 makes them more flexible to be built both as statically-linked ENGINEs
4275 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4276 Also, add stub code to each that makes building them as self-contained
4277 shared-libraries easier (see README.ENGINE).
4280 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4281 implementations into applications that are completely implemented in
4282 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4283 commands that can be used to configure what shared-library to load and
4284 to control aspects of the way it is handled. Also, made an update to
4285 the README.ENGINE file that brings its information up-to-date and
4286 provides some information and instructions on the "dynamic" ENGINE
4287 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4290 *) Make it possible to unload ranges of ERR strings with a new
4291 "ERR_unload_strings" function.
4294 *) Add a copy() function to EVP_MD.
4297 *) Make EVP_MD routines take a context pointer instead of just the
4298 md_data void pointer.
4301 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4302 that the digest can only process a single chunk of data
4303 (typically because it is provided by a piece of
4304 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4305 is only going to provide a single chunk of data, and hence the
4306 framework needn't accumulate the data for oneshot drivers.
4309 *) As with "ERR", make it possible to replace the underlying "ex_data"
4310 functions. This change also alters the storage and management of global
4311 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4312 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4313 index counters. The API functions that use this state have been changed
4314 to take a "class_index" rather than pointers to the class's local STACK
4315 and counter, and there is now an API function to dynamically create new
4316 classes. This centralisation allows us to (a) plug a lot of the
4317 thread-safety problems that existed, and (b) makes it possible to clean
4318 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4319 such data would previously have always leaked in application code and
4320 workarounds were in place to make the memory debugging turn a blind eye
4321 to it. Application code that doesn't use this new function will still
4322 leak as before, but their memory debugging output will announce it now
4323 rather than letting it slide.
4325 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4326 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4327 has a return value to indicate success or failure.
4330 *) Make it possible to replace the underlying "ERR" functions such that the
4331 global state (2 LHASH tables and 2 locks) is only used by the "default"
4332 implementation. This change also adds two functions to "get" and "set"
4333 the implementation prior to it being automatically set the first time
4334 any other ERR function takes place. Ie. an application can call "get",
4335 pass the return value to a module it has just loaded, and that module
4336 can call its own "set" function using that value. This means the
4337 module's "ERR" operations will use (and modify) the error state in the
4338 application and not in its own statically linked copy of OpenSSL code.
4341 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4342 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4343 the operation, and provides a more encapsulated way for external code
4344 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4345 to use these functions rather than manually incrementing the counts.
4347 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4350 *) Add EVP test program.
4353 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4356 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4357 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4358 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4359 These allow a CRL to be built without having to access X509_CRL fields
4360 directly. Modify 'ca' application to use new functions.
4363 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4364 bug workarounds. Rollback attack detection is a security feature.
4365 The problem will only arise on OpenSSL servers when TLSv1 is not
4366 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4367 Software authors not wanting to support TLSv1 will have special reasons
4368 for their choice and can explicitly enable this option.
4369 [Bodo Moeller, Lutz Jaenicke]
4371 *) Rationalise EVP so it can be extended: don't include a union of
4372 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4373 (similar to those existing for EVP_CIPHER_CTX).
4378 EVP_MD_CTX_init(&md); /* new function call */
4379 EVP_DigestInit(&md, EVP_sha1());
4380 EVP_DigestUpdate(&md, in, len);
4381 EVP_DigestFinal(&md, out, NULL);
4382 EVP_MD_CTX_cleanup(&md); /* new function call */
4386 *) Make DES key schedule conform to the usual scheme, as well as
4387 correcting its structure. This means that calls to DES functions
4388 now have to pass a pointer to a des_key_schedule instead of a
4389 plain des_key_schedule (which was actually always a pointer
4392 des_key_schedule ks;
4394 des_set_key_checked(..., &ks);
4395 des_ncbc_encrypt(..., &ks, ...);
4397 (Note that a later change renames 'des_...' into 'DES_...'.)
4400 *) Initial reduction of linker bloat: the use of some functions, such as
4401 PEM causes large amounts of unused functions to be linked in due to
4402 poor organisation. For example pem_all.c contains every PEM function
4403 which has a knock on effect of linking in large amounts of (unused)
4404 ASN1 code. Grouping together similar functions and splitting unrelated
4405 functions prevents this.
4408 *) Cleanup of EVP macros.
4411 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4412 correct _ecb suffix.
4415 *) Add initial OCSP responder support to ocsp application. The
4416 revocation information is handled using the text based index
4417 use by the ca application. The responder can either handle
4418 requests generated internally, supplied in files (for example
4419 via a CGI script) or using an internal minimal server.
4422 *) Add configuration choices to get zlib compression for TLS.
4425 *) Changes to Kerberos SSL for RFC 2712 compliance:
4426 1. Implemented real KerberosWrapper, instead of just using
4427 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4428 2. Implemented optional authenticator field of KerberosWrapper.
4430 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4431 and authenticator structs; see crypto/krb5/.
4433 Generalized Kerberos calls to support multiple Kerberos libraries.
4434 [Vern Staats <staatsvr@asc.hpc.mil>,
4435 Jeffrey Altman <jaltman@columbia.edu>
4436 via Richard Levitte]
4438 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4439 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4440 values for each of the key sizes rather than having just
4441 parameters (and 'speed' generating keys each time).
4444 *) Speed up EVP routines.
4447 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4448 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4449 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4450 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4452 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4453 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4454 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4457 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4459 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4462 *) Added the OS2-EMX target.
4463 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4465 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4466 to support NCONF routines in extension code. New function CONF_set_nconf()
4467 to allow functions which take an NCONF to also handle the old LHASH
4468 structure: this means that the old CONF compatible routines can be
4469 retained (in particular wrt extensions) without having to duplicate the
4470 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4473 *) Enhance the general user interface with mechanisms for inner control
4474 and with possibilities to have yes/no kind of prompts.
4477 *) Change all calls to low level digest routines in the library and
4478 applications to use EVP. Add missing calls to HMAC_cleanup() and
4479 don't assume HMAC_CTX can be copied using memcpy().
4480 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4482 *) Add the possibility to control engines through control names but with
4483 arbitrary arguments instead of just a string.
4484 Change the key loaders to take a UI_METHOD instead of a callback
4485 function pointer. NOTE: this breaks binary compatibility with earlier
4486 versions of OpenSSL [engine].
4487 Adapt the nCipher code for these new conditions and add a card insertion
4491 *) Enhance the general user interface with mechanisms to better support
4492 dialog box interfaces, application-defined prompts, the possibility
4493 to use defaults (for example default passwords from somewhere else)
4494 and interrupts/cancellations.
4497 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4498 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4501 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4502 tidy up some unnecessarily weird code in 'sk_new()').
4503 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4505 *) Change the key loading routines for ENGINEs to use the same kind
4506 callback (pem_password_cb) as all other routines that need this
4510 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4511 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4512 than this minimum value is recommended.
4515 *) New random seeder for OpenVMS, using the system process statistics
4516 that are easily reachable.
4519 *) Windows apparently can't transparently handle global
4520 variables defined in DLLs. Initialisations such as:
4522 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4524 wont compile. This is used by the any applications that need to
4525 declare their own ASN1 modules. This was fixed by adding the option
4526 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4527 needed for static libraries under Win32.
4530 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4531 setting of purpose and trust fields. New X509_STORE trust and
4532 purpose functions and tidy up setting in other SSL functions.
4535 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4536 structure. These are inherited by X509_STORE_CTX when it is
4537 initialised. This allows various defaults to be set in the
4538 X509_STORE structure (such as flags for CRL checking and custom
4539 purpose or trust settings) for functions which only use X509_STORE_CTX
4540 internally such as S/MIME.
4542 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4543 trust settings if they are not set in X509_STORE. This allows X509_STORE
4544 purposes and trust (in S/MIME for example) to override any set by default.
4546 Add command line options for CRL checking to smime, s_client and s_server
4550 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4551 are set then the CRL is looked up in the X509_STORE structure and
4552 its validity and signature checked, then if the certificate is found
4553 in the CRL the verify fails with a revoked error.
4555 Various new CRL related callbacks added to X509_STORE_CTX structure.
4557 Command line options added to 'verify' application to support this.
4559 This needs some additional work, such as being able to handle multiple
4560 CRLs with different times, extension based lookup (rather than just
4561 by subject name) and ultimately more complete V2 CRL extension
4565 *) Add a general user interface API (crypto/ui/). This is designed
4566 to replace things like des_read_password and friends (backward
4567 compatibility functions using this new API are provided).
4568 The purpose is to remove prompting functions from the DES code
4569 section as well as provide for prompting through dialog boxes in
4570 a window system and the like.
4573 *) Add "ex_data" support to ENGINE so implementations can add state at a
4574 per-structure level rather than having to store it globally.
4577 *) Make it possible for ENGINE structures to be copied when retrieved by
4578 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4579 This causes the "original" ENGINE structure to act like a template,
4580 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4581 operational state can be localised to each ENGINE structure, despite the
4582 fact they all share the same "methods". New ENGINE structures returned in
4583 this case have no functional references and the return value is the single
4584 structural reference. This matches the single structural reference returned
4585 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4589 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4590 needs to match any other type at all we need to manually clear the
4594 *) Changes to the "openssl engine" utility to include;
4595 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4596 about an ENGINE's available control commands.
4597 - executing control commands from command line arguments using the
4598 '-pre' and '-post' switches. '-post' is only used if '-t' is
4599 specified and the ENGINE is successfully initialised. The syntax for
4600 the individual commands are colon-separated, for example;
4601 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4604 *) New dynamic control command support for ENGINEs. ENGINEs can now
4605 declare their own commands (numbers), names (strings), descriptions,
4606 and input types for run-time discovery by calling applications. A
4607 subset of these commands are implicitly classed as "executable"
4608 depending on their input type, and only these can be invoked through
4609 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4610 can be based on user input, config files, etc). The distinction is
4611 that "executable" commands cannot return anything other than a boolean
4612 result and can only support numeric or string input, whereas some
4613 discoverable commands may only be for direct use through
4614 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4615 pointers, or other custom uses. The "executable" commands are to
4616 support parameterisations of ENGINE behaviour that can be
4617 unambiguously defined by ENGINEs and used consistently across any
4618 OpenSSL-based application. Commands have been added to all the
4619 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4620 control over shared-library paths without source code alterations.
4623 *) Changed all ENGINE implementations to dynamically allocate their
4624 ENGINEs rather than declaring them statically. Apart from this being
4625 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4626 this also allows the implementations to compile without using the
4627 internal engine_int.h header.
4630 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4631 'const' value. Any code that should be able to modify a RAND_METHOD
4632 should already have non-const pointers to it (ie. they should only
4633 modify their own ones).
4636 *) Made a variety of little tweaks to the ENGINE code.
4637 - "atalla" and "ubsec" string definitions were moved from header files
4638 to C code. "nuron" string definitions were placed in variables
4639 rather than hard-coded - allowing parameterisation of these values
4640 later on via ctrl() commands.
4641 - Removed unused "#if 0"'d code.
4642 - Fixed engine list iteration code so it uses ENGINE_free() to release
4643 structural references.
4644 - Constified the RAND_METHOD element of ENGINE structures.
4645 - Constified various get/set functions as appropriate and added
4646 missing functions (including a catch-all ENGINE_cpy that duplicates
4647 all ENGINE values onto a new ENGINE except reference counts/state).
4648 - Removed NULL parameter checks in get/set functions. Setting a method
4649 or function to NULL is a way of cancelling out a previously set
4650 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4651 and doesn't justify the extra error symbols and code.
4652 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4653 flags from engine_int.h to engine.h.
4654 - Changed prototypes for ENGINE handler functions (init(), finish(),
4655 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4658 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4659 to the algorithm using long division. The binary algorithm can be
4660 used only if the modulus is odd. On 32-bit systems, it is faster
4661 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4662 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4663 up to 450 bits. In 64-bit environments, the binary algorithm
4664 appears to be advantageous for much longer moduli; here we use it
4665 for moduli up to 2048 bits.
4668 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4669 could not support the combine flag in choice fields.
4672 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4673 extensions from a certificate request to the certificate.
4676 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4677 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4678 file: this allows the display of the certificate about to be
4679 signed to be customised, to allow certain fields to be included
4680 or excluded and extension details. The old system didn't display
4681 multicharacter strings properly, omitted fields not in the policy
4682 and couldn't display additional details such as extensions.
4685 *) Function EC_POINTs_mul for multiple scalar multiplication
4686 of an arbitrary number of elliptic curve points
4687 \sum scalars[i]*points[i],
4688 optionally including the generator defined for the EC_GROUP:
4689 scalar*generator + \sum scalars[i]*points[i].
4691 EC_POINT_mul is a simple wrapper function for the typical case
4692 that the point list has just one item (besides the optional
4696 *) First EC_METHODs for curves over GF(p):
4698 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4699 operations and provides various method functions that can also
4700 operate with faster implementations of modular arithmetic.
4702 EC_GFp_mont_method() reuses most functions that are part of
4703 EC_GFp_simple_method, but uses Montgomery arithmetic.
4705 [Bodo Moeller; point addition and point doubling
4706 implementation directly derived from source code provided by
4707 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4709 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4710 crypto/ec/ec_lib.c):
4712 Curves are EC_GROUP objects (with an optional group generator)
4713 based on EC_METHODs that are built into the library.
4715 Points are EC_POINT objects based on EC_GROUP objects.
4717 Most of the framework would be able to handle curves over arbitrary
4718 finite fields, but as there are no obvious types for fields other
4719 than GF(p), some functions are limited to that for now.
4722 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4723 that the file contains a complete HTTP response.
4726 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4727 change the def and num file printf format specifier from "%-40sXXX"
4728 to "%-39s XXX". The latter will always guarantee a space after the
4729 field while the former will cause them to run together if the field
4730 is 40 of more characters long.
4733 *) Constify the cipher and digest 'method' functions and structures
4734 and modify related functions to take constant EVP_MD and EVP_CIPHER
4738 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4739 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4742 *) Modify EVP_Digest*() routines so they now return values. Although the
4743 internal software routines can never fail additional hardware versions
4747 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4749 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4750 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4753 ERR_R_NESTED_ASN1_ERROR
4755 ERR_R_MISSING_ASN1_EOS
4756 were 4 .. 9, conflicting with
4757 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4759 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4760 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4762 Add new error code 'ERR_R_INTERNAL_ERROR'.
4765 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4769 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4770 sets the subject name for a new request or supersedes the
4771 subject name in a given request. Formats that can be parsed are
4772 'CN=Some Name, OU=myOU, C=IT'
4774 'CN=Some Name/OU=myOU/C=IT'.
4776 Add options '-batch' and '-verbose' to 'openssl req'.
4777 [Massimiliano Pala <madwolf@hackmasters.net>]
4779 *) Introduce the possibility to access global variables through
4780 functions on platform were that's the best way to handle exporting
4781 global variables in shared libraries. To enable this functionality,
4782 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4783 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4784 is normally done by Configure or something similar).
4786 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4787 in the source file (foo.c) like this:
4789 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4790 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4792 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4793 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4795 OPENSSL_DECLARE_GLOBAL(int,foo);
4796 #define foo OPENSSL_GLOBAL_REF(foo)
4797 OPENSSL_DECLARE_GLOBAL(double,bar);
4798 #define bar OPENSSL_GLOBAL_REF(bar)
4800 The #defines are very important, and therefore so is including the
4801 header file everywhere where the defined globals are used.
4803 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4804 of ASN.1 items, but that structure is a bit different.
4806 The largest change is in util/mkdef.pl which has been enhanced with
4807 better and easier to understand logic to choose which symbols should
4808 go into the Windows .def files as well as a number of fixes and code
4809 cleanup (among others, algorithm keywords are now sorted
4810 lexicographically to avoid constant rewrites).
4813 *) In BN_div() keep a copy of the sign of 'num' before writing the
4814 result to 'rm' because if rm==num the value will be overwritten
4815 and produce the wrong result if 'num' is negative: this caused
4816 problems with BN_mod() and BN_nnmod().
4819 *) Function OCSP_request_verify(). This checks the signature on an
4820 OCSP request and verifies the signer certificate. The signer
4821 certificate is just checked for a generic purpose and OCSP request
4825 *) Add OCSP_check_validity() function to check the validity of OCSP
4826 responses. OCSP responses are prepared in real time and may only
4827 be a few seconds old. Simply checking that the current time lies
4828 between thisUpdate and nextUpdate max reject otherwise valid responses
4829 caused by either OCSP responder or client clock inaccuracy. Instead
4830 we allow thisUpdate and nextUpdate to fall within a certain period of
4831 the current time. The age of the response can also optionally be
4832 checked. Two new options -validity_period and -status_age added to
4836 *) If signature or public key algorithm is unrecognized print out its
4837 OID rather that just UNKNOWN.
4840 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4841 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4842 ID to be generated from the issuer certificate alone which can then be
4843 passed to OCSP_id_issuer_cmp().
4846 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4847 ASN1 modules to export functions returning ASN1_ITEM pointers
4848 instead of the ASN1_ITEM structures themselves. This adds several
4849 new macros which allow the underlying ASN1 function/structure to
4850 be accessed transparently. As a result code should not use ASN1_ITEM
4851 references directly (such as &X509_it) but instead use the relevant
4852 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4853 use of the new ASN1 code on platforms where exporting structures
4854 is problematical (for example in shared libraries) but exporting
4855 functions returning pointers to structures is not.
4858 *) Add support for overriding the generation of SSL/TLS session IDs.
4859 These callbacks can be registered either in an SSL_CTX or per SSL.
4860 The purpose of this is to allow applications to control, if they wish,
4861 the arbitrary values chosen for use as session IDs, particularly as it
4862 can be useful for session caching in multiple-server environments. A
4863 command-line switch for testing this (and any client code that wishes
4864 to use such a feature) has been added to "s_server".
4865 [Geoff Thorpe, Lutz Jaenicke]
4867 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4868 of the form '#if defined(...) || defined(...) || ...' and
4869 '#if !defined(...) && !defined(...) && ...'. This also avoids
4870 the growing number of special cases it was previously handling.
4873 *) Make all configuration macros available for application by making
4874 sure they are available in opensslconf.h, by giving them names starting
4875 with "OPENSSL_" to avoid conflicts with other packages and by making
4876 sure e_os2.h will cover all platform-specific cases together with
4878 Additionally, it is now possible to define configuration/platform-
4879 specific names (called "system identities"). In the C code, these
4880 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4881 macro with the name beginning with "OPENSSL_SYS_", which is determined
4882 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4886 *) New option -set_serial to 'req' and 'x509' this allows the serial
4887 number to use to be specified on the command line. Previously self
4888 signed certificates were hard coded with serial number 0 and the
4889 CA options of 'x509' had to use a serial number in a file which was
4893 *) New options to 'ca' utility to support V2 CRL entry extensions.
4894 Currently CRL reason, invalidity date and hold instruction are
4895 supported. Add new CRL extensions to V3 code and some new objects.
4898 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4899 disable standard block padding (aka PKCS#5 padding) in the EVP
4900 API, which was previously mandatory. This means that the data is
4901 not padded in any way and so the total length much be a multiple
4902 of the block size, otherwise an error occurs.
4905 *) Initial (incomplete) OCSP SSL support.
4908 *) New function OCSP_parse_url(). This splits up a URL into its host,
4909 port and path components: primarily to parse OCSP URLs. New -url
4910 option to ocsp utility.
4913 *) New nonce behavior. The return value of OCSP_check_nonce() now
4914 reflects the various checks performed. Applications can decide
4915 whether to tolerate certain situations such as an absent nonce
4916 in a response when one was present in a request: the ocsp application
4917 just prints out a warning. New function OCSP_add1_basic_nonce()
4918 this is to allow responders to include a nonce in a response even if
4919 the request is nonce-less.
4922 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4923 skipped when using openssl x509 multiple times on a single input file,
4924 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4927 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4928 set string type: to handle setting ASN1_TIME structures. Fix ca
4929 utility to correctly initialize revocation date of CRLs.
4932 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4933 the clients preferred ciphersuites and rather use its own preferences.
4934 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4935 Internet Explorer by ensuring unchanged hash method during stepup.
4936 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4939 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4940 to aes and add a new 'exist' option to print out symbols that don't
4944 *) Additional options to ocsp utility to allow flags to be set and
4945 additional certificates supplied.
