5 Changes between 1.0.1b and 1.0.1c [xx XXX xxxx]
9 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
11 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
12 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
13 mean any application compiled against OpenSSL 1.0.0 headers setting
14 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
15 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
16 0x10000000L Any application which was previously compiled against
17 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
18 will need to be recompiled as a result. Letting be results in
19 inability to disable specifically TLS 1.1 and in client context,
20 in unlike event, limit maximum offered version to TLS 1.0 [see below].
23 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
24 disable just protocol X, but all protocols above X *if* there are
25 protocols *below* X still enabled. In more practical terms it means
26 that if application wants to disable TLS1.0 in favor of TLS1.1 and
27 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
28 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
32 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
34 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
35 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
36 in CRYPTO_realloc_clean.
38 Thanks to Tavis Ormandy, Google Security Team, for discovering this
39 issue and to Adam Langley <agl@chromium.org> for fixing it.
41 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
43 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
46 *) Workarounds for some broken servers that "hang" if a client hello
47 record length exceeds 255 bytes.
49 1. Do not use record version number > TLS 1.0 in initial client
50 hello: some (but not all) hanging servers will now work.
51 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
52 the number of ciphers sent in the client hello. This should be
53 set to an even number, such as 50, for example by passing:
54 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
55 Most broken servers should now work.
56 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
57 TLS 1.2 client support entirely.
60 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
63 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
65 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
66 STRING form instead of a DigestInfo.
69 *) The format used for MDC2 RSA signatures is inconsistent between EVP
70 and the RSA_sign/RSA_verify functions. This was made more apparent when
71 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
72 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
73 the correct format in RSA_verify so both forms transparently work.
76 *) Some servers which support TLS 1.0 can choke if we initially indicate
77 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
78 encrypted premaster secret. As a workaround use the maximum pemitted
79 client version in client hello, this should keep such servers happy
80 and still work with previous versions of OpenSSL.
83 *) Add support for TLS/DTLS heartbeats.
84 [Robin Seggelmann <seggelmann@fh-muenster.de>]
86 *) Add support for SCTP.
87 [Robin Seggelmann <seggelmann@fh-muenster.de>]
89 *) Improved PRNG seeding for VOS.
90 [Paul Green <Paul.Green@stratus.com>]
92 *) Extensive assembler packs updates, most notably:
94 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
95 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
96 - x86_64: bit-sliced AES implementation;
97 - ARM: NEON support, contemporary platforms optimizations;
98 - s390x: z196 support;
99 - *: GHASH and GF(2^m) multiplication implementations;
103 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
104 (removal of unnecessary code)
105 [Peter Sylvester <peter.sylvester@edelweb.fr>]
107 *) Add TLS key material exporter from RFC 5705.
110 *) Add DTLS-SRTP negotiation from RFC 5764.
113 *) Add Next Protocol Negotiation,
114 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
115 disabled with a no-npn flag to config or Configure. Code donated
117 [Adam Langley <agl@google.com> and Ben Laurie]
119 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
120 NIST-P256, NIST-P521, with constant-time single point multiplication on
121 typical inputs. Compiler support for the nonstandard type __uint128_t is
122 required to use this (present in gcc 4.4 and later, for 64-bit builds).
123 Code made available under Apache License version 2.0.
125 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
126 line to include this in your build of OpenSSL, and run "make depend" (or
127 "make update"). This enables the following EC_METHODs:
129 EC_GFp_nistp224_method()
130 EC_GFp_nistp256_method()
131 EC_GFp_nistp521_method()
133 EC_GROUP_new_by_curve_name() will automatically use these (while
134 EC_GROUP_new_curve_GFp() currently prefers the more flexible
136 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
138 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
139 all platforms. Move ssize_t definition from e_os.h to the public
140 header file e_os2.h as it now appears in public header file cms.h
143 *) New -sigopt option to the ca, req and x509 utilities. Additional
144 signature parameters can be passed using this option and in
148 *) Add RSA PSS signing function. This will generate and set the
149 appropriate AlgorithmIdentifiers for PSS based on those in the
150 corresponding EVP_MD_CTX structure. No application support yet.
153 *) Support for companion algorithm specific ASN1 signing routines.
154 New function ASN1_item_sign_ctx() signs a pre-initialised
155 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
156 the appropriate parameters.
159 *) Add new algorithm specific ASN1 verification initialisation function
160 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
161 handling will be the same no matter what EVP_PKEY_METHOD is used.
162 Add a PSS handler to support verification of PSS signatures: checked
163 against a number of sample certificates.
166 *) Add signature printing for PSS. Add PSS OIDs.
167 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
169 *) Add algorithm specific signature printing. An individual ASN1 method
170 can now print out signatures instead of the standard hex dump.
172 More complex signatures (e.g. PSS) can print out more meaningful
173 information. Include DSA version that prints out the signature
177 *) Password based recipient info support for CMS library: implementing
181 *) Split password based encryption into PBES2 and PBKDF2 functions. This
182 neatly separates the code into cipher and PBE sections and is required
183 for some algorithms that split PBES2 into separate pieces (such as
187 *) Session-handling fixes:
188 - Fix handling of connections that are resuming with a session ID,
189 but also support Session Tickets.
190 - Fix a bug that suppressed issuing of a new ticket if the client
191 presented a ticket with an expired session.
192 - Try to set the ticket lifetime hint to something reasonable.
193 - Make tickets shorter by excluding irrelevant information.
194 - On the client side, don't ignore renewed tickets.
195 [Adam Langley, Bodo Moeller (Google)]
197 *) Fix PSK session representation.
200 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
202 This work was sponsored by Intel.
205 *) Add GCM support to TLS library. Some custom code is needed to split
206 the IV between the fixed (from PRF) and explicit (from TLS record)
207 portions. This adds all GCM ciphersuites supported by RFC5288 and
208 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
209 add a special AESGCM string for GCM only.
212 *) Expand range of ctrls for AES GCM. Permit setting invocation
213 field on decrypt and retrieval of invocation field only on encrypt.
216 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
217 As required by RFC5289 these ciphersuites cannot be used if for
218 versions of TLS earlier than 1.2.
221 *) For FIPS capable OpenSSL interpret a NULL default public key method
222 as unset and return the appopriate default but do *not* set the default.
223 This means we can return the appopriate method in applications that
224 swicth between FIPS and non-FIPS modes.
227 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
228 ENGINE is used then we cannot handle that in the FIPS module so we
229 keep original code iff non-FIPS operations are allowed.
232 *) Add -attime option to openssl utilities.
233 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
235 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
238 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
239 FIPS EC methods unconditionally for now.
242 *) New build option no-ec2m to disable characteristic 2 code.
245 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
246 all cases can be covered as some introduce binary incompatibilities.
249 *) Redirect RSA operations to FIPS module including keygen,
250 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
253 *) Add similar low level API blocking to ciphers.
256 *) Low level digest APIs are not approved in FIPS mode: any attempt
257 to use these will cause a fatal error. Applications that *really* want
258 to use them can use the private_* version instead.
261 *) Redirect cipher operations to FIPS module for FIPS builds.
264 *) Redirect digest operations to FIPS module for FIPS builds.
267 *) Update build system to add "fips" flag which will link in fipscanister.o
268 for static and shared library builds embedding a signature if needed.
271 *) Output TLS supported curves in preference order instead of numerical
272 order. This is currently hardcoded for the highest order curves first.
273 This should be configurable so applications can judge speed vs strength.
276 *) Add TLS v1.2 server support for client authentication.
279 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
283 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
284 FIPS modules versions.
287 *) Add TLS v1.2 client side support for client authentication. Keep cache
288 of handshake records longer as we don't know the hash algorithm to use
289 until after the certificate request message is received.
292 *) Initial TLS v1.2 client support. Add a default signature algorithms
293 extension including all the algorithms we support. Parse new signature
294 format in client key exchange. Relax some ECC signing restrictions for
295 TLS v1.2 as indicated in RFC5246.
298 *) Add server support for TLS v1.2 signature algorithms extension. Switch
299 to new signature format when needed using client digest preference.
300 All server ciphersuites should now work correctly in TLS v1.2. No client
301 support yet and no support for client certificates.
304 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
305 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
306 ciphersuites. At present only RSA key exchange ciphersuites work with
307 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
308 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
309 and version checking.
312 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
313 with this defined it will not be affected by any changes to ssl internal
314 structures. Add several utility functions to allow openssl application
315 to work with OPENSSL_NO_SSL_INTERN defined.
319 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
321 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
324 *) Permit abbreviated handshakes when renegotiating using the function
325 SSL_renegotiate_abbreviated().
326 [Robin Seggelmann <seggelmann@fh-muenster.de>]
328 *) Add call to ENGINE_register_all_complete() to
329 ENGINE_load_builtin_engines(), so some implementations get used
330 automatically instead of needing explicit application support.
333 *) Add support for TLS key exporter as described in RFC5705.
334 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
336 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
337 a few changes are required:
339 Add SSL_OP_NO_TLSv1_1 flag.
341 Update version checking logic to handle version 1.1.
342 Add explicit IV handling (ported from DTLS code).
343 Add command line options to s_client/s_server.
346 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
348 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
349 in CMS and PKCS7 code. When RSA decryption fails use a random key for
350 content decryption and always return the same error. Note: this attack
351 needs on average 2^20 messages so it only affects automated senders. The
352 old behaviour can be reenabled in the CMS code by setting the
353 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
354 an MMA defence is not necessary.
355 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
356 this issue. (CVE-2012-0884)
359 *) Fix CVE-2011-4619: make sure we really are receiving a
360 client hello before rejecting multiple SGC restarts. Thanks to
361 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
364 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
366 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
367 Thanks to Antonio Martin, Enterprise Secure Access Research and
368 Development, Cisco Systems, Inc. for discovering this bug and
369 preparing a fix. (CVE-2012-0050)
372 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
374 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
375 of the Vaudenay padding oracle attack on CBC mode encryption
376 which enables an efficient plaintext recovery attack against
377 the OpenSSL implementation of DTLS. Their attack exploits timing
378 differences arising during decryption processing. A research
379 paper describing this attack can be found at:
380 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
381 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
382 Security Group at Royal Holloway, University of London
383 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
384 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
385 for preparing the fix. (CVE-2011-4108)
386 [Robin Seggelmann, Michael Tuexen]
388 *) Clear bytes used for block padding of SSL 3.0 records.
390 [Adam Langley (Google)]
392 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
393 Kadianakis <desnacked@gmail.com> for discovering this issue and
394 Adam Langley for preparing the fix. (CVE-2011-4619)
395 [Adam Langley (Google)]
397 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
398 [Andrey Kulikov <amdeich@gmail.com>]
400 *) Prevent malformed RFC3779 data triggering an assertion failure.
401 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
402 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
403 [Rob Austein <sra@hactrn.net>]
405 *) Improved PRNG seeding for VOS.
406 [Paul Green <Paul.Green@stratus.com>]
408 *) Fix ssl_ciph.c set-up race.
409 [Adam Langley (Google)]
411 *) Fix spurious failures in ecdsatest.c.
412 [Emilia Käsper (Google)]
414 *) Fix the BIO_f_buffer() implementation (which was mixing different
415 interpretations of the '..._len' fields).
416 [Adam Langley (Google)]
418 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
419 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
420 threads won't reuse the same blinding coefficients.
422 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
423 lock to call BN_BLINDING_invert_ex, and avoids one use of
424 BN_BLINDING_update for each BN_BLINDING structure (previously,
425 the last update always remained unused).
426 [Emilia Käsper (Google)]
428 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
429 [Bob Buckholz (Google)]
431 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
433 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
434 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
435 [Kaspar Brand <ossl@velox.ch>]
437 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
438 for multi-threaded use of ECDH. (CVE-2011-3210)
439 [Adam Langley (Google)]
441 *) Fix x509_name_ex_d2i memory leak on bad inputs.
444 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
445 signature public key algorithm by using OID xref utilities instead.
446 Before this you could only use some ECC ciphersuites with SHA1 only.
449 *) Add protection against ECDSA timing attacks as mentioned in the paper
450 by Billy Bob Brumley and Nicola Tuveri, see:
452 http://eprint.iacr.org/2011/232.pdf
454 [Billy Bob Brumley and Nicola Tuveri]
456 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
458 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
459 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
461 *) Fix bug in string printing code: if *any* escaping is enabled we must
462 escape the escape character (backslash) or the resulting string is
466 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
468 *) Disable code workaround for ancient and obsolete Netscape browsers
469 and servers: an attacker can use it in a ciphersuite downgrade attack.
470 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
473 *) Fixed J-PAKE implementation error, originally discovered by
474 Sebastien Martini, further info and confirmation from Stefan
475 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
478 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
480 *) Fix extension code to avoid race conditions which can result in a buffer
481 overrun vulnerability: resumed sessions must not be modified as they can
482 be shared by multiple threads. CVE-2010-3864
485 *) Fix WIN32 build system to correctly link an ENGINE directory into
489 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
491 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
493 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
495 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
497 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
498 context. The operation can be customised via the ctrl mechanism in
499 case ENGINEs want to include additional functionality.
502 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
505 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
506 output hashes compatible with older versions of OpenSSL.
507 [Willy Weisz <weisz@vcpc.univie.ac.at>]
509 *) Fix compression algorithm handling: if resuming a session use the
510 compression algorithm of the resumed session instead of determining
511 it from client hello again. Don't allow server to change algorithm.
514 *) Add load_crls() function to apps tidying load_certs() too. Add option
515 to verify utility to allow additional CRLs to be included.
518 *) Update OCSP request code to permit adding custom headers to the request:
519 some responders need this.
522 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
524 [Julia Lawall <julia@diku.dk>]
526 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
527 needlessly dereferenced structures, used obsolete functions and
528 didn't handle all updated verify codes correctly.
531 *) Disable MD2 in the default configuration.
534 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
535 indicate the initial BIO being pushed or popped. This makes it possible
536 to determine whether the BIO is the one explicitly called or as a result
537 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
538 it handles reference counts correctly and doesn't zero out the I/O bio
539 when it is not being explicitly popped. WARNING: applications which
540 included workarounds for the old buggy behaviour will need to be modified
541 or they could free up already freed BIOs.
544 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
545 renaming to all platforms (within the 0.9.8 branch, this was
546 done conditionally on Netware platforms to avoid a name clash).
547 [Guenter <lists@gknw.net>]
549 *) Add ECDHE and PSK support to DTLS.
550 [Michael Tuexen <tuexen@fh-muenster.de>]
552 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
556 *) Add "missing" function EVP_MD_flags() (without this the only way to
557 retrieve a digest flags is by accessing the structure directly. Update
558 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
559 or cipher is registered as in the "from" argument. Print out all
560 registered digests in the dgst usage message instead of manually
561 attempting to work them out.
564 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
565 this allows the use of compression and extensions. Change default cipher
566 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
567 by default unless an application cipher string requests it.
570 *) Alter match criteria in PKCS12_parse(). It used to try to use local
571 key ids to find matching certificates and keys but some PKCS#12 files
572 don't follow the (somewhat unwritten) rules and this strategy fails.
573 Now just gather all certificates together and the first private key
574 then look for the first certificate that matches the key.
577 *) Support use of registered digest and cipher names for dgst and cipher
578 commands instead of having to add each one as a special case. So now
585 openssl dgst -sha256 foo
587 and this works for ENGINE based algorithms too.
591 *) Update Gost ENGINE to support parameter files.
592 [Victor B. Wagner <vitus@cryptocom.ru>]
594 *) Support GeneralizedTime in ca utility.
595 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
597 *) Enhance the hash format used for certificate directory links. The new
598 form uses the canonical encoding (meaning equivalent names will work
599 even if they aren't identical) and uses SHA1 instead of MD5. This form
600 is incompatible with the older format and as a result c_rehash should
601 be used to rebuild symbolic links.
604 *) Make PKCS#8 the default write format for private keys, replacing the
605 traditional format. This form is standardised, more secure and doesn't
606 include an implicit MD5 dependency.
609 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
610 committed to OpenSSL should pass this lot as a minimum.
613 *) Add session ticket override functionality for use by EAP-FAST.
614 [Jouni Malinen <j@w1.fi>]
616 *) Modify HMAC functions to return a value. Since these can be implemented
617 in an ENGINE errors can occur.
620 *) Type-checked OBJ_bsearch_ex.
623 *) Type-checked OBJ_bsearch. Also some constification necessitated
624 by type-checking. Still to come: TXT_DB, bsearch(?),
625 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
629 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
630 seconds to a tm structure directly, instead of going through OS
631 specific date routines. This avoids any issues with OS routines such
632 as the year 2038 bug. New *_adj() functions for ASN1 time structures
633 and X509_time_adj_ex() to cover the extended range. The existing
634 X509_time_adj() is still usable and will no longer have any date issues.
637 *) Delta CRL support. New use deltas option which will attempt to locate
638 and search any appropriate delta CRLs available.
640 This work was sponsored by Google.
643 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
644 code and add additional score elements. Validate alternate CRL paths
645 as part of the CRL checking and indicate a new error "CRL path validation
646 error" in this case. Applications wanting additional details can use
647 the verify callback and check the new "parent" field. If this is not
648 NULL CRL path validation is taking place. Existing applications wont
649 see this because it requires extended CRL support which is off by
652 This work was sponsored by Google.
655 *) Support for freshest CRL extension.
657 This work was sponsored by Google.
660 *) Initial indirect CRL support. Currently only supported in the CRLs
661 passed directly and not via lookup. Process certificate issuer
662 CRL entry extension and lookup CRL entries by bother issuer name
663 and serial number. Check and process CRL issuer entry in IDP extension.
665 This work was sponsored by Google.
668 *) Add support for distinct certificate and CRL paths. The CRL issuer
669 certificate is validated separately in this case. Only enabled if
670 an extended CRL support flag is set: this flag will enable additional
671 CRL functionality in future.
673 This work was sponsored by Google.
676 *) Add support for policy mappings extension.
678 This work was sponsored by Google.
681 *) Fixes to pathlength constraint, self issued certificate handling,
682 policy processing to align with RFC3280 and PKITS tests.
684 This work was sponsored by Google.
687 *) Support for name constraints certificate extension. DN, email, DNS
688 and URI types are currently supported.
690 This work was sponsored by Google.
693 *) To cater for systems that provide a pointer-based thread ID rather
694 than numeric, deprecate the current numeric thread ID mechanism and
695 replace it with a structure and associated callback type. This
696 mechanism allows a numeric "hash" to be extracted from a thread ID in
697 either case, and on platforms where pointers are larger than 'long',
698 mixing is done to help ensure the numeric 'hash' is usable even if it
699 can't be guaranteed unique. The default mechanism is to use "&errno"
700 as a pointer-based thread ID to distinguish between threads.
702 Applications that want to provide their own thread IDs should now use
703 CRYPTO_THREADID_set_callback() to register a callback that will call
704 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
706 Note that ERR_remove_state() is now deprecated, because it is tied
707 to the assumption that thread IDs are numeric. ERR_remove_state(0)
708 to free the current thread's error state should be replaced by
709 ERR_remove_thread_state(NULL).
711 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
712 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
713 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
714 application was previously providing a numeric thread callback that
715 was inappropriate for distinguishing threads, then uniqueness might
716 have been obtained with &errno that happened immediately in the
717 intermediate development versions of OpenSSL; this is no longer the
718 case, the numeric thread callback will now override the automatic use
720 [Geoff Thorpe, with help from Bodo Moeller]
722 *) Initial support for different CRL issuing certificates. This covers a
723 simple case where the self issued certificates in the chain exist and
724 the real CRL issuer is higher in the existing chain.
726 This work was sponsored by Google.
729 *) Removed effectively defunct crypto/store from the build.
732 *) Revamp of STACK to provide stronger type-checking. Still to come:
733 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
734 ASN1_STRING, CONF_VALUE.
737 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
738 RAM on SSL connections. This option can save about 34k per idle SSL.
741 *) Revamp of LHASH to provide stronger type-checking. Still to come:
742 STACK, TXT_DB, bsearch, qsort.
745 *) Initial support for Cryptographic Message Syntax (aka CMS) based
746 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
747 support for data, signedData, compressedData, digestedData and
748 encryptedData, envelopedData types included. Scripts to check against
749 RFC4134 examples draft and interop and consistency checks of many
750 content types and variants.
753 *) Add options to enc utility to support use of zlib compression BIO.
756 *) Extend mk1mf to support importing of options and assembly language
757 files from Configure script, currently only included in VC-WIN32.
758 The assembly language rules can now optionally generate the source
759 files from the associated perl scripts.
762 *) Implement remaining functionality needed to support GOST ciphersuites.
763 Interop testing has been performed using CryptoPro implementations.
764 [Victor B. Wagner <vitus@cryptocom.ru>]
766 *) s390x assembler pack.
769 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
773 *) Implement Opaque PRF Input TLS extension as specified in
774 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
775 official specification yet and no extension type assignment by
776 IANA exists, this extension (for now) will have to be explicitly
777 enabled when building OpenSSL by providing the extension number
778 to use. For example, specify an option
780 -DTLSEXT_TYPE_opaque_prf_input=0x9527
782 to the "config" or "Configure" script to enable the extension,
783 assuming extension number 0x9527 (which is a completely arbitrary
784 and unofficial assignment based on the MD5 hash of the Internet
785 Draft). Note that by doing so, you potentially lose
786 interoperability with other TLS implementations since these might
787 be using the same extension number for other purposes.
789 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
790 opaque PRF input value to use in the handshake. This will create
791 an interal copy of the length-'len' string at 'src', and will
792 return non-zero for success.
794 To get more control and flexibility, provide a callback function
797 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
798 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
802 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
805 Callback function 'cb' will be called in handshakes, and is
806 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
807 Argument 'arg' is for application purposes (the value as given to
808 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
809 be provided to the callback function). The callback function
810 has to return non-zero to report success: usually 1 to use opaque
811 PRF input just if possible, or 2 to enforce use of the opaque PRF
812 input. In the latter case, the library will abort the handshake
813 if opaque PRF input is not successfully negotiated.
815 Arguments 'peerinput' and 'len' given to the callback function
816 will always be NULL and 0 in the case of a client. A server will
817 see the client's opaque PRF input through these variables if
818 available (NULL and 0 otherwise). Note that if the server
819 provides an opaque PRF input, the length must be the same as the
820 length of the client's opaque PRF input.
822 Note that the callback function will only be called when creating
823 a new session (session resumption can resume whatever was
824 previously negotiated), and will not be called in SSL 2.0
825 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
826 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
827 for applications that need to enforce opaque PRF input.
831 *) Update ssl code to support digests other than SHA1+MD5 for handshake
834 [Victor B. Wagner <vitus@cryptocom.ru>]
836 *) Add RFC4507 support to OpenSSL. This includes the corrections in
837 RFC4507bis. The encrypted ticket format is an encrypted encoded
838 SSL_SESSION structure, that way new session features are automatically
841 If a client application caches session in an SSL_SESSION structure
842 support is transparent because tickets are now stored in the encoded
845 The SSL_CTX structure automatically generates keys for ticket
846 protection in servers so again support should be possible
847 with no application modification.
849 If a client or server wishes to disable RFC4507 support then the option
850 SSL_OP_NO_TICKET can be set.
852 Add a TLS extension debugging callback to allow the contents of any client
853 or server extensions to be examined.
855 This work was sponsored by Google.
858 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
859 OpenSSL should now compile cleanly on gcc 4.2
860 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
862 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
863 support including streaming MAC support: this is required for GOST
865 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
867 *) Add option -stream to use PKCS#7 streaming in smime utility. New
868 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
869 to output in BER and PEM format.
872 *) Experimental support for use of HMAC via EVP_PKEY interface. This
873 allows HMAC to be handled via the EVP_DigestSign*() interface. The
874 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
875 ENGINE support for HMAC keys which are unextractable. New -mac and
876 -macopt options to dgst utility.
879 *) New option -sigopt to dgst utility. Update dgst to use
880 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
881 alternative signing paramaters such as X9.31 or PSS in the dgst
885 *) Change ssl_cipher_apply_rule(), the internal function that does
886 the work each time a ciphersuite string requests enabling
887 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
888 removing ("!foo+bar") a class of ciphersuites: Now it maintains
889 the order of disabled ciphersuites such that those ciphersuites
890 that most recently went from enabled to disabled not only stay
891 in order with respect to each other, but also have higher priority
892 than other disabled ciphersuites the next time ciphersuites are
895 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
896 the same ciphersuites as with "HIGH" alone, but in a specific
897 order where the PSK ciphersuites come first (since they are the
898 most recently disabled ciphersuites when "HIGH" is parsed).
900 Also, change ssl_create_cipher_list() (using this new
901 funcionality) such that between otherwise identical
902 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
906 *) Change ssl_create_cipher_list() so that it automatically
907 arranges the ciphersuites in reasonable order before starting
908 to process the rule string. Thus, the definition for "DEFAULT"
909 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
910 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
911 This makes it much easier to arrive at a reasonable default order
912 in applications for which anonymous ciphers are OK (meaning
913 that you can't actually use DEFAULT).
914 [Bodo Moeller; suggested by Victor Duchovni]
916 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
917 processing) into multiple integers instead of setting
918 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
919 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
920 (These masks as well as the individual bit definitions are hidden
921 away into the non-exported interface ssl/ssl_locl.h, so this
922 change to the definition of the SSL_CIPHER structure shouldn't
923 affect applications.) This give us more bits for each of these
924 categories, so there is no longer a need to coagulate AES128 and
925 AES256 into a single algorithm bit, and to coagulate Camellia128
926 and Camellia256 into a single algorithm bit, which has led to all
929 Thus, among other things, the kludge introduced in 0.9.7m and
930 0.9.8e for masking out AES256 independently of AES128 or masking
931 out Camellia256 independently of AES256 is not needed here in 0.9.9.
933 With the change, we also introduce new ciphersuite aliases that
934 so far were missing: "AES128", "AES256", "CAMELLIA128", and
938 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
939 Use the leftmost N bytes of the signature input if the input is
940 larger than the prime q (with N being the size in bytes of q).
943 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
944 it yet and it is largely untested.
947 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
950 *) Initial incomplete changes to avoid need for function casts in OpenSSL
951 some compilers (gcc 4.2 and later) reject their use. Safestack is
952 reimplemented. Update ASN1 to avoid use of legacy functions.
955 *) Win32/64 targets are linked with Winsock2.
958 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
959 to external functions. This can be used to increase CRL handling
960 efficiency especially when CRLs are very large by (for example) storing
961 the CRL revoked certificates in a database.
964 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
965 new CRLs added to a directory can be used. New command line option
966 -verify_return_error to s_client and s_server. This causes real errors
967 to be returned by the verify callback instead of carrying on no matter
968 what. This reflects the way a "real world" verify callback would behave.
971 *) GOST engine, supporting several GOST algorithms and public key formats.
972 Kindly donated by Cryptocom.
975 *) Partial support for Issuing Distribution Point CRL extension. CRLs
976 partitioned by DP are handled but no indirect CRL or reason partitioning
977 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
978 selected via a scoring technique which handles IDP and AKID in CRLs.
981 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
982 will ultimately be used for all verify operations: this will remove the
983 X509_STORE dependency on certificate verification and allow alternative
984 lookup methods. X509_STORE based implementations of these two callbacks.
987 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
988 Modify get_crl() to find a valid (unexpired) CRL if possible.
991 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
992 this would be called X509_CRL_cmp() but that name is already used by
993 a function that just compares CRL issuer names. Cache several CRL
994 extensions in X509_CRL structure and cache CRLDP in X509.
997 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
998 this maps equivalent X509_NAME structures into a consistent structure.
999 Name comparison can then be performed rapidly using memcmp().
1002 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1006 *) Allow digests to supply their own micalg string for S/MIME type using
1007 the ctrl EVP_MD_CTRL_MICALG.
1010 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1011 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1012 ctrl. It can then customise the structure before and/or after signing
1016 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1017 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1018 to free up any added signature OIDs.
1021 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1022 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1023 digest and cipher tables. New options added to openssl utility:
1024 list-message-digest-algorithms and list-cipher-algorithms.
1027 *) Change the array representation of binary polynomials: the list
1028 of degrees of non-zero coefficients is now terminated with -1.
1029 Previously it was terminated with 0, which was also part of the
1030 value; thus, the array representation was not applicable to
1031 polynomials where t^0 has coefficient zero. This change makes
1032 the array representation useful in a more general context.
1035 *) Various modifications and fixes to SSL/TLS cipher string
1036 handling. For ECC, the code now distinguishes between fixed ECDH
1037 with RSA certificates on the one hand and with ECDSA certificates
1038 on the other hand, since these are separate ciphersuites. The
1039 unused code for Fortezza ciphersuites has been removed.
1041 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1042 (not "ECDHE"). For consistency with the code for DH
1043 certificates, use of ECDH certificates is now considered ECDH
1044 authentication, not RSA or ECDSA authentication (the latter is
1045 merely the CA's signing algorithm and not actively used in the
1048 The temporary ciphersuite alias "ECCdraft" is no longer
1049 available, and ECC ciphersuites are no longer excluded from "ALL"
1050 and "DEFAULT". The following aliases now exist for RFC 4492
1051 ciphersuites, most of these by analogy with the DH case:
1053 kECDHr - ECDH cert, signed with RSA
1054 kECDHe - ECDH cert, signed with ECDSA
1055 kECDH - ECDH cert (signed with either RSA or ECDSA)
1056 kEECDH - ephemeral ECDH
1057 ECDH - ECDH cert or ephemeral ECDH
1063 AECDH - anonymous ECDH
1064 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1068 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1069 Use correct micalg parameters depending on digest(s) in signed message.
1072 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1073 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1076 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1077 an engine to register a method. Add ENGINE lookups for methods and
1078 functional reference processing.
1081 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1082 EVP_{Sign,Verify}* which allow an application to customise the signature
1086 *) New -resign option to smime utility. This adds one or more signers
1087 to an existing PKCS#7 signedData structure. Also -md option to use an
1088 alternative message digest algorithm for signing.
1091 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1092 create PKCS7 structures containing multiple signers. Update smime
1093 application to support multiple signers.
1096 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1100 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1101 Reorganize PBE internals to lookup from a static table using NIDs,
1102 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1103 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1104 PRF which will be automatically used with PBES2.
1107 *) Replace the algorithm specific calls to generate keys in "req" with the
1111 *) Update PKCS#7 enveloped data routines to use new API. This is now
1112 supported by any public key method supporting the encrypt operation. A
1113 ctrl is added to allow the public key algorithm to examine or modify
1114 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1118 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1119 a default digest type to use. In most cases this will be SHA1 but some
1120 algorithms (such as GOST) need to specify an alternative digest. The
1121 return value indicates how strong the prefernce is 1 means optional and
1122 2 is mandatory (that is it is the only supported type). Modify
1123 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1124 use the default md. Update openssl utilities to use the default digest
1125 type for signing if it is not explicitly indicated.
1128 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1129 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1130 signing method from the key type. This effectively removes the link
1131 between digests and public key types.
1134 *) Add an OID cross reference table and utility functions. Its purpose is to
1135 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1136 rsaEncryption. This will allow some of the algorithm specific hackery
1137 needed to use the correct OID to be removed.
1140 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1141 structures for PKCS7_sign(). They are now set up by the relevant public
1145 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1148 *) Add support for key derivation (agreement) in the API, DH method and
1152 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1153 public and private key formats. As a side effect these add additional
1154 command line functionality not previously available: DSA signatures can be
1155 generated and verified using pkeyutl and DH key support and generation in
1160 [Oliver Tappe <zooey@hirschkaefer.de>]
1162 *) New make target "install_html_docs" installs HTML renditions of the
1164 [Oliver Tappe <zooey@hirschkaefer.de>]
1166 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1167 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1168 support key and parameter generation and add initial key generation
1169 functionality for RSA.
1172 *) Add functions for main EVP_PKEY_method operations. The undocumented
1173 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1174 EVP_PKEY_{encrypt,decrypt}_old.
1177 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1178 key API, doesn't do much yet.
1181 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1182 public key algorithms. New option to openssl utility:
1183 "list-public-key-algorithms" to print out info.
1186 *) Implement the Supported Elliptic Curves Extension for
1187 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1190 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1191 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1194 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1195 utilities such as rsa, dsa, dsaparam etc except they process any key
1199 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1200 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1201 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1205 *) Initial support for pluggable public key ASN1.
1206 De-spaghettify the public key ASN1 handling. Move public and private
1207 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1208 algorithm specific handling to a single module within the relevant
1209 algorithm directory. Add functions to allow (near) opaque processing
1210 of public and private key structures.
1213 *) Implement the Supported Point Formats Extension for
1214 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1217 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1218 for the psk identity [hint] and the psk callback functions to the
1219 SSL_SESSION, SSL and SSL_CTX structure.
1222 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1226 SSL_CTX_use_psk_identity_hint
1227 SSL_get_psk_identity_hint
1228 SSL_get_psk_identity
1229 SSL_use_psk_identity_hint
1231 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1233 *) Add RFC 3161 compliant time stamp request creation, response generation
1234 and response verification functionality.
1235 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1237 *) Add initial support for TLS extensions, specifically for the server_name
1238 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1239 have new members for a host name. The SSL data structure has an
1240 additional member SSL_CTX *initial_ctx so that new sessions can be
1241 stored in that context to allow for session resumption, even after the
1242 SSL has been switched to a new SSL_CTX in reaction to a client's
1243 server_name extension.
1245 New functions (subject to change):
1247 SSL_get_servername()
1248 SSL_get_servername_type()
1251 New CTRL codes and macros (subject to change):
1253 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1254 - SSL_CTX_set_tlsext_servername_callback()
1255 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1256 - SSL_CTX_set_tlsext_servername_arg()
1257 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1259 openssl s_client has a new '-servername ...' option.
1261 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1262 '-key2 ...', '-servername_fatal' (subject to change). This allows
1263 testing the HostName extension for a specific single host name ('-cert'
1264 and '-key' remain fallbacks for handshakes without HostName
1265 negotiation). If the unrecogninzed_name alert has to be sent, this by
1266 default is a warning; it becomes fatal with the '-servername_fatal'
1269 [Peter Sylvester, Remy Allais, Christophe Renou]
1271 *) Whirlpool hash implementation is added.
1274 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1275 bn(64,32). Because of instruction set limitations it doesn't have
1276 any negative impact on performance. This was done mostly in order
1277 to make it possible to share assembler modules, such as bn_mul_mont
1278 implementations, between 32- and 64-bit builds without hassle.
1281 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1282 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1286 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1287 dedicated Montgomery multiplication procedure, is introduced.
1288 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1289 "64-bit" performance on certain 32-bit targets.
1292 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1293 in SSL structures. New SSL ctrl to set maximum send fragment size.
1294 Save memory by seeting the I/O buffer sizes dynamically instead of
1295 using the maximum available value.
1298 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1299 in addition to the text details.
1302 *) Very, very preliminary EXPERIMENTAL support for printing of general
1303 ASN1 structures. This currently produces rather ugly output and doesn't
1304 handle several customised structures at all.
1307 *) Integrated support for PVK file format and some related formats such
1308 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1309 these in the 'rsa' and 'dsa' utilities.
