5 Changes between 1.0.1 and 1.0.2 [xx XXX xxxx]
7 *) Enhance SSL/TLS certificate chain handling to support different
8 chains for each certificate instead of one chain in the parent SSL_CTX.
11 *) Support for fixed DH ciphersuite client authentication: where both
12 server and client use DH certificates with common parameters.
15 *) Support for fixed DH ciphersuites: those requiring DH server
19 Changes between 1.0.1 and 1.0.1a [xx XXX xxxx]
21 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
24 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
26 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
27 STRING form instead of a DigestInfo.
30 *) The format used for MDC2 RSA signatures is inconsistent between EVP
31 and the RSA_sign/RSA_verify functions. This was made more apparent when
32 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
33 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
34 the correct format in RSA_verify so both forms transparently work.
37 *) Some servers which support TLS 1.0 can choke if we initially indicate
38 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
39 encrypted premaster secret. As a workaround use the maximum pemitted
40 client version in client hello, this should keep such servers happy
41 and still work with previous versions of OpenSSL.
44 *) Add support for TLS/DTLS heartbeats.
45 [Robin Seggelmann <seggelmann@fh-muenster.de>]
47 *) Add support for SCTP.
48 [Robin Seggelmann <seggelmann@fh-muenster.de>]
50 *) Improved PRNG seeding for VOS.
51 [Paul Green <Paul.Green@stratus.com>]
53 *) Extensive assembler packs updates, most notably:
55 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
56 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
57 - x86_64: bit-sliced AES implementation;
58 - ARM: NEON support, contemporary platforms optimizations;
59 - s390x: z196 support;
60 - *: GHASH and GF(2^m) multiplication implementations;
64 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
65 (removal of unnecessary code)
66 [Peter Sylvester <peter.sylvester@edelweb.fr>]
68 *) Add TLS key material exporter from RFC 5705.
71 *) Add DTLS-SRTP negotiation from RFC 5764.
74 *) Add Next Protocol Negotiation,
75 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
76 disabled with a no-npn flag to config or Configure. Code donated
78 [Adam Langley <agl@google.com> and Ben Laurie]
80 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
81 NIST-P256, NIST-P521, with constant-time single point multiplication on
82 typical inputs. Compiler support for the nonstandard type __uint128_t is
83 required to use this (present in gcc 4.4 and later, for 64-bit builds).
84 Code made available under Apache License version 2.0.
86 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
87 line to include this in your build of OpenSSL, and run "make depend" (or
88 "make update"). This enables the following EC_METHODs:
90 EC_GFp_nistp224_method()
91 EC_GFp_nistp256_method()
92 EC_GFp_nistp521_method()
94 EC_GROUP_new_by_curve_name() will automatically use these (while
95 EC_GROUP_new_curve_GFp() currently prefers the more flexible
97 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
99 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
100 all platforms. Move ssize_t definition from e_os.h to the public
101 header file e_os2.h as it now appears in public header file cms.h
104 *) New -sigopt option to the ca, req and x509 utilities. Additional
105 signature parameters can be passed using this option and in
109 *) Add RSA PSS signing function. This will generate and set the
110 appropriate AlgorithmIdentifiers for PSS based on those in the
111 corresponding EVP_MD_CTX structure. No application support yet.
114 *) Support for companion algorithm specific ASN1 signing routines.
115 New function ASN1_item_sign_ctx() signs a pre-initialised
116 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
117 the appropriate parameters.
120 *) Add new algorithm specific ASN1 verification initialisation function
121 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
122 handling will be the same no matter what EVP_PKEY_METHOD is used.
123 Add a PSS handler to support verification of PSS signatures: checked
124 against a number of sample certificates.
127 *) Add signature printing for PSS. Add PSS OIDs.
128 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
130 *) Add algorithm specific signature printing. An individual ASN1 method
131 can now print out signatures instead of the standard hex dump.
133 More complex signatures (e.g. PSS) can print out more meaningful
134 information. Include DSA version that prints out the signature
138 *) Password based recipient info support for CMS library: implementing
142 *) Split password based encryption into PBES2 and PBKDF2 functions. This
143 neatly separates the code into cipher and PBE sections and is required
144 for some algorithms that split PBES2 into separate pieces (such as
148 *) Session-handling fixes:
149 - Fix handling of connections that are resuming with a session ID,
150 but also support Session Tickets.
151 - Fix a bug that suppressed issuing of a new ticket if the client
152 presented a ticket with an expired session.
153 - Try to set the ticket lifetime hint to something reasonable.
154 - Make tickets shorter by excluding irrelevant information.
155 - On the client side, don't ignore renewed tickets.
156 [Adam Langley, Bodo Moeller (Google)]
158 *) Fix PSK session representation.
161 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
163 This work was sponsored by Intel.
166 *) Add GCM support to TLS library. Some custom code is needed to split
167 the IV between the fixed (from PRF) and explicit (from TLS record)
168 portions. This adds all GCM ciphersuites supported by RFC5288 and
169 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
170 add a special AESGCM string for GCM only.
173 *) Expand range of ctrls for AES GCM. Permit setting invocation
174 field on decrypt and retrieval of invocation field only on encrypt.
177 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
178 As required by RFC5289 these ciphersuites cannot be used if for
179 versions of TLS earlier than 1.2.
182 *) For FIPS capable OpenSSL interpret a NULL default public key method
183 as unset and return the appopriate default but do *not* set the default.
184 This means we can return the appopriate method in applications that
185 swicth between FIPS and non-FIPS modes.
188 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
189 ENGINE is used then we cannot handle that in the FIPS module so we
190 keep original code iff non-FIPS operations are allowed.
193 *) Add -attime option to openssl utilities.
194 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
196 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
199 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
200 FIPS EC methods unconditionally for now.
203 *) New build option no-ec2m to disable characteristic 2 code.
206 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
207 all cases can be covered as some introduce binary incompatibilities.
210 *) Redirect RSA operations to FIPS module including keygen,
211 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
214 *) Add similar low level API blocking to ciphers.
217 *) Low level digest APIs are not approved in FIPS mode: any attempt
218 to use these will cause a fatal error. Applications that *really* want
219 to use them can use the private_* version instead.
222 *) Redirect cipher operations to FIPS module for FIPS builds.
225 *) Redirect digest operations to FIPS module for FIPS builds.
228 *) Update build system to add "fips" flag which will link in fipscanister.o
229 for static and shared library builds embedding a signature if needed.
232 *) Output TLS supported curves in preference order instead of numerical
233 order. This is currently hardcoded for the highest order curves first.
234 This should be configurable so applications can judge speed vs strength.
237 *) Add TLS v1.2 server support for client authentication.
240 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
244 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
245 FIPS modules versions.
248 *) Add TLS v1.2 client side support for client authentication. Keep cache
249 of handshake records longer as we don't know the hash algorithm to use
250 until after the certificate request message is received.
253 *) Initial TLS v1.2 client support. Add a default signature algorithms
254 extension including all the algorithms we support. Parse new signature
255 format in client key exchange. Relax some ECC signing restrictions for
256 TLS v1.2 as indicated in RFC5246.
259 *) Add server support for TLS v1.2 signature algorithms extension. Switch
260 to new signature format when needed using client digest preference.
261 All server ciphersuites should now work correctly in TLS v1.2. No client
262 support yet and no support for client certificates.
265 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
266 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
267 ciphersuites. At present only RSA key exchange ciphersuites work with
268 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
269 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
270 and version checking.
273 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
274 with this defined it will not be affected by any changes to ssl internal
275 structures. Add several utility functions to allow openssl application
276 to work with OPENSSL_NO_SSL_INTERN defined.
280 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
282 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
285 *) Permit abbreviated handshakes when renegotiating using the function
286 SSL_renegotiate_abbreviated().
287 [Robin Seggelmann <seggelmann@fh-muenster.de>]
289 *) Add call to ENGINE_register_all_complete() to
290 ENGINE_load_builtin_engines(), so some implementations get used
291 automatically instead of needing explicit application support.
294 *) Add support for TLS key exporter as described in RFC5705.
295 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
297 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
298 a few changes are required:
300 Add SSL_OP_NO_TLSv1_1 flag.
302 Update version checking logic to handle version 1.1.
303 Add explicit IV handling (ported from DTLS code).
304 Add command line options to s_client/s_server.
307 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
309 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
310 in CMS and PKCS7 code. When RSA decryption fails use a random key for
311 content decryption and always return the same error. Note: this attack
312 needs on average 2^20 messages so it only affects automated senders. The
313 old behaviour can be reenabled in the CMS code by setting the
314 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
315 an MMA defence is not necessary.
316 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
317 this issue. (CVE-2012-0884)
320 *) Fix CVE-2011-4619: make sure we really are receiving a
321 client hello before rejecting multiple SGC restarts. Thanks to
322 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
325 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
327 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
328 Thanks to Antonio Martin, Enterprise Secure Access Research and
329 Development, Cisco Systems, Inc. for discovering this bug and
330 preparing a fix. (CVE-2012-0050)
333 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
335 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
336 of the Vaudenay padding oracle attack on CBC mode encryption
337 which enables an efficient plaintext recovery attack against
338 the OpenSSL implementation of DTLS. Their attack exploits timing
339 differences arising during decryption processing. A research
340 paper describing this attack can be found at:
341 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
342 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
343 Security Group at Royal Holloway, University of London
344 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
345 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
346 for preparing the fix. (CVE-2011-4108)
347 [Robin Seggelmann, Michael Tuexen]
349 *) Clear bytes used for block padding of SSL 3.0 records.
351 [Adam Langley (Google)]
353 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
354 Kadianakis <desnacked@gmail.com> for discovering this issue and
355 Adam Langley for preparing the fix. (CVE-2011-4619)
356 [Adam Langley (Google)]
358 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
359 [Andrey Kulikov <amdeich@gmail.com>]
361 *) Prevent malformed RFC3779 data triggering an assertion failure.
362 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
363 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
364 [Rob Austein <sra@hactrn.net>]
366 *) Improved PRNG seeding for VOS.
367 [Paul Green <Paul.Green@stratus.com>]
369 *) Fix ssl_ciph.c set-up race.
370 [Adam Langley (Google)]
372 *) Fix spurious failures in ecdsatest.c.
373 [Emilia Käsper (Google)]
375 *) Fix the BIO_f_buffer() implementation (which was mixing different
376 interpretations of the '..._len' fields).
377 [Adam Langley (Google)]
379 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
380 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
381 threads won't reuse the same blinding coefficients.
383 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
384 lock to call BN_BLINDING_invert_ex, and avoids one use of
385 BN_BLINDING_update for each BN_BLINDING structure (previously,
386 the last update always remained unused).
387 [Emilia Käsper (Google)]
389 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
390 [Bob Buckholz (Google)]
392 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
394 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
395 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
396 [Kaspar Brand <ossl@velox.ch>]
398 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
399 for multi-threaded use of ECDH. (CVE-2011-3210)
400 [Adam Langley (Google)]
402 *) Fix x509_name_ex_d2i memory leak on bad inputs.
405 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
406 signature public key algorithm by using OID xref utilities instead.
407 Before this you could only use some ECC ciphersuites with SHA1 only.
410 *) Add protection against ECDSA timing attacks as mentioned in the paper
411 by Billy Bob Brumley and Nicola Tuveri, see:
413 http://eprint.iacr.org/2011/232.pdf
415 [Billy Bob Brumley and Nicola Tuveri]
417 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
419 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
420 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
422 *) Fix bug in string printing code: if *any* escaping is enabled we must
423 escape the escape character (backslash) or the resulting string is
427 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
429 *) Disable code workaround for ancient and obsolete Netscape browsers
430 and servers: an attacker can use it in a ciphersuite downgrade attack.
431 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
434 *) Fixed J-PAKE implementation error, originally discovered by
435 Sebastien Martini, further info and confirmation from Stefan
436 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
439 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
441 *) Fix extension code to avoid race conditions which can result in a buffer
442 overrun vulnerability: resumed sessions must not be modified as they can
443 be shared by multiple threads. CVE-2010-3864
446 *) Fix WIN32 build system to correctly link an ENGINE directory into
450 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
452 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
454 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
456 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
458 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
459 context. The operation can be customised via the ctrl mechanism in
460 case ENGINEs want to include additional functionality.
463 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
466 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
467 output hashes compatible with older versions of OpenSSL.
468 [Willy Weisz <weisz@vcpc.univie.ac.at>]
470 *) Fix compression algorithm handling: if resuming a session use the
471 compression algorithm of the resumed session instead of determining
472 it from client hello again. Don't allow server to change algorithm.
475 *) Add load_crls() function to apps tidying load_certs() too. Add option
476 to verify utility to allow additional CRLs to be included.
479 *) Update OCSP request code to permit adding custom headers to the request:
480 some responders need this.
483 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
485 [Julia Lawall <julia@diku.dk>]
487 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
488 needlessly dereferenced structures, used obsolete functions and
489 didn't handle all updated verify codes correctly.
492 *) Disable MD2 in the default configuration.
495 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
496 indicate the initial BIO being pushed or popped. This makes it possible
497 to determine whether the BIO is the one explicitly called or as a result
498 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
499 it handles reference counts correctly and doesn't zero out the I/O bio
500 when it is not being explicitly popped. WARNING: applications which
501 included workarounds for the old buggy behaviour will need to be modified
502 or they could free up already freed BIOs.
505 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
506 renaming to all platforms (within the 0.9.8 branch, this was
507 done conditionally on Netware platforms to avoid a name clash).
508 [Guenter <lists@gknw.net>]
510 *) Add ECDHE and PSK support to DTLS.
511 [Michael Tuexen <tuexen@fh-muenster.de>]
513 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
517 *) Add "missing" function EVP_MD_flags() (without this the only way to
518 retrieve a digest flags is by accessing the structure directly. Update
519 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
520 or cipher is registered as in the "from" argument. Print out all
521 registered digests in the dgst usage message instead of manually
522 attempting to work them out.
525 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
526 this allows the use of compression and extensions. Change default cipher
527 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
528 by default unless an application cipher string requests it.
531 *) Alter match criteria in PKCS12_parse(). It used to try to use local
532 key ids to find matching certificates and keys but some PKCS#12 files
533 don't follow the (somewhat unwritten) rules and this strategy fails.
534 Now just gather all certificates together and the first private key
535 then look for the first certificate that matches the key.
538 *) Support use of registered digest and cipher names for dgst and cipher
539 commands instead of having to add each one as a special case. So now
546 openssl dgst -sha256 foo
548 and this works for ENGINE based algorithms too.
552 *) Update Gost ENGINE to support parameter files.
553 [Victor B. Wagner <vitus@cryptocom.ru>]
555 *) Support GeneralizedTime in ca utility.
556 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
558 *) Enhance the hash format used for certificate directory links. The new
559 form uses the canonical encoding (meaning equivalent names will work
560 even if they aren't identical) and uses SHA1 instead of MD5. This form
561 is incompatible with the older format and as a result c_rehash should
562 be used to rebuild symbolic links.
565 *) Make PKCS#8 the default write format for private keys, replacing the
566 traditional format. This form is standardised, more secure and doesn't
567 include an implicit MD5 dependency.
570 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
571 committed to OpenSSL should pass this lot as a minimum.
574 *) Add session ticket override functionality for use by EAP-FAST.
575 [Jouni Malinen <j@w1.fi>]
577 *) Modify HMAC functions to return a value. Since these can be implemented
578 in an ENGINE errors can occur.
581 *) Type-checked OBJ_bsearch_ex.
584 *) Type-checked OBJ_bsearch. Also some constification necessitated
585 by type-checking. Still to come: TXT_DB, bsearch(?),
586 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
590 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
591 seconds to a tm structure directly, instead of going through OS
592 specific date routines. This avoids any issues with OS routines such
593 as the year 2038 bug. New *_adj() functions for ASN1 time structures
594 and X509_time_adj_ex() to cover the extended range. The existing
595 X509_time_adj() is still usable and will no longer have any date issues.
598 *) Delta CRL support. New use deltas option which will attempt to locate
599 and search any appropriate delta CRLs available.
601 This work was sponsored by Google.
604 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
605 code and add additional score elements. Validate alternate CRL paths
606 as part of the CRL checking and indicate a new error "CRL path validation
607 error" in this case. Applications wanting additional details can use
608 the verify callback and check the new "parent" field. If this is not
609 NULL CRL path validation is taking place. Existing applications wont
610 see this because it requires extended CRL support which is off by
613 This work was sponsored by Google.
616 *) Support for freshest CRL extension.
618 This work was sponsored by Google.
621 *) Initial indirect CRL support. Currently only supported in the CRLs
622 passed directly and not via lookup. Process certificate issuer
623 CRL entry extension and lookup CRL entries by bother issuer name
624 and serial number. Check and process CRL issuer entry in IDP extension.
626 This work was sponsored by Google.
629 *) Add support for distinct certificate and CRL paths. The CRL issuer
630 certificate is validated separately in this case. Only enabled if
631 an extended CRL support flag is set: this flag will enable additional
632 CRL functionality in future.
634 This work was sponsored by Google.
637 *) Add support for policy mappings extension.
639 This work was sponsored by Google.
642 *) Fixes to pathlength constraint, self issued certificate handling,
643 policy processing to align with RFC3280 and PKITS tests.
645 This work was sponsored by Google.
648 *) Support for name constraints certificate extension. DN, email, DNS
649 and URI types are currently supported.
651 This work was sponsored by Google.
654 *) To cater for systems that provide a pointer-based thread ID rather
655 than numeric, deprecate the current numeric thread ID mechanism and
656 replace it with a structure and associated callback type. This
657 mechanism allows a numeric "hash" to be extracted from a thread ID in
658 either case, and on platforms where pointers are larger than 'long',
659 mixing is done to help ensure the numeric 'hash' is usable even if it
660 can't be guaranteed unique. The default mechanism is to use "&errno"
661 as a pointer-based thread ID to distinguish between threads.
663 Applications that want to provide their own thread IDs should now use
664 CRYPTO_THREADID_set_callback() to register a callback that will call
665 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
667 Note that ERR_remove_state() is now deprecated, because it is tied
668 to the assumption that thread IDs are numeric. ERR_remove_state(0)
669 to free the current thread's error state should be replaced by
670 ERR_remove_thread_state(NULL).
672 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
673 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
674 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
675 application was previously providing a numeric thread callback that
676 was inappropriate for distinguishing threads, then uniqueness might
677 have been obtained with &errno that happened immediately in the
678 intermediate development versions of OpenSSL; this is no longer the
679 case, the numeric thread callback will now override the automatic use
681 [Geoff Thorpe, with help from Bodo Moeller]
683 *) Initial support for different CRL issuing certificates. This covers a
684 simple case where the self issued certificates in the chain exist and
685 the real CRL issuer is higher in the existing chain.
687 This work was sponsored by Google.
690 *) Removed effectively defunct crypto/store from the build.
693 *) Revamp of STACK to provide stronger type-checking. Still to come:
694 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
695 ASN1_STRING, CONF_VALUE.
698 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
699 RAM on SSL connections. This option can save about 34k per idle SSL.
702 *) Revamp of LHASH to provide stronger type-checking. Still to come:
703 STACK, TXT_DB, bsearch, qsort.
706 *) Initial support for Cryptographic Message Syntax (aka CMS) based
707 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
708 support for data, signedData, compressedData, digestedData and
709 encryptedData, envelopedData types included. Scripts to check against
710 RFC4134 examples draft and interop and consistency checks of many
711 content types and variants.
714 *) Add options to enc utility to support use of zlib compression BIO.
717 *) Extend mk1mf to support importing of options and assembly language
718 files from Configure script, currently only included in VC-WIN32.
719 The assembly language rules can now optionally generate the source
720 files from the associated perl scripts.
723 *) Implement remaining functionality needed to support GOST ciphersuites.
724 Interop testing has been performed using CryptoPro implementations.
725 [Victor B. Wagner <vitus@cryptocom.ru>]
727 *) s390x assembler pack.
730 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
734 *) Implement Opaque PRF Input TLS extension as specified in
735 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
736 official specification yet and no extension type assignment by
737 IANA exists, this extension (for now) will have to be explicitly
738 enabled when building OpenSSL by providing the extension number
739 to use. For example, specify an option
741 -DTLSEXT_TYPE_opaque_prf_input=0x9527
743 to the "config" or "Configure" script to enable the extension,
744 assuming extension number 0x9527 (which is a completely arbitrary
745 and unofficial assignment based on the MD5 hash of the Internet
746 Draft). Note that by doing so, you potentially lose
747 interoperability with other TLS implementations since these might
748 be using the same extension number for other purposes.
750 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
751 opaque PRF input value to use in the handshake. This will create
752 an interal copy of the length-'len' string at 'src', and will
753 return non-zero for success.
755 To get more control and flexibility, provide a callback function
758 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
759 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
763 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
766 Callback function 'cb' will be called in handshakes, and is
767 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
768 Argument 'arg' is for application purposes (the value as given to
769 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
770 be provided to the callback function). The callback function
771 has to return non-zero to report success: usually 1 to use opaque
772 PRF input just if possible, or 2 to enforce use of the opaque PRF
773 input. In the latter case, the library will abort the handshake
774 if opaque PRF input is not successfully negotiated.
776 Arguments 'peerinput' and 'len' given to the callback function
777 will always be NULL and 0 in the case of a client. A server will
778 see the client's opaque PRF input through these variables if
779 available (NULL and 0 otherwise). Note that if the server
780 provides an opaque PRF input, the length must be the same as the
781 length of the client's opaque PRF input.
783 Note that the callback function will only be called when creating
784 a new session (session resumption can resume whatever was
785 previously negotiated), and will not be called in SSL 2.0
786 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
787 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
788 for applications that need to enforce opaque PRF input.
792 *) Update ssl code to support digests other than SHA1+MD5 for handshake
795 [Victor B. Wagner <vitus@cryptocom.ru>]
797 *) Add RFC4507 support to OpenSSL. This includes the corrections in
798 RFC4507bis. The encrypted ticket format is an encrypted encoded
799 SSL_SESSION structure, that way new session features are automatically
802 If a client application caches session in an SSL_SESSION structure
803 support is transparent because tickets are now stored in the encoded
806 The SSL_CTX structure automatically generates keys for ticket
807 protection in servers so again support should be possible
808 with no application modification.
810 If a client or server wishes to disable RFC4507 support then the option
811 SSL_OP_NO_TICKET can be set.
813 Add a TLS extension debugging callback to allow the contents of any client
814 or server extensions to be examined.
816 This work was sponsored by Google.
819 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
820 OpenSSL should now compile cleanly on gcc 4.2
821 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
823 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
824 support including streaming MAC support: this is required for GOST
826 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
828 *) Add option -stream to use PKCS#7 streaming in smime utility. New
829 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
830 to output in BER and PEM format.
833 *) Experimental support for use of HMAC via EVP_PKEY interface. This
834 allows HMAC to be handled via the EVP_DigestSign*() interface. The
835 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
836 ENGINE support for HMAC keys which are unextractable. New -mac and
837 -macopt options to dgst utility.
840 *) New option -sigopt to dgst utility. Update dgst to use
841 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
842 alternative signing paramaters such as X9.31 or PSS in the dgst
846 *) Change ssl_cipher_apply_rule(), the internal function that does
847 the work each time a ciphersuite string requests enabling
848 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
849 removing ("!foo+bar") a class of ciphersuites: Now it maintains
850 the order of disabled ciphersuites such that those ciphersuites
851 that most recently went from enabled to disabled not only stay
852 in order with respect to each other, but also have higher priority
853 than other disabled ciphersuites the next time ciphersuites are
856 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
857 the same ciphersuites as with "HIGH" alone, but in a specific
858 order where the PSK ciphersuites come first (since they are the
859 most recently disabled ciphersuites when "HIGH" is parsed).
861 Also, change ssl_create_cipher_list() (using this new
862 funcionality) such that between otherwise identical
863 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
867 *) Change ssl_create_cipher_list() so that it automatically
868 arranges the ciphersuites in reasonable order before starting
869 to process the rule string. Thus, the definition for "DEFAULT"
870 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
871 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
872 This makes it much easier to arrive at a reasonable default order
873 in applications for which anonymous ciphers are OK (meaning
874 that you can't actually use DEFAULT).
875 [Bodo Moeller; suggested by Victor Duchovni]
877 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
878 processing) into multiple integers instead of setting
879 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
880 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
881 (These masks as well as the individual bit definitions are hidden
882 away into the non-exported interface ssl/ssl_locl.h, so this
883 change to the definition of the SSL_CIPHER structure shouldn't
884 affect applications.) This give us more bits for each of these
885 categories, so there is no longer a need to coagulate AES128 and
886 AES256 into a single algorithm bit, and to coagulate Camellia128
887 and Camellia256 into a single algorithm bit, which has led to all
890 Thus, among other things, the kludge introduced in 0.9.7m and
891 0.9.8e for masking out AES256 independently of AES128 or masking
892 out Camellia256 independently of AES256 is not needed here in 0.9.9.
894 With the change, we also introduce new ciphersuite aliases that
895 so far were missing: "AES128", "AES256", "CAMELLIA128", and
899 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
900 Use the leftmost N bytes of the signature input if the input is
901 larger than the prime q (with N being the size in bytes of q).
904 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
905 it yet and it is largely untested.
908 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
911 *) Initial incomplete changes to avoid need for function casts in OpenSSL
912 some compilers (gcc 4.2 and later) reject their use. Safestack is
913 reimplemented. Update ASN1 to avoid use of legacy functions.
916 *) Win32/64 targets are linked with Winsock2.
919 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
920 to external functions. This can be used to increase CRL handling
921 efficiency especially when CRLs are very large by (for example) storing
922 the CRL revoked certificates in a database.
925 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
926 new CRLs added to a directory can be used. New command line option
927 -verify_return_error to s_client and s_server. This causes real errors
928 to be returned by the verify callback instead of carrying on no matter
929 what. This reflects the way a "real world" verify callback would behave.
932 *) GOST engine, supporting several GOST algorithms and public key formats.
933 Kindly donated by Cryptocom.
936 *) Partial support for Issuing Distribution Point CRL extension. CRLs
937 partitioned by DP are handled but no indirect CRL or reason partitioning
938 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
939 selected via a scoring technique which handles IDP and AKID in CRLs.
942 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
943 will ultimately be used for all verify operations: this will remove the
944 X509_STORE dependency on certificate verification and allow alternative
945 lookup methods. X509_STORE based implementations of these two callbacks.
948 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
949 Modify get_crl() to find a valid (unexpired) CRL if possible.
952 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
953 this would be called X509_CRL_cmp() but that name is already used by
954 a function that just compares CRL issuer names. Cache several CRL
955 extensions in X509_CRL structure and cache CRLDP in X509.
958 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
959 this maps equivalent X509_NAME structures into a consistent structure.
960 Name comparison can then be performed rapidly using memcmp().
963 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
967 *) Allow digests to supply their own micalg string for S/MIME type using
968 the ctrl EVP_MD_CTRL_MICALG.
971 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
972 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
973 ctrl. It can then customise the structure before and/or after signing
977 *) New function OBJ_add_sigid() to allow application defined signature OIDs
978 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
979 to free up any added signature OIDs.
982 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
983 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
984 digest and cipher tables. New options added to openssl utility:
985 list-message-digest-algorithms and list-cipher-algorithms.
988 *) Change the array representation of binary polynomials: the list
989 of degrees of non-zero coefficients is now terminated with -1.
990 Previously it was terminated with 0, which was also part of the
991 value; thus, the array representation was not applicable to
992 polynomials where t^0 has coefficient zero. This change makes
993 the array representation useful in a more general context.
996 *) Various modifications and fixes to SSL/TLS cipher string
997 handling. For ECC, the code now distinguishes between fixed ECDH
998 with RSA certificates on the one hand and with ECDSA certificates
999 on the other hand, since these are separate ciphersuites. The
1000 unused code for Fortezza ciphersuites has been removed.
1002 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1003 (not "ECDHE"). For consistency with the code for DH
1004 certificates, use of ECDH certificates is now considered ECDH
1005 authentication, not RSA or ECDSA authentication (the latter is
1006 merely the CA's signing algorithm and not actively used in the
1009 The temporary ciphersuite alias "ECCdraft" is no longer
1010 available, and ECC ciphersuites are no longer excluded from "ALL"
1011 and "DEFAULT". The following aliases now exist for RFC 4492
1012 ciphersuites, most of these by analogy with the DH case:
1014 kECDHr - ECDH cert, signed with RSA
1015 kECDHe - ECDH cert, signed with ECDSA
1016 kECDH - ECDH cert (signed with either RSA or ECDSA)
1017 kEECDH - ephemeral ECDH
1018 ECDH - ECDH cert or ephemeral ECDH
1024 AECDH - anonymous ECDH
1025 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1029 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1030 Use correct micalg parameters depending on digest(s) in signed message.
1033 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1034 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1037 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1038 an engine to register a method. Add ENGINE lookups for methods and
1039 functional reference processing.
1042 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1043 EVP_{Sign,Verify}* which allow an application to customise the signature
1047 *) New -resign option to smime utility. This adds one or more signers
1048 to an existing PKCS#7 signedData structure. Also -md option to use an
1049 alternative message digest algorithm for signing.
1052 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1053 create PKCS7 structures containing multiple signers. Update smime
1054 application to support multiple signers.
1057 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1061 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1062 Reorganize PBE internals to lookup from a static table using NIDs,
1063 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1064 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1065 PRF which will be automatically used with PBES2.
1068 *) Replace the algorithm specific calls to generate keys in "req" with the
1072 *) Update PKCS#7 enveloped data routines to use new API. This is now
1073 supported by any public key method supporting the encrypt operation. A
1074 ctrl is added to allow the public key algorithm to examine or modify
1075 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1079 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1080 a default digest type to use. In most cases this will be SHA1 but some
1081 algorithms (such as GOST) need to specify an alternative digest. The
1082 return value indicates how strong the prefernce is 1 means optional and
1083 2 is mandatory (that is it is the only supported type). Modify
1084 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1085 use the default md. Update openssl utilities to use the default digest
1086 type for signing if it is not explicitly indicated.
1089 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1090 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1091 signing method from the key type. This effectively removes the link
1092 between digests and public key types.
1095 *) Add an OID cross reference table and utility functions. Its purpose is to
1096 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1097 rsaEncryption. This will allow some of the algorithm specific hackery
1098 needed to use the correct OID to be removed.
1101 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1102 structures for PKCS7_sign(). They are now set up by the relevant public
1106 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1109 *) Add support for key derivation (agreement) in the API, DH method and
1113 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1114 public and private key formats. As a side effect these add additional
1115 command line functionality not previously available: DSA signatures can be
1116 generated and verified using pkeyutl and DH key support and generation in
1121 [Oliver Tappe <zooey@hirschkaefer.de>]
1123 *) New make target "install_html_docs" installs HTML renditions of the
1125 [Oliver Tappe <zooey@hirschkaefer.de>]
1127 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1128 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1129 support key and parameter generation and add initial key generation
1130 functionality for RSA.
1133 *) Add functions for main EVP_PKEY_method operations. The undocumented
1134 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1135 EVP_PKEY_{encrypt,decrypt}_old.
1138 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1139 key API, doesn't do much yet.
1142 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1143 public key algorithms. New option to openssl utility:
1144 "list-public-key-algorithms" to print out info.
1147 *) Implement the Supported Elliptic Curves Extension for
1148 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1151 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1152 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1155 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1156 utilities such as rsa, dsa, dsaparam etc except they process any key
1160 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1161 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1162 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1166 *) Initial support for pluggable public key ASN1.
1167 De-spaghettify the public key ASN1 handling. Move public and private
1168 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1169 algorithm specific handling to a single module within the relevant
1170 algorithm directory. Add functions to allow (near) opaque processing
1171 of public and private key structures.
1174 *) Implement the Supported Point Formats Extension for
1175 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1178 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1179 for the psk identity [hint] and the psk callback functions to the
1180 SSL_SESSION, SSL and SSL_CTX structure.
1183 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1187 SSL_CTX_use_psk_identity_hint
1188 SSL_get_psk_identity_hint
1189 SSL_get_psk_identity
1190 SSL_use_psk_identity_hint
1192 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1194 *) Add RFC 3161 compliant time stamp request creation, response generation
1195 and response verification functionality.
1196 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1198 *) Add initial support for TLS extensions, specifically for the server_name
1199 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1200 have new members for a host name. The SSL data structure has an
1201 additional member SSL_CTX *initial_ctx so that new sessions can be
1202 stored in that context to allow for session resumption, even after the
1203 SSL has been switched to a new SSL_CTX in reaction to a client's
1204 server_name extension.
1206 New functions (subject to change):
1208 SSL_get_servername()
1209 SSL_get_servername_type()
1212 New CTRL codes and macros (subject to change):
1214 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1215 - SSL_CTX_set_tlsext_servername_callback()
1216 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1217 - SSL_CTX_set_tlsext_servername_arg()
1218 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1220 openssl s_client has a new '-servername ...' option.
1222 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1223 '-key2 ...', '-servername_fatal' (subject to change). This allows
1224 testing the HostName extension for a specific single host name ('-cert'
1225 and '-key' remain fallbacks for handshakes without HostName
1226 negotiation). If the unrecogninzed_name alert has to be sent, this by
1227 default is a warning; it becomes fatal with the '-servername_fatal'
1230 [Peter Sylvester, Remy Allais, Christophe Renou]
1232 *) Whirlpool hash implementation is added.
1235 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1236 bn(64,32). Because of instruction set limitations it doesn't have
1237 any negative impact on performance. This was done mostly in order
1238 to make it possible to share assembler modules, such as bn_mul_mont
1239 implementations, between 32- and 64-bit builds without hassle.
1242 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1243 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1247 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1248 dedicated Montgomery multiplication procedure, is introduced.
1249 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1250 "64-bit" performance on certain 32-bit targets.
1253 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1254 in SSL structures. New SSL ctrl to set maximum send fragment size.
1255 Save memory by seeting the I/O buffer sizes dynamically instead of
1256 using the maximum available value.
1259 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1260 in addition to the text details.
1263 *) Very, very preliminary EXPERIMENTAL support for printing of general
1264 ASN1 structures. This currently produces rather ugly output and doesn't
1265 handle several customised structures at all.
1268 *) Integrated support for PVK file format and some related formats such
1269 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1270 these in the 'rsa' and 'dsa' utilities.
1273 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1276 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1277 place for the (very old) "NETSCAPE" format certificates which are now
1278 handled using new ASN1 code equivalents.
1281 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1282 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1283 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1286 *) Modify CRL distribution points extension code to print out previously
1287 unsupported fields. Enhance extension setting code to allow setting of
1291 *) Add print and set support for Issuing Distribution Point CRL extension.
1294 *) Change 'Configure' script to enable Camellia by default.