4948 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4949 OCSP client a number of certificate to only verify the response
4953 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4954 handle the new API. Currently only ECB, CBC modes supported. Add new
4957 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4958 Encryption Standard (AES) Ciphersuites for Transport Layer
4959 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4960 not enabled by default and were not part of the "ALL" ciphersuite
4961 alias because they were not yet official; they could be
4962 explicitly requested by specifying the "AESdraft" ciphersuite
4963 group alias. In the final release of OpenSSL 0.9.7, the group
4964 alias is called "AES" and is part of "ALL".)
4965 [Ben Laurie, Steve Henson, Bodo Moeller]
4967 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4968 request to response.
4971 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4972 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4973 extract information from a certificate request. OCSP_response_create()
4974 creates a response and optionally adds a basic response structure.
4975 OCSP_basic_add1_status() adds a complete single response to a basic
4976 response and returns the OCSP_SINGLERESP structure just added (to allow
4977 extensions to be included for example). OCSP_basic_add1_cert() adds a
4978 certificate to a basic response and OCSP_basic_sign() signs a basic
4979 response with various flags. New helper functions ASN1_TIME_check()
4980 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4981 (converts ASN1_TIME to GeneralizedTime).
4984 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4985 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4986 structure from a certificate. X509_pubkey_digest() digests the public_key
4987 contents: this is used in various key identifiers.
4990 *) Make sk_sort() tolerate a NULL argument.
4991 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4993 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4994 passed by the function are trusted implicitly. If any of them signed the
4995 response then it is assumed to be valid and is not verified.
4998 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4999 to data. This was previously part of the PKCS7 ASN1 code. This
5000 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5001 [Steve Henson, reported by Kenneth R. Robinette
5002 <support@securenetterm.com>]
5004 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5005 routines: without these tracing memory leaks is very painful.
5006 Fix leaks in PKCS12 and PKCS7 routines.
5009 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5010 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5011 effectively meant GeneralizedTime would never be used. Now it
5012 is initialised to -1 but X509_time_adj() now has to check the value
5013 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5014 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5015 [Steve Henson, reported by Kenneth R. Robinette
5016 <support@securenetterm.com>]
5018 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5019 result in a zero length in the ASN1_INTEGER structure which was
5020 not consistent with the structure when d2i_ASN1_INTEGER() was used
5021 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5022 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5023 where it did not print out a minus for negative ASN1_INTEGER.
5026 *) Add summary printout to ocsp utility. The various functions which
5027 convert status values to strings have been renamed to:
5028 OCSP_response_status_str(), OCSP_cert_status_str() and
5029 OCSP_crl_reason_str() and are no longer static. New options
5030 to verify nonce values and to disable verification. OCSP response
5031 printout format cleaned up.
5034 *) Add additional OCSP certificate checks. These are those specified
5035 in RFC2560. This consists of two separate checks: the CA of the
5036 certificate being checked must either be the OCSP signer certificate
5037 or the issuer of the OCSP signer certificate. In the latter case the
5038 OCSP signer certificate must contain the OCSP signing extended key
5039 usage. This check is performed by attempting to match the OCSP
5040 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5041 in the OCSP_CERTID structures of the response.
5044 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5045 and related routines. This uses the standard OpenSSL certificate
5046 verify routines to perform initial checks (just CA validity) and
5047 to obtain the certificate chain. Then additional checks will be
5048 performed on the chain. Currently the root CA is checked to see
5049 if it is explicitly trusted for OCSP signing. This is used to set
5050 a root CA as a global signing root: that is any certificate that
5051 chains to that CA is an acceptable OCSP signing certificate.
5054 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5055 extensions from a separate configuration file.
5056 As when reading extensions from the main configuration file,
5057 the '-extensions ...' option may be used for specifying the
5059 [Massimiliano Pala <madwolf@comune.modena.it>]
5061 *) New OCSP utility. Allows OCSP requests to be generated or
5062 read. The request can be sent to a responder and the output
5063 parsed, outputed or printed in text form. Not complete yet:
5064 still needs to check the OCSP response validity.
5067 *) New subcommands for 'openssl ca':
5068 'openssl ca -status <serial>' prints the status of the cert with
5069 the given serial number (according to the index file).
5070 'openssl ca -updatedb' updates the expiry status of certificates
5072 [Massimiliano Pala <madwolf@comune.modena.it>]
5074 *) New '-newreq-nodes' command option to CA.pl. This is like
5075 '-newreq', but calls 'openssl req' with the '-nodes' option
5076 so that the resulting key is not encrypted.
5077 [Damien Miller <djm@mindrot.org>]
5079 *) New configuration for the GNU Hurd.
5080 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5082 *) Initial code to implement OCSP basic response verify. This
5083 is currently incomplete. Currently just finds the signer's
5084 certificate and verifies the signature on the response.
5087 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5088 value of OPENSSLDIR. This is available via the new '-d' option
5089 to 'openssl version', and is also included in 'openssl version -a'.
5092 *) Allowing defining memory allocation callbacks that will be given
5093 file name and line number information in additional arguments
5094 (a const char* and an int). The basic functionality remains, as
5095 well as the original possibility to just replace malloc(),
5096 realloc() and free() by functions that do not know about these
5097 additional arguments. To register and find out the current
5098 settings for extended allocation functions, the following
5099 functions are provided:
5101 CRYPTO_set_mem_ex_functions
5102 CRYPTO_set_locked_mem_ex_functions
5103 CRYPTO_get_mem_ex_functions
5104 CRYPTO_get_locked_mem_ex_functions
5106 These work the same way as CRYPTO_set_mem_functions and friends.
5107 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5108 extended allocation function is enabled.
5109 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5110 a conventional allocation function is enabled.
5111 [Richard Levitte, Bodo Moeller]
5113 *) Finish off removing the remaining LHASH function pointer casts.
5114 There should no longer be any prototype-casting required when using
5115 the LHASH abstraction, and any casts that remain are "bugs". See
5116 the callback types and macros at the head of lhash.h for details
5117 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5120 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5121 If /dev/[u]random devices are not available or do not return enough
5122 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5124 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5125 /etc/entropy will be queried once each in this sequence, quering stops
5126 when enough entropy was collected without querying more sockets.
5129 *) Change the Unix RAND_poll() variant to be able to poll several
5130 random devices, as specified by DEVRANDOM, until a sufficient amount
5131 of data has been collected. We spend at most 10 ms on each file
5132 (select timeout) and read in non-blocking mode. DEVRANDOM now
5133 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5134 (previously it was just the string "/dev/urandom"), so on typical
5135 platforms the 10 ms delay will never occur.
5136 Also separate out the Unix variant to its own file, rand_unix.c.
5137 For VMS, there's a currently-empty rand_vms.c.
5140 *) Move OCSP client related routines to ocsp_cl.c. These
5141 provide utility functions which an application needing
5142 to issue a request to an OCSP responder and analyse the
5143 response will typically need: as opposed to those which an
5144 OCSP responder itself would need which will be added later.
5146 OCSP_request_sign() signs an OCSP request with an API similar
5147 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5148 response. OCSP_response_get1_basic() extracts basic response
5149 from response. OCSP_resp_find_status(): finds and extracts status
5150 information from an OCSP_CERTID structure (which will be created
5151 when the request structure is built). These are built from lower
5152 level functions which work on OCSP_SINGLERESP structures but
5153 wont normally be used unless the application wishes to examine
5154 extensions in the OCSP response for example.
5156 Replace nonce routines with a pair of functions.
5157 OCSP_request_add1_nonce() adds a nonce value and optionally
5158 generates a random value. OCSP_check_nonce() checks the
5159 validity of the nonce in an OCSP response.
5162 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5163 This doesn't copy the supplied OCSP_CERTID and avoids the
5164 need to free up the newly created id. Change return type
5165 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5166 This can then be used to add extensions to the request.
5167 Deleted OCSP_request_new(), since most of its functionality
5168 is now in OCSP_REQUEST_new() (and the case insensitive name
5169 clash) apart from the ability to set the request name which
5170 will be added elsewhere.
5173 *) Update OCSP API. Remove obsolete extensions argument from
5174 various functions. Extensions are now handled using the new
5175 OCSP extension code. New simple OCSP HTTP function which
5176 can be used to send requests and parse the response.
5179 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5180 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5181 uses the special reorder version of SET OF to sort the attributes
5182 and reorder them to match the encoded order. This resolves a long
5183 standing problem: a verify on a PKCS7 structure just after signing
5184 it used to fail because the attribute order did not match the
5185 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5186 it uses the received order. This is necessary to tolerate some broken
5187 software that does not order SET OF. This is handled by encoding
5188 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5189 to produce the required SET OF.
5192 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5193 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5194 files to get correct declarations of the ASN.1 item variables.
5197 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5198 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5199 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5200 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5201 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5202 ASN1_ITEM and no wrapper functions.
5205 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5206 replace the old function pointer based I/O routines. Change most of
5207 the *_d2i_bio() and *_d2i_fp() functions to use these.
5210 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5211 lines, recognice more "algorithms" that can be deselected, and make
5212 it complain about algorithm deselection that isn't recognised.
5215 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5216 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5217 to use new functions. Add NO_ASN1_OLD which can be set to remove
5218 some old style ASN1 functions: this can be used to determine if old
5219 code will still work when these eventually go away.
5222 *) New extension functions for OCSP structures, these follow the
5223 same conventions as certificates and CRLs.
5226 *) New function X509V3_add1_i2d(). This automatically encodes and
5227 adds an extension. Its behaviour can be customised with various
5228 flags to append, replace or delete. Various wrappers added for
5229 certifcates and CRLs.
5232 *) Fix to avoid calling the underlying ASN1 print routine when
5233 an extension cannot be parsed. Correct a typo in the
5234 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5237 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5238 entries for variables.
5241 *) Add functionality to apps/openssl.c for detecting locking
5242 problems: As the program is single-threaded, all we have
5243 to do is register a locking callback using an array for
5244 storing which locks are currently held by the program.
5247 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5248 SSL_get_ex_data_X509_STORE_idx(), which is used in
5249 ssl_verify_cert_chain() and thus can be called at any time
5250 during TLS/SSL handshakes so that thread-safety is essential.
5251 Unfortunately, the ex_data design is not at all suited
5252 for multi-threaded use, so it probably should be abolished.
5255 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5256 [Broadcom, tweaked and integrated by Geoff Thorpe]
5258 *) Move common extension printing code to new function
5259 X509V3_print_extensions(). Reorganise OCSP print routines and
5260 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5263 *) New function X509_signature_print() to remove duplication in some
5267 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5268 set (this was treated exactly the same as SET OF previously). This
5269 is used to reorder the STACK representing the structure to match the
5270 encoding. This will be used to get round a problem where a PKCS7
5271 structure which was signed could not be verified because the STACK
5272 order did not reflect the encoded order.
5275 *) Reimplement the OCSP ASN1 module using the new code.
5278 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5279 for its ASN1 operations. The old style function pointers still exist
5280 for now but they will eventually go away.
5283 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5284 completely replaces the old ASN1 functionality with a table driven
5285 encoder and decoder which interprets an ASN1_ITEM structure describing
5286 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5287 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5288 has also been converted to the new form.
5291 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5292 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5293 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5294 for negative moduli.
5297 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5298 of not touching the result's sign bit.
5301 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5305 *) Changed the LHASH code to use prototypes for callbacks, and created
5306 macros to declare and implement thin (optionally static) functions
5307 that provide type-safety and avoid function pointer casting for the
5308 type-specific callbacks.
5311 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5313 [Veers Staats <staatsvr@asc.hpc.mil>,
5314 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5316 *) Reformat the FAQ so the different questions and answers can be divided
5317 in sections depending on the subject.
5320 *) Have the zlib compression code load ZLIB.DLL dynamically under
5324 *) New function BN_mod_sqrt for computing square roots modulo a prime
5325 (using the probabilistic Tonelli-Shanks algorithm unless
5326 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5327 be handled deterministically).
5328 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5330 *) Make BN_mod_inverse faster by explicitly handling small quotients
5331 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5332 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5335 *) New function BN_kronecker.
5338 *) Fix BN_gcd so that it works on negative inputs; the result is
5339 positive unless both parameters are zero.
5340 Previously something reasonably close to an infinite loop was
5341 possible because numbers could be growing instead of shrinking
5342 in the implementation of Euclid's algorithm.
5345 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5346 sign of the number in question.
5348 Fix BN_is_word(a,w) to work correctly for w == 0.
5350 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5351 because its test if the absolute value of 'a' equals 'w'.
5352 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5353 it exists mostly for use in the implementations of BN_is_zero(),
5354 BN_is_one(), and BN_is_word().
5357 *) New function BN_swap.
5360 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5361 the exponentiation functions are more likely to produce reasonable
5362 results on negative inputs.
5365 *) Change BN_mod_mul so that the result is always non-negative.
5366 Previously, it could be negative if one of the factors was negative;
5367 I don't think anyone really wanted that behaviour.
5370 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5371 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5372 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5373 and add new functions:
5382 BN_mod_lshift1_quick
5386 These functions always generate non-negative results.
5388 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5389 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5391 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5392 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5393 be reduced modulo m.
5394 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5397 The following entry accidentily appeared in the CHANGES file
5398 distributed with OpenSSL 0.9.7. The modifications described in
5399 it do *not* apply to OpenSSL 0.9.7.
5401 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5402 was actually never needed) and in BN_mul(). The removal in BN_mul()
5403 required a small change in bn_mul_part_recursive() and the addition
5404 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5405 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5406 bn_sub_words() and bn_add_words() except they take arrays with
5411 *) In 'openssl passwd', verify passwords read from the terminal
5412 unless the '-salt' option is used (which usually means that
5413 verification would just waste user's time since the resulting
5414 hash is going to be compared with some given password hash)
5415 or the new '-noverify' option is used.
5417 This is an incompatible change, but it does not affect
5418 non-interactive use of 'openssl passwd' (passwords on the command
5419 line, '-stdin' option, '-in ...' option) and thus should not
5423 *) Remove all references to RSAref, since there's no more need for it.
5426 *) Make DSO load along a path given through an environment variable
5427 (SHLIB_PATH) with shl_load().
5430 *) Constify the ENGINE code as a result of BIGNUM constification.
5431 Also constify the RSA code and most things related to it. In a
5432 few places, most notable in the depth of the ASN.1 code, ugly
5433 casts back to non-const were required (to be solved at a later
5437 *) Make it so the openssl application has all engines loaded by default.
5440 *) Constify the BIGNUM routines a little more.
5443 *) Add the following functions:
5445 ENGINE_load_cswift()
5447 ENGINE_load_atalla()
5449 ENGINE_load_builtin_engines()
5451 That way, an application can itself choose if external engines that
5452 are built-in in OpenSSL shall ever be used or not. The benefit is
5453 that applications won't have to be linked with libdl or other dso
5454 libraries unless it's really needed.
5456 Changed 'openssl engine' to load all engines on demand.
5457 Changed the engine header files to avoid the duplication of some
5458 declarations (they differed!).
5461 *) 'openssl engine' can now list capabilities.
5464 *) Better error reporting in 'openssl engine'.
5467 *) Never call load_dh_param(NULL) in s_server.
5470 *) Add engine application. It can currently list engines by name and
5471 identity, and test if they are actually available.
5474 *) Improve RPM specification file by forcing symbolic linking and making
5475 sure the installed documentation is also owned by root.root.
5476 [Damien Miller <djm@mindrot.org>]
5478 *) Give the OpenSSL applications more possibilities to make use of
5479 keys (public as well as private) handled by engines.
5482 *) Add OCSP code that comes from CertCo.
5485 *) Add VMS support for the Rijndael code.
5488 *) Added untested support for Nuron crypto accelerator.
5491 *) Add support for external cryptographic devices. This code was
5492 previously distributed separately as the "engine" branch.
5493 [Geoff Thorpe, Richard Levitte]
5495 *) Rework the filename-translation in the DSO code. It is now possible to
5496 have far greater control over how a "name" is turned into a filename
5497 depending on the operating environment and any oddities about the
5498 different shared library filenames on each system.
5501 *) Support threads on FreeBSD-elf in Configure.
5504 *) Fix for SHA1 assembly problem with MASM: it produces
5505 warnings about corrupt line number information when assembling
5506 with debugging information. This is caused by the overlapping
5508 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5511 NCONF_get_number() has no error checking at all. As a replacement,
5512 NCONF_get_number_e() is defined (_e for "error checking") and is
5513 promoted strongly. The old NCONF_get_number is kept around for
5514 binary backward compatibility.
5515 Make it possible for methods to load from something other than a BIO,
5516 by providing a function pointer that is given a name instead of a BIO.
5517 For example, this could be used to load configuration data from an
5521 *) Fix for non blocking accept BIOs. Added new I/O special reason
5522 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5523 with non blocking I/O was not possible because no retry code was
5524 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5528 *) Added the beginnings of Rijndael support.
5531 *) Fix for bug in DirectoryString mask setting. Add support for
5532 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5533 to allow certificate printing to more controllable, additional
5534 'certopt' option to 'x509' to allow new printing options to be
5538 *) Clean old EAY MD5 hack from e_os.h.
5541 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5543 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5544 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5545 [Joe Orton, Steve Henson]
5547 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5549 *) Fix additional bug revealed by the NISCC test suite:
5551 Stop bug triggering large recursion when presented with
5552 certain ASN.1 tags (CVE-2003-0851)
5555 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5557 *) Fix various bugs revealed by running the NISCC test suite:
5559 Stop out of bounds reads in the ASN1 code when presented with
5560 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5562 If verify callback ignores invalid public key errors don't try to check
5563 certificate signature with the NULL public key.
5567 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5568 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5572 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5573 extra data after the compression methods not only for TLS 1.0
5574 but also for SSL 3.0 (as required by the specification).
5575 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5577 *) Change X509_certificate_type() to mark the key as exported/exportable
5578 when it's 512 *bits* long, not 512 bytes.
5581 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5583 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5584 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5585 a protocol version number mismatch like a decryption error
5586 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5589 *) Turn on RSA blinding by default in the default implementation
5590 to avoid a timing attack. Applications that don't want it can call
5591 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5592 They would be ill-advised to do so in most cases.
5593 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5595 *) Change RSA blinding code so that it works when the PRNG is not
5596 seeded (in this case, the secret RSA exponent is abused as
5597 an unpredictable seed -- if it is not unpredictable, there
5598 is no point in blinding anyway). Make RSA blinding thread-safe
5599 by remembering the creator's thread ID in rsa->blinding and
5600 having all other threads use local one-time blinding factors
5601 (this requires more computation than sharing rsa->blinding, but
5602 avoids excessive locking; and if an RSA object is not shared
5603 between threads, blinding will still be very fast).
5606 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5608 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5609 via timing by performing a MAC computation even if incorrrect
5610 block cipher padding has been found. This is a countermeasure
5611 against active attacks where the attacker has to distinguish
5612 between bad padding and a MAC verification error. (CVE-2003-0078)
5614 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5615 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5616 Martin Vuagnoux (EPFL, Ilion)]
5618 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5620 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5621 memory from it's contents. This is done with a counter that will
5622 place alternating values in each byte. This can be used to solve
5623 two issues: 1) the removal of calls to memset() by highly optimizing
5624 compilers, and 2) cleansing with other values than 0, since those can
5625 be read through on certain media, for example a swap space on disk.
5628 *) Bugfix: client side session caching did not work with external caching,
5629 because the session->cipher setting was not restored when reloading
5630 from the external cache. This problem was masked, when
5631 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5632 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5635 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5636 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5637 [Zeev Lieber <zeev-l@yahoo.com>]
5639 *) Undo an undocumented change introduced in 0.9.6e which caused
5640 repeated calls to OpenSSL_add_all_ciphers() and
5641 OpenSSL_add_all_digests() to be ignored, even after calling
5645 *) Change the default configuration reader to deal with last line not
5646 being properly terminated.
5649 *) Change X509_NAME_cmp() so it applies the special rules on handling
5650 DN values that are of type PrintableString, as well as RDNs of type
5651 emailAddress where the value has the type ia5String.
5652 [stefank@valicert.com via Richard Levitte]
5654 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5655 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5656 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5657 the bitwise-OR of the two for use by the majority of applications
5658 wanting this behaviour, and update the docs. The documented
5659 behaviour and actual behaviour were inconsistent and had been
5660 changing anyway, so this is more a bug-fix than a behavioural
5662 [Geoff Thorpe, diagnosed by Nadav Har'El]
5664 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5665 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5668 *) Fix initialization code race conditions in
5669 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5670 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5671 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5672 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5673 ssl2_get_cipher_by_char(),
5674 ssl3_get_cipher_by_char().