1312 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1315 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1316 place for the (very old) "NETSCAPE" format certificates which are now
1317 handled using new ASN1 code equivalents.
1320 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1321 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1322 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1325 *) Modify CRL distribution points extension code to print out previously
1326 unsupported fields. Enhance extension setting code to allow setting of
1330 *) Add print and set support for Issuing Distribution Point CRL extension.
1333 *) Change 'Configure' script to enable Camellia by default.
1336 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1338 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1339 Thanks to Antonio Martin, Enterprise Secure Access Research and
1340 Development, Cisco Systems, Inc. for discovering this bug and
1341 preparing a fix. (CVE-2012-0050)
1344 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1346 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1347 of the Vaudenay padding oracle attack on CBC mode encryption
1348 which enables an efficient plaintext recovery attack against
1349 the OpenSSL implementation of DTLS. Their attack exploits timing
1350 differences arising during decryption processing. A research
1351 paper describing this attack can be found at:
1352 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1353 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1354 Security Group at Royal Holloway, University of London
1355 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1356 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1357 for preparing the fix. (CVE-2011-4108)
1358 [Robin Seggelmann, Michael Tuexen]
1360 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1361 [Ben Laurie, Kasper <ekasper@google.com>]
1363 *) Clear bytes used for block padding of SSL 3.0 records.
1365 [Adam Langley (Google)]
1367 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1368 Kadianakis <desnacked@gmail.com> for discovering this issue and
1369 Adam Langley for preparing the fix. (CVE-2011-4619)
1370 [Adam Langley (Google)]
1372 *) Prevent malformed RFC3779 data triggering an assertion failure.
1373 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1374 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1375 [Rob Austein <sra@hactrn.net>]
1377 *) Fix ssl_ciph.c set-up race.
1378 [Adam Langley (Google)]
1380 *) Fix spurious failures in ecdsatest.c.
1381 [Emilia Käsper (Google)]
1383 *) Fix the BIO_f_buffer() implementation (which was mixing different
1384 interpretations of the '..._len' fields).
1385 [Adam Langley (Google)]
1387 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1388 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1389 threads won't reuse the same blinding coefficients.
1391 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1392 lock to call BN_BLINDING_invert_ex, and avoids one use of
1393 BN_BLINDING_update for each BN_BLINDING structure (previously,
1394 the last update always remained unused).
1395 [Emilia Käsper (Google)]
1397 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1398 for multi-threaded use of ECDH.
1399 [Adam Langley (Google)]
1401 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1404 *) Add protection against ECDSA timing attacks as mentioned in the paper
1405 by Billy Bob Brumley and Nicola Tuveri, see:
1407 http://eprint.iacr.org/2011/232.pdf
1409 [Billy Bob Brumley and Nicola Tuveri]
1411 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1413 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1414 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1416 *) Fix bug in string printing code: if *any* escaping is enabled we must
1417 escape the escape character (backslash) or the resulting string is
1421 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1423 *) Disable code workaround for ancient and obsolete Netscape browsers
1424 and servers: an attacker can use it in a ciphersuite downgrade attack.
1425 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1428 *) Fixed J-PAKE implementation error, originally discovered by
1429 Sebastien Martini, further info and confirmation from Stefan
1430 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1433 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1435 *) Fix extension code to avoid race conditions which can result in a buffer
1436 overrun vulnerability: resumed sessions must not be modified as they can
1437 be shared by multiple threads. CVE-2010-3864
1440 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1443 *) Don't reencode certificate when calculating signature: cache and use
1444 the original encoding instead. This makes signature verification of
1445 some broken encodings work correctly.
1448 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1449 is also one of the inputs.
1450 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1452 *) Don't repeatedly append PBE algorithms to table if they already exist.
1453 Sort table on each new add. This effectively makes the table read only
1454 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1458 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1460 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1463 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1464 access or freeing data twice (CVE-2010-0742)
1465 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1467 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1468 common in certificates and some applications which only call
1469 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1473 Reduce copying into .apps and .test in makevms.com
1474 Don't try to use blank CA certificate in CA.com
1475 Allow use of C files from original directories in maketests.com
1476 [Steven M. Schweda" <sms@antinode.info>]
1478 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1480 *) When rejecting SSL/TLS records due to an incorrect version number, never
1481 update s->server with a new major version number. As of
1482 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1483 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1484 the previous behavior could result in a read attempt at NULL when
1485 receiving specific incorrect SSL/TLS records once record payload
1486 protection is active. (CVE-2010-0740)
1487 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1489 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1490 could be crashed if the relevant tables were not present (e.g. chrooted).
1491 [Tomas Hoger <thoger@redhat.com>]
1493 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1495 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1496 [Martin Olsson, Neel Mehta]
1498 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1499 accommodate for stack sorting, always a write lock!).
1502 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1503 excessive delays in the RAND_poll(): over a minute. As a workaround
1504 include a time check in the inner Heap32Next loop too.
1507 *) The code that handled flushing of data in SSL/TLS originally used the
1508 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1509 the problem outlined in PR#1949. The fix suggested there however can
1510 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1511 of Apache). So instead simplify the code to flush unconditionally.
1512 This should be fine since flushing with no data to flush is a no op.
1515 *) Handle TLS versions 2.0 and later properly and correctly use the
1516 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1517 off ancient servers have a habit of sticking around for a while...
1520 *) Modify compression code so it frees up structures without using the
1521 ex_data callbacks. This works around a problem where some applications
1522 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1523 restarting) then use compression (e.g. SSL with compression) later.
1524 This results in significant per-connection memory leaks and
1525 has caused some security issues including CVE-2008-1678 and
1529 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1530 change when encrypting or decrypting.
1533 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1534 connect and renegotiate with servers which do not support RI.
1535 Until RI is more widely deployed this option is enabled by default.
1538 *) Add "missing" ssl ctrls to clear options and mode.
1541 *) If client attempts to renegotiate and doesn't support RI respond with
1542 a no_renegotiation alert as required by RFC5746. Some renegotiating
1543 TLS clients will continue a connection gracefully when they receive
1544 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1545 waiting for a server hello which it will never receive. Now we treat a
1546 received no_renegotiation alert as a fatal error. This is because
1547 applications requesting a renegotiation might well expect it to succeed
1548 and would have no code in place to handle the server denying it so the
1549 only safe thing to do is to terminate the connection.
1552 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1553 peer supports secure renegotiation and 0 otherwise. Print out peer
1554 renegotiation support in s_client/s_server.
1557 *) Replace the highly broken and deprecated SPKAC certification method with
1558 the updated NID creation version. This should correctly handle UTF8.
1561 *) Implement RFC5746. Re-enable renegotiation but require the extension
1562 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1563 turns out to be a bad idea. It has been replaced by
1564 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1565 SSL_CTX_set_options(). This is really not recommended unless you
1566 know what you are doing.
1567 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1569 *) Fixes to stateless session resumption handling. Use initial_ctx when
1570 issuing and attempting to decrypt tickets in case it has changed during
1571 servername handling. Use a non-zero length session ID when attempting
1572 stateless session resumption: this makes it possible to determine if
1573 a resumption has occurred immediately after receiving server hello
1574 (several places in OpenSSL subtly assume this) instead of later in
1578 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1579 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1580 fixes for a few places where the return code is not checked
1582 [Julia Lawall <julia@diku.dk>]
1584 *) Add --strict-warnings option to Configure script to include devteam
1585 warnings in other configurations.
1588 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1589 makes it possible to install openssl libraries in locations which
1590 have names other than "lib", for example "/usr/lib64" which some
1592 [Steve Henson, based on patch from Jeremy Utley]
1594 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1595 X690 8.9.12 and can produce some misleading textual output of OIDs.
1596 [Steve Henson, reported by Dan Kaminsky]
1598 *) Delete MD2 from algorithm tables. This follows the recommendation in
1599 several standards that it is not used in new applications due to
1600 several cryptographic weaknesses. For binary compatibility reasons
1601 the MD2 API is still compiled in by default.
1604 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1608 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1609 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1611 [Guenter <lists@gknw.net>]
1613 *) Fix the server certificate chain building code to use X509_verify_cert(),
1614 it used to have an ad-hoc builder which was unable to cope with anything
1615 other than a simple chain.
1616 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1618 *) Don't check self signed certificate signatures in X509_verify_cert()
1619 by default (a flag can override this): it just wastes time without
1620 adding any security. As a useful side effect self signed root CAs
1621 with non-FIPS digests are now usable in FIPS mode.
1624 *) In dtls1_process_out_of_seq_message() the check if the current message
1625 is already buffered was missing. For every new message was memory
1626 allocated, allowing an attacker to perform an denial of service attack
1627 with sending out of seq handshake messages until there is no memory
1628 left. Additionally every future messege was buffered, even if the
1629 sequence number made no sense and would be part of another handshake.
1630 So only messages with sequence numbers less than 10 in advance will be
1631 buffered. (CVE-2009-1378)
1632 [Robin Seggelmann, discovered by Daniel Mentz]
1634 *) Records are buffered if they arrive with a future epoch to be
1635 processed after finishing the corresponding handshake. There is
1636 currently no limitation to this buffer allowing an attacker to perform
1637 a DOS attack with sending records with future epochs until there is no
1638 memory left. This patch adds the pqueue_size() function to detemine
1639 the size of a buffer and limits the record buffer to 100 entries.
1641 [Robin Seggelmann, discovered by Daniel Mentz]
1643 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1644 parent structure is freed. (CVE-2009-1379)
1647 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1648 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1651 [Ilya O. <vrghost@gmail.com>]
1653 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1655 *) Disable renegotiation completely - this fixes a severe security
1656 problem (CVE-2009-3555) at the cost of breaking all
1657 renegotiation. Renegotiation can be re-enabled by setting
1658 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1659 run-time. This is really not recommended unless you know what
1663 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1665 *) Don't set val to NULL when freeing up structures, it is freed up by
1666 underlying code. If sizeof(void *) > sizeof(long) this can result in
1667 zeroing past the valid field. (CVE-2009-0789)
1668 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1670 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1671 checked correctly. This would allow some invalid signed attributes to
1672 appear to verify correctly. (CVE-2009-0591)
1673 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1675 *) Reject UniversalString and BMPString types with invalid lengths. This
1676 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1677 a legal length. (CVE-2009-0590)
1680 *) Set S/MIME signing as the default purpose rather than setting it
1681 unconditionally. This allows applications to override it at the store
1685 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1686 to handle some structures.
1689 *) Improve efficiency of mem_gets: don't search whole buffer each time
1691 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1693 *) New -hex option for openssl rand.
1696 *) Print out UTF8String and NumericString when parsing ASN1.
1699 *) Support NumericString type for name components.
1702 *) Allow CC in the environment to override the automatically chosen
1703 compiler. Note that nothing is done to ensure flags work with the
1707 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1709 *) Properly check EVP_VerifyFinal() and similar return values
1711 [Ben Laurie, Bodo Moeller, Google Security Team]
1713 *) Enable TLS extensions by default.
1716 *) Allow the CHIL engine to be loaded, whether the application is
1717 multithreaded or not. (This does not release the developer from the
1718 obligation to set up the dynamic locking callbacks.)
1719 [Sander Temme <sander@temme.net>]
1721 *) Use correct exit code if there is an error in dgst command.
1722 [Steve Henson; problem pointed out by Roland Dirlewanger]
1724 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1725 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1728 *) Add experimental JPAKE support, including demo authentication in
1729 s_client and s_server.
1732 *) Set the comparison function in v3_addr_canonize().
1733 [Rob Austein <sra@hactrn.net>]
1735 *) Add support for XMPP STARTTLS in s_client.
1736 [Philip Paeps <philip@freebsd.org>]
1738 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1739 to ensure that even with this option, only ciphersuites in the
1740 server's preference list will be accepted. (Note that the option
1741 applies only when resuming a session, so the earlier behavior was
1742 just about the algorithm choice for symmetric cryptography.)
1745 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1747 *) Fix NULL pointer dereference if a DTLS server received
1748 ChangeCipherSpec as first record (CVE-2009-1386).
1751 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1752 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1755 *) The fix in 0.9.8c that supposedly got rid of unsafe
1756 double-checked locking was incomplete for RSA blinding,
1757 addressing just one layer of what turns out to have been
1758 doubly unsafe triple-checked locking.
1760 So now fix this for real by retiring the MONT_HELPER macro
1761 in crypto/rsa/rsa_eay.c.
1763 [Bodo Moeller; problem pointed out by Marius Schilder]
1765 *) Various precautionary measures:
1767 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1769 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1770 (NB: This would require knowledge of the secret session ticket key
1771 to exploit, in which case you'd be SOL either way.)
1773 - Change bn_nist.c so that it will properly handle input BIGNUMs
1774 outside the expected range.
1776 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1779 [Neel Mehta, Bodo Moeller]
1781 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1782 the load fails. Useful for distros.
1783 [Ben Laurie and the FreeBSD team]
1785 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1788 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1791 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1793 This work was sponsored by Logica.
1796 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1797 keystores. Support for SSL/TLS client authentication too.
1798 Not compiled unless enable-capieng specified to Configure.
1800 This work was sponsored by Logica.
1803 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1804 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1805 attribute creation routines such as certifcate requests and PKCS#12
1809 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1811 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1812 handshake which could lead to a cilent crash as found using the
1813 Codenomicon TLS test suite (CVE-2008-1672)
1814 [Steve Henson, Mark Cox]
1816 *) Fix double free in TLS server name extensions which could lead to
1817 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1820 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1822 Clear the error queue to ensure that error entries left from
1823 older function calls do not interfere with the correct operation.
1824 [Lutz Jaenicke, Erik de Castro Lopo]
1826 *) Remove root CA certificates of commercial CAs:
1828 The OpenSSL project does not recommend any specific CA and does not
1829 have any policy with respect to including or excluding any CA.
1830 Therefore it does not make any sense to ship an arbitrary selection
1831 of root CA certificates with the OpenSSL software.
1834 *) RSA OAEP patches to fix two separate invalid memory reads.
1835 The first one involves inputs when 'lzero' is greater than
1836 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1837 before the beginning of from). The second one involves inputs where
1838 the 'db' section contains nothing but zeroes (there is a one-byte
1839 invalid read after the end of 'db').
1840 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1842 *) Partial backport from 0.9.9-dev:
1844 Introduce bn_mul_mont (dedicated Montgomery multiplication
1845 procedure) as a candidate for BIGNUM assembler implementation.
1846 While 0.9.9-dev uses assembler for various architectures, only
1847 x86_64 is available by default here in the 0.9.8 branch, and
1848 32-bit x86 is available through a compile-time setting.
1850 To try the 32-bit x86 assembler implementation, use Configure
1851 option "enable-montasm" (which exists only for this backport).
1853 As "enable-montasm" for 32-bit x86 disclaims code stability
1854 anyway, in this constellation we activate additional code
1855 backported from 0.9.9-dev for further performance improvements,
1856 namely BN_from_montgomery_word. (To enable this otherwise,
1857 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1859 [Andy Polyakov (backport partially by Bodo Moeller)]
1861 *) Add TLS session ticket callback. This allows an application to set
1862 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1863 values. This is useful for key rollover for example where several key
1864 sets may exist with different names.
1867 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1868 This was broken until now in 0.9.8 releases, such that the only way
1869 a registered ENGINE could be used (assuming it initialises
1870 successfully on the host) was to explicitly set it as the default
1871 for the relevant algorithms. This is in contradiction with 0.9.7
1872 behaviour and the documentation. With this fix, when an ENGINE is
1873 registered into a given algorithm's table of implementations, the
1874 'uptodate' flag is reset so that auto-discovery will be used next
1875 time a new context for that algorithm attempts to select an
1877 [Ian Lister (tweaked by Geoff Thorpe)]
1879 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1880 implemention in the following ways:
1882 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1885 Lack of BER streaming support means one pass streaming processing is
1886 only supported if data is detached: setting the streaming flag is
1887 ignored for embedded content.
1889 CMS support is disabled by default and must be explicitly enabled
1890 with the enable-cms configuration option.
1893 *) Update the GMP engine glue to do direct copies between BIGNUM and
1894 mpz_t when openssl and GMP use the same limb size. Otherwise the
1895 existing "conversion via a text string export" trick is still used.
1896 [Paul Sheer <paulsheer@gmail.com>]
1898 *) Zlib compression BIO. This is a filter BIO which compressed and
1899 uncompresses any data passed through it.
1902 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1903 RFC3394 compatible AES key wrapping.
1906 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1907 sets string data without copying. X509_ALGOR_set0() and
1908 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1909 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1910 from an X509_ATTRIBUTE structure optionally checking it occurs only
1911 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1915 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1916 to get the expected BN_FLG_CONSTTIME behavior.
1917 [Bodo Moeller (Google)]
1921 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1922 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1923 - added some more tests to do_tests.pl
1924 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1925 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1926 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1927 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1928 - various changes to netware.pl to enable gcc-cross builds on Win32
1930 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1931 - various changes to fix missing prototype warnings
1932 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1933 - added AES, WHIRLPOOL and CPUID assembler code to build files
1934 - added missing AES assembler make rules to mk1mf.pl
1935 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1936 [Guenter Knauf <eflash@gmx.net>]
1938 *) Implement certificate status request TLS extension defined in RFC3546.
1939 A client can set the appropriate parameters and receive the encoded
1940 OCSP response via a callback. A server can query the supplied parameters
1941 and set the encoded OCSP response in the callback. Add simplified examples
1942 to s_client and s_server.
1945 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1947 *) Fix various bugs:
1948 + Binary incompatibility of ssl_ctx_st structure
1949 + DTLS interoperation with non-compliant servers
1950 + Don't call get_session_cb() without proposed session
1951 + Fix ia64 assembler code
1952 [Andy Polyakov, Steve Henson]
1954 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1956 *) DTLS Handshake overhaul. There were longstanding issues with
1957 OpenSSL DTLS implementation, which were making it impossible for
1958 RFC 4347 compliant client to communicate with OpenSSL server.
1959 Unfortunately just fixing these incompatibilities would "cut off"
1960 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1961 server keeps tolerating non RFC compliant syntax. The opposite is
1962 not true, 0.9.8f client can not communicate with earlier server.
1963 This update even addresses CVE-2007-4995.
1966 *) Changes to avoid need for function casts in OpenSSL: some compilers
1967 (gcc 4.2 and later) reject their use.
1968 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1971 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1972 RFC4507bis. The encrypted ticket format is an encrypted encoded
1973 SSL_SESSION structure, that way new session features are automatically
1976 If a client application caches session in an SSL_SESSION structure
1977 support is transparent because tickets are now stored in the encoded
1980 The SSL_CTX structure automatically generates keys for ticket
1981 protection in servers so again support should be possible
1982 with no application modification.
1984 If a client or server wishes to disable RFC4507 support then the option
1985 SSL_OP_NO_TICKET can be set.
1987 Add a TLS extension debugging callback to allow the contents of any client
1988 or server extensions to be examined.
1990 This work was sponsored by Google.
1993 *) Add initial support for TLS extensions, specifically for the server_name
1994 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1995 have new members for a host name. The SSL data structure has an
1996 additional member SSL_CTX *initial_ctx so that new sessions can be
1997 stored in that context to allow for session resumption, even after the
1998 SSL has been switched to a new SSL_CTX in reaction to a client's
1999 server_name extension.
2001 New functions (subject to change):
2003 SSL_get_servername()
2004 SSL_get_servername_type()
2007 New CTRL codes and macros (subject to change):
2009 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2010 - SSL_CTX_set_tlsext_servername_callback()
2011 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2012 - SSL_CTX_set_tlsext_servername_arg()
2013 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2015 openssl s_client has a new '-servername ...' option.
2017 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2018 '-key2 ...', '-servername_fatal' (subject to change). This allows
2019 testing the HostName extension for a specific single host name ('-cert'
2020 and '-key' remain fallbacks for handshakes without HostName
2021 negotiation). If the unrecogninzed_name alert has to be sent, this by
2022 default is a warning; it becomes fatal with the '-servername_fatal'
2025 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2027 *) Add AES and SSE2 assembly language support to VC++ build.
2030 *) Mitigate attack on final subtraction in Montgomery reduction.
2033 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2034 (which previously caused an internal error).
2037 *) Squeeze another 10% out of IGE mode when in != out.
2040 *) AES IGE mode speedup.
2041 [Dean Gaudet (Google)]
2043 *) Add the Korean symmetric 128-bit cipher SEED (see
2044 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2045 add SEED ciphersuites from RFC 4162:
2047 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2048 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2049 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2050 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2052 To minimize changes between patchlevels in the OpenSSL 0.9.8
2053 series, SEED remains excluded from compilation unless OpenSSL
2054 is configured with 'enable-seed'.
2055 [KISA, Bodo Moeller]
2057 *) Mitigate branch prediction attacks, which can be practical if a
2058 single processor is shared, allowing a spy process to extract
2059 information. For detailed background information, see
2060 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2061 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2062 and Necessary Software Countermeasures"). The core of the change
2063 are new versions BN_div_no_branch() and
2064 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2065 respectively, which are slower, but avoid the security-relevant
2066 conditional branches. These are automatically called by BN_div()
2067 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2068 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2069 remove a conditional branch.
2071 BN_FLG_CONSTTIME is the new name for the previous
2072 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2073 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2074 in the exponent causes BN_mod_exp_mont() to use the alternative
2075 implementation in BN_mod_exp_mont_consttime().) The old name
2076 remains as a deprecated alias.
2078 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2079 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2080 constant-time implementations for more than just exponentiation.
2081 Here too the old name is kept as a deprecated alias.
2083 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2084 the BN_BLINDING structure gets an independent copy of the
2085 modulus. This means that the previous "BIGNUM *m" argument to
2086 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2087 essentially becomes "const BIGNUM *m", although we can't actually
2088 change this in the header file before 0.9.9. It allows
2089 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2090 enable BN_FLG_CONSTTIME.
2092 [Matthew D Wood (Intel Corp)]
2094 *) In the SSL/TLS server implementation, be strict about session ID
2095 context matching (which matters if an application uses a single
2096 external cache for different purposes). Previously,
2097 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2098 set. This did ensure strict client verification, but meant that,
2099 with applications using a single external cache for quite
2100 different requirements, clients could circumvent ciphersuite
2101 restrictions for a given session ID context by starting a session
2102 in a different context.
2105 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2106 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2107 authentication-only ciphersuites.
2110 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2111 not complete and could lead to a possible single byte overflow
2112 (CVE-2007-5135) [Ben Laurie]
2114 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2116 *) Since AES128 and AES256 (and similarly Camellia128 and
2117 Camellia256) share a single mask bit in the logic of
2118 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2119 kludge to work properly if AES128 is available and AES256 isn't
2120 (or if Camellia128 is available and Camellia256 isn't).
2123 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2124 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2125 When a point or a seed is encoded in a BIT STRING, we need to
2126 prevent the removal of trailing zero bits to get the proper DER
2127 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2128 of a NamedBitList, for which trailing 0 bits need to be removed.)
2131 *) Have SSL/TLS server implementation tolerate "mismatched" record
2132 protocol version while receiving ClientHello even if the
2133 ClientHello is fragmented. (The server can't insist on the
2134 particular protocol version it has chosen before the ServerHello
2135 message has informed the client about his choice.)
2138 *) Add RFC 3779 support.
2139 [Rob Austein for ARIN, Ben Laurie]
2141 *) Load error codes if they are not already present instead of using a
2142 static variable. This allows them to be cleanly unloaded and reloaded.
2143 Improve header file function name parsing.
2146 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2147 or CAPABILITY handshake as required by RFCs.
2150 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2152 *) Introduce limits to prevent malicious keys being able to
2153 cause a denial of service. (CVE-2006-2940)
2154 [Steve Henson, Bodo Moeller]
2156 *) Fix ASN.1 parsing of certain invalid structures that can result
2157 in a denial of service. (CVE-2006-2937) [Steve Henson]
2159 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2160 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2162 *) Fix SSL client code which could crash if connecting to a
2163 malicious SSLv2 server. (CVE-2006-4343)
2164 [Tavis Ormandy and Will Drewry, Google Security Team]
2166 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2167 match only those. Before that, "AES256-SHA" would be interpreted
2168 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2169 the same strength classification in 0.9.7h) as we currently only
2170 have a single AES bit in the ciphersuite description bitmap.
2171 That change, however, also applied to ciphersuite strings such as
2172 "RC4-MD5" that intentionally matched multiple ciphersuites --
2173 namely, SSL 2.0 ciphersuites in addition to the more common ones
2174 from SSL 3.0/TLS 1.0.
2176 So we change the selection algorithm again: Naming an explicit
2177 ciphersuite selects this one ciphersuite, and any other similar
2178 ciphersuite (same bitmap) from *other* protocol versions.
2179 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2180 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2182 Since SSL 2.0 does not have any ciphersuites for which the
2183 128/256 bit distinction would be relevant, this works for now.
2184 The proper fix will be to use different bits for AES128 and
2185 AES256, which would have avoided the problems from the beginning;
2186 however, bits are scarce, so we can only do this in a new release
2187 (not just a patchlevel) when we can change the SSL_CIPHER
2188 definition to split the single 'unsigned long mask' bitmap into
2189 multiple values to extend the available space.
2193 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2195 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2196 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2198 *) Add AES IGE and biIGE modes.
2201 *) Change the Unix randomness entropy gathering to use poll() when
2202 possible instead of select(), since the latter has some
2203 undesirable limitations.
2204 [Darryl Miles via Richard Levitte and Bodo Moeller]
2206 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2207 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2208 cannot be implicitly activated as part of, e.g., the "AES" alias.
2209 However, please upgrade to OpenSSL 0.9.9[-dev] for
2210 non-experimental use of the ECC ciphersuites to get TLS extension
2211 support, which is required for curve and point format negotiation
2212 to avoid potential handshake problems.
2215 *) Disable rogue ciphersuites:
2217 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2218 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2219 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2221 The latter two were purportedly from
2222 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2225 Also deactivate the remaining ciphersuites from
2226 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2227 unofficial, and the ID has long expired.
2230 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2231 dual-core machines) and other potential thread-safety issues.
2234 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2235 versions), which is now available for royalty-free use
2236 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2237 Also, add Camellia TLS ciphersuites from RFC 4132.
2239 To minimize changes between patchlevels in the OpenSSL 0.9.8
2240 series, Camellia remains excluded from compilation unless OpenSSL
2241 is configured with 'enable-camellia'.
2244 *) Disable the padding bug check when compression is in use. The padding
2245 bug check assumes the first packet is of even length, this is not
2246 necessarily true if compresssion is enabled and can result in false
2247 positives causing handshake failure. The actual bug test is ancient
2248 code so it is hoped that implementations will either have fixed it by
2249 now or any which still have the bug do not support compression.
2252 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2254 *) When applying a cipher rule check to see if string match is an explicit
2255 cipher suite and only match that one cipher suite if it is.
2258 *) Link in manifests for VC++ if needed.
2259 [Austin Ziegler <halostatue@gmail.com>]
2261 *) Update support for ECC-based TLS ciphersuites according to
2262 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2263 TLS extensions, which are supported starting with the 0.9.9
2264 branch, not in the OpenSSL 0.9.8 branch).
2267 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2268 opaque EVP_CIPHER_CTX handling.
2271 *) Fixes and enhancements to zlib compression code. We now only use
2272 "zlib1.dll" and use the default __cdecl calling convention on Win32
2273 to conform with the standards mentioned here:
2274 http://www.zlib.net/DLL_FAQ.txt
2275 Static zlib linking now works on Windows and the new --with-zlib-include
2276 --with-zlib-lib options to Configure can be used to supply the location
2277 of the headers and library. Gracefully handle case where zlib library
2281 *) Several fixes and enhancements to the OID generation code. The old code
2282 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2283 handle numbers larger than ULONG_MAX, truncated printing and had a
2284 non standard OBJ_obj2txt() behaviour.
2287 *) Add support for building of engines under engine/ as shared libraries
2288 under VC++ build system.
2291 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2292 Hopefully, we will not see any false combination of paths any more.
2295 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2297 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2298 (part of SSL_OP_ALL). This option used to disable the
2299 countermeasure against man-in-the-middle protocol-version
2300 rollback in the SSL 2.0 server implementation, which is a bad
2301 idea. (CVE-2005-2969)
2303 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2304 for Information Security, National Institute of Advanced Industrial
2305 Science and Technology [AIST], Japan)]
2307 *) Add two function to clear and return the verify parameter flags.
2310 *) Keep cipherlists sorted in the source instead of sorting them at
2311 runtime, thus removing the need for a lock.
2314 *) Avoid some small subgroup attacks in Diffie-Hellman.
2315 [Nick Mathewson and Ben Laurie]
2317 *) Add functions for well-known primes.
2320 *) Extended Windows CE support.
2321 [Satoshi Nakamura and Andy Polyakov]
2323 *) Initialize SSL_METHOD structures at compile time instead of during
2324 runtime, thus removing the need for a lock.
2327 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2328 attempting to decrypt each encrypted key in turn. Add support to
2332 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2334 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2337 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2340 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2341 key into the same file any more.
2344 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2347 *) Add -utf8 command line and config file option to 'ca'.
2348 [Stefan <stf@udoma.org]
2350 *) Removed the macro des_crypt(), as it seems to conflict with some
2351 libraries. Use DES_crypt().
2354 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2355 involves renaming the source and generated shared-libs for
2356 both. The engines will accept the corrected or legacy ids
2357 ('ncipher' and '4758_cca' respectively) when binding. NB,
2358 this only applies when building 'shared'.
2359 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2361 *) Add attribute functions to EVP_PKEY structure. Modify
2362 PKCS12_create() to recognize a CSP name attribute and
2363 use it. Make -CSP option work again in pkcs12 utility.
2366 *) Add new functionality to the bn blinding code:
2367 - automatic re-creation of the BN_BLINDING parameters after
2368 a fixed number of uses (currently 32)
2369 - add new function for parameter creation
2370 - introduce flags to control the update behaviour of the
2371 BN_BLINDING parameters
2372 - hide BN_BLINDING structure
2373 Add a second BN_BLINDING slot to the RSA structure to improve
2374 performance when a single RSA object is shared among several
2378 *) Add support for DTLS.
2379 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2381 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2382 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2385 *) Remove buggy and incompletet DH cert support from
2386 ssl/ssl_rsa.c and ssl/s3_both.c
2389 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2390 the apps/openssl applications.
2393 *) Compile clean with "-Wall -Wmissing-prototypes
2394 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2395 DEBUG_SAFESTACK must also be set.
2398 *) Change ./Configure so that certain algorithms can be disabled by default.
2399 The new counterpiece to "no-xxx" is "enable-xxx".
2401 The patented RC5 and MDC2 algorithms will now be disabled unless
2402 "enable-rc5" and "enable-mdc2", respectively, are specified.
2404 (IDEA remains enabled despite being patented. This is because IDEA
2405 is frequently required for interoperability, and there is no license
2406 fee for non-commercial use. As before, "no-idea" can be used to
2407 avoid this algorithm.)
2411 *) Add processing of proxy certificates (see RFC 3820). This work was
2412 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2413 EGEE (Enabling Grids for E-science in Europe).
2416 *) RC4 performance overhaul on modern architectures/implementations, such
2417 as Intel P4, IA-64 and AMD64.
2420 *) New utility extract-section.pl. This can be used specify an alternative
2421 section number in a pod file instead of having to treat each file as
2422 a separate case in Makefile. This can be done by adding two lines to the
2425 =for comment openssl_section:XXX
2427 The blank line is mandatory.
2431 *) New arguments -certform, -keyform and -pass for s_client and s_server
2432 to allow alternative format key and certificate files and passphrase
2436 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2437 update associated structures and add various utility functions.
2439 Add new policy related verify parameters, include policy checking in
2440 standard verify code. Enhance 'smime' application with extra parameters
2441 to support policy checking and print out.
2444 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2445 Nehemiah processors. These extensions support AES encryption in hardware
2446 as well as RNG (though RNG support is currently disabled).
2447 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2449 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2452 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2453 [Andy Polyakov and a number of other people]
2455 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2456 implementation contributed by IBM.
2457 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2459 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2460 exponent rather than 'unsigned long'. There is a corresponding change to
2461 the new 'rsa_keygen' element of the RSA_METHOD structure.
2462 [Jelte Jansen, Geoff Thorpe]
2464 *) Functionality for creating the initial serial number file is now
2465 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2467 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2468 number file to 1, which is bound to cause problems. To avoid
2469 the problems while respecting compatibility between different 0.9.7
2470 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2471 CA.pl for serial number initialization. With the new release 0.9.8,
2472 we can fix the problem directly in the 'ca' utility.)
2475 *) Reduced header interdepencies by declaring more opaque objects in
2476 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2477 give fewer recursive includes, which could break lazy source code - so
2478 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2479 developers should define this symbol when building and using openssl to
2480 ensure they track the recommended behaviour, interfaces, [etc], but
2481 backwards-compatible behaviour prevails when this isn't defined.
2484 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2487 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2488 This will generate a random key of the appropriate length based on the
2489 cipher context. The EVP_CIPHER can provide its own random key generation
2490 routine to support keys of a specific form. This is used in the des and
2491 3des routines to generate a key of the correct parity. Update S/MIME
2492 code to use new functions and hence generate correct parity DES keys.
2493 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2494 valid (weak or incorrect parity).
2497 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2498 as looking them up. This is useful when the verified structure may contain
2499 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2500 present unless the new PKCS7_NO_CRL flag is asserted.
2503 *) Extend ASN1 oid configuration module. It now additionally accepts the
2506 shortName = some long name, 1.2.3.4
2509 *) Reimplemented the BN_CTX implementation. There is now no more static
2510 limitation on the number of variables it can handle nor the depth of the
2511 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2512 information can now expand as required, and rather than having a single
2513 static array of bignums, BN_CTX now uses a linked-list of such arrays
2514 allowing it to expand on demand whilst maintaining the usefulness of
2515 BN_CTX's "bundling".
2518 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2519 to allow all RSA operations to function using a single BN_CTX.
2522 *) Preliminary support for certificate policy evaluation and checking. This
2523 is initially intended to pass the tests outlined in "Conformance Testing
2524 of Relying Party Client Certificate Path Processing Logic" v1.07.
2527 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2528 remained unused and not that useful. A variety of other little bignum
2529 tweaks and fixes have also been made continuing on from the audit (see
2533 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2534 associated ASN1, EVP and SSL functions and old ASN1 macros.
2537 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2538 and this should never fail. So the return value from the use of
2539 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2540 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2543 *) BN_CTX_get() should return zero-valued bignums, providing the same
2544 initialised value as BN_new().
2545 [Geoff Thorpe, suggested by Ulf Möller]
2547 *) Support for inhibitAnyPolicy certificate extension.