1297 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1299 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1300 Thanks to Antonio Martin, Enterprise Secure Access Research and
1301 Development, Cisco Systems, Inc. for discovering this bug and
1302 preparing a fix. (CVE-2012-0050)
1305 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1307 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1308 of the Vaudenay padding oracle attack on CBC mode encryption
1309 which enables an efficient plaintext recovery attack against
1310 the OpenSSL implementation of DTLS. Their attack exploits timing
1311 differences arising during decryption processing. A research
1312 paper describing this attack can be found at:
1313 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1314 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1315 Security Group at Royal Holloway, University of London
1316 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1317 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1318 for preparing the fix. (CVE-2011-4108)
1319 [Robin Seggelmann, Michael Tuexen]
1321 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1322 [Ben Laurie, Kasper <ekasper@google.com>]
1324 *) Clear bytes used for block padding of SSL 3.0 records.
1326 [Adam Langley (Google)]
1328 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1329 Kadianakis <desnacked@gmail.com> for discovering this issue and
1330 Adam Langley for preparing the fix. (CVE-2011-4619)
1331 [Adam Langley (Google)]
1333 *) Prevent malformed RFC3779 data triggering an assertion failure.
1334 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1335 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1336 [Rob Austein <sra@hactrn.net>]
1338 *) Fix ssl_ciph.c set-up race.
1339 [Adam Langley (Google)]
1341 *) Fix spurious failures in ecdsatest.c.
1342 [Emilia Käsper (Google)]
1344 *) Fix the BIO_f_buffer() implementation (which was mixing different
1345 interpretations of the '..._len' fields).
1346 [Adam Langley (Google)]
1348 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1349 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1350 threads won't reuse the same blinding coefficients.
1352 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1353 lock to call BN_BLINDING_invert_ex, and avoids one use of
1354 BN_BLINDING_update for each BN_BLINDING structure (previously,
1355 the last update always remained unused).
1356 [Emilia Käsper (Google)]
1358 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1359 for multi-threaded use of ECDH.
1360 [Adam Langley (Google)]
1362 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1365 *) Add protection against ECDSA timing attacks as mentioned in the paper
1366 by Billy Bob Brumley and Nicola Tuveri, see:
1368 http://eprint.iacr.org/2011/232.pdf
1370 [Billy Bob Brumley and Nicola Tuveri]
1372 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1374 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1375 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1377 *) Fix bug in string printing code: if *any* escaping is enabled we must
1378 escape the escape character (backslash) or the resulting string is
1382 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1384 *) Disable code workaround for ancient and obsolete Netscape browsers
1385 and servers: an attacker can use it in a ciphersuite downgrade attack.
1386 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1389 *) Fixed J-PAKE implementation error, originally discovered by
1390 Sebastien Martini, further info and confirmation from Stefan
1391 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1394 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1396 *) Fix extension code to avoid race conditions which can result in a buffer
1397 overrun vulnerability: resumed sessions must not be modified as they can
1398 be shared by multiple threads. CVE-2010-3864
1401 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1404 *) Don't reencode certificate when calculating signature: cache and use
1405 the original encoding instead. This makes signature verification of
1406 some broken encodings work correctly.
1409 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1410 is also one of the inputs.
1411 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1413 *) Don't repeatedly append PBE algorithms to table if they already exist.
1414 Sort table on each new add. This effectively makes the table read only
1415 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1419 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1421 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1424 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1425 access or freeing data twice (CVE-2010-0742)
1426 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1428 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1429 common in certificates and some applications which only call
1430 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1434 Reduce copying into .apps and .test in makevms.com
1435 Don't try to use blank CA certificate in CA.com
1436 Allow use of C files from original directories in maketests.com
1437 [Steven M. Schweda" <sms@antinode.info>]
1439 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1441 *) When rejecting SSL/TLS records due to an incorrect version number, never
1442 update s->server with a new major version number. As of
1443 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1444 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1445 the previous behavior could result in a read attempt at NULL when
1446 receiving specific incorrect SSL/TLS records once record payload
1447 protection is active. (CVE-2010-0740)
1448 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1450 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1451 could be crashed if the relevant tables were not present (e.g. chrooted).
1452 [Tomas Hoger <thoger@redhat.com>]
1454 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1456 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1457 [Martin Olsson, Neel Mehta]
1459 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1460 accommodate for stack sorting, always a write lock!).
1463 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1464 excessive delays in the RAND_poll(): over a minute. As a workaround
1465 include a time check in the inner Heap32Next loop too.
1468 *) The code that handled flushing of data in SSL/TLS originally used the
1469 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1470 the problem outlined in PR#1949. The fix suggested there however can
1471 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1472 of Apache). So instead simplify the code to flush unconditionally.
1473 This should be fine since flushing with no data to flush is a no op.
1476 *) Handle TLS versions 2.0 and later properly and correctly use the
1477 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1478 off ancient servers have a habit of sticking around for a while...
1481 *) Modify compression code so it frees up structures without using the
1482 ex_data callbacks. This works around a problem where some applications
1483 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1484 restarting) then use compression (e.g. SSL with compression) later.
1485 This results in significant per-connection memory leaks and
1486 has caused some security issues including CVE-2008-1678 and
1490 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1491 change when encrypting or decrypting.
1494 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1495 connect and renegotiate with servers which do not support RI.
1496 Until RI is more widely deployed this option is enabled by default.
1499 *) Add "missing" ssl ctrls to clear options and mode.
1502 *) If client attempts to renegotiate and doesn't support RI respond with
1503 a no_renegotiation alert as required by RFC5746. Some renegotiating
1504 TLS clients will continue a connection gracefully when they receive
1505 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1506 waiting for a server hello which it will never receive. Now we treat a
1507 received no_renegotiation alert as a fatal error. This is because
1508 applications requesting a renegotiation might well expect it to succeed
1509 and would have no code in place to handle the server denying it so the
1510 only safe thing to do is to terminate the connection.
1513 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1514 peer supports secure renegotiation and 0 otherwise. Print out peer
1515 renegotiation support in s_client/s_server.
1518 *) Replace the highly broken and deprecated SPKAC certification method with
1519 the updated NID creation version. This should correctly handle UTF8.
1522 *) Implement RFC5746. Re-enable renegotiation but require the extension
1523 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1524 turns out to be a bad idea. It has been replaced by
1525 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1526 SSL_CTX_set_options(). This is really not recommended unless you
1527 know what you are doing.
1528 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1530 *) Fixes to stateless session resumption handling. Use initial_ctx when
1531 issuing and attempting to decrypt tickets in case it has changed during
1532 servername handling. Use a non-zero length session ID when attempting
1533 stateless session resumption: this makes it possible to determine if
1534 a resumption has occurred immediately after receiving server hello
1535 (several places in OpenSSL subtly assume this) instead of later in
1539 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1540 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1541 fixes for a few places where the return code is not checked
1543 [Julia Lawall <julia@diku.dk>]
1545 *) Add --strict-warnings option to Configure script to include devteam
1546 warnings in other configurations.
1549 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1550 makes it possible to install openssl libraries in locations which
1551 have names other than "lib", for example "/usr/lib64" which some
1553 [Steve Henson, based on patch from Jeremy Utley]
1555 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1556 X690 8.9.12 and can produce some misleading textual output of OIDs.
1557 [Steve Henson, reported by Dan Kaminsky]
1559 *) Delete MD2 from algorithm tables. This follows the recommendation in
1560 several standards that it is not used in new applications due to
1561 several cryptographic weaknesses. For binary compatibility reasons
1562 the MD2 API is still compiled in by default.
1565 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1569 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1570 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1572 [Guenter <lists@gknw.net>]
1574 *) Fix the server certificate chain building code to use X509_verify_cert(),
1575 it used to have an ad-hoc builder which was unable to cope with anything
1576 other than a simple chain.
1577 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1579 *) Don't check self signed certificate signatures in X509_verify_cert()
1580 by default (a flag can override this): it just wastes time without
1581 adding any security. As a useful side effect self signed root CAs
1582 with non-FIPS digests are now usable in FIPS mode.
1585 *) In dtls1_process_out_of_seq_message() the check if the current message
1586 is already buffered was missing. For every new message was memory
1587 allocated, allowing an attacker to perform an denial of service attack
1588 with sending out of seq handshake messages until there is no memory
1589 left. Additionally every future messege was buffered, even if the
1590 sequence number made no sense and would be part of another handshake.
1591 So only messages with sequence numbers less than 10 in advance will be
1592 buffered. (CVE-2009-1378)
1593 [Robin Seggelmann, discovered by Daniel Mentz]
1595 *) Records are buffered if they arrive with a future epoch to be
1596 processed after finishing the corresponding handshake. There is
1597 currently no limitation to this buffer allowing an attacker to perform
1598 a DOS attack with sending records with future epochs until there is no
1599 memory left. This patch adds the pqueue_size() function to detemine
1600 the size of a buffer and limits the record buffer to 100 entries.
1602 [Robin Seggelmann, discovered by Daniel Mentz]
1604 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1605 parent structure is freed. (CVE-2009-1379)
1608 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1609 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1612 [Ilya O. <vrghost@gmail.com>]
1614 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1616 *) Disable renegotiation completely - this fixes a severe security
1617 problem (CVE-2009-3555) at the cost of breaking all
1618 renegotiation. Renegotiation can be re-enabled by setting
1619 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1620 run-time. This is really not recommended unless you know what
1624 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1626 *) Don't set val to NULL when freeing up structures, it is freed up by
1627 underlying code. If sizeof(void *) > sizeof(long) this can result in
1628 zeroing past the valid field. (CVE-2009-0789)
1629 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1631 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1632 checked correctly. This would allow some invalid signed attributes to
1633 appear to verify correctly. (CVE-2009-0591)
1634 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1636 *) Reject UniversalString and BMPString types with invalid lengths. This
1637 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1638 a legal length. (CVE-2009-0590)
1641 *) Set S/MIME signing as the default purpose rather than setting it
1642 unconditionally. This allows applications to override it at the store
1646 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1647 to handle some structures.
1650 *) Improve efficiency of mem_gets: don't search whole buffer each time
1652 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1654 *) New -hex option for openssl rand.
1657 *) Print out UTF8String and NumericString when parsing ASN1.
1660 *) Support NumericString type for name components.
1663 *) Allow CC in the environment to override the automatically chosen
1664 compiler. Note that nothing is done to ensure flags work with the
1668 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1670 *) Properly check EVP_VerifyFinal() and similar return values
1672 [Ben Laurie, Bodo Moeller, Google Security Team]
1674 *) Enable TLS extensions by default.
1677 *) Allow the CHIL engine to be loaded, whether the application is
1678 multithreaded or not. (This does not release the developer from the
1679 obligation to set up the dynamic locking callbacks.)
1680 [Sander Temme <sander@temme.net>]
1682 *) Use correct exit code if there is an error in dgst command.
1683 [Steve Henson; problem pointed out by Roland Dirlewanger]
1685 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1686 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1689 *) Add experimental JPAKE support, including demo authentication in
1690 s_client and s_server.
1693 *) Set the comparison function in v3_addr_canonize().
1694 [Rob Austein <sra@hactrn.net>]
1696 *) Add support for XMPP STARTTLS in s_client.
1697 [Philip Paeps <philip@freebsd.org>]
1699 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1700 to ensure that even with this option, only ciphersuites in the
1701 server's preference list will be accepted. (Note that the option
1702 applies only when resuming a session, so the earlier behavior was
1703 just about the algorithm choice for symmetric cryptography.)
1706 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1708 *) Fix NULL pointer dereference if a DTLS server received
1709 ChangeCipherSpec as first record (CVE-2009-1386).
1712 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1713 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1716 *) The fix in 0.9.8c that supposedly got rid of unsafe
1717 double-checked locking was incomplete for RSA blinding,
1718 addressing just one layer of what turns out to have been
1719 doubly unsafe triple-checked locking.
1721 So now fix this for real by retiring the MONT_HELPER macro
1722 in crypto/rsa/rsa_eay.c.
1724 [Bodo Moeller; problem pointed out by Marius Schilder]
1726 *) Various precautionary measures:
1728 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1730 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1731 (NB: This would require knowledge of the secret session ticket key
1732 to exploit, in which case you'd be SOL either way.)
1734 - Change bn_nist.c so that it will properly handle input BIGNUMs
1735 outside the expected range.
1737 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1740 [Neel Mehta, Bodo Moeller]
1742 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1743 the load fails. Useful for distros.
1744 [Ben Laurie and the FreeBSD team]
1746 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1749 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1752 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1754 This work was sponsored by Logica.
1757 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1758 keystores. Support for SSL/TLS client authentication too.
1759 Not compiled unless enable-capieng specified to Configure.
1761 This work was sponsored by Logica.
1764 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1765 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1766 attribute creation routines such as certifcate requests and PKCS#12
1770 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1772 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1773 handshake which could lead to a cilent crash as found using the
1774 Codenomicon TLS test suite (CVE-2008-1672)
1775 [Steve Henson, Mark Cox]
1777 *) Fix double free in TLS server name extensions which could lead to
1778 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1781 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1783 Clear the error queue to ensure that error entries left from
1784 older function calls do not interfere with the correct operation.
1785 [Lutz Jaenicke, Erik de Castro Lopo]
1787 *) Remove root CA certificates of commercial CAs:
1789 The OpenSSL project does not recommend any specific CA and does not
1790 have any policy with respect to including or excluding any CA.
1791 Therefore it does not make any sense to ship an arbitrary selection
1792 of root CA certificates with the OpenSSL software.
1795 *) RSA OAEP patches to fix two separate invalid memory reads.
1796 The first one involves inputs when 'lzero' is greater than
1797 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1798 before the beginning of from). The second one involves inputs where
1799 the 'db' section contains nothing but zeroes (there is a one-byte
1800 invalid read after the end of 'db').
1801 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1803 *) Partial backport from 0.9.9-dev:
1805 Introduce bn_mul_mont (dedicated Montgomery multiplication
1806 procedure) as a candidate for BIGNUM assembler implementation.
1807 While 0.9.9-dev uses assembler for various architectures, only
1808 x86_64 is available by default here in the 0.9.8 branch, and
1809 32-bit x86 is available through a compile-time setting.
1811 To try the 32-bit x86 assembler implementation, use Configure
1812 option "enable-montasm" (which exists only for this backport).
1814 As "enable-montasm" for 32-bit x86 disclaims code stability
1815 anyway, in this constellation we activate additional code
1816 backported from 0.9.9-dev for further performance improvements,
1817 namely BN_from_montgomery_word. (To enable this otherwise,
1818 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1820 [Andy Polyakov (backport partially by Bodo Moeller)]
1822 *) Add TLS session ticket callback. This allows an application to set
1823 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1824 values. This is useful for key rollover for example where several key
1825 sets may exist with different names.
1828 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1829 This was broken until now in 0.9.8 releases, such that the only way
1830 a registered ENGINE could be used (assuming it initialises
1831 successfully on the host) was to explicitly set it as the default
1832 for the relevant algorithms. This is in contradiction with 0.9.7
1833 behaviour and the documentation. With this fix, when an ENGINE is
1834 registered into a given algorithm's table of implementations, the
1835 'uptodate' flag is reset so that auto-discovery will be used next
1836 time a new context for that algorithm attempts to select an
1838 [Ian Lister (tweaked by Geoff Thorpe)]
1840 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1841 implemention in the following ways:
1843 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1846 Lack of BER streaming support means one pass streaming processing is
1847 only supported if data is detached: setting the streaming flag is
1848 ignored for embedded content.
1850 CMS support is disabled by default and must be explicitly enabled
1851 with the enable-cms configuration option.
1854 *) Update the GMP engine glue to do direct copies between BIGNUM and
1855 mpz_t when openssl and GMP use the same limb size. Otherwise the
1856 existing "conversion via a text string export" trick is still used.
1857 [Paul Sheer <paulsheer@gmail.com>]
1859 *) Zlib compression BIO. This is a filter BIO which compressed and
1860 uncompresses any data passed through it.
1863 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1864 RFC3394 compatible AES key wrapping.
1867 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1868 sets string data without copying. X509_ALGOR_set0() and
1869 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1870 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1871 from an X509_ATTRIBUTE structure optionally checking it occurs only
1872 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1876 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1877 to get the expected BN_FLG_CONSTTIME behavior.
1878 [Bodo Moeller (Google)]
1882 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1883 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1884 - added some more tests to do_tests.pl
1885 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1886 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1887 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1888 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1889 - various changes to netware.pl to enable gcc-cross builds on Win32
1891 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1892 - various changes to fix missing prototype warnings
1893 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1894 - added AES, WHIRLPOOL and CPUID assembler code to build files
1895 - added missing AES assembler make rules to mk1mf.pl
1896 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1897 [Guenter Knauf <eflash@gmx.net>]
1899 *) Implement certificate status request TLS extension defined in RFC3546.
1900 A client can set the appropriate parameters and receive the encoded
1901 OCSP response via a callback. A server can query the supplied parameters
1902 and set the encoded OCSP response in the callback. Add simplified examples
1903 to s_client and s_server.
1906 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1908 *) Fix various bugs:
1909 + Binary incompatibility of ssl_ctx_st structure
1910 + DTLS interoperation with non-compliant servers
1911 + Don't call get_session_cb() without proposed session
1912 + Fix ia64 assembler code
1913 [Andy Polyakov, Steve Henson]
1915 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1917 *) DTLS Handshake overhaul. There were longstanding issues with
1918 OpenSSL DTLS implementation, which were making it impossible for
1919 RFC 4347 compliant client to communicate with OpenSSL server.
1920 Unfortunately just fixing these incompatibilities would "cut off"
1921 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1922 server keeps tolerating non RFC compliant syntax. The opposite is
1923 not true, 0.9.8f client can not communicate with earlier server.
1924 This update even addresses CVE-2007-4995.
1927 *) Changes to avoid need for function casts in OpenSSL: some compilers
1928 (gcc 4.2 and later) reject their use.
1929 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1932 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1933 RFC4507bis. The encrypted ticket format is an encrypted encoded
1934 SSL_SESSION structure, that way new session features are automatically
1937 If a client application caches session in an SSL_SESSION structure
1938 support is transparent because tickets are now stored in the encoded
1941 The SSL_CTX structure automatically generates keys for ticket
1942 protection in servers so again support should be possible
1943 with no application modification.
1945 If a client or server wishes to disable RFC4507 support then the option
1946 SSL_OP_NO_TICKET can be set.
1948 Add a TLS extension debugging callback to allow the contents of any client
1949 or server extensions to be examined.
1951 This work was sponsored by Google.
1954 *) Add initial support for TLS extensions, specifically for the server_name
1955 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1956 have new members for a host name. The SSL data structure has an
1957 additional member SSL_CTX *initial_ctx so that new sessions can be
1958 stored in that context to allow for session resumption, even after the
1959 SSL has been switched to a new SSL_CTX in reaction to a client's
1960 server_name extension.
1962 New functions (subject to change):
1964 SSL_get_servername()
1965 SSL_get_servername_type()
1968 New CTRL codes and macros (subject to change):
1970 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1971 - SSL_CTX_set_tlsext_servername_callback()
1972 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1973 - SSL_CTX_set_tlsext_servername_arg()
1974 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1976 openssl s_client has a new '-servername ...' option.
1978 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1979 '-key2 ...', '-servername_fatal' (subject to change). This allows
1980 testing the HostName extension for a specific single host name ('-cert'
1981 and '-key' remain fallbacks for handshakes without HostName
1982 negotiation). If the unrecogninzed_name alert has to be sent, this by
1983 default is a warning; it becomes fatal with the '-servername_fatal'
1986 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1988 *) Add AES and SSE2 assembly language support to VC++ build.
1991 *) Mitigate attack on final subtraction in Montgomery reduction.
1994 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1995 (which previously caused an internal error).
1998 *) Squeeze another 10% out of IGE mode when in != out.
2001 *) AES IGE mode speedup.
2002 [Dean Gaudet (Google)]
2004 *) Add the Korean symmetric 128-bit cipher SEED (see
2005 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2006 add SEED ciphersuites from RFC 4162:
2008 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2009 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2010 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2011 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2013 To minimize changes between patchlevels in the OpenSSL 0.9.8
2014 series, SEED remains excluded from compilation unless OpenSSL
2015 is configured with 'enable-seed'.
2016 [KISA, Bodo Moeller]
2018 *) Mitigate branch prediction attacks, which can be practical if a
2019 single processor is shared, allowing a spy process to extract
2020 information. For detailed background information, see
2021 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2022 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2023 and Necessary Software Countermeasures"). The core of the change
2024 are new versions BN_div_no_branch() and
2025 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2026 respectively, which are slower, but avoid the security-relevant
2027 conditional branches. These are automatically called by BN_div()
2028 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2029 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2030 remove a conditional branch.
2032 BN_FLG_CONSTTIME is the new name for the previous
2033 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2034 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2035 in the exponent causes BN_mod_exp_mont() to use the alternative
2036 implementation in BN_mod_exp_mont_consttime().) The old name
2037 remains as a deprecated alias.
2039 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2040 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2041 constant-time implementations for more than just exponentiation.
2042 Here too the old name is kept as a deprecated alias.
2044 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2045 the BN_BLINDING structure gets an independent copy of the
2046 modulus. This means that the previous "BIGNUM *m" argument to
2047 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2048 essentially becomes "const BIGNUM *m", although we can't actually
2049 change this in the header file before 0.9.9. It allows
2050 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2051 enable BN_FLG_CONSTTIME.
2053 [Matthew D Wood (Intel Corp)]
2055 *) In the SSL/TLS server implementation, be strict about session ID
2056 context matching (which matters if an application uses a single
2057 external cache for different purposes). Previously,
2058 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2059 set. This did ensure strict client verification, but meant that,
2060 with applications using a single external cache for quite
2061 different requirements, clients could circumvent ciphersuite
2062 restrictions for a given session ID context by starting a session
2063 in a different context.
2066 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2067 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2068 authentication-only ciphersuites.
2071 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2072 not complete and could lead to a possible single byte overflow
2073 (CVE-2007-5135) [Ben Laurie]
2075 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2077 *) Since AES128 and AES256 (and similarly Camellia128 and
2078 Camellia256) share a single mask bit in the logic of
2079 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2080 kludge to work properly if AES128 is available and AES256 isn't
2081 (or if Camellia128 is available and Camellia256 isn't).
2084 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2085 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2086 When a point or a seed is encoded in a BIT STRING, we need to
2087 prevent the removal of trailing zero bits to get the proper DER
2088 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2089 of a NamedBitList, for which trailing 0 bits need to be removed.)
2092 *) Have SSL/TLS server implementation tolerate "mismatched" record
2093 protocol version while receiving ClientHello even if the
2094 ClientHello is fragmented. (The server can't insist on the
2095 particular protocol version it has chosen before the ServerHello
2096 message has informed the client about his choice.)
2099 *) Add RFC 3779 support.
2100 [Rob Austein for ARIN, Ben Laurie]
2102 *) Load error codes if they are not already present instead of using a
2103 static variable. This allows them to be cleanly unloaded and reloaded.
2104 Improve header file function name parsing.
2107 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2108 or CAPABILITY handshake as required by RFCs.
2111 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2113 *) Introduce limits to prevent malicious keys being able to
2114 cause a denial of service. (CVE-2006-2940)
2115 [Steve Henson, Bodo Moeller]
2117 *) Fix ASN.1 parsing of certain invalid structures that can result
2118 in a denial of service. (CVE-2006-2937) [Steve Henson]
2120 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2121 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2123 *) Fix SSL client code which could crash if connecting to a
2124 malicious SSLv2 server. (CVE-2006-4343)
2125 [Tavis Ormandy and Will Drewry, Google Security Team]
2127 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2128 match only those. Before that, "AES256-SHA" would be interpreted
2129 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2130 the same strength classification in 0.9.7h) as we currently only
2131 have a single AES bit in the ciphersuite description bitmap.
2132 That change, however, also applied to ciphersuite strings such as
2133 "RC4-MD5" that intentionally matched multiple ciphersuites --
2134 namely, SSL 2.0 ciphersuites in addition to the more common ones
2135 from SSL 3.0/TLS 1.0.
2137 So we change the selection algorithm again: Naming an explicit
2138 ciphersuite selects this one ciphersuite, and any other similar
2139 ciphersuite (same bitmap) from *other* protocol versions.
2140 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2141 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2143 Since SSL 2.0 does not have any ciphersuites for which the
2144 128/256 bit distinction would be relevant, this works for now.
2145 The proper fix will be to use different bits for AES128 and
2146 AES256, which would have avoided the problems from the beginning;
2147 however, bits are scarce, so we can only do this in a new release
2148 (not just a patchlevel) when we can change the SSL_CIPHER
2149 definition to split the single 'unsigned long mask' bitmap into
2150 multiple values to extend the available space.
2154 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2156 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2157 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2159 *) Add AES IGE and biIGE modes.
2162 *) Change the Unix randomness entropy gathering to use poll() when
2163 possible instead of select(), since the latter has some
2164 undesirable limitations.
2165 [Darryl Miles via Richard Levitte and Bodo Moeller]
2167 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2168 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2169 cannot be implicitly activated as part of, e.g., the "AES" alias.
2170 However, please upgrade to OpenSSL 0.9.9[-dev] for
2171 non-experimental use of the ECC ciphersuites to get TLS extension
2172 support, which is required for curve and point format negotiation
2173 to avoid potential handshake problems.
2176 *) Disable rogue ciphersuites:
2178 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2179 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2180 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2182 The latter two were purportedly from
2183 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2186 Also deactivate the remaining ciphersuites from
2187 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2188 unofficial, and the ID has long expired.
2191 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2192 dual-core machines) and other potential thread-safety issues.
2195 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2196 versions), which is now available for royalty-free use
2197 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2198 Also, add Camellia TLS ciphersuites from RFC 4132.
2200 To minimize changes between patchlevels in the OpenSSL 0.9.8
2201 series, Camellia remains excluded from compilation unless OpenSSL
2202 is configured with 'enable-camellia'.
2205 *) Disable the padding bug check when compression is in use. The padding
2206 bug check assumes the first packet is of even length, this is not
2207 necessarily true if compresssion is enabled and can result in false
2208 positives causing handshake failure. The actual bug test is ancient
2209 code so it is hoped that implementations will either have fixed it by
2210 now or any which still have the bug do not support compression.
2213 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2215 *) When applying a cipher rule check to see if string match is an explicit
2216 cipher suite and only match that one cipher suite if it is.
2219 *) Link in manifests for VC++ if needed.
2220 [Austin Ziegler <halostatue@gmail.com>]
2222 *) Update support for ECC-based TLS ciphersuites according to
2223 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2224 TLS extensions, which are supported starting with the 0.9.9
2225 branch, not in the OpenSSL 0.9.8 branch).
2228 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2229 opaque EVP_CIPHER_CTX handling.
2232 *) Fixes and enhancements to zlib compression code. We now only use
2233 "zlib1.dll" and use the default __cdecl calling convention on Win32
2234 to conform with the standards mentioned here:
2235 http://www.zlib.net/DLL_FAQ.txt
2236 Static zlib linking now works on Windows and the new --with-zlib-include
2237 --with-zlib-lib options to Configure can be used to supply the location
2238 of the headers and library. Gracefully handle case where zlib library
2242 *) Several fixes and enhancements to the OID generation code. The old code
2243 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2244 handle numbers larger than ULONG_MAX, truncated printing and had a
2245 non standard OBJ_obj2txt() behaviour.
2248 *) Add support for building of engines under engine/ as shared libraries
2249 under VC++ build system.
2252 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2253 Hopefully, we will not see any false combination of paths any more.
2256 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2258 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2259 (part of SSL_OP_ALL). This option used to disable the
2260 countermeasure against man-in-the-middle protocol-version
2261 rollback in the SSL 2.0 server implementation, which is a bad
2262 idea. (CVE-2005-2969)
2264 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2265 for Information Security, National Institute of Advanced Industrial
2266 Science and Technology [AIST], Japan)]
2268 *) Add two function to clear and return the verify parameter flags.
2271 *) Keep cipherlists sorted in the source instead of sorting them at
2272 runtime, thus removing the need for a lock.
2275 *) Avoid some small subgroup attacks in Diffie-Hellman.
2276 [Nick Mathewson and Ben Laurie]
2278 *) Add functions for well-known primes.
2281 *) Extended Windows CE support.
2282 [Satoshi Nakamura and Andy Polyakov]
2284 *) Initialize SSL_METHOD structures at compile time instead of during
2285 runtime, thus removing the need for a lock.
2288 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2289 attempting to decrypt each encrypted key in turn. Add support to
2293 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2295 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2298 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2301 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2302 key into the same file any more.
2305 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2308 *) Add -utf8 command line and config file option to 'ca'.
2309 [Stefan <stf@udoma.org]
2311 *) Removed the macro des_crypt(), as it seems to conflict with some
2312 libraries. Use DES_crypt().
2315 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2316 involves renaming the source and generated shared-libs for
2317 both. The engines will accept the corrected or legacy ids
2318 ('ncipher' and '4758_cca' respectively) when binding. NB,
2319 this only applies when building 'shared'.
2320 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2322 *) Add attribute functions to EVP_PKEY structure. Modify
2323 PKCS12_create() to recognize a CSP name attribute and
2324 use it. Make -CSP option work again in pkcs12 utility.
2327 *) Add new functionality to the bn blinding code:
2328 - automatic re-creation of the BN_BLINDING parameters after
2329 a fixed number of uses (currently 32)
2330 - add new function for parameter creation
2331 - introduce flags to control the update behaviour of the
2332 BN_BLINDING parameters
2333 - hide BN_BLINDING structure
2334 Add a second BN_BLINDING slot to the RSA structure to improve
2335 performance when a single RSA object is shared among several
2339 *) Add support for DTLS.
2340 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2342 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2343 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2346 *) Remove buggy and incompletet DH cert support from
2347 ssl/ssl_rsa.c and ssl/s3_both.c
2350 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2351 the apps/openssl applications.
2354 *) Compile clean with "-Wall -Wmissing-prototypes
2355 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2356 DEBUG_SAFESTACK must also be set.
2359 *) Change ./Configure so that certain algorithms can be disabled by default.
2360 The new counterpiece to "no-xxx" is "enable-xxx".
2362 The patented RC5 and MDC2 algorithms will now be disabled unless
2363 "enable-rc5" and "enable-mdc2", respectively, are specified.
2365 (IDEA remains enabled despite being patented. This is because IDEA
2366 is frequently required for interoperability, and there is no license
2367 fee for non-commercial use. As before, "no-idea" can be used to
2368 avoid this algorithm.)
2372 *) Add processing of proxy certificates (see RFC 3820). This work was
2373 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2374 EGEE (Enabling Grids for E-science in Europe).
2377 *) RC4 performance overhaul on modern architectures/implementations, such
2378 as Intel P4, IA-64 and AMD64.
2381 *) New utility extract-section.pl. This can be used specify an alternative
2382 section number in a pod file instead of having to treat each file as
2383 a separate case in Makefile. This can be done by adding two lines to the
2386 =for comment openssl_section:XXX
2388 The blank line is mandatory.
2392 *) New arguments -certform, -keyform and -pass for s_client and s_server
2393 to allow alternative format key and certificate files and passphrase
2397 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2398 update associated structures and add various utility functions.
2400 Add new policy related verify parameters, include policy checking in
2401 standard verify code. Enhance 'smime' application with extra parameters
2402 to support policy checking and print out.
2405 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2406 Nehemiah processors. These extensions support AES encryption in hardware
2407 as well as RNG (though RNG support is currently disabled).
2408 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2410 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2413 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2414 [Andy Polyakov and a number of other people]
2416 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2417 implementation contributed by IBM.
2418 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2420 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2421 exponent rather than 'unsigned long'. There is a corresponding change to
2422 the new 'rsa_keygen' element of the RSA_METHOD structure.
2423 [Jelte Jansen, Geoff Thorpe]
2425 *) Functionality for creating the initial serial number file is now
2426 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2428 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2429 number file to 1, which is bound to cause problems. To avoid
2430 the problems while respecting compatibility between different 0.9.7
2431 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2432 CA.pl for serial number initialization. With the new release 0.9.8,
2433 we can fix the problem directly in the 'ca' utility.)
2436 *) Reduced header interdepencies by declaring more opaque objects in
2437 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2438 give fewer recursive includes, which could break lazy source code - so
2439 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2440 developers should define this symbol when building and using openssl to
2441 ensure they track the recommended behaviour, interfaces, [etc], but
2442 backwards-compatible behaviour prevails when this isn't defined.
2445 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2448 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2449 This will generate a random key of the appropriate length based on the
2450 cipher context. The EVP_CIPHER can provide its own random key generation
2451 routine to support keys of a specific form. This is used in the des and
2452 3des routines to generate a key of the correct parity. Update S/MIME
2453 code to use new functions and hence generate correct parity DES keys.
2454 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2455 valid (weak or incorrect parity).
2458 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2459 as looking them up. This is useful when the verified structure may contain
2460 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2461 present unless the new PKCS7_NO_CRL flag is asserted.
2464 *) Extend ASN1 oid configuration module. It now additionally accepts the
2467 shortName = some long name, 1.2.3.4
2470 *) Reimplemented the BN_CTX implementation. There is now no more static
2471 limitation on the number of variables it can handle nor the depth of the
2472 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2473 information can now expand as required, and rather than having a single
2474 static array of bignums, BN_CTX now uses a linked-list of such arrays
2475 allowing it to expand on demand whilst maintaining the usefulness of
2476 BN_CTX's "bundling".
2479 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2480 to allow all RSA operations to function using a single BN_CTX.
2483 *) Preliminary support for certificate policy evaluation and checking. This
2484 is initially intended to pass the tests outlined in "Conformance Testing
2485 of Relying Party Client Certificate Path Processing Logic" v1.07.
2488 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2489 remained unused and not that useful. A variety of other little bignum
2490 tweaks and fixes have also been made continuing on from the audit (see
2494 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2495 associated ASN1, EVP and SSL functions and old ASN1 macros.
2498 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2499 and this should never fail. So the return value from the use of
2500 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2501 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2504 *) BN_CTX_get() should return zero-valued bignums, providing the same
2505 initialised value as BN_new().
2506 [Geoff Thorpe, suggested by Ulf Möller]
2508 *) Support for inhibitAnyPolicy certificate extension.
2511 *) An audit of the BIGNUM code is underway, for which debugging code is
2512 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2513 is considered valid when processing BIGNUMs, and causes execution to
2514 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2515 further steps are taken to deliberately pollute unused data in BIGNUM
2516 structures to try and expose faulty code further on. For now, openssl will
2517 (in its default mode of operation) continue to tolerate the inconsistent
2518 forms that it has tolerated in the past, but authors and packagers should
2519 consider trying openssl and their own applications when compiled with
2520 these debugging symbols defined. It will help highlight potential bugs in
2521 their own code, and will improve the test coverage for OpenSSL itself. At
2522 some point, these tighter rules will become openssl's default to improve
2523 maintainability, though the assert()s and other overheads will remain only
2524 in debugging configurations. See bn.h for more details.
2525 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2527 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2528 that can only be obtained through BN_CTX_new() (which implicitly
2529 initialises it). The presence of this function only made it possible
2530 to overwrite an existing structure (and cause memory leaks).