5675 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5677 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5678 the cached sessions are flushed, as the remove_cb() might use ex_data
5679 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5680 (see [openssl.org #212]).
5681 [Geoff Thorpe, Lutz Jaenicke]
5683 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5684 length, instead of the encoding length to d2i_ASN1_OBJECT.
5687 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5689 *) [In 0.9.6g-engine release:]
5690 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5691 [Lynn Gazis <lgazis@rainbow.com>]
5693 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5695 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5696 and get fix the header length calculation.
5697 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5698 Alon Kantor <alonk@checkpoint.com> (and others),
5701 *) Use proper error handling instead of 'assertions' in buffer
5702 overflow checks added in 0.9.6e. This prevents DoS (the
5703 assertions could call abort()).
5704 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5706 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5708 *) Add various sanity checks to asn1_get_length() to reject
5709 the ASN1 length bytes if they exceed sizeof(long), will appear
5710 negative or the content length exceeds the length of the
5712 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5714 *) Fix cipher selection routines: ciphers without encryption had no flags
5715 for the cipher strength set and where therefore not handled correctly
5716 by the selection routines (PR #130).
5719 *) Fix EVP_dsa_sha macro.
5723 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5724 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5725 that was added in OpenSSL 0.9.6d.
5727 As the countermeasure turned out to be incompatible with some
5728 broken SSL implementations, the new option is part of SSL_OP_ALL.
5729 SSL_OP_ALL is usually employed when compatibility with weird SSL
5730 implementations is desired (e.g. '-bugs' option to 's_client' and
5731 's_server'), so the new option is automatically set in many
5735 *) Changes in security patch:
5737 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5738 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5739 Air Force Materiel Command, USAF, under agreement number
5742 *) Add various sanity checks to asn1_get_length() to reject
5743 the ASN1 length bytes if they exceed sizeof(long), will appear
5744 negative or the content length exceeds the length of the
5745 supplied buffer. (CVE-2002-0659)
5746 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5748 *) Assertions for various potential buffer overflows, not known to
5750 [Ben Laurie (CHATS)]
5752 *) Various temporary buffers to hold ASCII versions of integers were
5753 too small for 64 bit platforms. (CVE-2002-0655)
5754 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5756 *) Remote buffer overflow in SSL3 protocol - an attacker could
5757 supply an oversized session ID to a client. (CVE-2002-0656)
5758 [Ben Laurie (CHATS)]
5760 *) Remote buffer overflow in SSL2 protocol - an attacker could
5761 supply an oversized client master key. (CVE-2002-0656)
5762 [Ben Laurie (CHATS)]
5764 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5766 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5767 encoded as NULL) with id-dsa-with-sha1.
5768 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5770 *) Check various X509_...() return values in apps/req.c.
5771 [Nils Larsch <nla@trustcenter.de>]
5773 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5774 an end-of-file condition would erronously be flagged, when the CRLF
5775 was just at the end of a processed block. The bug was discovered when
5776 processing data through a buffering memory BIO handing the data to a
5777 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5778 <ptsekov@syntrex.com> and Nedelcho Stanev.
5781 *) Implement a countermeasure against a vulnerability recently found
5782 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5783 before application data chunks to avoid the use of known IVs
5784 with data potentially chosen by the attacker.
5787 *) Fix length checks in ssl3_get_client_hello().
5790 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5791 to prevent ssl3_read_internal() from incorrectly assuming that
5792 ssl3_read_bytes() found application data while handshake
5793 processing was enabled when in fact s->s3->in_read_app_data was
5794 merely automatically cleared during the initial handshake.
5795 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5797 *) Fix object definitions for Private and Enterprise: they were not
5798 recognized in their shortname (=lowercase) representation. Extend
5799 obj_dat.pl to issue an error when using undefined keywords instead
5800 of silently ignoring the problem (Svenning Sorensen
5801 <sss@sss.dnsalias.net>).
5804 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5805 generators, i.e. generators other than 2 and 5. (Previously, the
5806 code did not properly initialise the 'add' and 'rem' values to
5807 BN_generate_prime().)
5809 In the new general case, we do not insist that 'generator' is
5810 actually a primitive root: This requirement is rather pointless;
5811 a generator of the order-q subgroup is just as good, if not
5815 *) Map new X509 verification errors to alerts. Discovered and submitted by
5816 Tom Wu <tom@arcot.com>.
5819 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5820 returning non-zero before the data has been completely received
5821 when using non-blocking I/O.
5822 [Bodo Moeller; problem pointed out by John Hughes]
5824 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5825 [Ben Laurie, Lutz Jaenicke]
5827 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5828 Yoram Zahavi <YoramZ@gilian.com>).
5831 *) Add information about CygWin 1.3 and on, and preserve proper
5832 configuration for the versions before that.
5833 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5835 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5836 check whether we deal with a copy of a session and do not delete from
5837 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5838 <izhar@checkpoint.com>.
5841 *) Do not store session data into the internal session cache, if it
5842 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5843 flag is set). Proposed by Aslam <aslam@funk.com>.
5846 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5850 *) [In 0.9.6d-engine release:]
5851 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5852 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5854 *) Add the configuration target linux-s390x.
5855 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5857 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5858 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5859 variable as an indication that a ClientHello message has been
5860 received. As the flag value will be lost between multiple
5861 invocations of ssl3_accept when using non-blocking I/O, the
5862 function may not be aware that a handshake has actually taken
5863 place, thus preventing a new session from being added to the
5866 To avoid this problem, we now set s->new_session to 2 instead of
5867 using a local variable.
5868 [Lutz Jaenicke, Bodo Moeller]
5870 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5871 if the SSL_R_LENGTH_MISMATCH error is detected.
5872 [Geoff Thorpe, Bodo Moeller]
5874 *) New 'shared_ldflag' column in Configure platform table.
5877 *) Fix EVP_CIPHER_mode macro.
5878 ["Dan S. Camper" <dan@bti.net>]
5880 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5881 type, we must throw them away by setting rr->length to 0.
5882 [D P Chang <dpc@qualys.com>]
5884 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5886 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5887 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5888 worked incorrectly for those cases where range = 10..._2 and
5889 3*range is two bits longer than range.)
5892 *) Only add signing time to PKCS7 structures if it is not already
5896 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5897 OBJ_ld_ce should be OBJ_id_ce.
5898 Also some ip-pda OIDs in crypto/objects/objects.txt were
5899 incorrect (cf. RFC 3039).
5900 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5902 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5903 returns early because it has nothing to do.
5904 [Andy Schneider <andy.schneider@bjss.co.uk>]
5906 *) [In 0.9.6c-engine release:]
5907 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5908 [Andy Schneider <andy.schneider@bjss.co.uk>]
5910 *) [In 0.9.6c-engine release:]
5911 Add support for Cryptographic Appliance's keyserver technology.
5912 (Use engine 'keyclient')
5913 [Cryptographic Appliances and Geoff Thorpe]
5915 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5916 is called via tools/c89.sh because arguments have to be
5917 rearranged (all '-L' options must appear before the first object
5919 [Richard Shapiro <rshapiro@abinitio.com>]
5921 *) [In 0.9.6c-engine release:]
5922 Add support for Broadcom crypto accelerator cards, backported
5924 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5926 *) [In 0.9.6c-engine release:]
5927 Add support for SureWare crypto accelerator cards from
5928 Baltimore Technologies. (Use engine 'sureware')
5929 [Baltimore Technologies and Mark Cox]
5931 *) [In 0.9.6c-engine release:]
5932 Add support for crypto accelerator cards from Accelerated
5933 Encryption Processing, www.aep.ie. (Use engine 'aep')
5934 [AEP Inc. and Mark Cox]
5936 *) Add a configuration entry for gcc on UnixWare.
5937 [Gary Benson <gbenson@redhat.com>]
5939 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5940 messages are stored in a single piece (fixed-length part and
5941 variable-length part combined) and fix various bugs found on the way.
5944 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5945 instead. BIO_gethostbyname() does not know what timeouts are
5946 appropriate, so entries would stay in cache even when they have
5948 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5950 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5951 faced with a pathologically small ClientHello fragment that does
5952 not contain client_version: Instead of aborting with an error,
5953 simply choose the highest available protocol version (i.e.,
5954 TLS 1.0 unless it is disabled). In practice, ClientHello
5955 messages are never sent like this, but this change gives us
5956 strictly correct behaviour at least for TLS.
5959 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5960 never resets s->method to s->ctx->method when called from within
5961 one of the SSL handshake functions.
5962 [Bodo Moeller; problem pointed out by Niko Baric]
5964 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5965 (sent using the client's version number) if client_version is
5966 smaller than the protocol version in use. Also change
5967 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5968 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5969 the client will at least see that alert.
5972 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5976 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5977 client receives HelloRequest while in a handshake.
5978 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5980 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5981 should end in 'break', not 'goto end' which circuments various
5982 cleanups done in state SSL_ST_OK. But session related stuff
5983 must be disabled for SSL_ST_OK in the case that we just sent a
5986 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5987 before just sending a HelloRequest.
5988 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5990 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5991 reveal whether illegal block cipher padding was found or a MAC
5992 verification error occured. (Neither SSLerr() codes nor alerts
5993 are directly visible to potential attackers, but the information
5994 may leak via logfiles.)
5996 Similar changes are not required for the SSL 2.0 implementation
5997 because the number of padding bytes is sent in clear for SSL 2.0,
5998 and the extra bytes are just ignored. However ssl/s2_pkt.c
5999 failed to verify that the purported number of padding bytes is in
6003 *) Add OpenUNIX-8 support including shared libraries
6004 (Boyd Lynn Gerber <gerberb@zenez.com>).
6007 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6008 'wristwatch attack' using huge encoding parameters (cf.
6009 James H. Manger's CRYPTO 2001 paper). Note that the
6010 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6011 encoding parameters and hence was not vulnerable.
6014 *) BN_sqr() bug fix.
6015 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6017 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6018 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6019 followed by modular reduction.
6020 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6022 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6023 equivalent based on BN_pseudo_rand() instead of BN_rand().
6026 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6027 This function was broken, as the check for a new client hello message
6028 to handle SGC did not allow these large messages.
6029 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6032 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6035 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6036 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6039 *) Rework the configuration and shared library support for Tru64 Unix.
6040 The configuration part makes use of modern compiler features and
6041 still retains old compiler behavior for those that run older versions
6042 of the OS. The shared library support part includes a variant that
6043 uses the RPATH feature, and is available through the special
6044 configuration target "alpha-cc-rpath", which will never be selected
6046 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6048 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6049 with the same message size as in ssl3_get_certificate_request().
6050 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6051 messages might inadvertently be reject as too long.
6052 [Petr Lampa <lampa@fee.vutbr.cz>]
6054 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6057 *) Modified SSL library such that the verify_callback that has been set
6058 specificly for an SSL object with SSL_set_verify() is actually being
6059 used. Before the change, a verify_callback set with this function was
6060 ignored and the verify_callback() set in the SSL_CTX at the time of
6061 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6062 to allow the necessary settings.
6065 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6066 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6067 done automatically (in contradiction to the requirements of the C
6068 standard). This made problems when used from OpenSSH.
6071 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6072 dh->length and always used
6074 BN_rand_range(priv_key, dh->p).
6076 BN_rand_range() is not necessary for Diffie-Hellman, and this
6077 specific range makes Diffie-Hellman unnecessarily inefficient if
6078 dh->length (recommended exponent length) is much smaller than the
6079 length of dh->p. We could use BN_rand_range() if the order of
6080 the subgroup was stored in the DH structure, but we only have
6085 BN_rand(priv_key, l, ...)
6087 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6093 RSA_eay_public_encrypt
6094 RSA_eay_private_decrypt
6095 RSA_eay_private_encrypt (signing)
6096 RSA_eay_public_decrypt (signature verification)
6098 (default implementations for RSA_public_encrypt,
6099 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6100 always reject numbers >= n.
6103 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6104 to synchronize access to 'locking_thread'. This is necessary on
6105 systems where access to 'locking_thread' (an 'unsigned long'
6106 variable) is not atomic.
6109 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6110 *before* setting the 'crypto_lock_rand' flag. The previous code had
6111 a race condition if 0 is a valid thread ID.
6112 [Travis Vitek <vitek@roguewave.com>]
6114 *) Add support for shared libraries under Irix.
6115 [Albert Chin-A-Young <china@thewrittenword.com>]
6117 *) Add configuration option to build on Linux on both big-endian and
6119 [Ralf Baechle <ralf@uni-koblenz.de>]
6121 *) Add the possibility to create shared libraries on HP-UX.
6124 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6126 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6127 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6128 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6129 PRNG state recovery was possible based on the output of
6130 one PRNG request appropriately sized to gain knowledge on
6131 'md' followed by enough consecutive 1-byte PRNG requests
6132 to traverse all of 'state'.
6134 1. When updating 'md_local' (the current thread's copy of 'md')
6135 during PRNG output generation, hash all of the previous
6136 'md_local' value, not just the half used for PRNG output.
6138 2. Make the number of bytes from 'state' included into the hash
6139 independent from the number of PRNG bytes requested.
6141 The first measure alone would be sufficient to avoid
6142 Markku-Juhani's attack. (Actually it had never occurred
6143 to me that the half of 'md_local' used for chaining was the
6144 half from which PRNG output bytes were taken -- I had always
6145 assumed that the secret half would be used.) The second
6146 measure makes sure that additional data from 'state' is never
6147 mixed into 'md_local' in small portions; this heuristically
6148 further strengthens the PRNG.
6151 *) Fix crypto/bn/asm/mips3.s.
6154 *) When only the key is given to "enc", the IV is undefined. Print out
6155 an error message in this case.
6158 *) Handle special case when X509_NAME is empty in X509 printing routines.
6161 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6162 positive and less than q.
6165 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6166 used: it isn't thread safe and the add_lock_callback should handle
6168 [Paul Rose <Paul.Rose@bridge.com>]
6170 *) Verify that incoming data obeys the block size in
6171 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6175 [Ulf Möller, Bodo Möller]
6177 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6178 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6179 when fixing the server behaviour for backwards-compatible 'client
6180 hello' messages. (Note that the attack is impractical against
6181 SSL 3.0 and TLS 1.0 anyway because length and version checking
6182 means that the probability of guessing a valid ciphertext is
6183 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6186 Before 0.9.5, the countermeasure (hide the error by generating a
6187 random 'decryption result') did not work properly because
6188 ERR_clear_error() was missing, meaning that SSL_get_error() would
6189 detect the supposedly ignored error.
6191 Both problems are now fixed.
6194 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6195 (previously it was 1024).
6198 *) Fix for compatibility mode trust settings: ignore trust settings
6199 unless some valid trust or reject settings are present.
6202 *) Fix for blowfish EVP: its a variable length cipher.
6205 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6206 parameters in DSA public key structures and return an error in the
6207 DSA routines if parameters are absent.
6210 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6211 in the current directory if neither $RANDFILE nor $HOME was set.
6212 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6213 caused some confusion to Windows users who haven't defined $HOME.
6214 Thus RAND_file_name() is changed again: e_os.h can define a
6215 DEFAULT_HOME, which will be used if $HOME is not set.
6216 For Windows, we use "C:"; on other platforms, we still require
6217 environment variables.
6219 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6220 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6221 having multiple threads call RAND_poll() concurrently.
6224 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6225 combination of a flag and a thread ID variable.
6226 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6227 flag), *other* threads can enter ssleay_add_bytes without obeying
6228 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6229 that they do not hold after the first thread unsets add_do_not_lock).
6232 *) Change bctest again: '-x' expressions are not available in all
6236 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6238 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6239 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6241 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6242 the default extension for executables, if any. Also, make the perl
6243 scripts that use symlink() to test if it really exists and use "cp"
6244 if it doesn't. All this made OpenSSL compilable and installable in
6248 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6249 If SEQUENCE is length is indefinite just set c->slen to the total
6250 amount of data available.
6251 [Steve Henson, reported by shige@FreeBSD.org]
6252 [This change does not apply to 0.9.7.]
6254 *) Change bctest to avoid here-documents inside command substitution
6255 (workaround for FreeBSD /bin/sh bug).
6256 For compatibility with Ultrix, avoid shell functions (introduced
6257 in the bctest version that searches along $PATH).
6260 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6261 with des_encrypt() defined on some operating systems, like Solaris
6265 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6266 On the Importance of Eliminating Errors in Cryptographic
6267 Computations, J. Cryptology 14 (2001) 2, 101-119,
6268 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6271 *) MIPS assembler BIGNUM division bug fix.
6274 *) Disabled incorrect Alpha assembler code.
6277 *) Fix PKCS#7 decode routines so they correctly update the length
6278 after reading an EOC for the EXPLICIT tag.
6280 [This change does not apply to 0.9.7.]
6282 *) Fix bug in PKCS#12 key generation routines. This was triggered
6283 if a 3DES key was generated with a 0 initial byte. Include
6284 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6285 (but broken) behaviour.
6288 *) Enhance bctest to search for a working bc along $PATH and print
6290 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6292 *) Fix memory leaks in err.c: free err_data string if necessary;
6293 don't write to the wrong index in ERR_set_error_data.
6296 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6300 *) Replace rdtsc with _emit statements for VC++ version 5.
6301 [Jeremy Cooper <jeremy@baymoo.org>]
6303 *) Make it possible to reuse SSLv2 sessions.
6306 *) In copy_email() check for >= 0 as a return value for
6307 X509_NAME_get_index_by_NID() since 0 is a valid index.
6308 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6310 *) Avoid coredump with unsupported or invalid public keys by checking if
6311 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6312 PKCS7_verify() fails with non detached data.
6315 *) Don't use getenv in library functions when run as setuid/setgid.
6316 New function OPENSSL_issetugid().
6319 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6320 due to incorrect handling of multi-threading:
6322 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6324 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6326 3. Count how many times MemCheck_off() has been called so that
6327 nested use can be treated correctly. This also avoids
6328 inband-signalling in the previous code (which relied on the
6329 assumption that thread ID 0 is impossible).
6332 *) Add "-rand" option also to s_client and s_server.
6335 *) Fix CPU detection on Irix 6.x.
6336 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6337 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6339 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6342 [This change does not apply to 0.9.7.]
6344 *) Use the cached encoding of an X509_NAME structure rather than
6345 copying it. This is apparently the reason for the libsafe "errors"
6346 but the code is actually correct.
6349 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6350 Bleichenbacher's DSA attack.
6351 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6352 to be set and top=0 forces the highest bit to be set; top=-1 is new
6353 and leaves the highest bit random.
6354 [Ulf Moeller, Bodo Moeller]
6356 *) In the NCONF_...-based implementations for CONF_... queries
6357 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6358 a temporary CONF structure with the data component set to NULL
6359 (which gives segmentation faults in lh_retrieve).
6360 Instead, use NULL for the CONF pointer in CONF_get_string and
6361 CONF_get_number (which may use environment variables) and directly
6362 return NULL from CONF_get_section.
6365 *) Fix potential buffer overrun for EBCDIC.
6368 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6369 keyUsage if basicConstraints absent for a CA.
6372 *) Make SMIME_write_PKCS7() write mail header values with a format that
6373 is more generally accepted (no spaces before the semicolon), since
6374 some programs can't parse those values properly otherwise. Also make
6375 sure BIO's that break lines after each write do not create invalid
6379 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6380 macros previously used would not encode an empty SEQUENCE OF
6381 and break the signature.
6383 [This change does not apply to 0.9.7.]
6385 *) Zero the premaster secret after deriving the master secret in
6389 *) Add some EVP_add_digest_alias registrations (as found in
6390 OpenSSL_add_all_digests()) to SSL_library_init()
6391 aka OpenSSL_add_ssl_algorithms(). This provides improved
6392 compatibility with peers using X.509 certificates
6393 with unconventional AlgorithmIdentifier OIDs.
6396 *) Fix for Irix with NO_ASM.
6397 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6399 *) ./config script fixes.
6400 [Ulf Moeller, Richard Levitte]
6402 *) Fix 'openssl passwd -1'.