2550 *) An audit of the BIGNUM code is underway, for which debugging code is
2551 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2552 is considered valid when processing BIGNUMs, and causes execution to
2553 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2554 further steps are taken to deliberately pollute unused data in BIGNUM
2555 structures to try and expose faulty code further on. For now, openssl will
2556 (in its default mode of operation) continue to tolerate the inconsistent
2557 forms that it has tolerated in the past, but authors and packagers should
2558 consider trying openssl and their own applications when compiled with
2559 these debugging symbols defined. It will help highlight potential bugs in
2560 their own code, and will improve the test coverage for OpenSSL itself. At
2561 some point, these tighter rules will become openssl's default to improve
2562 maintainability, though the assert()s and other overheads will remain only
2563 in debugging configurations. See bn.h for more details.
2564 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2566 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2567 that can only be obtained through BN_CTX_new() (which implicitly
2568 initialises it). The presence of this function only made it possible
2569 to overwrite an existing structure (and cause memory leaks).
2572 *) Because of the callback-based approach for implementing LHASH as a
2573 template type, lh_insert() adds opaque objects to hash-tables and
2574 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2575 to clean up those corresponding objects before destroying the hash table
2576 (and losing the object pointers). So some over-zealous constifications in
2577 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2578 objects as "const" and the lh_doall[_arg] callback wrappers are not
2579 prototyped to have "const" restrictions on the object pointers they are
2580 given (and so aren't required to cast them away any more).
2583 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2584 (speed) prefers to use its own implementation. The two implementations
2585 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2586 its object type properly exposed (MS_TM) instead of casting to/from "char
2587 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2588 aren't necessarily the greatest nomenclatures - but this is what was used
2589 internally to the implementation so I've used that for now.
2592 *) Ensure that deprecated functions do not get compiled when
2593 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2594 the self-tests were still using deprecated key-generation functions so
2595 these have been updated also.
2598 *) Reorganise PKCS#7 code to separate the digest location functionality
2599 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2600 New function PKCS7_set_digest() to set the digest type for PKCS#7
2601 digestedData type. Add additional code to correctly generate the
2602 digestedData type and add support for this type in PKCS7 initialization
2606 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2607 structure of type "other".
2610 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2611 sure the loop does correctly stop and breaking ("division by zero")
2612 modulus operations are not performed. The (pre-generated) prime
2613 table crypto/bn/bn_prime.h was already correct, but it could not be
2614 re-generated on some platforms because of the "division by zero"
2615 situation in the script.
2616 [Ralf S. Engelschall]
2618 *) Update support for ECC-based TLS ciphersuites according to
2619 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2620 SHA-1 now is only used for "small" curves (where the
2621 representation of a field element takes up to 24 bytes); for
2622 larger curves, the field element resulting from ECDH is directly
2623 used as premaster secret.
2624 [Douglas Stebila (Sun Microsystems Laboratories)]
2626 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2627 curve secp160r1 to the tests.
2628 [Douglas Stebila (Sun Microsystems Laboratories)]
2630 *) Add the possibility to load symbols globally with DSO.
2631 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2633 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2634 control of the error stack.
2637 *) Add support for STORE in ENGINE.
2640 *) Add the STORE type. The intention is to provide a common interface
2641 to certificate and key stores, be they simple file-based stores, or
2642 HSM-type store, or LDAP stores, or...
2643 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2646 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2647 pass a list of arguments to any function as well as provide a way
2648 for a function to pass data back to the caller.
2651 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2652 works like BUF_strdup() but can be used to duplicate a portion of
2653 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2657 *) Add the function sk_find_ex() which works like sk_find(), but will
2658 return an index to an element even if an exact match couldn't be
2659 found. The index is guaranteed to point at the element where the
2660 searched-for key would be inserted to preserve sorting order.
2663 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2664 takes an extra flags argument for optional functionality. Currently,
2665 the following flags are defined:
2667 OBJ_BSEARCH_VALUE_ON_NOMATCH
2668 This one gets OBJ_bsearch_ex() to return a pointer to the first
2669 element where the comparing function returns a negative or zero
2672 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2673 This one gets OBJ_bsearch_ex() to return a pointer to the first
2674 element where the comparing function returns zero. This is useful
2675 if there are more than one element where the comparing function
2679 *) Make it possible to create self-signed certificates with 'openssl ca'
2680 in such a way that the self-signed certificate becomes part of the
2681 CA database and uses the same mechanisms for serial number generation
2682 as all other certificate signing. The new flag '-selfsign' enables
2683 this functionality. Adapt CA.sh and CA.pl.in.
2686 *) Add functionality to check the public key of a certificate request
2687 against a given private. This is useful to check that a certificate
2688 request can be signed by that key (self-signing).
2691 *) Make it possible to have multiple active certificates with the same
2692 subject in the CA index file. This is done only if the keyword
2693 'unique_subject' is set to 'no' in the main CA section (default
2694 if 'CA_default') of the configuration file. The value is saved
2695 with the database itself in a separate index attribute file,
2696 named like the index file with '.attr' appended to the name.
2699 *) Generate muti valued AVAs using '+' notation in config files for
2703 *) Support for nameConstraints certificate extension.
2706 *) Support for policyConstraints certificate extension.
2709 *) Support for policyMappings certificate extension.
2712 *) Make sure the default DSA_METHOD implementation only uses its
2713 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2714 and change its own handlers to be NULL so as to remove unnecessary
2715 indirection. This lets alternative implementations fallback to the
2716 default implementation more easily.
2719 *) Support for directoryName in GeneralName related extensions
2723 *) Make it possible to link applications using Makefile.shared.
2724 Make that possible even when linking against static libraries!
2727 *) Support for single pass processing for S/MIME signing. This now
2728 means that S/MIME signing can be done from a pipe, in addition
2729 cleartext signing (multipart/signed type) is effectively streaming
2730 and the signed data does not need to be all held in memory.
2732 This is done with a new flag PKCS7_STREAM. When this flag is set
2733 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2734 is done after the data is output (and digests calculated) in
2735 SMIME_write_PKCS7().
2738 *) Add full support for -rpath/-R, both in shared libraries and
2739 applications, at least on the platforms where it's known how
2743 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2744 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2745 will now compute a table of multiples of the generator that
2746 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2747 faster (notably in the case of a single point multiplication,
2748 scalar * generator).
2749 [Nils Larsch, Bodo Moeller]
2751 *) IPv6 support for certificate extensions. The various extensions
2752 which use the IP:a.b.c.d can now take IPv6 addresses using the
2753 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2757 *) Added an ENGINE that implements RSA by performing private key
2758 exponentiations with the GMP library. The conversions to and from
2759 GMP's mpz_t format aren't optimised nor are any montgomery forms
2760 cached, and on x86 it appears OpenSSL's own performance has caught up.
2761 However there are likely to be other architectures where GMP could
2762 provide a boost. This ENGINE is not built in by default, but it can be
2763 specified at Configure time and should be accompanied by the necessary
2764 linker additions, eg;
2765 ./config -DOPENSSL_USE_GMP -lgmp
2768 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2769 testing availability of engines with "-t" - the old behaviour is
2770 produced by increasing the feature's verbosity with "-tt".
2773 *) ECDSA routines: under certain error conditions uninitialized BN objects
2774 could be freed. Solution: make sure initialization is performed early
2775 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2779 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2780 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2781 software implementations. For DSA and DH, parameter generation can
2782 also be overriden by providing the appropriate method callbacks.
2785 *) Change the "progress" mechanism used in key-generation and
2786 primality testing to functions that take a new BN_GENCB pointer in
2787 place of callback/argument pairs. The new API functions have "_ex"
2788 postfixes and the older functions are reimplemented as wrappers for
2789 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2790 declarations of the old functions to help (graceful) attempts to
2791 migrate to the new functions. Also, the new key-generation API
2792 functions operate on a caller-supplied key-structure and return
2793 success/failure rather than returning a key or NULL - this is to
2794 help make "keygen" another member function of RSA_METHOD etc.
2796 Example for using the new callback interface:
2798 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2802 BN_GENCB_set(&my_cb, my_callback, my_arg);
2804 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2805 /* For the meaning of a, b in calls to my_callback(), see the
2806 * documentation of the function that calls the callback.
2807 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2808 * my_callback should return 1 if it wants BN_is_prime_ex()
2809 * to continue, or 0 to stop.
2814 *) Change the ZLIB compression method to be stateful, and make it
2815 available to TLS with the number defined in
2816 draft-ietf-tls-compression-04.txt.
2819 *) Add the ASN.1 structures and functions for CertificatePair, which
2820 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2822 CertificatePair ::= SEQUENCE {
2823 forward [0] Certificate OPTIONAL,
2824 reverse [1] Certificate OPTIONAL,
2825 -- at least one of the pair shall be present -- }
2827 Also implement the PEM functions to read and write certificate
2828 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2830 This needed to be defined, mostly for the sake of the LDAP
2831 attribute crossCertificatePair, but may prove useful elsewhere as
2835 *) Make it possible to inhibit symlinking of shared libraries in
2836 Makefile.shared, for Cygwin's sake.
2839 *) Extend the BIGNUM API by creating a function
2840 void BN_set_negative(BIGNUM *a, int neg);
2841 and a macro that behave like
2842 int BN_is_negative(const BIGNUM *a);
2844 to avoid the need to access 'a->neg' directly in applications.
2847 *) Implement fast modular reduction for pseudo-Mersenne primes
2848 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2849 EC_GROUP_new_curve_GFp() will now automatically use this
2851 [Nils Larsch <nla@trustcenter.de>]
2853 *) Add new lock type (CRYPTO_LOCK_BN).
2856 *) Change the ENGINE framework to automatically load engines
2857 dynamically from specific directories unless they could be
2858 found to already be built in or loaded. Move all the
2859 current engines except for the cryptodev one to a new
2861 The engines in engines/ are built as shared libraries if
2862 the "shared" options was given to ./Configure or ./config.
2863 Otherwise, they are inserted in libcrypto.a.
2864 /usr/local/ssl/engines is the default directory for dynamic
2865 engines, but that can be overriden at configure time through
2866 the usual use of --prefix and/or --openssldir, and at run
2867 time with the environment variable OPENSSL_ENGINES.
2868 [Geoff Thorpe and Richard Levitte]
2870 *) Add Makefile.shared, a helper makefile to build shared
2871 libraries. Addapt Makefile.org.
2874 *) Add version info to Win32 DLLs.
2875 [Peter 'Luna' Runestig" <peter@runestig.com>]
2877 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2878 can be added using this API to created arbitrary PKCS#12
2879 files while avoiding the low level API.
2881 New options to PKCS12_create(), key or cert can be NULL and
2882 will then be omitted from the output file. The encryption
2883 algorithm NIDs can be set to -1 for no encryption, the mac
2884 iteration count can be set to 0 to omit the mac.
2886 Enhance pkcs12 utility by making the -nokeys and -nocerts
2887 options work when creating a PKCS#12 file. New option -nomac
2888 to omit the mac, NONE can be set for an encryption algorithm.
2889 New code is modified to use the enhanced PKCS12_create()
2890 instead of the low level API.
2893 *) Extend ASN1 encoder to support indefinite length constructed
2894 encoding. This can output sequences tags and octet strings in
2895 this form. Modify pk7_asn1.c to support indefinite length
2896 encoding. This is experimental and needs additional code to
2897 be useful, such as an ASN1 bio and some enhanced streaming
2900 Extend template encode functionality so that tagging is passed
2901 down to the template encoder.
2904 *) Let 'openssl req' fail if an argument to '-newkey' is not
2905 recognized instead of using RSA as a default.
2908 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2909 As these are not official, they are not included in "ALL";
2910 the "ECCdraft" ciphersuite group alias can be used to select them.
2911 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2913 *) Add ECDH engine support.
2914 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2916 *) Add ECDH in new directory crypto/ecdh/.
2917 [Douglas Stebila (Sun Microsystems Laboratories)]
2919 *) Let BN_rand_range() abort with an error after 100 iterations
2920 without success (which indicates a broken PRNG).
2923 *) Change BN_mod_sqrt() so that it verifies that the input value
2924 is really the square of the return value. (Previously,
2925 BN_mod_sqrt would show GIGO behaviour.)
2928 *) Add named elliptic curves over binary fields from X9.62, SECG,
2929 and WAP/WTLS; add OIDs that were still missing.
2931 [Sheueling Chang Shantz and Douglas Stebila
2932 (Sun Microsystems Laboratories)]
2934 *) Extend the EC library for elliptic curves over binary fields
2935 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2938 EC_GF2m_simple_method
2942 EC_GROUP_new_curve_GF2m
2943 EC_GROUP_set_curve_GF2m
2944 EC_GROUP_get_curve_GF2m
2945 EC_POINT_set_affine_coordinates_GF2m
2946 EC_POINT_get_affine_coordinates_GF2m
2947 EC_POINT_set_compressed_coordinates_GF2m
2949 Point compression for binary fields is disabled by default for
2950 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2953 As binary polynomials are represented as BIGNUMs, various members
2954 of the EC_GROUP and EC_POINT data structures can be shared
2955 between the implementations for prime fields and binary fields;
2956 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2957 are essentially identical to their ..._GFp counterparts.
2958 (For simplicity, the '..._GFp' prefix has been dropped from
2959 various internal method names.)
2961 An internal 'field_div' method (similar to 'field_mul' and
2962 'field_sqr') has been added; this is used only for binary fields.
2964 [Sheueling Chang Shantz and Douglas Stebila
2965 (Sun Microsystems Laboratories)]
2967 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2968 through methods ('mul', 'precompute_mult').
2970 The generic implementations (now internally called 'ec_wNAF_mul'
2971 and 'ec_wNAF_precomputed_mult') remain the default if these
2972 methods are undefined.
2974 [Sheueling Chang Shantz and Douglas Stebila
2975 (Sun Microsystems Laboratories)]
2977 *) New function EC_GROUP_get_degree, which is defined through
2978 EC_METHOD. For curves over prime fields, this returns the bit
2979 length of the modulus.
2981 [Sheueling Chang Shantz and Douglas Stebila
2982 (Sun Microsystems Laboratories)]
2984 *) New functions EC_GROUP_dup, EC_POINT_dup.
2985 (These simply call ..._new and ..._copy).
2987 [Sheueling Chang Shantz and Douglas Stebila
2988 (Sun Microsystems Laboratories)]
2990 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2991 Polynomials are represented as BIGNUMs (where the sign bit is not
2992 used) in the following functions [macros]:
2995 BN_GF2m_sub [= BN_GF2m_add]
2996 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2997 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2998 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3000 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3001 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3002 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3003 BN_GF2m_cmp [= BN_ucmp]
3005 (Note that only the 'mod' functions are actually for fields GF(2^m).
3006 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3008 For some functions, an the irreducible polynomial defining a
3009 field can be given as an 'unsigned int[]' with strictly
3010 decreasing elements giving the indices of those bits that are set;
3011 i.e., p[] represents the polynomial
3012 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3014 p[0] > p[1] > ... > p[k] = 0.
3015 This applies to the following functions:
3020 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3021 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3023 BN_GF2m_mod_sqrt_arr
3024 BN_GF2m_mod_solve_quad_arr
3028 Conversion can be performed by the following functions:
3033 bntest.c has additional tests for binary polynomial arithmetic.
3035 Two implementations for BN_GF2m_mod_div() are available.
3036 The default algorithm simply uses BN_GF2m_mod_inv() and
3037 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3038 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3039 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3041 [Sheueling Chang Shantz and Douglas Stebila
3042 (Sun Microsystems Laboratories)]
3044 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3045 functionality is disabled at compile-time.
3046 [Douglas Stebila <douglas.stebila@sun.com>]
3048 *) Change default behaviour of 'openssl asn1parse' so that more
3049 information is visible when viewing, e.g., a certificate:
3051 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3052 mode the content of non-printable OCTET STRINGs is output in a
3053 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3054 avoid the appearance of a printable string.
3055 [Nils Larsch <nla@trustcenter.de>]
3057 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3059 EC_GROUP_set_asn1_flag()
3060 EC_GROUP_get_asn1_flag()
3061 EC_GROUP_set_point_conversion_form()
3062 EC_GROUP_get_point_conversion_form()
3063 These control ASN1 encoding details:
3064 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3065 has been set to OPENSSL_EC_NAMED_CURVE.
3066 - Points are encoded in uncompressed form by default; options for
3067 asn1_for are as for point2oct, namely
3068 POINT_CONVERSION_COMPRESSED
3069 POINT_CONVERSION_UNCOMPRESSED
3070 POINT_CONVERSION_HYBRID
3072 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3075 EC_GROUP_get0_seed()
3076 EC_GROUP_get_seed_len()
3077 This is used only for ASN1 purposes (so far).
3078 [Nils Larsch <nla@trustcenter.de>]
3080 *) Add 'field_type' member to EC_METHOD, which holds the NID
3081 of the appropriate field type OID. The new function
3082 EC_METHOD_get_field_type() returns this value.
3083 [Nils Larsch <nla@trustcenter.de>]
3088 EC_POINT_point2hex()
3089 EC_POINT_hex2point()
3090 providing useful interfaces to EC_POINT_point2oct() and
3091 EC_POINT_oct2point().
3092 [Nils Larsch <nla@trustcenter.de>]
3094 *) Change internals of the EC library so that the functions
3095 EC_GROUP_set_generator()
3096 EC_GROUP_get_generator()
3097 EC_GROUP_get_order()
3098 EC_GROUP_get_cofactor()
3099 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3100 to methods, which would lead to unnecessary code duplication when
3101 adding different types of curves.
3102 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3104 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3105 arithmetic, and such that modified wNAFs are generated
3106 (which avoid length expansion in many cases).
3109 *) Add a function EC_GROUP_check_discriminant() (defined via
3110 EC_METHOD) that verifies that the curve discriminant is non-zero.
3112 Add a function EC_GROUP_check() that makes some sanity tests
3113 on a EC_GROUP, its generator and order. This includes
3114 EC_GROUP_check_discriminant().
3115 [Nils Larsch <nla@trustcenter.de>]
3117 *) Add ECDSA in new directory crypto/ecdsa/.
3119 Add applications 'openssl ecparam' and 'openssl ecdsa'
3120 (these are based on 'openssl dsaparam' and 'openssl dsa').
3122 ECDSA support is also included in various other files across the
3123 library. Most notably,
3124 - 'openssl req' now has a '-newkey ecdsa:file' option;
3125 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3126 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3127 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3128 them suitable for ECDSA where domain parameters must be
3129 extracted before the specific public key;
3130 - ECDSA engine support has been added.
3131 [Nils Larsch <nla@trustcenter.de>]
3133 *) Include some named elliptic curves, and add OIDs from X9.62,
3134 SECG, and WAP/WTLS. Each curve can be obtained from the new
3136 EC_GROUP_new_by_curve_name(),
3137 and the list of available named curves can be obtained with
3138 EC_get_builtin_curves().
3139 Also add a 'curve_name' member to EC_GROUP objects, which can be
3141 EC_GROUP_set_curve_name()
3142 EC_GROUP_get_curve_name()
3143 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3145 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3146 was actually never needed) and in BN_mul(). The removal in BN_mul()
3147 required a small change in bn_mul_part_recursive() and the addition
3148 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3149 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3150 bn_sub_words() and bn_add_words() except they take arrays with
3154 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3156 *) Cleanse PEM buffers before freeing them since they may contain
3158 [Benjamin Bennett <ben@psc.edu>]
3160 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3161 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3162 authentication-only ciphersuites.
3165 *) Since AES128 and AES256 share a single mask bit in the logic of
3166 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3167 kludge to work properly if AES128 is available and AES256 isn't.
3170 *) Expand security boundary to match 1.1.1 module.
3173 *) Remove redundant features: hash file source, editing of test vectors
3174 modify fipsld to use external fips_premain.c signature.
3177 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3178 run algorithm test programs.
3181 *) Make algorithm test programs more tolerant of whitespace.
3184 *) Have SSL/TLS server implementation tolerate "mismatched" record
3185 protocol version while receiving ClientHello even if the
3186 ClientHello is fragmented. (The server can't insist on the
3187 particular protocol version it has chosen before the ServerHello
3188 message has informed the client about his choice.)
3191 *) Load error codes if they are not already present instead of using a
3192 static variable. This allows them to be cleanly unloaded and reloaded.
3195 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3197 *) Introduce limits to prevent malicious keys being able to
3198 cause a denial of service. (CVE-2006-2940)
3199 [Steve Henson, Bodo Moeller]
3201 *) Fix ASN.1 parsing of certain invalid structures that can result
3202 in a denial of service. (CVE-2006-2937) [Steve Henson]
3204 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3205 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3207 *) Fix SSL client code which could crash if connecting to a
3208 malicious SSLv2 server. (CVE-2006-4343)
3209 [Tavis Ormandy and Will Drewry, Google Security Team]
3211 *) Change ciphersuite string processing so that an explicit
3212 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3213 will no longer include "AES128-SHA"), and any other similar
3214 ciphersuite (same bitmap) from *other* protocol versions (so that
3215 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3216 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3217 changes from 0.9.8b and 0.9.8d.
3220 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3222 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3223 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3225 *) Change the Unix randomness entropy gathering to use poll() when
3226 possible instead of select(), since the latter has some
3227 undesirable limitations.
3228 [Darryl Miles via Richard Levitte and Bodo Moeller]
3230 *) Disable rogue ciphersuites:
3232 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3233 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3234 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3236 The latter two were purportedly from
3237 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3240 Also deactive the remaining ciphersuites from
3241 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3242 unofficial, and the ID has long expired.
3245 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3246 dual-core machines) and other potential thread-safety issues.
3249 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3251 *) Adapt fipsld and the build system to link against the validated FIPS
3252 module in FIPS mode.
3255 *) Fixes for VC++ 2005 build under Windows.
3258 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3259 from a Windows bash shell such as MSYS. It is autodetected from the
3260 "config" script when run from a VC++ environment. Modify standard VC++
3261 build to use fipscanister.o from the GNU make build.
3264 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3266 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3267 The value now differs depending on if you build for FIPS or not.
3268 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3269 safely run with a non-FIPSed libcrypto, as it may crash because of
3270 the difference induced by this change.
3273 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3275 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3276 (part of SSL_OP_ALL). This option used to disable the
3277 countermeasure against man-in-the-middle protocol-version
3278 rollback in the SSL 2.0 server implementation, which is a bad
3279 idea. (CVE-2005-2969)
3281 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3282 for Information Security, National Institute of Advanced Industrial
3283 Science and Technology [AIST], Japan)]
3285 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3286 mainly for FIPS compliance and not fully integrated at this stage.
3289 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3290 the exponentiation using a fixed-length exponent. (Otherwise,
3291 the information leaked through timing could expose the secret key
3292 after many signatures; cf. Bleichenbacher's attack on DSA with
3296 *) Make a new fixed-window mod_exp implementation the default for
3297 RSA, DSA, and DH private-key operations so that the sequence of
3298 squares and multiplies and the memory access pattern are
3299 independent of the particular secret key. This will mitigate
3300 cache-timing and potential related attacks.
3302 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3303 and this is automatically used by BN_mod_exp_mont() if the new flag
3304 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3305 will use this BN flag for private exponents unless the flag
3306 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3307 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3309 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3311 *) Change the client implementation for SSLv23_method() and
3312 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3313 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3314 (Previously, the SSL 2.0 backwards compatible Client Hello
3315 message format would be used even with SSL_OP_NO_SSLv2.)
3318 *) Add support for smime-type MIME parameter in S/MIME messages which some
3322 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3323 a threadsafe manner. Modify rsa code to use new function and add calls
3324 to dsa and dh code (which had race conditions before).
3327 *) Include the fixed error library code in the C error file definitions
3328 instead of fixing them up at runtime. This keeps the error code
3329 structures constant.
3332 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3334 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3337 *) Fixes for newer kerberos headers. NB: the casts are needed because
3338 the 'length' field is signed on one version and unsigned on another
3339 with no (?) obvious way to tell the difference, without these VC++
3340 complains. Also the "definition" of FAR (blank) is no longer included
3341 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3342 some needed definitions.
3345 *) Undo Cygwin change.
3348 *) Added support for proxy certificates according to RFC 3820.
3349 Because they may be a security thread to unaware applications,
3350 they must be explicitely allowed in run-time. See
3351 docs/HOWTO/proxy_certificates.txt for further information.
3354 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3356 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3357 server and client random values. Previously
3358 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3359 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3361 This change has negligible security impact because:
3363 1. Server and client random values still have 24 bytes of pseudo random
3366 2. Server and client random values are sent in the clear in the initial
3369 3. The master secret is derived using the premaster secret (48 bytes in
3370 size for static RSA ciphersuites) as well as client server and random
3373 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3376 [Stephen Henson, reported by UK NISCC]
3378 *) Use Windows randomness collection on Cygwin.
3381 *) Fix hang in EGD/PRNGD query when communication socket is closed
3382 prematurely by EGD/PRNGD.
3383 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3385 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3388 *) Back-port of selected performance improvements from development
3389 branch, as well as improved support for PowerPC platforms.
3392 *) Add lots of checks for memory allocation failure, error codes to indicate
3393 failure and freeing up memory if a failure occurs.
3394 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3396 *) Add new -passin argument to dgst.
3399 *) Perform some character comparisons of different types in X509_NAME_cmp:
3400 this is needed for some certificates that reencode DNs into UTF8Strings
3401 (in violation of RFC3280) and can't or wont issue name rollover
3405 *) Make an explicit check during certificate validation to see that
3406 the CA setting in each certificate on the chain is correct. As a
3407 side effect always do the following basic checks on extensions,
3408 not just when there's an associated purpose to the check:
3410 - if there is an unhandled critical extension (unless the user
3411 has chosen to ignore this fault)
3412 - if the path length has been exceeded (if one is set at all)
3413 - that certain extensions fit the associated purpose (if one has
3417 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3419 *) Avoid a race condition when CRLs are checked in a multi threaded
3420 environment. This would happen due to the reordering of the revoked
3421 entries during signature checking and serial number lookup. Now the
3422 encoding is cached and the serial number sort performed under a lock.
3423 Add new STACK function sk_is_sorted().
3426 *) Add Delta CRL to the extension code.
3429 *) Various fixes to s3_pkt.c so alerts are sent properly.
3430 [David Holmes <d.holmes@f5.com>]
3432 *) Reduce the chances of duplicate issuer name and serial numbers (in
3433 violation of RFC3280) using the OpenSSL certificate creation utilities.
3434 This is done by creating a random 64 bit value for the initial serial
3435 number when a serial number file is created or when a self signed
3436 certificate is created using 'openssl req -x509'. The initial serial
3437 number file is created using 'openssl x509 -next_serial' in CA.pl
3438 rather than being initialized to 1.
3441 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3443 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3444 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3445 [Joe Orton, Steve Henson]
3447 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3449 [Joe Orton, Steve Henson]
3451 *) Make it possible to have multiple active certificates with the same
3452 subject in the CA index file. This is done only if the keyword
3453 'unique_subject' is set to 'no' in the main CA section (default
3454 if 'CA_default') of the configuration file. The value is saved
3455 with the database itself in a separate index attribute file,
3456 named like the index file with '.attr' appended to the name.
3459 *) X509 verify fixes. Disable broken certificate workarounds when
3460 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3461 keyUsage extension present. Don't accept CRLs with unhandled critical
3462 extensions: since verify currently doesn't process CRL extensions this
3463 rejects a CRL with *any* critical extensions. Add new verify error codes
3467 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3468 A clarification of RFC2560 will require the use of OCTET STRINGs and
3469 some implementations cannot handle the current raw format. Since OpenSSL
3470 copies and compares OCSP nonces as opaque blobs without any attempt at
3471 parsing them this should not create any compatibility issues.
3474 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3475 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3476 this HMAC (and other) operations are several times slower than OpenSSL
3480 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3481 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3483 *) Use the correct content when signing type "other".
3486 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3488 *) Fix various bugs revealed by running the NISCC test suite:
3490 Stop out of bounds reads in the ASN1 code when presented with
3491 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3493 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3495 If verify callback ignores invalid public key errors don't try to check
3496 certificate signature with the NULL public key.
3500 *) New -ignore_err option in ocsp application to stop the server
3501 exiting on the first error in a request.
3504 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3505 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3509 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3510 extra data after the compression methods not only for TLS 1.0
3511 but also for SSL 3.0 (as required by the specification).
3512 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3514 *) Change X509_certificate_type() to mark the key as exported/exportable
3515 when it's 512 *bits* long, not 512 bytes.
3518 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3519 blocks during encryption.
3522 *) Various fixes to base64 BIO and non blocking I/O. On write
3523 flushes were not handled properly if the BIO retried. On read
3524 data was not being buffered properly and had various logic bugs.
3525 This also affects blocking I/O when the data being decoded is a
3529 *) Various S/MIME bugfixes and compatibility changes:
3530 output correct application/pkcs7 MIME type if
3531 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3532 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3533 of files as .eml work). Correctly handle very long lines in MIME
3537 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3539 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3540 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3541 a protocol version number mismatch like a decryption error
3542 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3545 *) Turn on RSA blinding by default in the default implementation
3546 to avoid a timing attack. Applications that don't want it can call
3547 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3548 They would be ill-advised to do so in most cases.
3549 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3551 *) Change RSA blinding code so that it works when the PRNG is not
3552 seeded (in this case, the secret RSA exponent is abused as
3553 an unpredictable seed -- if it is not unpredictable, there
3554 is no point in blinding anyway). Make RSA blinding thread-safe
3555 by remembering the creator's thread ID in rsa->blinding and
3556 having all other threads use local one-time blinding factors
3557 (this requires more computation than sharing rsa->blinding, but
3558 avoids excessive locking; and if an RSA object is not shared
3559 between threads, blinding will still be very fast).
3562 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3563 ENGINE as defaults for all supported algorithms irrespective of
3564 the 'flags' parameter. 'flags' is now honoured, so applications
3565 should make sure they are passing it correctly.
3568 *) Target "mingw" now allows native Windows code to be generated in
3569 the Cygwin environment as well as with the MinGW compiler.
3572 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3574 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3575 via timing by performing a MAC computation even if incorrrect
3576 block cipher padding has been found. This is a countermeasure
3577 against active attacks where the attacker has to distinguish
3578 between bad padding and a MAC verification error. (CVE-2003-0078)
3580 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3581 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3582 Martin Vuagnoux (EPFL, Ilion)]
3584 *) Make the no-err option work as intended. The intention with no-err
3585 is not to have the whole error stack handling routines removed from
3586 libcrypto, it's only intended to remove all the function name and
3587 reason texts, thereby removing some of the footprint that may not
3588 be interesting if those errors aren't displayed anyway.
3590 NOTE: it's still possible for any application or module to have it's
3591 own set of error texts inserted. The routines are there, just not
3592 used by default when no-err is given.
3595 *) Add support for FreeBSD on IA64.
3596 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3598 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3599 Kerberos function mit_des_cbc_cksum(). Before this change,
3600 the value returned by DES_cbc_cksum() was like the one from
3601 mit_des_cbc_cksum(), except the bytes were swapped.
3602 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3604 *) Allow an application to disable the automatic SSL chain building.
3605 Before this a rather primitive chain build was always performed in
3606 ssl3_output_cert_chain(): an application had no way to send the
3607 correct chain if the automatic operation produced an incorrect result.
3609 Now the chain builder is disabled if either:
3611 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3613 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3615 The reasoning behind this is that an application would not want the
3616 auto chain building to take place if extra chain certificates are
3617 present and it might also want a means of sending no additional
3618 certificates (for example the chain has two certificates and the
3622 *) Add the possibility to build without the ENGINE framework.
3623 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3625 *) Under Win32 gmtime() can return NULL: check return value in
3626 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3629 *) DSA routines: under certain error conditions uninitialized BN objects
3630 could be freed. Solution: make sure initialization is performed early
3631 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3632 Nils Larsch <nla@trustcenter.de> via PR#459)
3635 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3636 checked on reconnect on the client side, therefore session resumption
3637 could still fail with a "ssl session id is different" error. This
3638 behaviour is masked when SSL_OP_ALL is used due to
3639 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3640 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3641 followup to PR #377.
3644 *) IA-32 assembler support enhancements: unified ELF targets, support
3645 for SCO/Caldera platforms, fix for Cygwin shared build.
3648 *) Add support for FreeBSD on sparc64. As a consequence, support for
3649 FreeBSD on non-x86 processors is separate from x86 processors on
3650 the config script, much like the NetBSD support.
3651 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3653 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3655 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3658 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3659 code (06) was taken as the first octet of the session ID and the last
3660 octet was ignored consequently. As a result SSLv2 client side session
3661 caching could not have worked due to the session ID mismatch between
3663 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3667 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3668 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3672 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3673 seems that in spite of existing for more than a year, many application
3674 author have done nothing to provide the necessary callbacks, which
3675 means that this particular engine will not work properly anywhere.
3676 This is a very unfortunate situation which forces us, in the name
3677 of usability, to give the hw_ncipher.c a static lock, which is part
3679 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3680 appear in 0.9.8 or later. We EXPECT application authors to have
3681 dealt properly with this when 0.9.8 is released (unless we actually
3682 make such changes in the libcrypto locking code that changes will
3683 have to be made anyway).
3686 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3687 octets have been read, EOF or an error occurs. Without this change
3688 some truncated ASN1 structures will not produce an error.
3691 *) Disable Heimdal support, since it hasn't been fully implemented.
3692 Still give the possibility to force the use of Heimdal, but with
3693 warnings and a request that patches get sent to openssl-dev.
3696 *) Add the VC-CE target, introduce the WINCE sysname, and add
3697 INSTALL.WCE and appropriate conditionals to make it build.
3698 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3700 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3701 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3702 edit numbers of the version.
3703 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3705 *) Introduce safe string copy and catenation functions
3706 (BUF_strlcpy() and BUF_strlcat()).
3707 [Ben Laurie (CHATS) and Richard Levitte]
3709 *) Avoid using fixed-size buffers for one-line DNs.
3710 [Ben Laurie (CHATS)]
3712 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3713 resizing buffers containing secrets, and use where appropriate.
3714 [Ben Laurie (CHATS)]
3716 *) Avoid using fixed size buffers for configuration file location.
3717 [Ben Laurie (CHATS)]
3719 *) Avoid filename truncation for various CA files.
3720 [Ben Laurie (CHATS)]
3722 *) Use sizeof in preference to magic numbers.
3723 [Ben Laurie (CHATS)]
3725 *) Avoid filename truncation in cert requests.
3726 [Ben Laurie (CHATS)]
3728 *) Add assertions to check for (supposedly impossible) buffer
3730 [Ben Laurie (CHATS)]
3732 *) Don't cache truncated DNS entries in the local cache (this could
3733 potentially lead to a spoofing attack).
3734 [Ben Laurie (CHATS)]
3736 *) Fix various buffers to be large enough for hex/decimal
3737 representations in a platform independent manner.
3738 [Ben Laurie (CHATS)]
3740 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3741 resizing buffers containing secrets, and use where appropriate.
3742 [Ben Laurie (CHATS)]
3744 *) Add BIO_indent() to avoid much slightly worrying code to do
3746 [Ben Laurie (CHATS)]
3748 *) Convert sprintf()/BIO_puts() to BIO_printf().
3749 [Ben Laurie (CHATS)]
3751 *) buffer_gets() could terminate with the buffer only half
3753 [Ben Laurie (CHATS)]
3755 *) Add assertions to prevent user-supplied crypto functions from
3756 overflowing internal buffers by having large block sizes, etc.