2533 *) Because of the callback-based approach for implementing LHASH as a
2534 template type, lh_insert() adds opaque objects to hash-tables and
2535 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2536 to clean up those corresponding objects before destroying the hash table
2537 (and losing the object pointers). So some over-zealous constifications in
2538 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2539 objects as "const" and the lh_doall[_arg] callback wrappers are not
2540 prototyped to have "const" restrictions on the object pointers they are
2541 given (and so aren't required to cast them away any more).
2544 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2545 (speed) prefers to use its own implementation. The two implementations
2546 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2547 its object type properly exposed (MS_TM) instead of casting to/from "char
2548 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2549 aren't necessarily the greatest nomenclatures - but this is what was used
2550 internally to the implementation so I've used that for now.
2553 *) Ensure that deprecated functions do not get compiled when
2554 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2555 the self-tests were still using deprecated key-generation functions so
2556 these have been updated also.
2559 *) Reorganise PKCS#7 code to separate the digest location functionality
2560 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2561 New function PKCS7_set_digest() to set the digest type for PKCS#7
2562 digestedData type. Add additional code to correctly generate the
2563 digestedData type and add support for this type in PKCS7 initialization
2567 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2568 structure of type "other".
2571 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2572 sure the loop does correctly stop and breaking ("division by zero")
2573 modulus operations are not performed. The (pre-generated) prime
2574 table crypto/bn/bn_prime.h was already correct, but it could not be
2575 re-generated on some platforms because of the "division by zero"
2576 situation in the script.
2577 [Ralf S. Engelschall]
2579 *) Update support for ECC-based TLS ciphersuites according to
2580 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2581 SHA-1 now is only used for "small" curves (where the
2582 representation of a field element takes up to 24 bytes); for
2583 larger curves, the field element resulting from ECDH is directly
2584 used as premaster secret.
2585 [Douglas Stebila (Sun Microsystems Laboratories)]
2587 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2588 curve secp160r1 to the tests.
2589 [Douglas Stebila (Sun Microsystems Laboratories)]
2591 *) Add the possibility to load symbols globally with DSO.
2592 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2594 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2595 control of the error stack.
2598 *) Add support for STORE in ENGINE.
2601 *) Add the STORE type. The intention is to provide a common interface
2602 to certificate and key stores, be they simple file-based stores, or
2603 HSM-type store, or LDAP stores, or...
2604 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2607 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2608 pass a list of arguments to any function as well as provide a way
2609 for a function to pass data back to the caller.
2612 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2613 works like BUF_strdup() but can be used to duplicate a portion of
2614 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2618 *) Add the function sk_find_ex() which works like sk_find(), but will
2619 return an index to an element even if an exact match couldn't be
2620 found. The index is guaranteed to point at the element where the
2621 searched-for key would be inserted to preserve sorting order.
2624 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2625 takes an extra flags argument for optional functionality. Currently,
2626 the following flags are defined:
2628 OBJ_BSEARCH_VALUE_ON_NOMATCH
2629 This one gets OBJ_bsearch_ex() to return a pointer to the first
2630 element where the comparing function returns a negative or zero
2633 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2634 This one gets OBJ_bsearch_ex() to return a pointer to the first
2635 element where the comparing function returns zero. This is useful
2636 if there are more than one element where the comparing function
2640 *) Make it possible to create self-signed certificates with 'openssl ca'
2641 in such a way that the self-signed certificate becomes part of the
2642 CA database and uses the same mechanisms for serial number generation
2643 as all other certificate signing. The new flag '-selfsign' enables
2644 this functionality. Adapt CA.sh and CA.pl.in.
2647 *) Add functionality to check the public key of a certificate request
2648 against a given private. This is useful to check that a certificate
2649 request can be signed by that key (self-signing).
2652 *) Make it possible to have multiple active certificates with the same
2653 subject in the CA index file. This is done only if the keyword
2654 'unique_subject' is set to 'no' in the main CA section (default
2655 if 'CA_default') of the configuration file. The value is saved
2656 with the database itself in a separate index attribute file,
2657 named like the index file with '.attr' appended to the name.
2660 *) Generate muti valued AVAs using '+' notation in config files for
2664 *) Support for nameConstraints certificate extension.
2667 *) Support for policyConstraints certificate extension.
2670 *) Support for policyMappings certificate extension.
2673 *) Make sure the default DSA_METHOD implementation only uses its
2674 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2675 and change its own handlers to be NULL so as to remove unnecessary
2676 indirection. This lets alternative implementations fallback to the
2677 default implementation more easily.
2680 *) Support for directoryName in GeneralName related extensions
2684 *) Make it possible to link applications using Makefile.shared.
2685 Make that possible even when linking against static libraries!
2688 *) Support for single pass processing for S/MIME signing. This now
2689 means that S/MIME signing can be done from a pipe, in addition
2690 cleartext signing (multipart/signed type) is effectively streaming
2691 and the signed data does not need to be all held in memory.
2693 This is done with a new flag PKCS7_STREAM. When this flag is set
2694 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2695 is done after the data is output (and digests calculated) in
2696 SMIME_write_PKCS7().
2699 *) Add full support for -rpath/-R, both in shared libraries and
2700 applications, at least on the platforms where it's known how
2704 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2705 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2706 will now compute a table of multiples of the generator that
2707 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2708 faster (notably in the case of a single point multiplication,
2709 scalar * generator).
2710 [Nils Larsch, Bodo Moeller]
2712 *) IPv6 support for certificate extensions. The various extensions
2713 which use the IP:a.b.c.d can now take IPv6 addresses using the
2714 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2718 *) Added an ENGINE that implements RSA by performing private key
2719 exponentiations with the GMP library. The conversions to and from
2720 GMP's mpz_t format aren't optimised nor are any montgomery forms
2721 cached, and on x86 it appears OpenSSL's own performance has caught up.
2722 However there are likely to be other architectures where GMP could
2723 provide a boost. This ENGINE is not built in by default, but it can be
2724 specified at Configure time and should be accompanied by the necessary
2725 linker additions, eg;
2726 ./config -DOPENSSL_USE_GMP -lgmp
2729 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2730 testing availability of engines with "-t" - the old behaviour is
2731 produced by increasing the feature's verbosity with "-tt".
2734 *) ECDSA routines: under certain error conditions uninitialized BN objects
2735 could be freed. Solution: make sure initialization is performed early
2736 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2740 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2741 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2742 software implementations. For DSA and DH, parameter generation can
2743 also be overriden by providing the appropriate method callbacks.
2746 *) Change the "progress" mechanism used in key-generation and
2747 primality testing to functions that take a new BN_GENCB pointer in
2748 place of callback/argument pairs. The new API functions have "_ex"
2749 postfixes and the older functions are reimplemented as wrappers for
2750 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2751 declarations of the old functions to help (graceful) attempts to
2752 migrate to the new functions. Also, the new key-generation API
2753 functions operate on a caller-supplied key-structure and return
2754 success/failure rather than returning a key or NULL - this is to
2755 help make "keygen" another member function of RSA_METHOD etc.
2757 Example for using the new callback interface:
2759 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2763 BN_GENCB_set(&my_cb, my_callback, my_arg);
2765 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2766 /* For the meaning of a, b in calls to my_callback(), see the
2767 * documentation of the function that calls the callback.
2768 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2769 * my_callback should return 1 if it wants BN_is_prime_ex()
2770 * to continue, or 0 to stop.
2775 *) Change the ZLIB compression method to be stateful, and make it
2776 available to TLS with the number defined in
2777 draft-ietf-tls-compression-04.txt.
2780 *) Add the ASN.1 structures and functions for CertificatePair, which
2781 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2783 CertificatePair ::= SEQUENCE {
2784 forward [0] Certificate OPTIONAL,
2785 reverse [1] Certificate OPTIONAL,
2786 -- at least one of the pair shall be present -- }
2788 Also implement the PEM functions to read and write certificate
2789 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2791 This needed to be defined, mostly for the sake of the LDAP
2792 attribute crossCertificatePair, but may prove useful elsewhere as
2796 *) Make it possible to inhibit symlinking of shared libraries in
2797 Makefile.shared, for Cygwin's sake.
2800 *) Extend the BIGNUM API by creating a function
2801 void BN_set_negative(BIGNUM *a, int neg);
2802 and a macro that behave like
2803 int BN_is_negative(const BIGNUM *a);
2805 to avoid the need to access 'a->neg' directly in applications.
2808 *) Implement fast modular reduction for pseudo-Mersenne primes
2809 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2810 EC_GROUP_new_curve_GFp() will now automatically use this
2812 [Nils Larsch <nla@trustcenter.de>]
2814 *) Add new lock type (CRYPTO_LOCK_BN).
2817 *) Change the ENGINE framework to automatically load engines
2818 dynamically from specific directories unless they could be
2819 found to already be built in or loaded. Move all the
2820 current engines except for the cryptodev one to a new
2822 The engines in engines/ are built as shared libraries if
2823 the "shared" options was given to ./Configure or ./config.
2824 Otherwise, they are inserted in libcrypto.a.
2825 /usr/local/ssl/engines is the default directory for dynamic
2826 engines, but that can be overriden at configure time through
2827 the usual use of --prefix and/or --openssldir, and at run
2828 time with the environment variable OPENSSL_ENGINES.
2829 [Geoff Thorpe and Richard Levitte]
2831 *) Add Makefile.shared, a helper makefile to build shared
2832 libraries. Addapt Makefile.org.
2835 *) Add version info to Win32 DLLs.
2836 [Peter 'Luna' Runestig" <peter@runestig.com>]
2838 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2839 can be added using this API to created arbitrary PKCS#12
2840 files while avoiding the low level API.
2842 New options to PKCS12_create(), key or cert can be NULL and
2843 will then be omitted from the output file. The encryption
2844 algorithm NIDs can be set to -1 for no encryption, the mac
2845 iteration count can be set to 0 to omit the mac.
2847 Enhance pkcs12 utility by making the -nokeys and -nocerts
2848 options work when creating a PKCS#12 file. New option -nomac
2849 to omit the mac, NONE can be set for an encryption algorithm.
2850 New code is modified to use the enhanced PKCS12_create()
2851 instead of the low level API.
2854 *) Extend ASN1 encoder to support indefinite length constructed
2855 encoding. This can output sequences tags and octet strings in
2856 this form. Modify pk7_asn1.c to support indefinite length
2857 encoding. This is experimental and needs additional code to
2858 be useful, such as an ASN1 bio and some enhanced streaming
2861 Extend template encode functionality so that tagging is passed
2862 down to the template encoder.
2865 *) Let 'openssl req' fail if an argument to '-newkey' is not
2866 recognized instead of using RSA as a default.
2869 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2870 As these are not official, they are not included in "ALL";
2871 the "ECCdraft" ciphersuite group alias can be used to select them.
2872 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2874 *) Add ECDH engine support.
2875 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2877 *) Add ECDH in new directory crypto/ecdh/.
2878 [Douglas Stebila (Sun Microsystems Laboratories)]
2880 *) Let BN_rand_range() abort with an error after 100 iterations
2881 without success (which indicates a broken PRNG).
2884 *) Change BN_mod_sqrt() so that it verifies that the input value
2885 is really the square of the return value. (Previously,
2886 BN_mod_sqrt would show GIGO behaviour.)
2889 *) Add named elliptic curves over binary fields from X9.62, SECG,
2890 and WAP/WTLS; add OIDs that were still missing.
2892 [Sheueling Chang Shantz and Douglas Stebila
2893 (Sun Microsystems Laboratories)]
2895 *) Extend the EC library for elliptic curves over binary fields
2896 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2899 EC_GF2m_simple_method
2903 EC_GROUP_new_curve_GF2m
2904 EC_GROUP_set_curve_GF2m
2905 EC_GROUP_get_curve_GF2m
2906 EC_POINT_set_affine_coordinates_GF2m
2907 EC_POINT_get_affine_coordinates_GF2m
2908 EC_POINT_set_compressed_coordinates_GF2m
2910 Point compression for binary fields is disabled by default for
2911 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2914 As binary polynomials are represented as BIGNUMs, various members
2915 of the EC_GROUP and EC_POINT data structures can be shared
2916 between the implementations for prime fields and binary fields;
2917 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2918 are essentially identical to their ..._GFp counterparts.
2919 (For simplicity, the '..._GFp' prefix has been dropped from
2920 various internal method names.)
2922 An internal 'field_div' method (similar to 'field_mul' and
2923 'field_sqr') has been added; this is used only for binary fields.
2925 [Sheueling Chang Shantz and Douglas Stebila
2926 (Sun Microsystems Laboratories)]
2928 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2929 through methods ('mul', 'precompute_mult').
2931 The generic implementations (now internally called 'ec_wNAF_mul'
2932 and 'ec_wNAF_precomputed_mult') remain the default if these
2933 methods are undefined.
2935 [Sheueling Chang Shantz and Douglas Stebila
2936 (Sun Microsystems Laboratories)]
2938 *) New function EC_GROUP_get_degree, which is defined through
2939 EC_METHOD. For curves over prime fields, this returns the bit
2940 length of the modulus.
2942 [Sheueling Chang Shantz and Douglas Stebila
2943 (Sun Microsystems Laboratories)]
2945 *) New functions EC_GROUP_dup, EC_POINT_dup.
2946 (These simply call ..._new and ..._copy).
2948 [Sheueling Chang Shantz and Douglas Stebila
2949 (Sun Microsystems Laboratories)]
2951 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2952 Polynomials are represented as BIGNUMs (where the sign bit is not
2953 used) in the following functions [macros]:
2956 BN_GF2m_sub [= BN_GF2m_add]
2957 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2958 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2959 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2961 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2962 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2963 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2964 BN_GF2m_cmp [= BN_ucmp]
2966 (Note that only the 'mod' functions are actually for fields GF(2^m).
2967 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2969 For some functions, an the irreducible polynomial defining a
2970 field can be given as an 'unsigned int[]' with strictly
2971 decreasing elements giving the indices of those bits that are set;
2972 i.e., p[] represents the polynomial
2973 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2975 p[0] > p[1] > ... > p[k] = 0.
2976 This applies to the following functions:
2981 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2982 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2984 BN_GF2m_mod_sqrt_arr
2985 BN_GF2m_mod_solve_quad_arr
2989 Conversion can be performed by the following functions:
2994 bntest.c has additional tests for binary polynomial arithmetic.
2996 Two implementations for BN_GF2m_mod_div() are available.
2997 The default algorithm simply uses BN_GF2m_mod_inv() and
2998 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2999 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3000 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3002 [Sheueling Chang Shantz and Douglas Stebila
3003 (Sun Microsystems Laboratories)]
3005 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3006 functionality is disabled at compile-time.
3007 [Douglas Stebila <douglas.stebila@sun.com>]
3009 *) Change default behaviour of 'openssl asn1parse' so that more
3010 information is visible when viewing, e.g., a certificate:
3012 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3013 mode the content of non-printable OCTET STRINGs is output in a
3014 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3015 avoid the appearance of a printable string.
3016 [Nils Larsch <nla@trustcenter.de>]
3018 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3020 EC_GROUP_set_asn1_flag()
3021 EC_GROUP_get_asn1_flag()
3022 EC_GROUP_set_point_conversion_form()
3023 EC_GROUP_get_point_conversion_form()
3024 These control ASN1 encoding details:
3025 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3026 has been set to OPENSSL_EC_NAMED_CURVE.
3027 - Points are encoded in uncompressed form by default; options for
3028 asn1_for are as for point2oct, namely
3029 POINT_CONVERSION_COMPRESSED
3030 POINT_CONVERSION_UNCOMPRESSED
3031 POINT_CONVERSION_HYBRID
3033 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3036 EC_GROUP_get0_seed()
3037 EC_GROUP_get_seed_len()
3038 This is used only for ASN1 purposes (so far).
3039 [Nils Larsch <nla@trustcenter.de>]
3041 *) Add 'field_type' member to EC_METHOD, which holds the NID
3042 of the appropriate field type OID. The new function
3043 EC_METHOD_get_field_type() returns this value.
3044 [Nils Larsch <nla@trustcenter.de>]
3049 EC_POINT_point2hex()
3050 EC_POINT_hex2point()
3051 providing useful interfaces to EC_POINT_point2oct() and
3052 EC_POINT_oct2point().
3053 [Nils Larsch <nla@trustcenter.de>]
3055 *) Change internals of the EC library so that the functions
3056 EC_GROUP_set_generator()
3057 EC_GROUP_get_generator()
3058 EC_GROUP_get_order()
3059 EC_GROUP_get_cofactor()
3060 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3061 to methods, which would lead to unnecessary code duplication when
3062 adding different types of curves.
3063 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3065 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3066 arithmetic, and such that modified wNAFs are generated
3067 (which avoid length expansion in many cases).
3070 *) Add a function EC_GROUP_check_discriminant() (defined via
3071 EC_METHOD) that verifies that the curve discriminant is non-zero.
3073 Add a function EC_GROUP_check() that makes some sanity tests
3074 on a EC_GROUP, its generator and order. This includes
3075 EC_GROUP_check_discriminant().
3076 [Nils Larsch <nla@trustcenter.de>]
3078 *) Add ECDSA in new directory crypto/ecdsa/.
3080 Add applications 'openssl ecparam' and 'openssl ecdsa'
3081 (these are based on 'openssl dsaparam' and 'openssl dsa').
3083 ECDSA support is also included in various other files across the
3084 library. Most notably,
3085 - 'openssl req' now has a '-newkey ecdsa:file' option;
3086 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3087 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3088 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3089 them suitable for ECDSA where domain parameters must be
3090 extracted before the specific public key;
3091 - ECDSA engine support has been added.
3092 [Nils Larsch <nla@trustcenter.de>]
3094 *) Include some named elliptic curves, and add OIDs from X9.62,
3095 SECG, and WAP/WTLS. Each curve can be obtained from the new
3097 EC_GROUP_new_by_curve_name(),
3098 and the list of available named curves can be obtained with
3099 EC_get_builtin_curves().
3100 Also add a 'curve_name' member to EC_GROUP objects, which can be
3102 EC_GROUP_set_curve_name()
3103 EC_GROUP_get_curve_name()
3104 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3106 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3107 was actually never needed) and in BN_mul(). The removal in BN_mul()
3108 required a small change in bn_mul_part_recursive() and the addition
3109 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3110 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3111 bn_sub_words() and bn_add_words() except they take arrays with
3115 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3117 *) Cleanse PEM buffers before freeing them since they may contain
3119 [Benjamin Bennett <ben@psc.edu>]
3121 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3122 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3123 authentication-only ciphersuites.
3126 *) Since AES128 and AES256 share a single mask bit in the logic of
3127 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3128 kludge to work properly if AES128 is available and AES256 isn't.
3131 *) Expand security boundary to match 1.1.1 module.
3134 *) Remove redundant features: hash file source, editing of test vectors
3135 modify fipsld to use external fips_premain.c signature.
3138 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3139 run algorithm test programs.
3142 *) Make algorithm test programs more tolerant of whitespace.
3145 *) Have SSL/TLS server implementation tolerate "mismatched" record
3146 protocol version while receiving ClientHello even if the
3147 ClientHello is fragmented. (The server can't insist on the
3148 particular protocol version it has chosen before the ServerHello
3149 message has informed the client about his choice.)
3152 *) Load error codes if they are not already present instead of using a
3153 static variable. This allows them to be cleanly unloaded and reloaded.
3156 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3158 *) Introduce limits to prevent malicious keys being able to
3159 cause a denial of service. (CVE-2006-2940)
3160 [Steve Henson, Bodo Moeller]
3162 *) Fix ASN.1 parsing of certain invalid structures that can result
3163 in a denial of service. (CVE-2006-2937) [Steve Henson]
3165 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3166 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3168 *) Fix SSL client code which could crash if connecting to a
3169 malicious SSLv2 server. (CVE-2006-4343)
3170 [Tavis Ormandy and Will Drewry, Google Security Team]
3172 *) Change ciphersuite string processing so that an explicit
3173 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3174 will no longer include "AES128-SHA"), and any other similar
3175 ciphersuite (same bitmap) from *other* protocol versions (so that
3176 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3177 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3178 changes from 0.9.8b and 0.9.8d.
3181 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3183 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3184 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3186 *) Change the Unix randomness entropy gathering to use poll() when
3187 possible instead of select(), since the latter has some
3188 undesirable limitations.
3189 [Darryl Miles via Richard Levitte and Bodo Moeller]
3191 *) Disable rogue ciphersuites:
3193 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3194 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3195 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3197 The latter two were purportedly from
3198 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3201 Also deactive the remaining ciphersuites from
3202 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3203 unofficial, and the ID has long expired.
3206 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3207 dual-core machines) and other potential thread-safety issues.
3210 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3212 *) Adapt fipsld and the build system to link against the validated FIPS
3213 module in FIPS mode.
3216 *) Fixes for VC++ 2005 build under Windows.
3219 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3220 from a Windows bash shell such as MSYS. It is autodetected from the
3221 "config" script when run from a VC++ environment. Modify standard VC++
3222 build to use fipscanister.o from the GNU make build.
3225 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3227 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3228 The value now differs depending on if you build for FIPS or not.
3229 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3230 safely run with a non-FIPSed libcrypto, as it may crash because of
3231 the difference induced by this change.
3234 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3236 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3237 (part of SSL_OP_ALL). This option used to disable the
3238 countermeasure against man-in-the-middle protocol-version
3239 rollback in the SSL 2.0 server implementation, which is a bad
3240 idea. (CVE-2005-2969)
3242 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3243 for Information Security, National Institute of Advanced Industrial
3244 Science and Technology [AIST], Japan)]
3246 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3247 mainly for FIPS compliance and not fully integrated at this stage.
3250 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3251 the exponentiation using a fixed-length exponent. (Otherwise,
3252 the information leaked through timing could expose the secret key
3253 after many signatures; cf. Bleichenbacher's attack on DSA with
3257 *) Make a new fixed-window mod_exp implementation the default for
3258 RSA, DSA, and DH private-key operations so that the sequence of
3259 squares and multiplies and the memory access pattern are
3260 independent of the particular secret key. This will mitigate
3261 cache-timing and potential related attacks.
3263 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3264 and this is automatically used by BN_mod_exp_mont() if the new flag
3265 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3266 will use this BN flag for private exponents unless the flag
3267 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3268 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3270 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3272 *) Change the client implementation for SSLv23_method() and
3273 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3274 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3275 (Previously, the SSL 2.0 backwards compatible Client Hello
3276 message format would be used even with SSL_OP_NO_SSLv2.)
3279 *) Add support for smime-type MIME parameter in S/MIME messages which some
3283 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3284 a threadsafe manner. Modify rsa code to use new function and add calls
3285 to dsa and dh code (which had race conditions before).
3288 *) Include the fixed error library code in the C error file definitions
3289 instead of fixing them up at runtime. This keeps the error code
3290 structures constant.
3293 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3295 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3298 *) Fixes for newer kerberos headers. NB: the casts are needed because
3299 the 'length' field is signed on one version and unsigned on another
3300 with no (?) obvious way to tell the difference, without these VC++
3301 complains. Also the "definition" of FAR (blank) is no longer included
3302 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3303 some needed definitions.
3306 *) Undo Cygwin change.
3309 *) Added support for proxy certificates according to RFC 3820.
3310 Because they may be a security thread to unaware applications,
3311 they must be explicitely allowed in run-time. See
3312 docs/HOWTO/proxy_certificates.txt for further information.
3315 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3317 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3318 server and client random values. Previously
3319 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3320 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3322 This change has negligible security impact because:
3324 1. Server and client random values still have 24 bytes of pseudo random
3327 2. Server and client random values are sent in the clear in the initial
3330 3. The master secret is derived using the premaster secret (48 bytes in
3331 size for static RSA ciphersuites) as well as client server and random
3334 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3337 [Stephen Henson, reported by UK NISCC]
3339 *) Use Windows randomness collection on Cygwin.
3342 *) Fix hang in EGD/PRNGD query when communication socket is closed
3343 prematurely by EGD/PRNGD.
3344 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3346 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3349 *) Back-port of selected performance improvements from development
3350 branch, as well as improved support for PowerPC platforms.
3353 *) Add lots of checks for memory allocation failure, error codes to indicate
3354 failure and freeing up memory if a failure occurs.
3355 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3357 *) Add new -passin argument to dgst.
3360 *) Perform some character comparisons of different types in X509_NAME_cmp:
3361 this is needed for some certificates that reencode DNs into UTF8Strings
3362 (in violation of RFC3280) and can't or wont issue name rollover
3366 *) Make an explicit check during certificate validation to see that
3367 the CA setting in each certificate on the chain is correct. As a
3368 side effect always do the following basic checks on extensions,
3369 not just when there's an associated purpose to the check:
3371 - if there is an unhandled critical extension (unless the user
3372 has chosen to ignore this fault)
3373 - if the path length has been exceeded (if one is set at all)
3374 - that certain extensions fit the associated purpose (if one has
3378 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3380 *) Avoid a race condition when CRLs are checked in a multi threaded
3381 environment. This would happen due to the reordering of the revoked
3382 entries during signature checking and serial number lookup. Now the
3383 encoding is cached and the serial number sort performed under a lock.
3384 Add new STACK function sk_is_sorted().
3387 *) Add Delta CRL to the extension code.
3390 *) Various fixes to s3_pkt.c so alerts are sent properly.
3391 [David Holmes <d.holmes@f5.com>]
3393 *) Reduce the chances of duplicate issuer name and serial numbers (in
3394 violation of RFC3280) using the OpenSSL certificate creation utilities.
3395 This is done by creating a random 64 bit value for the initial serial
3396 number when a serial number file is created or when a self signed
3397 certificate is created using 'openssl req -x509'. The initial serial
3398 number file is created using 'openssl x509 -next_serial' in CA.pl
3399 rather than being initialized to 1.
3402 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3404 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3405 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3406 [Joe Orton, Steve Henson]
3408 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3410 [Joe Orton, Steve Henson]
3412 *) Make it possible to have multiple active certificates with the same
3413 subject in the CA index file. This is done only if the keyword
3414 'unique_subject' is set to 'no' in the main CA section (default
3415 if 'CA_default') of the configuration file. The value is saved
3416 with the database itself in a separate index attribute file,
3417 named like the index file with '.attr' appended to the name.
3420 *) X509 verify fixes. Disable broken certificate workarounds when
3421 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3422 keyUsage extension present. Don't accept CRLs with unhandled critical
3423 extensions: since verify currently doesn't process CRL extensions this
3424 rejects a CRL with *any* critical extensions. Add new verify error codes
3428 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3429 A clarification of RFC2560 will require the use of OCTET STRINGs and
3430 some implementations cannot handle the current raw format. Since OpenSSL
3431 copies and compares OCSP nonces as opaque blobs without any attempt at
3432 parsing them this should not create any compatibility issues.
3435 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3436 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3437 this HMAC (and other) operations are several times slower than OpenSSL
3441 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3442 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3444 *) Use the correct content when signing type "other".
3447 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3449 *) Fix various bugs revealed by running the NISCC test suite:
3451 Stop out of bounds reads in the ASN1 code when presented with
3452 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3454 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3456 If verify callback ignores invalid public key errors don't try to check
3457 certificate signature with the NULL public key.
3461 *) New -ignore_err option in ocsp application to stop the server
3462 exiting on the first error in a request.
3465 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3466 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3470 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3471 extra data after the compression methods not only for TLS 1.0
3472 but also for SSL 3.0 (as required by the specification).
3473 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3475 *) Change X509_certificate_type() to mark the key as exported/exportable
3476 when it's 512 *bits* long, not 512 bytes.
3479 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3480 blocks during encryption.
3483 *) Various fixes to base64 BIO and non blocking I/O. On write
3484 flushes were not handled properly if the BIO retried. On read
3485 data was not being buffered properly and had various logic bugs.
3486 This also affects blocking I/O when the data being decoded is a
3490 *) Various S/MIME bugfixes and compatibility changes:
3491 output correct application/pkcs7 MIME type if
3492 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3493 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3494 of files as .eml work). Correctly handle very long lines in MIME
3498 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3500 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3501 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3502 a protocol version number mismatch like a decryption error
3503 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3506 *) Turn on RSA blinding by default in the default implementation
3507 to avoid a timing attack. Applications that don't want it can call
3508 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3509 They would be ill-advised to do so in most cases.
3510 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3512 *) Change RSA blinding code so that it works when the PRNG is not
3513 seeded (in this case, the secret RSA exponent is abused as
3514 an unpredictable seed -- if it is not unpredictable, there
3515 is no point in blinding anyway). Make RSA blinding thread-safe
3516 by remembering the creator's thread ID in rsa->blinding and
3517 having all other threads use local one-time blinding factors
3518 (this requires more computation than sharing rsa->blinding, but
3519 avoids excessive locking; and if an RSA object is not shared
3520 between threads, blinding will still be very fast).
3523 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3524 ENGINE as defaults for all supported algorithms irrespective of
3525 the 'flags' parameter. 'flags' is now honoured, so applications
3526 should make sure they are passing it correctly.
3529 *) Target "mingw" now allows native Windows code to be generated in
3530 the Cygwin environment as well as with the MinGW compiler.
3533 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3535 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3536 via timing by performing a MAC computation even if incorrrect
3537 block cipher padding has been found. This is a countermeasure
3538 against active attacks where the attacker has to distinguish
3539 between bad padding and a MAC verification error. (CVE-2003-0078)
3541 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3542 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3543 Martin Vuagnoux (EPFL, Ilion)]
3545 *) Make the no-err option work as intended. The intention with no-err
3546 is not to have the whole error stack handling routines removed from
3547 libcrypto, it's only intended to remove all the function name and
3548 reason texts, thereby removing some of the footprint that may not
3549 be interesting if those errors aren't displayed anyway.
3551 NOTE: it's still possible for any application or module to have it's
3552 own set of error texts inserted. The routines are there, just not
3553 used by default when no-err is given.
3556 *) Add support for FreeBSD on IA64.
3557 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3559 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3560 Kerberos function mit_des_cbc_cksum(). Before this change,
3561 the value returned by DES_cbc_cksum() was like the one from
3562 mit_des_cbc_cksum(), except the bytes were swapped.
3563 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3565 *) Allow an application to disable the automatic SSL chain building.
3566 Before this a rather primitive chain build was always performed in
3567 ssl3_output_cert_chain(): an application had no way to send the
3568 correct chain if the automatic operation produced an incorrect result.
3570 Now the chain builder is disabled if either:
3572 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3574 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3576 The reasoning behind this is that an application would not want the
3577 auto chain building to take place if extra chain certificates are
3578 present and it might also want a means of sending no additional
3579 certificates (for example the chain has two certificates and the
3583 *) Add the possibility to build without the ENGINE framework.
3584 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3586 *) Under Win32 gmtime() can return NULL: check return value in
3587 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3590 *) DSA routines: under certain error conditions uninitialized BN objects
3591 could be freed. Solution: make sure initialization is performed early
3592 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3593 Nils Larsch <nla@trustcenter.de> via PR#459)
3596 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3597 checked on reconnect on the client side, therefore session resumption
3598 could still fail with a "ssl session id is different" error. This
3599 behaviour is masked when SSL_OP_ALL is used due to
3600 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3601 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3602 followup to PR #377.
3605 *) IA-32 assembler support enhancements: unified ELF targets, support
3606 for SCO/Caldera platforms, fix for Cygwin shared build.
3609 *) Add support for FreeBSD on sparc64. As a consequence, support for
3610 FreeBSD on non-x86 processors is separate from x86 processors on
3611 the config script, much like the NetBSD support.
3612 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3614 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3616 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3619 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3620 code (06) was taken as the first octet of the session ID and the last
3621 octet was ignored consequently. As a result SSLv2 client side session
3622 caching could not have worked due to the session ID mismatch between
3624 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3628 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3629 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3633 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3634 seems that in spite of existing for more than a year, many application
3635 author have done nothing to provide the necessary callbacks, which
3636 means that this particular engine will not work properly anywhere.
3637 This is a very unfortunate situation which forces us, in the name
3638 of usability, to give the hw_ncipher.c a static lock, which is part
3640 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3641 appear in 0.9.8 or later. We EXPECT application authors to have
3642 dealt properly with this when 0.9.8 is released (unless we actually
3643 make such changes in the libcrypto locking code that changes will
3644 have to be made anyway).
3647 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3648 octets have been read, EOF or an error occurs. Without this change
3649 some truncated ASN1 structures will not produce an error.
3652 *) Disable Heimdal support, since it hasn't been fully implemented.
3653 Still give the possibility to force the use of Heimdal, but with
3654 warnings and a request that patches get sent to openssl-dev.
3657 *) Add the VC-CE target, introduce the WINCE sysname, and add
3658 INSTALL.WCE and appropriate conditionals to make it build.
3659 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3661 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3662 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3663 edit numbers of the version.
3664 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3666 *) Introduce safe string copy and catenation functions
3667 (BUF_strlcpy() and BUF_strlcat()).
3668 [Ben Laurie (CHATS) and Richard Levitte]
3670 *) Avoid using fixed-size buffers for one-line DNs.
3671 [Ben Laurie (CHATS)]
3673 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3674 resizing buffers containing secrets, and use where appropriate.
3675 [Ben Laurie (CHATS)]
3677 *) Avoid using fixed size buffers for configuration file location.
3678 [Ben Laurie (CHATS)]
3680 *) Avoid filename truncation for various CA files.
3681 [Ben Laurie (CHATS)]
3683 *) Use sizeof in preference to magic numbers.
3684 [Ben Laurie (CHATS)]
3686 *) Avoid filename truncation in cert requests.
3687 [Ben Laurie (CHATS)]
3689 *) Add assertions to check for (supposedly impossible) buffer
3691 [Ben Laurie (CHATS)]
3693 *) Don't cache truncated DNS entries in the local cache (this could
3694 potentially lead to a spoofing attack).
3695 [Ben Laurie (CHATS)]
3697 *) Fix various buffers to be large enough for hex/decimal
3698 representations in a platform independent manner.
3699 [Ben Laurie (CHATS)]
3701 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3702 resizing buffers containing secrets, and use where appropriate.
3703 [Ben Laurie (CHATS)]
3705 *) Add BIO_indent() to avoid much slightly worrying code to do
3707 [Ben Laurie (CHATS)]
3709 *) Convert sprintf()/BIO_puts() to BIO_printf().
3710 [Ben Laurie (CHATS)]
3712 *) buffer_gets() could terminate with the buffer only half
3714 [Ben Laurie (CHATS)]
3716 *) Add assertions to prevent user-supplied crypto functions from
3717 overflowing internal buffers by having large block sizes, etc.
3718 [Ben Laurie (CHATS)]
3720 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3722 [Ben Laurie (CHATS)]
3724 *) Eliminate unused copy of key in RC4.
3725 [Ben Laurie (CHATS)]
3727 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3728 [Ben Laurie (CHATS)]
3730 *) Fix off-by-one error in EGD path.
3731 [Ben Laurie (CHATS)]
3733 *) If RANDFILE path is too long, ignore instead of truncating.