6405 *) Change PKCS12_key_gen_asc() so it can cope with non null
6406 terminated strings whose length is passed in the passlen
6407 parameter, for example from PEM callbacks. This was done
6408 by adding an extra length parameter to asc2uni().
6409 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6411 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6412 call failed, free the DSA structure.
6415 *) Fix to uni2asc() to cope with zero length Unicode strings.
6416 These are present in some PKCS#12 files.
6419 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6420 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6421 when writing a 32767 byte record.
6422 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6424 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6425 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6427 (RSA objects have a reference count access to which is protected
6428 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6429 so they are meant to be shared between threads.)
6430 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6431 "Reddie, Steven" <Steven.Reddie@ca.com>]
6433 *) Fix a deadlock in CRYPTO_mem_leaks().
6436 *) Use better test patterns in bntest.
6439 *) rand_win.c fix for Borland C.
6442 *) BN_rshift bugfix for n == 0.
6445 *) Add a 'bctest' script that checks for some known 'bc' bugs
6446 so that 'make test' does not abort just because 'bc' is broken.
6449 *) Store verify_result within SSL_SESSION also for client side to
6450 avoid potential security hole. (Re-used sessions on the client side
6451 always resulted in verify_result==X509_V_OK, not using the original
6452 result of the server certificate verification.)
6455 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6456 SSL3_RT_APPLICATION_DATA, return 0.
6457 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6461 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6462 releases, have been re-implemented by renaming the previous
6463 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6464 and ssl3_read_internal, respectively, and adding 'peek' parameters
6465 to them. The new ssl[23]_{read,peek} functions are calls to
6466 ssl[23]_read_internal with the 'peek' flag set appropriately.
6467 A 'peek' parameter has also been added to ssl3_read_bytes, which
6468 does the actual work for ssl3_read_internal.
6471 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6472 the method-specific "init()" handler. Also clean up ex_data after
6473 calling the method-specific "finish()" handler. Previously, this was
6474 happening the other way round.
6477 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6478 The previous value, 12, was not always sufficient for BN_mod_exp().
6481 *) Make sure that shared libraries get the internal name engine with
6482 the full version number and not just 0. This should mark the
6483 shared libraries as not backward compatible. Of course, this should
6484 be changed again when we can guarantee backward binary compatibility.
6487 *) Fix typo in get_cert_by_subject() in by_dir.c
6488 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6490 *) Rework the system to generate shared libraries:
6492 - Make note of the expected extension for the shared libraries and
6493 if there is a need for symbolic links from for example libcrypto.so.0
6494 to libcrypto.so.0.9.7. There is extended info in Configure for
6497 - Make as few rebuilds of the shared libraries as possible.
6499 - Still avoid linking the OpenSSL programs with the shared libraries.
6501 - When installing, install the shared libraries separately from the
6505 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6507 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6508 and not in SSL_clear because the latter is also used by the
6509 accept/connect functions; previously, the settings made by
6510 SSL_set_read_ahead would be lost during the handshake.
6511 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6513 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6514 Previously, it would create entries for disableed algorithms no
6518 *) Added several new manual pages for SSL_* function.
6521 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6523 *) In ssl23_get_client_hello, generate an error message when faced
6524 with an initial SSL 3.0/TLS record that is too small to contain the
6525 first two bytes of the ClientHello message, i.e. client_version.
6526 (Note that this is a pathologic case that probably has never happened
6527 in real life.) The previous approach was to use the version number
6528 from the record header as a substitute; but our protocol choice
6529 should not depend on that one because it is not authenticated
6530 by the Finished messages.
6533 *) More robust randomness gathering functions for Windows.
6534 [Jeffrey Altman <jaltman@columbia.edu>]
6536 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6537 not set then we don't setup the error code for issuer check errors
6538 to avoid possibly overwriting other errors which the callback does
6539 handle. If an application does set the flag then we assume it knows
6540 what it is doing and can handle the new informational codes
6544 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6545 a general "ANY" type, as such it should be able to decode anything
6546 including tagged types. However it didn't check the class so it would
6547 wrongly interpret tagged types in the same way as their universal
6548 counterpart and unknown types were just rejected. Changed so that the
6549 tagged and unknown types are handled in the same way as a SEQUENCE:
6550 that is the encoding is stored intact. There is also a new type
6551 "V_ASN1_OTHER" which is used when the class is not universal, in this
6552 case we have no idea what the actual type is so we just lump them all
6556 *) On VMS, stdout may very well lead to a file that is written to
6557 in a record-oriented fashion. That means that every write() will
6558 write a separate record, which will be read separately by the
6559 programs trying to read from it. This can be very confusing.
6561 The solution is to put a BIO filter in the way that will buffer
6562 text until a linefeed is reached, and then write everything a
6563 line at a time, so every record written will be an actual line,
6564 not chunks of lines and not (usually doesn't happen, but I've
6565 seen it once) several lines in one record. BIO_f_linebuffer() is
6568 Currently, it's a VMS-only method, because that's where it has
6569 been tested well enough.
6572 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6573 it can return incorrect results.
6574 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6575 but it was in 0.9.6-beta[12].)
6578 *) Disable the check for content being present when verifying detached
6579 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6580 include zero length content when signing messages.
6583 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6584 BIO_ctrl (for BIO pairs).
6587 *) Add DSO method for VMS.
6590 *) Bug fix: Montgomery multiplication could produce results with the
6594 *) Add RPM specification openssl.spec and modify it to build three
6595 packages. The default package contains applications, application
6596 documentation and run-time libraries. The devel package contains
6597 include files, static libraries and function documentation. The
6598 doc package contains the contents of the doc directory. The original
6599 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6602 *) Add a large number of documentation files for many SSL routines.
6603 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6605 *) Add a configuration entry for Sony News 4.
6606 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6608 *) Don't set the two most significant bits to one when generating a
6609 random number < q in the DSA library.
6612 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6613 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6614 the underlying transport is blocking) if a handshake took place.
6615 (The default behaviour is needed by applications such as s_client
6616 and s_server that use select() to determine when to use SSL_read;
6617 but for applications that know in advance when to expect data, it
6618 just makes things more complicated.)
6621 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6625 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6626 work better on such systems.
6627 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6629 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6630 Update PKCS12_parse() so it copies the friendlyName and the
6631 keyid to the certificates aux info.
6634 *) Fix bug in PKCS7_verify() which caused an infinite loop
6635 if there was more than one signature.
6636 [Sven Uszpelkat <su@celocom.de>]
6638 *) Major change in util/mkdef.pl to include extra information
6639 about each symbol, as well as presentig variables as well
6640 as functions. This change means that there's n more need
6641 to rebuild the .num files when some algorithms are excluded.
6644 *) Allow the verify time to be set by an application,
6645 rather than always using the current time.
6648 *) Phase 2 verify code reorganisation. The certificate
6649 verify code now looks up an issuer certificate by a
6650 number of criteria: subject name, authority key id
6651 and key usage. It also verifies self signed certificates
6652 by the same criteria. The main comparison function is
6653 X509_check_issued() which performs these checks.
6655 Lot of changes were necessary in order to support this
6656 without completely rewriting the lookup code.
6658 Authority and subject key identifier are now cached.
6660 The LHASH 'certs' is X509_STORE has now been replaced
6661 by a STACK_OF(X509_OBJECT). This is mainly because an
6662 LHASH can't store or retrieve multiple objects with
6663 the same hash value.
6665 As a result various functions (which were all internal
6666 use only) have changed to handle the new X509_STORE
6667 structure. This will break anything that messed round
6668 with X509_STORE internally.
6670 The functions X509_STORE_add_cert() now checks for an
6671 exact match, rather than just subject name.
6673 The X509_STORE API doesn't directly support the retrieval
6674 of multiple certificates matching a given criteria, however
6675 this can be worked round by performing a lookup first
6676 (which will fill the cache with candidate certificates)
6677 and then examining the cache for matches. This is probably
6678 the best we can do without throwing out X509_LOOKUP
6679 entirely (maybe later...).
6681 The X509_VERIFY_CTX structure has been enhanced considerably.
6683 All certificate lookup operations now go via a get_issuer()
6684 callback. Although this currently uses an X509_STORE it
6685 can be replaced by custom lookups. This is a simple way
6686 to bypass the X509_STORE hackery necessary to make this
6687 work and makes it possible to use more efficient techniques
6688 in future. A very simple version which uses a simple
6689 STACK for its trusted certificate store is also provided
6690 using X509_STORE_CTX_trusted_stack().
6692 The verify_cb() and verify() callbacks now have equivalents
6693 in the X509_STORE_CTX structure.
6695 X509_STORE_CTX also has a 'flags' field which can be used
6696 to customise the verify behaviour.
6699 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6700 excludes S/MIME capabilities.
6703 *) When a certificate request is read in keep a copy of the
6704 original encoding of the signed data and use it when outputing
6705 again. Signatures then use the original encoding rather than
6706 a decoded, encoded version which may cause problems if the
6707 request is improperly encoded.
6710 *) For consistency with other BIO_puts implementations, call
6711 buffer_write(b, ...) directly in buffer_puts instead of calling
6714 In BIO_puts, increment b->num_write as in BIO_write.
6715 [Peter.Sylvester@EdelWeb.fr]
6717 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6718 BN_zero, we may not return a BIGNUM with an array consisting of
6722 *) Avoid calling abort() from within the library when problems are
6723 detected, except if preprocessor symbols have been defined
6724 (such as REF_CHECK, BN_DEBUG etc.).
6727 *) New openssl application 'rsautl'. This utility can be
6728 used for low level RSA operations. DER public key
6729 BIO/fp routines also added.
6732 *) New Configure entry and patches for compiling on QNX 4.
6733 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6735 *) A demo state-machine implementation was sponsored by
6736 Nuron (http://www.nuron.com/) and is now available in
6737 demos/state_machine.
6740 *) New options added to the 'dgst' utility for signature
6741 generation and verification.
6744 *) Unrecognized PKCS#7 content types are now handled via a
6745 catch all ASN1_TYPE structure. This allows unsupported
6746 types to be stored as a "blob" and an application can
6747 encode and decode it manually.
6750 *) Fix various signed/unsigned issues to make a_strex.c
6752 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6754 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6755 length if passed a buffer. ASN1_INTEGER_to_BN failed
6756 if passed a NULL BN and its argument was negative.
6757 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6759 *) Modification to PKCS#7 encoding routines to output definite
6760 length encoding. Since currently the whole structures are in
6761 memory there's not real point in using indefinite length
6762 constructed encoding. However if OpenSSL is compiled with
6763 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6766 *) Added BIO_vprintf() and BIO_vsnprintf().
6769 *) Added more prefixes to parse for in the the strings written
6770 through a logging bio, to cover all the levels that are available
6771 through syslog. The prefixes are now:
6773 PANIC, EMERG, EMR => LOG_EMERG
6774 ALERT, ALR => LOG_ALERT
6775 CRIT, CRI => LOG_CRIT
6776 ERROR, ERR => LOG_ERR
6777 WARNING, WARN, WAR => LOG_WARNING
6778 NOTICE, NOTE, NOT => LOG_NOTICE
6779 INFO, INF => LOG_INFO
6780 DEBUG, DBG => LOG_DEBUG
6782 and as before, if none of those prefixes are present at the
6783 beginning of the string, LOG_ERR is chosen.
6785 On Win32, the LOG_* levels are mapped according to this:
6787 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6788 LOG_WARNING => EVENTLOG_WARNING_TYPE
6789 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6793 *) Made it possible to reconfigure with just the configuration
6794 argument "reconf" or "reconfigure". The command line arguments
6795 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6796 and are retrieved from there when reconfiguring.
6800 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6802 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6805 *) The obj_dat.pl script was messing up the sorting of object
6806 names. The reason was that it compared the quoted version
6807 of strings as a result "OCSP" > "OCSP Signing" because
6808 " > SPACE. Changed script to store unquoted versions of
6809 names and add quotes on output. It was also omitting some
6810 names from the lookup table if they were given a default
6811 value (that is if SN is missing it is given the same
6812 value as LN and vice versa), these are now added on the
6813 grounds that if an object has a name we should be able to
6814 look it up. Finally added warning output when duplicate
6815 short or long names are found.
6818 *) Changes needed for Tandem NSK.
6819 [Scott Uroff <scott@xypro.com>]
6821 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6822 RSA_padding_check_SSLv23(), special padding was never detected
6823 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6824 version rollback attacks was not effective.
6826 In s23_clnt.c, don't use special rollback-attack detection padding
6827 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6828 client; similarly, in s23_srvr.c, don't do the rollback check if
6829 SSL 2.0 is the only protocol enabled in the server.
6832 *) Make it possible to get hexdumps of unprintable data with 'openssl
6833 asn1parse'. By implication, the functions ASN1_parse_dump() and
6834 BIO_dump_indent() are added.
6837 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6838 these print out strings and name structures based on various
6839 flags including RFC2253 support and proper handling of
6840 multibyte characters. Added options to the 'x509' utility
6841 to allow the various flags to be set.
6844 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6845 Also change the functions X509_cmp_current_time() and
6846 X509_gmtime_adj() work with an ASN1_TIME structure,
6847 this will enable certificates using GeneralizedTime in validity
6848 dates to be checked.
6851 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6852 negative public key encodings) on by default,
6853 NO_NEG_PUBKEY_BUG can be set to disable it.
6856 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6857 content octets. An i2c_ASN1_OBJECT is unnecessary because
6858 the encoding can be trivially obtained from the structure.
6861 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6862 not read locks (CRYPTO_r_[un]lock).
6865 *) A first attempt at creating official support for shared
6866 libraries through configuration. I've kept it so the
6867 default is static libraries only, and the OpenSSL programs
6868 are always statically linked for now, but there are
6869 preparations for dynamic linking in place.
6870 This has been tested on Linux and Tru64.
6873 *) Randomness polling function for Win9x, as described in:
6874 Peter Gutmann, Software Generation of Practically Strong
6878 *) Fix so PRNG is seeded in req if using an already existing
6882 *) New options to smime application. -inform and -outform
6883 allow alternative formats for the S/MIME message including
6884 PEM and DER. The -content option allows the content to be
6885 specified separately. This should allow things like Netscape
6886 form signing output easier to verify.
6889 *) Fix the ASN1 encoding of tags using the 'long form'.
6892 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6893 STRING types. These convert content octets to and from the
6894 underlying type. The actual tag and length octets are
6895 already assumed to have been read in and checked. These
6896 are needed because all other string types have virtually
6897 identical handling apart from the tag. By having versions
6898 of the ASN1 functions that just operate on content octets
6899 IMPLICIT tagging can be handled properly. It also allows
6900 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6901 and ASN1_INTEGER are identical apart from the tag.
6904 *) Change the handling of OID objects as follows:
6906 - New object identifiers are inserted in objects.txt, following
6907 the syntax given in objects.README.
6908 - objects.pl is used to process obj_mac.num and create a new
6910 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6913 This is currently kind of a hack, and the perl code in objects.pl
6914 isn't very elegant, but it works as I intended. The simplest way
6915 to check that it worked correctly is to look in obj_dat.h and
6916 check the array nid_objs and make sure the objects haven't moved
6917 around (this is important!). Additions are OK, as well as
6918 consistent name changes.
6921 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6924 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6925 The given file adds to whatever has already been seeded into the
6926 random pool through the RANDFILE configuration file option or
6927 environment variable, or the default random state file.
6930 *) mkstack.pl now sorts each macro group into lexical order.
6931 Previously the output order depended on the order the files
6932 appeared in the directory, resulting in needless rewriting
6936 *) Patches to make OpenSSL compile under Win32 again. Mostly
6937 work arounds for the VC++ problem that it treats func() as
6938 func(void). Also stripped out the parts of mkdef.pl that
6939 added extra typesafe functions: these no longer exist.
6942 *) Reorganisation of the stack code. The macros are now all
6943 collected in safestack.h . Each macro is defined in terms of
6944 a "stack macro" of the form SKM_<name>(type, a, b). The
6945 DEBUG_SAFESTACK is now handled in terms of function casts,
6946 this has the advantage of retaining type safety without the
6947 use of additional functions. If DEBUG_SAFESTACK is not defined
6948 then the non typesafe macros are used instead. Also modified the
6949 mkstack.pl script to handle the new form. Needs testing to see
6950 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6951 the default if no major problems. Similar behaviour for ASN1_SET_OF
6952 and PKCS12_STACK_OF.
6955 *) When some versions of IIS use the 'NET' form of private key the
6956 key derivation algorithm is different. Normally MD5(password) is
6957 used as a 128 bit RC4 key. In the modified case
6958 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6959 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6960 as the old Netscape_RSA functions except they have an additional
6961 'sgckey' parameter which uses the modified algorithm. Also added
6962 an -sgckey command line option to the rsa utility. Thanks to
6963 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6964 algorithm to openssl-dev.
6967 *) The evp_local.h macros were using 'c.##kname' which resulted in
6968 invalid expansion on some systems (SCO 5.0.5 for example).
6969 Corrected to 'c.kname'.
6970 [Phillip Porch <root@theporch.com>]
6972 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6973 a STACK of email addresses from a certificate or request, these look
6974 in the subject name and the subject alternative name extensions and
6975 omit any duplicate addresses.
6978 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6979 This makes DSA verification about 2 % faster.
6982 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6983 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6984 plus overhead for 1024 bit moduli).
6985 This makes exponentiations about 0.5 % faster for 1024 bit
6986 exponents (as measured by "openssl speed rsa2048").
6989 *) Rename memory handling macros to avoid conflicts with other
6991 Malloc => OPENSSL_malloc
6992 Malloc_locked => OPENSSL_malloc_locked
6993 Realloc => OPENSSL_realloc
6994 Free => OPENSSL_free
6997 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6998 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7001 *) CygWin32 support.
7002 [John Jarvie <jjarvie@newsguy.com>]
7004 *) The type-safe stack code has been rejigged. It is now only compiled
7005 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7006 by default all type-specific stack functions are "#define"d back to
7007 standard stack functions. This results in more streamlined output
7008 but retains the type-safety checking possibilities of the original
7012 *) The STACK code has been cleaned up, and certain type declarations
7013 that didn't make a lot of sense have been brought in line. This has
7014 also involved a cleanup of sorts in safestack.h to more correctly
7015 map type-safe stack functions onto their plain stack counterparts.
7016 This work has also resulted in a variety of "const"ifications of
7017 lots of the code, especially "_cmp" operations which should normally
7018 be prototyped with "const" parameters anyway.
7021 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7022 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7023 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7024 where all of 'md' is used each time the PRNG is used, but 'state'
7025 is used only indexed by a cyclic counter. As entropy may not be
7026 well distributed from the beginning, 'md' is important as a
7027 chaining variable. However, the output function chains only half
7028 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7029 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7030 in all of 'state' being rewritten, with the new values depending
7031 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7034 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7035 the handshake is continued after ssl_verify_cert_chain();
7036 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7037 can lead to 'unexplainable' connection aborts later.
7038 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7040 *) Major EVP API cipher revision.
7041 Add hooks for extra EVP features. This allows various cipher
7042 parameters to be set in the EVP interface. Support added for variable
7043 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7044 setting of RC2 and RC5 parameters.
7046 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7049 Remove lots of duplicated code from the EVP library. For example *every*
7050 cipher init() function handles the 'iv' in the same way according to the
7051 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7052 for CFB and OFB modes they zero ctx->num.
7054 New functionality allows removal of S/MIME code RC2 hack.
7056 Most of the routines have the same form and so can be declared in terms
7059 By shifting this to the top level EVP_CipherInit() it can be removed from
7060 all individual ciphers. If the cipher wants to handle IVs or keys
7061 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7064 Change lots of functions like EVP_EncryptUpdate() to now return a
7065 value: although software versions of the algorithms cannot fail
7066 any installed hardware versions can.
7069 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7070 this option is set, tolerate broken clients that send the negotiated
7071 protocol version number instead of the requested protocol version
7075 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7076 i.e. non-zero for export ciphersuites, zero otherwise.
7077 Previous versions had this flag inverted, inconsistent with
7078 rsa_tmp_cb (..._TMP_RSA_CB).
7079 [Bodo Moeller; problem reported by Amit Chopra]
7081 *) Add missing DSA library text string. Work around for some IIS
7082 key files with invalid SEQUENCE encoding.
7085 *) Add a document (doc/standards.txt) that list all kinds of standards
7086 and so on that are implemented in OpenSSL.