3757 [Ben Laurie (CHATS)]
3759 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3761 [Ben Laurie (CHATS)]
3763 *) Eliminate unused copy of key in RC4.
3764 [Ben Laurie (CHATS)]
3766 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3767 [Ben Laurie (CHATS)]
3769 *) Fix off-by-one error in EGD path.
3770 [Ben Laurie (CHATS)]
3772 *) If RANDFILE path is too long, ignore instead of truncating.
3773 [Ben Laurie (CHATS)]
3775 *) Eliminate unused and incorrectly sized X.509 structure
3777 [Ben Laurie (CHATS)]
3779 *) Eliminate unused and dangerous function knumber().
3780 [Ben Laurie (CHATS)]
3782 *) Eliminate unused and dangerous structure, KSSL_ERR.
3783 [Ben Laurie (CHATS)]
3785 *) Protect against overlong session ID context length in an encoded
3786 session object. Since these are local, this does not appear to be
3788 [Ben Laurie (CHATS)]
3790 *) Change from security patch (see 0.9.6e below) that did not affect
3791 the 0.9.6 release series:
3793 Remote buffer overflow in SSL3 protocol - an attacker could
3794 supply an oversized master key in Kerberos-enabled versions.
3796 [Ben Laurie (CHATS)]
3798 *) Change the SSL kerb5 codes to match RFC 2712.
3801 *) Make -nameopt work fully for req and add -reqopt switch.
3802 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3804 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3805 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3807 *) Make sure tests can be performed even if the corresponding algorithms
3808 have been removed entirely. This was also the last step to make
3809 OpenSSL compilable with DJGPP under all reasonable conditions.
3810 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3812 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3813 to allow version independent disabling of normally unselected ciphers,
3814 which may be activated as a side-effect of selecting a single cipher.
3816 (E.g., cipher list string "RSA" enables ciphersuites that are left
3817 out of "ALL" because they do not provide symmetric encryption.
3818 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3819 [Lutz Jaenicke, Bodo Moeller]
3821 *) Add appropriate support for separate platform-dependent build
3822 directories. The recommended way to make a platform-dependent
3823 build directory is the following (tested on Linux), maybe with
3826 # Place yourself outside of the OpenSSL source tree. In
3827 # this example, the environment variable OPENSSL_SOURCE
3828 # is assumed to contain the absolute OpenSSL source directory.
3829 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3830 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3831 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3832 mkdir -p `dirname $F`
3833 ln -s $OPENSSL_SOURCE/$F $F
3836 To be absolutely sure not to disturb the source tree, a "make clean"
3837 is a good thing. If it isn't successfull, don't worry about it,
3838 it probably means the source directory is very clean.
3841 *) Make sure any ENGINE control commands make local copies of string
3842 pointers passed to them whenever necessary. Otherwise it is possible
3843 the caller may have overwritten (or deallocated) the original string
3844 data when a later ENGINE operation tries to use the stored values.
3845 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3847 *) Improve diagnostics in file reading and command-line digests.
3848 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3850 *) Add AES modes CFB and OFB to the object database. Correct an
3851 error in AES-CFB decryption.
3854 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3855 allows existing EVP_CIPHER_CTX structures to be reused after
3856 calling EVP_*Final(). This behaviour is used by encryption
3857 BIOs and some applications. This has the side effect that
3858 applications must explicitly clean up cipher contexts with
3859 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3862 *) Check the values of dna and dnb in bn_mul_recursive before calling
3863 bn_mul_comba (a non zero value means the a or b arrays do not contain
3864 n2 elements) and fallback to bn_mul_normal if either is not zero.
3867 *) Fix escaping of non-ASCII characters when using the -subj option
3868 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3871 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3872 form for "surname", serialNumber has no short form.
3873 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3874 therefore remove "mail" short name for "internet 7".
3875 The OID for unique identifiers in X509 certificates is
3876 x500UniqueIdentifier, not uniqueIdentifier.
3877 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3880 *) Add an "init" command to the ENGINE config module and auto initialize
3881 ENGINEs. Without any "init" command the ENGINE will be initialized
3882 after all ctrl commands have been executed on it. If init=1 the
3883 ENGINE is initailized at that point (ctrls before that point are run
3884 on the uninitialized ENGINE and after on the initialized one). If
3885 init=0 then the ENGINE will not be iniatialized at all.
3888 *) Fix the 'app_verify_callback' interface so that the user-defined
3889 argument is actually passed to the callback: In the
3890 SSL_CTX_set_cert_verify_callback() prototype, the callback
3891 declaration has been changed from
3894 int (*cb)(X509_STORE_CTX *,void *);
3895 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3896 i=s->ctx->app_verify_callback(&ctx)
3897 has been changed into
3898 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3900 To update applications using SSL_CTX_set_cert_verify_callback(),
3901 a dummy argument can be added to their callback functions.
3902 [D. K. Smetters <smetters@parc.xerox.com>]
3904 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3905 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3907 *) Add and OPENSSL_LOAD_CONF define which will cause
3908 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3909 This allows older applications to transparently support certain
3910 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3911 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3912 load the config file and OPENSSL_add_all_algorithms_conf() which will
3913 always load it have also been added.
3916 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3917 Adjust NIDs and EVP layer.
3918 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3920 *) Config modules support in openssl utility.
3922 Most commands now load modules from the config file,
3923 though in a few (such as version) this isn't done
3924 because it couldn't be used for anything.
3926 In the case of ca and req the config file used is
3927 the same as the utility itself: that is the -config
3928 command line option can be used to specify an
3932 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3933 use "openssl_conf" if filename is NULL use default openssl config file.
3936 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3937 config section name. Add a new flag to tolerate a missing config file
3938 and move code to CONF_modules_load_file().
3941 *) Support for crypto accelerator cards from Accelerated Encryption
3942 Processing, www.aep.ie. (Use engine 'aep')
3943 The support was copied from 0.9.6c [engine] and adapted/corrected
3944 to work with the new engine framework.
3945 [AEP Inc. and Richard Levitte]
3947 *) Support for SureWare crypto accelerator cards from Baltimore
3948 Technologies. (Use engine 'sureware')
3949 The support was copied from 0.9.6c [engine] and adapted
3950 to work with the new engine framework.
3953 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3954 make the newer ENGINE framework commands for the CHIL engine work.
3955 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3957 *) Make it possible to produce shared libraries on ReliantUNIX.
3958 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3960 *) Add the configuration target debug-linux-ppro.
3961 Make 'openssl rsa' use the general key loading routines
3962 implemented in apps.c, and make those routines able to
3963 handle the key format FORMAT_NETSCAPE and the variant
3965 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3967 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3968 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3970 *) Add -keyform to rsautl, and document -engine.
3971 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3973 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3974 BIO_R_NO_SUCH_FILE error code rather than the generic
3975 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3978 *) Add new functions
3980 ERR_peek_last_error_line
3981 ERR_peek_last_error_line_data.
3982 These are similar to
3985 ERR_peek_error_line_data,
3986 but report on the latest error recorded rather than the first one
3987 still in the error queue.
3988 [Ben Laurie, Bodo Moeller]
3990 *) default_algorithms option in ENGINE config module. This allows things
3992 default_algorithms = ALL
3993 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3996 *) Prelminary ENGINE config module.
3999 *) New experimental application configuration code.
4002 *) Change the AES code to follow the same name structure as all other
4003 symmetric ciphers, and behave the same way. Move everything to
4004 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4005 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4007 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4008 [Ben Laurie and Theo de Raadt]
4010 *) Add option to output public keys in req command.
4011 [Massimiliano Pala madwolf@openca.org]
4013 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4014 (up to about 10% better than before for P-192 and P-224).
4017 *) New functions/macros
4019 SSL_CTX_set_msg_callback(ctx, cb)
4020 SSL_CTX_set_msg_callback_arg(ctx, arg)
4021 SSL_set_msg_callback(ssl, cb)
4022 SSL_set_msg_callback_arg(ssl, arg)
4024 to request calling a callback function
4026 void cb(int write_p, int version, int content_type,
4027 const void *buf, size_t len, SSL *ssl, void *arg)
4029 whenever a protocol message has been completely received
4030 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4031 protocol version according to which the SSL library interprets
4032 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4033 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4034 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4035 specification (change_cipher_spec(20), alert(21), handshake(22)).
4036 'buf' and 'len' point to the actual message, 'ssl' to the
4037 SSL object, and 'arg' is the application-defined value set by
4038 SSL[_CTX]_set_msg_callback_arg().
4040 'openssl s_client' and 'openssl s_server' have new '-msg' options
4041 to enable a callback that displays all protocol messages.
4044 *) Change the shared library support so shared libraries are built as
4045 soon as the corresponding static library is finished, and thereby get
4046 openssl and the test programs linked against the shared library.
4047 This still only happens when the keyword "shard" has been given to
4048 the configuration scripts.
4050 NOTE: shared library support is still an experimental thing, and
4051 backward binary compatibility is still not guaranteed.
4052 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4054 *) Add support for Subject Information Access extension.
4055 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4057 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4058 additional bytes when new memory had to be allocated, not just
4059 when reusing an existing buffer.
4062 *) New command line and configuration option 'utf8' for the req command.
4063 This allows field values to be specified as UTF8 strings.
4066 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4067 runs for the former and machine-readable output for the latter.
4070 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4071 of the e-mail address in the DN (i.e., it will go into a certificate
4072 extension only). The new configuration file option 'email_in_dn = no'
4073 has the same effect.
4074 [Massimiliano Pala madwolf@openca.org]
4076 *) Change all functions with names starting with des_ to be starting
4077 with DES_ instead. Add wrappers that are compatible with libdes,
4078 but are named _ossl_old_des_*. Finally, add macros that map the
4079 des_* symbols to the corresponding _ossl_old_des_* if libdes
4080 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4081 desired, the des_* symbols will be mapped to DES_*, with one
4084 Since we provide two compatibility mappings, the user needs to
4085 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4086 compatibility is desired. The default (i.e., when that macro
4087 isn't defined) is OpenSSL 0.9.6c compatibility.
4089 There are also macros that enable and disable the support of old
4090 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4091 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4092 are defined, the default will apply: to support the old des routines.
4094 In either case, one must include openssl/des.h to get the correct
4095 definitions. Do not try to just include openssl/des_old.h, that
4098 NOTE: This is a major break of an old API into a new one. Software
4099 authors are encouraged to switch to the DES_ style functions. Some
4100 time in the future, des_old.h and the libdes compatibility functions
4101 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4102 default), and then completely removed.
4105 *) Test for certificates which contain unsupported critical extensions.
4106 If such a certificate is found during a verify operation it is
4107 rejected by default: this behaviour can be overridden by either
4108 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4109 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4110 X509_supported_extension() has also been added which returns 1 if a
4111 particular extension is supported.
4114 *) Modify the behaviour of EVP cipher functions in similar way to digests
4115 to retain compatibility with existing code.
4118 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4119 compatibility with existing code. In particular the 'ctx' parameter does
4120 not have to be to be initialized before the call to EVP_DigestInit() and
4121 it is tidied up after a call to EVP_DigestFinal(). New function
4122 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4123 EVP_MD_CTX_copy() changed to not require the destination to be
4124 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4125 requires the destination to be valid.
4127 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4128 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4131 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4132 so that complete 'Handshake' protocol structures are kept in memory
4133 instead of overwriting 'msg_type' and 'length' with 'body' data.
4136 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4137 [Massimo Santin via Richard Levitte]
4139 *) Major restructuring to the underlying ENGINE code. This includes
4140 reduction of linker bloat, separation of pure "ENGINE" manipulation
4141 (initialisation, etc) from functionality dealing with implementations
4142 of specific crypto iterfaces. This change also introduces integrated
4143 support for symmetric ciphers and digest implementations - so ENGINEs
4144 can now accelerate these by providing EVP_CIPHER and EVP_MD
4145 implementations of their own. This is detailed in crypto/engine/README
4146 as it couldn't be adequately described here. However, there are a few
4147 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4148 were changed in the original introduction of ENGINE code have now
4149 reverted back - the hooking from this code to ENGINE is now a good
4150 deal more passive and at run-time, operations deal directly with
4151 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4152 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4153 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4154 they were not being used by the framework as there is no concept of a
4155 BIGNUM_METHOD and they could not be generalised to the new
4156 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4157 ENGINE_cpy() has been removed as it cannot be consistently defined in
4161 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4164 *) Change mkdef.pl to sort symbols that get the same entry number,
4165 and make sure the automatically generated functions ERR_load_*
4166 become part of libeay.num as well.
4169 *) New function SSL_renegotiate_pending(). This returns true once
4170 renegotiation has been requested (either SSL_renegotiate() call
4171 or HelloRequest/ClientHello receveived from the peer) and becomes
4172 false once a handshake has been completed.
4173 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4174 sends a HelloRequest, but does not ensure that a handshake takes
4175 place. SSL_renegotiate_pending() is useful for checking if the
4176 client has followed the request.)
4179 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4180 By default, clients may request session resumption even during
4181 renegotiation (if session ID contexts permit); with this option,
4182 session resumption is possible only in the first handshake.
4184 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4185 more bits available for options that should not be part of
4186 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4189 *) Add some demos for certificate and certificate request creation.
4192 *) Make maximum certificate chain size accepted from the peer application
4193 settable (SSL*_get/set_max_cert_list()), as proposed by
4194 "Douglas E. Engert" <deengert@anl.gov>.
4197 *) Add support for shared libraries for Unixware-7
4198 (Boyd Lynn Gerber <gerberb@zenez.com>).
4201 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4202 be done prior to destruction. Use this to unload error strings from
4203 ENGINEs that load their own error strings. NB: This adds two new API
4204 functions to "get" and "set" this destroy handler in an ENGINE.
4207 *) Alter all existing ENGINE implementations (except "openssl" and
4208 "openbsd") to dynamically instantiate their own error strings. This
4209 makes them more flexible to be built both as statically-linked ENGINEs
4210 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4211 Also, add stub code to each that makes building them as self-contained
4212 shared-libraries easier (see README.ENGINE).
4215 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4216 implementations into applications that are completely implemented in
4217 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4218 commands that can be used to configure what shared-library to load and
4219 to control aspects of the way it is handled. Also, made an update to
4220 the README.ENGINE file that brings its information up-to-date and
4221 provides some information and instructions on the "dynamic" ENGINE
4222 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4225 *) Make it possible to unload ranges of ERR strings with a new
4226 "ERR_unload_strings" function.
4229 *) Add a copy() function to EVP_MD.
4232 *) Make EVP_MD routines take a context pointer instead of just the
4233 md_data void pointer.
4236 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4237 that the digest can only process a single chunk of data
4238 (typically because it is provided by a piece of
4239 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4240 is only going to provide a single chunk of data, and hence the
4241 framework needn't accumulate the data for oneshot drivers.
4244 *) As with "ERR", make it possible to replace the underlying "ex_data"
4245 functions. This change also alters the storage and management of global
4246 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4247 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4248 index counters. The API functions that use this state have been changed
4249 to take a "class_index" rather than pointers to the class's local STACK
4250 and counter, and there is now an API function to dynamically create new
4251 classes. This centralisation allows us to (a) plug a lot of the
4252 thread-safety problems that existed, and (b) makes it possible to clean
4253 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4254 such data would previously have always leaked in application code and
4255 workarounds were in place to make the memory debugging turn a blind eye
4256 to it. Application code that doesn't use this new function will still
4257 leak as before, but their memory debugging output will announce it now
4258 rather than letting it slide.
4260 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4261 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4262 has a return value to indicate success or failure.
4265 *) Make it possible to replace the underlying "ERR" functions such that the
4266 global state (2 LHASH tables and 2 locks) is only used by the "default"
4267 implementation. This change also adds two functions to "get" and "set"
4268 the implementation prior to it being automatically set the first time
4269 any other ERR function takes place. Ie. an application can call "get",
4270 pass the return value to a module it has just loaded, and that module
4271 can call its own "set" function using that value. This means the
4272 module's "ERR" operations will use (and modify) the error state in the
4273 application and not in its own statically linked copy of OpenSSL code.
4276 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4277 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4278 the operation, and provides a more encapsulated way for external code
4279 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4280 to use these functions rather than manually incrementing the counts.
4282 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4285 *) Add EVP test program.
4288 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4291 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4292 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4293 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4294 These allow a CRL to be built without having to access X509_CRL fields
4295 directly. Modify 'ca' application to use new functions.
4298 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4299 bug workarounds. Rollback attack detection is a security feature.
4300 The problem will only arise on OpenSSL servers when TLSv1 is not
4301 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4302 Software authors not wanting to support TLSv1 will have special reasons
4303 for their choice and can explicitly enable this option.
4304 [Bodo Moeller, Lutz Jaenicke]
4306 *) Rationalise EVP so it can be extended: don't include a union of
4307 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4308 (similar to those existing for EVP_CIPHER_CTX).
4313 EVP_MD_CTX_init(&md); /* new function call */
4314 EVP_DigestInit(&md, EVP_sha1());
4315 EVP_DigestUpdate(&md, in, len);
4316 EVP_DigestFinal(&md, out, NULL);
4317 EVP_MD_CTX_cleanup(&md); /* new function call */
4321 *) Make DES key schedule conform to the usual scheme, as well as
4322 correcting its structure. This means that calls to DES functions
4323 now have to pass a pointer to a des_key_schedule instead of a
4324 plain des_key_schedule (which was actually always a pointer
4327 des_key_schedule ks;
4329 des_set_key_checked(..., &ks);
4330 des_ncbc_encrypt(..., &ks, ...);
4332 (Note that a later change renames 'des_...' into 'DES_...'.)
4335 *) Initial reduction of linker bloat: the use of some functions, such as
4336 PEM causes large amounts of unused functions to be linked in due to
4337 poor organisation. For example pem_all.c contains every PEM function
4338 which has a knock on effect of linking in large amounts of (unused)
4339 ASN1 code. Grouping together similar functions and splitting unrelated
4340 functions prevents this.
4343 *) Cleanup of EVP macros.
4346 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4347 correct _ecb suffix.
4350 *) Add initial OCSP responder support to ocsp application. The
4351 revocation information is handled using the text based index
4352 use by the ca application. The responder can either handle
4353 requests generated internally, supplied in files (for example
4354 via a CGI script) or using an internal minimal server.
4357 *) Add configuration choices to get zlib compression for TLS.
4360 *) Changes to Kerberos SSL for RFC 2712 compliance:
4361 1. Implemented real KerberosWrapper, instead of just using
4362 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4363 2. Implemented optional authenticator field of KerberosWrapper.
4365 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4366 and authenticator structs; see crypto/krb5/.
4368 Generalized Kerberos calls to support multiple Kerberos libraries.
4369 [Vern Staats <staatsvr@asc.hpc.mil>,
4370 Jeffrey Altman <jaltman@columbia.edu>
4371 via Richard Levitte]
4373 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4374 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4375 values for each of the key sizes rather than having just
4376 parameters (and 'speed' generating keys each time).
4379 *) Speed up EVP routines.
4382 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4383 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4384 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4385 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4387 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4388 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4389 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4392 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4394 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4397 *) Added the OS2-EMX target.
4398 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4400 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4401 to support NCONF routines in extension code. New function CONF_set_nconf()
4402 to allow functions which take an NCONF to also handle the old LHASH
4403 structure: this means that the old CONF compatible routines can be
4404 retained (in particular wrt extensions) without having to duplicate the
4405 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4408 *) Enhance the general user interface with mechanisms for inner control
4409 and with possibilities to have yes/no kind of prompts.
4412 *) Change all calls to low level digest routines in the library and
4413 applications to use EVP. Add missing calls to HMAC_cleanup() and
4414 don't assume HMAC_CTX can be copied using memcpy().
4415 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4417 *) Add the possibility to control engines through control names but with
4418 arbitrary arguments instead of just a string.
4419 Change the key loaders to take a UI_METHOD instead of a callback
4420 function pointer. NOTE: this breaks binary compatibility with earlier
4421 versions of OpenSSL [engine].
4422 Adapt the nCipher code for these new conditions and add a card insertion
4426 *) Enhance the general user interface with mechanisms to better support
4427 dialog box interfaces, application-defined prompts, the possibility
4428 to use defaults (for example default passwords from somewhere else)
4429 and interrupts/cancellations.
4432 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4433 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4436 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4437 tidy up some unnecessarily weird code in 'sk_new()').
4438 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4440 *) Change the key loading routines for ENGINEs to use the same kind
4441 callback (pem_password_cb) as all other routines that need this
4445 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4446 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4447 than this minimum value is recommended.
4450 *) New random seeder for OpenVMS, using the system process statistics
4451 that are easily reachable.
4454 *) Windows apparently can't transparently handle global
4455 variables defined in DLLs. Initialisations such as:
4457 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4459 wont compile. This is used by the any applications that need to
4460 declare their own ASN1 modules. This was fixed by adding the option
4461 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4462 needed for static libraries under Win32.
4465 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4466 setting of purpose and trust fields. New X509_STORE trust and
4467 purpose functions and tidy up setting in other SSL functions.
4470 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4471 structure. These are inherited by X509_STORE_CTX when it is
4472 initialised. This allows various defaults to be set in the
4473 X509_STORE structure (such as flags for CRL checking and custom
4474 purpose or trust settings) for functions which only use X509_STORE_CTX
4475 internally such as S/MIME.
4477 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4478 trust settings if they are not set in X509_STORE. This allows X509_STORE
4479 purposes and trust (in S/MIME for example) to override any set by default.
4481 Add command line options for CRL checking to smime, s_client and s_server
4485 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4486 are set then the CRL is looked up in the X509_STORE structure and
4487 its validity and signature checked, then if the certificate is found
4488 in the CRL the verify fails with a revoked error.
4490 Various new CRL related callbacks added to X509_STORE_CTX structure.
4492 Command line options added to 'verify' application to support this.
4494 This needs some additional work, such as being able to handle multiple
4495 CRLs with different times, extension based lookup (rather than just
4496 by subject name) and ultimately more complete V2 CRL extension
4500 *) Add a general user interface API (crypto/ui/). This is designed
4501 to replace things like des_read_password and friends (backward
4502 compatibility functions using this new API are provided).
4503 The purpose is to remove prompting functions from the DES code
4504 section as well as provide for prompting through dialog boxes in
4505 a window system and the like.
4508 *) Add "ex_data" support to ENGINE so implementations can add state at a
4509 per-structure level rather than having to store it globally.
4512 *) Make it possible for ENGINE structures to be copied when retrieved by
4513 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4514 This causes the "original" ENGINE structure to act like a template,
4515 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4516 operational state can be localised to each ENGINE structure, despite the
4517 fact they all share the same "methods". New ENGINE structures returned in
4518 this case have no functional references and the return value is the single
4519 structural reference. This matches the single structural reference returned
4520 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4524 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4525 needs to match any other type at all we need to manually clear the
4529 *) Changes to the "openssl engine" utility to include;
4530 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4531 about an ENGINE's available control commands.
4532 - executing control commands from command line arguments using the
4533 '-pre' and '-post' switches. '-post' is only used if '-t' is
4534 specified and the ENGINE is successfully initialised. The syntax for
4535 the individual commands are colon-separated, for example;
4536 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4539 *) New dynamic control command support for ENGINEs. ENGINEs can now
4540 declare their own commands (numbers), names (strings), descriptions,
4541 and input types for run-time discovery by calling applications. A
4542 subset of these commands are implicitly classed as "executable"
4543 depending on their input type, and only these can be invoked through
4544 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4545 can be based on user input, config files, etc). The distinction is
4546 that "executable" commands cannot return anything other than a boolean
4547 result and can only support numeric or string input, whereas some
4548 discoverable commands may only be for direct use through
4549 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4550 pointers, or other custom uses. The "executable" commands are to
4551 support parameterisations of ENGINE behaviour that can be
4552 unambiguously defined by ENGINEs and used consistently across any
4553 OpenSSL-based application. Commands have been added to all the
4554 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4555 control over shared-library paths without source code alterations.
4558 *) Changed all ENGINE implementations to dynamically allocate their
4559 ENGINEs rather than declaring them statically. Apart from this being
4560 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4561 this also allows the implementations to compile without using the
4562 internal engine_int.h header.
4565 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4566 'const' value. Any code that should be able to modify a RAND_METHOD
4567 should already have non-const pointers to it (ie. they should only
4568 modify their own ones).
4571 *) Made a variety of little tweaks to the ENGINE code.
4572 - "atalla" and "ubsec" string definitions were moved from header files
4573 to C code. "nuron" string definitions were placed in variables
4574 rather than hard-coded - allowing parameterisation of these values
4575 later on via ctrl() commands.
4576 - Removed unused "#if 0"'d code.
4577 - Fixed engine list iteration code so it uses ENGINE_free() to release
4578 structural references.
4579 - Constified the RAND_METHOD element of ENGINE structures.
4580 - Constified various get/set functions as appropriate and added
4581 missing functions (including a catch-all ENGINE_cpy that duplicates
4582 all ENGINE values onto a new ENGINE except reference counts/state).
4583 - Removed NULL parameter checks in get/set functions. Setting a method
4584 or function to NULL is a way of cancelling out a previously set
4585 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4586 and doesn't justify the extra error symbols and code.
4587 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4588 flags from engine_int.h to engine.h.
4589 - Changed prototypes for ENGINE handler functions (init(), finish(),
4590 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4593 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4594 to the algorithm using long division. The binary algorithm can be
4595 used only if the modulus is odd. On 32-bit systems, it is faster
4596 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4597 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4598 up to 450 bits. In 64-bit environments, the binary algorithm
4599 appears to be advantageous for much longer moduli; here we use it
4600 for moduli up to 2048 bits.
4603 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4604 could not support the combine flag in choice fields.
4607 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4608 extensions from a certificate request to the certificate.
4611 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4612 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4613 file: this allows the display of the certificate about to be
4614 signed to be customised, to allow certain fields to be included
4615 or excluded and extension details. The old system didn't display
4616 multicharacter strings properly, omitted fields not in the policy
4617 and couldn't display additional details such as extensions.
4620 *) Function EC_POINTs_mul for multiple scalar multiplication
4621 of an arbitrary number of elliptic curve points
4622 \sum scalars[i]*points[i],
4623 optionally including the generator defined for the EC_GROUP:
4624 scalar*generator + \sum scalars[i]*points[i].
4626 EC_POINT_mul is a simple wrapper function for the typical case
4627 that the point list has just one item (besides the optional
4631 *) First EC_METHODs for curves over GF(p):
4633 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4634 operations and provides various method functions that can also
4635 operate with faster implementations of modular arithmetic.
4637 EC_GFp_mont_method() reuses most functions that are part of
4638 EC_GFp_simple_method, but uses Montgomery arithmetic.
4640 [Bodo Moeller; point addition and point doubling
4641 implementation directly derived from source code provided by
4642 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4644 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4645 crypto/ec/ec_lib.c):
4647 Curves are EC_GROUP objects (with an optional group generator)
4648 based on EC_METHODs that are built into the library.
4650 Points are EC_POINT objects based on EC_GROUP objects.
4652 Most of the framework would be able to handle curves over arbitrary
4653 finite fields, but as there are no obvious types for fields other
4654 than GF(p), some functions are limited to that for now.
4657 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4658 that the file contains a complete HTTP response.
4661 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4662 change the def and num file printf format specifier from "%-40sXXX"
4663 to "%-39s XXX". The latter will always guarantee a space after the
4664 field while the former will cause them to run together if the field
4665 is 40 of more characters long.
4668 *) Constify the cipher and digest 'method' functions and structures
4669 and modify related functions to take constant EVP_MD and EVP_CIPHER
4673 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4674 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4677 *) Modify EVP_Digest*() routines so they now return values. Although the
4678 internal software routines can never fail additional hardware versions
4682 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4684 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4685 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4688 ERR_R_NESTED_ASN1_ERROR
4690 ERR_R_MISSING_ASN1_EOS
4691 were 4 .. 9, conflicting with
4692 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4694 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4695 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4697 Add new error code 'ERR_R_INTERNAL_ERROR'.
4700 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4704 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4705 sets the subject name for a new request or supersedes the
4706 subject name in a given request. Formats that can be parsed are
4707 'CN=Some Name, OU=myOU, C=IT'
4709 'CN=Some Name/OU=myOU/C=IT'.
4711 Add options '-batch' and '-verbose' to 'openssl req'.
4712 [Massimiliano Pala <madwolf@hackmasters.net>]
4714 *) Introduce the possibility to access global variables through
4715 functions on platform were that's the best way to handle exporting
4716 global variables in shared libraries. To enable this functionality,
4717 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4718 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4719 is normally done by Configure or something similar).
4721 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4722 in the source file (foo.c) like this:
4724 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4725 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4727 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4728 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4730 OPENSSL_DECLARE_GLOBAL(int,foo);
4731 #define foo OPENSSL_GLOBAL_REF(foo)
4732 OPENSSL_DECLARE_GLOBAL(double,bar);
4733 #define bar OPENSSL_GLOBAL_REF(bar)
4735 The #defines are very important, and therefore so is including the
4736 header file everywhere where the defined globals are used.
4738 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4739 of ASN.1 items, but that structure is a bit different.
4741 The largest change is in util/mkdef.pl which has been enhanced with
4742 better and easier to understand logic to choose which symbols should
4743 go into the Windows .def files as well as a number of fixes and code
4744 cleanup (among others, algorithm keywords are now sorted
4745 lexicographically to avoid constant rewrites).
4748 *) In BN_div() keep a copy of the sign of 'num' before writing the
4749 result to 'rm' because if rm==num the value will be overwritten
4750 and produce the wrong result if 'num' is negative: this caused
4751 problems with BN_mod() and BN_nnmod().
4754 *) Function OCSP_request_verify(). This checks the signature on an
4755 OCSP request and verifies the signer certificate. The signer
4756 certificate is just checked for a generic purpose and OCSP request
4760 *) Add OCSP_check_validity() function to check the validity of OCSP
4761 responses. OCSP responses are prepared in real time and may only
4762 be a few seconds old. Simply checking that the current time lies
4763 between thisUpdate and nextUpdate max reject otherwise valid responses
4764 caused by either OCSP responder or client clock inaccuracy. Instead
4765 we allow thisUpdate and nextUpdate to fall within a certain period of
4766 the current time. The age of the response can also optionally be
4767 checked. Two new options -validity_period and -status_age added to
4771 *) If signature or public key algorithm is unrecognized print out its
4772 OID rather that just UNKNOWN.
4775 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4776 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4777 ID to be generated from the issuer certificate alone which can then be
4778 passed to OCSP_id_issuer_cmp().
4781 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4782 ASN1 modules to export functions returning ASN1_ITEM pointers
4783 instead of the ASN1_ITEM structures themselves. This adds several
4784 new macros which allow the underlying ASN1 function/structure to
4785 be accessed transparently. As a result code should not use ASN1_ITEM
4786 references directly (such as &X509_it) but instead use the relevant
4787 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4788 use of the new ASN1 code on platforms where exporting structures
4789 is problematical (for example in shared libraries) but exporting
4790 functions returning pointers to structures is not.
4793 *) Add support for overriding the generation of SSL/TLS session IDs.
4794 These callbacks can be registered either in an SSL_CTX or per SSL.
4795 The purpose of this is to allow applications to control, if they wish,
4796 the arbitrary values chosen for use as session IDs, particularly as it
4797 can be useful for session caching in multiple-server environments. A
4798 command-line switch for testing this (and any client code that wishes
4799 to use such a feature) has been added to "s_server".
4800 [Geoff Thorpe, Lutz Jaenicke]
4802 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4803 of the form '#if defined(...) || defined(...) || ...' and
4804 '#if !defined(...) && !defined(...) && ...'. This also avoids
4805 the growing number of special cases it was previously handling.
4808 *) Make all configuration macros available for application by making
4809 sure they are available in opensslconf.h, by giving them names starting
4810 with "OPENSSL_" to avoid conflicts with other packages and by making
4811 sure e_os2.h will cover all platform-specific cases together with
4813 Additionally, it is now possible to define configuration/platform-
4814 specific names (called "system identities"). In the C code, these
4815 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4816 macro with the name beginning with "OPENSSL_SYS_", which is determined
4817 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4821 *) New option -set_serial to 'req' and 'x509' this allows the serial
4822 number to use to be specified on the command line. Previously self
4823 signed certificates were hard coded with serial number 0 and the
4824 CA options of 'x509' had to use a serial number in a file which was
4828 *) New options to 'ca' utility to support V2 CRL entry extensions.
4829 Currently CRL reason, invalidity date and hold instruction are
4830 supported. Add new CRL extensions to V3 code and some new objects.
4833 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4834 disable standard block padding (aka PKCS#5 padding) in the EVP
4835 API, which was previously mandatory. This means that the data is
4836 not padded in any way and so the total length much be a multiple
4837 of the block size, otherwise an error occurs.
4840 *) Initial (incomplete) OCSP SSL support.
4843 *) New function OCSP_parse_url(). This splits up a URL into its host,
4844 port and path components: primarily to parse OCSP URLs. New -url
4845 option to ocsp utility.
4848 *) New nonce behavior. The return value of OCSP_check_nonce() now
4849 reflects the various checks performed. Applications can decide
4850 whether to tolerate certain situations such as an absent nonce
4851 in a response when one was present in a request: the ocsp application
4852 just prints out a warning. New function OCSP_add1_basic_nonce()
4853 this is to allow responders to include a nonce in a response even if
4854 the request is nonce-less.
4857 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4858 skipped when using openssl x509 multiple times on a single input file,
4859 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4862 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4863 set string type: to handle setting ASN1_TIME structures. Fix ca
4864 utility to correctly initialize revocation date of CRLs.
4867 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4868 the clients preferred ciphersuites and rather use its own preferences.
4869 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4870 Internet Explorer by ensuring unchanged hash method during stepup.
4871 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4874 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4875 to aes and add a new 'exist' option to print out symbols that don't
4879 *) Additional options to ocsp utility to allow flags to be set and
4880 additional certificates supplied.
4883 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4884 OCSP client a number of certificate to only verify the response
4888 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4889 handle the new API. Currently only ECB, CBC modes supported. Add new
4892 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4893 Encryption Standard (AES) Ciphersuites for Transport Layer
4894 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4895 not enabled by default and were not part of the "ALL" ciphersuite
4896 alias because they were not yet official; they could be
4897 explicitly requested by specifying the "AESdraft" ciphersuite
4898 group alias. In the final release of OpenSSL 0.9.7, the group
4899 alias is called "AES" and is part of "ALL".)
4900 [Ben Laurie, Steve Henson, Bodo Moeller]
4902 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4903 request to response.
4906 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4907 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4908 extract information from a certificate request. OCSP_response_create()
4909 creates a response and optionally adds a basic response structure.
4910 OCSP_basic_add1_status() adds a complete single response to a basic
4911 response and returns the OCSP_SINGLERESP structure just added (to allow
4912 extensions to be included for example). OCSP_basic_add1_cert() adds a
4913 certificate to a basic response and OCSP_basic_sign() signs a basic
4914 response with various flags. New helper functions ASN1_TIME_check()
4915 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4916 (converts ASN1_TIME to GeneralizedTime).