3734 [Ben Laurie (CHATS)]
3736 *) Eliminate unused and incorrectly sized X.509 structure
3738 [Ben Laurie (CHATS)]
3740 *) Eliminate unused and dangerous function knumber().
3741 [Ben Laurie (CHATS)]
3743 *) Eliminate unused and dangerous structure, KSSL_ERR.
3744 [Ben Laurie (CHATS)]
3746 *) Protect against overlong session ID context length in an encoded
3747 session object. Since these are local, this does not appear to be
3749 [Ben Laurie (CHATS)]
3751 *) Change from security patch (see 0.9.6e below) that did not affect
3752 the 0.9.6 release series:
3754 Remote buffer overflow in SSL3 protocol - an attacker could
3755 supply an oversized master key in Kerberos-enabled versions.
3757 [Ben Laurie (CHATS)]
3759 *) Change the SSL kerb5 codes to match RFC 2712.
3762 *) Make -nameopt work fully for req and add -reqopt switch.
3763 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3765 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3766 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3768 *) Make sure tests can be performed even if the corresponding algorithms
3769 have been removed entirely. This was also the last step to make
3770 OpenSSL compilable with DJGPP under all reasonable conditions.
3771 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3773 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3774 to allow version independent disabling of normally unselected ciphers,
3775 which may be activated as a side-effect of selecting a single cipher.
3777 (E.g., cipher list string "RSA" enables ciphersuites that are left
3778 out of "ALL" because they do not provide symmetric encryption.
3779 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3780 [Lutz Jaenicke, Bodo Moeller]
3782 *) Add appropriate support for separate platform-dependent build
3783 directories. The recommended way to make a platform-dependent
3784 build directory is the following (tested on Linux), maybe with
3787 # Place yourself outside of the OpenSSL source tree. In
3788 # this example, the environment variable OPENSSL_SOURCE
3789 # is assumed to contain the absolute OpenSSL source directory.
3790 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3791 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3792 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3793 mkdir -p `dirname $F`
3794 ln -s $OPENSSL_SOURCE/$F $F
3797 To be absolutely sure not to disturb the source tree, a "make clean"
3798 is a good thing. If it isn't successfull, don't worry about it,
3799 it probably means the source directory is very clean.
3802 *) Make sure any ENGINE control commands make local copies of string
3803 pointers passed to them whenever necessary. Otherwise it is possible
3804 the caller may have overwritten (or deallocated) the original string
3805 data when a later ENGINE operation tries to use the stored values.
3806 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3808 *) Improve diagnostics in file reading and command-line digests.
3809 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3811 *) Add AES modes CFB and OFB to the object database. Correct an
3812 error in AES-CFB decryption.
3815 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3816 allows existing EVP_CIPHER_CTX structures to be reused after
3817 calling EVP_*Final(). This behaviour is used by encryption
3818 BIOs and some applications. This has the side effect that
3819 applications must explicitly clean up cipher contexts with
3820 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3823 *) Check the values of dna and dnb in bn_mul_recursive before calling
3824 bn_mul_comba (a non zero value means the a or b arrays do not contain
3825 n2 elements) and fallback to bn_mul_normal if either is not zero.
3828 *) Fix escaping of non-ASCII characters when using the -subj option
3829 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3832 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3833 form for "surname", serialNumber has no short form.
3834 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3835 therefore remove "mail" short name for "internet 7".
3836 The OID for unique identifiers in X509 certificates is
3837 x500UniqueIdentifier, not uniqueIdentifier.
3838 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3841 *) Add an "init" command to the ENGINE config module and auto initialize
3842 ENGINEs. Without any "init" command the ENGINE will be initialized
3843 after all ctrl commands have been executed on it. If init=1 the
3844 ENGINE is initailized at that point (ctrls before that point are run
3845 on the uninitialized ENGINE and after on the initialized one). If
3846 init=0 then the ENGINE will not be iniatialized at all.
3849 *) Fix the 'app_verify_callback' interface so that the user-defined
3850 argument is actually passed to the callback: In the
3851 SSL_CTX_set_cert_verify_callback() prototype, the callback
3852 declaration has been changed from
3855 int (*cb)(X509_STORE_CTX *,void *);
3856 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3857 i=s->ctx->app_verify_callback(&ctx)
3858 has been changed into
3859 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3861 To update applications using SSL_CTX_set_cert_verify_callback(),
3862 a dummy argument can be added to their callback functions.
3863 [D. K. Smetters <smetters@parc.xerox.com>]
3865 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3866 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3868 *) Add and OPENSSL_LOAD_CONF define which will cause
3869 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3870 This allows older applications to transparently support certain
3871 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3872 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3873 load the config file and OPENSSL_add_all_algorithms_conf() which will
3874 always load it have also been added.
3877 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3878 Adjust NIDs and EVP layer.
3879 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3881 *) Config modules support in openssl utility.
3883 Most commands now load modules from the config file,
3884 though in a few (such as version) this isn't done
3885 because it couldn't be used for anything.
3887 In the case of ca and req the config file used is
3888 the same as the utility itself: that is the -config
3889 command line option can be used to specify an
3893 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3894 use "openssl_conf" if filename is NULL use default openssl config file.
3897 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3898 config section name. Add a new flag to tolerate a missing config file
3899 and move code to CONF_modules_load_file().
3902 *) Support for crypto accelerator cards from Accelerated Encryption
3903 Processing, www.aep.ie. (Use engine 'aep')
3904 The support was copied from 0.9.6c [engine] and adapted/corrected
3905 to work with the new engine framework.
3906 [AEP Inc. and Richard Levitte]
3908 *) Support for SureWare crypto accelerator cards from Baltimore
3909 Technologies. (Use engine 'sureware')
3910 The support was copied from 0.9.6c [engine] and adapted
3911 to work with the new engine framework.
3914 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3915 make the newer ENGINE framework commands for the CHIL engine work.
3916 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3918 *) Make it possible to produce shared libraries on ReliantUNIX.
3919 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3921 *) Add the configuration target debug-linux-ppro.
3922 Make 'openssl rsa' use the general key loading routines
3923 implemented in apps.c, and make those routines able to
3924 handle the key format FORMAT_NETSCAPE and the variant
3926 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3928 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3929 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3931 *) Add -keyform to rsautl, and document -engine.
3932 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3934 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3935 BIO_R_NO_SUCH_FILE error code rather than the generic
3936 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3939 *) Add new functions
3941 ERR_peek_last_error_line
3942 ERR_peek_last_error_line_data.
3943 These are similar to
3946 ERR_peek_error_line_data,
3947 but report on the latest error recorded rather than the first one
3948 still in the error queue.
3949 [Ben Laurie, Bodo Moeller]
3951 *) default_algorithms option in ENGINE config module. This allows things
3953 default_algorithms = ALL
3954 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3957 *) Prelminary ENGINE config module.
3960 *) New experimental application configuration code.
3963 *) Change the AES code to follow the same name structure as all other
3964 symmetric ciphers, and behave the same way. Move everything to
3965 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3966 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3968 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3969 [Ben Laurie and Theo de Raadt]
3971 *) Add option to output public keys in req command.
3972 [Massimiliano Pala madwolf@openca.org]
3974 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3975 (up to about 10% better than before for P-192 and P-224).
3978 *) New functions/macros
3980 SSL_CTX_set_msg_callback(ctx, cb)
3981 SSL_CTX_set_msg_callback_arg(ctx, arg)
3982 SSL_set_msg_callback(ssl, cb)
3983 SSL_set_msg_callback_arg(ssl, arg)
3985 to request calling a callback function
3987 void cb(int write_p, int version, int content_type,
3988 const void *buf, size_t len, SSL *ssl, void *arg)
3990 whenever a protocol message has been completely received
3991 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3992 protocol version according to which the SSL library interprets
3993 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3994 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3995 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3996 specification (change_cipher_spec(20), alert(21), handshake(22)).
3997 'buf' and 'len' point to the actual message, 'ssl' to the
3998 SSL object, and 'arg' is the application-defined value set by
3999 SSL[_CTX]_set_msg_callback_arg().
4001 'openssl s_client' and 'openssl s_server' have new '-msg' options
4002 to enable a callback that displays all protocol messages.
4005 *) Change the shared library support so shared libraries are built as
4006 soon as the corresponding static library is finished, and thereby get
4007 openssl and the test programs linked against the shared library.
4008 This still only happens when the keyword "shard" has been given to
4009 the configuration scripts.
4011 NOTE: shared library support is still an experimental thing, and
4012 backward binary compatibility is still not guaranteed.
4013 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4015 *) Add support for Subject Information Access extension.
4016 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4018 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4019 additional bytes when new memory had to be allocated, not just
4020 when reusing an existing buffer.
4023 *) New command line and configuration option 'utf8' for the req command.
4024 This allows field values to be specified as UTF8 strings.
4027 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4028 runs for the former and machine-readable output for the latter.
4031 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4032 of the e-mail address in the DN (i.e., it will go into a certificate
4033 extension only). The new configuration file option 'email_in_dn = no'
4034 has the same effect.
4035 [Massimiliano Pala madwolf@openca.org]
4037 *) Change all functions with names starting with des_ to be starting
4038 with DES_ instead. Add wrappers that are compatible with libdes,
4039 but are named _ossl_old_des_*. Finally, add macros that map the
4040 des_* symbols to the corresponding _ossl_old_des_* if libdes
4041 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4042 desired, the des_* symbols will be mapped to DES_*, with one
4045 Since we provide two compatibility mappings, the user needs to
4046 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4047 compatibility is desired. The default (i.e., when that macro
4048 isn't defined) is OpenSSL 0.9.6c compatibility.
4050 There are also macros that enable and disable the support of old
4051 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4052 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4053 are defined, the default will apply: to support the old des routines.
4055 In either case, one must include openssl/des.h to get the correct
4056 definitions. Do not try to just include openssl/des_old.h, that
4059 NOTE: This is a major break of an old API into a new one. Software
4060 authors are encouraged to switch to the DES_ style functions. Some
4061 time in the future, des_old.h and the libdes compatibility functions
4062 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4063 default), and then completely removed.
4066 *) Test for certificates which contain unsupported critical extensions.
4067 If such a certificate is found during a verify operation it is
4068 rejected by default: this behaviour can be overridden by either
4069 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4070 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4071 X509_supported_extension() has also been added which returns 1 if a
4072 particular extension is supported.
4075 *) Modify the behaviour of EVP cipher functions in similar way to digests
4076 to retain compatibility with existing code.
4079 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4080 compatibility with existing code. In particular the 'ctx' parameter does
4081 not have to be to be initialized before the call to EVP_DigestInit() and
4082 it is tidied up after a call to EVP_DigestFinal(). New function
4083 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4084 EVP_MD_CTX_copy() changed to not require the destination to be
4085 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4086 requires the destination to be valid.
4088 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4089 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4092 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4093 so that complete 'Handshake' protocol structures are kept in memory
4094 instead of overwriting 'msg_type' and 'length' with 'body' data.
4097 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4098 [Massimo Santin via Richard Levitte]
4100 *) Major restructuring to the underlying ENGINE code. This includes
4101 reduction of linker bloat, separation of pure "ENGINE" manipulation
4102 (initialisation, etc) from functionality dealing with implementations
4103 of specific crypto iterfaces. This change also introduces integrated
4104 support for symmetric ciphers and digest implementations - so ENGINEs
4105 can now accelerate these by providing EVP_CIPHER and EVP_MD
4106 implementations of their own. This is detailed in crypto/engine/README
4107 as it couldn't be adequately described here. However, there are a few
4108 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4109 were changed in the original introduction of ENGINE code have now
4110 reverted back - the hooking from this code to ENGINE is now a good
4111 deal more passive and at run-time, operations deal directly with
4112 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4113 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4114 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4115 they were not being used by the framework as there is no concept of a
4116 BIGNUM_METHOD and they could not be generalised to the new
4117 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4118 ENGINE_cpy() has been removed as it cannot be consistently defined in
4122 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4125 *) Change mkdef.pl to sort symbols that get the same entry number,
4126 and make sure the automatically generated functions ERR_load_*
4127 become part of libeay.num as well.
4130 *) New function SSL_renegotiate_pending(). This returns true once
4131 renegotiation has been requested (either SSL_renegotiate() call
4132 or HelloRequest/ClientHello receveived from the peer) and becomes
4133 false once a handshake has been completed.
4134 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4135 sends a HelloRequest, but does not ensure that a handshake takes
4136 place. SSL_renegotiate_pending() is useful for checking if the
4137 client has followed the request.)
4140 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4141 By default, clients may request session resumption even during
4142 renegotiation (if session ID contexts permit); with this option,
4143 session resumption is possible only in the first handshake.
4145 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4146 more bits available for options that should not be part of
4147 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4150 *) Add some demos for certificate and certificate request creation.
4153 *) Make maximum certificate chain size accepted from the peer application
4154 settable (SSL*_get/set_max_cert_list()), as proposed by
4155 "Douglas E. Engert" <deengert@anl.gov>.
4158 *) Add support for shared libraries for Unixware-7
4159 (Boyd Lynn Gerber <gerberb@zenez.com>).
4162 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4163 be done prior to destruction. Use this to unload error strings from
4164 ENGINEs that load their own error strings. NB: This adds two new API
4165 functions to "get" and "set" this destroy handler in an ENGINE.
4168 *) Alter all existing ENGINE implementations (except "openssl" and
4169 "openbsd") to dynamically instantiate their own error strings. This
4170 makes them more flexible to be built both as statically-linked ENGINEs
4171 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4172 Also, add stub code to each that makes building them as self-contained
4173 shared-libraries easier (see README.ENGINE).
4176 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4177 implementations into applications that are completely implemented in
4178 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4179 commands that can be used to configure what shared-library to load and
4180 to control aspects of the way it is handled. Also, made an update to
4181 the README.ENGINE file that brings its information up-to-date and
4182 provides some information and instructions on the "dynamic" ENGINE
4183 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4186 *) Make it possible to unload ranges of ERR strings with a new
4187 "ERR_unload_strings" function.
4190 *) Add a copy() function to EVP_MD.
4193 *) Make EVP_MD routines take a context pointer instead of just the
4194 md_data void pointer.
4197 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4198 that the digest can only process a single chunk of data
4199 (typically because it is provided by a piece of
4200 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4201 is only going to provide a single chunk of data, and hence the
4202 framework needn't accumulate the data for oneshot drivers.
4205 *) As with "ERR", make it possible to replace the underlying "ex_data"
4206 functions. This change also alters the storage and management of global
4207 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4208 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4209 index counters. The API functions that use this state have been changed
4210 to take a "class_index" rather than pointers to the class's local STACK
4211 and counter, and there is now an API function to dynamically create new
4212 classes. This centralisation allows us to (a) plug a lot of the
4213 thread-safety problems that existed, and (b) makes it possible to clean
4214 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4215 such data would previously have always leaked in application code and
4216 workarounds were in place to make the memory debugging turn a blind eye
4217 to it. Application code that doesn't use this new function will still
4218 leak as before, but their memory debugging output will announce it now
4219 rather than letting it slide.
4221 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4222 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4223 has a return value to indicate success or failure.
4226 *) Make it possible to replace the underlying "ERR" functions such that the
4227 global state (2 LHASH tables and 2 locks) is only used by the "default"
4228 implementation. This change also adds two functions to "get" and "set"
4229 the implementation prior to it being automatically set the first time
4230 any other ERR function takes place. Ie. an application can call "get",
4231 pass the return value to a module it has just loaded, and that module
4232 can call its own "set" function using that value. This means the
4233 module's "ERR" operations will use (and modify) the error state in the
4234 application and not in its own statically linked copy of OpenSSL code.
4237 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4238 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4239 the operation, and provides a more encapsulated way for external code
4240 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4241 to use these functions rather than manually incrementing the counts.
4243 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4246 *) Add EVP test program.
4249 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4252 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4253 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4254 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4255 These allow a CRL to be built without having to access X509_CRL fields
4256 directly. Modify 'ca' application to use new functions.
4259 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4260 bug workarounds. Rollback attack detection is a security feature.
4261 The problem will only arise on OpenSSL servers when TLSv1 is not
4262 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4263 Software authors not wanting to support TLSv1 will have special reasons
4264 for their choice and can explicitly enable this option.
4265 [Bodo Moeller, Lutz Jaenicke]
4267 *) Rationalise EVP so it can be extended: don't include a union of
4268 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4269 (similar to those existing for EVP_CIPHER_CTX).
4274 EVP_MD_CTX_init(&md); /* new function call */
4275 EVP_DigestInit(&md, EVP_sha1());
4276 EVP_DigestUpdate(&md, in, len);
4277 EVP_DigestFinal(&md, out, NULL);
4278 EVP_MD_CTX_cleanup(&md); /* new function call */
4282 *) Make DES key schedule conform to the usual scheme, as well as
4283 correcting its structure. This means that calls to DES functions
4284 now have to pass a pointer to a des_key_schedule instead of a
4285 plain des_key_schedule (which was actually always a pointer
4288 des_key_schedule ks;
4290 des_set_key_checked(..., &ks);
4291 des_ncbc_encrypt(..., &ks, ...);
4293 (Note that a later change renames 'des_...' into 'DES_...'.)
4296 *) Initial reduction of linker bloat: the use of some functions, such as
4297 PEM causes large amounts of unused functions to be linked in due to
4298 poor organisation. For example pem_all.c contains every PEM function
4299 which has a knock on effect of linking in large amounts of (unused)
4300 ASN1 code. Grouping together similar functions and splitting unrelated
4301 functions prevents this.
4304 *) Cleanup of EVP macros.
4307 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4308 correct _ecb suffix.
4311 *) Add initial OCSP responder support to ocsp application. The
4312 revocation information is handled using the text based index
4313 use by the ca application. The responder can either handle
4314 requests generated internally, supplied in files (for example
4315 via a CGI script) or using an internal minimal server.
4318 *) Add configuration choices to get zlib compression for TLS.
4321 *) Changes to Kerberos SSL for RFC 2712 compliance:
4322 1. Implemented real KerberosWrapper, instead of just using
4323 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4324 2. Implemented optional authenticator field of KerberosWrapper.
4326 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4327 and authenticator structs; see crypto/krb5/.
4329 Generalized Kerberos calls to support multiple Kerberos libraries.
4330 [Vern Staats <staatsvr@asc.hpc.mil>,
4331 Jeffrey Altman <jaltman@columbia.edu>
4332 via Richard Levitte]
4334 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4335 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4336 values for each of the key sizes rather than having just
4337 parameters (and 'speed' generating keys each time).
4340 *) Speed up EVP routines.
4343 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4344 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4345 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4346 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4348 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4349 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4350 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4353 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4355 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4358 *) Added the OS2-EMX target.
4359 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4361 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4362 to support NCONF routines in extension code. New function CONF_set_nconf()
4363 to allow functions which take an NCONF to also handle the old LHASH
4364 structure: this means that the old CONF compatible routines can be
4365 retained (in particular wrt extensions) without having to duplicate the
4366 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4369 *) Enhance the general user interface with mechanisms for inner control
4370 and with possibilities to have yes/no kind of prompts.
4373 *) Change all calls to low level digest routines in the library and
4374 applications to use EVP. Add missing calls to HMAC_cleanup() and
4375 don't assume HMAC_CTX can be copied using memcpy().
4376 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4378 *) Add the possibility to control engines through control names but with
4379 arbitrary arguments instead of just a string.
4380 Change the key loaders to take a UI_METHOD instead of a callback
4381 function pointer. NOTE: this breaks binary compatibility with earlier
4382 versions of OpenSSL [engine].
4383 Adapt the nCipher code for these new conditions and add a card insertion
4387 *) Enhance the general user interface with mechanisms to better support
4388 dialog box interfaces, application-defined prompts, the possibility
4389 to use defaults (for example default passwords from somewhere else)
4390 and interrupts/cancellations.
4393 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4394 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4397 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4398 tidy up some unnecessarily weird code in 'sk_new()').
4399 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4401 *) Change the key loading routines for ENGINEs to use the same kind
4402 callback (pem_password_cb) as all other routines that need this
4406 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4407 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4408 than this minimum value is recommended.
4411 *) New random seeder for OpenVMS, using the system process statistics
4412 that are easily reachable.
4415 *) Windows apparently can't transparently handle global
4416 variables defined in DLLs. Initialisations such as:
4418 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4420 wont compile. This is used by the any applications that need to
4421 declare their own ASN1 modules. This was fixed by adding the option
4422 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4423 needed for static libraries under Win32.
4426 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4427 setting of purpose and trust fields. New X509_STORE trust and
4428 purpose functions and tidy up setting in other SSL functions.
4431 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4432 structure. These are inherited by X509_STORE_CTX when it is
4433 initialised. This allows various defaults to be set in the
4434 X509_STORE structure (such as flags for CRL checking and custom
4435 purpose or trust settings) for functions which only use X509_STORE_CTX
4436 internally such as S/MIME.
4438 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4439 trust settings if they are not set in X509_STORE. This allows X509_STORE
4440 purposes and trust (in S/MIME for example) to override any set by default.
4442 Add command line options for CRL checking to smime, s_client and s_server
4446 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4447 are set then the CRL is looked up in the X509_STORE structure and
4448 its validity and signature checked, then if the certificate is found
4449 in the CRL the verify fails with a revoked error.
4451 Various new CRL related callbacks added to X509_STORE_CTX structure.
4453 Command line options added to 'verify' application to support this.
4455 This needs some additional work, such as being able to handle multiple
4456 CRLs with different times, extension based lookup (rather than just
4457 by subject name) and ultimately more complete V2 CRL extension
4461 *) Add a general user interface API (crypto/ui/). This is designed
4462 to replace things like des_read_password and friends (backward
4463 compatibility functions using this new API are provided).
4464 The purpose is to remove prompting functions from the DES code
4465 section as well as provide for prompting through dialog boxes in
4466 a window system and the like.
4469 *) Add "ex_data" support to ENGINE so implementations can add state at a
4470 per-structure level rather than having to store it globally.
4473 *) Make it possible for ENGINE structures to be copied when retrieved by
4474 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4475 This causes the "original" ENGINE structure to act like a template,
4476 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4477 operational state can be localised to each ENGINE structure, despite the
4478 fact they all share the same "methods". New ENGINE structures returned in
4479 this case have no functional references and the return value is the single
4480 structural reference. This matches the single structural reference returned
4481 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4485 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4486 needs to match any other type at all we need to manually clear the
4490 *) Changes to the "openssl engine" utility to include;
4491 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4492 about an ENGINE's available control commands.
4493 - executing control commands from command line arguments using the
4494 '-pre' and '-post' switches. '-post' is only used if '-t' is
4495 specified and the ENGINE is successfully initialised. The syntax for
4496 the individual commands are colon-separated, for example;
4497 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4500 *) New dynamic control command support for ENGINEs. ENGINEs can now
4501 declare their own commands (numbers), names (strings), descriptions,
4502 and input types for run-time discovery by calling applications. A
4503 subset of these commands are implicitly classed as "executable"
4504 depending on their input type, and only these can be invoked through
4505 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4506 can be based on user input, config files, etc). The distinction is
4507 that "executable" commands cannot return anything other than a boolean
4508 result and can only support numeric or string input, whereas some
4509 discoverable commands may only be for direct use through
4510 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4511 pointers, or other custom uses. The "executable" commands are to
4512 support parameterisations of ENGINE behaviour that can be
4513 unambiguously defined by ENGINEs and used consistently across any
4514 OpenSSL-based application. Commands have been added to all the
4515 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4516 control over shared-library paths without source code alterations.
4519 *) Changed all ENGINE implementations to dynamically allocate their
4520 ENGINEs rather than declaring them statically. Apart from this being
4521 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4522 this also allows the implementations to compile without using the
4523 internal engine_int.h header.
4526 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4527 'const' value. Any code that should be able to modify a RAND_METHOD
4528 should already have non-const pointers to it (ie. they should only
4529 modify their own ones).
4532 *) Made a variety of little tweaks to the ENGINE code.
4533 - "atalla" and "ubsec" string definitions were moved from header files
4534 to C code. "nuron" string definitions were placed in variables
4535 rather than hard-coded - allowing parameterisation of these values
4536 later on via ctrl() commands.
4537 - Removed unused "#if 0"'d code.
4538 - Fixed engine list iteration code so it uses ENGINE_free() to release
4539 structural references.
4540 - Constified the RAND_METHOD element of ENGINE structures.
4541 - Constified various get/set functions as appropriate and added
4542 missing functions (including a catch-all ENGINE_cpy that duplicates
4543 all ENGINE values onto a new ENGINE except reference counts/state).
4544 - Removed NULL parameter checks in get/set functions. Setting a method
4545 or function to NULL is a way of cancelling out a previously set
4546 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4547 and doesn't justify the extra error symbols and code.
4548 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4549 flags from engine_int.h to engine.h.
4550 - Changed prototypes for ENGINE handler functions (init(), finish(),
4551 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4554 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4555 to the algorithm using long division. The binary algorithm can be
4556 used only if the modulus is odd. On 32-bit systems, it is faster
4557 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4558 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4559 up to 450 bits. In 64-bit environments, the binary algorithm
4560 appears to be advantageous for much longer moduli; here we use it
4561 for moduli up to 2048 bits.
4564 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4565 could not support the combine flag in choice fields.
4568 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4569 extensions from a certificate request to the certificate.
4572 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4573 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4574 file: this allows the display of the certificate about to be
4575 signed to be customised, to allow certain fields to be included
4576 or excluded and extension details. The old system didn't display
4577 multicharacter strings properly, omitted fields not in the policy
4578 and couldn't display additional details such as extensions.
4581 *) Function EC_POINTs_mul for multiple scalar multiplication
4582 of an arbitrary number of elliptic curve points
4583 \sum scalars[i]*points[i],
4584 optionally including the generator defined for the EC_GROUP:
4585 scalar*generator + \sum scalars[i]*points[i].
4587 EC_POINT_mul is a simple wrapper function for the typical case
4588 that the point list has just one item (besides the optional
4592 *) First EC_METHODs for curves over GF(p):
4594 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4595 operations and provides various method functions that can also
4596 operate with faster implementations of modular arithmetic.
4598 EC_GFp_mont_method() reuses most functions that are part of
4599 EC_GFp_simple_method, but uses Montgomery arithmetic.
4601 [Bodo Moeller; point addition and point doubling
4602 implementation directly derived from source code provided by
4603 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4605 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4606 crypto/ec/ec_lib.c):
4608 Curves are EC_GROUP objects (with an optional group generator)
4609 based on EC_METHODs that are built into the library.
4611 Points are EC_POINT objects based on EC_GROUP objects.
4613 Most of the framework would be able to handle curves over arbitrary
4614 finite fields, but as there are no obvious types for fields other
4615 than GF(p), some functions are limited to that for now.
4618 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4619 that the file contains a complete HTTP response.
4622 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4623 change the def and num file printf format specifier from "%-40sXXX"
4624 to "%-39s XXX". The latter will always guarantee a space after the
4625 field while the former will cause them to run together if the field
4626 is 40 of more characters long.
4629 *) Constify the cipher and digest 'method' functions and structures
4630 and modify related functions to take constant EVP_MD and EVP_CIPHER
4634 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4635 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4638 *) Modify EVP_Digest*() routines so they now return values. Although the
4639 internal software routines can never fail additional hardware versions
4643 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4645 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4646 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4649 ERR_R_NESTED_ASN1_ERROR
4651 ERR_R_MISSING_ASN1_EOS
4652 were 4 .. 9, conflicting with
4653 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4655 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4656 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4658 Add new error code 'ERR_R_INTERNAL_ERROR'.
4661 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4665 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4666 sets the subject name for a new request or supersedes the
4667 subject name in a given request. Formats that can be parsed are
4668 'CN=Some Name, OU=myOU, C=IT'
4670 'CN=Some Name/OU=myOU/C=IT'.
4672 Add options '-batch' and '-verbose' to 'openssl req'.
4673 [Massimiliano Pala <madwolf@hackmasters.net>]
4675 *) Introduce the possibility to access global variables through
4676 functions on platform were that's the best way to handle exporting
4677 global variables in shared libraries. To enable this functionality,
4678 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4679 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4680 is normally done by Configure or something similar).
4682 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4683 in the source file (foo.c) like this:
4685 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4686 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4688 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4689 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4691 OPENSSL_DECLARE_GLOBAL(int,foo);
4692 #define foo OPENSSL_GLOBAL_REF(foo)
4693 OPENSSL_DECLARE_GLOBAL(double,bar);
4694 #define bar OPENSSL_GLOBAL_REF(bar)
4696 The #defines are very important, and therefore so is including the
4697 header file everywhere where the defined globals are used.
4699 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4700 of ASN.1 items, but that structure is a bit different.
4702 The largest change is in util/mkdef.pl which has been enhanced with
4703 better and easier to understand logic to choose which symbols should
4704 go into the Windows .def files as well as a number of fixes and code
4705 cleanup (among others, algorithm keywords are now sorted
4706 lexicographically to avoid constant rewrites).
4709 *) In BN_div() keep a copy of the sign of 'num' before writing the
4710 result to 'rm' because if rm==num the value will be overwritten
4711 and produce the wrong result if 'num' is negative: this caused
4712 problems with BN_mod() and BN_nnmod().
4715 *) Function OCSP_request_verify(). This checks the signature on an
4716 OCSP request and verifies the signer certificate. The signer
4717 certificate is just checked for a generic purpose and OCSP request
4721 *) Add OCSP_check_validity() function to check the validity of OCSP
4722 responses. OCSP responses are prepared in real time and may only
4723 be a few seconds old. Simply checking that the current time lies
4724 between thisUpdate and nextUpdate max reject otherwise valid responses
4725 caused by either OCSP responder or client clock inaccuracy. Instead
4726 we allow thisUpdate and nextUpdate to fall within a certain period of
4727 the current time. The age of the response can also optionally be
4728 checked. Two new options -validity_period and -status_age added to
4732 *) If signature or public key algorithm is unrecognized print out its
4733 OID rather that just UNKNOWN.
4736 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4737 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4738 ID to be generated from the issuer certificate alone which can then be
4739 passed to OCSP_id_issuer_cmp().
4742 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4743 ASN1 modules to export functions returning ASN1_ITEM pointers
4744 instead of the ASN1_ITEM structures themselves. This adds several
4745 new macros which allow the underlying ASN1 function/structure to
4746 be accessed transparently. As a result code should not use ASN1_ITEM
4747 references directly (such as &X509_it) but instead use the relevant
4748 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4749 use of the new ASN1 code on platforms where exporting structures
4750 is problematical (for example in shared libraries) but exporting
4751 functions returning pointers to structures is not.
4754 *) Add support for overriding the generation of SSL/TLS session IDs.
4755 These callbacks can be registered either in an SSL_CTX or per SSL.
4756 The purpose of this is to allow applications to control, if they wish,
4757 the arbitrary values chosen for use as session IDs, particularly as it
4758 can be useful for session caching in multiple-server environments. A
4759 command-line switch for testing this (and any client code that wishes
4760 to use such a feature) has been added to "s_server".
4761 [Geoff Thorpe, Lutz Jaenicke]
4763 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4764 of the form '#if defined(...) || defined(...) || ...' and
4765 '#if !defined(...) && !defined(...) && ...'. This also avoids
4766 the growing number of special cases it was previously handling.
4769 *) Make all configuration macros available for application by making
4770 sure they are available in opensslconf.h, by giving them names starting
4771 with "OPENSSL_" to avoid conflicts with other packages and by making
4772 sure e_os2.h will cover all platform-specific cases together with
4774 Additionally, it is now possible to define configuration/platform-
4775 specific names (called "system identities"). In the C code, these
4776 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4777 macro with the name beginning with "OPENSSL_SYS_", which is determined
4778 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4782 *) New option -set_serial to 'req' and 'x509' this allows the serial
4783 number to use to be specified on the command line. Previously self
4784 signed certificates were hard coded with serial number 0 and the
4785 CA options of 'x509' had to use a serial number in a file which was
4789 *) New options to 'ca' utility to support V2 CRL entry extensions.
4790 Currently CRL reason, invalidity date and hold instruction are
4791 supported. Add new CRL extensions to V3 code and some new objects.
4794 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4795 disable standard block padding (aka PKCS#5 padding) in the EVP
4796 API, which was previously mandatory. This means that the data is
4797 not padded in any way and so the total length much be a multiple
4798 of the block size, otherwise an error occurs.
4801 *) Initial (incomplete) OCSP SSL support.
4804 *) New function OCSP_parse_url(). This splits up a URL into its host,
4805 port and path components: primarily to parse OCSP URLs. New -url
4806 option to ocsp utility.
4809 *) New nonce behavior. The return value of OCSP_check_nonce() now
4810 reflects the various checks performed. Applications can decide
4811 whether to tolerate certain situations such as an absent nonce
4812 in a response when one was present in a request: the ocsp application
4813 just prints out a warning. New function OCSP_add1_basic_nonce()
4814 this is to allow responders to include a nonce in a response even if
4815 the request is nonce-less.
4818 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4819 skipped when using openssl x509 multiple times on a single input file,
4820 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4823 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4824 set string type: to handle setting ASN1_TIME structures. Fix ca
4825 utility to correctly initialize revocation date of CRLs.
4828 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4829 the clients preferred ciphersuites and rather use its own preferences.
4830 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4831 Internet Explorer by ensuring unchanged hash method during stepup.
4832 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4835 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4836 to aes and add a new 'exist' option to print out symbols that don't
4840 *) Additional options to ocsp utility to allow flags to be set and
4841 additional certificates supplied.
4844 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4845 OCSP client a number of certificate to only verify the response
4849 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4850 handle the new API. Currently only ECB, CBC modes supported. Add new
4853 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4854 Encryption Standard (AES) Ciphersuites for Transport Layer
4855 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4856 not enabled by default and were not part of the "ALL" ciphersuite
4857 alias because they were not yet official; they could be
4858 explicitly requested by specifying the "AESdraft" ciphersuite
4859 group alias. In the final release of OpenSSL 0.9.7, the group
4860 alias is called "AES" and is part of "ALL".)
4861 [Ben Laurie, Steve Henson, Bodo Moeller]
4863 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4864 request to response.
4867 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4868 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4869 extract information from a certificate request. OCSP_response_create()
4870 creates a response and optionally adds a basic response structure.
4871 OCSP_basic_add1_status() adds a complete single response to a basic
4872 response and returns the OCSP_SINGLERESP structure just added (to allow
4873 extensions to be included for example). OCSP_basic_add1_cert() adds a
4874 certificate to a basic response and OCSP_basic_sign() signs a basic
4875 response with various flags. New helper functions ASN1_TIME_check()
4876 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4877 (converts ASN1_TIME to GeneralizedTime).
4880 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4881 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4882 structure from a certificate. X509_pubkey_digest() digests the public_key
4883 contents: this is used in various key identifiers.
4886 *) Make sk_sort() tolerate a NULL argument.
4887 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4889 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4890 passed by the function are trusted implicitly. If any of them signed the
4891 response then it is assumed to be valid and is not verified.