7089 *) Enhance c_rehash script. Old version would mishandle certificates
7090 with the same subject name hash and wouldn't handle CRLs at all.
7091 Added -fingerprint option to crl utility, to support new c_rehash
7095 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7098 *) Fix for SSL server purpose checking. Server checking was
7099 rejecting certificates which had extended key usage present
7100 but no ssl client purpose.
7101 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7103 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7104 is a little unclear about how a blank password is handled.
7105 Since the password in encoded as a BMPString with terminating
7106 double NULL a zero length password would end up as just the
7107 double NULL. However no password at all is different and is
7108 handled differently in the PKCS#12 key generation code. NS
7109 treats a blank password as zero length. MSIE treats it as no
7110 password on export: but it will try both on import. We now do
7111 the same: PKCS12_parse() tries zero length and no password if
7112 the password is set to "" or NULL (NULL is now a valid password:
7113 it wasn't before) as does the pkcs12 application.
7116 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7117 perror when PEM_read_bio_X509_REQ fails, the error message must
7118 be obtained from the error queue.
7121 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7122 it in ERR_remove_state if appropriate, and change ERR_get_state
7123 accordingly to avoid race conditions (this is necessary because
7124 thread_hash is no longer constant once set).
7127 *) Bugfix for linux-elf makefile.one.
7130 *) RSA_get_default_method() will now cause a default
7131 RSA_METHOD to be chosen if one doesn't exist already.
7132 Previously this was only set during a call to RSA_new()
7133 or RSA_new_method(NULL) meaning it was possible for
7134 RSA_get_default_method() to return NULL.
7137 *) Added native name translation to the existing DSO code
7138 that will convert (if the flag to do so is set) filenames
7139 that are sufficiently small and have no path information
7140 into a canonical native form. Eg. "blah" converted to
7141 "libblah.so" or "blah.dll" etc.
7144 *) New function ERR_error_string_n(e, buf, len) which is like
7145 ERR_error_string(e, buf), but writes at most 'len' bytes
7146 including the 0 terminator. For ERR_error_string_n, 'buf'
7148 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7150 *) CONF library reworked to become more general. A new CONF
7151 configuration file reader "class" is implemented as well as a
7152 new functions (NCONF_*, for "New CONF") to handle it. The now
7153 old CONF_* functions are still there, but are reimplemented to
7154 work in terms of the new functions. Also, a set of functions
7155 to handle the internal storage of the configuration data is
7156 provided to make it easier to write new configuration file
7157 reader "classes" (I can definitely see something reading a
7158 configuration file in XML format, for example), called _CONF_*,
7159 or "the configuration storage API"...
7161 The new configuration file reading functions are:
7163 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7164 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7166 NCONF_default, NCONF_WIN32
7168 NCONF_dump_fp, NCONF_dump_bio
7170 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7171 NCONF_new creates a new CONF object. This works in the same way
7172 as other interfaces in OpenSSL, like the BIO interface.
7173 NCONF_dump_* dump the internal storage of the configuration file,
7174 which is useful for debugging. All other functions take the same
7175 arguments as the old CONF_* functions wth the exception of the
7176 first that must be a `CONF *' instead of a `LHASH *'.
7178 To make it easer to use the new classes with the old CONF_* functions,
7179 the function CONF_set_default_method is provided.
7182 *) Add '-tls1' option to 'openssl ciphers', which was already
7183 mentioned in the documentation but had not been implemented.
7184 (This option is not yet really useful because even the additional
7185 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7188 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7189 OpenSSL-based applications) load shared libraries and bind to
7190 them in a portable way.
7191 [Geoff Thorpe, with contributions from Richard Levitte]
7193 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7195 *) Make sure _lrotl and _lrotr are only used with MSVC.
7197 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7198 (the default implementation of RAND_status).
7200 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7201 to '-clrext' (= clear extensions), as intended and documented.
7202 [Bodo Moeller; inconsistency pointed out by Michael Attili
7205 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7206 was larger than the MD block size.
7207 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7209 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7210 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7211 using the passed key: if the passed key was a private key the result
7212 of X509_print(), for example, would be to print out all the private key
7216 *) des_quad_cksum() byte order bug fix.
7217 [Ulf Möller, using the problem description in krb4-0.9.7, where
7218 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7220 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7222 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7224 *) For easily testing in shell scripts whether some command
7225 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7226 returns with exit code 0 iff no command of the given name is available.
7227 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7228 the output goes to stdout and nothing is printed to stderr.
7229 Additional arguments are always ignored.
7231 Since for each cipher there is a command of the same name,
7232 the 'no-cipher' compilation switches can be tested this way.
7234 ('openssl no-XXX' is not able to detect pseudo-commands such
7235 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7238 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7241 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7242 is set; it will be thrown away anyway because each handshake creates
7244 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7245 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7246 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7247 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7250 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7251 'Q' and 'R' lose their special meanings (quit/renegotiate).
7252 This is part of what -quiet does; unlike -quiet, -ign_eof
7253 does not suppress any output.
7256 *) Add compatibility options to the purpose and trust code. The
7257 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7258 accepts a certificate or CA, this was the previous behaviour,
7259 with all the associated security issues.
7261 X509_TRUST_COMPAT is the old trust behaviour: only and
7262 automatically trust self signed roots in certificate store. A
7263 new trust setting X509_TRUST_DEFAULT is used to specify that
7264 a purpose has no associated trust setting and it should instead
7265 use the value in the default purpose.
7268 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7269 and fix a memory leak.
7272 *) In util/mkerr.pl (which implements 'make errors'), preserve
7273 reason strings from the previous version of the .c file, as
7274 the default to have only downcase letters (and digits) in
7275 automatically generated reasons codes is not always appropriate.
7278 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7279 using strerror. Previously, ERR_reason_error_string() returned
7280 library names as reason strings for SYSerr; but SYSerr is a special
7281 case where small numbers are errno values, not library numbers.
7284 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7285 converts DSA parameters into DH parameters. (When creating parameters,
7286 DSA_generate_parameters is used.)
7289 *) Include 'length' (recommended exponent length) in C code generated
7290 by 'openssl dhparam -C'.
7293 *) The second argument to set_label in perlasm was already being used
7294 so couldn't be used as a "file scope" flag. Moved to third argument
7298 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7299 instead of RAND_bytes for encryption IVs and salts.
7302 *) Include RAND_status() into RAND_METHOD instead of implementing
7303 it only for md_rand.c Otherwise replacing the PRNG by calling
7304 RAND_set_rand_method would be impossible.
7307 *) Don't let DSA_generate_key() enter an infinite loop if the random
7308 number generation fails.
7311 *) New 'rand' application for creating pseudo-random output.
7314 *) Added configuration support for Linux/IA64
7315 [Rolf Haberrecker <rolf@suse.de>]
7317 *) Assembler module support for Mingw32.
7320 *) Shared library support for HPUX (in shlib/).
7321 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7323 *) Shared library support for Solaris gcc.
7324 [Lutz Behnke <behnke@trustcenter.de>]
7326 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7328 *) PKCS7_encrypt() was adding text MIME headers twice because they
7329 were added manually and by SMIME_crlf_copy().
7332 *) In bntest.c don't call BN_rand with zero bits argument.
7333 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7335 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7336 case was implemented. This caused BN_div_recp() to fail occasionally.
7339 *) Add an optional second argument to the set_label() in the perl
7340 assembly language builder. If this argument exists and is set
7341 to 1 it signals that the assembler should use a symbol whose
7342 scope is the entire file, not just the current function. This
7343 is needed with MASM which uses the format label:: for this scope.
7344 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7346 *) Change the ASN1 types so they are typedefs by default. Before
7347 almost all types were #define'd to ASN1_STRING which was causing
7348 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7352 *) Change names of new functions to the new get1/get0 naming
7353 convention: After 'get1', the caller owns a reference count
7354 and has to call ..._free; 'get0' returns a pointer to some
7355 data structure without incrementing reference counters.
7356 (Some of the existing 'get' functions increment a reference
7357 counter, some don't.)
7358 Similarly, 'set1' and 'add1' functions increase reference
7359 counters or duplicate objects.
7362 *) Allow for the possibility of temp RSA key generation failure:
7363 the code used to assume it always worked and crashed on failure.
7366 *) Fix potential buffer overrun problem in BIO_printf().
7367 [Ulf Möller, using public domain code by Patrick Powell; problem
7368 pointed out by David Sacerdote <das33@cornell.edu>]
7370 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7371 RAND_egd() and RAND_status(). In the command line application,
7372 the EGD socket can be specified like a seed file using RANDFILE
7376 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7377 Some CAs (e.g. Verisign) distribute certificates in this form.
7380 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7381 list to exclude them. This means that no special compilation option
7382 is needed to use anonymous DH: it just needs to be included in the
7386 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7387 EVP_MD_type. The old functionality is available in a new macro called
7388 EVP_MD_md(). Change code that uses it and update docs.
7391 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7392 where the 'void *' argument is replaced by a function pointer argument.
7393 Previously 'void *' was abused to point to functions, which works on
7394 many platforms, but is not correct. As these functions are usually
7395 called by macros defined in OpenSSL header files, most source code
7396 should work without changes.
7399 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7400 sections with information on -D... compiler switches used for
7401 compiling the library so that applications can see them. To enable
7402 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7403 must be defined. E.g.,
7404 #define OPENSSL_ALGORITHM_DEFINES
7405 #include <openssl/opensslconf.h>
7406 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7407 [Richard Levitte, Ulf and Bodo Möller]
7409 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7413 *) Change the 'other' type in certificate aux info to a STACK_OF
7414 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7415 the required ASN1 format: arbitrary types determined by an OID.
7418 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7419 argument to 'req'. This is not because the function is newer or
7420 better than others it just uses the work 'NEW' in the certificate
7421 request header lines. Some software needs this.
7424 *) Reorganise password command line arguments: now passwords can be
7425 obtained from various sources. Delete the PEM_cb function and make
7426 it the default behaviour: i.e. if the callback is NULL and the
7427 usrdata argument is not NULL interpret it as a null terminated pass
7428 phrase. If usrdata and the callback are NULL then the pass phrase
7429 is prompted for as usual.
7432 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7433 the support is automatically enabled. The resulting binaries will
7434 autodetect the card and use it if present.
7435 [Ben Laurie and Compaq Inc.]
7437 *) Work around for Netscape hang bug. This sends certificate request
7438 and server done in one record. Since this is perfectly legal in the
7439 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7440 the bugs/SSLv3 entry for more info.
7443 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7446 *) Add -rand argument to smime and pkcs12 applications and read/write
7450 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7453 *) Add command line password options to the remaining applications.
7456 *) Bug fix for BN_div_recp() for numerators with an even number of
7460 *) More tests in bntest.c, and changed test_bn output.
7463 *) ./config recognizes MacOS X now.
7466 *) Bug fix for BN_div() when the first words of num and divsor are
7467 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7470 *) Add support for various broken PKCS#8 formats, and command line
7471 options to produce them.
7474 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7475 get temporary BIGNUMs from a BN_CTX.
7478 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7482 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7483 include a #define from the old name to the new. The original intent
7484 was that statically linked binaries could for example just call
7485 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7486 link with digests. This never worked becayse SSLeay_add_all_digests()
7487 and SSLeay_add_all_ciphers() were in the same source file so calling
7488 one would link with the other. They are now in separate source files.
7491 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7494 *) Use a less unusual form of the Miller-Rabin primality test (it used
7495 a binary algorithm for exponentiation integrated into the Miller-Rabin
7496 loop, our standard modexp algorithms are faster).
7499 *) Support for the EBCDIC character set completed.
7500 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7502 *) Source code cleanups: use const where appropriate, eliminate casts,
7503 use void * instead of char * in lhash.
7506 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7507 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7508 this the server could overwrite ephemeral keys that the client
7512 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7513 using 50 iterations of the Rabin-Miller test.
7515 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7516 iterations of the Rabin-Miller test as required by the appendix
7517 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7518 As BN_is_prime_fasttest includes trial division, DSA parameter
7519 generation becomes much faster.
7521 This implies a change for the callback functions in DSA_is_prime
7522 and DSA_generate_parameters: The callback function is called once
7523 for each positive witness in the Rabin-Miller test, not just
7524 occasionally in the inner loop; and the parameters to the
7525 callback function now provide an iteration count for the outer
7526 loop rather than for the current invocation of the inner loop.
7527 DSA_generate_parameters additionally can call the callback
7528 function with an 'iteration count' of -1, meaning that a
7529 candidate has passed the trial division test (when q is generated
7530 from an application-provided seed, trial division is skipped).
7533 *) New function BN_is_prime_fasttest that optionally does trial
7534 division before starting the Rabin-Miller test and has
7535 an additional BN_CTX * argument (whereas BN_is_prime always
7536 has to allocate at least one BN_CTX).
7537 'callback(1, -1, cb_arg)' is called when a number has passed the
7538 trial division stage.
7541 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7545 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7548 *) New function BN_pseudo_rand().
7551 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7552 bignum version of BN_from_montgomery() with the working code from
7553 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7557 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7558 made it impossible to use the same SSL_SESSION data structure in
7559 SSL2 clients in multiple threads.
7562 *) The return value of RAND_load_file() no longer counts bytes obtained
7563 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7564 to seed the PRNG (previously an explicit byte count was required).
7565 [Ulf Möller, Bodo Möller]
7567 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7568 used (char *) instead of (void *) and had casts all over the place.
7571 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7574 *) Retain source code compatibility for BN_prime_checks macro:
7575 BN_is_prime(..., BN_prime_checks, ...) now uses
7576 BN_prime_checks_for_size to determine the appropriate number of
7577 Rabin-Miller iterations.
7580 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7581 DH_CHECK_P_NOT_SAFE_PRIME.
7582 (Check if this is true? OpenPGP calls them "strong".)
7585 *) Merge the functionality of "dh" and "gendh" programs into a new program
7586 "dhparam". The old programs are retained for now but will handle DH keys
7587 (instead of parameters) in future.
7590 *) Make the ciphers, s_server and s_client programs check the return values
7591 when a new cipher list is set.
7594 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7595 ciphers. Before when the 56bit ciphers were enabled the sorting was
7598 The syntax for the cipher sorting has been extended to support sorting by
7599 cipher-strength (using the strength_bits hard coded in the tables).
7600 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7602 Fix a bug in the cipher-command parser: when supplying a cipher command
7603 string with an "undefined" symbol (neither command nor alphanumeric
7604 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7605 an error is flagged.
7607 Due to the strength-sorting extension, the code of the
7608 ssl_create_cipher_list() function was completely rearranged. I hope that
7609 the readability was also increased :-)
7610 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7612 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7613 for the first serial number and places 2 in the serial number file. This
7614 avoids problems when the root CA is created with serial number zero and
7615 the first user certificate has the same issuer name and serial number
7619 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7620 the new code. Add documentation for this stuff.
7623 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7624 X509_*() to X509at_*() on the grounds that they don't handle X509
7625 structures and behave in an analagous way to the X509v3 functions:
7626 they shouldn't be called directly but wrapper functions should be used
7629 So we also now have some wrapper functions that call the X509at functions
7630 when passed certificate requests. (TO DO: similar things can be done with
7631 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7632 things. Some of these need some d2i or i2d and print functionality
7633 because they handle more complex structures.)
7636 *) Add missing #ifndefs that caused missing symbols when building libssl
7637 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7638 NO_RSA in ssl/s2*.c.
7639 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7641 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7642 has a return value which indicates the quality of the random data
7643 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7644 error queue. New function RAND_pseudo_bytes() generates output that is
7645 guaranteed to be unique but not unpredictable. RAND_add is like
7646 RAND_seed, but takes an extra argument for an entropy estimate
7647 (RAND_seed always assumes full entropy).
7650 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7651 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7652 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7653 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7654 false-positive rate of at most 2^-80 for random input.
7657 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7660 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7661 in the 0.9.5 release), this returns the chain
7662 from an X509_CTX structure with a dup of the stack and all
7663 the X509 reference counts upped: so the stack will exist
7664 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7667 Also make SSL_SESSION_print() print out the verify return
7671 *) Add manpage for the pkcs12 command. Also change the default
7672 behaviour so MAC iteration counts are used unless the new
7673 -nomaciter option is used. This improves file security and
7674 only older versions of MSIE (4.0 for example) need it.
7677 *) Honor the no-xxx Configure options when creating .DEF files.
7680 *) Add PKCS#10 attributes to field table: challengePassword,
7681 unstructuredName and unstructuredAddress. These are taken from
7682 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7683 international characters are used.
7685 More changes to X509_ATTRIBUTE code: allow the setting of types
7686 based on strings. Remove the 'loc' parameter when adding
7687 attributes because these will be a SET OF encoding which is sorted
7691 *) Initial changes to the 'req' utility to allow request generation
7692 automation. This will allow an application to just generate a template
7693 file containing all the field values and have req construct the
7696 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7697 used all over the place including certificate requests and PKCS#7
7698 structures. They are currently handled manually where necessary with
7699 some primitive wrappers for PKCS#7. The new functions behave in a
7700 manner analogous to the X509 extension functions: they allow
7701 attributes to be looked up by NID and added.
7703 Later something similar to the X509V3 code would be desirable to
7704 automatically handle the encoding, decoding and printing of the
7705 more complex types. The string types like challengePassword can
7706 be handled by the string table functions.
7708 Also modified the multi byte string table handling. Now there is
7709 a 'global mask' which masks out certain types. The table itself
7710 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7711 is useful when for example there is only one permissible type
7712 (as in countryName) and using the mask might result in no valid
7716 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7717 SSL_get_peer_finished to allow applications to obtain the latest
7718 Finished messages sent to the peer or expected from the peer,
7719 respectively. (SSL_get_peer_finished is usually the Finished message
7720 actually received from the peer, otherwise the protocol will be aborted.)
7722 As the Finished message are message digests of the complete handshake
7723 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7724 be used for external authentication procedures when the authentication
7725 provided by SSL/TLS is not desired or is not enough.
7728 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7729 the host supports BWX extension and if Compaq C is present on the
7730 $PATH. Just exploiting of the BWX extension results in 20-30%
7731 performance kick for some algorithms, e.g. DES and RC4 to mention
7732 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7736 *) Add support for MS "fast SGC". This is arguably a violation of the
7737 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7738 weak crypto and after checking the certificate is SGC a second one
7739 with strong crypto. MS SGC stops the first handshake after receiving
7740 the server certificate message and sends a second client hello. Since
7741 a server will typically do all the time consuming operations before
7742 expecting any further messages from the client (server key exchange
7743 is the most expensive) there is little difference between the two.
7745 To get OpenSSL to support MS SGC we have to permit a second client
7746 hello message after we have sent server done. In addition we have to
7747 reset the MAC if we do get this second client hello.
7750 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7751 if a DER encoded private key is RSA or DSA traditional format. Changed
7752 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7753 format DER encoded private key. Newer code should use PKCS#8 format which
7754 has the key type encoded in the ASN1 structure. Added DER private key
7755 support to pkcs8 application.
7758 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7759 ciphersuites has been selected (as required by the SSL 3/TLS 1
7760 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7761 is set, we interpret this as a request to violate the specification
7762 (the worst that can happen is a handshake failure, and 'correct'
7763 behaviour would result in a handshake failure anyway).
7766 *) In SSL_CTX_add_session, take into account that there might be multiple
7767 SSL_SESSION structures with the same session ID (e.g. when two threads
7768 concurrently obtain them from an external cache).
7769 The internal cache can handle only one SSL_SESSION with a given ID,
7770 so if there's a conflict, we now throw out the old one to achieve
7774 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7775 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7776 some routines that use cipher OIDs: some ciphers do not have OIDs
7777 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7781 *) Simplify the trust setting structure and code. Now we just have
7782 two sequences of OIDs for trusted and rejected settings. These will
7783 typically have values the same as the extended key usage extension
7784 and any application specific purposes.
7786 The trust checking code now has a default behaviour: it will just
7787 check for an object with the same NID as the passed id. Functions can
7788 be provided to override either the default behaviour or the behaviour
7789 for a given id. SSL client, server and email already have functions
7790 in place for compatibility: they check the NID and also return "trusted"
7791 if the certificate is self signed.
7794 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7795 traditional format into an EVP_PKEY structure.