4919 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4920 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4921 structure from a certificate. X509_pubkey_digest() digests the public_key
4922 contents: this is used in various key identifiers.
4925 *) Make sk_sort() tolerate a NULL argument.
4926 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4928 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4929 passed by the function are trusted implicitly. If any of them signed the
4930 response then it is assumed to be valid and is not verified.
4933 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4934 to data. This was previously part of the PKCS7 ASN1 code. This
4935 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4936 [Steve Henson, reported by Kenneth R. Robinette
4937 <support@securenetterm.com>]
4939 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4940 routines: without these tracing memory leaks is very painful.
4941 Fix leaks in PKCS12 and PKCS7 routines.
4944 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4945 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4946 effectively meant GeneralizedTime would never be used. Now it
4947 is initialised to -1 but X509_time_adj() now has to check the value
4948 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4949 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4950 [Steve Henson, reported by Kenneth R. Robinette
4951 <support@securenetterm.com>]
4953 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4954 result in a zero length in the ASN1_INTEGER structure which was
4955 not consistent with the structure when d2i_ASN1_INTEGER() was used
4956 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4957 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4958 where it did not print out a minus for negative ASN1_INTEGER.
4961 *) Add summary printout to ocsp utility. The various functions which
4962 convert status values to strings have been renamed to:
4963 OCSP_response_status_str(), OCSP_cert_status_str() and
4964 OCSP_crl_reason_str() and are no longer static. New options
4965 to verify nonce values and to disable verification. OCSP response
4966 printout format cleaned up.
4969 *) Add additional OCSP certificate checks. These are those specified
4970 in RFC2560. This consists of two separate checks: the CA of the
4971 certificate being checked must either be the OCSP signer certificate
4972 or the issuer of the OCSP signer certificate. In the latter case the
4973 OCSP signer certificate must contain the OCSP signing extended key
4974 usage. This check is performed by attempting to match the OCSP
4975 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4976 in the OCSP_CERTID structures of the response.
4979 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4980 and related routines. This uses the standard OpenSSL certificate
4981 verify routines to perform initial checks (just CA validity) and
4982 to obtain the certificate chain. Then additional checks will be
4983 performed on the chain. Currently the root CA is checked to see
4984 if it is explicitly trusted for OCSP signing. This is used to set
4985 a root CA as a global signing root: that is any certificate that
4986 chains to that CA is an acceptable OCSP signing certificate.
4989 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4990 extensions from a separate configuration file.
4991 As when reading extensions from the main configuration file,
4992 the '-extensions ...' option may be used for specifying the
4994 [Massimiliano Pala <madwolf@comune.modena.it>]
4996 *) New OCSP utility. Allows OCSP requests to be generated or
4997 read. The request can be sent to a responder and the output
4998 parsed, outputed or printed in text form. Not complete yet:
4999 still needs to check the OCSP response validity.
5002 *) New subcommands for 'openssl ca':
5003 'openssl ca -status <serial>' prints the status of the cert with
5004 the given serial number (according to the index file).
5005 'openssl ca -updatedb' updates the expiry status of certificates
5007 [Massimiliano Pala <madwolf@comune.modena.it>]
5009 *) New '-newreq-nodes' command option to CA.pl. This is like
5010 '-newreq', but calls 'openssl req' with the '-nodes' option
5011 so that the resulting key is not encrypted.
5012 [Damien Miller <djm@mindrot.org>]
5014 *) New configuration for the GNU Hurd.
5015 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5017 *) Initial code to implement OCSP basic response verify. This
5018 is currently incomplete. Currently just finds the signer's
5019 certificate and verifies the signature on the response.
5022 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5023 value of OPENSSLDIR. This is available via the new '-d' option
5024 to 'openssl version', and is also included in 'openssl version -a'.
5027 *) Allowing defining memory allocation callbacks that will be given
5028 file name and line number information in additional arguments
5029 (a const char* and an int). The basic functionality remains, as
5030 well as the original possibility to just replace malloc(),
5031 realloc() and free() by functions that do not know about these
5032 additional arguments. To register and find out the current
5033 settings for extended allocation functions, the following
5034 functions are provided:
5036 CRYPTO_set_mem_ex_functions
5037 CRYPTO_set_locked_mem_ex_functions
5038 CRYPTO_get_mem_ex_functions
5039 CRYPTO_get_locked_mem_ex_functions
5041 These work the same way as CRYPTO_set_mem_functions and friends.
5042 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5043 extended allocation function is enabled.
5044 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5045 a conventional allocation function is enabled.
5046 [Richard Levitte, Bodo Moeller]
5048 *) Finish off removing the remaining LHASH function pointer casts.
5049 There should no longer be any prototype-casting required when using
5050 the LHASH abstraction, and any casts that remain are "bugs". See
5051 the callback types and macros at the head of lhash.h for details
5052 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5055 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5056 If /dev/[u]random devices are not available or do not return enough
5057 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5059 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5060 /etc/entropy will be queried once each in this sequence, quering stops
5061 when enough entropy was collected without querying more sockets.
5064 *) Change the Unix RAND_poll() variant to be able to poll several
5065 random devices, as specified by DEVRANDOM, until a sufficient amount
5066 of data has been collected. We spend at most 10 ms on each file
5067 (select timeout) and read in non-blocking mode. DEVRANDOM now
5068 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5069 (previously it was just the string "/dev/urandom"), so on typical
5070 platforms the 10 ms delay will never occur.
5071 Also separate out the Unix variant to its own file, rand_unix.c.
5072 For VMS, there's a currently-empty rand_vms.c.
5075 *) Move OCSP client related routines to ocsp_cl.c. These
5076 provide utility functions which an application needing
5077 to issue a request to an OCSP responder and analyse the
5078 response will typically need: as opposed to those which an
5079 OCSP responder itself would need which will be added later.
5081 OCSP_request_sign() signs an OCSP request with an API similar
5082 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5083 response. OCSP_response_get1_basic() extracts basic response
5084 from response. OCSP_resp_find_status(): finds and extracts status
5085 information from an OCSP_CERTID structure (which will be created
5086 when the request structure is built). These are built from lower
5087 level functions which work on OCSP_SINGLERESP structures but
5088 wont normally be used unless the application wishes to examine
5089 extensions in the OCSP response for example.
5091 Replace nonce routines with a pair of functions.
5092 OCSP_request_add1_nonce() adds a nonce value and optionally
5093 generates a random value. OCSP_check_nonce() checks the
5094 validity of the nonce in an OCSP response.
5097 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5098 This doesn't copy the supplied OCSP_CERTID and avoids the
5099 need to free up the newly created id. Change return type
5100 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5101 This can then be used to add extensions to the request.
5102 Deleted OCSP_request_new(), since most of its functionality
5103 is now in OCSP_REQUEST_new() (and the case insensitive name
5104 clash) apart from the ability to set the request name which
5105 will be added elsewhere.
5108 *) Update OCSP API. Remove obsolete extensions argument from
5109 various functions. Extensions are now handled using the new
5110 OCSP extension code. New simple OCSP HTTP function which
5111 can be used to send requests and parse the response.
5114 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5115 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5116 uses the special reorder version of SET OF to sort the attributes
5117 and reorder them to match the encoded order. This resolves a long
5118 standing problem: a verify on a PKCS7 structure just after signing
5119 it used to fail because the attribute order did not match the
5120 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5121 it uses the received order. This is necessary to tolerate some broken
5122 software that does not order SET OF. This is handled by encoding
5123 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5124 to produce the required SET OF.
5127 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5128 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5129 files to get correct declarations of the ASN.1 item variables.
5132 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5133 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5134 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5135 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5136 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5137 ASN1_ITEM and no wrapper functions.
5140 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5141 replace the old function pointer based I/O routines. Change most of
5142 the *_d2i_bio() and *_d2i_fp() functions to use these.
5145 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5146 lines, recognice more "algorithms" that can be deselected, and make
5147 it complain about algorithm deselection that isn't recognised.
5150 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5151 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5152 to use new functions. Add NO_ASN1_OLD which can be set to remove
5153 some old style ASN1 functions: this can be used to determine if old
5154 code will still work when these eventually go away.
5157 *) New extension functions for OCSP structures, these follow the
5158 same conventions as certificates and CRLs.
5161 *) New function X509V3_add1_i2d(). This automatically encodes and
5162 adds an extension. Its behaviour can be customised with various
5163 flags to append, replace or delete. Various wrappers added for
5164 certifcates and CRLs.
5167 *) Fix to avoid calling the underlying ASN1 print routine when
5168 an extension cannot be parsed. Correct a typo in the
5169 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5172 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5173 entries for variables.
5176 *) Add functionality to apps/openssl.c for detecting locking
5177 problems: As the program is single-threaded, all we have
5178 to do is register a locking callback using an array for
5179 storing which locks are currently held by the program.
5182 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5183 SSL_get_ex_data_X509_STORE_idx(), which is used in
5184 ssl_verify_cert_chain() and thus can be called at any time
5185 during TLS/SSL handshakes so that thread-safety is essential.
5186 Unfortunately, the ex_data design is not at all suited
5187 for multi-threaded use, so it probably should be abolished.
5190 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5191 [Broadcom, tweaked and integrated by Geoff Thorpe]
5193 *) Move common extension printing code to new function
5194 X509V3_print_extensions(). Reorganise OCSP print routines and
5195 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5198 *) New function X509_signature_print() to remove duplication in some
5202 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5203 set (this was treated exactly the same as SET OF previously). This
5204 is used to reorder the STACK representing the structure to match the
5205 encoding. This will be used to get round a problem where a PKCS7
5206 structure which was signed could not be verified because the STACK
5207 order did not reflect the encoded order.
5210 *) Reimplement the OCSP ASN1 module using the new code.
5213 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5214 for its ASN1 operations. The old style function pointers still exist
5215 for now but they will eventually go away.
5218 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5219 completely replaces the old ASN1 functionality with a table driven
5220 encoder and decoder which interprets an ASN1_ITEM structure describing
5221 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5222 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5223 has also been converted to the new form.
5226 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5227 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5228 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5229 for negative moduli.
5232 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5233 of not touching the result's sign bit.
5236 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5240 *) Changed the LHASH code to use prototypes for callbacks, and created
5241 macros to declare and implement thin (optionally static) functions
5242 that provide type-safety and avoid function pointer casting for the
5243 type-specific callbacks.
5246 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5248 [Veers Staats <staatsvr@asc.hpc.mil>,
5249 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5251 *) Reformat the FAQ so the different questions and answers can be divided
5252 in sections depending on the subject.
5255 *) Have the zlib compression code load ZLIB.DLL dynamically under
5259 *) New function BN_mod_sqrt for computing square roots modulo a prime
5260 (using the probabilistic Tonelli-Shanks algorithm unless
5261 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5262 be handled deterministically).
5263 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5265 *) Make BN_mod_inverse faster by explicitly handling small quotients
5266 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5267 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5270 *) New function BN_kronecker.
5273 *) Fix BN_gcd so that it works on negative inputs; the result is
5274 positive unless both parameters are zero.
5275 Previously something reasonably close to an infinite loop was
5276 possible because numbers could be growing instead of shrinking
5277 in the implementation of Euclid's algorithm.
5280 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5281 sign of the number in question.
5283 Fix BN_is_word(a,w) to work correctly for w == 0.
5285 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5286 because its test if the absolute value of 'a' equals 'w'.
5287 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5288 it exists mostly for use in the implementations of BN_is_zero(),
5289 BN_is_one(), and BN_is_word().
5292 *) New function BN_swap.
5295 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5296 the exponentiation functions are more likely to produce reasonable
5297 results on negative inputs.
5300 *) Change BN_mod_mul so that the result is always non-negative.
5301 Previously, it could be negative if one of the factors was negative;
5302 I don't think anyone really wanted that behaviour.
5305 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5306 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5307 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5308 and add new functions:
5317 BN_mod_lshift1_quick
5321 These functions always generate non-negative results.
5323 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5324 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5326 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5327 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5328 be reduced modulo m.
5329 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5332 The following entry accidentily appeared in the CHANGES file
5333 distributed with OpenSSL 0.9.7. The modifications described in
5334 it do *not* apply to OpenSSL 0.9.7.
5336 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5337 was actually never needed) and in BN_mul(). The removal in BN_mul()
5338 required a small change in bn_mul_part_recursive() and the addition
5339 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5340 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5341 bn_sub_words() and bn_add_words() except they take arrays with
5346 *) In 'openssl passwd', verify passwords read from the terminal
5347 unless the '-salt' option is used (which usually means that
5348 verification would just waste user's time since the resulting
5349 hash is going to be compared with some given password hash)
5350 or the new '-noverify' option is used.
5352 This is an incompatible change, but it does not affect
5353 non-interactive use of 'openssl passwd' (passwords on the command
5354 line, '-stdin' option, '-in ...' option) and thus should not
5358 *) Remove all references to RSAref, since there's no more need for it.
5361 *) Make DSO load along a path given through an environment variable
5362 (SHLIB_PATH) with shl_load().
5365 *) Constify the ENGINE code as a result of BIGNUM constification.
5366 Also constify the RSA code and most things related to it. In a
5367 few places, most notable in the depth of the ASN.1 code, ugly
5368 casts back to non-const were required (to be solved at a later
5372 *) Make it so the openssl application has all engines loaded by default.
5375 *) Constify the BIGNUM routines a little more.
5378 *) Add the following functions:
5380 ENGINE_load_cswift()
5382 ENGINE_load_atalla()
5384 ENGINE_load_builtin_engines()
5386 That way, an application can itself choose if external engines that
5387 are built-in in OpenSSL shall ever be used or not. The benefit is
5388 that applications won't have to be linked with libdl or other dso
5389 libraries unless it's really needed.
5391 Changed 'openssl engine' to load all engines on demand.
5392 Changed the engine header files to avoid the duplication of some
5393 declarations (they differed!).
5396 *) 'openssl engine' can now list capabilities.
5399 *) Better error reporting in 'openssl engine'.
5402 *) Never call load_dh_param(NULL) in s_server.
5405 *) Add engine application. It can currently list engines by name and
5406 identity, and test if they are actually available.
5409 *) Improve RPM specification file by forcing symbolic linking and making
5410 sure the installed documentation is also owned by root.root.
5411 [Damien Miller <djm@mindrot.org>]
5413 *) Give the OpenSSL applications more possibilities to make use of
5414 keys (public as well as private) handled by engines.
5417 *) Add OCSP code that comes from CertCo.
5420 *) Add VMS support for the Rijndael code.
5423 *) Added untested support for Nuron crypto accelerator.
5426 *) Add support for external cryptographic devices. This code was
5427 previously distributed separately as the "engine" branch.
5428 [Geoff Thorpe, Richard Levitte]
5430 *) Rework the filename-translation in the DSO code. It is now possible to
5431 have far greater control over how a "name" is turned into a filename
5432 depending on the operating environment and any oddities about the
5433 different shared library filenames on each system.
5436 *) Support threads on FreeBSD-elf in Configure.
5439 *) Fix for SHA1 assembly problem with MASM: it produces
5440 warnings about corrupt line number information when assembling
5441 with debugging information. This is caused by the overlapping
5443 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5446 NCONF_get_number() has no error checking at all. As a replacement,
5447 NCONF_get_number_e() is defined (_e for "error checking") and is
5448 promoted strongly. The old NCONF_get_number is kept around for
5449 binary backward compatibility.
5450 Make it possible for methods to load from something other than a BIO,
5451 by providing a function pointer that is given a name instead of a BIO.
5452 For example, this could be used to load configuration data from an
5456 *) Fix for non blocking accept BIOs. Added new I/O special reason
5457 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5458 with non blocking I/O was not possible because no retry code was
5459 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5463 *) Added the beginnings of Rijndael support.
5466 *) Fix for bug in DirectoryString mask setting. Add support for
5467 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5468 to allow certificate printing to more controllable, additional
5469 'certopt' option to 'x509' to allow new printing options to be
5473 *) Clean old EAY MD5 hack from e_os.h.
5476 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5478 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5479 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5480 [Joe Orton, Steve Henson]
5482 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5484 *) Fix additional bug revealed by the NISCC test suite:
5486 Stop bug triggering large recursion when presented with
5487 certain ASN.1 tags (CVE-2003-0851)
5490 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5492 *) Fix various bugs revealed by running the NISCC test suite:
5494 Stop out of bounds reads in the ASN1 code when presented with
5495 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5497 If verify callback ignores invalid public key errors don't try to check
5498 certificate signature with the NULL public key.
5502 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5503 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5507 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5508 extra data after the compression methods not only for TLS 1.0
5509 but also for SSL 3.0 (as required by the specification).
5510 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5512 *) Change X509_certificate_type() to mark the key as exported/exportable
5513 when it's 512 *bits* long, not 512 bytes.
5516 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5518 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5519 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5520 a protocol version number mismatch like a decryption error
5521 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5524 *) Turn on RSA blinding by default in the default implementation
5525 to avoid a timing attack. Applications that don't want it can call
5526 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5527 They would be ill-advised to do so in most cases.
5528 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5530 *) Change RSA blinding code so that it works when the PRNG is not
5531 seeded (in this case, the secret RSA exponent is abused as
5532 an unpredictable seed -- if it is not unpredictable, there
5533 is no point in blinding anyway). Make RSA blinding thread-safe
5534 by remembering the creator's thread ID in rsa->blinding and
5535 having all other threads use local one-time blinding factors
5536 (this requires more computation than sharing rsa->blinding, but
5537 avoids excessive locking; and if an RSA object is not shared
5538 between threads, blinding will still be very fast).
5541 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5543 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5544 via timing by performing a MAC computation even if incorrrect
5545 block cipher padding has been found. This is a countermeasure
5546 against active attacks where the attacker has to distinguish
5547 between bad padding and a MAC verification error. (CVE-2003-0078)
5549 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5550 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5551 Martin Vuagnoux (EPFL, Ilion)]
5553 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5555 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5556 memory from it's contents. This is done with a counter that will
5557 place alternating values in each byte. This can be used to solve
5558 two issues: 1) the removal of calls to memset() by highly optimizing
5559 compilers, and 2) cleansing with other values than 0, since those can
5560 be read through on certain media, for example a swap space on disk.
5563 *) Bugfix: client side session caching did not work with external caching,
5564 because the session->cipher setting was not restored when reloading
5565 from the external cache. This problem was masked, when
5566 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5567 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5570 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5571 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5572 [Zeev Lieber <zeev-l@yahoo.com>]
5574 *) Undo an undocumented change introduced in 0.9.6e which caused
5575 repeated calls to OpenSSL_add_all_ciphers() and
5576 OpenSSL_add_all_digests() to be ignored, even after calling
5580 *) Change the default configuration reader to deal with last line not
5581 being properly terminated.
5584 *) Change X509_NAME_cmp() so it applies the special rules on handling
5585 DN values that are of type PrintableString, as well as RDNs of type
5586 emailAddress where the value has the type ia5String.
5587 [stefank@valicert.com via Richard Levitte]
5589 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5590 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5591 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5592 the bitwise-OR of the two for use by the majority of applications
5593 wanting this behaviour, and update the docs. The documented
5594 behaviour and actual behaviour were inconsistent and had been
5595 changing anyway, so this is more a bug-fix than a behavioural
5597 [Geoff Thorpe, diagnosed by Nadav Har'El]
5599 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5600 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5603 *) Fix initialization code race conditions in
5604 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5605 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5606 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5607 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5608 ssl2_get_cipher_by_char(),
5609 ssl3_get_cipher_by_char().
5610 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5612 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5613 the cached sessions are flushed, as the remove_cb() might use ex_data
5614 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5615 (see [openssl.org #212]).
5616 [Geoff Thorpe, Lutz Jaenicke]
5618 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5619 length, instead of the encoding length to d2i_ASN1_OBJECT.
5622 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5624 *) [In 0.9.6g-engine release:]
5625 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5626 [Lynn Gazis <lgazis@rainbow.com>]
5628 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5630 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5631 and get fix the header length calculation.
5632 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5633 Alon Kantor <alonk@checkpoint.com> (and others),
5636 *) Use proper error handling instead of 'assertions' in buffer
5637 overflow checks added in 0.9.6e. This prevents DoS (the
5638 assertions could call abort()).
5639 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5641 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5643 *) Add various sanity checks to asn1_get_length() to reject
5644 the ASN1 length bytes if they exceed sizeof(long), will appear
5645 negative or the content length exceeds the length of the
5647 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5649 *) Fix cipher selection routines: ciphers without encryption had no flags
5650 for the cipher strength set and where therefore not handled correctly
5651 by the selection routines (PR #130).
5654 *) Fix EVP_dsa_sha macro.
5658 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5659 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5660 that was added in OpenSSL 0.9.6d.
5662 As the countermeasure turned out to be incompatible with some
5663 broken SSL implementations, the new option is part of SSL_OP_ALL.
5664 SSL_OP_ALL is usually employed when compatibility with weird SSL
5665 implementations is desired (e.g. '-bugs' option to 's_client' and
5666 's_server'), so the new option is automatically set in many
5670 *) Changes in security patch:
5672 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5673 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5674 Air Force Materiel Command, USAF, under agreement number
5677 *) Add various sanity checks to asn1_get_length() to reject
5678 the ASN1 length bytes if they exceed sizeof(long), will appear
5679 negative or the content length exceeds the length of the
5680 supplied buffer. (CVE-2002-0659)
5681 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5683 *) Assertions for various potential buffer overflows, not known to
5685 [Ben Laurie (CHATS)]
5687 *) Various temporary buffers to hold ASCII versions of integers were
5688 too small for 64 bit platforms. (CVE-2002-0655)
5689 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5691 *) Remote buffer overflow in SSL3 protocol - an attacker could
5692 supply an oversized session ID to a client. (CVE-2002-0656)
5693 [Ben Laurie (CHATS)]
5695 *) Remote buffer overflow in SSL2 protocol - an attacker could
5696 supply an oversized client master key. (CVE-2002-0656)
5697 [Ben Laurie (CHATS)]
5699 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5701 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5702 encoded as NULL) with id-dsa-with-sha1.
5703 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5705 *) Check various X509_...() return values in apps/req.c.
5706 [Nils Larsch <nla@trustcenter.de>]
5708 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5709 an end-of-file condition would erronously be flagged, when the CRLF
5710 was just at the end of a processed block. The bug was discovered when
5711 processing data through a buffering memory BIO handing the data to a
5712 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5713 <ptsekov@syntrex.com> and Nedelcho Stanev.
5716 *) Implement a countermeasure against a vulnerability recently found
5717 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5718 before application data chunks to avoid the use of known IVs
5719 with data potentially chosen by the attacker.
5722 *) Fix length checks in ssl3_get_client_hello().
5725 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5726 to prevent ssl3_read_internal() from incorrectly assuming that
5727 ssl3_read_bytes() found application data while handshake
5728 processing was enabled when in fact s->s3->in_read_app_data was
5729 merely automatically cleared during the initial handshake.
5730 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5732 *) Fix object definitions for Private and Enterprise: they were not
5733 recognized in their shortname (=lowercase) representation. Extend
5734 obj_dat.pl to issue an error when using undefined keywords instead
5735 of silently ignoring the problem (Svenning Sorensen
5736 <sss@sss.dnsalias.net>).
5739 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5740 generators, i.e. generators other than 2 and 5. (Previously, the
5741 code did not properly initialise the 'add' and 'rem' values to
5742 BN_generate_prime().)
5744 In the new general case, we do not insist that 'generator' is
5745 actually a primitive root: This requirement is rather pointless;
5746 a generator of the order-q subgroup is just as good, if not
5750 *) Map new X509 verification errors to alerts. Discovered and submitted by
5751 Tom Wu <tom@arcot.com>.
5754 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5755 returning non-zero before the data has been completely received
5756 when using non-blocking I/O.
5757 [Bodo Moeller; problem pointed out by John Hughes]
5759 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5760 [Ben Laurie, Lutz Jaenicke]
5762 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5763 Yoram Zahavi <YoramZ@gilian.com>).
5766 *) Add information about CygWin 1.3 and on, and preserve proper
5767 configuration for the versions before that.
5768 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5770 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5771 check whether we deal with a copy of a session and do not delete from
5772 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5773 <izhar@checkpoint.com>.
5776 *) Do not store session data into the internal session cache, if it
5777 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5778 flag is set). Proposed by Aslam <aslam@funk.com>.
5781 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5785 *) [In 0.9.6d-engine release:]
5786 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5787 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5789 *) Add the configuration target linux-s390x.
5790 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5792 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5793 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5794 variable as an indication that a ClientHello message has been
5795 received. As the flag value will be lost between multiple
5796 invocations of ssl3_accept when using non-blocking I/O, the
5797 function may not be aware that a handshake has actually taken
5798 place, thus preventing a new session from being added to the
5801 To avoid this problem, we now set s->new_session to 2 instead of
5802 using a local variable.
5803 [Lutz Jaenicke, Bodo Moeller]
5805 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5806 if the SSL_R_LENGTH_MISMATCH error is detected.
5807 [Geoff Thorpe, Bodo Moeller]
5809 *) New 'shared_ldflag' column in Configure platform table.
5812 *) Fix EVP_CIPHER_mode macro.
5813 ["Dan S. Camper" <dan@bti.net>]
5815 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5816 type, we must throw them away by setting rr->length to 0.
5817 [D P Chang <dpc@qualys.com>]
5819 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5821 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5822 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5823 worked incorrectly for those cases where range = 10..._2 and
5824 3*range is two bits longer than range.)
5827 *) Only add signing time to PKCS7 structures if it is not already
5831 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5832 OBJ_ld_ce should be OBJ_id_ce.
5833 Also some ip-pda OIDs in crypto/objects/objects.txt were
5834 incorrect (cf. RFC 3039).
5835 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5837 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5838 returns early because it has nothing to do.
5839 [Andy Schneider <andy.schneider@bjss.co.uk>]
5841 *) [In 0.9.6c-engine release:]
5842 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5843 [Andy Schneider <andy.schneider@bjss.co.uk>]
5845 *) [In 0.9.6c-engine release:]
5846 Add support for Cryptographic Appliance's keyserver technology.
5847 (Use engine 'keyclient')
5848 [Cryptographic Appliances and Geoff Thorpe]
5850 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5851 is called via tools/c89.sh because arguments have to be
5852 rearranged (all '-L' options must appear before the first object
5854 [Richard Shapiro <rshapiro@abinitio.com>]
5856 *) [In 0.9.6c-engine release:]
5857 Add support for Broadcom crypto accelerator cards, backported
5859 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5861 *) [In 0.9.6c-engine release:]
5862 Add support for SureWare crypto accelerator cards from
5863 Baltimore Technologies. (Use engine 'sureware')
5864 [Baltimore Technologies and Mark Cox]
5866 *) [In 0.9.6c-engine release:]
5867 Add support for crypto accelerator cards from Accelerated
5868 Encryption Processing, www.aep.ie. (Use engine 'aep')
5869 [AEP Inc. and Mark Cox]
5871 *) Add a configuration entry for gcc on UnixWare.
5872 [Gary Benson <gbenson@redhat.com>]
5874 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5875 messages are stored in a single piece (fixed-length part and
5876 variable-length part combined) and fix various bugs found on the way.
5879 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5880 instead. BIO_gethostbyname() does not know what timeouts are
5881 appropriate, so entries would stay in cache even when they have
5883 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5885 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5886 faced with a pathologically small ClientHello fragment that does
5887 not contain client_version: Instead of aborting with an error,
5888 simply choose the highest available protocol version (i.e.,
5889 TLS 1.0 unless it is disabled). In practice, ClientHello
5890 messages are never sent like this, but this change gives us
5891 strictly correct behaviour at least for TLS.
5894 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5895 never resets s->method to s->ctx->method when called from within
5896 one of the SSL handshake functions.
5897 [Bodo Moeller; problem pointed out by Niko Baric]
5899 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5900 (sent using the client's version number) if client_version is
5901 smaller than the protocol version in use. Also change
5902 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5903 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5904 the client will at least see that alert.
5907 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5911 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5912 client receives HelloRequest while in a handshake.
5913 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5915 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5916 should end in 'break', not 'goto end' which circuments various
5917 cleanups done in state SSL_ST_OK. But session related stuff
5918 must be disabled for SSL_ST_OK in the case that we just sent a
5921 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5922 before just sending a HelloRequest.
5923 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5925 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5926 reveal whether illegal block cipher padding was found or a MAC
5927 verification error occured. (Neither SSLerr() codes nor alerts
5928 are directly visible to potential attackers, but the information
5929 may leak via logfiles.)
5931 Similar changes are not required for the SSL 2.0 implementation
5932 because the number of padding bytes is sent in clear for SSL 2.0,
5933 and the extra bytes are just ignored. However ssl/s2_pkt.c
5934 failed to verify that the purported number of padding bytes is in
5938 *) Add OpenUNIX-8 support including shared libraries
5939 (Boyd Lynn Gerber <gerberb@zenez.com>).
5942 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5943 'wristwatch attack' using huge encoding parameters (cf.
5944 James H. Manger's CRYPTO 2001 paper). Note that the
5945 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5946 encoding parameters and hence was not vulnerable.
5949 *) BN_sqr() bug fix.
5950 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5952 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5953 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5954 followed by modular reduction.
5955 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5957 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5958 equivalent based on BN_pseudo_rand() instead of BN_rand().
5961 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5962 This function was broken, as the check for a new client hello message
5963 to handle SGC did not allow these large messages.
5964 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5967 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5970 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5971 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5974 *) Rework the configuration and shared library support for Tru64 Unix.
5975 The configuration part makes use of modern compiler features and
5976 still retains old compiler behavior for those that run older versions
5977 of the OS. The shared library support part includes a variant that
5978 uses the RPATH feature, and is available through the special
5979 configuration target "alpha-cc-rpath", which will never be selected
5981 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5983 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5984 with the same message size as in ssl3_get_certificate_request().
5985 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5986 messages might inadvertently be reject as too long.
5987 [Petr Lampa <lampa@fee.vutbr.cz>]
5989 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5992 *) Modified SSL library such that the verify_callback that has been set
5993 specificly for an SSL object with SSL_set_verify() is actually being
5994 used. Before the change, a verify_callback set with this function was
5995 ignored and the verify_callback() set in the SSL_CTX at the time of
5996 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5997 to allow the necessary settings.
6000 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6001 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6002 done automatically (in contradiction to the requirements of the C
6003 standard). This made problems when used from OpenSSH.
6006 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6007 dh->length and always used
6009 BN_rand_range(priv_key, dh->p).
6011 BN_rand_range() is not necessary for Diffie-Hellman, and this
6012 specific range makes Diffie-Hellman unnecessarily inefficient if
6013 dh->length (recommended exponent length) is much smaller than the
6014 length of dh->p. We could use BN_rand_range() if the order of
6015 the subgroup was stored in the DH structure, but we only have
6020 BN_rand(priv_key, l, ...)
6022 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6028 RSA_eay_public_encrypt
6029 RSA_eay_private_decrypt
6030 RSA_eay_private_encrypt (signing)
6031 RSA_eay_public_decrypt (signature verification)
6033 (default implementations for RSA_public_encrypt,
6034 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6035 always reject numbers >= n.
6038 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6039 to synchronize access to 'locking_thread'. This is necessary on
6040 systems where access to 'locking_thread' (an 'unsigned long'
6041 variable) is not atomic.
6044 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6045 *before* setting the 'crypto_lock_rand' flag. The previous code had
6046 a race condition if 0 is a valid thread ID.
6047 [Travis Vitek <vitek@roguewave.com>]
6049 *) Add support for shared libraries under Irix.
6050 [Albert Chin-A-Young <china@thewrittenword.com>]
6052 *) Add configuration option to build on Linux on both big-endian and
6054 [Ralf Baechle <ralf@uni-koblenz.de>]
6056 *) Add the possibility to create shared libraries on HP-UX.
6059 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6061 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6062 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6063 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6064 PRNG state recovery was possible based on the output of
6065 one PRNG request appropriately sized to gain knowledge on
6066 'md' followed by enough consecutive 1-byte PRNG requests
6067 to traverse all of 'state'.
6069 1. When updating 'md_local' (the current thread's copy of 'md')
6070 during PRNG output generation, hash all of the previous
6071 'md_local' value, not just the half used for PRNG output.
6073 2. Make the number of bytes from 'state' included into the hash
6074 independent from the number of PRNG bytes requested.
6076 The first measure alone would be sufficient to avoid
6077 Markku-Juhani's attack. (Actually it had never occurred
6078 to me that the half of 'md_local' used for chaining was the
6079 half from which PRNG output bytes were taken -- I had always
6080 assumed that the secret half would be used.) The second
6081 measure makes sure that additional data from 'state' is never
6082 mixed into 'md_local' in small portions; this heuristically
6083 further strengthens the PRNG.
6086 *) Fix crypto/bn/asm/mips3.s.
6089 *) When only the key is given to "enc", the IV is undefined. Print out
6090 an error message in this case.
6093 *) Handle special case when X509_NAME is empty in X509 printing routines.
6096 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6097 positive and less than q.
6100 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6101 used: it isn't thread safe and the add_lock_callback should handle
6103 [Paul Rose <Paul.Rose@bridge.com>]
6105 *) Verify that incoming data obeys the block size in
6106 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6110 [Ulf Möller, Bodo Möller]
6112 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6113 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6114 when fixing the server behaviour for backwards-compatible 'client
6115 hello' messages. (Note that the attack is impractical against
6116 SSL 3.0 and TLS 1.0 anyway because length and version checking
6117 means that the probability of guessing a valid ciphertext is
6118 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6121 Before 0.9.5, the countermeasure (hide the error by generating a
6122 random 'decryption result') did not work properly because
6123 ERR_clear_error() was missing, meaning that SSL_get_error() would
6124 detect the supposedly ignored error.
6126 Both problems are now fixed.
6129 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6130 (previously it was 1024).
6133 *) Fix for compatibility mode trust settings: ignore trust settings
6134 unless some valid trust or reject settings are present.
6137 *) Fix for blowfish EVP: its a variable length cipher.
6140 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6141 parameters in DSA public key structures and return an error in the
6142 DSA routines if parameters are absent.
6145 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6146 in the current directory if neither $RANDFILE nor $HOME was set.
6147 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6148 caused some confusion to Windows users who haven't defined $HOME.
6149 Thus RAND_file_name() is changed again: e_os.h can define a
6150 DEFAULT_HOME, which will be used if $HOME is not set.
6151 For Windows, we use "C:"; on other platforms, we still require
6152 environment variables.
6154 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6155 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6156 having multiple threads call RAND_poll() concurrently.
6159 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6160 combination of a flag and a thread ID variable.
6161 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6162 flag), *other* threads can enter ssleay_add_bytes without obeying
6163 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6164 that they do not hold after the first thread unsets add_do_not_lock).
6167 *) Change bctest again: '-x' expressions are not available in all
6171 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6173 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6174 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6176 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6177 the default extension for executables, if any. Also, make the perl
6178 scripts that use symlink() to test if it really exists and use "cp"
6179 if it doesn't. All this made OpenSSL compilable and installable in
6183 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6184 If SEQUENCE is length is indefinite just set c->slen to the total
6185 amount of data available.