4894 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4895 to data. This was previously part of the PKCS7 ASN1 code. This
4896 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4897 [Steve Henson, reported by Kenneth R. Robinette
4898 <support@securenetterm.com>]
4900 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4901 routines: without these tracing memory leaks is very painful.
4902 Fix leaks in PKCS12 and PKCS7 routines.
4905 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4906 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4907 effectively meant GeneralizedTime would never be used. Now it
4908 is initialised to -1 but X509_time_adj() now has to check the value
4909 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4910 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4911 [Steve Henson, reported by Kenneth R. Robinette
4912 <support@securenetterm.com>]
4914 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4915 result in a zero length in the ASN1_INTEGER structure which was
4916 not consistent with the structure when d2i_ASN1_INTEGER() was used
4917 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4918 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4919 where it did not print out a minus for negative ASN1_INTEGER.
4922 *) Add summary printout to ocsp utility. The various functions which
4923 convert status values to strings have been renamed to:
4924 OCSP_response_status_str(), OCSP_cert_status_str() and
4925 OCSP_crl_reason_str() and are no longer static. New options
4926 to verify nonce values and to disable verification. OCSP response
4927 printout format cleaned up.
4930 *) Add additional OCSP certificate checks. These are those specified
4931 in RFC2560. This consists of two separate checks: the CA of the
4932 certificate being checked must either be the OCSP signer certificate
4933 or the issuer of the OCSP signer certificate. In the latter case the
4934 OCSP signer certificate must contain the OCSP signing extended key
4935 usage. This check is performed by attempting to match the OCSP
4936 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4937 in the OCSP_CERTID structures of the response.
4940 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4941 and related routines. This uses the standard OpenSSL certificate
4942 verify routines to perform initial checks (just CA validity) and
4943 to obtain the certificate chain. Then additional checks will be
4944 performed on the chain. Currently the root CA is checked to see
4945 if it is explicitly trusted for OCSP signing. This is used to set
4946 a root CA as a global signing root: that is any certificate that
4947 chains to that CA is an acceptable OCSP signing certificate.
4950 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4951 extensions from a separate configuration file.
4952 As when reading extensions from the main configuration file,
4953 the '-extensions ...' option may be used for specifying the
4955 [Massimiliano Pala <madwolf@comune.modena.it>]
4957 *) New OCSP utility. Allows OCSP requests to be generated or
4958 read. The request can be sent to a responder and the output
4959 parsed, outputed or printed in text form. Not complete yet:
4960 still needs to check the OCSP response validity.
4963 *) New subcommands for 'openssl ca':
4964 'openssl ca -status <serial>' prints the status of the cert with
4965 the given serial number (according to the index file).
4966 'openssl ca -updatedb' updates the expiry status of certificates
4968 [Massimiliano Pala <madwolf@comune.modena.it>]
4970 *) New '-newreq-nodes' command option to CA.pl. This is like
4971 '-newreq', but calls 'openssl req' with the '-nodes' option
4972 so that the resulting key is not encrypted.
4973 [Damien Miller <djm@mindrot.org>]
4975 *) New configuration for the GNU Hurd.
4976 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4978 *) Initial code to implement OCSP basic response verify. This
4979 is currently incomplete. Currently just finds the signer's
4980 certificate and verifies the signature on the response.
4983 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4984 value of OPENSSLDIR. This is available via the new '-d' option
4985 to 'openssl version', and is also included in 'openssl version -a'.
4988 *) Allowing defining memory allocation callbacks that will be given
4989 file name and line number information in additional arguments
4990 (a const char* and an int). The basic functionality remains, as
4991 well as the original possibility to just replace malloc(),
4992 realloc() and free() by functions that do not know about these
4993 additional arguments. To register and find out the current
4994 settings for extended allocation functions, the following
4995 functions are provided:
4997 CRYPTO_set_mem_ex_functions
4998 CRYPTO_set_locked_mem_ex_functions
4999 CRYPTO_get_mem_ex_functions
5000 CRYPTO_get_locked_mem_ex_functions
5002 These work the same way as CRYPTO_set_mem_functions and friends.
5003 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5004 extended allocation function is enabled.
5005 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5006 a conventional allocation function is enabled.
5007 [Richard Levitte, Bodo Moeller]
5009 *) Finish off removing the remaining LHASH function pointer casts.
5010 There should no longer be any prototype-casting required when using
5011 the LHASH abstraction, and any casts that remain are "bugs". See
5012 the callback types and macros at the head of lhash.h for details
5013 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5016 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5017 If /dev/[u]random devices are not available or do not return enough
5018 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5020 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5021 /etc/entropy will be queried once each in this sequence, quering stops
5022 when enough entropy was collected without querying more sockets.
5025 *) Change the Unix RAND_poll() variant to be able to poll several
5026 random devices, as specified by DEVRANDOM, until a sufficient amount
5027 of data has been collected. We spend at most 10 ms on each file
5028 (select timeout) and read in non-blocking mode. DEVRANDOM now
5029 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5030 (previously it was just the string "/dev/urandom"), so on typical
5031 platforms the 10 ms delay will never occur.
5032 Also separate out the Unix variant to its own file, rand_unix.c.
5033 For VMS, there's a currently-empty rand_vms.c.
5036 *) Move OCSP client related routines to ocsp_cl.c. These
5037 provide utility functions which an application needing
5038 to issue a request to an OCSP responder and analyse the
5039 response will typically need: as opposed to those which an
5040 OCSP responder itself would need which will be added later.
5042 OCSP_request_sign() signs an OCSP request with an API similar
5043 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5044 response. OCSP_response_get1_basic() extracts basic response
5045 from response. OCSP_resp_find_status(): finds and extracts status
5046 information from an OCSP_CERTID structure (which will be created
5047 when the request structure is built). These are built from lower
5048 level functions which work on OCSP_SINGLERESP structures but
5049 wont normally be used unless the application wishes to examine
5050 extensions in the OCSP response for example.
5052 Replace nonce routines with a pair of functions.
5053 OCSP_request_add1_nonce() adds a nonce value and optionally
5054 generates a random value. OCSP_check_nonce() checks the
5055 validity of the nonce in an OCSP response.
5058 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5059 This doesn't copy the supplied OCSP_CERTID and avoids the
5060 need to free up the newly created id. Change return type
5061 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5062 This can then be used to add extensions to the request.
5063 Deleted OCSP_request_new(), since most of its functionality
5064 is now in OCSP_REQUEST_new() (and the case insensitive name
5065 clash) apart from the ability to set the request name which
5066 will be added elsewhere.
5069 *) Update OCSP API. Remove obsolete extensions argument from
5070 various functions. Extensions are now handled using the new
5071 OCSP extension code. New simple OCSP HTTP function which
5072 can be used to send requests and parse the response.
5075 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5076 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5077 uses the special reorder version of SET OF to sort the attributes
5078 and reorder them to match the encoded order. This resolves a long
5079 standing problem: a verify on a PKCS7 structure just after signing
5080 it used to fail because the attribute order did not match the
5081 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5082 it uses the received order. This is necessary to tolerate some broken
5083 software that does not order SET OF. This is handled by encoding
5084 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5085 to produce the required SET OF.
5088 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5089 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5090 files to get correct declarations of the ASN.1 item variables.
5093 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5094 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5095 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5096 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5097 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5098 ASN1_ITEM and no wrapper functions.
5101 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5102 replace the old function pointer based I/O routines. Change most of
5103 the *_d2i_bio() and *_d2i_fp() functions to use these.
5106 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5107 lines, recognice more "algorithms" that can be deselected, and make
5108 it complain about algorithm deselection that isn't recognised.
5111 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5112 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5113 to use new functions. Add NO_ASN1_OLD which can be set to remove
5114 some old style ASN1 functions: this can be used to determine if old
5115 code will still work when these eventually go away.
5118 *) New extension functions for OCSP structures, these follow the
5119 same conventions as certificates and CRLs.
5122 *) New function X509V3_add1_i2d(). This automatically encodes and
5123 adds an extension. Its behaviour can be customised with various
5124 flags to append, replace or delete. Various wrappers added for
5125 certifcates and CRLs.
5128 *) Fix to avoid calling the underlying ASN1 print routine when
5129 an extension cannot be parsed. Correct a typo in the
5130 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5133 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5134 entries for variables.
5137 *) Add functionality to apps/openssl.c for detecting locking
5138 problems: As the program is single-threaded, all we have
5139 to do is register a locking callback using an array for
5140 storing which locks are currently held by the program.
5143 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5144 SSL_get_ex_data_X509_STORE_idx(), which is used in
5145 ssl_verify_cert_chain() and thus can be called at any time
5146 during TLS/SSL handshakes so that thread-safety is essential.
5147 Unfortunately, the ex_data design is not at all suited
5148 for multi-threaded use, so it probably should be abolished.
5151 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5152 [Broadcom, tweaked and integrated by Geoff Thorpe]
5154 *) Move common extension printing code to new function
5155 X509V3_print_extensions(). Reorganise OCSP print routines and
5156 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5159 *) New function X509_signature_print() to remove duplication in some
5163 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5164 set (this was treated exactly the same as SET OF previously). This
5165 is used to reorder the STACK representing the structure to match the
5166 encoding. This will be used to get round a problem where a PKCS7
5167 structure which was signed could not be verified because the STACK
5168 order did not reflect the encoded order.
5171 *) Reimplement the OCSP ASN1 module using the new code.
5174 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5175 for its ASN1 operations. The old style function pointers still exist
5176 for now but they will eventually go away.
5179 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5180 completely replaces the old ASN1 functionality with a table driven
5181 encoder and decoder which interprets an ASN1_ITEM structure describing
5182 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5183 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5184 has also been converted to the new form.
5187 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5188 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5189 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5190 for negative moduli.
5193 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5194 of not touching the result's sign bit.
5197 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5201 *) Changed the LHASH code to use prototypes for callbacks, and created
5202 macros to declare and implement thin (optionally static) functions
5203 that provide type-safety and avoid function pointer casting for the
5204 type-specific callbacks.
5207 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5209 [Veers Staats <staatsvr@asc.hpc.mil>,
5210 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5212 *) Reformat the FAQ so the different questions and answers can be divided
5213 in sections depending on the subject.
5216 *) Have the zlib compression code load ZLIB.DLL dynamically under
5220 *) New function BN_mod_sqrt for computing square roots modulo a prime
5221 (using the probabilistic Tonelli-Shanks algorithm unless
5222 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5223 be handled deterministically).
5224 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5226 *) Make BN_mod_inverse faster by explicitly handling small quotients
5227 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5228 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5231 *) New function BN_kronecker.
5234 *) Fix BN_gcd so that it works on negative inputs; the result is
5235 positive unless both parameters are zero.
5236 Previously something reasonably close to an infinite loop was
5237 possible because numbers could be growing instead of shrinking
5238 in the implementation of Euclid's algorithm.
5241 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5242 sign of the number in question.
5244 Fix BN_is_word(a,w) to work correctly for w == 0.
5246 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5247 because its test if the absolute value of 'a' equals 'w'.
5248 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5249 it exists mostly for use in the implementations of BN_is_zero(),
5250 BN_is_one(), and BN_is_word().
5253 *) New function BN_swap.
5256 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5257 the exponentiation functions are more likely to produce reasonable
5258 results on negative inputs.
5261 *) Change BN_mod_mul so that the result is always non-negative.
5262 Previously, it could be negative if one of the factors was negative;
5263 I don't think anyone really wanted that behaviour.
5266 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5267 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5268 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5269 and add new functions:
5278 BN_mod_lshift1_quick
5282 These functions always generate non-negative results.
5284 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5285 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5287 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5288 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5289 be reduced modulo m.
5290 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5293 The following entry accidentily appeared in the CHANGES file
5294 distributed with OpenSSL 0.9.7. The modifications described in
5295 it do *not* apply to OpenSSL 0.9.7.
5297 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5298 was actually never needed) and in BN_mul(). The removal in BN_mul()
5299 required a small change in bn_mul_part_recursive() and the addition
5300 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5301 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5302 bn_sub_words() and bn_add_words() except they take arrays with
5307 *) In 'openssl passwd', verify passwords read from the terminal
5308 unless the '-salt' option is used (which usually means that
5309 verification would just waste user's time since the resulting
5310 hash is going to be compared with some given password hash)
5311 or the new '-noverify' option is used.
5313 This is an incompatible change, but it does not affect
5314 non-interactive use of 'openssl passwd' (passwords on the command
5315 line, '-stdin' option, '-in ...' option) and thus should not
5319 *) Remove all references to RSAref, since there's no more need for it.
5322 *) Make DSO load along a path given through an environment variable
5323 (SHLIB_PATH) with shl_load().
5326 *) Constify the ENGINE code as a result of BIGNUM constification.
5327 Also constify the RSA code and most things related to it. In a
5328 few places, most notable in the depth of the ASN.1 code, ugly
5329 casts back to non-const were required (to be solved at a later
5333 *) Make it so the openssl application has all engines loaded by default.
5336 *) Constify the BIGNUM routines a little more.
5339 *) Add the following functions:
5341 ENGINE_load_cswift()
5343 ENGINE_load_atalla()
5345 ENGINE_load_builtin_engines()
5347 That way, an application can itself choose if external engines that
5348 are built-in in OpenSSL shall ever be used or not. The benefit is
5349 that applications won't have to be linked with libdl or other dso
5350 libraries unless it's really needed.
5352 Changed 'openssl engine' to load all engines on demand.
5353 Changed the engine header files to avoid the duplication of some
5354 declarations (they differed!).
5357 *) 'openssl engine' can now list capabilities.
5360 *) Better error reporting in 'openssl engine'.
5363 *) Never call load_dh_param(NULL) in s_server.
5366 *) Add engine application. It can currently list engines by name and
5367 identity, and test if they are actually available.
5370 *) Improve RPM specification file by forcing symbolic linking and making
5371 sure the installed documentation is also owned by root.root.
5372 [Damien Miller <djm@mindrot.org>]
5374 *) Give the OpenSSL applications more possibilities to make use of
5375 keys (public as well as private) handled by engines.
5378 *) Add OCSP code that comes from CertCo.
5381 *) Add VMS support for the Rijndael code.
5384 *) Added untested support for Nuron crypto accelerator.
5387 *) Add support for external cryptographic devices. This code was
5388 previously distributed separately as the "engine" branch.
5389 [Geoff Thorpe, Richard Levitte]
5391 *) Rework the filename-translation in the DSO code. It is now possible to
5392 have far greater control over how a "name" is turned into a filename
5393 depending on the operating environment and any oddities about the
5394 different shared library filenames on each system.
5397 *) Support threads on FreeBSD-elf in Configure.
5400 *) Fix for SHA1 assembly problem with MASM: it produces
5401 warnings about corrupt line number information when assembling
5402 with debugging information. This is caused by the overlapping
5404 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5407 NCONF_get_number() has no error checking at all. As a replacement,
5408 NCONF_get_number_e() is defined (_e for "error checking") and is
5409 promoted strongly. The old NCONF_get_number is kept around for
5410 binary backward compatibility.
5411 Make it possible for methods to load from something other than a BIO,
5412 by providing a function pointer that is given a name instead of a BIO.
5413 For example, this could be used to load configuration data from an
5417 *) Fix for non blocking accept BIOs. Added new I/O special reason
5418 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5419 with non blocking I/O was not possible because no retry code was
5420 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5424 *) Added the beginnings of Rijndael support.
5427 *) Fix for bug in DirectoryString mask setting. Add support for
5428 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5429 to allow certificate printing to more controllable, additional
5430 'certopt' option to 'x509' to allow new printing options to be
5434 *) Clean old EAY MD5 hack from e_os.h.
5437 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5439 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5440 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5441 [Joe Orton, Steve Henson]
5443 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5445 *) Fix additional bug revealed by the NISCC test suite:
5447 Stop bug triggering large recursion when presented with
5448 certain ASN.1 tags (CVE-2003-0851)
5451 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5453 *) Fix various bugs revealed by running the NISCC test suite:
5455 Stop out of bounds reads in the ASN1 code when presented with
5456 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5458 If verify callback ignores invalid public key errors don't try to check
5459 certificate signature with the NULL public key.
5463 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5464 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5468 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5469 extra data after the compression methods not only for TLS 1.0
5470 but also for SSL 3.0 (as required by the specification).
5471 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5473 *) Change X509_certificate_type() to mark the key as exported/exportable
5474 when it's 512 *bits* long, not 512 bytes.
5477 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5479 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5480 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5481 a protocol version number mismatch like a decryption error
5482 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5485 *) Turn on RSA blinding by default in the default implementation
5486 to avoid a timing attack. Applications that don't want it can call
5487 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5488 They would be ill-advised to do so in most cases.
5489 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5491 *) Change RSA blinding code so that it works when the PRNG is not
5492 seeded (in this case, the secret RSA exponent is abused as
5493 an unpredictable seed -- if it is not unpredictable, there
5494 is no point in blinding anyway). Make RSA blinding thread-safe
5495 by remembering the creator's thread ID in rsa->blinding and
5496 having all other threads use local one-time blinding factors
5497 (this requires more computation than sharing rsa->blinding, but
5498 avoids excessive locking; and if an RSA object is not shared
5499 between threads, blinding will still be very fast).
5502 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5504 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5505 via timing by performing a MAC computation even if incorrrect
5506 block cipher padding has been found. This is a countermeasure
5507 against active attacks where the attacker has to distinguish
5508 between bad padding and a MAC verification error. (CVE-2003-0078)
5510 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5511 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5512 Martin Vuagnoux (EPFL, Ilion)]
5514 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5516 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5517 memory from it's contents. This is done with a counter that will
5518 place alternating values in each byte. This can be used to solve
5519 two issues: 1) the removal of calls to memset() by highly optimizing
5520 compilers, and 2) cleansing with other values than 0, since those can
5521 be read through on certain media, for example a swap space on disk.
5524 *) Bugfix: client side session caching did not work with external caching,
5525 because the session->cipher setting was not restored when reloading
5526 from the external cache. This problem was masked, when
5527 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5528 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5531 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5532 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5533 [Zeev Lieber <zeev-l@yahoo.com>]
5535 *) Undo an undocumented change introduced in 0.9.6e which caused
5536 repeated calls to OpenSSL_add_all_ciphers() and
5537 OpenSSL_add_all_digests() to be ignored, even after calling
5541 *) Change the default configuration reader to deal with last line not
5542 being properly terminated.
5545 *) Change X509_NAME_cmp() so it applies the special rules on handling
5546 DN values that are of type PrintableString, as well as RDNs of type
5547 emailAddress where the value has the type ia5String.
5548 [stefank@valicert.com via Richard Levitte]
5550 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5551 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5552 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5553 the bitwise-OR of the two for use by the majority of applications
5554 wanting this behaviour, and update the docs. The documented
5555 behaviour and actual behaviour were inconsistent and had been
5556 changing anyway, so this is more a bug-fix than a behavioural
5558 [Geoff Thorpe, diagnosed by Nadav Har'El]
5560 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5561 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5564 *) Fix initialization code race conditions in
5565 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5566 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5567 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5568 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5569 ssl2_get_cipher_by_char(),
5570 ssl3_get_cipher_by_char().
5571 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5573 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5574 the cached sessions are flushed, as the remove_cb() might use ex_data
5575 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5576 (see [openssl.org #212]).
5577 [Geoff Thorpe, Lutz Jaenicke]
5579 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5580 length, instead of the encoding length to d2i_ASN1_OBJECT.
5583 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5585 *) [In 0.9.6g-engine release:]
5586 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5587 [Lynn Gazis <lgazis@rainbow.com>]
5589 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5591 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5592 and get fix the header length calculation.
5593 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5594 Alon Kantor <alonk@checkpoint.com> (and others),
5597 *) Use proper error handling instead of 'assertions' in buffer
5598 overflow checks added in 0.9.6e. This prevents DoS (the
5599 assertions could call abort()).
5600 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5602 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5604 *) Add various sanity checks to asn1_get_length() to reject
5605 the ASN1 length bytes if they exceed sizeof(long), will appear
5606 negative or the content length exceeds the length of the
5608 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5610 *) Fix cipher selection routines: ciphers without encryption had no flags
5611 for the cipher strength set and where therefore not handled correctly
5612 by the selection routines (PR #130).
5615 *) Fix EVP_dsa_sha macro.
5619 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5620 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5621 that was added in OpenSSL 0.9.6d.
5623 As the countermeasure turned out to be incompatible with some
5624 broken SSL implementations, the new option is part of SSL_OP_ALL.
5625 SSL_OP_ALL is usually employed when compatibility with weird SSL
5626 implementations is desired (e.g. '-bugs' option to 's_client' and
5627 's_server'), so the new option is automatically set in many
5631 *) Changes in security patch:
5633 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5634 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5635 Air Force Materiel Command, USAF, under agreement number
5638 *) Add various sanity checks to asn1_get_length() to reject
5639 the ASN1 length bytes if they exceed sizeof(long), will appear
5640 negative or the content length exceeds the length of the
5641 supplied buffer. (CVE-2002-0659)
5642 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5644 *) Assertions for various potential buffer overflows, not known to
5646 [Ben Laurie (CHATS)]
5648 *) Various temporary buffers to hold ASCII versions of integers were
5649 too small for 64 bit platforms. (CVE-2002-0655)
5650 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5652 *) Remote buffer overflow in SSL3 protocol - an attacker could
5653 supply an oversized session ID to a client. (CVE-2002-0656)
5654 [Ben Laurie (CHATS)]
5656 *) Remote buffer overflow in SSL2 protocol - an attacker could
5657 supply an oversized client master key. (CVE-2002-0656)
5658 [Ben Laurie (CHATS)]
5660 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5662 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5663 encoded as NULL) with id-dsa-with-sha1.
5664 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5666 *) Check various X509_...() return values in apps/req.c.
5667 [Nils Larsch <nla@trustcenter.de>]
5669 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5670 an end-of-file condition would erronously be flagged, when the CRLF
5671 was just at the end of a processed block. The bug was discovered when
5672 processing data through a buffering memory BIO handing the data to a
5673 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5674 <ptsekov@syntrex.com> and Nedelcho Stanev.
5677 *) Implement a countermeasure against a vulnerability recently found
5678 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5679 before application data chunks to avoid the use of known IVs
5680 with data potentially chosen by the attacker.
5683 *) Fix length checks in ssl3_get_client_hello().
5686 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5687 to prevent ssl3_read_internal() from incorrectly assuming that
5688 ssl3_read_bytes() found application data while handshake
5689 processing was enabled when in fact s->s3->in_read_app_data was
5690 merely automatically cleared during the initial handshake.
5691 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5693 *) Fix object definitions for Private and Enterprise: they were not
5694 recognized in their shortname (=lowercase) representation. Extend
5695 obj_dat.pl to issue an error when using undefined keywords instead
5696 of silently ignoring the problem (Svenning Sorensen
5697 <sss@sss.dnsalias.net>).
5700 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5701 generators, i.e. generators other than 2 and 5. (Previously, the
5702 code did not properly initialise the 'add' and 'rem' values to
5703 BN_generate_prime().)
5705 In the new general case, we do not insist that 'generator' is
5706 actually a primitive root: This requirement is rather pointless;
5707 a generator of the order-q subgroup is just as good, if not
5711 *) Map new X509 verification errors to alerts. Discovered and submitted by
5712 Tom Wu <tom@arcot.com>.
5715 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5716 returning non-zero before the data has been completely received
5717 when using non-blocking I/O.
5718 [Bodo Moeller; problem pointed out by John Hughes]
5720 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5721 [Ben Laurie, Lutz Jaenicke]
5723 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5724 Yoram Zahavi <YoramZ@gilian.com>).
5727 *) Add information about CygWin 1.3 and on, and preserve proper
5728 configuration for the versions before that.
5729 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5731 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5732 check whether we deal with a copy of a session and do not delete from
5733 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5734 <izhar@checkpoint.com>.
5737 *) Do not store session data into the internal session cache, if it
5738 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5739 flag is set). Proposed by Aslam <aslam@funk.com>.
5742 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5746 *) [In 0.9.6d-engine release:]
5747 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5748 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5750 *) Add the configuration target linux-s390x.
5751 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5753 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5754 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5755 variable as an indication that a ClientHello message has been
5756 received. As the flag value will be lost between multiple
5757 invocations of ssl3_accept when using non-blocking I/O, the
5758 function may not be aware that a handshake has actually taken
5759 place, thus preventing a new session from being added to the
5762 To avoid this problem, we now set s->new_session to 2 instead of
5763 using a local variable.
5764 [Lutz Jaenicke, Bodo Moeller]
5766 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5767 if the SSL_R_LENGTH_MISMATCH error is detected.
5768 [Geoff Thorpe, Bodo Moeller]
5770 *) New 'shared_ldflag' column in Configure platform table.
5773 *) Fix EVP_CIPHER_mode macro.
5774 ["Dan S. Camper" <dan@bti.net>]
5776 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5777 type, we must throw them away by setting rr->length to 0.
5778 [D P Chang <dpc@qualys.com>]
5780 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5782 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5783 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5784 worked incorrectly for those cases where range = 10..._2 and
5785 3*range is two bits longer than range.)
5788 *) Only add signing time to PKCS7 structures if it is not already
5792 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5793 OBJ_ld_ce should be OBJ_id_ce.
5794 Also some ip-pda OIDs in crypto/objects/objects.txt were
5795 incorrect (cf. RFC 3039).
5796 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5798 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5799 returns early because it has nothing to do.
5800 [Andy Schneider <andy.schneider@bjss.co.uk>]
5802 *) [In 0.9.6c-engine release:]
5803 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5804 [Andy Schneider <andy.schneider@bjss.co.uk>]
5806 *) [In 0.9.6c-engine release:]
5807 Add support for Cryptographic Appliance's keyserver technology.
5808 (Use engine 'keyclient')
5809 [Cryptographic Appliances and Geoff Thorpe]
5811 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5812 is called via tools/c89.sh because arguments have to be
5813 rearranged (all '-L' options must appear before the first object
5815 [Richard Shapiro <rshapiro@abinitio.com>]
5817 *) [In 0.9.6c-engine release:]
5818 Add support for Broadcom crypto accelerator cards, backported
5820 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5822 *) [In 0.9.6c-engine release:]
5823 Add support for SureWare crypto accelerator cards from
5824 Baltimore Technologies. (Use engine 'sureware')
5825 [Baltimore Technologies and Mark Cox]
5827 *) [In 0.9.6c-engine release:]
5828 Add support for crypto accelerator cards from Accelerated
5829 Encryption Processing, www.aep.ie. (Use engine 'aep')
5830 [AEP Inc. and Mark Cox]
5832 *) Add a configuration entry for gcc on UnixWare.
5833 [Gary Benson <gbenson@redhat.com>]
5835 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5836 messages are stored in a single piece (fixed-length part and
5837 variable-length part combined) and fix various bugs found on the way.
5840 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5841 instead. BIO_gethostbyname() does not know what timeouts are
5842 appropriate, so entries would stay in cache even when they have
5844 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5846 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5847 faced with a pathologically small ClientHello fragment that does
5848 not contain client_version: Instead of aborting with an error,
5849 simply choose the highest available protocol version (i.e.,
5850 TLS 1.0 unless it is disabled). In practice, ClientHello
5851 messages are never sent like this, but this change gives us
5852 strictly correct behaviour at least for TLS.
5855 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5856 never resets s->method to s->ctx->method when called from within
5857 one of the SSL handshake functions.
5858 [Bodo Moeller; problem pointed out by Niko Baric]
5860 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5861 (sent using the client's version number) if client_version is
5862 smaller than the protocol version in use. Also change
5863 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5864 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5865 the client will at least see that alert.
5868 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5872 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5873 client receives HelloRequest while in a handshake.
5874 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5876 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5877 should end in 'break', not 'goto end' which circuments various
5878 cleanups done in state SSL_ST_OK. But session related stuff
5879 must be disabled for SSL_ST_OK in the case that we just sent a
5882 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5883 before just sending a HelloRequest.
5884 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5886 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5887 reveal whether illegal block cipher padding was found or a MAC
5888 verification error occured. (Neither SSLerr() codes nor alerts
5889 are directly visible to potential attackers, but the information
5890 may leak via logfiles.)
5892 Similar changes are not required for the SSL 2.0 implementation
5893 because the number of padding bytes is sent in clear for SSL 2.0,
5894 and the extra bytes are just ignored. However ssl/s2_pkt.c
5895 failed to verify that the purported number of padding bytes is in
5899 *) Add OpenUNIX-8 support including shared libraries
5900 (Boyd Lynn Gerber <gerberb@zenez.com>).
5903 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5904 'wristwatch attack' using huge encoding parameters (cf.
5905 James H. Manger's CRYPTO 2001 paper). Note that the
5906 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5907 encoding parameters and hence was not vulnerable.
5910 *) BN_sqr() bug fix.
5911 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5913 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5914 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5915 followed by modular reduction.
5916 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5918 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5919 equivalent based on BN_pseudo_rand() instead of BN_rand().
5922 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5923 This function was broken, as the check for a new client hello message
5924 to handle SGC did not allow these large messages.
5925 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5928 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5931 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5932 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5935 *) Rework the configuration and shared library support for Tru64 Unix.
5936 The configuration part makes use of modern compiler features and
5937 still retains old compiler behavior for those that run older versions
5938 of the OS. The shared library support part includes a variant that
5939 uses the RPATH feature, and is available through the special
5940 configuration target "alpha-cc-rpath", which will never be selected
5942 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5944 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5945 with the same message size as in ssl3_get_certificate_request().
5946 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5947 messages might inadvertently be reject as too long.
5948 [Petr Lampa <lampa@fee.vutbr.cz>]
5950 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5953 *) Modified SSL library such that the verify_callback that has been set
5954 specificly for an SSL object with SSL_set_verify() is actually being
5955 used. Before the change, a verify_callback set with this function was
5956 ignored and the verify_callback() set in the SSL_CTX at the time of
5957 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5958 to allow the necessary settings.
5961 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5962 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5963 done automatically (in contradiction to the requirements of the C
5964 standard). This made problems when used from OpenSSH.
5967 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5968 dh->length and always used
5970 BN_rand_range(priv_key, dh->p).
5972 BN_rand_range() is not necessary for Diffie-Hellman, and this
5973 specific range makes Diffie-Hellman unnecessarily inefficient if
5974 dh->length (recommended exponent length) is much smaller than the
5975 length of dh->p. We could use BN_rand_range() if the order of
5976 the subgroup was stored in the DH structure, but we only have
5981 BN_rand(priv_key, l, ...)
5983 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5989 RSA_eay_public_encrypt
5990 RSA_eay_private_decrypt
5991 RSA_eay_private_encrypt (signing)
5992 RSA_eay_public_decrypt (signature verification)
5994 (default implementations for RSA_public_encrypt,
5995 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5996 always reject numbers >= n.
5999 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6000 to synchronize access to 'locking_thread'. This is necessary on
6001 systems where access to 'locking_thread' (an 'unsigned long'
6002 variable) is not atomic.
6005 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6006 *before* setting the 'crypto_lock_rand' flag. The previous code had
6007 a race condition if 0 is a valid thread ID.
6008 [Travis Vitek <vitek@roguewave.com>]
6010 *) Add support for shared libraries under Irix.
6011 [Albert Chin-A-Young <china@thewrittenword.com>]
6013 *) Add configuration option to build on Linux on both big-endian and
6015 [Ralf Baechle <ralf@uni-koblenz.de>]
6017 *) Add the possibility to create shared libraries on HP-UX.
6020 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6022 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6023 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6024 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6025 PRNG state recovery was possible based on the output of
6026 one PRNG request appropriately sized to gain knowledge on
6027 'md' followed by enough consecutive 1-byte PRNG requests
6028 to traverse all of 'state'.
6030 1. When updating 'md_local' (the current thread's copy of 'md')
6031 during PRNG output generation, hash all of the previous
6032 'md_local' value, not just the half used for PRNG output.
6034 2. Make the number of bytes from 'state' included into the hash
6035 independent from the number of PRNG bytes requested.
6037 The first measure alone would be sufficient to avoid
6038 Markku-Juhani's attack. (Actually it had never occurred
6039 to me that the half of 'md_local' used for chaining was the
6040 half from which PRNG output bytes were taken -- I had always
6041 assumed that the secret half would be used.) The second
6042 measure makes sure that additional data from 'state' is never
6043 mixed into 'md_local' in small portions; this heuristically
6044 further strengthens the PRNG.
6047 *) Fix crypto/bn/asm/mips3.s.
6050 *) When only the key is given to "enc", the IV is undefined. Print out
6051 an error message in this case.
6054 *) Handle special case when X509_NAME is empty in X509 printing routines.
6057 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6058 positive and less than q.
6061 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6062 used: it isn't thread safe and the add_lock_callback should handle
6064 [Paul Rose <Paul.Rose@bridge.com>]
6066 *) Verify that incoming data obeys the block size in
6067 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6071 [Ulf Möller, Bodo Möller]
6073 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6074 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6075 when fixing the server behaviour for backwards-compatible 'client
6076 hello' messages. (Note that the attack is impractical against
6077 SSL 3.0 and TLS 1.0 anyway because length and version checking
6078 means that the probability of guessing a valid ciphertext is
6079 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6082 Before 0.9.5, the countermeasure (hide the error by generating a
6083 random 'decryption result') did not work properly because
6084 ERR_clear_error() was missing, meaning that SSL_get_error() would
6085 detect the supposedly ignored error.
6087 Both problems are now fixed.
6090 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6091 (previously it was 1024).
6094 *) Fix for compatibility mode trust settings: ignore trust settings
6095 unless some valid trust or reject settings are present.
6098 *) Fix for blowfish EVP: its a variable length cipher.
6101 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6102 parameters in DSA public key structures and return an error in the
6103 DSA routines if parameters are absent.
6106 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6107 in the current directory if neither $RANDFILE nor $HOME was set.
6108 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6109 caused some confusion to Windows users who haven't defined $HOME.
6110 Thus RAND_file_name() is changed again: e_os.h can define a
6111 DEFAULT_HOME, which will be used if $HOME is not set.
6112 For Windows, we use "C:"; on other platforms, we still require
6113 environment variables.
6115 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6116 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6117 having multiple threads call RAND_poll() concurrently.
6120 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6121 combination of a flag and a thread ID variable.
6122 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6123 flag), *other* threads can enter ssleay_add_bytes without obeying
6124 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6125 that they do not hold after the first thread unsets add_do_not_lock).
6128 *) Change bctest again: '-x' expressions are not available in all
6132 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6134 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6135 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6137 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6138 the default extension for executables, if any. Also, make the perl
6139 scripts that use symlink() to test if it really exists and use "cp"
6140 if it doesn't. All this made OpenSSL compilable and installable in
6144 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6145 If SEQUENCE is length is indefinite just set c->slen to the total
6146 amount of data available.
6147 [Steve Henson, reported by shige@FreeBSD.org]
6148 [This change does not apply to 0.9.7.]
6150 *) Change bctest to avoid here-documents inside command substitution
6151 (workaround for FreeBSD /bin/sh bug).
6152 For compatibility with Ultrix, avoid shell functions (introduced
6153 in the bctest version that searches along $PATH).
6156 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6157 with des_encrypt() defined on some operating systems, like Solaris
6161 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6162 On the Importance of Eliminating Errors in Cryptographic
6163 Computations, J. Cryptology 14 (2001) 2, 101-119,
6164 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6167 *) MIPS assembler BIGNUM division bug fix.
6170 *) Disabled incorrect Alpha assembler code.
6173 *) Fix PKCS#7 decode routines so they correctly update the length
6174 after reading an EOC for the EXPLICIT tag.
6176 [This change does not apply to 0.9.7.]
6178 *) Fix bug in PKCS#12 key generation routines. This was triggered
6179 if a 3DES key was generated with a 0 initial byte. Include
6180 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6181 (but broken) behaviour.
6184 *) Enhance bctest to search for a working bc along $PATH and print
6186 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6188 *) Fix memory leaks in err.c: free err_data string if necessary;
6189 don't write to the wrong index in ERR_set_error_data.
6192 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6196 *) Replace rdtsc with _emit statements for VC++ version 5.
6197 [Jeremy Cooper <jeremy@baymoo.org>]
6199 *) Make it possible to reuse SSLv2 sessions.
6202 *) In copy_email() check for >= 0 as a return value for
6203 X509_NAME_get_index_by_NID() since 0 is a valid index.
6204 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6206 *) Avoid coredump with unsupported or invalid public keys by checking if
6207 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6208 PKCS7_verify() fails with non detached data.
6211 *) Don't use getenv in library functions when run as setuid/setgid.
6212 New function OPENSSL_issetugid().
6215 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6216 due to incorrect handling of multi-threading:
6218 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6220 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6222 3. Count how many times MemCheck_off() has been called so that
6223 nested use can be treated correctly. This also avoids
6224 inband-signalling in the previous code (which relied on the
6225 assumption that thread ID 0 is impossible).
6228 *) Add "-rand" option also to s_client and s_server.
6231 *) Fix CPU detection on Irix 6.x.
6232 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6233 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6235 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6238 [This change does not apply to 0.9.7.]
6240 *) Use the cached encoding of an X509_NAME structure rather than
6241 copying it. This is apparently the reason for the libsafe "errors"
6242 but the code is actually correct.
6245 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6246 Bleichenbacher's DSA attack.
6247 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6248 to be set and top=0 forces the highest bit to be set; top=-1 is new
6249 and leaves the highest bit random.
6250 [Ulf Moeller, Bodo Moeller]
6252 *) In the NCONF_...-based implementations for CONF_... queries
6253 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6254 a temporary CONF structure with the data component set to NULL
6255 (which gives segmentation faults in lh_retrieve).
6256 Instead, use NULL for the CONF pointer in CONF_get_string and
6257 CONF_get_number (which may use environment variables) and directly
6258 return NULL from CONF_get_section.
6261 *) Fix potential buffer overrun for EBCDIC.
6264 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6265 keyUsage if basicConstraints absent for a CA.
6268 *) Make SMIME_write_PKCS7() write mail header values with a format that
6269 is more generally accepted (no spaces before the semicolon), since
6270 some programs can't parse those values properly otherwise. Also make
6271 sure BIO's that break lines after each write do not create invalid
6275 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6276 macros previously used would not encode an empty SEQUENCE OF
6277 and break the signature.
6279 [This change does not apply to 0.9.7.]
6281 *) Zero the premaster secret after deriving the master secret in
6285 *) Add some EVP_add_digest_alias registrations (as found in
6286 OpenSSL_add_all_digests()) to SSL_library_init()
6287 aka OpenSSL_add_ssl_algorithms(). This provides improved
6288 compatibility with peers using X.509 certificates
6289 with unconventional AlgorithmIdentifier OIDs.
6292 *) Fix for Irix with NO_ASM.
6293 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6295 *) ./config script fixes.
6296 [Ulf Moeller, Richard Levitte]
6298 *) Fix 'openssl passwd -1'.
6301 *) Change PKCS12_key_gen_asc() so it can cope with non null
6302 terminated strings whose length is passed in the passlen
6303 parameter, for example from PEM callbacks. This was done
6304 by adding an extra length parameter to asc2uni().
6305 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6307 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6308 call failed, free the DSA structure.
6311 *) Fix to uni2asc() to cope with zero length Unicode strings.
6312 These are present in some PKCS#12 files.
6315 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6316 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6317 when writing a 32767 byte record.
6318 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6320 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6321 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6323 (RSA objects have a reference count access to which is protected
6324 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6325 so they are meant to be shared between threads.)
6326 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6327 "Reddie, Steven" <Steven.Reddie@ca.com>]
6329 *) Fix a deadlock in CRYPTO_mem_leaks().
6332 *) Use better test patterns in bntest.
6335 *) rand_win.c fix for Borland C.
6338 *) BN_rshift bugfix for n == 0.
6341 *) Add a 'bctest' script that checks for some known 'bc' bugs
6342 so that 'make test' does not abort just because 'bc' is broken.
6345 *) Store verify_result within SSL_SESSION also for client side to
6346 avoid potential security hole. (Re-used sessions on the client side
6347 always resulted in verify_result==X509_V_OK, not using the original
6348 result of the server certificate verification.)
6351 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6352 SSL3_RT_APPLICATION_DATA, return 0.
6353 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6357 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6358 releases, have been re-implemented by renaming the previous
6359 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6360 and ssl3_read_internal, respectively, and adding 'peek' parameters
6361 to them. The new ssl[23]_{read,peek} functions are calls to
6362 ssl[23]_read_internal with the 'peek' flag set appropriately.
6363 A 'peek' parameter has also been added to ssl3_read_bytes, which
6364 does the actual work for ssl3_read_internal.
6367 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6368 the method-specific "init()" handler. Also clean up ex_data after
6369 calling the method-specific "finish()" handler. Previously, this was
6370 happening the other way round.
6373 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6374 The previous value, 12, was not always sufficient for BN_mod_exp().
6377 *) Make sure that shared libraries get the internal name engine with
6378 the full version number and not just 0. This should mark the
6379 shared libraries as not backward compatible. Of course, this should
6380 be changed again when we can guarantee backward binary compatibility.
6383 *) Fix typo in get_cert_by_subject() in by_dir.c
6384 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6386 *) Rework the system to generate shared libraries:
6388 - Make note of the expected extension for the shared libraries and
6389 if there is a need for symbolic links from for example libcrypto.so.0
6390 to libcrypto.so.0.9.7. There is extended info in Configure for
6393 - Make as few rebuilds of the shared libraries as possible.
6395 - Still avoid linking the OpenSSL programs with the shared libraries.
6397 - When installing, install the shared libraries separately from the
6401 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6403 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6404 and not in SSL_clear because the latter is also used by the
6405 accept/connect functions; previously, the settings made by
6406 SSL_set_read_ahead would be lost during the handshake.
6407 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6409 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6410 Previously, it would create entries for disableed algorithms no
6414 *) Added several new manual pages for SSL_* function.
6417 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6419 *) In ssl23_get_client_hello, generate an error message when faced
6420 with an initial SSL 3.0/TLS record that is too small to contain the
6421 first two bytes of the ClientHello message, i.e. client_version.
6422 (Note that this is a pathologic case that probably has never happened
6423 in real life.) The previous approach was to use the version number
6424 from the record header as a substitute; but our protocol choice
6425 should not depend on that one because it is not authenticated
6426 by the Finished messages.
6429 *) More robust randomness gathering functions for Windows.
6430 [Jeffrey Altman <jaltman@columbia.edu>]
6432 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6433 not set then we don't setup the error code for issuer check errors
6434 to avoid possibly overwriting other errors which the callback does
6435 handle. If an application does set the flag then we assume it knows
6436 what it is doing and can handle the new informational codes
6440 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6441 a general "ANY" type, as such it should be able to decode anything
6442 including tagged types. However it didn't check the class so it would
6443 wrongly interpret tagged types in the same way as their universal
6444 counterpart and unknown types were just rejected. Changed so that the
6445 tagged and unknown types are handled in the same way as a SEQUENCE:
6446 that is the encoding is stored intact. There is also a new type
6447 "V_ASN1_OTHER" which is used when the class is not universal, in this
6448 case we have no idea what the actual type is so we just lump them all
6452 *) On VMS, stdout may very well lead to a file that is written to
6453 in a record-oriented fashion. That means that every write() will
6454 write a separate record, which will be read separately by the
6455 programs trying to read from it. This can be very confusing.
6457 The solution is to put a BIO filter in the way that will buffer
6458 text until a linefeed is reached, and then write everything a
6459 line at a time, so every record written will be an actual line,
6460 not chunks of lines and not (usually doesn't happen, but I've
6461 seen it once) several lines in one record. BIO_f_linebuffer() is
6464 Currently, it's a VMS-only method, because that's where it has
6465 been tested well enough.
6468 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6469 it can return incorrect results.
6470 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6471 but it was in 0.9.6-beta[12].)
6474 *) Disable the check for content being present when verifying detached
6475 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6476 include zero length content when signing messages.
6479 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6480 BIO_ctrl (for BIO pairs).
6483 *) Add DSO method for VMS.
6486 *) Bug fix: Montgomery multiplication could produce results with the
6490 *) Add RPM specification openssl.spec and modify it to build three
6491 packages. The default package contains applications, application
6492 documentation and run-time libraries. The devel package contains
6493 include files, static libraries and function documentation. The
6494 doc package contains the contents of the doc directory. The original
6495 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6498 *) Add a large number of documentation files for many SSL routines.
6499 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6501 *) Add a configuration entry for Sony News 4.
6502 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6504 *) Don't set the two most significant bits to one when generating a
6505 random number < q in the DSA library.
6508 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6509 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6510 the underlying transport is blocking) if a handshake took place.
6511 (The default behaviour is needed by applications such as s_client
6512 and s_server that use select() to determine when to use SSL_read;
6513 but for applications that know in advance when to expect data, it
6514 just makes things more complicated.)
6517 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6521 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6522 work better on such systems.
6523 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6525 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6526 Update PKCS12_parse() so it copies the friendlyName and the
6527 keyid to the certificates aux info.
6530 *) Fix bug in PKCS7_verify() which caused an infinite loop
6531 if there was more than one signature.
6532 [Sven Uszpelkat <su@celocom.de>]
6534 *) Major change in util/mkdef.pl to include extra information
6535 about each symbol, as well as presentig variables as well
6536 as functions. This change means that there's n more need
6537 to rebuild the .num files when some algorithms are excluded.
6540 *) Allow the verify time to be set by an application,
6541 rather than always using the current time.
6544 *) Phase 2 verify code reorganisation. The certificate
6545 verify code now looks up an issuer certificate by a
6546 number of criteria: subject name, authority key id
6547 and key usage. It also verifies self signed certificates
6548 by the same criteria. The main comparison function is
6549 X509_check_issued() which performs these checks.
6551 Lot of changes were necessary in order to support this
6552 without completely rewriting the lookup code.
6554 Authority and subject key identifier are now cached.
6556 The LHASH 'certs' is X509_STORE has now been replaced
6557 by a STACK_OF(X509_OBJECT). This is mainly because an
6558 LHASH can't store or retrieve multiple objects with
6559 the same hash value.
6561 As a result various functions (which were all internal
6562 use only) have changed to handle the new X509_STORE
6563 structure. This will break anything that messed round
6564 with X509_STORE internally.
6566 The functions X509_STORE_add_cert() now checks for an
6567 exact match, rather than just subject name.
6569 The X509_STORE API doesn't directly support the retrieval
6570 of multiple certificates matching a given criteria, however
6571 this can be worked round by performing a lookup first
6572 (which will fill the cache with candidate certificates)
6573 and then examining the cache for matches. This is probably
6574 the best we can do without throwing out X509_LOOKUP
6575 entirely (maybe later...).
6577 The X509_VERIFY_CTX structure has been enhanced considerably.
6579 All certificate lookup operations now go via a get_issuer()
6580 callback. Although this currently uses an X509_STORE it
6581 can be replaced by custom lookups. This is a simple way
6582 to bypass the X509_STORE hackery necessary to make this
6583 work and makes it possible to use more efficient techniques
6584 in future. A very simple version which uses a simple
6585 STACK for its trusted certificate store is also provided
6586 using X509_STORE_CTX_trusted_stack().
6588 The verify_cb() and verify() callbacks now have equivalents
6589 in the X509_STORE_CTX structure.
6591 X509_STORE_CTX also has a 'flags' field which can be used
6592 to customise the verify behaviour.
6595 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6596 excludes S/MIME capabilities.
6599 *) When a certificate request is read in keep a copy of the
6600 original encoding of the signed data and use it when outputing
6601 again. Signatures then use the original encoding rather than
6602 a decoded, encoded version which may cause problems if the
6603 request is improperly encoded.
6606 *) For consistency with other BIO_puts implementations, call
6607 buffer_write(b, ...) directly in buffer_puts instead of calling
6610 In BIO_puts, increment b->num_write as in BIO_write.
6611 [Peter.Sylvester@EdelWeb.fr]
6613 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6614 BN_zero, we may not return a BIGNUM with an array consisting of
6618 *) Avoid calling abort() from within the library when problems are
6619 detected, except if preprocessor symbols have been defined
6620 (such as REF_CHECK, BN_DEBUG etc.).
6623 *) New openssl application 'rsautl'. This utility can be
6624 used for low level RSA operations. DER public key
6625 BIO/fp routines also added.
6628 *) New Configure entry and patches for compiling on QNX 4.
6629 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6631 *) A demo state-machine implementation was sponsored by
6632 Nuron (http://www.nuron.com/) and is now available in
6633 demos/state_machine.
6636 *) New options added to the 'dgst' utility for signature
6637 generation and verification.
6640 *) Unrecognized PKCS#7 content types are now handled via a
6641 catch all ASN1_TYPE structure. This allows unsupported
6642 types to be stored as a "blob" and an application can
6643 encode and decode it manually.
6646 *) Fix various signed/unsigned issues to make a_strex.c
6648 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6650 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6651 length if passed a buffer. ASN1_INTEGER_to_BN failed
6652 if passed a NULL BN and its argument was negative.
6653 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6655 *) Modification to PKCS#7 encoding routines to output definite
6656 length encoding. Since currently the whole structures are in
6657 memory there's not real point in using indefinite length
6658 constructed encoding. However if OpenSSL is compiled with
6659 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6662 *) Added BIO_vprintf() and BIO_vsnprintf().
6665 *) Added more prefixes to parse for in the the strings written
6666 through a logging bio, to cover all the levels that are available
6667 through syslog. The prefixes are now:
6669 PANIC, EMERG, EMR => LOG_EMERG
6670 ALERT, ALR => LOG_ALERT
6671 CRIT, CRI => LOG_CRIT
6672 ERROR, ERR => LOG_ERR
6673 WARNING, WARN, WAR => LOG_WARNING
6674 NOTICE, NOTE, NOT => LOG_NOTICE
6675 INFO, INF => LOG_INFO
6676 DEBUG, DBG => LOG_DEBUG
6678 and as before, if none of those prefixes are present at the
6679 beginning of the string, LOG_ERR is chosen.
6681 On Win32, the LOG_* levels are mapped according to this:
6683 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6684 LOG_WARNING => EVENTLOG_WARNING_TYPE
6685 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6689 *) Made it possible to reconfigure with just the configuration
6690 argument "reconf" or "reconfigure". The command line arguments
6691 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6692 and are retrieved from there when reconfiguring.
6696 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6698 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6701 *) The obj_dat.pl script was messing up the sorting of object
6702 names. The reason was that it compared the quoted version
6703 of strings as a result "OCSP" > "OCSP Signing" because
6704 " > SPACE. Changed script to store unquoted versions of
6705 names and add quotes on output. It was also omitting some
6706 names from the lookup table if they were given a default
6707 value (that is if SN is missing it is given the same
6708 value as LN and vice versa), these are now added on the
6709 grounds that if an object has a name we should be able to
6710 look it up. Finally added warning output when duplicate
6711 short or long names are found.
6714 *) Changes needed for Tandem NSK.
6715 [Scott Uroff <scott@xypro.com>]
6717 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6718 RSA_padding_check_SSLv23(), special padding was never detected
6719 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6720 version rollback attacks was not effective.
6722 In s23_clnt.c, don't use special rollback-attack detection padding
6723 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6724 client; similarly, in s23_srvr.c, don't do the rollback check if
6725 SSL 2.0 is the only protocol enabled in the server.
6728 *) Make it possible to get hexdumps of unprintable data with 'openssl
6729 asn1parse'. By implication, the functions ASN1_parse_dump() and
6730 BIO_dump_indent() are added.
6733 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6734 these print out strings and name structures based on various
6735 flags including RFC2253 support and proper handling of
6736 multibyte characters. Added options to the 'x509' utility
6737 to allow the various flags to be set.
6740 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6741 Also change the functions X509_cmp_current_time() and
6742 X509_gmtime_adj() work with an ASN1_TIME structure,
6743 this will enable certificates using GeneralizedTime in validity
6744 dates to be checked.
6747 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6748 negative public key encodings) on by default,
6749 NO_NEG_PUBKEY_BUG can be set to disable it.
6752 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6753 content octets. An i2c_ASN1_OBJECT is unnecessary because
6754 the encoding can be trivially obtained from the structure.
6757 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6758 not read locks (CRYPTO_r_[un]lock).
6761 *) A first attempt at creating official support for shared
6762 libraries through configuration. I've kept it so the
6763 default is static libraries only, and the OpenSSL programs
6764 are always statically linked for now, but there are
6765 preparations for dynamic linking in place.
6766 This has been tested on Linux and Tru64.
6769 *) Randomness polling function for Win9x, as described in:
6770 Peter Gutmann, Software Generation of Practically Strong
6774 *) Fix so PRNG is seeded in req if using an already existing
6778 *) New options to smime application. -inform and -outform
6779 allow alternative formats for the S/MIME message including
6780 PEM and DER. The -content option allows the content to be
6781 specified separately. This should allow things like Netscape
6782 form signing output easier to verify.
6785 *) Fix the ASN1 encoding of tags using the 'long form'.
6788 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6789 STRING types. These convert content octets to and from the
6790 underlying type. The actual tag and length octets are
6791 already assumed to have been read in and checked. These
6792 are needed because all other string types have virtually
6793 identical handling apart from the tag. By having versions
6794 of the ASN1 functions that just operate on content octets
6795 IMPLICIT tagging can be handled properly. It also allows
6796 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6797 and ASN1_INTEGER are identical apart from the tag.
6800 *) Change the handling of OID objects as follows:
6802 - New object identifiers are inserted in objects.txt, following
6803 the syntax given in objects.README.
6804 - objects.pl is used to process obj_mac.num and create a new
6806 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6809 This is currently kind of a hack, and the perl code in objects.pl
6810 isn't very elegant, but it works as I intended. The simplest way
6811 to check that it worked correctly is to look in obj_dat.h and
6812 check the array nid_objs and make sure the objects haven't moved
6813 around (this is important!). Additions are OK, as well as
6814 consistent name changes.
6817 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6820 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6821 The given file adds to whatever has already been seeded into the
6822 random pool through the RANDFILE configuration file option or
6823 environment variable, or the default random state file.
6826 *) mkstack.pl now sorts each macro group into lexical order.
6827 Previously the output order depended on the order the files
6828 appeared in the directory, resulting in needless rewriting
6832 *) Patches to make OpenSSL compile under Win32 again. Mostly
6833 work arounds for the VC++ problem that it treats func() as
6834 func(void). Also stripped out the parts of mkdef.pl that
6835 added extra typesafe functions: these no longer exist.
6838 *) Reorganisation of the stack code. The macros are now all
6839 collected in safestack.h . Each macro is defined in terms of
6840 a "stack macro" of the form SKM_<name>(type, a, b). The
6841 DEBUG_SAFESTACK is now handled in terms of function casts,
6842 this has the advantage of retaining type safety without the
6843 use of additional functions. If DEBUG_SAFESTACK is not defined
6844 then the non typesafe macros are used instead. Also modified the
6845 mkstack.pl script to handle the new form. Needs testing to see
6846 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6847 the default if no major problems. Similar behaviour for ASN1_SET_OF
6848 and PKCS12_STACK_OF.
6851 *) When some versions of IIS use the 'NET' form of private key the
6852 key derivation algorithm is different. Normally MD5(password) is
6853 used as a 128 bit RC4 key. In the modified case
6854 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6855 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6856 as the old Netscape_RSA functions except they have an additional
6857 'sgckey' parameter which uses the modified algorithm. Also added
6858 an -sgckey command line option to the rsa utility. Thanks to
6859 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6860 algorithm to openssl-dev.
6863 *) The evp_local.h macros were using 'c.##kname' which resulted in
6864 invalid expansion on some systems (SCO 5.0.5 for example).
6865 Corrected to 'c.kname'.
6866 [Phillip Porch <root@theporch.com>]
6868 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6869 a STACK of email addresses from a certificate or request, these look
6870 in the subject name and the subject alternative name extensions and
6871 omit any duplicate addresses.
6874 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6875 This makes DSA verification about 2 % faster.
6878 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6879 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6880 plus overhead for 1024 bit moduli).
6881 This makes exponentiations about 0.5 % faster for 1024 bit
6882 exponents (as measured by "openssl speed rsa2048").
6885 *) Rename memory handling macros to avoid conflicts with other
6887 Malloc => OPENSSL_malloc
6888 Malloc_locked => OPENSSL_malloc_locked
6889 Realloc => OPENSSL_realloc
6890 Free => OPENSSL_free
6893 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6894 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6897 *) CygWin32 support.
6898 [John Jarvie <jjarvie@newsguy.com>]
6900 *) The type-safe stack code has been rejigged. It is now only compiled
6901 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6902 by default all type-specific stack functions are "#define"d back to
6903 standard stack functions. This results in more streamlined output
6904 but retains the type-safety checking possibilities of the original
6908 *) The STACK code has been cleaned up, and certain type declarations
6909 that didn't make a lot of sense have been brought in line. This has
6910 also involved a cleanup of sorts in safestack.h to more correctly
6911 map type-safe stack functions onto their plain stack counterparts.
6912 This work has also resulted in a variety of "const"ifications of
6913 lots of the code, especially "_cmp" operations which should normally
6914 be prototyped with "const" parameters anyway.
6917 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6918 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6919 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6920 where all of 'md' is used each time the PRNG is used, but 'state'
6921 is used only indexed by a cyclic counter. As entropy may not be
6922 well distributed from the beginning, 'md' is important as a
6923 chaining variable. However, the output function chains only half
6924 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6925 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6926 in all of 'state' being rewritten, with the new values depending
6927 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6930 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6931 the handshake is continued after ssl_verify_cert_chain();
6932 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6933 can lead to 'unexplainable' connection aborts later.
6934 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6936 *) Major EVP API cipher revision.
6937 Add hooks for extra EVP features. This allows various cipher
6938 parameters to be set in the EVP interface. Support added for variable
6939 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6940 setting of RC2 and RC5 parameters.
6942 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6945 Remove lots of duplicated code from the EVP library. For example *every*
6946 cipher init() function handles the 'iv' in the same way according to the
6947 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6948 for CFB and OFB modes they zero ctx->num.
6950 New functionality allows removal of S/MIME code RC2 hack.
6952 Most of the routines have the same form and so can be declared in terms
6955 By shifting this to the top level EVP_CipherInit() it can be removed from
6956 all individual ciphers. If the cipher wants to handle IVs or keys
6957 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6960 Change lots of functions like EVP_EncryptUpdate() to now return a
6961 value: although software versions of the algorithms cannot fail
6962 any installed hardware versions can.
6965 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6966 this option is set, tolerate broken clients that send the negotiated
6967 protocol version number instead of the requested protocol version
6971 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6972 i.e. non-zero for export ciphersuites, zero otherwise.
6973 Previous versions had this flag inverted, inconsistent with
6974 rsa_tmp_cb (..._TMP_RSA_CB).
6975 [Bodo Moeller; problem reported by Amit Chopra]
6977 *) Add missing DSA library text string. Work around for some IIS
6978 key files with invalid SEQUENCE encoding.
6981 *) Add a document (doc/standards.txt) that list all kinds of standards
6982 and so on that are implemented in OpenSSL.
6985 *) Enhance c_rehash script. Old version would mishandle certificates
6986 with the same subject name hash and wouldn't handle CRLs at all.
6987 Added -fingerprint option to crl utility, to support new c_rehash
6991 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6994 *) Fix for SSL server purpose checking. Server checking was
6995 rejecting certificates which had extended key usage present
6996 but no ssl client purpose.
6997 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6999 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7000 is a little unclear about how a blank password is handled.
7001 Since the password in encoded as a BMPString with terminating
7002 double NULL a zero length password would end up as just the
7003 double NULL. However no password at all is different and is
7004 handled differently in the PKCS#12 key generation code. NS
7005 treats a blank password as zero length. MSIE treats it as no
7006 password on export: but it will try both on import. We now do
7007 the same: PKCS12_parse() tries zero length and no password if
7008 the password is set to "" or NULL (NULL is now a valid password:
7009 it wasn't before) as does the pkcs12 application.
7012 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7013 perror when PEM_read_bio_X509_REQ fails, the error message must
7014 be obtained from the error queue.
7017 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7018 it in ERR_remove_state if appropriate, and change ERR_get_state
7019 accordingly to avoid race conditions (this is necessary because
7020 thread_hash is no longer constant once set).
7023 *) Bugfix for linux-elf makefile.one.
7026 *) RSA_get_default_method() will now cause a default
7027 RSA_METHOD to be chosen if one doesn't exist already.
7028 Previously this was only set during a call to RSA_new()
7029 or RSA_new_method(NULL) meaning it was possible for
7030 RSA_get_default_method() to return NULL.
7033 *) Added native name translation to the existing DSO code
7034 that will convert (if the flag to do so is set) filenames
7035 that are sufficiently small and have no path information
7036 into a canonical native form. Eg. "blah" converted to
7037 "libblah.so" or "blah.dll" etc.
7040 *) New function ERR_error_string_n(e, buf, len) which is like
7041 ERR_error_string(e, buf), but writes at most 'len' bytes
7042 including the 0 terminator. For ERR_error_string_n, 'buf'
7044 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7046 *) CONF library reworked to become more general. A new CONF
7047 configuration file reader "class" is implemented as well as a
7048 new functions (NCONF_*, for "New CONF") to handle it. The now
7049 old CONF_* functions are still there, but are reimplemented to
7050 work in terms of the new functions. Also, a set of functions
7051 to handle the internal storage of the configuration data is
7052 provided to make it easier to write new configuration file
7053 reader "classes" (I can definitely see something reading a
7054 configuration file in XML format, for example), called _CONF_*,
7055 or "the configuration storage API"...
7057 The new configuration file reading functions are:
7059 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7060 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7062 NCONF_default, NCONF_WIN32
7064 NCONF_dump_fp, NCONF_dump_bio
7066 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7067 NCONF_new creates a new CONF object. This works in the same way
7068 as other interfaces in OpenSSL, like the BIO interface.
7069 NCONF_dump_* dump the internal storage of the configuration file,
7070 which is useful for debugging. All other functions take the same
7071 arguments as the old CONF_* functions wth the exception of the
7072 first that must be a `CONF *' instead of a `LHASH *'.
7074 To make it easer to use the new classes with the old CONF_* functions,
7075 the function CONF_set_default_method is provided.
7078 *) Add '-tls1' option to 'openssl ciphers', which was already
7079 mentioned in the documentation but had not been implemented.
7080 (This option is not yet really useful because even the additional
7081 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7084 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7085 OpenSSL-based applications) load shared libraries and bind to
7086 them in a portable way.
7087 [Geoff Thorpe, with contributions from Richard Levitte]
7089 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7091 *) Make sure _lrotl and _lrotr are only used with MSVC.
7093 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7094 (the default implementation of RAND_status).
7096 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7097 to '-clrext' (= clear extensions), as intended and documented.
7098 [Bodo Moeller; inconsistency pointed out by Michael Attili
7101 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7102 was larger than the MD block size.
7103 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7105 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7106 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7107 using the passed key: if the passed key was a private key the result
7108 of X509_print(), for example, would be to print out all the private key
7112 *) des_quad_cksum() byte order bug fix.
7113 [Ulf Möller, using the problem description in krb4-0.9.7, where
7114 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7116 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7118 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7120 *) For easily testing in shell scripts whether some command
7121 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7122 returns with exit code 0 iff no command of the given name is available.
7123 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7124 the output goes to stdout and nothing is printed to stderr.
7125 Additional arguments are always ignored.
7127 Since for each cipher there is a command of the same name,
7128 the 'no-cipher' compilation switches can be tested this way.
7130 ('openssl no-XXX' is not able to detect pseudo-commands such
7131 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7134 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7137 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7138 is set; it will be thrown away anyway because each handshake creates
7140 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7141 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7142 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7143 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7146 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7147 'Q' and 'R' lose their special meanings (quit/renegotiate).
7148 This is part of what -quiet does; unlike -quiet, -ign_eof
7149 does not suppress any output.
7152 *) Add compatibility options to the purpose and trust code. The
7153 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7154 accepts a certificate or CA, this was the previous behaviour,
7155 with all the associated security issues.
7157 X509_TRUST_COMPAT is the old trust behaviour: only and
7158 automatically trust self signed roots in certificate store. A
7159 new trust setting X509_TRUST_DEFAULT is used to specify that
7160 a purpose has no associated trust setting and it should instead
7161 use the value in the default purpose.
7164 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7165 and fix a memory leak.
7168 *) In util/mkerr.pl (which implements 'make errors'), preserve
7169 reason strings from the previous version of the .c file, as
7170 the default to have only downcase letters (and digits) in
7171 automatically generated reasons codes is not always appropriate.
7174 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7175 using strerror. Previously, ERR_reason_error_string() returned
7176 library names as reason strings for SYSerr; but SYSerr is a special
7177 case where small numbers are errno values, not library numbers.
7180 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7181 converts DSA parameters into DH parameters. (When creating parameters,
7182 DSA_generate_parameters is used.)
7185 *) Include 'length' (recommended exponent length) in C code generated
7186 by 'openssl dhparam -C'.
7189 *) The second argument to set_label in perlasm was already being used
7190 so couldn't be used as a "file scope" flag. Moved to third argument
7194 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7195 instead of RAND_bytes for encryption IVs and salts.
7198 *) Include RAND_status() into RAND_METHOD instead of implementing
7199 it only for md_rand.c Otherwise replacing the PRNG by calling
7200 RAND_set_rand_method would be impossible.
7203 *) Don't let DSA_generate_key() enter an infinite loop if the random
7204 number generation fails.
7207 *) New 'rand' application for creating pseudo-random output.
7210 *) Added configuration support for Linux/IA64
7211 [Rolf Haberrecker <rolf@suse.de>]
7213 *) Assembler module support for Mingw32.
7216 *) Shared library support for HPUX (in shlib/).
7217 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7219 *) Shared library support for Solaris gcc.
7220 [Lutz Behnke <behnke@trustcenter.de>]
7222 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7224 *) PKCS7_encrypt() was adding text MIME headers twice because they
7225 were added manually and by SMIME_crlf_copy().
7228 *) In bntest.c don't call BN_rand with zero bits argument.
7229 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7231 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7232 case was implemented. This caused BN_div_recp() to fail occasionally.
7235 *) Add an optional second argument to the set_label() in the perl
7236 assembly language builder. If this argument exists and is set
7237 to 1 it signals that the assembler should use a symbol whose
7238 scope is the entire file, not just the current function. This
7239 is needed with MASM which uses the format label:: for this scope.
7240 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7242 *) Change the ASN1 types so they are typedefs by default. Before
7243 almost all types were #define'd to ASN1_STRING which was causing
7244 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7248 *) Change names of new functions to the new get1/get0 naming
7249 convention: After 'get1', the caller owns a reference count
7250 and has to call ..._free; 'get0' returns a pointer to some
7251 data structure without incrementing reference counters.
7252 (Some of the existing 'get' functions increment a reference
7253 counter, some don't.)
7254 Similarly, 'set1' and 'add1' functions increase reference
7255 counters or duplicate objects.
7258 *) Allow for the possibility of temp RSA key generation failure:
7259 the code used to assume it always worked and crashed on failure.
7262 *) Fix potential buffer overrun problem in BIO_printf().
7263 [Ulf Möller, using public domain code by Patrick Powell; problem
7264 pointed out by David Sacerdote <das33@cornell.edu>]
7266 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7267 RAND_egd() and RAND_status(). In the command line application,
7268 the EGD socket can be specified like a seed file using RANDFILE
7272 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7273 Some CAs (e.g. Verisign) distribute certificates in this form.
7276 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7277 list to exclude them. This means that no special compilation option
7278 is needed to use anonymous DH: it just needs to be included in the
7282 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7283 EVP_MD_type. The old functionality is available in a new macro called
7284 EVP_MD_md(). Change code that uses it and update docs.
7287 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7288 where the 'void *' argument is replaced by a function pointer argument.
7289 Previously 'void *' was abused to point to functions, which works on
7290 many platforms, but is not correct. As these functions are usually
7291 called by macros defined in OpenSSL header files, most source code
7292 should work without changes.
7295 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7296 sections with information on -D... compiler switches used for
7297 compiling the library so that applications can see them. To enable
7298 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7299 must be defined. E.g.,
7300 #define OPENSSL_ALGORITHM_DEFINES
7301 #include <openssl/opensslconf.h>
7302 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7303 [Richard Levitte, Ulf and Bodo Möller]
7305 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7309 *) Change the 'other' type in certificate aux info to a STACK_OF
7310 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7311 the required ASN1 format: arbitrary types determined by an OID.
7314 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7315 argument to 'req'. This is not because the function is newer or
7316 better than others it just uses the work 'NEW' in the certificate
7317 request header lines. Some software needs this.
7320 *) Reorganise password command line arguments: now passwords can be
7321 obtained from various sources. Delete the PEM_cb function and make
7322 it the default behaviour: i.e. if the callback is NULL and the
7323 usrdata argument is not NULL interpret it as a null terminated pass
7324 phrase. If usrdata and the callback are NULL then the pass phrase
7325 is prompted for as usual.
7328 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7329 the support is automatically enabled. The resulting binaries will
7330 autodetect the card and use it if present.
7331 [Ben Laurie and Compaq Inc.]
7333 *) Work around for Netscape hang bug. This sends certificate request
7334 and server done in one record. Since this is perfectly legal in the
7335 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7336 the bugs/SSLv3 entry for more info.
7339 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7342 *) Add -rand argument to smime and pkcs12 applications and read/write
7346 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7349 *) Add command line password options to the remaining applications.
7352 *) Bug fix for BN_div_recp() for numerators with an even number of
7356 *) More tests in bntest.c, and changed test_bn output.
7359 *) ./config recognizes MacOS X now.
7362 *) Bug fix for BN_div() when the first words of num and divsor are
7363 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7366 *) Add support for various broken PKCS#8 formats, and command line
7367 options to produce them.
7370 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7371 get temporary BIGNUMs from a BN_CTX.
7374 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7378 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7379 include a #define from the old name to the new. The original intent
7380 was that statically linked binaries could for example just call
7381 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7382 link with digests. This never worked becayse SSLeay_add_all_digests()
7383 and SSLeay_add_all_ciphers() were in the same source file so calling
7384 one would link with the other. They are now in separate source files.
7387 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7390 *) Use a less unusual form of the Miller-Rabin primality test (it used
7391 a binary algorithm for exponentiation integrated into the Miller-Rabin
7392 loop, our standard modexp algorithms are faster).
7395 *) Support for the EBCDIC character set completed.
7396 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7398 *) Source code cleanups: use const where appropriate, eliminate casts,
7399 use void * instead of char * in lhash.
7402 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7403 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7404 this the server could overwrite ephemeral keys that the client
7408 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7409 using 50 iterations of the Rabin-Miller test.
7411 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7412 iterations of the Rabin-Miller test as required by the appendix
7413 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7414 As BN_is_prime_fasttest includes trial division, DSA parameter
7415 generation becomes much faster.
7417 This implies a change for the callback functions in DSA_is_prime
7418 and DSA_generate_parameters: The callback function is called once
7419 for each positive witness in the Rabin-Miller test, not just
7420 occasionally in the inner loop; and the parameters to the
7421 callback function now provide an iteration count for the outer
7422 loop rather than for the current invocation of the inner loop.
7423 DSA_generate_parameters additionally can call the callback
7424 function with an 'iteration count' of -1, meaning that a
7425 candidate has passed the trial division test (when q is generated
7426 from an application-provided seed, trial division is skipped).
7429 *) New function BN_is_prime_fasttest that optionally does trial
7430 division before starting the Rabin-Miller test and has
7431 an additional BN_CTX * argument (whereas BN_is_prime always
7432 has to allocate at least one BN_CTX).
7433 'callback(1, -1, cb_arg)' is called when a number has passed the
7434 trial division stage.
7437 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7441 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7444 *) New function BN_pseudo_rand().
7447 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7448 bignum version of BN_from_montgomery() with the working code from
7449 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7453 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7454 made it impossible to use the same SSL_SESSION data structure in
7455 SSL2 clients in multiple threads.
7458 *) The return value of RAND_load_file() no longer counts bytes obtained
7459 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7460 to seed the PRNG (previously an explicit byte count was required).
7461 [Ulf Möller, Bodo Möller]
7463 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7464 used (char *) instead of (void *) and had casts all over the place.
7467 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7470 *) Retain source code compatibility for BN_prime_checks macro:
7471 BN_is_prime(..., BN_prime_checks, ...) now uses
7472 BN_prime_checks_for_size to determine the appropriate number of
7473 Rabin-Miller iterations.
7476 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7477 DH_CHECK_P_NOT_SAFE_PRIME.
7478 (Check if this is true? OpenPGP calls them "strong".)
7481 *) Merge the functionality of "dh" and "gendh" programs into a new program
7482 "dhparam". The old programs are retained for now but will handle DH keys
7483 (instead of parameters) in future.
7486 *) Make the ciphers, s_server and s_client programs check the return values
7487 when a new cipher list is set.
7490 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7491 ciphers. Before when the 56bit ciphers were enabled the sorting was
7494 The syntax for the cipher sorting has been extended to support sorting by
7495 cipher-strength (using the strength_bits hard coded in the tables).
7496 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7498 Fix a bug in the cipher-command parser: when supplying a cipher command
7499 string with an "undefined" symbol (neither command nor alphanumeric
7500 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7501 an error is flagged.
7503 Due to the strength-sorting extension, the code of the
7504 ssl_create_cipher_list() function was completely rearranged. I hope that
7505 the readability was also increased :-)
7506 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7508 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7509 for the first serial number and places 2 in the serial number file. This
7510 avoids problems when the root CA is created with serial number zero and
7511 the first user certificate has the same issuer name and serial number
7515 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7516 the new code. Add documentation for this stuff.
7519 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7520 X509_*() to X509at_*() on the grounds that they don't handle X509
7521 structures and behave in an analagous way to the X509v3 functions:
7522 they shouldn't be called directly but wrapper functions should be used
7525 So we also now have some wrapper functions that call the X509at functions
7526 when passed certificate requests. (TO DO: similar things can be done with
7527 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7528 things. Some of these need some d2i or i2d and print functionality
7529 because they handle more complex structures.)
7532 *) Add missing #ifndefs that caused missing symbols when building libssl
7533 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7534 NO_RSA in ssl/s2*.c.
7535 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7537 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7538 has a return value which indicates the quality of the random data
7539 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7540 error queue. New function RAND_pseudo_bytes() generates output that is
7541 guaranteed to be unique but not unpredictable. RAND_add is like
7542 RAND_seed, but takes an extra argument for an entropy estimate
7543 (RAND_seed always assumes full entropy).
7546 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7547 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7548 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7549 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7550 false-positive rate of at most 2^-80 for random input.
7553 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7556 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7557 in the 0.9.5 release), this returns the chain
7558 from an X509_CTX structure with a dup of the stack and all
7559 the X509 reference counts upped: so the stack will exist
7560 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7563 Also make SSL_SESSION_print() print out the verify return
7567 *) Add manpage for the pkcs12 command. Also change the default
7568 behaviour so MAC iteration counts are used unless the new
7569 -nomaciter option is used. This improves file security and
7570 only older versions of MSIE (4.0 for example) need it.
7573 *) Honor the no-xxx Configure options when creating .DEF files.
7576 *) Add PKCS#10 attributes to field table: challengePassword,
7577 unstructuredName and unstructuredAddress. These are taken from
7578 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7579 international characters are used.
7581 More changes to X509_ATTRIBUTE code: allow the setting of types
7582 based on strings. Remove the 'loc' parameter when adding
7583 attributes because these will be a SET OF encoding which is sorted
7587 *) Initial changes to the 'req' utility to allow request generation
7588 automation. This will allow an application to just generate a template
7589 file containing all the field values and have req construct the
7592 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7593 used all over the place including certificate requests and PKCS#7
7594 structures. They are currently handled manually where necessary with
7595 some primitive wrappers for PKCS#7. The new functions behave in a
7596 manner analogous to the X509 extension functions: they allow
7597 attributes to be looked up by NID and added.
7599 Later something similar to the X509V3 code would be desirable to
7600 automatically handle the encoding, decoding and printing of the
7601 more complex types. The string types like challengePassword can
7602 be handled by the string table functions.
7604 Also modified the multi byte string table handling. Now there is
7605 a 'global mask' which masks out certain types. The table itself
7606 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7607 is useful when for example there is only one permissible type
7608 (as in countryName) and using the mask might result in no valid
7612 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7613 SSL_get_peer_finished to allow applications to obtain the latest
7614 Finished messages sent to the peer or expected from the peer,
7615 respectively. (SSL_get_peer_finished is usually the Finished message
7616 actually received from the peer, otherwise the protocol will be aborted.)
7618 As the Finished message are message digests of the complete handshake
7619 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7620 be used for external authentication procedures when the authentication
7621 provided by SSL/TLS is not desired or is not enough.
7624 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7625 the host supports BWX extension and if Compaq C is present on the
7626 $PATH. Just exploiting of the BWX extension results in 20-30%
7627 performance kick for some algorithms, e.g. DES and RC4 to mention
7628 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7632 *) Add support for MS "fast SGC". This is arguably a violation of the
7633 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7634 weak crypto and after checking the certificate is SGC a second one
7635 with strong crypto. MS SGC stops the first handshake after receiving
7636 the server certificate message and sends a second client hello. Since
7637 a server will typically do all the time consuming operations before
7638 expecting any further messages from the client (server key exchange
7639 is the most expensive) there is little difference between the two.
7641 To get OpenSSL to support MS SGC we have to permit a second client
7642 hello message after we have sent server done. In addition we have to
7643 reset the MAC if we do get this second client hello.
7646 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7647 if a DER encoded private key is RSA or DSA traditional format. Changed
7648 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7649 format DER encoded private key. Newer code should use PKCS#8 format which
7650 has the key type encoded in the ASN1 structure. Added DER private key
7651 support to pkcs8 application.
7654 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7655 ciphersuites has been selected (as required by the SSL 3/TLS 1
7656 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7657 is set, we interpret this as a request to violate the specification
7658 (the worst that can happen is a handshake failure, and 'correct'
7659 behaviour would result in a handshake failure anyway).
7662 *) In SSL_CTX_add_session, take into account that there might be multiple
7663 SSL_SESSION structures with the same session ID (e.g. when two threads
7664 concurrently obtain them from an external cache).
7665 The internal cache can handle only one SSL_SESSION with a given ID,
7666 so if there's a conflict, we now throw out the old one to achieve
7670 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7671 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7672 some routines that use cipher OIDs: some ciphers do not have OIDs
7673 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7677 *) Simplify the trust setting structure and code. Now we just have
7678 two sequences of OIDs for trusted and rejected settings. These will
7679 typically have values the same as the extended key usage extension
7680 and any application specific purposes.
7682 The trust checking code now has a default behaviour: it will just
7683 check for an object with the same NID as the passed id. Functions can
7684 be provided to override either the default behaviour or the behaviour
7685 for a given id. SSL client, server and email already have functions
7686 in place for compatibility: they check the NID and also return "trusted"
7687 if the certificate is self signed.
7690 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7691 traditional format into an EVP_PKEY structure.
7694 *) Add a password callback function PEM_cb() which either prompts for
7695 a password if usr_data is NULL or otherwise assumes it is a null
7696 terminated password. Allow passwords to be passed on command line
7697 environment or config files in a few more utilities.
7700 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7701 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7702 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7703 Update documentation.
7706 *) Support for ASN1 "NULL" type. This could be handled before by using
7707 ASN1_TYPE but there wasn't any function that would try to read a NULL
7708 and produce an error if it couldn't. For compatibility we also have
7709 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7710 don't allocate anything because they don't need to.
7713 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7715 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7717 *) Rebuild of the memory allocation routines used by OpenSSL code and
7718 possibly others as well. The purpose is to make an interface that
7719 provide hooks so anyone can build a separate set of allocation and
7720 deallocation routines to be used by OpenSSL, for example memory
7721 pool implementations, or something else, which was previously hard
7722 since Malloc(), Realloc() and Free() were defined as macros having
7723 the values malloc, realloc and free, respectively (except for Win32
7724 compilations). The same is provided for memory debugging code.
7725 OpenSSL already comes with functionality to find memory leaks, but
7726 this gives people a chance to debug other memory problems.
7728 With these changes, a new set of functions and macros have appeared:
7730 CRYPTO_set_mem_debug_functions() [F]
7731 CRYPTO_get_mem_debug_functions() [F]
7732 CRYPTO_dbg_set_options() [F]
7733 CRYPTO_dbg_get_options() [F]
7734 CRYPTO_malloc_debug_init() [M]
7736 The memory debug functions are NULL by default, unless the library
7737 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7738 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7739 gives the standard debugging functions that come with OpenSSL) or
7740 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7741 provided by the library user) must be used. When the standard
7742 debugging functions are used, CRYPTO_dbg_set_options can be used to
7743 request additional information:
7744 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7745 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7747 Also, things like CRYPTO_set_mem_functions will always give the
7748 expected result (the new set of functions is used for allocation
7749 and deallocation) at all times, regardless of platform and compiler
7752 To finish it up, some functions that were never use in any other
7753 way than through macros have a new API and new semantic:
7756 CRYPTO_dbg_realloc()
7759 All macros of value have retained their old syntax.
7760 [Richard Levitte and Bodo Moeller]
7762 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7763 ordering of SMIMECapabilities wasn't in "strength order" and there
7764 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7768 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7769 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7770 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7772 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7773 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7774 functionality to handle multipart/signed properly) and a utility
7775 called 'smime' to call all this stuff. This is based on code I
7776 originally wrote for Celo who have kindly allowed it to be
7777 included in OpenSSL.
7780 *) Add variants des_set_key_checked and des_set_key_unchecked of
7781 des_set_key (aka des_key_sched). Global variable des_check_key
7782 decides which of these is called by des_set_key; this way
7783 des_check_key behaves as it always did, but applications and
7784 the library itself, which was buggy for des_check_key == 1,
7785 have a cleaner way to pick the version they need.
7788 *) New function PKCS12_newpass() which changes the password of a
7792 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7793 dynamic mix. In both cases the ids can be used as an index into the
7794 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7795 functions so they accept a list of the field values and the
7796 application doesn't need to directly manipulate the X509_TRUST
7800 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7804 *) Modify the way the V3 extension code looks up extensions. This now
7805 works in a similar way to the object code: we have some "standard"
7806 extensions in a static table which is searched with OBJ_bsearch()
7807 and the application can add dynamic ones if needed. The file
7808 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7809 updated whenever a new extension is added to the core code and kept
7810 in ext_nid order. There is a simple program 'tabtest.c' which checks
7811 this. New extensions are not added too often so this file can readily
7812 be maintained manually.
7814 There are two big advantages in doing things this way. The extensions
7815 can be looked up immediately and no longer need to be "added" using
7816 X509V3_add_standard_extensions(): this function now does nothing.
7817 [Side note: I get *lots* of email saying the extension code doesn't
7818 work because people forget to call this function]
7819 Also no dynamic allocation is done unless new extensions are added:
7820 so if we don't add custom extensions there is no need to call
7821 X509V3_EXT_cleanup().
7824 *) Modify enc utility's salting as follows: make salting the default. Add a
7825 magic header, so unsalted files fail gracefully instead of just decrypting
7826 to garbage. This is because not salting is a big security hole, so people
7827 should be discouraged from doing it.
7830 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7831 digest to be passed on the command line but it only used this
7832 parameter when signing a certificate. Modified so all relevant
7833 operations are affected by the digest parameter including the
7834 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7835 DSA key was used because it didn't fix the digest.
7838 *) Initial certificate chain verify code. Currently tests the untrusted
7839 certificates for consistency with the verify purpose (which is set
7840 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7842 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7843 this is because it will reject chains with invalid extensions whereas
7844 every previous version of OpenSSL and SSLeay made no checks at all.
7846 Trust code: checks the root CA for the relevant trust settings. Trust
7847 settings have an initial value consistent with the verify purpose: e.g.
7848 if the verify purpose is for SSL client use it expects the CA to be
7849 trusted for SSL client use. However the default value can be changed to
7850 permit custom trust settings: one example of this would be to only trust
7851 certificates from a specific "secure" set of CAs.
7853 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7854 which should be used for version portability: especially since the
7855 verify structure is likely to change more often now.
7857 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7858 to set them. If not set then assume SSL clients will verify SSL servers
7861 Two new options to the verify program: -untrusted allows a set of
7862 untrusted certificates to be passed in and -purpose which sets the
7863 intended purpose of the certificate. If a purpose is set then the
7864 new chain verify code is used to check extension consistency.
7867 *) Support for the authority information access extension.
7870 *) Modify RSA and DSA PEM read routines to transparently handle
7871 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7872 public keys in a format compatible with certificate
7873 SubjectPublicKeyInfo structures. Unfortunately there were already
7874 functions called *_PublicKey_* which used various odd formats so
7875 these are retained for compatibility: however the DSA variants were
7876 never in a public release so they have been deleted. Changed dsa/rsa
7877 utilities to handle the new format: note no releases ever handled public
7878 keys so we should be OK.
7880 The primary motivation for this change is to avoid the same fiasco
7881 that dogs private keys: there are several incompatible private key
7882 formats some of which are standard and some OpenSSL specific and
7883 require various evil hacks to allow partial transparent handling and
7884 even then it doesn't work with DER formats. Given the option anything
7885 other than PKCS#8 should be dumped: but the other formats have to
7886 stay in the name of compatibility.
7888 With public keys and the benefit of hindsight one standard format
7889 is used which works with EVP_PKEY, RSA or DSA structures: though
7890 it clearly returns an error if you try to read the wrong kind of key.
7892 Added a -pubkey option to the 'x509' utility to output the public key.
7893 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7894 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7895 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7896 that do the same as the EVP_PKEY_assign_*() except they up the
7897 reference count of the added key (they don't "swallow" the
7901 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7902 CRLs would fail if the file contained no certificates or no CRLs:
7903 added a new function to read in both types and return the number
7904 read: this means that if none are read it will be an error. The
7905 DER versions of the certificate and CRL reader would always fail
7906 because it isn't possible to mix certificates and CRLs in DER format
7907 without choking one or the other routine. Changed this to just read
7908 a certificate: this is the best we can do. Also modified the code
7909 in apps/verify.c to take notice of return codes: it was previously
7910 attempting to read in certificates from NULL pointers and ignoring
7911 any errors: this is one reason why the cert and CRL reader seemed
7912 to work. It doesn't check return codes from the default certificate
7913 routines: these may well fail if the certificates aren't installed.
7916 *) Code to support otherName option in GeneralName.
7919 *) First update to verify code. Change the verify utility
7920 so it warns if it is passed a self signed certificate:
7921 for consistency with the normal behaviour. X509_verify
7922 has been modified to it will now verify a self signed
7923 certificate if *exactly* the same certificate appears
7924 in the store: it was previously impossible to trust a
7925 single self signed certificate. This means that:
7926 openssl verify ss.pem
7927 now gives a warning about a self signed certificate but
7928 openssl verify -CAfile ss.pem ss.pem
7932 *) For servers, store verify_result in SSL_SESSION data structure
7933 (and add it to external session representation).
7934 This is needed when client certificate verifications fails,
7935 but an application-provided verification callback (set by
7936 SSL_CTX_set_cert_verify_callback) allows accepting the session
7937 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7938 but returns 1): When the session is reused, we have to set
7939 ssl->verify_result to the appropriate error code to avoid
7941 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7943 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7944 case in PKCS7_dataInit() where the signed PKCS7 structure
7945 didn't contain any existing data because it was being created.
7946 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7948 *) Add a salt to the key derivation routines in enc.c. This
7949 forms the first 8 bytes of the encrypted file. Also add a
7950 -S option to allow a salt to be input on the command line.
7953 *) New function X509_cmp(). Oddly enough there wasn't a function
7954 to compare two certificates. We do this by working out the SHA1
7955 hash and comparing that. X509_cmp() will be needed by the trust
7959 *) SSL_get1_session() is like SSL_get_session(), but increments
7960 the reference count in the SSL_SESSION returned.
7961 [Geoff Thorpe <geoff@eu.c2.net>]
7963 *) Fix for 'req': it was adding a null to request attributes.
7964 Also change the X509_LOOKUP and X509_INFO code to handle
7965 certificate auxiliary information.
7968 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7972 *) Add the possibility to add extra information to the memory leak
7973 detecting output, to form tracebacks, showing from where each
7974 allocation was originated: CRYPTO_push_info("constant string") adds
7975 the string plus current file name and line number to a per-thread
7976 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7977 is like calling CYRPTO_pop_info() until the stack is empty.
7978 Also updated memory leak detection code to be multi-thread-safe.
7981 *) Add options -text and -noout to pkcs7 utility and delete the
7982 encryption options which never did anything. Update docs.
7985 *) Add options to some of the utilities to allow the pass phrase
7986 to be included on either the command line (not recommended on
7987 OSes like Unix) or read from the environment. Update the
7988 manpages and fix a few bugs.
7991 *) Add a few manpages for some of the openssl commands.
7994 *) Fix the -revoke option in ca. It was freeing up memory twice,
7995 leaking and not finding already revoked certificates.
7998 *) Extensive changes to support certificate auxiliary information.
7999 This involves the use of X509_CERT_AUX structure and X509_AUX
8000 functions. An X509_AUX function such as PEM_read_X509_AUX()
8001 can still read in a certificate file in the usual way but it
8002 will also read in any additional "auxiliary information". By
8003 doing things this way a fair degree of compatibility can be
8004 retained: existing certificates can have this information added
8005 using the new 'x509' options.
8007 Current auxiliary information includes an "alias" and some trust
8008 settings. The trust settings will ultimately be used in enhanced
8009 certificate chain verification routines: currently a certificate
8010 can only be trusted if it is self signed and then it is trusted
8014 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8015 The problem was that one of the replacement routines had not been working
8016 since SSLeay releases. For now the offending routine has been replaced
8017 with non-optimised assembler. Even so, this now gives around 95%
8018 performance improvement for 1024 bit RSA signs.
8021 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8022 handling. Most clients have the effective key size in bits equal to
8023 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8024 A few however don't do this and instead use the size of the decrypted key
8025 to determine the RC2 key length and the AlgorithmIdentifier to determine
8026 the effective key length. In this case the effective key length can still
8027 be 40 bits but the key length can be 168 bits for example. This is fixed
8028 by manually forcing an RC2 key into the EVP_PKEY structure because the
8029 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8030 the key length and effective key length are equal.
8033 *) Add a bunch of functions that should simplify the creation of
8034 X509_NAME structures. Now you should be able to do:
8035 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8036 and have it automatically work out the correct field type and fill in
8037 the structures. The more adventurous can try:
8038 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8039 and it will (hopefully) work out the correct multibyte encoding.
8042 *) Change the 'req' utility to use the new field handling and multibyte
8043 copy routines. Before the DN field creation was handled in an ad hoc
8044 way in req, ca, and x509 which was rather broken and didn't support
8045 BMPStrings or UTF8Strings. Since some software doesn't implement
8046 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8047 using the dirstring_type option. See the new comment in the default
8048 openssl.cnf for more info.
8051 *) Make crypto/rand/md_rand.c more robust:
8052 - Assure unique random numbers after fork().
8053 - Make sure that concurrent threads access the global counter and
8054 md serializably so that we never lose entropy in them
8055 or use exactly the same state in multiple threads.
8056 Access to the large state is not always serializable because
8057 the additional locking could be a performance killer, and
8058 md should be large enough anyway.
8061 *) New file apps/app_rand.c with commonly needed functionality
8062 for handling the random seed file.
8064 Use the random seed file in some applications that previously did not:
8066 dsaparam -genkey (which also ignored its '-rand' option),
8069 x509 (when signing).
8070 Except on systems with /dev/urandom, it is crucial to have a random
8071 seed file at least for key creation, DSA signing, and for DH exchanges;
8072 for RSA signatures we could do without one.
8074 gendh and gendsa (unlike genrsa) used to read only the first byte
8075 of each file listed in the '-rand' option. The function as previously
8076 found in genrsa is now in app_rand.c and is used by all programs
8077 that support '-rand'.
8080 *) In RAND_write_file, use mode 0600 for creating files;
8081 don't just chmod when it may be too late.
8084 *) Report an error from X509_STORE_load_locations
8085 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8088 *) New function ASN1_mbstring_copy() this copies a string in either
8089 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8090 into an ASN1_STRING type. A mask of permissible types is passed
8091 and it chooses the "minimal" type to use or an error if not type
8095 *) Add function equivalents to the various macros in asn1.h. The old
8096 macros are retained with an M_ prefix. Code inside the library can
8097 use the M_ macros. External code (including the openssl utility)
8098 should *NOT* in order to be "shared library friendly".
8101 *) Add various functions that can check a certificate's extensions
8102 to see if it usable for various purposes such as SSL client,
8103 server or S/MIME and CAs of these types. This is currently
8104 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8105 verification. Also added a -purpose flag to x509 utility to
8106 print out all the purposes.
8109 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8113 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8114 for, obtain and decode and extension and obtain its critical flag.
8115 This allows all the necessary extension code to be handled in a
8116 single function call.
8119 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8120 platforms. See crypto/rc4/rc4_enc.c for further details.
8123 *) New -noout option to asn1parse. This causes no output to be produced
8124 its main use is when combined with -strparse and -out to extract data
8125 from a file (which may not be in ASN.1 format).
8128 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8129 when producing the local key id.
8130 [Richard Levitte <levitte@stacken.kth.se>]
8132 *) New option -dhparam in s_server. This allows a DH parameter file to be
8133 stated explicitly. If it is not stated then it tries the first server
8134 certificate file. The previous behaviour hard coded the filename
8138 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8139 a public key to be input or output. For example:
8140 openssl rsa -in key.pem -pubout -out pubkey.pem
8141 Also added necessary DSA public key functions to handle this.
8144 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8145 in the message. This was handled by allowing
8146 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8147 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8149 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8150 to the end of the strings whereas this didn't. This would cause problems
8151 if strings read with d2i_ASN1_bytes() were later modified.
8152 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8154 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8155 data and it contains EOF it will end up returning an error. This is
8156 caused by input 46 bytes long. The cause is due to the way base64
8157 BIOs find the start of base64 encoded data. They do this by trying a
8158 trial decode on each line until they find one that works. When they
8159 do a flag is set and it starts again knowing it can pass all the
8160 data directly through the decoder. Unfortunately it doesn't reset
8161 the context it uses. This means that if EOF is reached an attempt
8162 is made to pass two EOFs through the context and this causes the
8163 resulting error. This can also cause other problems as well. As is
8164 usual with these problems it takes *ages* to find and the fix is
8165 trivial: move one line.
8166 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8168 *) Ugly workaround to get s_client and s_server working under Windows. The
8169 old code wouldn't work because it needed to select() on sockets and the
8170 tty (for keypresses and to see if data could be written). Win32 only
8171 supports select() on sockets so we select() with a 1s timeout on the
8172 sockets and then see if any characters are waiting to be read, if none
8173 are present then we retry, we also assume we can always write data to
8174 the tty. This isn't nice because the code then blocks until we've
8175 received a complete line of data and it is effectively polling the
8176 keyboard at 1s intervals: however it's quite a bit better than not
8177 working at all :-) A dedicated Windows application might handle this
8178 with an event loop for example.
8181 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8182 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8183 will be called when RSA_sign() and RSA_verify() are used. This is useful
8184 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8185 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8186 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8187 This necessitated the support of an extra signature type NID_md5_sha1
8188 for SSL signatures and modifications to the SSL library to use it instead
8189 of calling RSA_public_decrypt() and RSA_private_encrypt().
8192 *) Add new -verify -CAfile and -CApath options to the crl program, these
8193 will lookup a CRL issuers certificate and verify the signature in a
8194 similar way to the verify program. Tidy up the crl program so it
8195 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8196 less strict. It will now permit CRL extensions even if it is not
8197 a V2 CRL: this will allow it to tolerate some broken CRLs.
8200 *) Initialize all non-automatic variables each time one of the openssl
8201 sub-programs is started (this is necessary as they may be started
8202 multiple times from the "OpenSSL>" prompt).
8203 [Lennart Bang, Bodo Moeller]
8205 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8206 removing all other RSA functionality (this is what NO_RSA does). This
8207 is so (for example) those in the US can disable those operations covered
8208 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8212 *) Non-copying interface to BIO pairs.
8213 (still largely untested)
8216 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8217 ASCII string. This was handled independently in various places before.
8220 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8221 UTF8 strings a character at a time.
8224 *) Use client_version from client hello to select the protocol
8225 (s23_srvr.c) and for RSA client key exchange verification
8226 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8229 *) Add various utility functions to handle SPKACs, these were previously
8230 handled by poking round in the structure internals. Added new function
8231 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8232 print, verify and generate SPKACs. Based on an original idea from
8233 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8236 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8239 *) Allow the config file extension section to be overwritten on the
8240 command line. Based on an original idea from Massimiliano Pala
8241 <madwolf@comune.modena.it>. The new option is called -extensions
8242 and can be applied to ca, req and x509. Also -reqexts to override
8243 the request extensions in req and -crlexts to override the crl extensions
8247 *) Add new feature to the SPKAC handling in ca. Now you can include
8248 the same field multiple times by preceding it by "XXXX." for example:
8251 this is the same syntax as used in the req config file.
8254 *) Allow certificate extensions to be added to certificate requests. These
8255 are specified in a 'req_extensions' option of the req section of the
8256 config file. They can be printed out with the -text option to req but
8257 are otherwise ignored at present.
8260 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8261 data read consists of only the final block it would not decrypted because
8262 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8263 A misplaced 'break' also meant the decrypted final block might not be
8264 copied until the next read.
8267 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8268 a few extra parameters to the DH structure: these will be useful if
8269 for example we want the value of 'q' or implement X9.42 DH.
8272 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8273 provides hooks that allow the default DSA functions or functions on a
8274 "per key" basis to be replaced. This allows hardware acceleration and
8275 hardware key storage to be handled without major modification to the
8276 library. Also added low level modexp hooks and CRYPTO_EX structure and
8277 associated functions.
8280 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8281 as "read only": it can't be written to and the buffer it points to will
8282 not be freed. Reading from a read only BIO is much more efficient than
8283 a normal memory BIO. This was added because there are several times when
8284 an area of memory needs to be read from a BIO. The previous method was
8285 to create a memory BIO and write the data to it, this results in two
8286 copies of the data and an O(n^2) reading algorithm. There is a new
8287 function BIO_new_mem_buf() which creates a read only memory BIO from
8288 an area of memory. Also modified the PKCS#7 routines to use read only
8292 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8293 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8294 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8295 but a retry condition occured while trying to read the rest.
8298 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8299 NID_pkcs7_encrypted by default: this was wrong since this should almost
8300 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8301 the encrypted data type: this is a more sensible place to put it and it
8302 allows the PKCS#12 code to be tidied up that duplicated this
8306 *) Changed obj_dat.pl script so it takes its input and output files on
8307 the command line. This should avoid shell escape redirection problems
8311 *) Initial support for certificate extension requests, these are included
8312 in things like Xenroll certificate requests. Included functions to allow
8313 extensions to be obtained and added.
8316 *) -crlf option to s_client and s_server for sending newlines as
8317 CRLF (as required by many protocols).
8320 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8322 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8323 [Ralf S. Engelschall]
8325 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8326 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8328 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8332 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8333 DH parameters/keys (q is lost during that conversion, but the resulting
8334 DH parameters contain its length).
8336 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8337 much faster than DH_generate_parameters (which creates parameters
8338 where p = 2*q + 1), and also the smaller q makes DH computations
8339 much more efficient (160-bit exponentiation instead of 1024-bit
8340 exponentiation); so this provides a convenient way to support DHE
8341 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8342 utter importance to use
8343 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8345 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8346 when such DH parameters are used, because otherwise small subgroup
8347 attacks may become possible!
8350 *) Avoid memory leak in i2d_DHparams.
8353 *) Allow the -k option to be used more than once in the enc program:
8354 this allows the same encrypted message to be read by multiple recipients.
8357 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8358 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8359 it will always use the numerical form of the OID, even if it has a short
8363 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8364 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8365 otherwise bn_mod_exp was called. In the case of hardware keys for example
8366 no private key components need be present and it might store extra data
8367 in the RSA structure, which cannot be accessed from bn_mod_exp.
8368 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8369 private key operations.
8372 *) Added support for SPARC Linux.
8375 *) pem_password_cb function type incompatibly changed from
8376 typedef int pem_password_cb(char *buf, int size, int rwflag);
8378 ....(char *buf, int size, int rwflag, void *userdata);
8379 so that applications can pass data to their callbacks:
8380 The PEM[_ASN1]_{read,write}... functions and macros now take an
8381 additional void * argument, which is just handed through whenever
8382 the password callback is called.
8383 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8385 New function SSL_CTX_set_default_passwd_cb_userdata.
8387 Compatibility note: As many C implementations push function arguments
8388 onto the stack in reverse order, the new library version is likely to
8389 interoperate with programs that have been compiled with the old
8390 pem_password_cb definition (PEM_whatever takes some data that
8391 happens to be on the stack as its last argument, and the callback
8392 just ignores this garbage); but there is no guarantee whatsoever that
8395 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8396 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8397 problems not only on Windows, but also on some Unix platforms.
8398 To avoid problematic command lines, these definitions are now in an
8399 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8400 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8403 *) MIPS III/IV assembler module is reimplemented.
8406 *) More DES library cleanups: remove references to srand/rand and
8407 delete an unused file.
8410 *) Add support for the the free Netwide assembler (NASM) under Win32,
8411 since not many people have MASM (ml) and it can be hard to obtain.
8412 This is currently experimental but it seems to work OK and pass all
8413 the tests. Check out INSTALL.W32 for info.
8416 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8417 without temporary keys kept an extra copy of the server key,
8418 and connections with temporary keys did not free everything in case
8422 *) New function RSA_check_key and new openssl rsa option -check
8423 for verifying the consistency of RSA keys.
8424 [Ulf Moeller, Bodo Moeller]
8426 *) Various changes to make Win32 compile work:
8427 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8428 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8429 comparison" warnings.
8430 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8433 *) Add a debugging option to PKCS#5 v2 key generation function: when
8434 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8435 derived keys are printed to stderr.
8438 *) Copy the flags in ASN1_STRING_dup().
8439 [Roman E. Pavlov <pre@mo.msk.ru>]
8441 *) The x509 application mishandled signing requests containing DSA
8442 keys when the signing key was also DSA and the parameters didn't match.
8444 It was supposed to omit the parameters when they matched the signing key:
8445 the verifying software was then supposed to automatically use the CA's
8446 parameters if they were absent from the end user certificate.
8448 Omitting parameters is no longer recommended. The test was also
8449 the wrong way round! This was probably due to unusual behaviour in
8450 EVP_cmp_parameters() which returns 1 if the parameters match.
8451 This meant that parameters were omitted when they *didn't* match and
8452 the certificate was useless. Certificates signed with 'ca' didn't have
8454 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8456 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8457 The interface is as follows:
8458 Applications can use
8459 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8460 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8461 "off" is now the default.
8462 The library internally uses
8463 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8464 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8465 to disable memory-checking temporarily.
8467 Some inconsistent states that previously were possible (and were
8468 even the default) are now avoided.
8470 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8471 with each memory chunk allocated; this is occasionally more helpful
8472 than just having a counter.
8474 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8476 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8480 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8481 which largely parallels "options", but is for changing API behaviour,
8482 whereas "options" are about protocol behaviour.
8483 Initial "mode" flags are:
8485 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8486 a single record has been written.
8487 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8488 retries use the same buffer location.
8489 (But all of the contents must be
8493 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8496 *) Fix problems with no-hmac etc.
8497 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8499 *) New functions RSA_get_default_method(), RSA_set_method() and
8500 RSA_get_method(). These allows replacement of RSA_METHODs without having
8501 to mess around with the internals of an RSA structure.
8504 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8505 Also really enable memory leak checks in openssl.c and in some
8507 [Chad C. Mulligan, Bodo Moeller]
8509 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8510 up the length of negative integers. This has now been simplified to just
8511 store the length when it is first determined and use it later, rather
8512 than trying to keep track of where data is copied and updating it to
8514 [Steve Henson, reported by Brien Wheeler
8515 <bwheeler@authentica-security.com>]
8517 *) Add a new function PKCS7_signatureVerify. This allows the verification
8518 of a PKCS#7 signature but with the signing certificate passed to the
8519 function itself. This contrasts with PKCS7_dataVerify which assumes the
8520 certificate is present in the PKCS#7 structure. This isn't always the
8521 case: certificates can be omitted from a PKCS#7 structure and be
8522 distributed by "out of band" means (such as a certificate database).
8525 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8526 function prototypes in pem.h, also change util/mkdef.pl to add the
8527 necessary function names.
8530 *) mk1mf.pl (used by Windows builds) did not properly read the
8531 options set by Configure in the top level Makefile, and Configure
8532 was not even able to write more than one option correctly.
8533 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8536 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8537 file to be loaded from a BIO or FILE pointer. The BIO version will
8538 for example allow memory BIOs to contain config info.
8541 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8542 Whoever hopes to achieve shared-library compatibility across versions
8543 must use this, not the compile-time macro.
8544 (Exercise 0.9.4: Which is the minimum library version required by
8546 Note: All this applies only to multi-threaded programs, others don't
8550 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8551 through a BIO pair triggered the default case, i.e.
8552 SSLerr(...,SSL_R_UNKNOWN_STATE).
8555 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8556 can use the SSL library even if none of the specific BIOs is
8560 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8561 for the encoded length.
8562 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8564 *) Add initial documentation of the X509V3 functions.
8567 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8568 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8569 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8570 secure PKCS#8 private key format with a high iteration count.
8573 *) Fix determination of Perl interpreter: A perl or perl5
8574 _directory_ in $PATH was also accepted as the interpreter.
8575 [Ralf S. Engelschall]
8577 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8578 wrong with it but it was very old and did things like calling
8579 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8583 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8584 to use the new extension code.
8587 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8588 with macros. This should make it easier to change their form, add extra
8589 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8593 *) Add to configuration table a new entry that can specify an alternative
8594 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8595 according to Mark Crispin <MRC@Panda.COM>.
8599 *) DES CBC did not update the IV. Weird.
8602 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8603 Changing the behaviour of the former might break existing programs --
8604 where IV updating is needed, des_ncbc_encrypt can be used.
8607 *) When bntest is run from "make test" it drives bc to check its
8608 calculations, as well as internally checking them. If an internal check
8609 fails, it needs to cause bc to give a non-zero result or make test carries
8610 on without noticing the failure. Fixed.
8613 *) DES library cleanups.
8616 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8617 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8618 ciphers. NOTE: although the key derivation function has been verified
8619 against some published test vectors it has not been extensively tested
8620 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8624 *) Instead of "mkdir -p", which is not fully portable, use new
8625 Perl script "util/mkdir-p.pl".
8628 *) Rewrite the way password based encryption (PBE) is handled. It used to
8629 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8630 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8631 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8632 the 'parameter' field of the AlgorithmIdentifier is passed to the
8633 underlying key generation function so it must do its own ASN1 parsing.
8634 This has also changed the EVP_PBE_CipherInit() function which now has a
8635 'parameter' argument instead of literal salt and iteration count values
8636 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8639 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8640 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8641 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8642 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8643 value was just used as a "magic string" and not used directly its
8644 value doesn't matter.
8647 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8651 *) "linux-sparc64" configuration (ultrapenguin).
8652 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8653 "linux-sparc" configuration.
8654 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8656 *) config now generates no-xxx options for missing ciphers.
8659 *) Support the EBCDIC character set (work in progress).
8660 File ebcdic.c not yet included because it has a different license.
8661 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8663 *) Support BS2000/OSD-POSIX.
8664 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8666 *) Make callbacks for key generation use void * instead of char *.
8669 *) Make S/MIME samples compile (not yet tested).
8672 *) Additional typesafe stacks.
8675 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8679 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8681 *) New configuration variant "sco5-gcc".
8683 *) Updated some demos.
8684 [Sean O Riordain, Wade Scholine]
8686 *) Add missing BIO_free at exit of pkcs12 application.
8689 *) Fix memory leak in conf.c.
8692 *) Updates for Win32 to assembler version of MD5.
8695 *) Set #! path to perl in apps/der_chop to where we found it
8696 instead of using a fixed path.
8699 *) SHA library changes for irix64-mips4-cc.
8702 *) Improvements for VMS support.
8706 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8708 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8709 This also avoids the problems with SC4.2 and unpatched SC5.
8710 [Andy Polyakov <appro@fy.chalmers.se>]
8712 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8713 These are required because of the typesafe stack would otherwise break
8714 existing code. If old code used a structure member which used to be STACK
8715 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8716 sk_num or sk_value it would produce an error because the num, data members
8717 are not present in STACK_OF. Now it just produces a warning. sk_set
8718 replaces the old method of assigning a value to sk_value
8719 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8720 that does this will no longer work (and should use sk_set instead) but
8721 this could be regarded as a "questionable" behaviour anyway.
8724 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8725 correctly handle encrypted S/MIME data.
8728 *) Change type of various DES function arguments from des_cblock
8729 (which means, in function argument declarations, pointer to char)
8730 to des_cblock * (meaning pointer to array with 8 char elements),
8731 which allows the compiler to do more typechecking; it was like
8732 that back in SSLeay, but with lots of ugly casts.
8734 Introduce new type const_des_cblock.
8737 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8738 problems: find RecipientInfo structure that matches recipient certificate
8739 and initialise the ASN1 structures properly based on passed cipher.
8742 *) Belatedly make the BN tests actually check the results.
8745 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8746 to and from BNs: it was completely broken. New compilation option
8747 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8748 key elements as negative integers.
8751 *) Reorganize and speed up MD5.
8752 [Andy Polyakov <appro@fy.chalmers.se>]
8755 [Richard Levitte <richard@levitte.org>]
8757 *) New option -out to asn1parse to allow the parsed structure to be
8758 output to a file. This is most useful when combined with the -strparse
8759 option to examine the output of things like OCTET STRINGS.
8762 *) Make SSL library a little more fool-proof by not requiring any longer
8763 that SSL_set_{accept,connect}_state be called before
8764 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8765 in many applications because usually everything *appeared* to work as
8766 intended anyway -- now it really works as intended).
8769 *) Move openssl.cnf out of lib/.
8772 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8773 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8774 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8775 [Ralf S. Engelschall]
8777 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8778 handle PKCS#7 enveloped data properly.
8779 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8781 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8782 copying pointers. The cert_st handling is changed by this in
8783 various ways (and thus what used to be known as ctx->default_cert
8784 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8785 any longer when s->cert does not give us what we need).
8786 ssl_cert_instantiate becomes obsolete by this change.
8787 As soon as we've got the new code right (possibly it already is?),
8788 we have solved a couple of bugs of the earlier code where s->cert
8789 was used as if it could not have been shared with other SSL structures.
8791 Note that using the SSL API in certain dirty ways now will result
8792 in different behaviour than observed with earlier library versions:
8793 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8794 does not influence s as it used to.
8796 In order to clean up things more thoroughly, inside SSL_SESSION
8797 we don't use CERT any longer, but a new structure SESS_CERT
8798 that holds per-session data (if available); currently, this is
8799 the peer's certificate chain and, for clients, the server's certificate
8800 and temporary key. CERT holds only those values that can have
8801 meaningful defaults in an SSL_CTX.
8804 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8805 from the internal representation. Various PKCS#7 fixes: remove some
8806 evil casts and set the enc_dig_alg field properly based on the signing
8810 *) Allow PKCS#12 password to be set from the command line or the
8811 environment. Let 'ca' get its config file name from the environment
8812 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8816 *) Allow certificate policies extension to use an IA5STRING for the
8817 organization field. This is contrary to the PKIX definition but
8818 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8822 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8823 without disallowing inline assembler and the like for non-pedantic builds.
8826 *) Support Borland C++ builder.
8827 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8832 *) SHA-1 cleanups and performance enhancements.
8833 [Andy Polyakov <appro@fy.chalmers.se>]
8835 *) Sparc v8plus assembler for the bignum library.
8836 [Andy Polyakov <appro@fy.chalmers.se>]
8838 *) Accept any -xxx and +xxx compiler options in Configure.
8841 *) Update HPUX configuration.
8844 *) Add missing sk_<type>_unshift() function to safestack.h
8845 [Ralf S. Engelschall]
8847 *) New function SSL_CTX_use_certificate_chain_file that sets the
8848 "extra_cert"s in addition to the certificate. (This makes sense
8849 only for "PEM" format files, as chains as a whole are not
8853 *) Support verify_depth from the SSL API.
8854 x509_vfy.c had what can be considered an off-by-one-error:
8855 Its depth (which was not part of the external interface)
8856 was actually counting the number of certificates in a chain;
8857 now it really counts the depth.
8860 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8861 instead of X509err, which often resulted in confusing error
8862 messages since the error codes are not globally unique
8863 (e.g. an alleged error in ssl3_accept when a certificate
8864 didn't match the private key).
8866 *) New function SSL_CTX_set_session_id_context that allows to set a default
8867 value (so that you don't need SSL_set_session_id_context for each
8868 connection using the SSL_CTX).
8871 *) OAEP decoding bug fix.
8874 *) Support INSTALL_PREFIX for package builders, as proposed by
8878 *) New Configure options "threads" and "no-threads". For systems
8879 where the proper compiler options are known (currently Solaris
8880 and Linux), "threads" is the default.
8883 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8886 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8887 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8888 such as /usr/local/bin.
8891 *) "make linux-shared" to build shared libraries.
8892 [Niels Poppe <niels@netbox.org>]
8894 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8897 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8898 extension adding in x509 utility.
8901 *) Remove NOPROTO sections and error code comments.
8904 *) Partial rewrite of the DEF file generator to now parse the ANSI
8908 *) New Configure options --prefix=DIR and --openssldir=DIR.
8911 *) Complete rewrite of the error code script(s). It is all now handled
8912 by one script at the top level which handles error code gathering,
8913 header rewriting and C source file generation. It should be much better
8914 than the old method: it now uses a modified version of Ulf's parser to
8915 read the ANSI prototypes in all header files (thus the old K&R definitions
8916 aren't needed for error creation any more) and do a better job of
8917 translating function codes into names. The old 'ASN1 error code imbedded
8918 in a comment' is no longer necessary and it doesn't use .err files which
8919 have now been deleted. Also the error code call doesn't have to appear all
8920 on one line (which resulted in some large lines...).
8923 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8926 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8927 0 (which usually indicates a closed connection), but continue reading.
8930 *) Fix some race conditions.
8933 *) Add support for CRL distribution points extension. Add Certificate
8934 Policies and CRL distribution points documentation.
8937 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8940 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8941 8 of keying material. Merlin has also confirmed interop with this fix
8942 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8943 [Merlin Hughes <merlin@baltimore.ie>]
8945 *) Fix lots of warnings.
8946 [Richard Levitte <levitte@stacken.kth.se>]
8948 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8949 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8950 [Richard Levitte <levitte@stacken.kth.se>]
8952 *) Fix problems with sizeof(long) == 8.
8953 [Andy Polyakov <appro@fy.chalmers.se>]
8955 *) Change functions to ANSI C.
8958 *) Fix typos in error codes.
8959 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8961 *) Remove defunct assembler files from Configure.
8964 *) SPARC v8 assembler BIGNUM implementation.
8965 [Andy Polyakov <appro@fy.chalmers.se>]
8967 *) Support for Certificate Policies extension: both print and set.
8968 Various additions to support the r2i method this uses.
8971 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8972 return a const string when you are expecting an allocated buffer.
8975 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8976 types DirectoryString and DisplayText.
8979 *) Add code to allow r2i extensions to access the configuration database,
8980 add an LHASH database driver and add several ctx helper functions.
8983 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8984 fail when they extended the size of a BIGNUM.
8987 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8988 support typesafe stack.
8991 *) Fix typo in SSL_[gs]et_options().
8992 [Nils Frostberg <nils@medcom.se>]
8994 *) Delete various functions and files that belonged to the (now obsolete)
8995 old X509V3 handling code.
8998 *) New Configure option "rsaref".
9001 *) Don't auto-generate pem.h.
9004 *) Introduce type-safe ASN.1 SETs.
9007 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9008 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9010 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9011 that links with OpenSSL (well at least cause lots of warnings), but fear
9012 not: the conversion is trivial, and it eliminates loads of evil casts. A
9013 few STACKed things have been converted already. Feel free to convert more.
9014 In the fullness of time, I'll do away with the STACK type altogether.
9017 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9018 specified in <certfile> by updating the entry in the index.txt file.
9019 This way one no longer has to edit the index.txt file manually for
9020 revoking a certificate. The -revoke option does the gory details now.
9021 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9023 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9024 `-text' option at all and this way the `-noout -text' combination was
9025 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9026 [Ralf S. Engelschall]
9028 *) Make sure a corresponding plain text error message exists for the
9029 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9030 verify callback function determined that a certificate was revoked.
9031 [Ralf S. Engelschall]
9033 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9034 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9035 all available cipers including rc5, which was forgotten until now.
9036 In order to let the testing shell script know which algorithms
9037 are available, a new (up to now undocumented) command
9038 "openssl list-cipher-commands" is used.
9041 *) Bugfix: s_client occasionally would sleep in select() when
9042 it should have checked SSL_pending() first.
9045 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9046 the raw DSA values prior to ASN.1 encoding.
9049 *) Tweaks to Configure
9050 [Niels Poppe <niels@netbox.org>]
9052 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9056 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9059 *) New config option to avoid instructions that are illegal on the 80386.
9060 The default code is faster, but requires at least a 486.
9063 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9064 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9065 same as SSL2_VERSION anyway.
9068 *) New "-showcerts" option for s_client.
9071 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9072 application. Various cleanups and fixes.
9075 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9076 modify error routines to work internally. Add error codes and PBE init
9077 to library startup routines.
9080 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9081 packing functions to asn1 and evp. Changed function names and error
9082 codes along the way.
9085 *) PKCS12 integration: and so it begins... First of several patches to
9086 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9087 objects to objects.h
9090 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9091 and display support for Thawte strong extranet extension.
9094 *) Add LinuxPPC support.
9095 [Jeff Dubrule <igor@pobox.org>]
9097 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9098 bn_div_words in alpha.s.
9099 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9101 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9102 OAEP isn't supported when OpenSSL is built with RSAref.
9103 [Ulf Moeller <ulf@fitug.de>]
9105 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9106 so they no longer are missing under -DNOPROTO.
9107 [Soren S. Jorvang <soren@t.dk>]
9110 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9112 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9113 doesn't work when the session is reused. Coming soon!
9116 *) Fix a security hole, that allows sessions to be reused in the wrong
9117 context thus bypassing client cert protection! All software that uses
9118 client certs and session caches in multiple contexts NEEDS PATCHING to
9119 allow session reuse! A fuller solution is in the works.
9120 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9122 *) Some more source tree cleanups (removed obsolete files
9123 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9124 permission on "config" script to be executable) and a fix for the INSTALL
9126 [Ulf Moeller <ulf@fitug.de>]
9128 *) Remove some legacy and erroneous uses of malloc, free instead of
9130 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9132 *) Make rsa_oaep_test return non-zero on error.
9133 [Ulf Moeller <ulf@fitug.de>]
9135 *) Add support for native Solaris shared libraries. Configure
9136 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9137 if someone would make that last step automatic.
9138 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9140 *) ctx_size was not built with the right compiler during "make links". Fixed.
9143 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9144 except NULL ciphers". This means the default cipher list will no longer
9145 enable NULL ciphers. They need to be specifically enabled e.g. with
9146 the string "DEFAULT:eNULL".
9149 *) Fix to RSA private encryption routines: if p < q then it would
9150 occasionally produce an invalid result. This will only happen with
9151 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9154 *) Be less restrictive and allow also `perl util/perlpath.pl
9155 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9156 because this way one can also use an interpreter named `perl5' (which is
9157 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9158 installed as `perl').
9159 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9161 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9162 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9164 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9165 advapi32.lib to Win32 build and change the pem test comparision
9166 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9167 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9168 and crypto/des/ede_cbcm_enc.c.
9171 *) DES quad checksum was broken on big-endian architectures. Fixed.
9174 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9175 Win32 test batch file so it (might) work again. The Win32 test batch file
9176 is horrible: I feel ill....
9179 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9180 in e_os.h. Audit of header files to check ANSI and non ANSI
9181 sections: 10 functions were absent from non ANSI section and not exported
9182 from Windows DLLs. Fixed up libeay.num for new functions.
9185 *) Make `openssl version' output lines consistent.
9186 [Ralf S. Engelschall]
9188 *) Fix Win32 symbol export lists for BIO functions: Added
9189 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9190 to ms/libeay{16,32}.def.
9191 [Ralf S. Engelschall]
9193 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9194 fine under Unix and passes some trivial tests I've now added. But the
9195 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9196 added to make sure no one expects that this stuff really works in the
9197 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9198 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9200 [Ralf S. Engelschall]
9202 *) Fix the generation of two part addresses in perl.
9203 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9205 *) Add config entry for Linux on MIPS.
9206 [John Tobey <jtobey@channel1.com>]
9208 *) Make links whenever Configure is run, unless we are on Windoze.
9211 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9212 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9216 *) Add a useful kludge to allow package maintainers to specify compiler and
9217 other platforms details on the command line without having to patch the
9218 Configure script everytime: One now can use ``perl Configure
9219 <id>:<details>'', i.e. platform ids are allowed to have details appended
9220 to them (seperated by colons). This is treated as there would be a static
9221 pre-configured entry in Configure's %table under key <id> with value
9222 <details> and ``perl Configure <id>'' is called. So, when you want to
9223 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9224 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9225 now, which overrides the FreeBSD-elf entry on-the-fly.
9226 [Ralf S. Engelschall]
9228 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9231 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9232 on the `perl Configure ...' command line. This way one can compile
9233 OpenSSL libraries with Position Independent Code (PIC) which is needed
9234 for linking it into DSOs.
9235 [Ralf S. Engelschall]
9237 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9241 *) Cleaned up the LICENSE document: The official contact for any license
9242 questions now is the OpenSSL core team under openssl-core@openssl.org.
9243 And add a paragraph about the dual-license situation to make sure people
9244 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9245 to the OpenSSL toolkit.
9246 [Ralf S. Engelschall]
9248 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9249 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9250 Additonally cleaned up the `make links' target: Remove unnecessary
9251 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9252 to speed processing and no longer clutter the display with confusing
9253 stuff. Instead only the actually done links are displayed.
9254 [Ralf S. Engelschall]
9256 *) Permit null encryption ciphersuites, used for authentication only. It used
9257 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9258 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9262 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9263 signed attributes when verifying signatures (this would break them),
9264 the detached data encoding was wrong and public keys obtained using
9265 X509_get_pubkey() weren't freed.
9268 *) Add text documentation for the BUFFER functions. Also added a work around
9269 to a Win95 console bug. This was triggered by the password read stuff: the
9270 last character typed gets carried over to the next fread(). If you were
9271 generating a new cert request using 'req' for example then the last
9272 character of the passphrase would be CR which would then enter the first
9276 *) Added the new `Includes OpenSSL Cryptography Software' button as
9277 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9278 button and can be used by applications based on OpenSSL to show the
9279 relationship to the OpenSSL project.
9280 [Ralf S. Engelschall]
9282 *) Remove confusing variables in function signatures in files
9283 ssl/ssl_lib.c and ssl/ssl.h.
9284 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9286 *) Don't install bss_file.c under PREFIX/include/
9287 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9289 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9290 functions that return function pointers and has support for NT specific
9291 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9292 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9293 unsigned to signed types: this was killing the Win32 compile.
9296 *) Add new certificate file to stack functions,
9297 SSL_add_dir_cert_subjects_to_stack() and
9298 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9299 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9300 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9301 This means that Apache-SSL and similar packages don't have to mess around
9302 to add as many CAs as they want to the preferred list.
9305 *) Experiment with doxygen documentation. Currently only partially applied to
9307 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9308 openssl.doxy as the configuration file.
9311 *) Get rid of remaining C++-style comments which strict C compilers hate.
9312 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9314 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9315 compiled in by default: it has problems with large keys.
9318 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9319 DH private keys and/or callback functions which directly correspond to
9320 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9321 is needed for applications which have to configure certificates on a
9322 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9324 For the RSA certificate situation is makes no difference, but
9325 for the DSA certificate situation this fixes the "no shared cipher"
9326 problem where the OpenSSL cipher selection procedure failed because the
9327 temporary keys were not overtaken from the context and the API provided
9328 no way to reconfigure them.
9329 The new functions now let applications reconfigure the stuff and they
9330 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9331 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9332 non-public-API function ssl_cert_instantiate() is used as a helper
9333 function and also to reduce code redundancy inside ssl_rsa.c.
9334 [Ralf S. Engelschall]
9336 *) Move s_server -dcert and -dkey options out of the undocumented feature
9337 area because they are useful for the DSA situation and should be
9338 recognized by the users.
9339 [Ralf S. Engelschall]
9341 *) Fix the cipher decision scheme for export ciphers: the export bits are
9342 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9343 SSL_EXP_MASK. So, the original variable has to be used instead of the
9344 already masked variable.
9345 [Richard Levitte <levitte@stacken.kth.se>]
9347 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9348 [Richard Levitte <levitte@stacken.kth.se>]
9350 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9351 from `int' to `unsigned int' because it's a length and initialized by
9352 EVP_DigestFinal() which expects an `unsigned int *'.
9353 [Richard Levitte <levitte@stacken.kth.se>]
9355 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9356 script. Instead use the usual Shell->Perl transition trick.
9357 [Ralf S. Engelschall]
9359 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9360 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9361 -noout -modulus' as it's already the case for `openssl rsa -noout
9362 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9363 currently the public key is printed (a decision which was already done by
9364 `openssl dsa -modulus' in the past) which serves a similar purpose.
9365 Additionally the NO_RSA no longer completely removes the whole -modulus
9366 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9368 [Ralf S. Engelschall]
9370 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9371 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9372 [Arne Ansper <arne@ats.cyber.ee>]
9374 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9375 to be added. Now both 'req' and 'ca' can use new objects defined in the
9379 *) Add cool BIO that does syslog (or event log on NT).
9380 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9382 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9383 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9384 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9385 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9388 *) Add preliminary config info for new extension code.
9391 *) Make RSA_NO_PADDING really use no padding.
9392 [Ulf Moeller <ulf@fitug.de>]
9394 *) Generate errors when private/public key check is done.
9397 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9398 for some CRL extensions and new objects added.
9401 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9402 key usage extension and fuller support for authority key id.
9405 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9406 padding method for RSA, which is recommended for new applications in PKCS
9407 #1 v2.0 (RFC 2437, October 1998).
9408 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9409 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9410 against Bleichbacher's attack on RSA.
9411 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9414 *) Updates to the new SSL compression code
9415 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9417 *) Fix so that the version number in the master secret, when passed
9418 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9419 (because the server will not accept higher), that the version number
9420 is 0x03,0x01, not 0x03,0x00
9421 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9423 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9424 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9425 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9428 *) Support for RAW extensions where an arbitrary extension can be
9429 created by including its DER encoding. See apps/openssl.cnf for
9433 *) Make sure latest Perl versions don't interpret some generated C array
9434 code as Perl array code in the crypto/err/err_genc.pl script.
9435 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9437 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9438 not many people have the assembler. Various Win32 compilation fixes and
9439 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9443 *) Modify configure script 'Configure' to automatically create crypto/date.h
9444 file under Win32 and also build pem.h from pem.org. New script
9445 util/mkfiles.pl to create the MINFO file on environments that can't do a
9446 'make files': perl util/mkfiles.pl >MINFO should work.
9449 *) Major rework of DES function declarations, in the pursuit of correctness
9450 and purity. As a result, many evil casts evaporated, and some weirdness,
9451 too. You may find this causes warnings in your code. Zapping your evil
9452 casts will probably fix them. Mostly.
9455 *) Fix for a typo in asn1.h. Bug fix to object creation script
9456 obj_dat.pl. It considered a zero in an object definition to mean
9457 "end of object": none of the objects in objects.h have any zeros
9458 so it wasn't spotted.
9459 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9461 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9462 Masking (CBCM). In the absence of test vectors, the best I have been able
9463 to do is check that the decrypt undoes the encrypt, so far. Send me test
9464 vectors if you have them.
9467 *) Correct calculation of key length for export ciphers (too much space was
9468 allocated for null ciphers). This has not been tested!
9471 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9472 message is now correct (it understands "crypto" and "ssl" on its
9473 command line). There is also now an "update" option. This will update
9474 the util/ssleay.num and util/libeay.num files with any new functions.
9476 perl util/mkdef.pl crypto ssl update
9477 it will update them.
9480 *) Overhauled the Perl interface (perl/*):
9481 - ported BN stuff to OpenSSL's different BN library
9482 - made the perl/ source tree CVS-aware
9483 - renamed the package from SSLeay to OpenSSL (the files still contain
9484 their history because I've copied them in the repository)
9485 - removed obsolete files (the test scripts will be replaced
9486 by better Test::Harness variants in the future)
9487 [Ralf S. Engelschall]
9489 *) First cut for a very conservative source tree cleanup:
9490 1. merge various obsolete readme texts into doc/ssleay.txt
9491 where we collect the old documents and readme texts.
9492 2. remove the first part of files where I'm already sure that we no
9493 longer need them because of three reasons: either they are just temporary
9494 files which were left by Eric or they are preserved original files where
9495 I've verified that the diff is also available in the CVS via "cvs diff
9496 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9497 the crypto/md/ stuff).
9498 [Ralf S. Engelschall]
9500 *) More extension code. Incomplete support for subject and issuer alt
9501 name, issuer and authority key id. Change the i2v function parameters
9502 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9503 what that's for :-) Fix to ASN1 macro which messed up
9504 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9507 *) Preliminary support for ENUMERATED type. This is largely copied from the
9511 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9512 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9514 *) Make sure `make rehash' target really finds the `openssl' program.
9515 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9517 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9518 like to hear about it if this slows down other processors.
9521 *) Add CygWin32 platform information to Configure script.
9522 [Alan Batie <batie@aahz.jf.intel.com>]
9524 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9525 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9527 *) New program nseq to manipulate netscape certificate sequences
9530 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9534 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9535 but the BN code had some problems that would cause failures when
9536 doing certificate verification and some other functions.
9537 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9539 *) Add ASN1 and PEM code to support netscape certificate sequences.
9542 *) Add ASN1 and PEM code to support netscape certificate sequences.
9545 *) Add several PKIX and private extended key usage OIDs.
9548 *) Modify the 'ca' program to handle the new extension code. Modify
9549 openssl.cnf for new extension format, add comments.
9552 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9553 and add a sample to openssl.cnf so req -x509 now adds appropriate
9557 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9558 error code, add initial support to X509_print() and x509 application.
9561 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9562 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9563 stuff is currently isolated and isn't even compiled yet.
9566 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9567 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9568 Removed the versions check from X509 routines when loading extensions:
9569 this allows certain broken certificates that don't set the version
9570 properly to be processed.
9573 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9574 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9575 can still be regenerated with "make depend".
9578 *) Spelling mistake in C version of CAST-128.
9579 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9581 *) Changes to the error generation code. The perl script err-code.pl
9582 now reads in the old error codes and retains the old numbers, only
9583 adding new ones if necessary. It also only changes the .err files if new
9584 codes are added. The makefiles have been modified to only insert errors
9585 when needed (to avoid needlessly modifying header files). This is done
9586 by only inserting errors if the .err file is newer than the auto generated
9587 C file. To rebuild all the error codes from scratch (the old behaviour)
9588 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9589 or delete all the .err files.
9592 *) CAST-128 was incorrectly implemented for short keys. The C version has
9593 been fixed, but is untested. The assembler versions are also fixed, but
9594 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9595 to regenerate it if needed.
9596 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9597 Hagino <itojun@kame.net>]
9599 *) File was opened incorrectly in randfile.c.
9600 [Ulf Möller <ulf@fitug.de>]
9602 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9603 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9604 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9605 al: it's just almost always a UTCTime. Note this patch adds new error
9606 codes so do a "make errors" if there are problems.
9609 *) Correct Linux 1 recognition in config.
9610 [Ulf Möller <ulf@fitug.de>]
9612 *) Remove pointless MD5 hash when using DSA keys in ca.
9613 [Anonymous <nobody@replay.com>]
9615 *) Generate an error if given an empty string as a cert directory. Also
9616 generate an error if handed NULL (previously returned 0 to indicate an
9617 error, but didn't set one).
9618 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9620 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9623 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9624 parameters. This was causing a warning which killed off the Win32 compile.
9627 *) Remove C++ style comments from crypto/bn/bn_local.h.
9628 [Neil Costigan <neil.costigan@celocom.com>]
9630 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9631 based on a text string, looking up short and long names and finally
9632 "dot" format. The "dot" format stuff didn't work. Added new function
9633 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9634 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9635 OID is not part of the table.
9638 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9639 X509_LOOKUP_by_alias().
9642 *) Sort openssl functions by name.
9645 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9646 encryption from sample DSA keys (in case anyone is interested the password
9650 *) Make _all_ *_free functions accept a NULL pointer.
9651 [Frans Heymans <fheymans@isaserver.be>]
9653 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9655 [Anonymous <nobody@replay.com>]
9657 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9658 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9660 *) Don't blow it for numeric -newkey arguments to apps/req.
9661 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9663 *) Temp key "for export" tests were wrong in s3_srvr.c.
9664 [Anonymous <nobody@replay.com>]
9666 *) Add prototype for temp key callback functions
9667 SSL_CTX_set_tmp_{rsa,dh}_callback().
9670 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9671 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9674 *) X509_name_add_entry() freed the wrong thing after an error.
9675 [Arne Ansper <arne@ats.cyber.ee>]
9677 *) rsa_eay.c would attempt to free a NULL context.
9678 [Arne Ansper <arne@ats.cyber.ee>]
9680 *) BIO_s_socket() had a broken should_retry() on Windoze.
9681 [Arne Ansper <arne@ats.cyber.ee>]
9683 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9684 [Arne Ansper <arne@ats.cyber.ee>]
9686 *) Make sure the already existing X509_STORE->depth variable is initialized
9687 in X509_STORE_new(), but document the fact that this variable is still
9688 unused in the certificate verification process.
9689 [Ralf S. Engelschall]
9691 *) Fix the various library and apps files to free up pkeys obtained from
9692 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9695 *) Fix reference counting in X509_PUBKEY_get(). This makes
9696 demos/maurice/example2.c work, amongst others, probably.
9697 [Steve Henson and Ben Laurie]
9699 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9700 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9701 are no longer created. This way we have a single and consistent command
9702 line interface `openssl <command>', similar to `cvs <command>'.
9703 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9705 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9706 BIT STRING wrapper always have zero unused bits.
9709 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9712 *) Make the top-level INSTALL documentation easier to understand.
9715 *) Makefiles updated to exit if an error occurs in a sub-directory
9716 make (including if user presses ^C) [Paul Sutton]
9718 *) Make Montgomery context stuff explicit in RSA data structure.
9721 *) Fix build order of pem and err to allow for generated pem.h.
9724 *) Fix renumbering bug in X509_NAME_delete_entry().
9727 *) Enhanced the err-ins.pl script so it makes the error library number
9728 global and can add a library name. This is needed for external ASN1 and
9729 other error libraries.
9732 *) Fixed sk_insert which never worked properly.
9735 *) Fix ASN1 macros so they can handle indefinite length construted
9736 EXPLICIT tags. Some non standard certificates use these: they can now
9740 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9741 into a single doc/ssleay.txt bundle. This way the information is still
9742 preserved but no longer messes up this directory. Now it's new room for
9743 the new set of documenation files.
9744 [Ralf S. Engelschall]
9746 *) SETs were incorrectly DER encoded. This was a major pain, because they
9747 shared code with SEQUENCEs, which aren't coded the same. This means that
9748 almost everything to do with SETs or SEQUENCEs has either changed name or
9749 number of arguments.
9750 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9752 *) Fix test data to work with the above.
9755 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9756 was already fixed by Eric for 0.9.1 it seems.
9757 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9759 *) Autodetect FreeBSD3.
9762 *) Fix various bugs in Configure. This affects the following platforms:
9766 unixware-2.0-pentium
9770 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9771 before they are needed.
9774 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9778 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9780 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9781 changed SSLeay to OpenSSL in version strings.
9782 [Ralf S. Engelschall]
9784 *) Some fixups to the top-level documents.
9787 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9788 because the symlink to include/ was missing.
9789 [Ralf S. Engelschall]
9791 *) Incorporated the popular no-RSA/DSA-only patches
9792 which allow to compile a RSA-free SSLeay.
9793 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9795 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9796 when "ssleay" is still not found.
9797 [Ralf S. Engelschall]
9799 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9800 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9802 *) Updated the README file.
9803 [Ralf S. Engelschall]
9805 *) Added various .cvsignore files in the CVS repository subdirs
9806 to make a "cvs update" really silent.
9807 [Ralf S. Engelschall]
9809 *) Recompiled the error-definition header files and added
9810 missing symbols to the Win32 linker tables.
9811 [Ralf S. Engelschall]
9813 *) Cleaned up the top-level documents;
9814 o new files: CHANGES and LICENSE
9815 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9816 o merged COPYRIGHT into LICENSE
9817 o removed obsolete TODO file
9818 o renamed MICROSOFT to INSTALL.W32
9819 [Ralf S. Engelschall]
9821 *) Removed dummy files from the 0.9.1b source tree:
9822 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9823 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9824 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9825 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9826 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9827 [Ralf S. Engelschall]
9829 *) Added various platform portability fixes.
9832 *) The Genesis of the OpenSSL rpject:
9833 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9834 Young and Tim J. Hudson created while they were working for C2Net until
9836 [The OpenSSL Project]
9839 Changes between 0.9.0b and 0.9.1b [not released]
9841 *) Updated a few CA certificates under certs/
9844 *) Changed some BIGNUM api stuff.
9847 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9848 DGUX x86, Linux Alpha, etc.
9851 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9852 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9856 *) Add -strparse option to asn1pars program which parses nested
9858 [Dr Stephen Henson <shenson@bigfoot.com>]
9860 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9863 *) DSA fix for "ca" program.
9866 *) Added "-genkey" option to "dsaparam" program.
9869 *) Added RIPE MD160 (rmd160) message digest.
9872 *) Added -a (all) option to "ssleay version" command.
9875 *) Added PLATFORM define which is the id given to Configure.
9878 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9881 *) Extended the ASN.1 parser routines.
9884 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9887 *) Added a BN_CTX to the BN library.
9890 *) Fixed the weak key values in DES library
9893 *) Changed API in EVP library for cipher aliases.
9896 *) Added support for RC2/64bit cipher.
9899 *) Converted the lhash library to the crypto/mem.c functions.
9902 *) Added more recognized ASN.1 object ids.
9905 *) Added more RSA padding checks for SSL/TLS.
9908 *) Added BIO proxy/filter functionality.
9911 *) Added extra_certs to SSL_CTX which can be used
9912 send extra CA certificates to the client in the CA cert chain sending
9913 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9916 *) Now Fortezza is denied in the authentication phase because
9917 this is key exchange mechanism is not supported by SSLeay at all.
9920 *) Additional PKCS1 checks.
9923 *) Support the string "TLSv1" for all TLS v1 ciphers.
9926 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9927 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9930 *) Fixed a few memory leaks.
9933 *) Fixed various code and comment typos.
9936 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9937 bytes sent in the client random.
9938 [Edward Bishop <ebishop@spyglass.com>]