7798 *) Add a password callback function PEM_cb() which either prompts for
7799 a password if usr_data is NULL or otherwise assumes it is a null
7800 terminated password. Allow passwords to be passed on command line
7801 environment or config files in a few more utilities.
7804 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7805 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7806 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7807 Update documentation.
7810 *) Support for ASN1 "NULL" type. This could be handled before by using
7811 ASN1_TYPE but there wasn't any function that would try to read a NULL
7812 and produce an error if it couldn't. For compatibility we also have
7813 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7814 don't allocate anything because they don't need to.
7817 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7819 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7821 *) Rebuild of the memory allocation routines used by OpenSSL code and
7822 possibly others as well. The purpose is to make an interface that
7823 provide hooks so anyone can build a separate set of allocation and
7824 deallocation routines to be used by OpenSSL, for example memory
7825 pool implementations, or something else, which was previously hard
7826 since Malloc(), Realloc() and Free() were defined as macros having
7827 the values malloc, realloc and free, respectively (except for Win32
7828 compilations). The same is provided for memory debugging code.
7829 OpenSSL already comes with functionality to find memory leaks, but
7830 this gives people a chance to debug other memory problems.
7832 With these changes, a new set of functions and macros have appeared:
7834 CRYPTO_set_mem_debug_functions() [F]
7835 CRYPTO_get_mem_debug_functions() [F]
7836 CRYPTO_dbg_set_options() [F]
7837 CRYPTO_dbg_get_options() [F]
7838 CRYPTO_malloc_debug_init() [M]
7840 The memory debug functions are NULL by default, unless the library
7841 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7842 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7843 gives the standard debugging functions that come with OpenSSL) or
7844 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7845 provided by the library user) must be used. When the standard
7846 debugging functions are used, CRYPTO_dbg_set_options can be used to
7847 request additional information:
7848 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7849 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7851 Also, things like CRYPTO_set_mem_functions will always give the
7852 expected result (the new set of functions is used for allocation
7853 and deallocation) at all times, regardless of platform and compiler
7856 To finish it up, some functions that were never use in any other
7857 way than through macros have a new API and new semantic:
7860 CRYPTO_dbg_realloc()
7863 All macros of value have retained their old syntax.
7864 [Richard Levitte and Bodo Moeller]
7866 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7867 ordering of SMIMECapabilities wasn't in "strength order" and there
7868 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7872 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7873 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7874 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7876 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7877 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7878 functionality to handle multipart/signed properly) and a utility
7879 called 'smime' to call all this stuff. This is based on code I
7880 originally wrote for Celo who have kindly allowed it to be
7881 included in OpenSSL.
7884 *) Add variants des_set_key_checked and des_set_key_unchecked of
7885 des_set_key (aka des_key_sched). Global variable des_check_key
7886 decides which of these is called by des_set_key; this way
7887 des_check_key behaves as it always did, but applications and
7888 the library itself, which was buggy for des_check_key == 1,
7889 have a cleaner way to pick the version they need.
7892 *) New function PKCS12_newpass() which changes the password of a
7896 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7897 dynamic mix. In both cases the ids can be used as an index into the
7898 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7899 functions so they accept a list of the field values and the
7900 application doesn't need to directly manipulate the X509_TRUST
7904 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7908 *) Modify the way the V3 extension code looks up extensions. This now
7909 works in a similar way to the object code: we have some "standard"
7910 extensions in a static table which is searched with OBJ_bsearch()
7911 and the application can add dynamic ones if needed. The file
7912 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7913 updated whenever a new extension is added to the core code and kept
7914 in ext_nid order. There is a simple program 'tabtest.c' which checks
7915 this. New extensions are not added too often so this file can readily
7916 be maintained manually.
7918 There are two big advantages in doing things this way. The extensions
7919 can be looked up immediately and no longer need to be "added" using
7920 X509V3_add_standard_extensions(): this function now does nothing.
7921 [Side note: I get *lots* of email saying the extension code doesn't
7922 work because people forget to call this function]
7923 Also no dynamic allocation is done unless new extensions are added:
7924 so if we don't add custom extensions there is no need to call
7925 X509V3_EXT_cleanup().
7928 *) Modify enc utility's salting as follows: make salting the default. Add a
7929 magic header, so unsalted files fail gracefully instead of just decrypting
7930 to garbage. This is because not salting is a big security hole, so people
7931 should be discouraged from doing it.
7934 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7935 digest to be passed on the command line but it only used this
7936 parameter when signing a certificate. Modified so all relevant
7937 operations are affected by the digest parameter including the
7938 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7939 DSA key was used because it didn't fix the digest.
7942 *) Initial certificate chain verify code. Currently tests the untrusted
7943 certificates for consistency with the verify purpose (which is set
7944 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7946 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7947 this is because it will reject chains with invalid extensions whereas
7948 every previous version of OpenSSL and SSLeay made no checks at all.
7950 Trust code: checks the root CA for the relevant trust settings. Trust
7951 settings have an initial value consistent with the verify purpose: e.g.
7952 if the verify purpose is for SSL client use it expects the CA to be
7953 trusted for SSL client use. However the default value can be changed to
7954 permit custom trust settings: one example of this would be to only trust
7955 certificates from a specific "secure" set of CAs.
7957 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7958 which should be used for version portability: especially since the
7959 verify structure is likely to change more often now.
7961 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7962 to set them. If not set then assume SSL clients will verify SSL servers
7965 Two new options to the verify program: -untrusted allows a set of
7966 untrusted certificates to be passed in and -purpose which sets the
7967 intended purpose of the certificate. If a purpose is set then the
7968 new chain verify code is used to check extension consistency.
7971 *) Support for the authority information access extension.
7974 *) Modify RSA and DSA PEM read routines to transparently handle
7975 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7976 public keys in a format compatible with certificate
7977 SubjectPublicKeyInfo structures. Unfortunately there were already
7978 functions called *_PublicKey_* which used various odd formats so
7979 these are retained for compatibility: however the DSA variants were
7980 never in a public release so they have been deleted. Changed dsa/rsa
7981 utilities to handle the new format: note no releases ever handled public
7982 keys so we should be OK.
7984 The primary motivation for this change is to avoid the same fiasco
7985 that dogs private keys: there are several incompatible private key
7986 formats some of which are standard and some OpenSSL specific and
7987 require various evil hacks to allow partial transparent handling and
7988 even then it doesn't work with DER formats. Given the option anything
7989 other than PKCS#8 should be dumped: but the other formats have to
7990 stay in the name of compatibility.
7992 With public keys and the benefit of hindsight one standard format
7993 is used which works with EVP_PKEY, RSA or DSA structures: though
7994 it clearly returns an error if you try to read the wrong kind of key.
7996 Added a -pubkey option to the 'x509' utility to output the public key.
7997 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7998 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7999 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8000 that do the same as the EVP_PKEY_assign_*() except they up the
8001 reference count of the added key (they don't "swallow" the
8005 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8006 CRLs would fail if the file contained no certificates or no CRLs:
8007 added a new function to read in both types and return the number
8008 read: this means that if none are read it will be an error. The
8009 DER versions of the certificate and CRL reader would always fail
8010 because it isn't possible to mix certificates and CRLs in DER format
8011 without choking one or the other routine. Changed this to just read
8012 a certificate: this is the best we can do. Also modified the code
8013 in apps/verify.c to take notice of return codes: it was previously
8014 attempting to read in certificates from NULL pointers and ignoring
8015 any errors: this is one reason why the cert and CRL reader seemed
8016 to work. It doesn't check return codes from the default certificate
8017 routines: these may well fail if the certificates aren't installed.
8020 *) Code to support otherName option in GeneralName.
8023 *) First update to verify code. Change the verify utility
8024 so it warns if it is passed a self signed certificate:
8025 for consistency with the normal behaviour. X509_verify
8026 has been modified to it will now verify a self signed
8027 certificate if *exactly* the same certificate appears
8028 in the store: it was previously impossible to trust a
8029 single self signed certificate. This means that:
8030 openssl verify ss.pem
8031 now gives a warning about a self signed certificate but
8032 openssl verify -CAfile ss.pem ss.pem
8036 *) For servers, store verify_result in SSL_SESSION data structure
8037 (and add it to external session representation).
8038 This is needed when client certificate verifications fails,
8039 but an application-provided verification callback (set by
8040 SSL_CTX_set_cert_verify_callback) allows accepting the session
8041 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8042 but returns 1): When the session is reused, we have to set
8043 ssl->verify_result to the appropriate error code to avoid
8045 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8047 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8048 case in PKCS7_dataInit() where the signed PKCS7 structure
8049 didn't contain any existing data because it was being created.
8050 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8052 *) Add a salt to the key derivation routines in enc.c. This
8053 forms the first 8 bytes of the encrypted file. Also add a
8054 -S option to allow a salt to be input on the command line.
8057 *) New function X509_cmp(). Oddly enough there wasn't a function
8058 to compare two certificates. We do this by working out the SHA1
8059 hash and comparing that. X509_cmp() will be needed by the trust
8063 *) SSL_get1_session() is like SSL_get_session(), but increments
8064 the reference count in the SSL_SESSION returned.
8065 [Geoff Thorpe <geoff@eu.c2.net>]
8067 *) Fix for 'req': it was adding a null to request attributes.
8068 Also change the X509_LOOKUP and X509_INFO code to handle
8069 certificate auxiliary information.
8072 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8076 *) Add the possibility to add extra information to the memory leak
8077 detecting output, to form tracebacks, showing from where each
8078 allocation was originated: CRYPTO_push_info("constant string") adds
8079 the string plus current file name and line number to a per-thread
8080 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8081 is like calling CYRPTO_pop_info() until the stack is empty.
8082 Also updated memory leak detection code to be multi-thread-safe.
8085 *) Add options -text and -noout to pkcs7 utility and delete the
8086 encryption options which never did anything. Update docs.
8089 *) Add options to some of the utilities to allow the pass phrase
8090 to be included on either the command line (not recommended on
8091 OSes like Unix) or read from the environment. Update the
8092 manpages and fix a few bugs.
8095 *) Add a few manpages for some of the openssl commands.
8098 *) Fix the -revoke option in ca. It was freeing up memory twice,
8099 leaking and not finding already revoked certificates.
8102 *) Extensive changes to support certificate auxiliary information.
8103 This involves the use of X509_CERT_AUX structure and X509_AUX
8104 functions. An X509_AUX function such as PEM_read_X509_AUX()
8105 can still read in a certificate file in the usual way but it
8106 will also read in any additional "auxiliary information". By
8107 doing things this way a fair degree of compatibility can be
8108 retained: existing certificates can have this information added
8109 using the new 'x509' options.
8111 Current auxiliary information includes an "alias" and some trust
8112 settings. The trust settings will ultimately be used in enhanced
8113 certificate chain verification routines: currently a certificate
8114 can only be trusted if it is self signed and then it is trusted
8118 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8119 The problem was that one of the replacement routines had not been working
8120 since SSLeay releases. For now the offending routine has been replaced
8121 with non-optimised assembler. Even so, this now gives around 95%
8122 performance improvement for 1024 bit RSA signs.
8125 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8126 handling. Most clients have the effective key size in bits equal to
8127 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8128 A few however don't do this and instead use the size of the decrypted key
8129 to determine the RC2 key length and the AlgorithmIdentifier to determine
8130 the effective key length. In this case the effective key length can still
8131 be 40 bits but the key length can be 168 bits for example. This is fixed
8132 by manually forcing an RC2 key into the EVP_PKEY structure because the
8133 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8134 the key length and effective key length are equal.
8137 *) Add a bunch of functions that should simplify the creation of
8138 X509_NAME structures. Now you should be able to do:
8139 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8140 and have it automatically work out the correct field type and fill in
8141 the structures. The more adventurous can try:
8142 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8143 and it will (hopefully) work out the correct multibyte encoding.
8146 *) Change the 'req' utility to use the new field handling and multibyte
8147 copy routines. Before the DN field creation was handled in an ad hoc
8148 way in req, ca, and x509 which was rather broken and didn't support
8149 BMPStrings or UTF8Strings. Since some software doesn't implement
8150 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8151 using the dirstring_type option. See the new comment in the default
8152 openssl.cnf for more info.
8155 *) Make crypto/rand/md_rand.c more robust:
8156 - Assure unique random numbers after fork().
8157 - Make sure that concurrent threads access the global counter and
8158 md serializably so that we never lose entropy in them
8159 or use exactly the same state in multiple threads.
8160 Access to the large state is not always serializable because
8161 the additional locking could be a performance killer, and
8162 md should be large enough anyway.
8165 *) New file apps/app_rand.c with commonly needed functionality
8166 for handling the random seed file.
8168 Use the random seed file in some applications that previously did not:
8170 dsaparam -genkey (which also ignored its '-rand' option),
8173 x509 (when signing).
8174 Except on systems with /dev/urandom, it is crucial to have a random
8175 seed file at least for key creation, DSA signing, and for DH exchanges;
8176 for RSA signatures we could do without one.
8178 gendh and gendsa (unlike genrsa) used to read only the first byte
8179 of each file listed in the '-rand' option. The function as previously
8180 found in genrsa is now in app_rand.c and is used by all programs
8181 that support '-rand'.
8184 *) In RAND_write_file, use mode 0600 for creating files;
8185 don't just chmod when it may be too late.
8188 *) Report an error from X509_STORE_load_locations
8189 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8192 *) New function ASN1_mbstring_copy() this copies a string in either
8193 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8194 into an ASN1_STRING type. A mask of permissible types is passed
8195 and it chooses the "minimal" type to use or an error if not type
8199 *) Add function equivalents to the various macros in asn1.h. The old
8200 macros are retained with an M_ prefix. Code inside the library can
8201 use the M_ macros. External code (including the openssl utility)
8202 should *NOT* in order to be "shared library friendly".
8205 *) Add various functions that can check a certificate's extensions
8206 to see if it usable for various purposes such as SSL client,
8207 server or S/MIME and CAs of these types. This is currently
8208 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8209 verification. Also added a -purpose flag to x509 utility to
8210 print out all the purposes.
8213 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8217 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8218 for, obtain and decode and extension and obtain its critical flag.
8219 This allows all the necessary extension code to be handled in a
8220 single function call.
8223 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8224 platforms. See crypto/rc4/rc4_enc.c for further details.
8227 *) New -noout option to asn1parse. This causes no output to be produced
8228 its main use is when combined with -strparse and -out to extract data
8229 from a file (which may not be in ASN.1 format).
8232 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8233 when producing the local key id.
8234 [Richard Levitte <levitte@stacken.kth.se>]
8236 *) New option -dhparam in s_server. This allows a DH parameter file to be
8237 stated explicitly. If it is not stated then it tries the first server
8238 certificate file. The previous behaviour hard coded the filename
8242 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8243 a public key to be input or output. For example:
8244 openssl rsa -in key.pem -pubout -out pubkey.pem
8245 Also added necessary DSA public key functions to handle this.
8248 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8249 in the message. This was handled by allowing
8250 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8251 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8253 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8254 to the end of the strings whereas this didn't. This would cause problems
8255 if strings read with d2i_ASN1_bytes() were later modified.
8256 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8258 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8259 data and it contains EOF it will end up returning an error. This is
8260 caused by input 46 bytes long. The cause is due to the way base64
8261 BIOs find the start of base64 encoded data. They do this by trying a
8262 trial decode on each line until they find one that works. When they
8263 do a flag is set and it starts again knowing it can pass all the
8264 data directly through the decoder. Unfortunately it doesn't reset
8265 the context it uses. This means that if EOF is reached an attempt
8266 is made to pass two EOFs through the context and this causes the
8267 resulting error. This can also cause other problems as well. As is
8268 usual with these problems it takes *ages* to find and the fix is
8269 trivial: move one line.
8270 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8272 *) Ugly workaround to get s_client and s_server working under Windows. The
8273 old code wouldn't work because it needed to select() on sockets and the
8274 tty (for keypresses and to see if data could be written). Win32 only
8275 supports select() on sockets so we select() with a 1s timeout on the
8276 sockets and then see if any characters are waiting to be read, if none
8277 are present then we retry, we also assume we can always write data to
8278 the tty. This isn't nice because the code then blocks until we've
8279 received a complete line of data and it is effectively polling the
8280 keyboard at 1s intervals: however it's quite a bit better than not
8281 working at all :-) A dedicated Windows application might handle this
8282 with an event loop for example.
8285 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8286 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8287 will be called when RSA_sign() and RSA_verify() are used. This is useful
8288 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8289 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8290 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8291 This necessitated the support of an extra signature type NID_md5_sha1
8292 for SSL signatures and modifications to the SSL library to use it instead
8293 of calling RSA_public_decrypt() and RSA_private_encrypt().
8296 *) Add new -verify -CAfile and -CApath options to the crl program, these
8297 will lookup a CRL issuers certificate and verify the signature in a
8298 similar way to the verify program. Tidy up the crl program so it
8299 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8300 less strict. It will now permit CRL extensions even if it is not
8301 a V2 CRL: this will allow it to tolerate some broken CRLs.
8304 *) Initialize all non-automatic variables each time one of the openssl
8305 sub-programs is started (this is necessary as they may be started
8306 multiple times from the "OpenSSL>" prompt).
8307 [Lennart Bang, Bodo Moeller]
8309 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8310 removing all other RSA functionality (this is what NO_RSA does). This
8311 is so (for example) those in the US can disable those operations covered
8312 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8316 *) Non-copying interface to BIO pairs.
8317 (still largely untested)
8320 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8321 ASCII string. This was handled independently in various places before.
8324 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8325 UTF8 strings a character at a time.
8328 *) Use client_version from client hello to select the protocol
8329 (s23_srvr.c) and for RSA client key exchange verification
8330 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8333 *) Add various utility functions to handle SPKACs, these were previously
8334 handled by poking round in the structure internals. Added new function
8335 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8336 print, verify and generate SPKACs. Based on an original idea from
8337 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8340 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8343 *) Allow the config file extension section to be overwritten on the
8344 command line. Based on an original idea from Massimiliano Pala
8345 <madwolf@comune.modena.it>. The new option is called -extensions
8346 and can be applied to ca, req and x509. Also -reqexts to override
8347 the request extensions in req and -crlexts to override the crl extensions
8351 *) Add new feature to the SPKAC handling in ca. Now you can include
8352 the same field multiple times by preceding it by "XXXX." for example:
8355 this is the same syntax as used in the req config file.
8358 *) Allow certificate extensions to be added to certificate requests. These
8359 are specified in a 'req_extensions' option of the req section of the
8360 config file. They can be printed out with the -text option to req but
8361 are otherwise ignored at present.
8364 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8365 data read consists of only the final block it would not decrypted because
8366 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8367 A misplaced 'break' also meant the decrypted final block might not be
8368 copied until the next read.
8371 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8372 a few extra parameters to the DH structure: these will be useful if
8373 for example we want the value of 'q' or implement X9.42 DH.
8376 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8377 provides hooks that allow the default DSA functions or functions on a
8378 "per key" basis to be replaced. This allows hardware acceleration and
8379 hardware key storage to be handled without major modification to the
8380 library. Also added low level modexp hooks and CRYPTO_EX structure and
8381 associated functions.
8384 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8385 as "read only": it can't be written to and the buffer it points to will
8386 not be freed. Reading from a read only BIO is much more efficient than
8387 a normal memory BIO. This was added because there are several times when
8388 an area of memory needs to be read from a BIO. The previous method was
8389 to create a memory BIO and write the data to it, this results in two
8390 copies of the data and an O(n^2) reading algorithm. There is a new
8391 function BIO_new_mem_buf() which creates a read only memory BIO from
8392 an area of memory. Also modified the PKCS#7 routines to use read only
8396 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8397 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8398 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8399 but a retry condition occured while trying to read the rest.
8402 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8403 NID_pkcs7_encrypted by default: this was wrong since this should almost
8404 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8405 the encrypted data type: this is a more sensible place to put it and it
8406 allows the PKCS#12 code to be tidied up that duplicated this
8410 *) Changed obj_dat.pl script so it takes its input and output files on
8411 the command line. This should avoid shell escape redirection problems
8415 *) Initial support for certificate extension requests, these are included
8416 in things like Xenroll certificate requests. Included functions to allow
8417 extensions to be obtained and added.