6186 [Steve Henson, reported by shige@FreeBSD.org]
6187 [This change does not apply to 0.9.7.]
6189 *) Change bctest to avoid here-documents inside command substitution
6190 (workaround for FreeBSD /bin/sh bug).
6191 For compatibility with Ultrix, avoid shell functions (introduced
6192 in the bctest version that searches along $PATH).
6195 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6196 with des_encrypt() defined on some operating systems, like Solaris
6200 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6201 On the Importance of Eliminating Errors in Cryptographic
6202 Computations, J. Cryptology 14 (2001) 2, 101-119,
6203 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6206 *) MIPS assembler BIGNUM division bug fix.
6209 *) Disabled incorrect Alpha assembler code.
6212 *) Fix PKCS#7 decode routines so they correctly update the length
6213 after reading an EOC for the EXPLICIT tag.
6215 [This change does not apply to 0.9.7.]
6217 *) Fix bug in PKCS#12 key generation routines. This was triggered
6218 if a 3DES key was generated with a 0 initial byte. Include
6219 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6220 (but broken) behaviour.
6223 *) Enhance bctest to search for a working bc along $PATH and print
6225 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6227 *) Fix memory leaks in err.c: free err_data string if necessary;
6228 don't write to the wrong index in ERR_set_error_data.
6231 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6235 *) Replace rdtsc with _emit statements for VC++ version 5.
6236 [Jeremy Cooper <jeremy@baymoo.org>]
6238 *) Make it possible to reuse SSLv2 sessions.
6241 *) In copy_email() check for >= 0 as a return value for
6242 X509_NAME_get_index_by_NID() since 0 is a valid index.
6243 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6245 *) Avoid coredump with unsupported or invalid public keys by checking if
6246 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6247 PKCS7_verify() fails with non detached data.
6250 *) Don't use getenv in library functions when run as setuid/setgid.
6251 New function OPENSSL_issetugid().
6254 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6255 due to incorrect handling of multi-threading:
6257 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6259 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6261 3. Count how many times MemCheck_off() has been called so that
6262 nested use can be treated correctly. This also avoids
6263 inband-signalling in the previous code (which relied on the
6264 assumption that thread ID 0 is impossible).
6267 *) Add "-rand" option also to s_client and s_server.
6270 *) Fix CPU detection on Irix 6.x.
6271 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6272 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6274 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6277 [This change does not apply to 0.9.7.]
6279 *) Use the cached encoding of an X509_NAME structure rather than
6280 copying it. This is apparently the reason for the libsafe "errors"
6281 but the code is actually correct.
6284 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6285 Bleichenbacher's DSA attack.
6286 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6287 to be set and top=0 forces the highest bit to be set; top=-1 is new
6288 and leaves the highest bit random.
6289 [Ulf Moeller, Bodo Moeller]
6291 *) In the NCONF_...-based implementations for CONF_... queries
6292 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6293 a temporary CONF structure with the data component set to NULL
6294 (which gives segmentation faults in lh_retrieve).
6295 Instead, use NULL for the CONF pointer in CONF_get_string and
6296 CONF_get_number (which may use environment variables) and directly
6297 return NULL from CONF_get_section.
6300 *) Fix potential buffer overrun for EBCDIC.
6303 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6304 keyUsage if basicConstraints absent for a CA.
6307 *) Make SMIME_write_PKCS7() write mail header values with a format that
6308 is more generally accepted (no spaces before the semicolon), since
6309 some programs can't parse those values properly otherwise. Also make
6310 sure BIO's that break lines after each write do not create invalid
6314 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6315 macros previously used would not encode an empty SEQUENCE OF
6316 and break the signature.
6318 [This change does not apply to 0.9.7.]
6320 *) Zero the premaster secret after deriving the master secret in
6324 *) Add some EVP_add_digest_alias registrations (as found in
6325 OpenSSL_add_all_digests()) to SSL_library_init()
6326 aka OpenSSL_add_ssl_algorithms(). This provides improved
6327 compatibility with peers using X.509 certificates
6328 with unconventional AlgorithmIdentifier OIDs.
6331 *) Fix for Irix with NO_ASM.
6332 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6334 *) ./config script fixes.
6335 [Ulf Moeller, Richard Levitte]
6337 *) Fix 'openssl passwd -1'.
6340 *) Change PKCS12_key_gen_asc() so it can cope with non null
6341 terminated strings whose length is passed in the passlen
6342 parameter, for example from PEM callbacks. This was done
6343 by adding an extra length parameter to asc2uni().
6344 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6346 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6347 call failed, free the DSA structure.
6350 *) Fix to uni2asc() to cope with zero length Unicode strings.
6351 These are present in some PKCS#12 files.
6354 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6355 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6356 when writing a 32767 byte record.
6357 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6359 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6360 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6362 (RSA objects have a reference count access to which is protected
6363 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6364 so they are meant to be shared between threads.)
6365 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6366 "Reddie, Steven" <Steven.Reddie@ca.com>]
6368 *) Fix a deadlock in CRYPTO_mem_leaks().
6371 *) Use better test patterns in bntest.
6374 *) rand_win.c fix for Borland C.
6377 *) BN_rshift bugfix for n == 0.
6380 *) Add a 'bctest' script that checks for some known 'bc' bugs
6381 so that 'make test' does not abort just because 'bc' is broken.
6384 *) Store verify_result within SSL_SESSION also for client side to
6385 avoid potential security hole. (Re-used sessions on the client side
6386 always resulted in verify_result==X509_V_OK, not using the original
6387 result of the server certificate verification.)
6390 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6391 SSL3_RT_APPLICATION_DATA, return 0.
6392 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6396 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6397 releases, have been re-implemented by renaming the previous
6398 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6399 and ssl3_read_internal, respectively, and adding 'peek' parameters
6400 to them. The new ssl[23]_{read,peek} functions are calls to
6401 ssl[23]_read_internal with the 'peek' flag set appropriately.
6402 A 'peek' parameter has also been added to ssl3_read_bytes, which
6403 does the actual work for ssl3_read_internal.
6406 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6407 the method-specific "init()" handler. Also clean up ex_data after
6408 calling the method-specific "finish()" handler. Previously, this was
6409 happening the other way round.
6412 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6413 The previous value, 12, was not always sufficient for BN_mod_exp().
6416 *) Make sure that shared libraries get the internal name engine with
6417 the full version number and not just 0. This should mark the
6418 shared libraries as not backward compatible. Of course, this should
6419 be changed again when we can guarantee backward binary compatibility.
6422 *) Fix typo in get_cert_by_subject() in by_dir.c
6423 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6425 *) Rework the system to generate shared libraries:
6427 - Make note of the expected extension for the shared libraries and
6428 if there is a need for symbolic links from for example libcrypto.so.0
6429 to libcrypto.so.0.9.7. There is extended info in Configure for
6432 - Make as few rebuilds of the shared libraries as possible.
6434 - Still avoid linking the OpenSSL programs with the shared libraries.
6436 - When installing, install the shared libraries separately from the
6440 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6442 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6443 and not in SSL_clear because the latter is also used by the
6444 accept/connect functions; previously, the settings made by
6445 SSL_set_read_ahead would be lost during the handshake.
6446 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6448 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6449 Previously, it would create entries for disableed algorithms no
6453 *) Added several new manual pages for SSL_* function.
6456 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6458 *) In ssl23_get_client_hello, generate an error message when faced
6459 with an initial SSL 3.0/TLS record that is too small to contain the
6460 first two bytes of the ClientHello message, i.e. client_version.
6461 (Note that this is a pathologic case that probably has never happened
6462 in real life.) The previous approach was to use the version number
6463 from the record header as a substitute; but our protocol choice
6464 should not depend on that one because it is not authenticated
6465 by the Finished messages.
6468 *) More robust randomness gathering functions for Windows.
6469 [Jeffrey Altman <jaltman@columbia.edu>]
6471 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6472 not set then we don't setup the error code for issuer check errors
6473 to avoid possibly overwriting other errors which the callback does
6474 handle. If an application does set the flag then we assume it knows
6475 what it is doing and can handle the new informational codes
6479 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6480 a general "ANY" type, as such it should be able to decode anything
6481 including tagged types. However it didn't check the class so it would
6482 wrongly interpret tagged types in the same way as their universal
6483 counterpart and unknown types were just rejected. Changed so that the
6484 tagged and unknown types are handled in the same way as a SEQUENCE:
6485 that is the encoding is stored intact. There is also a new type
6486 "V_ASN1_OTHER" which is used when the class is not universal, in this
6487 case we have no idea what the actual type is so we just lump them all
6491 *) On VMS, stdout may very well lead to a file that is written to
6492 in a record-oriented fashion. That means that every write() will
6493 write a separate record, which will be read separately by the
6494 programs trying to read from it. This can be very confusing.
6496 The solution is to put a BIO filter in the way that will buffer
6497 text until a linefeed is reached, and then write everything a
6498 line at a time, so every record written will be an actual line,
6499 not chunks of lines and not (usually doesn't happen, but I've
6500 seen it once) several lines in one record. BIO_f_linebuffer() is
6503 Currently, it's a VMS-only method, because that's where it has
6504 been tested well enough.
6507 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6508 it can return incorrect results.
6509 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6510 but it was in 0.9.6-beta[12].)
6513 *) Disable the check for content being present when verifying detached
6514 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6515 include zero length content when signing messages.
6518 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6519 BIO_ctrl (for BIO pairs).
6522 *) Add DSO method for VMS.
6525 *) Bug fix: Montgomery multiplication could produce results with the
6529 *) Add RPM specification openssl.spec and modify it to build three
6530 packages. The default package contains applications, application
6531 documentation and run-time libraries. The devel package contains
6532 include files, static libraries and function documentation. The
6533 doc package contains the contents of the doc directory. The original
6534 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6537 *) Add a large number of documentation files for many SSL routines.
6538 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6540 *) Add a configuration entry for Sony News 4.
6541 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6543 *) Don't set the two most significant bits to one when generating a
6544 random number < q in the DSA library.
6547 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6548 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6549 the underlying transport is blocking) if a handshake took place.
6550 (The default behaviour is needed by applications such as s_client
6551 and s_server that use select() to determine when to use SSL_read;
6552 but for applications that know in advance when to expect data, it
6553 just makes things more complicated.)
6556 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6560 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6561 work better on such systems.
6562 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6564 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6565 Update PKCS12_parse() so it copies the friendlyName and the
6566 keyid to the certificates aux info.
6569 *) Fix bug in PKCS7_verify() which caused an infinite loop
6570 if there was more than one signature.
6571 [Sven Uszpelkat <su@celocom.de>]
6573 *) Major change in util/mkdef.pl to include extra information
6574 about each symbol, as well as presentig variables as well
6575 as functions. This change means that there's n more need
6576 to rebuild the .num files when some algorithms are excluded.
6579 *) Allow the verify time to be set by an application,
6580 rather than always using the current time.
6583 *) Phase 2 verify code reorganisation. The certificate
6584 verify code now looks up an issuer certificate by a
6585 number of criteria: subject name, authority key id
6586 and key usage. It also verifies self signed certificates
6587 by the same criteria. The main comparison function is
6588 X509_check_issued() which performs these checks.
6590 Lot of changes were necessary in order to support this
6591 without completely rewriting the lookup code.
6593 Authority and subject key identifier are now cached.
6595 The LHASH 'certs' is X509_STORE has now been replaced
6596 by a STACK_OF(X509_OBJECT). This is mainly because an
6597 LHASH can't store or retrieve multiple objects with
6598 the same hash value.
6600 As a result various functions (which were all internal
6601 use only) have changed to handle the new X509_STORE
6602 structure. This will break anything that messed round
6603 with X509_STORE internally.
6605 The functions X509_STORE_add_cert() now checks for an
6606 exact match, rather than just subject name.
6608 The X509_STORE API doesn't directly support the retrieval
6609 of multiple certificates matching a given criteria, however
6610 this can be worked round by performing a lookup first
6611 (which will fill the cache with candidate certificates)
6612 and then examining the cache for matches. This is probably
6613 the best we can do without throwing out X509_LOOKUP
6614 entirely (maybe later...).
6616 The X509_VERIFY_CTX structure has been enhanced considerably.
6618 All certificate lookup operations now go via a get_issuer()
6619 callback. Although this currently uses an X509_STORE it
6620 can be replaced by custom lookups. This is a simple way
6621 to bypass the X509_STORE hackery necessary to make this
6622 work and makes it possible to use more efficient techniques
6623 in future. A very simple version which uses a simple
6624 STACK for its trusted certificate store is also provided
6625 using X509_STORE_CTX_trusted_stack().
6627 The verify_cb() and verify() callbacks now have equivalents
6628 in the X509_STORE_CTX structure.
6630 X509_STORE_CTX also has a 'flags' field which can be used
6631 to customise the verify behaviour.
6634 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6635 excludes S/MIME capabilities.
6638 *) When a certificate request is read in keep a copy of the
6639 original encoding of the signed data and use it when outputing
6640 again. Signatures then use the original encoding rather than
6641 a decoded, encoded version which may cause problems if the
6642 request is improperly encoded.
6645 *) For consistency with other BIO_puts implementations, call
6646 buffer_write(b, ...) directly in buffer_puts instead of calling
6649 In BIO_puts, increment b->num_write as in BIO_write.
6650 [Peter.Sylvester@EdelWeb.fr]
6652 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6653 BN_zero, we may not return a BIGNUM with an array consisting of
6657 *) Avoid calling abort() from within the library when problems are
6658 detected, except if preprocessor symbols have been defined
6659 (such as REF_CHECK, BN_DEBUG etc.).
6662 *) New openssl application 'rsautl'. This utility can be
6663 used for low level RSA operations. DER public key
6664 BIO/fp routines also added.
6667 *) New Configure entry and patches for compiling on QNX 4.
6668 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6670 *) A demo state-machine implementation was sponsored by
6671 Nuron (http://www.nuron.com/) and is now available in
6672 demos/state_machine.
6675 *) New options added to the 'dgst' utility for signature
6676 generation and verification.
6679 *) Unrecognized PKCS#7 content types are now handled via a
6680 catch all ASN1_TYPE structure. This allows unsupported
6681 types to be stored as a "blob" and an application can
6682 encode and decode it manually.
6685 *) Fix various signed/unsigned issues to make a_strex.c
6687 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6689 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6690 length if passed a buffer. ASN1_INTEGER_to_BN failed
6691 if passed a NULL BN and its argument was negative.
6692 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6694 *) Modification to PKCS#7 encoding routines to output definite
6695 length encoding. Since currently the whole structures are in
6696 memory there's not real point in using indefinite length
6697 constructed encoding. However if OpenSSL is compiled with
6698 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6701 *) Added BIO_vprintf() and BIO_vsnprintf().
6704 *) Added more prefixes to parse for in the the strings written
6705 through a logging bio, to cover all the levels that are available
6706 through syslog. The prefixes are now:
6708 PANIC, EMERG, EMR => LOG_EMERG
6709 ALERT, ALR => LOG_ALERT
6710 CRIT, CRI => LOG_CRIT
6711 ERROR, ERR => LOG_ERR
6712 WARNING, WARN, WAR => LOG_WARNING
6713 NOTICE, NOTE, NOT => LOG_NOTICE
6714 INFO, INF => LOG_INFO
6715 DEBUG, DBG => LOG_DEBUG
6717 and as before, if none of those prefixes are present at the
6718 beginning of the string, LOG_ERR is chosen.
6720 On Win32, the LOG_* levels are mapped according to this:
6722 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6723 LOG_WARNING => EVENTLOG_WARNING_TYPE
6724 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6728 *) Made it possible to reconfigure with just the configuration
6729 argument "reconf" or "reconfigure". The command line arguments
6730 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6731 and are retrieved from there when reconfiguring.
6735 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6737 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6740 *) The obj_dat.pl script was messing up the sorting of object
6741 names. The reason was that it compared the quoted version
6742 of strings as a result "OCSP" > "OCSP Signing" because
6743 " > SPACE. Changed script to store unquoted versions of
6744 names and add quotes on output. It was also omitting some
6745 names from the lookup table if they were given a default
6746 value (that is if SN is missing it is given the same
6747 value as LN and vice versa), these are now added on the
6748 grounds that if an object has a name we should be able to
6749 look it up. Finally added warning output when duplicate
6750 short or long names are found.
6753 *) Changes needed for Tandem NSK.
6754 [Scott Uroff <scott@xypro.com>]
6756 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6757 RSA_padding_check_SSLv23(), special padding was never detected
6758 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6759 version rollback attacks was not effective.
6761 In s23_clnt.c, don't use special rollback-attack detection padding
6762 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6763 client; similarly, in s23_srvr.c, don't do the rollback check if
6764 SSL 2.0 is the only protocol enabled in the server.
6767 *) Make it possible to get hexdumps of unprintable data with 'openssl
6768 asn1parse'. By implication, the functions ASN1_parse_dump() and
6769 BIO_dump_indent() are added.
6772 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6773 these print out strings and name structures based on various
6774 flags including RFC2253 support and proper handling of
6775 multibyte characters. Added options to the 'x509' utility
6776 to allow the various flags to be set.
6779 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6780 Also change the functions X509_cmp_current_time() and
6781 X509_gmtime_adj() work with an ASN1_TIME structure,
6782 this will enable certificates using GeneralizedTime in validity
6783 dates to be checked.
6786 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6787 negative public key encodings) on by default,
6788 NO_NEG_PUBKEY_BUG can be set to disable it.
6791 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6792 content octets. An i2c_ASN1_OBJECT is unnecessary because
6793 the encoding can be trivially obtained from the structure.
6796 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6797 not read locks (CRYPTO_r_[un]lock).
6800 *) A first attempt at creating official support for shared
6801 libraries through configuration. I've kept it so the
6802 default is static libraries only, and the OpenSSL programs
6803 are always statically linked for now, but there are
6804 preparations for dynamic linking in place.
6805 This has been tested on Linux and Tru64.
6808 *) Randomness polling function for Win9x, as described in:
6809 Peter Gutmann, Software Generation of Practically Strong
6813 *) Fix so PRNG is seeded in req if using an already existing
6817 *) New options to smime application. -inform and -outform
6818 allow alternative formats for the S/MIME message including
6819 PEM and DER. The -content option allows the content to be
6820 specified separately. This should allow things like Netscape
6821 form signing output easier to verify.
6824 *) Fix the ASN1 encoding of tags using the 'long form'.
6827 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6828 STRING types. These convert content octets to and from the
6829 underlying type. The actual tag and length octets are
6830 already assumed to have been read in and checked. These
6831 are needed because all other string types have virtually
6832 identical handling apart from the tag. By having versions
6833 of the ASN1 functions that just operate on content octets
6834 IMPLICIT tagging can be handled properly. It also allows
6835 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6836 and ASN1_INTEGER are identical apart from the tag.
6839 *) Change the handling of OID objects as follows:
6841 - New object identifiers are inserted in objects.txt, following
6842 the syntax given in objects.README.
6843 - objects.pl is used to process obj_mac.num and create a new
6845 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6848 This is currently kind of a hack, and the perl code in objects.pl
6849 isn't very elegant, but it works as I intended. The simplest way
6850 to check that it worked correctly is to look in obj_dat.h and
6851 check the array nid_objs and make sure the objects haven't moved
6852 around (this is important!). Additions are OK, as well as
6853 consistent name changes.
6856 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6859 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6860 The given file adds to whatever has already been seeded into the
6861 random pool through the RANDFILE configuration file option or
6862 environment variable, or the default random state file.
6865 *) mkstack.pl now sorts each macro group into lexical order.
6866 Previously the output order depended on the order the files
6867 appeared in the directory, resulting in needless rewriting
6871 *) Patches to make OpenSSL compile under Win32 again. Mostly
6872 work arounds for the VC++ problem that it treats func() as
6873 func(void). Also stripped out the parts of mkdef.pl that
6874 added extra typesafe functions: these no longer exist.
6877 *) Reorganisation of the stack code. The macros are now all
6878 collected in safestack.h . Each macro is defined in terms of
6879 a "stack macro" of the form SKM_<name>(type, a, b). The
6880 DEBUG_SAFESTACK is now handled in terms of function casts,
6881 this has the advantage of retaining type safety without the
6882 use of additional functions. If DEBUG_SAFESTACK is not defined
6883 then the non typesafe macros are used instead. Also modified the
6884 mkstack.pl script to handle the new form. Needs testing to see
6885 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6886 the default if no major problems. Similar behaviour for ASN1_SET_OF
6887 and PKCS12_STACK_OF.
6890 *) When some versions of IIS use the 'NET' form of private key the
6891 key derivation algorithm is different. Normally MD5(password) is
6892 used as a 128 bit RC4 key. In the modified case
6893 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6894 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6895 as the old Netscape_RSA functions except they have an additional
6896 'sgckey' parameter which uses the modified algorithm. Also added
6897 an -sgckey command line option to the rsa utility. Thanks to
6898 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6899 algorithm to openssl-dev.
6902 *) The evp_local.h macros were using 'c.##kname' which resulted in
6903 invalid expansion on some systems (SCO 5.0.5 for example).
6904 Corrected to 'c.kname'.
6905 [Phillip Porch <root@theporch.com>]
6907 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6908 a STACK of email addresses from a certificate or request, these look
6909 in the subject name and the subject alternative name extensions and
6910 omit any duplicate addresses.
6913 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6914 This makes DSA verification about 2 % faster.
6917 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6918 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6919 plus overhead for 1024 bit moduli).
6920 This makes exponentiations about 0.5 % faster for 1024 bit
6921 exponents (as measured by "openssl speed rsa2048").
6924 *) Rename memory handling macros to avoid conflicts with other
6926 Malloc => OPENSSL_malloc
6927 Malloc_locked => OPENSSL_malloc_locked
6928 Realloc => OPENSSL_realloc
6929 Free => OPENSSL_free
6932 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6933 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6936 *) CygWin32 support.
6937 [John Jarvie <jjarvie@newsguy.com>]
6939 *) The type-safe stack code has been rejigged. It is now only compiled
6940 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6941 by default all type-specific stack functions are "#define"d back to
6942 standard stack functions. This results in more streamlined output
6943 but retains the type-safety checking possibilities of the original
6947 *) The STACK code has been cleaned up, and certain type declarations
6948 that didn't make a lot of sense have been brought in line. This has
6949 also involved a cleanup of sorts in safestack.h to more correctly
6950 map type-safe stack functions onto their plain stack counterparts.
6951 This work has also resulted in a variety of "const"ifications of
6952 lots of the code, especially "_cmp" operations which should normally
6953 be prototyped with "const" parameters anyway.
6956 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6957 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6958 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6959 where all of 'md' is used each time the PRNG is used, but 'state'
6960 is used only indexed by a cyclic counter. As entropy may not be
6961 well distributed from the beginning, 'md' is important as a
6962 chaining variable. However, the output function chains only half
6963 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6964 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6965 in all of 'state' being rewritten, with the new values depending
6966 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6969 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6970 the handshake is continued after ssl_verify_cert_chain();
6971 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6972 can lead to 'unexplainable' connection aborts later.
6973 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6975 *) Major EVP API cipher revision.
6976 Add hooks for extra EVP features. This allows various cipher
6977 parameters to be set in the EVP interface. Support added for variable
6978 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6979 setting of RC2 and RC5 parameters.
6981 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6984 Remove lots of duplicated code from the EVP library. For example *every*
6985 cipher init() function handles the 'iv' in the same way according to the
6986 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6987 for CFB and OFB modes they zero ctx->num.
6989 New functionality allows removal of S/MIME code RC2 hack.
6991 Most of the routines have the same form and so can be declared in terms
6994 By shifting this to the top level EVP_CipherInit() it can be removed from
6995 all individual ciphers. If the cipher wants to handle IVs or keys
6996 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6999 Change lots of functions like EVP_EncryptUpdate() to now return a
7000 value: although software versions of the algorithms cannot fail
7001 any installed hardware versions can.
7004 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7005 this option is set, tolerate broken clients that send the negotiated
7006 protocol version number instead of the requested protocol version
7010 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7011 i.e. non-zero for export ciphersuites, zero otherwise.
7012 Previous versions had this flag inverted, inconsistent with
7013 rsa_tmp_cb (..._TMP_RSA_CB).
7014 [Bodo Moeller; problem reported by Amit Chopra]
7016 *) Add missing DSA library text string. Work around for some IIS
7017 key files with invalid SEQUENCE encoding.
7020 *) Add a document (doc/standards.txt) that list all kinds of standards
7021 and so on that are implemented in OpenSSL.
7024 *) Enhance c_rehash script. Old version would mishandle certificates
7025 with the same subject name hash and wouldn't handle CRLs at all.
7026 Added -fingerprint option to crl utility, to support new c_rehash
7030 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7033 *) Fix for SSL server purpose checking. Server checking was
7034 rejecting certificates which had extended key usage present
7035 but no ssl client purpose.
7036 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7038 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7039 is a little unclear about how a blank password is handled.
7040 Since the password in encoded as a BMPString with terminating
7041 double NULL a zero length password would end up as just the
7042 double NULL. However no password at all is different and is
7043 handled differently in the PKCS#12 key generation code. NS
7044 treats a blank password as zero length. MSIE treats it as no
7045 password on export: but it will try both on import. We now do
7046 the same: PKCS12_parse() tries zero length and no password if
7047 the password is set to "" or NULL (NULL is now a valid password:
7048 it wasn't before) as does the pkcs12 application.
7051 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7052 perror when PEM_read_bio_X509_REQ fails, the error message must
7053 be obtained from the error queue.
7056 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7057 it in ERR_remove_state if appropriate, and change ERR_get_state
7058 accordingly to avoid race conditions (this is necessary because
7059 thread_hash is no longer constant once set).
7062 *) Bugfix for linux-elf makefile.one.
7065 *) RSA_get_default_method() will now cause a default
7066 RSA_METHOD to be chosen if one doesn't exist already.
7067 Previously this was only set during a call to RSA_new()
7068 or RSA_new_method(NULL) meaning it was possible for
7069 RSA_get_default_method() to return NULL.
7072 *) Added native name translation to the existing DSO code
7073 that will convert (if the flag to do so is set) filenames
7074 that are sufficiently small and have no path information
7075 into a canonical native form. Eg. "blah" converted to
7076 "libblah.so" or "blah.dll" etc.
7079 *) New function ERR_error_string_n(e, buf, len) which is like
7080 ERR_error_string(e, buf), but writes at most 'len' bytes
7081 including the 0 terminator. For ERR_error_string_n, 'buf'
7083 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7085 *) CONF library reworked to become more general. A new CONF
7086 configuration file reader "class" is implemented as well as a
7087 new functions (NCONF_*, for "New CONF") to handle it. The now
7088 old CONF_* functions are still there, but are reimplemented to
7089 work in terms of the new functions. Also, a set of functions
7090 to handle the internal storage of the configuration data is
7091 provided to make it easier to write new configuration file
7092 reader "classes" (I can definitely see something reading a
7093 configuration file in XML format, for example), called _CONF_*,
7094 or "the configuration storage API"...
7096 The new configuration file reading functions are:
7098 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7099 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7101 NCONF_default, NCONF_WIN32
7103 NCONF_dump_fp, NCONF_dump_bio
7105 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7106 NCONF_new creates a new CONF object. This works in the same way
7107 as other interfaces in OpenSSL, like the BIO interface.
7108 NCONF_dump_* dump the internal storage of the configuration file,
7109 which is useful for debugging. All other functions take the same
7110 arguments as the old CONF_* functions wth the exception of the
7111 first that must be a `CONF *' instead of a `LHASH *'.
7113 To make it easer to use the new classes with the old CONF_* functions,
7114 the function CONF_set_default_method is provided.
7117 *) Add '-tls1' option to 'openssl ciphers', which was already
7118 mentioned in the documentation but had not been implemented.
7119 (This option is not yet really useful because even the additional
7120 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7123 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7124 OpenSSL-based applications) load shared libraries and bind to
7125 them in a portable way.
7126 [Geoff Thorpe, with contributions from Richard Levitte]
7128 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7130 *) Make sure _lrotl and _lrotr are only used with MSVC.
7132 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7133 (the default implementation of RAND_status).
7135 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7136 to '-clrext' (= clear extensions), as intended and documented.
7137 [Bodo Moeller; inconsistency pointed out by Michael Attili
7140 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7141 was larger than the MD block size.
7142 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7144 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7145 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7146 using the passed key: if the passed key was a private key the result
7147 of X509_print(), for example, would be to print out all the private key
7151 *) des_quad_cksum() byte order bug fix.
7152 [Ulf Möller, using the problem description in krb4-0.9.7, where
7153 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7155 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7157 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7159 *) For easily testing in shell scripts whether some command
7160 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7161 returns with exit code 0 iff no command of the given name is available.
7162 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7163 the output goes to stdout and nothing is printed to stderr.
7164 Additional arguments are always ignored.
7166 Since for each cipher there is a command of the same name,
7167 the 'no-cipher' compilation switches can be tested this way.
7169 ('openssl no-XXX' is not able to detect pseudo-commands such
7170 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7173 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7176 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7177 is set; it will be thrown away anyway because each handshake creates
7179 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7180 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7181 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7182 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7185 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7186 'Q' and 'R' lose their special meanings (quit/renegotiate).
7187 This is part of what -quiet does; unlike -quiet, -ign_eof
7188 does not suppress any output.
7191 *) Add compatibility options to the purpose and trust code. The
7192 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7193 accepts a certificate or CA, this was the previous behaviour,
7194 with all the associated security issues.
7196 X509_TRUST_COMPAT is the old trust behaviour: only and
7197 automatically trust self signed roots in certificate store. A
7198 new trust setting X509_TRUST_DEFAULT is used to specify that
7199 a purpose has no associated trust setting and it should instead
7200 use the value in the default purpose.
7203 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7204 and fix a memory leak.
7207 *) In util/mkerr.pl (which implements 'make errors'), preserve
7208 reason strings from the previous version of the .c file, as
7209 the default to have only downcase letters (and digits) in
7210 automatically generated reasons codes is not always appropriate.
7213 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7214 using strerror. Previously, ERR_reason_error_string() returned
7215 library names as reason strings for SYSerr; but SYSerr is a special
7216 case where small numbers are errno values, not library numbers.
7219 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7220 converts DSA parameters into DH parameters. (When creating parameters,
7221 DSA_generate_parameters is used.)
7224 *) Include 'length' (recommended exponent length) in C code generated
7225 by 'openssl dhparam -C'.
7228 *) The second argument to set_label in perlasm was already being used
7229 so couldn't be used as a "file scope" flag. Moved to third argument
7233 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7234 instead of RAND_bytes for encryption IVs and salts.
7237 *) Include RAND_status() into RAND_METHOD instead of implementing
7238 it only for md_rand.c Otherwise replacing the PRNG by calling
7239 RAND_set_rand_method would be impossible.
7242 *) Don't let DSA_generate_key() enter an infinite loop if the random
7243 number generation fails.
7246 *) New 'rand' application for creating pseudo-random output.
7249 *) Added configuration support for Linux/IA64
7250 [Rolf Haberrecker <rolf@suse.de>]
7252 *) Assembler module support for Mingw32.
7255 *) Shared library support for HPUX (in shlib/).
7256 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7258 *) Shared library support for Solaris gcc.
7259 [Lutz Behnke <behnke@trustcenter.de>]
7261 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7263 *) PKCS7_encrypt() was adding text MIME headers twice because they
7264 were added manually and by SMIME_crlf_copy().
7267 *) In bntest.c don't call BN_rand with zero bits argument.
7268 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7270 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7271 case was implemented. This caused BN_div_recp() to fail occasionally.
7274 *) Add an optional second argument to the set_label() in the perl
7275 assembly language builder. If this argument exists and is set
7276 to 1 it signals that the assembler should use a symbol whose
7277 scope is the entire file, not just the current function. This
7278 is needed with MASM which uses the format label:: for this scope.
7279 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7281 *) Change the ASN1 types so they are typedefs by default. Before
7282 almost all types were #define'd to ASN1_STRING which was causing
7283 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7287 *) Change names of new functions to the new get1/get0 naming
7288 convention: After 'get1', the caller owns a reference count
7289 and has to call ..._free; 'get0' returns a pointer to some
7290 data structure without incrementing reference counters.
7291 (Some of the existing 'get' functions increment a reference
7292 counter, some don't.)
7293 Similarly, 'set1' and 'add1' functions increase reference
7294 counters or duplicate objects.
7297 *) Allow for the possibility of temp RSA key generation failure:
7298 the code used to assume it always worked and crashed on failure.
7301 *) Fix potential buffer overrun problem in BIO_printf().
7302 [Ulf Möller, using public domain code by Patrick Powell; problem
7303 pointed out by David Sacerdote <das33@cornell.edu>]
7305 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7306 RAND_egd() and RAND_status(). In the command line application,
7307 the EGD socket can be specified like a seed file using RANDFILE
7311 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7312 Some CAs (e.g. Verisign) distribute certificates in this form.
7315 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7316 list to exclude them. This means that no special compilation option
7317 is needed to use anonymous DH: it just needs to be included in the
7321 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7322 EVP_MD_type. The old functionality is available in a new macro called
7323 EVP_MD_md(). Change code that uses it and update docs.
7326 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7327 where the 'void *' argument is replaced by a function pointer argument.
7328 Previously 'void *' was abused to point to functions, which works on
7329 many platforms, but is not correct. As these functions are usually
7330 called by macros defined in OpenSSL header files, most source code
7331 should work without changes.
7334 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7335 sections with information on -D... compiler switches used for
7336 compiling the library so that applications can see them. To enable
7337 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7338 must be defined. E.g.,
7339 #define OPENSSL_ALGORITHM_DEFINES
7340 #include <openssl/opensslconf.h>
7341 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7342 [Richard Levitte, Ulf and Bodo Möller]
7344 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7348 *) Change the 'other' type in certificate aux info to a STACK_OF
7349 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7350 the required ASN1 format: arbitrary types determined by an OID.
7353 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7354 argument to 'req'. This is not because the function is newer or
7355 better than others it just uses the work 'NEW' in the certificate
7356 request header lines. Some software needs this.
7359 *) Reorganise password command line arguments: now passwords can be
7360 obtained from various sources. Delete the PEM_cb function and make
7361 it the default behaviour: i.e. if the callback is NULL and the
7362 usrdata argument is not NULL interpret it as a null terminated pass
7363 phrase. If usrdata and the callback are NULL then the pass phrase
7364 is prompted for as usual.
7367 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7368 the support is automatically enabled. The resulting binaries will
7369 autodetect the card and use it if present.
7370 [Ben Laurie and Compaq Inc.]
7372 *) Work around for Netscape hang bug. This sends certificate request
7373 and server done in one record. Since this is perfectly legal in the
7374 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7375 the bugs/SSLv3 entry for more info.
7378 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7381 *) Add -rand argument to smime and pkcs12 applications and read/write
7385 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7388 *) Add command line password options to the remaining applications.
7391 *) Bug fix for BN_div_recp() for numerators with an even number of
7395 *) More tests in bntest.c, and changed test_bn output.
7398 *) ./config recognizes MacOS X now.
7401 *) Bug fix for BN_div() when the first words of num and divsor are
7402 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7405 *) Add support for various broken PKCS#8 formats, and command line
7406 options to produce them.
7409 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7410 get temporary BIGNUMs from a BN_CTX.
7413 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7417 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7418 include a #define from the old name to the new. The original intent
7419 was that statically linked binaries could for example just call
7420 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7421 link with digests. This never worked becayse SSLeay_add_all_digests()
7422 and SSLeay_add_all_ciphers() were in the same source file so calling
7423 one would link with the other. They are now in separate source files.
7426 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7429 *) Use a less unusual form of the Miller-Rabin primality test (it used
7430 a binary algorithm for exponentiation integrated into the Miller-Rabin
7431 loop, our standard modexp algorithms are faster).
7434 *) Support for the EBCDIC character set completed.
7435 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7437 *) Source code cleanups: use const where appropriate, eliminate casts,
7438 use void * instead of char * in lhash.
7441 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7442 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7443 this the server could overwrite ephemeral keys that the client
7447 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7448 using 50 iterations of the Rabin-Miller test.
7450 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7451 iterations of the Rabin-Miller test as required by the appendix
7452 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7453 As BN_is_prime_fasttest includes trial division, DSA parameter
7454 generation becomes much faster.
7456 This implies a change for the callback functions in DSA_is_prime
7457 and DSA_generate_parameters: The callback function is called once
7458 for each positive witness in the Rabin-Miller test, not just
7459 occasionally in the inner loop; and the parameters to the
7460 callback function now provide an iteration count for the outer
7461 loop rather than for the current invocation of the inner loop.
7462 DSA_generate_parameters additionally can call the callback
7463 function with an 'iteration count' of -1, meaning that a
7464 candidate has passed the trial division test (when q is generated
7465 from an application-provided seed, trial division is skipped).
7468 *) New function BN_is_prime_fasttest that optionally does trial
7469 division before starting the Rabin-Miller test and has
7470 an additional BN_CTX * argument (whereas BN_is_prime always
7471 has to allocate at least one BN_CTX).
7472 'callback(1, -1, cb_arg)' is called when a number has passed the
7473 trial division stage.
7476 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7480 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7483 *) New function BN_pseudo_rand().
7486 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7487 bignum version of BN_from_montgomery() with the working code from
7488 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7492 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7493 made it impossible to use the same SSL_SESSION data structure in
7494 SSL2 clients in multiple threads.
7497 *) The return value of RAND_load_file() no longer counts bytes obtained
7498 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7499 to seed the PRNG (previously an explicit byte count was required).
7500 [Ulf Möller, Bodo Möller]
7502 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7503 used (char *) instead of (void *) and had casts all over the place.
7506 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7509 *) Retain source code compatibility for BN_prime_checks macro:
7510 BN_is_prime(..., BN_prime_checks, ...) now uses
7511 BN_prime_checks_for_size to determine the appropriate number of
7512 Rabin-Miller iterations.
7515 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7516 DH_CHECK_P_NOT_SAFE_PRIME.
7517 (Check if this is true? OpenPGP calls them "strong".)
7520 *) Merge the functionality of "dh" and "gendh" programs into a new program
7521 "dhparam". The old programs are retained for now but will handle DH keys
7522 (instead of parameters) in future.
7525 *) Make the ciphers, s_server and s_client programs check the return values
7526 when a new cipher list is set.
7529 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7530 ciphers. Before when the 56bit ciphers were enabled the sorting was
7533 The syntax for the cipher sorting has been extended to support sorting by
7534 cipher-strength (using the strength_bits hard coded in the tables).
7535 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7537 Fix a bug in the cipher-command parser: when supplying a cipher command
7538 string with an "undefined" symbol (neither command nor alphanumeric
7539 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7540 an error is flagged.
7542 Due to the strength-sorting extension, the code of the
7543 ssl_create_cipher_list() function was completely rearranged. I hope that
7544 the readability was also increased :-)
7545 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7547 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7548 for the first serial number and places 2 in the serial number file. This
7549 avoids problems when the root CA is created with serial number zero and
7550 the first user certificate has the same issuer name and serial number
7554 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7555 the new code. Add documentation for this stuff.
7558 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7559 X509_*() to X509at_*() on the grounds that they don't handle X509
7560 structures and behave in an analagous way to the X509v3 functions:
7561 they shouldn't be called directly but wrapper functions should be used
7564 So we also now have some wrapper functions that call the X509at functions
7565 when passed certificate requests. (TO DO: similar things can be done with
7566 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7567 things. Some of these need some d2i or i2d and print functionality
7568 because they handle more complex structures.)
7571 *) Add missing #ifndefs that caused missing symbols when building libssl
7572 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7573 NO_RSA in ssl/s2*.c.
7574 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7576 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7577 has a return value which indicates the quality of the random data
7578 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7579 error queue. New function RAND_pseudo_bytes() generates output that is
7580 guaranteed to be unique but not unpredictable. RAND_add is like
7581 RAND_seed, but takes an extra argument for an entropy estimate
7582 (RAND_seed always assumes full entropy).
7585 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7586 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7587 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7588 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7589 false-positive rate of at most 2^-80 for random input.
7592 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7595 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7596 in the 0.9.5 release), this returns the chain
7597 from an X509_CTX structure with a dup of the stack and all
7598 the X509 reference counts upped: so the stack will exist
7599 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7602 Also make SSL_SESSION_print() print out the verify return
7606 *) Add manpage for the pkcs12 command. Also change the default
7607 behaviour so MAC iteration counts are used unless the new
7608 -nomaciter option is used. This improves file security and
7609 only older versions of MSIE (4.0 for example) need it.
7612 *) Honor the no-xxx Configure options when creating .DEF files.
7615 *) Add PKCS#10 attributes to field table: challengePassword,
7616 unstructuredName and unstructuredAddress. These are taken from
7617 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7618 international characters are used.
7620 More changes to X509_ATTRIBUTE code: allow the setting of types
7621 based on strings. Remove the 'loc' parameter when adding
7622 attributes because these will be a SET OF encoding which is sorted
7626 *) Initial changes to the 'req' utility to allow request generation
7627 automation. This will allow an application to just generate a template
7628 file containing all the field values and have req construct the
7631 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7632 used all over the place including certificate requests and PKCS#7
7633 structures. They are currently handled manually where necessary with
7634 some primitive wrappers for PKCS#7. The new functions behave in a
7635 manner analogous to the X509 extension functions: they allow
7636 attributes to be looked up by NID and added.
7638 Later something similar to the X509V3 code would be desirable to
7639 automatically handle the encoding, decoding and printing of the
7640 more complex types. The string types like challengePassword can
7641 be handled by the string table functions.
7643 Also modified the multi byte string table handling. Now there is
7644 a 'global mask' which masks out certain types. The table itself
7645 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7646 is useful when for example there is only one permissible type
7647 (as in countryName) and using the mask might result in no valid
7651 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7652 SSL_get_peer_finished to allow applications to obtain the latest
7653 Finished messages sent to the peer or expected from the peer,
7654 respectively. (SSL_get_peer_finished is usually the Finished message
7655 actually received from the peer, otherwise the protocol will be aborted.)
7657 As the Finished message are message digests of the complete handshake
7658 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7659 be used for external authentication procedures when the authentication
7660 provided by SSL/TLS is not desired or is not enough.
7663 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7664 the host supports BWX extension and if Compaq C is present on the
7665 $PATH. Just exploiting of the BWX extension results in 20-30%
7666 performance kick for some algorithms, e.g. DES and RC4 to mention
7667 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7671 *) Add support for MS "fast SGC". This is arguably a violation of the
7672 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7673 weak crypto and after checking the certificate is SGC a second one
7674 with strong crypto. MS SGC stops the first handshake after receiving
7675 the server certificate message and sends a second client hello. Since
7676 a server will typically do all the time consuming operations before
7677 expecting any further messages from the client (server key exchange
7678 is the most expensive) there is little difference between the two.
7680 To get OpenSSL to support MS SGC we have to permit a second client
7681 hello message after we have sent server done. In addition we have to
7682 reset the MAC if we do get this second client hello.
7685 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7686 if a DER encoded private key is RSA or DSA traditional format. Changed
7687 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7688 format DER encoded private key. Newer code should use PKCS#8 format which
7689 has the key type encoded in the ASN1 structure. Added DER private key
7690 support to pkcs8 application.
7693 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7694 ciphersuites has been selected (as required by the SSL 3/TLS 1
7695 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7696 is set, we interpret this as a request to violate the specification
7697 (the worst that can happen is a handshake failure, and 'correct'
7698 behaviour would result in a handshake failure anyway).
7701 *) In SSL_CTX_add_session, take into account that there might be multiple
7702 SSL_SESSION structures with the same session ID (e.g. when two threads
7703 concurrently obtain them from an external cache).
7704 The internal cache can handle only one SSL_SESSION with a given ID,
7705 so if there's a conflict, we now throw out the old one to achieve
7709 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7710 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7711 some routines that use cipher OIDs: some ciphers do not have OIDs
7712 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7716 *) Simplify the trust setting structure and code. Now we just have
7717 two sequences of OIDs for trusted and rejected settings. These will
7718 typically have values the same as the extended key usage extension
7719 and any application specific purposes.
7721 The trust checking code now has a default behaviour: it will just
7722 check for an object with the same NID as the passed id. Functions can
7723 be provided to override either the default behaviour or the behaviour
7724 for a given id. SSL client, server and email already have functions
7725 in place for compatibility: they check the NID and also return "trusted"
7726 if the certificate is self signed.
7729 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7730 traditional format into an EVP_PKEY structure.
7733 *) Add a password callback function PEM_cb() which either prompts for
7734 a password if usr_data is NULL or otherwise assumes it is a null
7735 terminated password. Allow passwords to be passed on command line
7736 environment or config files in a few more utilities.
7739 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7740 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7741 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7742 Update documentation.
7745 *) Support for ASN1 "NULL" type. This could be handled before by using
7746 ASN1_TYPE but there wasn't any function that would try to read a NULL
7747 and produce an error if it couldn't. For compatibility we also have
7748 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7749 don't allocate anything because they don't need to.
7752 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7754 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7756 *) Rebuild of the memory allocation routines used by OpenSSL code and
7757 possibly others as well. The purpose is to make an interface that
7758 provide hooks so anyone can build a separate set of allocation and
7759 deallocation routines to be used by OpenSSL, for example memory
7760 pool implementations, or something else, which was previously hard
7761 since Malloc(), Realloc() and Free() were defined as macros having
7762 the values malloc, realloc and free, respectively (except for Win32
7763 compilations). The same is provided for memory debugging code.
7764 OpenSSL already comes with functionality to find memory leaks, but
7765 this gives people a chance to debug other memory problems.
7767 With these changes, a new set of functions and macros have appeared:
7769 CRYPTO_set_mem_debug_functions() [F]
7770 CRYPTO_get_mem_debug_functions() [F]
7771 CRYPTO_dbg_set_options() [F]
7772 CRYPTO_dbg_get_options() [F]
7773 CRYPTO_malloc_debug_init() [M]
7775 The memory debug functions are NULL by default, unless the library
7776 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7777 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7778 gives the standard debugging functions that come with OpenSSL) or
7779 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7780 provided by the library user) must be used. When the standard
7781 debugging functions are used, CRYPTO_dbg_set_options can be used to
7782 request additional information:
7783 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7784 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7786 Also, things like CRYPTO_set_mem_functions will always give the
7787 expected result (the new set of functions is used for allocation
7788 and deallocation) at all times, regardless of platform and compiler
7791 To finish it up, some functions that were never use in any other
7792 way than through macros have a new API and new semantic:
7795 CRYPTO_dbg_realloc()
7798 All macros of value have retained their old syntax.
7799 [Richard Levitte and Bodo Moeller]
7801 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7802 ordering of SMIMECapabilities wasn't in "strength order" and there
7803 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7807 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7808 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7809 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7811 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7812 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7813 functionality to handle multipart/signed properly) and a utility
7814 called 'smime' to call all this stuff. This is based on code I
7815 originally wrote for Celo who have kindly allowed it to be
7816 included in OpenSSL.
7819 *) Add variants des_set_key_checked and des_set_key_unchecked of
7820 des_set_key (aka des_key_sched). Global variable des_check_key
7821 decides which of these is called by des_set_key; this way
7822 des_check_key behaves as it always did, but applications and
7823 the library itself, which was buggy for des_check_key == 1,
7824 have a cleaner way to pick the version they need.
7827 *) New function PKCS12_newpass() which changes the password of a
7831 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7832 dynamic mix. In both cases the ids can be used as an index into the
7833 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7834 functions so they accept a list of the field values and the
7835 application doesn't need to directly manipulate the X509_TRUST
7839 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7843 *) Modify the way the V3 extension code looks up extensions. This now
7844 works in a similar way to the object code: we have some "standard"
7845 extensions in a static table which is searched with OBJ_bsearch()
7846 and the application can add dynamic ones if needed. The file
7847 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7848 updated whenever a new extension is added to the core code and kept
7849 in ext_nid order. There is a simple program 'tabtest.c' which checks
7850 this. New extensions are not added too often so this file can readily
7851 be maintained manually.
7853 There are two big advantages in doing things this way. The extensions
7854 can be looked up immediately and no longer need to be "added" using
7855 X509V3_add_standard_extensions(): this function now does nothing.
7856 [Side note: I get *lots* of email saying the extension code doesn't
7857 work because people forget to call this function]
7858 Also no dynamic allocation is done unless new extensions are added:
7859 so if we don't add custom extensions there is no need to call
7860 X509V3_EXT_cleanup().
7863 *) Modify enc utility's salting as follows: make salting the default. Add a
7864 magic header, so unsalted files fail gracefully instead of just decrypting
7865 to garbage. This is because not salting is a big security hole, so people
7866 should be discouraged from doing it.
7869 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7870 digest to be passed on the command line but it only used this
7871 parameter when signing a certificate. Modified so all relevant
7872 operations are affected by the digest parameter including the
7873 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7874 DSA key was used because it didn't fix the digest.
7877 *) Initial certificate chain verify code. Currently tests the untrusted
7878 certificates for consistency with the verify purpose (which is set
7879 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7881 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7882 this is because it will reject chains with invalid extensions whereas
7883 every previous version of OpenSSL and SSLeay made no checks at all.
7885 Trust code: checks the root CA for the relevant trust settings. Trust
7886 settings have an initial value consistent with the verify purpose: e.g.
7887 if the verify purpose is for SSL client use it expects the CA to be
7888 trusted for SSL client use. However the default value can be changed to
7889 permit custom trust settings: one example of this would be to only trust
7890 certificates from a specific "secure" set of CAs.
7892 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7893 which should be used for version portability: especially since the
7894 verify structure is likely to change more often now.
7896 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7897 to set them. If not set then assume SSL clients will verify SSL servers
7900 Two new options to the verify program: -untrusted allows a set of
7901 untrusted certificates to be passed in and -purpose which sets the
7902 intended purpose of the certificate. If a purpose is set then the
7903 new chain verify code is used to check extension consistency.
7906 *) Support for the authority information access extension.
7909 *) Modify RSA and DSA PEM read routines to transparently handle
7910 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7911 public keys in a format compatible with certificate
7912 SubjectPublicKeyInfo structures. Unfortunately there were already
7913 functions called *_PublicKey_* which used various odd formats so
7914 these are retained for compatibility: however the DSA variants were
7915 never in a public release so they have been deleted. Changed dsa/rsa
7916 utilities to handle the new format: note no releases ever handled public
7917 keys so we should be OK.
7919 The primary motivation for this change is to avoid the same fiasco
7920 that dogs private keys: there are several incompatible private key
7921 formats some of which are standard and some OpenSSL specific and
7922 require various evil hacks to allow partial transparent handling and
7923 even then it doesn't work with DER formats. Given the option anything
7924 other than PKCS#8 should be dumped: but the other formats have to
7925 stay in the name of compatibility.
7927 With public keys and the benefit of hindsight one standard format
7928 is used which works with EVP_PKEY, RSA or DSA structures: though
7929 it clearly returns an error if you try to read the wrong kind of key.
7931 Added a -pubkey option to the 'x509' utility to output the public key.
7932 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7933 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7934 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7935 that do the same as the EVP_PKEY_assign_*() except they up the
7936 reference count of the added key (they don't "swallow" the
7940 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7941 CRLs would fail if the file contained no certificates or no CRLs:
7942 added a new function to read in both types and return the number
7943 read: this means that if none are read it will be an error. The
7944 DER versions of the certificate and CRL reader would always fail
7945 because it isn't possible to mix certificates and CRLs in DER format
7946 without choking one or the other routine. Changed this to just read
7947 a certificate: this is the best we can do. Also modified the code
7948 in apps/verify.c to take notice of return codes: it was previously
7949 attempting to read in certificates from NULL pointers and ignoring
7950 any errors: this is one reason why the cert and CRL reader seemed
7951 to work. It doesn't check return codes from the default certificate
7952 routines: these may well fail if the certificates aren't installed.
7955 *) Code to support otherName option in GeneralName.
7958 *) First update to verify code. Change the verify utility
7959 so it warns if it is passed a self signed certificate:
7960 for consistency with the normal behaviour. X509_verify
7961 has been modified to it will now verify a self signed
7962 certificate if *exactly* the same certificate appears
7963 in the store: it was previously impossible to trust a
7964 single self signed certificate. This means that:
7965 openssl verify ss.pem
7966 now gives a warning about a self signed certificate but
7967 openssl verify -CAfile ss.pem ss.pem
7971 *) For servers, store verify_result in SSL_SESSION data structure
7972 (and add it to external session representation).
7973 This is needed when client certificate verifications fails,
7974 but an application-provided verification callback (set by
7975 SSL_CTX_set_cert_verify_callback) allows accepting the session
7976 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7977 but returns 1): When the session is reused, we have to set
7978 ssl->verify_result to the appropriate error code to avoid
7980 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7982 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7983 case in PKCS7_dataInit() where the signed PKCS7 structure
7984 didn't contain any existing data because it was being created.
7985 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7987 *) Add a salt to the key derivation routines in enc.c. This
7988 forms the first 8 bytes of the encrypted file. Also add a
7989 -S option to allow a salt to be input on the command line.
7992 *) New function X509_cmp(). Oddly enough there wasn't a function
7993 to compare two certificates. We do this by working out the SHA1
7994 hash and comparing that. X509_cmp() will be needed by the trust
7998 *) SSL_get1_session() is like SSL_get_session(), but increments
7999 the reference count in the SSL_SESSION returned.
8000 [Geoff Thorpe <geoff@eu.c2.net>]
8002 *) Fix for 'req': it was adding a null to request attributes.
8003 Also change the X509_LOOKUP and X509_INFO code to handle
8004 certificate auxiliary information.
8007 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8011 *) Add the possibility to add extra information to the memory leak
8012 detecting output, to form tracebacks, showing from where each
8013 allocation was originated: CRYPTO_push_info("constant string") adds
8014 the string plus current file name and line number to a per-thread
8015 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8016 is like calling CYRPTO_pop_info() until the stack is empty.
8017 Also updated memory leak detection code to be multi-thread-safe.
8020 *) Add options -text and -noout to pkcs7 utility and delete the
8021 encryption options which never did anything. Update docs.
8024 *) Add options to some of the utilities to allow the pass phrase
8025 to be included on either the command line (not recommended on
8026 OSes like Unix) or read from the environment. Update the
8027 manpages and fix a few bugs.
8030 *) Add a few manpages for some of the openssl commands.
8033 *) Fix the -revoke option in ca. It was freeing up memory twice,
8034 leaking and not finding already revoked certificates.
8037 *) Extensive changes to support certificate auxiliary information.
8038 This involves the use of X509_CERT_AUX structure and X509_AUX
8039 functions. An X509_AUX function such as PEM_read_X509_AUX()
8040 can still read in a certificate file in the usual way but it
8041 will also read in any additional "auxiliary information". By
8042 doing things this way a fair degree of compatibility can be
8043 retained: existing certificates can have this information added
8044 using the new 'x509' options.
8046 Current auxiliary information includes an "alias" and some trust
8047 settings. The trust settings will ultimately be used in enhanced
8048 certificate chain verification routines: currently a certificate
8049 can only be trusted if it is self signed and then it is trusted
8053 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8054 The problem was that one of the replacement routines had not been working
8055 since SSLeay releases. For now the offending routine has been replaced
8056 with non-optimised assembler. Even so, this now gives around 95%
8057 performance improvement for 1024 bit RSA signs.
8060 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8061 handling. Most clients have the effective key size in bits equal to
8062 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8063 A few however don't do this and instead use the size of the decrypted key
8064 to determine the RC2 key length and the AlgorithmIdentifier to determine
8065 the effective key length. In this case the effective key length can still
8066 be 40 bits but the key length can be 168 bits for example. This is fixed
8067 by manually forcing an RC2 key into the EVP_PKEY structure because the
8068 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8069 the key length and effective key length are equal.
8072 *) Add a bunch of functions that should simplify the creation of
8073 X509_NAME structures. Now you should be able to do:
8074 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8075 and have it automatically work out the correct field type and fill in
8076 the structures. The more adventurous can try:
8077 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8078 and it will (hopefully) work out the correct multibyte encoding.
8081 *) Change the 'req' utility to use the new field handling and multibyte
8082 copy routines. Before the DN field creation was handled in an ad hoc
8083 way in req, ca, and x509 which was rather broken and didn't support
8084 BMPStrings or UTF8Strings. Since some software doesn't implement
8085 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8086 using the dirstring_type option. See the new comment in the default
8087 openssl.cnf for more info.
8090 *) Make crypto/rand/md_rand.c more robust:
8091 - Assure unique random numbers after fork().
8092 - Make sure that concurrent threads access the global counter and
8093 md serializably so that we never lose entropy in them
8094 or use exactly the same state in multiple threads.
8095 Access to the large state is not always serializable because
8096 the additional locking could be a performance killer, and
8097 md should be large enough anyway.
8100 *) New file apps/app_rand.c with commonly needed functionality
8101 for handling the random seed file.
8103 Use the random seed file in some applications that previously did not:
8105 dsaparam -genkey (which also ignored its '-rand' option),
8108 x509 (when signing).
8109 Except on systems with /dev/urandom, it is crucial to have a random
8110 seed file at least for key creation, DSA signing, and for DH exchanges;
8111 for RSA signatures we could do without one.
8113 gendh and gendsa (unlike genrsa) used to read only the first byte
8114 of each file listed in the '-rand' option. The function as previously
8115 found in genrsa is now in app_rand.c and is used by all programs
8116 that support '-rand'.
8119 *) In RAND_write_file, use mode 0600 for creating files;
8120 don't just chmod when it may be too late.
8123 *) Report an error from X509_STORE_load_locations
8124 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8127 *) New function ASN1_mbstring_copy() this copies a string in either
8128 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8129 into an ASN1_STRING type. A mask of permissible types is passed
8130 and it chooses the "minimal" type to use or an error if not type
8134 *) Add function equivalents to the various macros in asn1.h. The old
8135 macros are retained with an M_ prefix. Code inside the library can
8136 use the M_ macros. External code (including the openssl utility)
8137 should *NOT* in order to be "shared library friendly".
8140 *) Add various functions that can check a certificate's extensions
8141 to see if it usable for various purposes such as SSL client,
8142 server or S/MIME and CAs of these types. This is currently
8143 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8144 verification. Also added a -purpose flag to x509 utility to
8145 print out all the purposes.
8148 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8152 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8153 for, obtain and decode and extension and obtain its critical flag.
8154 This allows all the necessary extension code to be handled in a
8155 single function call.
8158 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8159 platforms. See crypto/rc4/rc4_enc.c for further details.
8162 *) New -noout option to asn1parse. This causes no output to be produced
8163 its main use is when combined with -strparse and -out to extract data
8164 from a file (which may not be in ASN.1 format).
8167 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8168 when producing the local key id.
8169 [Richard Levitte <levitte@stacken.kth.se>]
8171 *) New option -dhparam in s_server. This allows a DH parameter file to be
8172 stated explicitly. If it is not stated then it tries the first server
8173 certificate file. The previous behaviour hard coded the filename
8177 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8178 a public key to be input or output. For example:
8179 openssl rsa -in key.pem -pubout -out pubkey.pem
8180 Also added necessary DSA public key functions to handle this.
8183 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8184 in the message. This was handled by allowing
8185 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8186 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8188 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8189 to the end of the strings whereas this didn't. This would cause problems
8190 if strings read with d2i_ASN1_bytes() were later modified.
8191 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8193 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8194 data and it contains EOF it will end up returning an error. This is
8195 caused by input 46 bytes long. The cause is due to the way base64
8196 BIOs find the start of base64 encoded data. They do this by trying a
8197 trial decode on each line until they find one that works. When they
8198 do a flag is set and it starts again knowing it can pass all the
8199 data directly through the decoder. Unfortunately it doesn't reset
8200 the context it uses. This means that if EOF is reached an attempt
8201 is made to pass two EOFs through the context and this causes the
8202 resulting error. This can also cause other problems as well. As is
8203 usual with these problems it takes *ages* to find and the fix is
8204 trivial: move one line.
8205 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8207 *) Ugly workaround to get s_client and s_server working under Windows. The
8208 old code wouldn't work because it needed to select() on sockets and the
8209 tty (for keypresses and to see if data could be written). Win32 only
8210 supports select() on sockets so we select() with a 1s timeout on the
8211 sockets and then see if any characters are waiting to be read, if none
8212 are present then we retry, we also assume we can always write data to
8213 the tty. This isn't nice because the code then blocks until we've
8214 received a complete line of data and it is effectively polling the
8215 keyboard at 1s intervals: however it's quite a bit better than not
8216 working at all :-) A dedicated Windows application might handle this
8217 with an event loop for example.
8220 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8221 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8222 will be called when RSA_sign() and RSA_verify() are used. This is useful
8223 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8224 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8225 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8226 This necessitated the support of an extra signature type NID_md5_sha1
8227 for SSL signatures and modifications to the SSL library to use it instead
8228 of calling RSA_public_decrypt() and RSA_private_encrypt().
8231 *) Add new -verify -CAfile and -CApath options to the crl program, these
8232 will lookup a CRL issuers certificate and verify the signature in a
8233 similar way to the verify program. Tidy up the crl program so it
8234 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8235 less strict. It will now permit CRL extensions even if it is not
8236 a V2 CRL: this will allow it to tolerate some broken CRLs.
8239 *) Initialize all non-automatic variables each time one of the openssl
8240 sub-programs is started (this is necessary as they may be started
8241 multiple times from the "OpenSSL>" prompt).
8242 [Lennart Bang, Bodo Moeller]
8244 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8245 removing all other RSA functionality (this is what NO_RSA does). This
8246 is so (for example) those in the US can disable those operations covered
8247 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8251 *) Non-copying interface to BIO pairs.
8252 (still largely untested)
8255 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8256 ASCII string. This was handled independently in various places before.
8259 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8260 UTF8 strings a character at a time.
8263 *) Use client_version from client hello to select the protocol
8264 (s23_srvr.c) and for RSA client key exchange verification
8265 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8268 *) Add various utility functions to handle SPKACs, these were previously
8269 handled by poking round in the structure internals. Added new function
8270 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8271 print, verify and generate SPKACs. Based on an original idea from
8272 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8275 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8278 *) Allow the config file extension section to be overwritten on the
8279 command line. Based on an original idea from Massimiliano Pala
8280 <madwolf@comune.modena.it>. The new option is called -extensions
8281 and can be applied to ca, req and x509. Also -reqexts to override
8282 the request extensions in req and -crlexts to override the crl extensions
8286 *) Add new feature to the SPKAC handling in ca. Now you can include
8287 the same field multiple times by preceding it by "XXXX." for example:
8290 this is the same syntax as used in the req config file.
8293 *) Allow certificate extensions to be added to certificate requests. These
8294 are specified in a 'req_extensions' option of the req section of the
8295 config file. They can be printed out with the -text option to req but
8296 are otherwise ignored at present.
8299 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8300 data read consists of only the final block it would not decrypted because
8301 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8302 A misplaced 'break' also meant the decrypted final block might not be
8303 copied until the next read.
8306 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8307 a few extra parameters to the DH structure: these will be useful if
8308 for example we want the value of 'q' or implement X9.42 DH.
8311 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8312 provides hooks that allow the default DSA functions or functions on a
8313 "per key" basis to be replaced. This allows hardware acceleration and
8314 hardware key storage to be handled without major modification to the
8315 library. Also added low level modexp hooks and CRYPTO_EX structure and
8316 associated functions.
8319 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8320 as "read only": it can't be written to and the buffer it points to will
8321 not be freed. Reading from a read only BIO is much more efficient than
8322 a normal memory BIO. This was added because there are several times when
8323 an area of memory needs to be read from a BIO. The previous method was
8324 to create a memory BIO and write the data to it, this results in two
8325 copies of the data and an O(n^2) reading algorithm. There is a new
8326 function BIO_new_mem_buf() which creates a read only memory BIO from
8327 an area of memory. Also modified the PKCS#7 routines to use read only
8331 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8332 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8333 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8334 but a retry condition occured while trying to read the rest.
8337 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8338 NID_pkcs7_encrypted by default: this was wrong since this should almost
8339 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8340 the encrypted data type: this is a more sensible place to put it and it
8341 allows the PKCS#12 code to be tidied up that duplicated this
8345 *) Changed obj_dat.pl script so it takes its input and output files on
8346 the command line. This should avoid shell escape redirection problems
8350 *) Initial support for certificate extension requests, these are included
8351 in things like Xenroll certificate requests. Included functions to allow
8352 extensions to be obtained and added.
8355 *) -crlf option to s_client and s_server for sending newlines as
8356 CRLF (as required by many protocols).
8359 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8361 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8362 [Ralf S. Engelschall]
8364 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8365 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8367 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8371 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8372 DH parameters/keys (q is lost during that conversion, but the resulting
8373 DH parameters contain its length).
8375 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8376 much faster than DH_generate_parameters (which creates parameters
8377 where p = 2*q + 1), and also the smaller q makes DH computations
8378 much more efficient (160-bit exponentiation instead of 1024-bit
8379 exponentiation); so this provides a convenient way to support DHE
8380 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8381 utter importance to use
8382 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8384 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8385 when such DH parameters are used, because otherwise small subgroup
8386 attacks may become possible!
8389 *) Avoid memory leak in i2d_DHparams.
8392 *) Allow the -k option to be used more than once in the enc program:
8393 this allows the same encrypted message to be read by multiple recipients.
8396 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8397 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8398 it will always use the numerical form of the OID, even if it has a short
8402 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8403 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8404 otherwise bn_mod_exp was called. In the case of hardware keys for example
8405 no private key components need be present and it might store extra data
8406 in the RSA structure, which cannot be accessed from bn_mod_exp.
8407 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8408 private key operations.
8411 *) Added support for SPARC Linux.
8414 *) pem_password_cb function type incompatibly changed from
8415 typedef int pem_password_cb(char *buf, int size, int rwflag);
8417 ....(char *buf, int size, int rwflag, void *userdata);
8418 so that applications can pass data to their callbacks:
8419 The PEM[_ASN1]_{read,write}... functions and macros now take an
8420 additional void * argument, which is just handed through whenever
8421 the password callback is called.
8422 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8424 New function SSL_CTX_set_default_passwd_cb_userdata.
8426 Compatibility note: As many C implementations push function arguments
8427 onto the stack in reverse order, the new library version is likely to
8428 interoperate with programs that have been compiled with the old
8429 pem_password_cb definition (PEM_whatever takes some data that
8430 happens to be on the stack as its last argument, and the callback
8431 just ignores this garbage); but there is no guarantee whatsoever that
8434 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8435 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8436 problems not only on Windows, but also on some Unix platforms.
8437 To avoid problematic command lines, these definitions are now in an
8438 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8439 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8442 *) MIPS III/IV assembler module is reimplemented.
8445 *) More DES library cleanups: remove references to srand/rand and
8446 delete an unused file.
8449 *) Add support for the the free Netwide assembler (NASM) under Win32,
8450 since not many people have MASM (ml) and it can be hard to obtain.
8451 This is currently experimental but it seems to work OK and pass all
8452 the tests. Check out INSTALL.W32 for info.
8455 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8456 without temporary keys kept an extra copy of the server key,
8457 and connections with temporary keys did not free everything in case
8461 *) New function RSA_check_key and new openssl rsa option -check
8462 for verifying the consistency of RSA keys.
8463 [Ulf Moeller, Bodo Moeller]
8465 *) Various changes to make Win32 compile work:
8466 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8467 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8468 comparison" warnings.
8469 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8472 *) Add a debugging option to PKCS#5 v2 key generation function: when
8473 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8474 derived keys are printed to stderr.
8477 *) Copy the flags in ASN1_STRING_dup().
8478 [Roman E. Pavlov <pre@mo.msk.ru>]
8480 *) The x509 application mishandled signing requests containing DSA
8481 keys when the signing key was also DSA and the parameters didn't match.
8483 It was supposed to omit the parameters when they matched the signing key:
8484 the verifying software was then supposed to automatically use the CA's
8485 parameters if they were absent from the end user certificate.
8487 Omitting parameters is no longer recommended. The test was also
8488 the wrong way round! This was probably due to unusual behaviour in
8489 EVP_cmp_parameters() which returns 1 if the parameters match.
8490 This meant that parameters were omitted when they *didn't* match and
8491 the certificate was useless. Certificates signed with 'ca' didn't have
8493 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8495 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8496 The interface is as follows:
8497 Applications can use
8498 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8499 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8500 "off" is now the default.
8501 The library internally uses
8502 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8503 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8504 to disable memory-checking temporarily.
8506 Some inconsistent states that previously were possible (and were
8507 even the default) are now avoided.
8509 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8510 with each memory chunk allocated; this is occasionally more helpful
8511 than just having a counter.
8513 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8515 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8519 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8520 which largely parallels "options", but is for changing API behaviour,
8521 whereas "options" are about protocol behaviour.
8522 Initial "mode" flags are:
8524 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8525 a single record has been written.
8526 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8527 retries use the same buffer location.
8528 (But all of the contents must be
8532 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8535 *) Fix problems with no-hmac etc.
8536 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8538 *) New functions RSA_get_default_method(), RSA_set_method() and
8539 RSA_get_method(). These allows replacement of RSA_METHODs without having
8540 to mess around with the internals of an RSA structure.
8543 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8544 Also really enable memory leak checks in openssl.c and in some
8546 [Chad C. Mulligan, Bodo Moeller]
8548 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8549 up the length of negative integers. This has now been simplified to just
8550 store the length when it is first determined and use it later, rather
8551 than trying to keep track of where data is copied and updating it to
8553 [Steve Henson, reported by Brien Wheeler
8554 <bwheeler@authentica-security.com>]
8556 *) Add a new function PKCS7_signatureVerify. This allows the verification
8557 of a PKCS#7 signature but with the signing certificate passed to the
8558 function itself. This contrasts with PKCS7_dataVerify which assumes the
8559 certificate is present in the PKCS#7 structure. This isn't always the
8560 case: certificates can be omitted from a PKCS#7 structure and be
8561 distributed by "out of band" means (such as a certificate database).