8420 *) -crlf option to s_client and s_server for sending newlines as
8421 CRLF (as required by many protocols).
8424 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8426 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8427 [Ralf S. Engelschall]
8429 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8430 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8432 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8436 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8437 DH parameters/keys (q is lost during that conversion, but the resulting
8438 DH parameters contain its length).
8440 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8441 much faster than DH_generate_parameters (which creates parameters
8442 where p = 2*q + 1), and also the smaller q makes DH computations
8443 much more efficient (160-bit exponentiation instead of 1024-bit
8444 exponentiation); so this provides a convenient way to support DHE
8445 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8446 utter importance to use
8447 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8449 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8450 when such DH parameters are used, because otherwise small subgroup
8451 attacks may become possible!
8454 *) Avoid memory leak in i2d_DHparams.
8457 *) Allow the -k option to be used more than once in the enc program:
8458 this allows the same encrypted message to be read by multiple recipients.
8461 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8462 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8463 it will always use the numerical form of the OID, even if it has a short
8467 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8468 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8469 otherwise bn_mod_exp was called. In the case of hardware keys for example
8470 no private key components need be present and it might store extra data
8471 in the RSA structure, which cannot be accessed from bn_mod_exp.
8472 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8473 private key operations.
8476 *) Added support for SPARC Linux.
8479 *) pem_password_cb function type incompatibly changed from
8480 typedef int pem_password_cb(char *buf, int size, int rwflag);
8482 ....(char *buf, int size, int rwflag, void *userdata);
8483 so that applications can pass data to their callbacks:
8484 The PEM[_ASN1]_{read,write}... functions and macros now take an
8485 additional void * argument, which is just handed through whenever
8486 the password callback is called.
8487 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8489 New function SSL_CTX_set_default_passwd_cb_userdata.
8491 Compatibility note: As many C implementations push function arguments
8492 onto the stack in reverse order, the new library version is likely to
8493 interoperate with programs that have been compiled with the old
8494 pem_password_cb definition (PEM_whatever takes some data that
8495 happens to be on the stack as its last argument, and the callback
8496 just ignores this garbage); but there is no guarantee whatsoever that
8499 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8500 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8501 problems not only on Windows, but also on some Unix platforms.
8502 To avoid problematic command lines, these definitions are now in an
8503 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8504 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8507 *) MIPS III/IV assembler module is reimplemented.
8510 *) More DES library cleanups: remove references to srand/rand and
8511 delete an unused file.
8514 *) Add support for the the free Netwide assembler (NASM) under Win32,
8515 since not many people have MASM (ml) and it can be hard to obtain.
8516 This is currently experimental but it seems to work OK and pass all
8517 the tests. Check out INSTALL.W32 for info.
8520 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8521 without temporary keys kept an extra copy of the server key,
8522 and connections with temporary keys did not free everything in case
8526 *) New function RSA_check_key and new openssl rsa option -check
8527 for verifying the consistency of RSA keys.
8528 [Ulf Moeller, Bodo Moeller]
8530 *) Various changes to make Win32 compile work:
8531 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8532 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8533 comparison" warnings.
8534 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8537 *) Add a debugging option to PKCS#5 v2 key generation function: when
8538 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8539 derived keys are printed to stderr.
8542 *) Copy the flags in ASN1_STRING_dup().
8543 [Roman E. Pavlov <pre@mo.msk.ru>]
8545 *) The x509 application mishandled signing requests containing DSA
8546 keys when the signing key was also DSA and the parameters didn't match.
8548 It was supposed to omit the parameters when they matched the signing key:
8549 the verifying software was then supposed to automatically use the CA's
8550 parameters if they were absent from the end user certificate.
8552 Omitting parameters is no longer recommended. The test was also
8553 the wrong way round! This was probably due to unusual behaviour in
8554 EVP_cmp_parameters() which returns 1 if the parameters match.
8555 This meant that parameters were omitted when they *didn't* match and
8556 the certificate was useless. Certificates signed with 'ca' didn't have
8558 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8560 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8561 The interface is as follows:
8562 Applications can use
8563 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8564 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8565 "off" is now the default.
8566 The library internally uses
8567 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8568 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8569 to disable memory-checking temporarily.
8571 Some inconsistent states that previously were possible (and were
8572 even the default) are now avoided.
8574 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8575 with each memory chunk allocated; this is occasionally more helpful
8576 than just having a counter.
8578 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8580 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8584 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8585 which largely parallels "options", but is for changing API behaviour,
8586 whereas "options" are about protocol behaviour.
8587 Initial "mode" flags are:
8589 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8590 a single record has been written.
8591 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8592 retries use the same buffer location.
8593 (But all of the contents must be
8597 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8600 *) Fix problems with no-hmac etc.
8601 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8603 *) New functions RSA_get_default_method(), RSA_set_method() and
8604 RSA_get_method(). These allows replacement of RSA_METHODs without having
8605 to mess around with the internals of an RSA structure.
8608 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8609 Also really enable memory leak checks in openssl.c and in some
8611 [Chad C. Mulligan, Bodo Moeller]
8613 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8614 up the length of negative integers. This has now been simplified to just
8615 store the length when it is first determined and use it later, rather
8616 than trying to keep track of where data is copied and updating it to
8618 [Steve Henson, reported by Brien Wheeler
8619 <bwheeler@authentica-security.com>]
8621 *) Add a new function PKCS7_signatureVerify. This allows the verification
8622 of a PKCS#7 signature but with the signing certificate passed to the
8623 function itself. This contrasts with PKCS7_dataVerify which assumes the
8624 certificate is present in the PKCS#7 structure. This isn't always the
8625 case: certificates can be omitted from a PKCS#7 structure and be
8626 distributed by "out of band" means (such as a certificate database).
8629 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8630 function prototypes in pem.h, also change util/mkdef.pl to add the
8631 necessary function names.
8634 *) mk1mf.pl (used by Windows builds) did not properly read the
8635 options set by Configure in the top level Makefile, and Configure
8636 was not even able to write more than one option correctly.
8637 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8640 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8641 file to be loaded from a BIO or FILE pointer. The BIO version will
8642 for example allow memory BIOs to contain config info.
8645 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8646 Whoever hopes to achieve shared-library compatibility across versions
8647 must use this, not the compile-time macro.
8648 (Exercise 0.9.4: Which is the minimum library version required by
8650 Note: All this applies only to multi-threaded programs, others don't
8654 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8655 through a BIO pair triggered the default case, i.e.
8656 SSLerr(...,SSL_R_UNKNOWN_STATE).
8659 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8660 can use the SSL library even if none of the specific BIOs is
8664 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8665 for the encoded length.
8666 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8668 *) Add initial documentation of the X509V3 functions.
8671 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8672 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8673 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8674 secure PKCS#8 private key format with a high iteration count.
8677 *) Fix determination of Perl interpreter: A perl or perl5
8678 _directory_ in $PATH was also accepted as the interpreter.
8679 [Ralf S. Engelschall]
8681 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8682 wrong with it but it was very old and did things like calling
8683 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8687 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8688 to use the new extension code.
8691 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8692 with macros. This should make it easier to change their form, add extra
8693 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8697 *) Add to configuration table a new entry that can specify an alternative
8698 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8699 according to Mark Crispin <MRC@Panda.COM>.
8703 *) DES CBC did not update the IV. Weird.
8706 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8707 Changing the behaviour of the former might break existing programs --
8708 where IV updating is needed, des_ncbc_encrypt can be used.
8711 *) When bntest is run from "make test" it drives bc to check its
8712 calculations, as well as internally checking them. If an internal check
8713 fails, it needs to cause bc to give a non-zero result or make test carries
8714 on without noticing the failure. Fixed.
8717 *) DES library cleanups.
8720 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8721 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8722 ciphers. NOTE: although the key derivation function has been verified
8723 against some published test vectors it has not been extensively tested
8724 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8728 *) Instead of "mkdir -p", which is not fully portable, use new
8729 Perl script "util/mkdir-p.pl".
8732 *) Rewrite the way password based encryption (PBE) is handled. It used to
8733 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8734 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8735 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8736 the 'parameter' field of the AlgorithmIdentifier is passed to the
8737 underlying key generation function so it must do its own ASN1 parsing.
8738 This has also changed the EVP_PBE_CipherInit() function which now has a
8739 'parameter' argument instead of literal salt and iteration count values
8740 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8743 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8744 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8745 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8746 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8747 value was just used as a "magic string" and not used directly its
8748 value doesn't matter.
8751 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8755 *) "linux-sparc64" configuration (ultrapenguin).
8756 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8757 "linux-sparc" configuration.
8758 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8760 *) config now generates no-xxx options for missing ciphers.
8763 *) Support the EBCDIC character set (work in progress).
8764 File ebcdic.c not yet included because it has a different license.
8765 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8767 *) Support BS2000/OSD-POSIX.
8768 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8770 *) Make callbacks for key generation use void * instead of char *.
8773 *) Make S/MIME samples compile (not yet tested).
8776 *) Additional typesafe stacks.
8779 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8783 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8785 *) New configuration variant "sco5-gcc".
8787 *) Updated some demos.
8788 [Sean O Riordain, Wade Scholine]
8790 *) Add missing BIO_free at exit of pkcs12 application.
8793 *) Fix memory leak in conf.c.
8796 *) Updates for Win32 to assembler version of MD5.
8799 *) Set #! path to perl in apps/der_chop to where we found it
8800 instead of using a fixed path.
8803 *) SHA library changes for irix64-mips4-cc.
8806 *) Improvements for VMS support.
8810 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8812 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8813 This also avoids the problems with SC4.2 and unpatched SC5.
8814 [Andy Polyakov <appro@fy.chalmers.se>]
8816 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8817 These are required because of the typesafe stack would otherwise break
8818 existing code. If old code used a structure member which used to be STACK
8819 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8820 sk_num or sk_value it would produce an error because the num, data members
8821 are not present in STACK_OF. Now it just produces a warning. sk_set
8822 replaces the old method of assigning a value to sk_value
8823 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8824 that does this will no longer work (and should use sk_set instead) but
8825 this could be regarded as a "questionable" behaviour anyway.
8828 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8829 correctly handle encrypted S/MIME data.
8832 *) Change type of various DES function arguments from des_cblock
8833 (which means, in function argument declarations, pointer to char)
8834 to des_cblock * (meaning pointer to array with 8 char elements),
8835 which allows the compiler to do more typechecking; it was like
8836 that back in SSLeay, but with lots of ugly casts.
8838 Introduce new type const_des_cblock.
8841 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8842 problems: find RecipientInfo structure that matches recipient certificate
8843 and initialise the ASN1 structures properly based on passed cipher.
8846 *) Belatedly make the BN tests actually check the results.
8849 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8850 to and from BNs: it was completely broken. New compilation option
8851 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8852 key elements as negative integers.
8855 *) Reorganize and speed up MD5.
8856 [Andy Polyakov <appro@fy.chalmers.se>]
8859 [Richard Levitte <richard@levitte.org>]
8861 *) New option -out to asn1parse to allow the parsed structure to be
8862 output to a file. This is most useful when combined with the -strparse
8863 option to examine the output of things like OCTET STRINGS.
8866 *) Make SSL library a little more fool-proof by not requiring any longer
8867 that SSL_set_{accept,connect}_state be called before
8868 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8869 in many applications because usually everything *appeared* to work as
8870 intended anyway -- now it really works as intended).
8873 *) Move openssl.cnf out of lib/.
8876 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8877 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8878 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8879 [Ralf S. Engelschall]
8881 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8882 handle PKCS#7 enveloped data properly.
8883 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8885 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8886 copying pointers. The cert_st handling is changed by this in
8887 various ways (and thus what used to be known as ctx->default_cert
8888 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8889 any longer when s->cert does not give us what we need).
8890 ssl_cert_instantiate becomes obsolete by this change.
8891 As soon as we've got the new code right (possibly it already is?),
8892 we have solved a couple of bugs of the earlier code where s->cert
8893 was used as if it could not have been shared with other SSL structures.
8895 Note that using the SSL API in certain dirty ways now will result
8896 in different behaviour than observed with earlier library versions:
8897 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8898 does not influence s as it used to.
8900 In order to clean up things more thoroughly, inside SSL_SESSION
8901 we don't use CERT any longer, but a new structure SESS_CERT
8902 that holds per-session data (if available); currently, this is
8903 the peer's certificate chain and, for clients, the server's certificate
8904 and temporary key. CERT holds only those values that can have
8905 meaningful defaults in an SSL_CTX.
8908 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8909 from the internal representation. Various PKCS#7 fixes: remove some
8910 evil casts and set the enc_dig_alg field properly based on the signing
8914 *) Allow PKCS#12 password to be set from the command line or the
8915 environment. Let 'ca' get its config file name from the environment
8916 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8920 *) Allow certificate policies extension to use an IA5STRING for the
8921 organization field. This is contrary to the PKIX definition but
8922 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8926 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8927 without disallowing inline assembler and the like for non-pedantic builds.
8930 *) Support Borland C++ builder.
8931 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8936 *) SHA-1 cleanups and performance enhancements.
8937 [Andy Polyakov <appro@fy.chalmers.se>]
8939 *) Sparc v8plus assembler for the bignum library.
8940 [Andy Polyakov <appro@fy.chalmers.se>]
8942 *) Accept any -xxx and +xxx compiler options in Configure.
8945 *) Update HPUX configuration.
8948 *) Add missing sk_<type>_unshift() function to safestack.h
8949 [Ralf S. Engelschall]
8951 *) New function SSL_CTX_use_certificate_chain_file that sets the
8952 "extra_cert"s in addition to the certificate. (This makes sense
8953 only for "PEM" format files, as chains as a whole are not
8957 *) Support verify_depth from the SSL API.
8958 x509_vfy.c had what can be considered an off-by-one-error:
8959 Its depth (which was not part of the external interface)
8960 was actually counting the number of certificates in a chain;
8961 now it really counts the depth.
8964 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8965 instead of X509err, which often resulted in confusing error
8966 messages since the error codes are not globally unique
8967 (e.g. an alleged error in ssl3_accept when a certificate
8968 didn't match the private key).
8970 *) New function SSL_CTX_set_session_id_context that allows to set a default
8971 value (so that you don't need SSL_set_session_id_context for each
8972 connection using the SSL_CTX).
8975 *) OAEP decoding bug fix.
8978 *) Support INSTALL_PREFIX for package builders, as proposed by
8982 *) New Configure options "threads" and "no-threads". For systems
8983 where the proper compiler options are known (currently Solaris
8984 and Linux), "threads" is the default.
8987 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8990 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8991 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8992 such as /usr/local/bin.
8995 *) "make linux-shared" to build shared libraries.
8996 [Niels Poppe <niels@netbox.org>]
8998 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9001 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9002 extension adding in x509 utility.
9005 *) Remove NOPROTO sections and error code comments.
9008 *) Partial rewrite of the DEF file generator to now parse the ANSI
9012 *) New Configure options --prefix=DIR and --openssldir=DIR.
9015 *) Complete rewrite of the error code script(s). It is all now handled
9016 by one script at the top level which handles error code gathering,
9017 header rewriting and C source file generation. It should be much better
9018 than the old method: it now uses a modified version of Ulf's parser to
9019 read the ANSI prototypes in all header files (thus the old K&R definitions
9020 aren't needed for error creation any more) and do a better job of
9021 translating function codes into names. The old 'ASN1 error code imbedded
9022 in a comment' is no longer necessary and it doesn't use .err files which
9023 have now been deleted. Also the error code call doesn't have to appear all
9024 on one line (which resulted in some large lines...).
9027 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9030 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9031 0 (which usually indicates a closed connection), but continue reading.
9034 *) Fix some race conditions.
9037 *) Add support for CRL distribution points extension. Add Certificate
9038 Policies and CRL distribution points documentation.
9041 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9044 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9045 8 of keying material. Merlin has also confirmed interop with this fix
9046 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9047 [Merlin Hughes <merlin@baltimore.ie>]
9049 *) Fix lots of warnings.
9050 [Richard Levitte <levitte@stacken.kth.se>]
9052 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9053 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9054 [Richard Levitte <levitte@stacken.kth.se>]
9056 *) Fix problems with sizeof(long) == 8.
9057 [Andy Polyakov <appro@fy.chalmers.se>]
9059 *) Change functions to ANSI C.
9062 *) Fix typos in error codes.
9063 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9065 *) Remove defunct assembler files from Configure.
9068 *) SPARC v8 assembler BIGNUM implementation.
9069 [Andy Polyakov <appro@fy.chalmers.se>]
9071 *) Support for Certificate Policies extension: both print and set.
9072 Various additions to support the r2i method this uses.
9075 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9076 return a const string when you are expecting an allocated buffer.
9079 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9080 types DirectoryString and DisplayText.
9083 *) Add code to allow r2i extensions to access the configuration database,
9084 add an LHASH database driver and add several ctx helper functions.
9087 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9088 fail when they extended the size of a BIGNUM.
9091 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9092 support typesafe stack.
9095 *) Fix typo in SSL_[gs]et_options().
9096 [Nils Frostberg <nils@medcom.se>]
9098 *) Delete various functions and files that belonged to the (now obsolete)
9099 old X509V3 handling code.
9102 *) New Configure option "rsaref".
9105 *) Don't auto-generate pem.h.
9108 *) Introduce type-safe ASN.1 SETs.
9111 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9112 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9114 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9115 that links with OpenSSL (well at least cause lots of warnings), but fear
9116 not: the conversion is trivial, and it eliminates loads of evil casts. A
9117 few STACKed things have been converted already. Feel free to convert more.
9118 In the fullness of time, I'll do away with the STACK type altogether.
9121 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9122 specified in <certfile> by updating the entry in the index.txt file.
9123 This way one no longer has to edit the index.txt file manually for
9124 revoking a certificate. The -revoke option does the gory details now.
9125 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9127 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9128 `-text' option at all and this way the `-noout -text' combination was
9129 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9130 [Ralf S. Engelschall]
9132 *) Make sure a corresponding plain text error message exists for the
9133 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9134 verify callback function determined that a certificate was revoked.
9135 [Ralf S. Engelschall]
9137 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9138 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9139 all available cipers including rc5, which was forgotten until now.
9140 In order to let the testing shell script know which algorithms
9141 are available, a new (up to now undocumented) command
9142 "openssl list-cipher-commands" is used.
9145 *) Bugfix: s_client occasionally would sleep in select() when
9146 it should have checked SSL_pending() first.
9149 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9150 the raw DSA values prior to ASN.1 encoding.
9153 *) Tweaks to Configure
9154 [Niels Poppe <niels@netbox.org>]
9156 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9160 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9163 *) New config option to avoid instructions that are illegal on the 80386.
9164 The default code is faster, but requires at least a 486.
9167 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9168 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9169 same as SSL2_VERSION anyway.
9172 *) New "-showcerts" option for s_client.
9175 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9176 application. Various cleanups and fixes.
9179 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9180 modify error routines to work internally. Add error codes and PBE init
9181 to library startup routines.
9184 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9185 packing functions to asn1 and evp. Changed function names and error
9186 codes along the way.
9189 *) PKCS12 integration: and so it begins... First of several patches to
9190 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9191 objects to objects.h
9194 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9195 and display support for Thawte strong extranet extension.
9198 *) Add LinuxPPC support.
9199 [Jeff Dubrule <igor@pobox.org>]
9201 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9202 bn_div_words in alpha.s.
9203 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9205 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9206 OAEP isn't supported when OpenSSL is built with RSAref.
9207 [Ulf Moeller <ulf@fitug.de>]
9209 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9210 so they no longer are missing under -DNOPROTO.
9211 [Soren S. Jorvang <soren@t.dk>]
9214 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9216 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9217 doesn't work when the session is reused. Coming soon!
9220 *) Fix a security hole, that allows sessions to be reused in the wrong
9221 context thus bypassing client cert protection! All software that uses
9222 client certs and session caches in multiple contexts NEEDS PATCHING to
9223 allow session reuse! A fuller solution is in the works.
9224 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9226 *) Some more source tree cleanups (removed obsolete files
9227 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9228 permission on "config" script to be executable) and a fix for the INSTALL
9230 [Ulf Moeller <ulf@fitug.de>]
9232 *) Remove some legacy and erroneous uses of malloc, free instead of
9234 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9236 *) Make rsa_oaep_test return non-zero on error.
9237 [Ulf Moeller <ulf@fitug.de>]
9239 *) Add support for native Solaris shared libraries. Configure
9240 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9241 if someone would make that last step automatic.
9242 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9244 *) ctx_size was not built with the right compiler during "make links". Fixed.
9247 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9248 except NULL ciphers". This means the default cipher list will no longer
9249 enable NULL ciphers. They need to be specifically enabled e.g. with
9250 the string "DEFAULT:eNULL".
9253 *) Fix to RSA private encryption routines: if p < q then it would
9254 occasionally produce an invalid result. This will only happen with
9255 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9258 *) Be less restrictive and allow also `perl util/perlpath.pl
9259 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9260 because this way one can also use an interpreter named `perl5' (which is
9261 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9262 installed as `perl').
9263 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9265 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9266 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9268 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9269 advapi32.lib to Win32 build and change the pem test comparision
9270 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9271 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9272 and crypto/des/ede_cbcm_enc.c.
9275 *) DES quad checksum was broken on big-endian architectures. Fixed.
9278 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9279 Win32 test batch file so it (might) work again. The Win32 test batch file
9280 is horrible: I feel ill....
9283 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9284 in e_os.h. Audit of header files to check ANSI and non ANSI
9285 sections: 10 functions were absent from non ANSI section and not exported
9286 from Windows DLLs. Fixed up libeay.num for new functions.
9289 *) Make `openssl version' output lines consistent.
9290 [Ralf S. Engelschall]
9292 *) Fix Win32 symbol export lists for BIO functions: Added
9293 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9294 to ms/libeay{16,32}.def.
9295 [Ralf S. Engelschall]
9297 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9298 fine under Unix and passes some trivial tests I've now added. But the
9299 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9300 added to make sure no one expects that this stuff really works in the
9301 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9302 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9304 [Ralf S. Engelschall]
9306 *) Fix the generation of two part addresses in perl.
9307 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9309 *) Add config entry for Linux on MIPS.
9310 [John Tobey <jtobey@channel1.com>]
9312 *) Make links whenever Configure is run, unless we are on Windoze.
9315 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9316 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9320 *) Add a useful kludge to allow package maintainers to specify compiler and
9321 other platforms details on the command line without having to patch the
9322 Configure script everytime: One now can use ``perl Configure
9323 <id>:<details>'', i.e. platform ids are allowed to have details appended
9324 to them (seperated by colons). This is treated as there would be a static
9325 pre-configured entry in Configure's %table under key <id> with value
9326 <details> and ``perl Configure <id>'' is called. So, when you want to
9327 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9328 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9329 now, which overrides the FreeBSD-elf entry on-the-fly.
9330 [Ralf S. Engelschall]
9332 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9335 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9336 on the `perl Configure ...' command line. This way one can compile
9337 OpenSSL libraries with Position Independent Code (PIC) which is needed
9338 for linking it into DSOs.
9339 [Ralf S. Engelschall]
9341 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9345 *) Cleaned up the LICENSE document: The official contact for any license
9346 questions now is the OpenSSL core team under openssl-core@openssl.org.
9347 And add a paragraph about the dual-license situation to make sure people
9348 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9349 to the OpenSSL toolkit.
9350 [Ralf S. Engelschall]
9352 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9353 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9354 Additonally cleaned up the `make links' target: Remove unnecessary
9355 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9356 to speed processing and no longer clutter the display with confusing
9357 stuff. Instead only the actually done links are displayed.
9358 [Ralf S. Engelschall]
9360 *) Permit null encryption ciphersuites, used for authentication only. It used
9361 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9362 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9366 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9367 signed attributes when verifying signatures (this would break them),
9368 the detached data encoding was wrong and public keys obtained using
9369 X509_get_pubkey() weren't freed.
9372 *) Add text documentation for the BUFFER functions. Also added a work around
9373 to a Win95 console bug. This was triggered by the password read stuff: the
9374 last character typed gets carried over to the next fread(). If you were
9375 generating a new cert request using 'req' for example then the last
9376 character of the passphrase would be CR which would then enter the first
9380 *) Added the new `Includes OpenSSL Cryptography Software' button as
9381 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9382 button and can be used by applications based on OpenSSL to show the
9383 relationship to the OpenSSL project.
9384 [Ralf S. Engelschall]
9386 *) Remove confusing variables in function signatures in files
9387 ssl/ssl_lib.c and ssl/ssl.h.
9388 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9390 *) Don't install bss_file.c under PREFIX/include/
9391 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9393 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9394 functions that return function pointers and has support for NT specific
9395 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9396 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9397 unsigned to signed types: this was killing the Win32 compile.
9400 *) Add new certificate file to stack functions,
9401 SSL_add_dir_cert_subjects_to_stack() and
9402 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9403 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9404 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9405 This means that Apache-SSL and similar packages don't have to mess around
9406 to add as many CAs as they want to the preferred list.
9409 *) Experiment with doxygen documentation. Currently only partially applied to
9411 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9412 openssl.doxy as the configuration file.
9415 *) Get rid of remaining C++-style comments which strict C compilers hate.
9416 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9418 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9419 compiled in by default: it has problems with large keys.
9422 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9423 DH private keys and/or callback functions which directly correspond to
9424 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9425 is needed for applications which have to configure certificates on a
9426 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9428 For the RSA certificate situation is makes no difference, but
9429 for the DSA certificate situation this fixes the "no shared cipher"
9430 problem where the OpenSSL cipher selection procedure failed because the
9431 temporary keys were not overtaken from the context and the API provided
9432 no way to reconfigure them.
9433 The new functions now let applications reconfigure the stuff and they
9434 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9435 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9436 non-public-API function ssl_cert_instantiate() is used as a helper
9437 function and also to reduce code redundancy inside ssl_rsa.c.
9438 [Ralf S. Engelschall]
9440 *) Move s_server -dcert and -dkey options out of the undocumented feature
9441 area because they are useful for the DSA situation and should be
9442 recognized by the users.
9443 [Ralf S. Engelschall]
9445 *) Fix the cipher decision scheme for export ciphers: the export bits are
9446 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9447 SSL_EXP_MASK. So, the original variable has to be used instead of the
9448 already masked variable.
9449 [Richard Levitte <levitte@stacken.kth.se>]
9451 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9452 [Richard Levitte <levitte@stacken.kth.se>]
9454 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9455 from `int' to `unsigned int' because it's a length and initialized by
9456 EVP_DigestFinal() which expects an `unsigned int *'.
9457 [Richard Levitte <levitte@stacken.kth.se>]
9459 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9460 script. Instead use the usual Shell->Perl transition trick.
9461 [Ralf S. Engelschall]
9463 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9464 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9465 -noout -modulus' as it's already the case for `openssl rsa -noout
9466 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9467 currently the public key is printed (a decision which was already done by
9468 `openssl dsa -modulus' in the past) which serves a similar purpose.
9469 Additionally the NO_RSA no longer completely removes the whole -modulus
9470 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9472 [Ralf S. Engelschall]
9474 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9475 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9476 [Arne Ansper <arne@ats.cyber.ee>]
9478 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9479 to be added. Now both 'req' and 'ca' can use new objects defined in the
9483 *) Add cool BIO that does syslog (or event log on NT).
9484 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9486 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9487 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9488 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9489 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9492 *) Add preliminary config info for new extension code.
9495 *) Make RSA_NO_PADDING really use no padding.
9496 [Ulf Moeller <ulf@fitug.de>]
9498 *) Generate errors when private/public key check is done.
9501 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9502 for some CRL extensions and new objects added.
9505 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9506 key usage extension and fuller support for authority key id.
9509 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9510 padding method for RSA, which is recommended for new applications in PKCS
9511 #1 v2.0 (RFC 2437, October 1998).
9512 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9513 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9514 against Bleichbacher's attack on RSA.
9515 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9518 *) Updates to the new SSL compression code
9519 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9521 *) Fix so that the version number in the master secret, when passed
9522 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9523 (because the server will not accept higher), that the version number
9524 is 0x03,0x01, not 0x03,0x00
9525 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9527 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9528 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9529 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9532 *) Support for RAW extensions where an arbitrary extension can be
9533 created by including its DER encoding. See apps/openssl.cnf for
9537 *) Make sure latest Perl versions don't interpret some generated C array
9538 code as Perl array code in the crypto/err/err_genc.pl script.
9539 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9541 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9542 not many people have the assembler. Various Win32 compilation fixes and
9543 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9547 *) Modify configure script 'Configure' to automatically create crypto/date.h
9548 file under Win32 and also build pem.h from pem.org. New script
9549 util/mkfiles.pl to create the MINFO file on environments that can't do a
9550 'make files': perl util/mkfiles.pl >MINFO should work.
9553 *) Major rework of DES function declarations, in the pursuit of correctness
9554 and purity. As a result, many evil casts evaporated, and some weirdness,
9555 too. You may find this causes warnings in your code. Zapping your evil
9556 casts will probably fix them. Mostly.
9559 *) Fix for a typo in asn1.h. Bug fix to object creation script
9560 obj_dat.pl. It considered a zero in an object definition to mean
9561 "end of object": none of the objects in objects.h have any zeros
9562 so it wasn't spotted.
9563 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9565 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9566 Masking (CBCM). In the absence of test vectors, the best I have been able
9567 to do is check that the decrypt undoes the encrypt, so far. Send me test
9568 vectors if you have them.
9571 *) Correct calculation of key length for export ciphers (too much space was
9572 allocated for null ciphers). This has not been tested!
9575 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9576 message is now correct (it understands "crypto" and "ssl" on its
9577 command line). There is also now an "update" option. This will update
9578 the util/ssleay.num and util/libeay.num files with any new functions.
9580 perl util/mkdef.pl crypto ssl update
9581 it will update them.
9584 *) Overhauled the Perl interface (perl/*):
9585 - ported BN stuff to OpenSSL's different BN library
9586 - made the perl/ source tree CVS-aware
9587 - renamed the package from SSLeay to OpenSSL (the files still contain
9588 their history because I've copied them in the repository)
9589 - removed obsolete files (the test scripts will be replaced
9590 by better Test::Harness variants in the future)
9591 [Ralf S. Engelschall]
9593 *) First cut for a very conservative source tree cleanup:
9594 1. merge various obsolete readme texts into doc/ssleay.txt
9595 where we collect the old documents and readme texts.
9596 2. remove the first part of files where I'm already sure that we no
9597 longer need them because of three reasons: either they are just temporary
9598 files which were left by Eric or they are preserved original files where
9599 I've verified that the diff is also available in the CVS via "cvs diff
9600 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9601 the crypto/md/ stuff).
9602 [Ralf S. Engelschall]
9604 *) More extension code. Incomplete support for subject and issuer alt
9605 name, issuer and authority key id. Change the i2v function parameters
9606 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9607 what that's for :-) Fix to ASN1 macro which messed up
9608 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9611 *) Preliminary support for ENUMERATED type. This is largely copied from the
9615 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9616 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9618 *) Make sure `make rehash' target really finds the `openssl' program.
9619 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9621 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9622 like to hear about it if this slows down other processors.
9625 *) Add CygWin32 platform information to Configure script.
9626 [Alan Batie <batie@aahz.jf.intel.com>]
9628 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9629 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9631 *) New program nseq to manipulate netscape certificate sequences
9634 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9638 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9639 but the BN code had some problems that would cause failures when
9640 doing certificate verification and some other functions.
9641 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9643 *) Add ASN1 and PEM code to support netscape certificate sequences.
9646 *) Add ASN1 and PEM code to support netscape certificate sequences.
9649 *) Add several PKIX and private extended key usage OIDs.
9652 *) Modify the 'ca' program to handle the new extension code. Modify
9653 openssl.cnf for new extension format, add comments.
9656 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9657 and add a sample to openssl.cnf so req -x509 now adds appropriate
9661 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9662 error code, add initial support to X509_print() and x509 application.
9665 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9666 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9667 stuff is currently isolated and isn't even compiled yet.
9670 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9671 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9672 Removed the versions check from X509 routines when loading extensions:
9673 this allows certain broken certificates that don't set the version
9674 properly to be processed.
9677 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9678 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9679 can still be regenerated with "make depend".
9682 *) Spelling mistake in C version of CAST-128.
9683 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9685 *) Changes to the error generation code. The perl script err-code.pl
9686 now reads in the old error codes and retains the old numbers, only
9687 adding new ones if necessary. It also only changes the .err files if new
9688 codes are added. The makefiles have been modified to only insert errors
9689 when needed (to avoid needlessly modifying header files). This is done
9690 by only inserting errors if the .err file is newer than the auto generated
9691 C file. To rebuild all the error codes from scratch (the old behaviour)
9692 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9693 or delete all the .err files.
9696 *) CAST-128 was incorrectly implemented for short keys. The C version has
9697 been fixed, but is untested. The assembler versions are also fixed, but
9698 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9699 to regenerate it if needed.
9700 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9701 Hagino <itojun@kame.net>]
9703 *) File was opened incorrectly in randfile.c.
9704 [Ulf Möller <ulf@fitug.de>]
9706 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9707 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9708 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9709 al: it's just almost always a UTCTime. Note this patch adds new error
9710 codes so do a "make errors" if there are problems.
9713 *) Correct Linux 1 recognition in config.
9714 [Ulf Möller <ulf@fitug.de>]
9716 *) Remove pointless MD5 hash when using DSA keys in ca.
9717 [Anonymous <nobody@replay.com>]
9719 *) Generate an error if given an empty string as a cert directory. Also
9720 generate an error if handed NULL (previously returned 0 to indicate an
9721 error, but didn't set one).
9722 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9724 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9727 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9728 parameters. This was causing a warning which killed off the Win32 compile.
9731 *) Remove C++ style comments from crypto/bn/bn_local.h.
9732 [Neil Costigan <neil.costigan@celocom.com>]
9734 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9735 based on a text string, looking up short and long names and finally
9736 "dot" format. The "dot" format stuff didn't work. Added new function
9737 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9738 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9739 OID is not part of the table.
9742 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9743 X509_LOOKUP_by_alias().
9746 *) Sort openssl functions by name.
9749 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9750 encryption from sample DSA keys (in case anyone is interested the password
9754 *) Make _all_ *_free functions accept a NULL pointer.
9755 [Frans Heymans <fheymans@isaserver.be>]
9757 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9759 [Anonymous <nobody@replay.com>]
9761 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9762 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9764 *) Don't blow it for numeric -newkey arguments to apps/req.
9765 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9767 *) Temp key "for export" tests were wrong in s3_srvr.c.
9768 [Anonymous <nobody@replay.com>]
9770 *) Add prototype for temp key callback functions
9771 SSL_CTX_set_tmp_{rsa,dh}_callback().
9774 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9775 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9778 *) X509_name_add_entry() freed the wrong thing after an error.
9779 [Arne Ansper <arne@ats.cyber.ee>]
9781 *) rsa_eay.c would attempt to free a NULL context.
9782 [Arne Ansper <arne@ats.cyber.ee>]
9784 *) BIO_s_socket() had a broken should_retry() on Windoze.
9785 [Arne Ansper <arne@ats.cyber.ee>]
9787 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9788 [Arne Ansper <arne@ats.cyber.ee>]
9790 *) Make sure the already existing X509_STORE->depth variable is initialized
9791 in X509_STORE_new(), but document the fact that this variable is still
9792 unused in the certificate verification process.
9793 [Ralf S. Engelschall]
9795 *) Fix the various library and apps files to free up pkeys obtained from
9796 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9799 *) Fix reference counting in X509_PUBKEY_get(). This makes
9800 demos/maurice/example2.c work, amongst others, probably.
9801 [Steve Henson and Ben Laurie]
9803 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9804 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9805 are no longer created. This way we have a single and consistent command
9806 line interface `openssl <command>', similar to `cvs <command>'.
9807 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9809 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9810 BIT STRING wrapper always have zero unused bits.
9813 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9816 *) Make the top-level INSTALL documentation easier to understand.
9819 *) Makefiles updated to exit if an error occurs in a sub-directory
9820 make (including if user presses ^C) [Paul Sutton]
9822 *) Make Montgomery context stuff explicit in RSA data structure.
9825 *) Fix build order of pem and err to allow for generated pem.h.
9828 *) Fix renumbering bug in X509_NAME_delete_entry().
9831 *) Enhanced the err-ins.pl script so it makes the error library number
9832 global and can add a library name. This is needed for external ASN1 and
9833 other error libraries.
9836 *) Fixed sk_insert which never worked properly.
9839 *) Fix ASN1 macros so they can handle indefinite length construted
9840 EXPLICIT tags. Some non standard certificates use these: they can now
9844 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9845 into a single doc/ssleay.txt bundle. This way the information is still
9846 preserved but no longer messes up this directory. Now it's new room for
9847 the new set of documenation files.
9848 [Ralf S. Engelschall]
9850 *) SETs were incorrectly DER encoded. This was a major pain, because they
9851 shared code with SEQUENCEs, which aren't coded the same. This means that
9852 almost everything to do with SETs or SEQUENCEs has either changed name or
9853 number of arguments.
9854 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9856 *) Fix test data to work with the above.
9859 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9860 was already fixed by Eric for 0.9.1 it seems.
9861 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9863 *) Autodetect FreeBSD3.
9866 *) Fix various bugs in Configure. This affects the following platforms:
9870 unixware-2.0-pentium
9874 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9875 before they are needed.
9878 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9882 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9884 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9885 changed SSLeay to OpenSSL in version strings.
9886 [Ralf S. Engelschall]
9888 *) Some fixups to the top-level documents.
9891 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9892 because the symlink to include/ was missing.
9893 [Ralf S. Engelschall]
9895 *) Incorporated the popular no-RSA/DSA-only patches
9896 which allow to compile a RSA-free SSLeay.
9897 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9899 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9900 when "ssleay" is still not found.
9901 [Ralf S. Engelschall]
9903 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9904 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9906 *) Updated the README file.
9907 [Ralf S. Engelschall]
9909 *) Added various .cvsignore files in the CVS repository subdirs
9910 to make a "cvs update" really silent.
9911 [Ralf S. Engelschall]
9913 *) Recompiled the error-definition header files and added
9914 missing symbols to the Win32 linker tables.
9915 [Ralf S. Engelschall]
9917 *) Cleaned up the top-level documents;
9918 o new files: CHANGES and LICENSE
9919 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9920 o merged COPYRIGHT into LICENSE
9921 o removed obsolete TODO file
9922 o renamed MICROSOFT to INSTALL.W32
9923 [Ralf S. Engelschall]
9925 *) Removed dummy files from the 0.9.1b source tree:
9926 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9927 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9928 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9929 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9930 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9931 [Ralf S. Engelschall]
9933 *) Added various platform portability fixes.
9936 *) The Genesis of the OpenSSL rpject:
9937 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9938 Young and Tim J. Hudson created while they were working for C2Net until
9940 [The OpenSSL Project]
9943 Changes between 0.9.0b and 0.9.1b [not released]
9945 *) Updated a few CA certificates under certs/
9948 *) Changed some BIGNUM api stuff.
9951 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9952 DGUX x86, Linux Alpha, etc.
9955 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9956 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9960 *) Add -strparse option to asn1pars program which parses nested
9962 [Dr Stephen Henson <shenson@bigfoot.com>]
9964 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9967 *) DSA fix for "ca" program.
9970 *) Added "-genkey" option to "dsaparam" program.
9973 *) Added RIPE MD160 (rmd160) message digest.
9976 *) Added -a (all) option to "ssleay version" command.
9979 *) Added PLATFORM define which is the id given to Configure.
9982 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9985 *) Extended the ASN.1 parser routines.
9988 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9991 *) Added a BN_CTX to the BN library.
9994 *) Fixed the weak key values in DES library
9997 *) Changed API in EVP library for cipher aliases.
10000 *) Added support for RC2/64bit cipher.
10003 *) Converted the lhash library to the crypto/mem.c functions.
10006 *) Added more recognized ASN.1 object ids.
10009 *) Added more RSA padding checks for SSL/TLS.
10012 *) Added BIO proxy/filter functionality.
10015 *) Added extra_certs to SSL_CTX which can be used
10016 send extra CA certificates to the client in the CA cert chain sending
10017 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10020 *) Now Fortezza is denied in the authentication phase because
10021 this is key exchange mechanism is not supported by SSLeay at all.
10024 *) Additional PKCS1 checks.
10027 *) Support the string "TLSv1" for all TLS v1 ciphers.
10030 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10031 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10034 *) Fixed a few memory leaks.
10037 *) Fixed various code and comment typos.
10040 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10041 bytes sent in the client random.
10042 [Edward Bishop <ebishop@spyglass.com>]