8564 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8565 function prototypes in pem.h, also change util/mkdef.pl to add the
8566 necessary function names.
8569 *) mk1mf.pl (used by Windows builds) did not properly read the
8570 options set by Configure in the top level Makefile, and Configure
8571 was not even able to write more than one option correctly.
8572 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8575 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8576 file to be loaded from a BIO or FILE pointer. The BIO version will
8577 for example allow memory BIOs to contain config info.
8580 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8581 Whoever hopes to achieve shared-library compatibility across versions
8582 must use this, not the compile-time macro.
8583 (Exercise 0.9.4: Which is the minimum library version required by
8585 Note: All this applies only to multi-threaded programs, others don't
8589 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8590 through a BIO pair triggered the default case, i.e.
8591 SSLerr(...,SSL_R_UNKNOWN_STATE).
8594 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8595 can use the SSL library even if none of the specific BIOs is
8599 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8600 for the encoded length.
8601 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8603 *) Add initial documentation of the X509V3 functions.
8606 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8607 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8608 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8609 secure PKCS#8 private key format with a high iteration count.
8612 *) Fix determination of Perl interpreter: A perl or perl5
8613 _directory_ in $PATH was also accepted as the interpreter.
8614 [Ralf S. Engelschall]
8616 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8617 wrong with it but it was very old and did things like calling
8618 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8622 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8623 to use the new extension code.
8626 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8627 with macros. This should make it easier to change their form, add extra
8628 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8632 *) Add to configuration table a new entry that can specify an alternative
8633 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8634 according to Mark Crispin <MRC@Panda.COM>.
8638 *) DES CBC did not update the IV. Weird.
8641 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8642 Changing the behaviour of the former might break existing programs --
8643 where IV updating is needed, des_ncbc_encrypt can be used.
8646 *) When bntest is run from "make test" it drives bc to check its
8647 calculations, as well as internally checking them. If an internal check
8648 fails, it needs to cause bc to give a non-zero result or make test carries
8649 on without noticing the failure. Fixed.
8652 *) DES library cleanups.
8655 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8656 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8657 ciphers. NOTE: although the key derivation function has been verified
8658 against some published test vectors it has not been extensively tested
8659 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8663 *) Instead of "mkdir -p", which is not fully portable, use new
8664 Perl script "util/mkdir-p.pl".
8667 *) Rewrite the way password based encryption (PBE) is handled. It used to
8668 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8669 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8670 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8671 the 'parameter' field of the AlgorithmIdentifier is passed to the
8672 underlying key generation function so it must do its own ASN1 parsing.
8673 This has also changed the EVP_PBE_CipherInit() function which now has a
8674 'parameter' argument instead of literal salt and iteration count values
8675 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8678 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8679 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8680 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8681 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8682 value was just used as a "magic string" and not used directly its
8683 value doesn't matter.
8686 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8690 *) "linux-sparc64" configuration (ultrapenguin).
8691 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8692 "linux-sparc" configuration.
8693 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8695 *) config now generates no-xxx options for missing ciphers.
8698 *) Support the EBCDIC character set (work in progress).
8699 File ebcdic.c not yet included because it has a different license.
8700 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8702 *) Support BS2000/OSD-POSIX.
8703 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8705 *) Make callbacks for key generation use void * instead of char *.
8708 *) Make S/MIME samples compile (not yet tested).
8711 *) Additional typesafe stacks.
8714 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8718 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8720 *) New configuration variant "sco5-gcc".
8722 *) Updated some demos.
8723 [Sean O Riordain, Wade Scholine]
8725 *) Add missing BIO_free at exit of pkcs12 application.
8728 *) Fix memory leak in conf.c.
8731 *) Updates for Win32 to assembler version of MD5.
8734 *) Set #! path to perl in apps/der_chop to where we found it
8735 instead of using a fixed path.
8738 *) SHA library changes for irix64-mips4-cc.
8741 *) Improvements for VMS support.
8745 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8747 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8748 This also avoids the problems with SC4.2 and unpatched SC5.
8749 [Andy Polyakov <appro@fy.chalmers.se>]
8751 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8752 These are required because of the typesafe stack would otherwise break
8753 existing code. If old code used a structure member which used to be STACK
8754 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8755 sk_num or sk_value it would produce an error because the num, data members
8756 are not present in STACK_OF. Now it just produces a warning. sk_set
8757 replaces the old method of assigning a value to sk_value
8758 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8759 that does this will no longer work (and should use sk_set instead) but
8760 this could be regarded as a "questionable" behaviour anyway.
8763 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8764 correctly handle encrypted S/MIME data.
8767 *) Change type of various DES function arguments from des_cblock
8768 (which means, in function argument declarations, pointer to char)
8769 to des_cblock * (meaning pointer to array with 8 char elements),
8770 which allows the compiler to do more typechecking; it was like
8771 that back in SSLeay, but with lots of ugly casts.
8773 Introduce new type const_des_cblock.
8776 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8777 problems: find RecipientInfo structure that matches recipient certificate
8778 and initialise the ASN1 structures properly based on passed cipher.
8781 *) Belatedly make the BN tests actually check the results.
8784 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8785 to and from BNs: it was completely broken. New compilation option
8786 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8787 key elements as negative integers.
8790 *) Reorganize and speed up MD5.
8791 [Andy Polyakov <appro@fy.chalmers.se>]
8794 [Richard Levitte <richard@levitte.org>]
8796 *) New option -out to asn1parse to allow the parsed structure to be
8797 output to a file. This is most useful when combined with the -strparse
8798 option to examine the output of things like OCTET STRINGS.
8801 *) Make SSL library a little more fool-proof by not requiring any longer
8802 that SSL_set_{accept,connect}_state be called before
8803 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8804 in many applications because usually everything *appeared* to work as
8805 intended anyway -- now it really works as intended).
8808 *) Move openssl.cnf out of lib/.
8811 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8812 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8813 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8814 [Ralf S. Engelschall]
8816 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8817 handle PKCS#7 enveloped data properly.
8818 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8820 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8821 copying pointers. The cert_st handling is changed by this in
8822 various ways (and thus what used to be known as ctx->default_cert
8823 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8824 any longer when s->cert does not give us what we need).
8825 ssl_cert_instantiate becomes obsolete by this change.
8826 As soon as we've got the new code right (possibly it already is?),
8827 we have solved a couple of bugs of the earlier code where s->cert
8828 was used as if it could not have been shared with other SSL structures.
8830 Note that using the SSL API in certain dirty ways now will result
8831 in different behaviour than observed with earlier library versions:
8832 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8833 does not influence s as it used to.
8835 In order to clean up things more thoroughly, inside SSL_SESSION
8836 we don't use CERT any longer, but a new structure SESS_CERT
8837 that holds per-session data (if available); currently, this is
8838 the peer's certificate chain and, for clients, the server's certificate
8839 and temporary key. CERT holds only those values that can have
8840 meaningful defaults in an SSL_CTX.
8843 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8844 from the internal representation. Various PKCS#7 fixes: remove some
8845 evil casts and set the enc_dig_alg field properly based on the signing
8849 *) Allow PKCS#12 password to be set from the command line or the
8850 environment. Let 'ca' get its config file name from the environment
8851 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8855 *) Allow certificate policies extension to use an IA5STRING for the
8856 organization field. This is contrary to the PKIX definition but
8857 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8861 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8862 without disallowing inline assembler and the like for non-pedantic builds.
8865 *) Support Borland C++ builder.
8866 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8871 *) SHA-1 cleanups and performance enhancements.
8872 [Andy Polyakov <appro@fy.chalmers.se>]
8874 *) Sparc v8plus assembler for the bignum library.
8875 [Andy Polyakov <appro@fy.chalmers.se>]
8877 *) Accept any -xxx and +xxx compiler options in Configure.
8880 *) Update HPUX configuration.
8883 *) Add missing sk_<type>_unshift() function to safestack.h
8884 [Ralf S. Engelschall]
8886 *) New function SSL_CTX_use_certificate_chain_file that sets the
8887 "extra_cert"s in addition to the certificate. (This makes sense
8888 only for "PEM" format files, as chains as a whole are not
8892 *) Support verify_depth from the SSL API.
8893 x509_vfy.c had what can be considered an off-by-one-error:
8894 Its depth (which was not part of the external interface)
8895 was actually counting the number of certificates in a chain;
8896 now it really counts the depth.
8899 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8900 instead of X509err, which often resulted in confusing error
8901 messages since the error codes are not globally unique
8902 (e.g. an alleged error in ssl3_accept when a certificate
8903 didn't match the private key).
8905 *) New function SSL_CTX_set_session_id_context that allows to set a default
8906 value (so that you don't need SSL_set_session_id_context for each
8907 connection using the SSL_CTX).
8910 *) OAEP decoding bug fix.
8913 *) Support INSTALL_PREFIX for package builders, as proposed by
8917 *) New Configure options "threads" and "no-threads". For systems
8918 where the proper compiler options are known (currently Solaris
8919 and Linux), "threads" is the default.
8922 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8925 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8926 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8927 such as /usr/local/bin.
8930 *) "make linux-shared" to build shared libraries.
8931 [Niels Poppe <niels@netbox.org>]
8933 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8936 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8937 extension adding in x509 utility.
8940 *) Remove NOPROTO sections and error code comments.
8943 *) Partial rewrite of the DEF file generator to now parse the ANSI
8947 *) New Configure options --prefix=DIR and --openssldir=DIR.
8950 *) Complete rewrite of the error code script(s). It is all now handled
8951 by one script at the top level which handles error code gathering,
8952 header rewriting and C source file generation. It should be much better
8953 than the old method: it now uses a modified version of Ulf's parser to
8954 read the ANSI prototypes in all header files (thus the old K&R definitions
8955 aren't needed for error creation any more) and do a better job of
8956 translating function codes into names. The old 'ASN1 error code imbedded
8957 in a comment' is no longer necessary and it doesn't use .err files which
8958 have now been deleted. Also the error code call doesn't have to appear all
8959 on one line (which resulted in some large lines...).
8962 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8965 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8966 0 (which usually indicates a closed connection), but continue reading.
8969 *) Fix some race conditions.
8972 *) Add support for CRL distribution points extension. Add Certificate
8973 Policies and CRL distribution points documentation.
8976 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8979 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8980 8 of keying material. Merlin has also confirmed interop with this fix
8981 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8982 [Merlin Hughes <merlin@baltimore.ie>]
8984 *) Fix lots of warnings.
8985 [Richard Levitte <levitte@stacken.kth.se>]
8987 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8988 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8989 [Richard Levitte <levitte@stacken.kth.se>]
8991 *) Fix problems with sizeof(long) == 8.
8992 [Andy Polyakov <appro@fy.chalmers.se>]
8994 *) Change functions to ANSI C.
8997 *) Fix typos in error codes.
8998 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9000 *) Remove defunct assembler files from Configure.
9003 *) SPARC v8 assembler BIGNUM implementation.
9004 [Andy Polyakov <appro@fy.chalmers.se>]
9006 *) Support for Certificate Policies extension: both print and set.
9007 Various additions to support the r2i method this uses.
9010 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9011 return a const string when you are expecting an allocated buffer.
9014 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9015 types DirectoryString and DisplayText.
9018 *) Add code to allow r2i extensions to access the configuration database,
9019 add an LHASH database driver and add several ctx helper functions.
9022 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9023 fail when they extended the size of a BIGNUM.
9026 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9027 support typesafe stack.
9030 *) Fix typo in SSL_[gs]et_options().
9031 [Nils Frostberg <nils@medcom.se>]
9033 *) Delete various functions and files that belonged to the (now obsolete)
9034 old X509V3 handling code.
9037 *) New Configure option "rsaref".
9040 *) Don't auto-generate pem.h.
9043 *) Introduce type-safe ASN.1 SETs.
9046 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9047 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9049 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9050 that links with OpenSSL (well at least cause lots of warnings), but fear
9051 not: the conversion is trivial, and it eliminates loads of evil casts. A
9052 few STACKed things have been converted already. Feel free to convert more.
9053 In the fullness of time, I'll do away with the STACK type altogether.
9056 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9057 specified in <certfile> by updating the entry in the index.txt file.
9058 This way one no longer has to edit the index.txt file manually for
9059 revoking a certificate. The -revoke option does the gory details now.
9060 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9062 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9063 `-text' option at all and this way the `-noout -text' combination was
9064 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9065 [Ralf S. Engelschall]
9067 *) Make sure a corresponding plain text error message exists for the
9068 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9069 verify callback function determined that a certificate was revoked.
9070 [Ralf S. Engelschall]
9072 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9073 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9074 all available cipers including rc5, which was forgotten until now.
9075 In order to let the testing shell script know which algorithms
9076 are available, a new (up to now undocumented) command
9077 "openssl list-cipher-commands" is used.
9080 *) Bugfix: s_client occasionally would sleep in select() when
9081 it should have checked SSL_pending() first.
9084 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9085 the raw DSA values prior to ASN.1 encoding.
9088 *) Tweaks to Configure
9089 [Niels Poppe <niels@netbox.org>]
9091 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9095 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9098 *) New config option to avoid instructions that are illegal on the 80386.
9099 The default code is faster, but requires at least a 486.
9102 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9103 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9104 same as SSL2_VERSION anyway.
9107 *) New "-showcerts" option for s_client.
9110 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9111 application. Various cleanups and fixes.
9114 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9115 modify error routines to work internally. Add error codes and PBE init
9116 to library startup routines.
9119 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9120 packing functions to asn1 and evp. Changed function names and error
9121 codes along the way.
9124 *) PKCS12 integration: and so it begins... First of several patches to
9125 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9126 objects to objects.h
9129 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9130 and display support for Thawte strong extranet extension.
9133 *) Add LinuxPPC support.
9134 [Jeff Dubrule <igor@pobox.org>]
9136 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9137 bn_div_words in alpha.s.
9138 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9140 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9141 OAEP isn't supported when OpenSSL is built with RSAref.
9142 [Ulf Moeller <ulf@fitug.de>]
9144 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9145 so they no longer are missing under -DNOPROTO.
9146 [Soren S. Jorvang <soren@t.dk>]
9149 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9151 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9152 doesn't work when the session is reused. Coming soon!
9155 *) Fix a security hole, that allows sessions to be reused in the wrong
9156 context thus bypassing client cert protection! All software that uses
9157 client certs and session caches in multiple contexts NEEDS PATCHING to
9158 allow session reuse! A fuller solution is in the works.
9159 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9161 *) Some more source tree cleanups (removed obsolete files
9162 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9163 permission on "config" script to be executable) and a fix for the INSTALL
9165 [Ulf Moeller <ulf@fitug.de>]
9167 *) Remove some legacy and erroneous uses of malloc, free instead of
9169 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9171 *) Make rsa_oaep_test return non-zero on error.
9172 [Ulf Moeller <ulf@fitug.de>]
9174 *) Add support for native Solaris shared libraries. Configure
9175 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9176 if someone would make that last step automatic.
9177 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9179 *) ctx_size was not built with the right compiler during "make links". Fixed.
9182 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9183 except NULL ciphers". This means the default cipher list will no longer
9184 enable NULL ciphers. They need to be specifically enabled e.g. with
9185 the string "DEFAULT:eNULL".
9188 *) Fix to RSA private encryption routines: if p < q then it would
9189 occasionally produce an invalid result. This will only happen with
9190 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9193 *) Be less restrictive and allow also `perl util/perlpath.pl
9194 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9195 because this way one can also use an interpreter named `perl5' (which is
9196 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9197 installed as `perl').
9198 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9200 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9201 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9203 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9204 advapi32.lib to Win32 build and change the pem test comparision
9205 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9206 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9207 and crypto/des/ede_cbcm_enc.c.
9210 *) DES quad checksum was broken on big-endian architectures. Fixed.
9213 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9214 Win32 test batch file so it (might) work again. The Win32 test batch file
9215 is horrible: I feel ill....
9218 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9219 in e_os.h. Audit of header files to check ANSI and non ANSI
9220 sections: 10 functions were absent from non ANSI section and not exported
9221 from Windows DLLs. Fixed up libeay.num for new functions.
9224 *) Make `openssl version' output lines consistent.
9225 [Ralf S. Engelschall]
9227 *) Fix Win32 symbol export lists for BIO functions: Added
9228 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9229 to ms/libeay{16,32}.def.
9230 [Ralf S. Engelschall]
9232 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9233 fine under Unix and passes some trivial tests I've now added. But the
9234 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9235 added to make sure no one expects that this stuff really works in the
9236 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9237 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9239 [Ralf S. Engelschall]
9241 *) Fix the generation of two part addresses in perl.
9242 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9244 *) Add config entry for Linux on MIPS.
9245 [John Tobey <jtobey@channel1.com>]
9247 *) Make links whenever Configure is run, unless we are on Windoze.
9250 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9251 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9255 *) Add a useful kludge to allow package maintainers to specify compiler and
9256 other platforms details on the command line without having to patch the
9257 Configure script everytime: One now can use ``perl Configure
9258 <id>:<details>'', i.e. platform ids are allowed to have details appended
9259 to them (seperated by colons). This is treated as there would be a static
9260 pre-configured entry in Configure's %table under key <id> with value
9261 <details> and ``perl Configure <id>'' is called. So, when you want to
9262 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9263 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9264 now, which overrides the FreeBSD-elf entry on-the-fly.
9265 [Ralf S. Engelschall]
9267 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9270 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9271 on the `perl Configure ...' command line. This way one can compile
9272 OpenSSL libraries with Position Independent Code (PIC) which is needed
9273 for linking it into DSOs.
9274 [Ralf S. Engelschall]
9276 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9280 *) Cleaned up the LICENSE document: The official contact for any license
9281 questions now is the OpenSSL core team under openssl-core@openssl.org.
9282 And add a paragraph about the dual-license situation to make sure people
9283 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9284 to the OpenSSL toolkit.
9285 [Ralf S. Engelschall]
9287 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9288 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9289 Additonally cleaned up the `make links' target: Remove unnecessary
9290 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9291 to speed processing and no longer clutter the display with confusing
9292 stuff. Instead only the actually done links are displayed.
9293 [Ralf S. Engelschall]
9295 *) Permit null encryption ciphersuites, used for authentication only. It used
9296 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9297 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9301 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9302 signed attributes when verifying signatures (this would break them),
9303 the detached data encoding was wrong and public keys obtained using
9304 X509_get_pubkey() weren't freed.
9307 *) Add text documentation for the BUFFER functions. Also added a work around
9308 to a Win95 console bug. This was triggered by the password read stuff: the
9309 last character typed gets carried over to the next fread(). If you were
9310 generating a new cert request using 'req' for example then the last
9311 character of the passphrase would be CR which would then enter the first
9315 *) Added the new `Includes OpenSSL Cryptography Software' button as
9316 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9317 button and can be used by applications based on OpenSSL to show the
9318 relationship to the OpenSSL project.
9319 [Ralf S. Engelschall]
9321 *) Remove confusing variables in function signatures in files
9322 ssl/ssl_lib.c and ssl/ssl.h.
9323 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9325 *) Don't install bss_file.c under PREFIX/include/
9326 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9328 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9329 functions that return function pointers and has support for NT specific
9330 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9331 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9332 unsigned to signed types: this was killing the Win32 compile.
9335 *) Add new certificate file to stack functions,
9336 SSL_add_dir_cert_subjects_to_stack() and
9337 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9338 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9339 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9340 This means that Apache-SSL and similar packages don't have to mess around
9341 to add as many CAs as they want to the preferred list.
9344 *) Experiment with doxygen documentation. Currently only partially applied to
9346 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9347 openssl.doxy as the configuration file.
9350 *) Get rid of remaining C++-style comments which strict C compilers hate.
9351 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9353 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9354 compiled in by default: it has problems with large keys.
9357 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9358 DH private keys and/or callback functions which directly correspond to
9359 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9360 is needed for applications which have to configure certificates on a
9361 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9363 For the RSA certificate situation is makes no difference, but
9364 for the DSA certificate situation this fixes the "no shared cipher"
9365 problem where the OpenSSL cipher selection procedure failed because the
9366 temporary keys were not overtaken from the context and the API provided
9367 no way to reconfigure them.
9368 The new functions now let applications reconfigure the stuff and they
9369 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9370 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9371 non-public-API function ssl_cert_instantiate() is used as a helper
9372 function and also to reduce code redundancy inside ssl_rsa.c.
9373 [Ralf S. Engelschall]
9375 *) Move s_server -dcert and -dkey options out of the undocumented feature
9376 area because they are useful for the DSA situation and should be
9377 recognized by the users.
9378 [Ralf S. Engelschall]
9380 *) Fix the cipher decision scheme for export ciphers: the export bits are
9381 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9382 SSL_EXP_MASK. So, the original variable has to be used instead of the
9383 already masked variable.
9384 [Richard Levitte <levitte@stacken.kth.se>]
9386 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9387 [Richard Levitte <levitte@stacken.kth.se>]
9389 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9390 from `int' to `unsigned int' because it's a length and initialized by
9391 EVP_DigestFinal() which expects an `unsigned int *'.
9392 [Richard Levitte <levitte@stacken.kth.se>]
9394 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9395 script. Instead use the usual Shell->Perl transition trick.
9396 [Ralf S. Engelschall]
9398 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9399 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9400 -noout -modulus' as it's already the case for `openssl rsa -noout
9401 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9402 currently the public key is printed (a decision which was already done by
9403 `openssl dsa -modulus' in the past) which serves a similar purpose.
9404 Additionally the NO_RSA no longer completely removes the whole -modulus
9405 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9407 [Ralf S. Engelschall]
9409 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9410 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9411 [Arne Ansper <arne@ats.cyber.ee>]
9413 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9414 to be added. Now both 'req' and 'ca' can use new objects defined in the
9418 *) Add cool BIO that does syslog (or event log on NT).
9419 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9421 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9422 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9423 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9424 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9427 *) Add preliminary config info for new extension code.
9430 *) Make RSA_NO_PADDING really use no padding.
9431 [Ulf Moeller <ulf@fitug.de>]
9433 *) Generate errors when private/public key check is done.
9436 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9437 for some CRL extensions and new objects added.
9440 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9441 key usage extension and fuller support for authority key id.
9444 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9445 padding method for RSA, which is recommended for new applications in PKCS
9446 #1 v2.0 (RFC 2437, October 1998).
9447 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9448 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9449 against Bleichbacher's attack on RSA.
9450 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9453 *) Updates to the new SSL compression code
9454 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9456 *) Fix so that the version number in the master secret, when passed
9457 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9458 (because the server will not accept higher), that the version number
9459 is 0x03,0x01, not 0x03,0x00
9460 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9462 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9463 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9464 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9467 *) Support for RAW extensions where an arbitrary extension can be
9468 created by including its DER encoding. See apps/openssl.cnf for
9472 *) Make sure latest Perl versions don't interpret some generated C array
9473 code as Perl array code in the crypto/err/err_genc.pl script.
9474 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9476 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9477 not many people have the assembler. Various Win32 compilation fixes and
9478 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9482 *) Modify configure script 'Configure' to automatically create crypto/date.h
9483 file under Win32 and also build pem.h from pem.org. New script
9484 util/mkfiles.pl to create the MINFO file on environments that can't do a
9485 'make files': perl util/mkfiles.pl >MINFO should work.
9488 *) Major rework of DES function declarations, in the pursuit of correctness
9489 and purity. As a result, many evil casts evaporated, and some weirdness,
9490 too. You may find this causes warnings in your code. Zapping your evil
9491 casts will probably fix them. Mostly.
9494 *) Fix for a typo in asn1.h. Bug fix to object creation script
9495 obj_dat.pl. It considered a zero in an object definition to mean
9496 "end of object": none of the objects in objects.h have any zeros
9497 so it wasn't spotted.
9498 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9500 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9501 Masking (CBCM). In the absence of test vectors, the best I have been able
9502 to do is check that the decrypt undoes the encrypt, so far. Send me test
9503 vectors if you have them.
9506 *) Correct calculation of key length for export ciphers (too much space was
9507 allocated for null ciphers). This has not been tested!
9510 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9511 message is now correct (it understands "crypto" and "ssl" on its
9512 command line). There is also now an "update" option. This will update
9513 the util/ssleay.num and util/libeay.num files with any new functions.
9515 perl util/mkdef.pl crypto ssl update
9516 it will update them.
9519 *) Overhauled the Perl interface (perl/*):
9520 - ported BN stuff to OpenSSL's different BN library
9521 - made the perl/ source tree CVS-aware
9522 - renamed the package from SSLeay to OpenSSL (the files still contain
9523 their history because I've copied them in the repository)
9524 - removed obsolete files (the test scripts will be replaced
9525 by better Test::Harness variants in the future)
9526 [Ralf S. Engelschall]
9528 *) First cut for a very conservative source tree cleanup:
9529 1. merge various obsolete readme texts into doc/ssleay.txt
9530 where we collect the old documents and readme texts.
9531 2. remove the first part of files where I'm already sure that we no
9532 longer need them because of three reasons: either they are just temporary
9533 files which were left by Eric or they are preserved original files where
9534 I've verified that the diff is also available in the CVS via "cvs diff
9535 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9536 the crypto/md/ stuff).
9537 [Ralf S. Engelschall]
9539 *) More extension code. Incomplete support for subject and issuer alt
9540 name, issuer and authority key id. Change the i2v function parameters
9541 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9542 what that's for :-) Fix to ASN1 macro which messed up
9543 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9546 *) Preliminary support for ENUMERATED type. This is largely copied from the
9550 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9551 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9553 *) Make sure `make rehash' target really finds the `openssl' program.
9554 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9556 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9557 like to hear about it if this slows down other processors.
9560 *) Add CygWin32 platform information to Configure script.
9561 [Alan Batie <batie@aahz.jf.intel.com>]
9563 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9564 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9566 *) New program nseq to manipulate netscape certificate sequences
9569 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9573 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9574 but the BN code had some problems that would cause failures when
9575 doing certificate verification and some other functions.
9576 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9578 *) Add ASN1 and PEM code to support netscape certificate sequences.
9581 *) Add ASN1 and PEM code to support netscape certificate sequences.
9584 *) Add several PKIX and private extended key usage OIDs.
9587 *) Modify the 'ca' program to handle the new extension code. Modify
9588 openssl.cnf for new extension format, add comments.
9591 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9592 and add a sample to openssl.cnf so req -x509 now adds appropriate
9596 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9597 error code, add initial support to X509_print() and x509 application.
9600 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9601 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9602 stuff is currently isolated and isn't even compiled yet.
9605 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9606 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9607 Removed the versions check from X509 routines when loading extensions:
9608 this allows certain broken certificates that don't set the version
9609 properly to be processed.
9612 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9613 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9614 can still be regenerated with "make depend".
9617 *) Spelling mistake in C version of CAST-128.
9618 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9620 *) Changes to the error generation code. The perl script err-code.pl
9621 now reads in the old error codes and retains the old numbers, only
9622 adding new ones if necessary. It also only changes the .err files if new
9623 codes are added. The makefiles have been modified to only insert errors
9624 when needed (to avoid needlessly modifying header files). This is done
9625 by only inserting errors if the .err file is newer than the auto generated
9626 C file. To rebuild all the error codes from scratch (the old behaviour)
9627 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9628 or delete all the .err files.
9631 *) CAST-128 was incorrectly implemented for short keys. The C version has
9632 been fixed, but is untested. The assembler versions are also fixed, but
9633 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9634 to regenerate it if needed.
9635 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9636 Hagino <itojun@kame.net>]
9638 *) File was opened incorrectly in randfile.c.
9639 [Ulf Möller <ulf@fitug.de>]
9641 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9642 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9643 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9644 al: it's just almost always a UTCTime. Note this patch adds new error
9645 codes so do a "make errors" if there are problems.
9648 *) Correct Linux 1 recognition in config.
9649 [Ulf Möller <ulf@fitug.de>]
9651 *) Remove pointless MD5 hash when using DSA keys in ca.
9652 [Anonymous <nobody@replay.com>]
9654 *) Generate an error if given an empty string as a cert directory. Also
9655 generate an error if handed NULL (previously returned 0 to indicate an
9656 error, but didn't set one).
9657 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9659 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9662 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9663 parameters. This was causing a warning which killed off the Win32 compile.
9666 *) Remove C++ style comments from crypto/bn/bn_local.h.
9667 [Neil Costigan <neil.costigan@celocom.com>]
9669 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9670 based on a text string, looking up short and long names and finally
9671 "dot" format. The "dot" format stuff didn't work. Added new function
9672 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9673 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9674 OID is not part of the table.
9677 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9678 X509_LOOKUP_by_alias().
9681 *) Sort openssl functions by name.
9684 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9685 encryption from sample DSA keys (in case anyone is interested the password
9689 *) Make _all_ *_free functions accept a NULL pointer.
9690 [Frans Heymans <fheymans@isaserver.be>]
9692 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9694 [Anonymous <nobody@replay.com>]
9696 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9697 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9699 *) Don't blow it for numeric -newkey arguments to apps/req.
9700 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9702 *) Temp key "for export" tests were wrong in s3_srvr.c.
9703 [Anonymous <nobody@replay.com>]
9705 *) Add prototype for temp key callback functions
9706 SSL_CTX_set_tmp_{rsa,dh}_callback().
9709 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9710 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9713 *) X509_name_add_entry() freed the wrong thing after an error.
9714 [Arne Ansper <arne@ats.cyber.ee>]
9716 *) rsa_eay.c would attempt to free a NULL context.
9717 [Arne Ansper <arne@ats.cyber.ee>]
9719 *) BIO_s_socket() had a broken should_retry() on Windoze.
9720 [Arne Ansper <arne@ats.cyber.ee>]
9722 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9723 [Arne Ansper <arne@ats.cyber.ee>]
9725 *) Make sure the already existing X509_STORE->depth variable is initialized
9726 in X509_STORE_new(), but document the fact that this variable is still
9727 unused in the certificate verification process.
9728 [Ralf S. Engelschall]
9730 *) Fix the various library and apps files to free up pkeys obtained from
9731 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9734 *) Fix reference counting in X509_PUBKEY_get(). This makes
9735 demos/maurice/example2.c work, amongst others, probably.
9736 [Steve Henson and Ben Laurie]
9738 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9739 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9740 are no longer created. This way we have a single and consistent command
9741 line interface `openssl <command>', similar to `cvs <command>'.
9742 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9744 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9745 BIT STRING wrapper always have zero unused bits.
9748 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9751 *) Make the top-level INSTALL documentation easier to understand.
9754 *) Makefiles updated to exit if an error occurs in a sub-directory
9755 make (including if user presses ^C) [Paul Sutton]
9757 *) Make Montgomery context stuff explicit in RSA data structure.
9760 *) Fix build order of pem and err to allow for generated pem.h.
9763 *) Fix renumbering bug in X509_NAME_delete_entry().
9766 *) Enhanced the err-ins.pl script so it makes the error library number
9767 global and can add a library name. This is needed for external ASN1 and
9768 other error libraries.
9771 *) Fixed sk_insert which never worked properly.
9774 *) Fix ASN1 macros so they can handle indefinite length construted
9775 EXPLICIT tags. Some non standard certificates use these: they can now
9779 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9780 into a single doc/ssleay.txt bundle. This way the information is still
9781 preserved but no longer messes up this directory. Now it's new room for
9782 the new set of documenation files.
9783 [Ralf S. Engelschall]
9785 *) SETs were incorrectly DER encoded. This was a major pain, because they
9786 shared code with SEQUENCEs, which aren't coded the same. This means that
9787 almost everything to do with SETs or SEQUENCEs has either changed name or
9788 number of arguments.
9789 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9791 *) Fix test data to work with the above.
9794 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9795 was already fixed by Eric for 0.9.1 it seems.
9796 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9798 *) Autodetect FreeBSD3.
9801 *) Fix various bugs in Configure. This affects the following platforms:
9805 unixware-2.0-pentium
9809 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9810 before they are needed.
9813 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9817 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9819 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9820 changed SSLeay to OpenSSL in version strings.
9821 [Ralf S. Engelschall]
9823 *) Some fixups to the top-level documents.
9826 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9827 because the symlink to include/ was missing.
9828 [Ralf S. Engelschall]
9830 *) Incorporated the popular no-RSA/DSA-only patches
9831 which allow to compile a RSA-free SSLeay.
9832 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9834 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9835 when "ssleay" is still not found.
9836 [Ralf S. Engelschall]
9838 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9839 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9841 *) Updated the README file.
9842 [Ralf S. Engelschall]
9844 *) Added various .cvsignore files in the CVS repository subdirs
9845 to make a "cvs update" really silent.
9846 [Ralf S. Engelschall]
9848 *) Recompiled the error-definition header files and added
9849 missing symbols to the Win32 linker tables.
9850 [Ralf S. Engelschall]
9852 *) Cleaned up the top-level documents;
9853 o new files: CHANGES and LICENSE
9854 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9855 o merged COPYRIGHT into LICENSE
9856 o removed obsolete TODO file
9857 o renamed MICROSOFT to INSTALL.W32
9858 [Ralf S. Engelschall]
9860 *) Removed dummy files from the 0.9.1b source tree:
9861 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9862 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9863 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9864 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9865 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9866 [Ralf S. Engelschall]
9868 *) Added various platform portability fixes.
9871 *) The Genesis of the OpenSSL rpject:
9872 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9873 Young and Tim J. Hudson created while they were working for C2Net until
9875 [The OpenSSL Project]
9878 Changes between 0.9.0b and 0.9.1b [not released]
9880 *) Updated a few CA certificates under certs/
9883 *) Changed some BIGNUM api stuff.
9886 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9887 DGUX x86, Linux Alpha, etc.
9890 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9891 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9895 *) Add -strparse option to asn1pars program which parses nested
9897 [Dr Stephen Henson <shenson@bigfoot.com>]
9899 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9902 *) DSA fix for "ca" program.
9905 *) Added "-genkey" option to "dsaparam" program.
9908 *) Added RIPE MD160 (rmd160) message digest.
9911 *) Added -a (all) option to "ssleay version" command.
9914 *) Added PLATFORM define which is the id given to Configure.
9917 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9920 *) Extended the ASN.1 parser routines.
9923 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9926 *) Added a BN_CTX to the BN library.
9929 *) Fixed the weak key values in DES library
9932 *) Changed API in EVP library for cipher aliases.
9935 *) Added support for RC2/64bit cipher.
9938 *) Converted the lhash library to the crypto/mem.c functions.
9941 *) Added more recognized ASN.1 object ids.
9944 *) Added more RSA padding checks for SSL/TLS.
9947 *) Added BIO proxy/filter functionality.
9950 *) Added extra_certs to SSL_CTX which can be used
9951 send extra CA certificates to the client in the CA cert chain sending
9952 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9955 *) Now Fortezza is denied in the authentication phase because
9956 this is key exchange mechanism is not supported by SSLeay at all.
9959 *) Additional PKCS1 checks.
9962 *) Support the string "TLSv1" for all TLS v1 ciphers.
9965 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9966 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9969 *) Fixed a few memory leaks.
9972 *) Fixed various code and comment typos.
9975 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9976 bytes sent in the client random.
9977 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob