5 Changes between 0.9.8b and 0.9.9 [xx XXX xxxx]
7 *) Replace the algorithm specific calls to generate keys in "req" with the
11 *) Update PKCS#7 enveloped data routines to use new API. This is now
12 supported by any public key method supporting the encrypt operation. A
13 ctrl is added to allow the public key algorithm to examine or modify
14 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
18 *) Add a ctrl to asn1 method to allow a public key algorithm to express
19 a default digest type to use. In most cases this will be SHA1 but some
20 algorithms (such as GOST) need to specify an alternative digest. The
21 return value indicates how strong the prefernce is 1 means optional and
22 2 is mandatory (that is it is the only supported type). Modify
23 ASN1_item_sign() to accept a NULL digest argument to indicate it should
24 use the default md. Update openssl utilities to use the default digest
25 type for signing if it is not explicitly indicated.
28 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
29 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
30 signing method from the key type. This effectively removes the link
31 between digests and public key types.
34 *) Add an OID cross reference table and utility functions. Its purpose is to
35 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
36 rsaEncryption. This will allow some of the algorithm specific hackery
37 needed to use the correct OID to be removed.
40 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
41 structures for PKCS7_sign(). They are now set up by the relevant public
45 *) Add provisional EC pkey method with support for ECDSA and ECDH.
48 *) Add support for key derivation (agreement) in the API, DH method and
52 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
53 public and private key formats. As a side effect these add additional
54 command line functionality not previously available: DSA signatures can be
55 generated and verified using pkeyutl and DH key support and generation in
60 [Oliver Tappe <zooey@hirschkaefer.de>]
62 *) New make target "install_html_docs" installs HTML renditions of the
64 [Oliver Tappe <zooey@hirschkaefer.de>]
66 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
67 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
68 support key and parameter generation and add initial key generation
69 functionality for RSA.
72 *) Add functions for main EVP_PKEY_method operations. The undocumented
73 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
74 EVP_PKEY_{encrypt,decrypt}_old.
77 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
78 key API, doesn't do much yet.
81 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
82 public key algorithms. New option to openssl utility:
83 "list-public-key-algorithms" to print out info.
86 *) Implement the Supported Elliptic Curves Extension for
87 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
90 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
91 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
94 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
95 utilities such as rsa, dsa, dsaparam etc except they process any key
99 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
100 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
101 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
105 *) Initial support for pluggable public key ASN1.
106 De-spaghettify the public key ASN1 handling. Move public and private
107 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
108 algorithm specific handling to a single module within the relevant
109 algorithm directory. Add functions to allow (near) opaque processing
110 of public and private key structures.
113 *) Implement the Supported Point Formats Extension for
114 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
117 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
118 for the psk identity [hint] and the psk callback functions to the
119 SSL_SESSION, SSL and SSL_CTX structure.
122 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
126 SSL_CTX_use_psk_identity_hint
127 SSL_get_psk_identity_hint
129 SSL_use_psk_identity_hint
131 [Mika Kousa and Pasi Eronen of Nokia Corporation]
133 *) Add RFC 3161 compliant time stamp request creation, response generation
134 and response verification functionality.
135 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
137 *) Add initial support for TLS extensions, specifically for the server_name
138 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
139 have new members for a host name. The SSL data structure has an
140 additional member SSL_CTX *initial_ctx so that new sessions can be
141 stored in that context to allow for session resumption, even after the
142 SSL has been switched to a new SSL_CTX in reaction to a client's
143 server_name extension.
145 New functions (subject to change):
148 SSL_get_servername_type()
151 New CTRL codes and macros (subject to change):
153 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
154 - SSL_CTX_set_tlsext_servername_callback()
155 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
156 - SSL_CTX_set_tlsext_servername_arg()
157 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_hostname()
159 openssl s_client has a new '-servername ...' option.
161 openssl s_server has new options '-servername_host ...', '-cert2 ...',
162 '-key2 ...', '-servername_fatal' (subject to change). This allows
163 testing the HostName extension for a specific single host name ('-cert'
164 and '-key' remain fallbacks for handshakes without HostName
165 negotiation). If the unrecogninzed_name alert has to be sent, this by
166 default is a warning; it becomes fatal with the '-servername_fatal'
169 [Peter Sylvester, Remy Allais, Christophe Renou]
171 *) Whirlpool hash implementation is added.
174 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
175 bn(64,32). Because of instruction set limitations it doesn't have
176 any negative impact on performance. This was done mostly in order
177 to make it possible to share assembler modules, such as bn_mul_mont
178 implementations, between 32- and 64-bit builds without hassle.
181 *) Disable rogue ciphersuites:
183 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
184 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
185 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
187 The latter two were purportedly from
188 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
191 Other ciphersuites from draft-ietf-tls-56-bit-ciphersuites-01.txt
192 remain enabled for now, but are just as unofficial, and the ID
193 has long expired; these will probably disappear soon.
196 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
197 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
201 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
202 dedicated Montgomery multiplication procedure, is introduced.
203 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
204 "64-bit" performance on certain 32-bit targets.
207 *) New option SSL_OP_NO_COMP to disable use of compression selectively
208 in SSL structures. New SSL ctrl to set maximum send fragment size.
209 Save memory by seeting the I/O buffer sizes dynamically instead of
210 using the maximum available value.
213 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
214 in addition to the text details.
217 *) Very, very preliminary EXPERIMENTAL support for printing of general
218 ASN1 structures. This currently produces rather ugly output and doesn't
219 handle several customised structures at all.
222 *) Integrated support for PVK file format and some related formats such
223 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
224 these in the 'rsa' and 'dsa' utilities.
227 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
230 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
231 place for the (very old) "NETSCAPE" format certificates which are now
232 handled using new ASN1 code equivalents.
235 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
236 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
237 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
240 *) Modify CRL distribution points extension code to print out previously
241 unsupported fields. Enhance extension setting code to allow setting of
245 *) Add print and set support for Issuing Distribution Point CRL extension.
248 Changes between 0.9.8b and 0.9.8c [xx XXX xxxx]
250 *) Disable the padding bug check when compression is in use. The padding
251 bug check assumes the first packet is of even length, this is not
252 necessarily true if compresssion is enabled and can result in false
253 positives causing handshake failure. The actual bug test is ancient
254 code so it is hoped that implementations will either have fixed it by
255 now or any which still have the bug do not support compression.
258 Changes between 0.9.8a and 0.9.8b [04 May 2006]
260 *) When applying a cipher rule check to see if string match is an explicit
261 cipher suite and only match that one cipher suite if it is.
264 *) Link in manifests for VC++ if needed.
265 [Austin Ziegler <halostatue@gmail.com>]
267 *) Update support for ECC-based TLS ciphersuites according to
268 draft-ietf-tls-ecc-12.txt with proposed changes (but without
269 TLS extensions, which are supported starting with the 0.9.9
270 branch, not in the OpenSSL 0.9.8 branch).
273 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
274 opaque EVP_CIPHER_CTX handling.
277 *) Fixes and enhancements to zlib compression code. We now only use
278 "zlib1.dll" and use the default __cdecl calling convention on Win32
279 to conform with the standards mentioned here:
280 http://www.zlib.net/DLL_FAQ.txt
281 Static zlib linking now works on Windows and the new --with-zlib-include
282 --with-zlib-lib options to Configure can be used to supply the location
283 of the headers and library. Gracefully handle case where zlib library
287 *) Several fixes and enhancements to the OID generation code. The old code
288 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
289 handle numbers larger than ULONG_MAX, truncated printing and had a
290 non standard OBJ_obj2txt() behaviour.
293 *) Add support for building of engines under engine/ as shared libraries
294 under VC++ build system.
297 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
298 Hopefully, we will not see any false combination of paths any more.
301 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
303 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
304 (part of SSL_OP_ALL). This option used to disable the
305 countermeasure against man-in-the-middle protocol-version
306 rollback in the SSL 2.0 server implementation, which is a bad
307 idea. (CVE-2005-2969)
309 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
310 for Information Security, National Institute of Advanced Industrial
311 Science and Technology [AIST], Japan)]
313 *) Add two function to clear and return the verify parameter flags.
316 *) Keep cipherlists sorted in the source instead of sorting them at
317 runtime, thus removing the need for a lock.
320 *) Avoid some small subgroup attacks in Diffie-Hellman.
321 [Nick Mathewson and Ben Laurie]
323 *) Add functions for well-known primes.
326 *) Extended Windows CE support.
327 [Satoshi Nakamura and Andy Polyakov]
329 *) Initialize SSL_METHOD structures at compile time instead of during
330 runtime, thus removing the need for a lock.
333 *) Make PKCS7_decrypt() work even if no certificate is supplied by
334 attempting to decrypt each encrypted key in turn. Add support to
338 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
340 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
343 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
344 key into the same file any more.
347 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
350 *) Add -utf8 command line and config file option to 'ca'.
351 [Stefan <stf@udoma.org]
353 *) Removed the macro des_crypt(), as it seems to conflict with some
354 libraries. Use DES_crypt().
357 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
358 involves renaming the source and generated shared-libs for
359 both. The engines will accept the corrected or legacy ids
360 ('ncipher' and '4758_cca' respectively) when binding. NB,
361 this only applies when building 'shared'.
362 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
364 *) Add attribute functions to EVP_PKEY structure. Modify
365 PKCS12_create() to recognize a CSP name attribute and
366 use it. Make -CSP option work again in pkcs12 utility.
369 *) Add new functionality to the bn blinding code:
370 - automatic re-creation of the BN_BLINDING parameters after
371 a fixed number of uses (currently 32)
372 - add new function for parameter creation
373 - introduce flags to control the update behaviour of the
374 BN_BLINDING parameters
375 - hide BN_BLINDING structure
376 Add a second BN_BLINDING slot to the RSA structure to improve
377 performance when a single RSA object is shared among several
381 *) Add support for DTLS.
382 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
384 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
385 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
388 *) Remove buggy and incompletet DH cert support from
389 ssl/ssl_rsa.c and ssl/s3_both.c
392 *) Use SHA-1 instead of MD5 as the default digest algorithm for
393 the apps/openssl applications.
396 *) Compile clean with "-Wall -Wmissing-prototypes
397 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
398 DEBUG_SAFESTACK must also be set.
401 *) Change ./Configure so that certain algorithms can be disabled by default.
402 The new counterpiece to "no-xxx" is "enable-xxx".
404 The patented RC5 and MDC2 algorithms will now be disabled unless
405 "enable-rc5" and "enable-mdc2", respectively, are specified.
407 (IDEA remains enabled despite being patented. This is because IDEA
408 is frequently required for interoperability, and there is no license
409 fee for non-commercial use. As before, "no-idea" can be used to
410 avoid this algorithm.)
414 *) Add processing of proxy certificates (see RFC 3820). This work was
415 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
416 EGEE (Enabling Grids for E-science in Europe).
419 *) RC4 performance overhaul on modern architectures/implementations, such
420 as Intel P4, IA-64 and AMD64.
423 *) New utility extract-section.pl. This can be used specify an alternative
424 section number in a pod file instead of having to treat each file as
425 a separate case in Makefile. This can be done by adding two lines to the
428 =for comment openssl_section:XXX
430 The blank line is mandatory.
434 *) New arguments -certform, -keyform and -pass for s_client and s_server
435 to allow alternative format key and certificate files and passphrase
439 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
440 update associated structures and add various utility functions.
442 Add new policy related verify parameters, include policy checking in
443 standard verify code. Enhance 'smime' application with extra parameters
444 to support policy checking and print out.
447 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
448 Nehemiah processors. These extensions support AES encryption in hardware
449 as well as RNG (though RNG support is currently disabled).
450 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
452 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
455 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
456 [Andy Polyakov and a number of other people]
458 *) Improved PowerPC platform support. Most notably BIGNUM assembler
459 implementation contributed by IBM.
460 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
462 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
463 exponent rather than 'unsigned long'. There is a corresponding change to
464 the new 'rsa_keygen' element of the RSA_METHOD structure.
465 [Jelte Jansen, Geoff Thorpe]
467 *) Functionality for creating the initial serial number file is now
468 moved from CA.pl to the 'ca' utility with a new option -create_serial.
470 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
471 number file to 1, which is bound to cause problems. To avoid
472 the problems while respecting compatibility between different 0.9.7
473 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
474 CA.pl for serial number initialization. With the new release 0.9.8,
475 we can fix the problem directly in the 'ca' utility.)
478 *) Reduced header interdepencies by declaring more opaque objects in
479 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
480 give fewer recursive includes, which could break lazy source code - so
481 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
482 developers should define this symbol when building and using openssl to
483 ensure they track the recommended behaviour, interfaces, [etc], but
484 backwards-compatible behaviour prevails when this isn't defined.
487 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
490 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
491 This will generate a random key of the appropriate length based on the
492 cipher context. The EVP_CIPHER can provide its own random key generation
493 routine to support keys of a specific form. This is used in the des and
494 3des routines to generate a key of the correct parity. Update S/MIME
495 code to use new functions and hence generate correct parity DES keys.
496 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
497 valid (weak or incorrect parity).
500 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
501 as looking them up. This is useful when the verified structure may contain
502 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
503 present unless the new PKCS7_NO_CRL flag is asserted.
506 *) Extend ASN1 oid configuration module. It now additionally accepts the
509 shortName = some long name, 1.2.3.4
512 *) Reimplemented the BN_CTX implementation. There is now no more static
513 limitation on the number of variables it can handle nor the depth of the
514 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
515 information can now expand as required, and rather than having a single
516 static array of bignums, BN_CTX now uses a linked-list of such arrays
517 allowing it to expand on demand whilst maintaining the usefulness of
521 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
522 to allow all RSA operations to function using a single BN_CTX.
525 *) Preliminary support for certificate policy evaluation and checking. This
526 is initially intended to pass the tests outlined in "Conformance Testing
527 of Relying Party Client Certificate Path Processing Logic" v1.07.
530 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
531 remained unused and not that useful. A variety of other little bignum
532 tweaks and fixes have also been made continuing on from the audit (see
536 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
537 associated ASN1, EVP and SSL functions and old ASN1 macros.
540 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
541 and this should never fail. So the return value from the use of
542 BN_set_word() (which can fail due to needless expansion) is now deprecated;
543 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
546 *) BN_CTX_get() should return zero-valued bignums, providing the same
547 initialised value as BN_new().
548 [Geoff Thorpe, suggested by Ulf Möller]
550 *) Support for inhibitAnyPolicy certificate extension.
553 *) An audit of the BIGNUM code is underway, for which debugging code is
554 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
555 is considered valid when processing BIGNUMs, and causes execution to
556 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
557 further steps are taken to deliberately pollute unused data in BIGNUM
558 structures to try and expose faulty code further on. For now, openssl will
559 (in its default mode of operation) continue to tolerate the inconsistent
560 forms that it has tolerated in the past, but authors and packagers should
561 consider trying openssl and their own applications when compiled with
562 these debugging symbols defined. It will help highlight potential bugs in
563 their own code, and will improve the test coverage for OpenSSL itself. At
564 some point, these tighter rules will become openssl's default to improve
565 maintainability, though the assert()s and other overheads will remain only
566 in debugging configurations. See bn.h for more details.
567 [Geoff Thorpe, Nils Larsch, Ulf Möller]
569 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
570 that can only be obtained through BN_CTX_new() (which implicitly
571 initialises it). The presence of this function only made it possible
572 to overwrite an existing structure (and cause memory leaks).
575 *) Because of the callback-based approach for implementing LHASH as a
576 template type, lh_insert() adds opaque objects to hash-tables and
577 lh_doall() or lh_doall_arg() are typically used with a destructor callback
578 to clean up those corresponding objects before destroying the hash table
579 (and losing the object pointers). So some over-zealous constifications in
580 LHASH have been relaxed so that lh_insert() does not take (nor store) the
581 objects as "const" and the lh_doall[_arg] callback wrappers are not
582 prototyped to have "const" restrictions on the object pointers they are
583 given (and so aren't required to cast them away any more).
586 *) The tmdiff.h API was so ugly and minimal that our own timing utility
587 (speed) prefers to use its own implementation. The two implementations
588 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
589 its object type properly exposed (MS_TM) instead of casting to/from "char
590 *". This may still change yet if someone realises MS_TM and "ms_time_***"
591 aren't necessarily the greatest nomenclatures - but this is what was used
592 internally to the implementation so I've used that for now.
595 *) Ensure that deprecated functions do not get compiled when
596 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
597 the self-tests were still using deprecated key-generation functions so
598 these have been updated also.
601 *) Reorganise PKCS#7 code to separate the digest location functionality
602 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
603 New function PKCS7_set_digest() to set the digest type for PKCS#7
604 digestedData type. Add additional code to correctly generate the
605 digestedData type and add support for this type in PKCS7 initialization
609 *) New function PKCS7_set0_type_other() this initializes a PKCS7
610 structure of type "other".
613 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
614 sure the loop does correctly stop and breaking ("division by zero")
615 modulus operations are not performed. The (pre-generated) prime
616 table crypto/bn/bn_prime.h was already correct, but it could not be
617 re-generated on some platforms because of the "division by zero"
618 situation in the script.
619 [Ralf S. Engelschall]
621 *) Update support for ECC-based TLS ciphersuites according to
622 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
623 SHA-1 now is only used for "small" curves (where the
624 representation of a field element takes up to 24 bytes); for
625 larger curves, the field element resulting from ECDH is directly
626 used as premaster secret.
627 [Douglas Stebila (Sun Microsystems Laboratories)]
629 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
630 curve secp160r1 to the tests.
631 [Douglas Stebila (Sun Microsystems Laboratories)]
633 *) Add the possibility to load symbols globally with DSO.
634 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
636 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
637 control of the error stack.
640 *) Add support for STORE in ENGINE.
643 *) Add the STORE type. The intention is to provide a common interface
644 to certificate and key stores, be they simple file-based stores, or
645 HSM-type store, or LDAP stores, or...
646 NOTE: The code is currently UNTESTED and isn't really used anywhere.
649 *) Add a generic structure called OPENSSL_ITEM. This can be used to
650 pass a list of arguments to any function as well as provide a way
651 for a function to pass data back to the caller.
654 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
655 works like BUF_strdup() but can be used to duplicate a portion of
656 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
660 *) Add the function sk_find_ex() which works like sk_find(), but will
661 return an index to an element even if an exact match couldn't be
662 found. The index is guaranteed to point at the element where the
663 searched-for key would be inserted to preserve sorting order.
666 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
667 takes an extra flags argument for optional functionality. Currently,
668 the following flags are defined:
670 OBJ_BSEARCH_VALUE_ON_NOMATCH
671 This one gets OBJ_bsearch_ex() to return a pointer to the first
672 element where the comparing function returns a negative or zero
675 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
676 This one gets OBJ_bsearch_ex() to return a pointer to the first
677 element where the comparing function returns zero. This is useful
678 if there are more than one element where the comparing function
682 *) Make it possible to create self-signed certificates with 'openssl ca'
683 in such a way that the self-signed certificate becomes part of the
684 CA database and uses the same mechanisms for serial number generation
685 as all other certificate signing. The new flag '-selfsign' enables
686 this functionality. Adapt CA.sh and CA.pl.in.
689 *) Add functionality to check the public key of a certificate request
690 against a given private. This is useful to check that a certificate
691 request can be signed by that key (self-signing).
694 *) Make it possible to have multiple active certificates with the same
695 subject in the CA index file. This is done only if the keyword
696 'unique_subject' is set to 'no' in the main CA section (default
697 if 'CA_default') of the configuration file. The value is saved
698 with the database itself in a separate index attribute file,
699 named like the index file with '.attr' appended to the name.
702 *) Generate muti valued AVAs using '+' notation in config files for
706 *) Support for nameConstraints certificate extension.
709 *) Support for policyConstraints certificate extension.
712 *) Support for policyMappings certificate extension.
715 *) Make sure the default DSA_METHOD implementation only uses its
716 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
717 and change its own handlers to be NULL so as to remove unnecessary
718 indirection. This lets alternative implementations fallback to the
719 default implementation more easily.
722 *) Support for directoryName in GeneralName related extensions
726 *) Make it possible to link applications using Makefile.shared.
727 Make that possible even when linking against static libraries!
730 *) Support for single pass processing for S/MIME signing. This now
731 means that S/MIME signing can be done from a pipe, in addition
732 cleartext signing (multipart/signed type) is effectively streaming
733 and the signed data does not need to be all held in memory.
735 This is done with a new flag PKCS7_STREAM. When this flag is set
736 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
737 is done after the data is output (and digests calculated) in
741 *) Add full support for -rpath/-R, both in shared libraries and
742 applications, at least on the platforms where it's known how
746 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
747 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
748 will now compute a table of multiples of the generator that
749 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
750 faster (notably in the case of a single point multiplication,
752 [Nils Larsch, Bodo Moeller]
754 *) IPv6 support for certificate extensions. The various extensions
755 which use the IP:a.b.c.d can now take IPv6 addresses using the
756 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
760 *) Added an ENGINE that implements RSA by performing private key
761 exponentiations with the GMP library. The conversions to and from
762 GMP's mpz_t format aren't optimised nor are any montgomery forms
763 cached, and on x86 it appears OpenSSL's own performance has caught up.
764 However there are likely to be other architectures where GMP could
765 provide a boost. This ENGINE is not built in by default, but it can be
766 specified at Configure time and should be accompanied by the necessary
767 linker additions, eg;
768 ./config -DOPENSSL_USE_GMP -lgmp
771 *) "openssl engine" will not display ENGINE/DSO load failure errors when
772 testing availability of engines with "-t" - the old behaviour is
773 produced by increasing the feature's verbosity with "-tt".
776 *) ECDSA routines: under certain error conditions uninitialized BN objects
777 could be freed. Solution: make sure initialization is performed early
778 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
782 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
783 and DH_METHOD (eg. by ENGINE implementations) to override the normal
784 software implementations. For DSA and DH, parameter generation can
785 also be overriden by providing the appropriate method callbacks.
788 *) Change the "progress" mechanism used in key-generation and
789 primality testing to functions that take a new BN_GENCB pointer in
790 place of callback/argument pairs. The new API functions have "_ex"
791 postfixes and the older functions are reimplemented as wrappers for
792 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
793 declarations of the old functions to help (graceful) attempts to
794 migrate to the new functions. Also, the new key-generation API
795 functions operate on a caller-supplied key-structure and return
796 success/failure rather than returning a key or NULL - this is to
797 help make "keygen" another member function of RSA_METHOD etc.
799 Example for using the new callback interface:
801 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
805 BN_GENCB_set(&my_cb, my_callback, my_arg);
807 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
808 /* For the meaning of a, b in calls to my_callback(), see the
809 * documentation of the function that calls the callback.
810 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
811 * my_callback should return 1 if it wants BN_is_prime_ex()
812 * to continue, or 0 to stop.
817 *) Change the ZLIB compression method to be stateful, and make it
818 available to TLS with the number defined in
819 draft-ietf-tls-compression-04.txt.
822 *) Add the ASN.1 structures and functions for CertificatePair, which
823 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
825 CertificatePair ::= SEQUENCE {
826 forward [0] Certificate OPTIONAL,
827 reverse [1] Certificate OPTIONAL,
828 -- at least one of the pair shall be present -- }
830 Also implement the PEM functions to read and write certificate
831 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
833 This needed to be defined, mostly for the sake of the LDAP
834 attribute crossCertificatePair, but may prove useful elsewhere as
838 *) Make it possible to inhibit symlinking of shared libraries in
839 Makefile.shared, for Cygwin's sake.
842 *) Extend the BIGNUM API by creating a function
843 void BN_set_negative(BIGNUM *a, int neg);
844 and a macro that behave like
845 int BN_is_negative(const BIGNUM *a);
847 to avoid the need to access 'a->neg' directly in applications.
850 *) Implement fast modular reduction for pseudo-Mersenne primes
851 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
852 EC_GROUP_new_curve_GFp() will now automatically use this
854 [Nils Larsch <nla@trustcenter.de>]
856 *) Add new lock type (CRYPTO_LOCK_BN).
859 *) Change the ENGINE framework to automatically load engines
860 dynamically from specific directories unless they could be
861 found to already be built in or loaded. Move all the
862 current engines except for the cryptodev one to a new
864 The engines in engines/ are built as shared libraries if
865 the "shared" options was given to ./Configure or ./config.
866 Otherwise, they are inserted in libcrypto.a.
867 /usr/local/ssl/engines is the default directory for dynamic
868 engines, but that can be overriden at configure time through
869 the usual use of --prefix and/or --openssldir, and at run
870 time with the environment variable OPENSSL_ENGINES.
871 [Geoff Thorpe and Richard Levitte]
873 *) Add Makefile.shared, a helper makefile to build shared
874 libraries. Addapt Makefile.org.
877 *) Add version info to Win32 DLLs.
878 [Peter 'Luna' Runestig" <peter@runestig.com>]
880 *) Add new 'medium level' PKCS#12 API. Certificates and keys
881 can be added using this API to created arbitrary PKCS#12
882 files while avoiding the low level API.
884 New options to PKCS12_create(), key or cert can be NULL and
885 will then be omitted from the output file. The encryption
886 algorithm NIDs can be set to -1 for no encryption, the mac
887 iteration count can be set to 0 to omit the mac.
889 Enhance pkcs12 utility by making the -nokeys and -nocerts
890 options work when creating a PKCS#12 file. New option -nomac
891 to omit the mac, NONE can be set for an encryption algorithm.
892 New code is modified to use the enhanced PKCS12_create()
893 instead of the low level API.
896 *) Extend ASN1 encoder to support indefinite length constructed
897 encoding. This can output sequences tags and octet strings in
898 this form. Modify pk7_asn1.c to support indefinite length
899 encoding. This is experimental and needs additional code to
900 be useful, such as an ASN1 bio and some enhanced streaming
903 Extend template encode functionality so that tagging is passed
904 down to the template encoder.
907 *) Let 'openssl req' fail if an argument to '-newkey' is not
908 recognized instead of using RSA as a default.
911 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
912 As these are not official, they are not included in "ALL";
913 the "ECCdraft" ciphersuite group alias can be used to select them.
914 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
916 *) Add ECDH engine support.
917 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
919 *) Add ECDH in new directory crypto/ecdh/.
920 [Douglas Stebila (Sun Microsystems Laboratories)]
922 *) Let BN_rand_range() abort with an error after 100 iterations
923 without success (which indicates a broken PRNG).
926 *) Change BN_mod_sqrt() so that it verifies that the input value
927 is really the square of the return value. (Previously,
928 BN_mod_sqrt would show GIGO behaviour.)
931 *) Add named elliptic curves over binary fields from X9.62, SECG,
932 and WAP/WTLS; add OIDs that were still missing.
934 [Sheueling Chang Shantz and Douglas Stebila
935 (Sun Microsystems Laboratories)]
937 *) Extend the EC library for elliptic curves over binary fields
938 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
941 EC_GF2m_simple_method
945 EC_GROUP_new_curve_GF2m
946 EC_GROUP_set_curve_GF2m
947 EC_GROUP_get_curve_GF2m
948 EC_POINT_set_affine_coordinates_GF2m
949 EC_POINT_get_affine_coordinates_GF2m
950 EC_POINT_set_compressed_coordinates_GF2m
952 Point compression for binary fields is disabled by default for
953 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
956 As binary polynomials are represented as BIGNUMs, various members
957 of the EC_GROUP and EC_POINT data structures can be shared
958 between the implementations for prime fields and binary fields;
959 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
960 are essentially identical to their ..._GFp counterparts.
961 (For simplicity, the '..._GFp' prefix has been dropped from
962 various internal method names.)
964 An internal 'field_div' method (similar to 'field_mul' and
965 'field_sqr') has been added; this is used only for binary fields.
967 [Sheueling Chang Shantz and Douglas Stebila
968 (Sun Microsystems Laboratories)]
970 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
971 through methods ('mul', 'precompute_mult').
973 The generic implementations (now internally called 'ec_wNAF_mul'
974 and 'ec_wNAF_precomputed_mult') remain the default if these
975 methods are undefined.
977 [Sheueling Chang Shantz and Douglas Stebila
978 (Sun Microsystems Laboratories)]
980 *) New function EC_GROUP_get_degree, which is defined through
981 EC_METHOD. For curves over prime fields, this returns the bit
982 length of the modulus.
984 [Sheueling Chang Shantz and Douglas Stebila
985 (Sun Microsystems Laboratories)]
987 *) New functions EC_GROUP_dup, EC_POINT_dup.
988 (These simply call ..._new and ..._copy).
990 [Sheueling Chang Shantz and Douglas Stebila
991 (Sun Microsystems Laboratories)]
993 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
994 Polynomials are represented as BIGNUMs (where the sign bit is not
995 used) in the following functions [macros]:
998 BN_GF2m_sub [= BN_GF2m_add]
999 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1000 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1001 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1003 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1004 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1005 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1006 BN_GF2m_cmp [= BN_ucmp]
1008 (Note that only the 'mod' functions are actually for fields GF(2^m).
1009 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1011 For some functions, an the irreducible polynomial defining a
1012 field can be given as an 'unsigned int[]' with strictly
1013 decreasing elements giving the indices of those bits that are set;
1014 i.e., p[] represents the polynomial
1015 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1017 p[0] > p[1] > ... > p[k] = 0.
1018 This applies to the following functions:
1023 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1024 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1026 BN_GF2m_mod_sqrt_arr
1027 BN_GF2m_mod_solve_quad_arr
1031 Conversion can be performed by the following functions:
1036 bntest.c has additional tests for binary polynomial arithmetic.
1038 Two implementations for BN_GF2m_mod_div() are available.
1039 The default algorithm simply uses BN_GF2m_mod_inv() and
1040 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1041 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1042 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1044 [Sheueling Chang Shantz and Douglas Stebila
1045 (Sun Microsystems Laboratories)]
1047 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1048 functionality is disabled at compile-time.
1049 [Douglas Stebila <douglas.stebila@sun.com>]
1051 *) Change default behaviour of 'openssl asn1parse' so that more
1052 information is visible when viewing, e.g., a certificate:
1054 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1055 mode the content of non-printable OCTET STRINGs is output in a
1056 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1057 avoid the appearance of a printable string.
1058 [Nils Larsch <nla@trustcenter.de>]
1060 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1062 EC_GROUP_set_asn1_flag()
1063 EC_GROUP_get_asn1_flag()
1064 EC_GROUP_set_point_conversion_form()
1065 EC_GROUP_get_point_conversion_form()
1066 These control ASN1 encoding details:
1067 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1068 has been set to OPENSSL_EC_NAMED_CURVE.
1069 - Points are encoded in uncompressed form by default; options for
1070 asn1_for are as for point2oct, namely
1071 POINT_CONVERSION_COMPRESSED
1072 POINT_CONVERSION_UNCOMPRESSED
1073 POINT_CONVERSION_HYBRID
1075 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1078 EC_GROUP_get0_seed()
1079 EC_GROUP_get_seed_len()
1080 This is used only for ASN1 purposes (so far).
1081 [Nils Larsch <nla@trustcenter.de>]
1083 *) Add 'field_type' member to EC_METHOD, which holds the NID
1084 of the appropriate field type OID. The new function
1085 EC_METHOD_get_field_type() returns this value.
1086 [Nils Larsch <nla@trustcenter.de>]
1091 EC_POINT_point2hex()
1092 EC_POINT_hex2point()
1093 providing useful interfaces to EC_POINT_point2oct() and
1094 EC_POINT_oct2point().
1095 [Nils Larsch <nla@trustcenter.de>]
1097 *) Change internals of the EC library so that the functions
1098 EC_GROUP_set_generator()
1099 EC_GROUP_get_generator()
1100 EC_GROUP_get_order()
1101 EC_GROUP_get_cofactor()
1102 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1103 to methods, which would lead to unnecessary code duplication when
1104 adding different types of curves.
1105 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1107 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1108 arithmetic, and such that modified wNAFs are generated
1109 (which avoid length expansion in many cases).
1112 *) Add a function EC_GROUP_check_discriminant() (defined via
1113 EC_METHOD) that verifies that the curve discriminant is non-zero.
1115 Add a function EC_GROUP_check() that makes some sanity tests
1116 on a EC_GROUP, its generator and order. This includes
1117 EC_GROUP_check_discriminant().
1118 [Nils Larsch <nla@trustcenter.de>]
1120 *) Add ECDSA in new directory crypto/ecdsa/.
1122 Add applications 'openssl ecparam' and 'openssl ecdsa'
1123 (these are based on 'openssl dsaparam' and 'openssl dsa').
1125 ECDSA support is also included in various other files across the
1126 library. Most notably,
1127 - 'openssl req' now has a '-newkey ecdsa:file' option;
1128 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1129 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1130 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1131 them suitable for ECDSA where domain parameters must be
1132 extracted before the specific public key;
1133 - ECDSA engine support has been added.
1134 [Nils Larsch <nla@trustcenter.de>]
1136 *) Include some named elliptic curves, and add OIDs from X9.62,
1137 SECG, and WAP/WTLS. Each curve can be obtained from the new
1139 EC_GROUP_new_by_curve_name(),
1140 and the list of available named curves can be obtained with
1141 EC_get_builtin_curves().
1142 Also add a 'curve_name' member to EC_GROUP objects, which can be
1144 EC_GROUP_set_curve_name()
1145 EC_GROUP_get_curve_name()
1146 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1148 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1149 was actually never needed) and in BN_mul(). The removal in BN_mul()
1150 required a small change in bn_mul_part_recursive() and the addition
1151 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1152 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1153 bn_sub_words() and bn_add_words() except they take arrays with
1157 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1159 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1160 The value now differs depending on if you build for FIPS or not.
1161 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1162 safely run with a non-FIPSed libcrypto, as it may crash because of
1163 the difference induced by this change.
1166 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1168 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1169 (part of SSL_OP_ALL). This option used to disable the
1170 countermeasure against man-in-the-middle protocol-version
1171 rollback in the SSL 2.0 server implementation, which is a bad
1172 idea. (CVE-2005-2969)
1174 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1175 for Information Security, National Institute of Advanced Industrial
1176 Science and Technology [AIST], Japan)]
1178 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1179 mainly for FIPS compliance and not fully integrated at this stage.
1182 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1183 the exponentiation using a fixed-length exponent. (Otherwise,
1184 the information leaked through timing could expose the secret key
1185 after many signatures; cf. Bleichenbacher's attack on DSA with
1189 *) Make a new fixed-window mod_exp implementation the default for
1190 RSA, DSA, and DH private-key operations so that the sequence of
1191 squares and multiplies and the memory access pattern are
1192 independent of the particular secret key. This will mitigate
1193 cache-timing and potential related attacks.
1195 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1196 and this is automatically used by BN_mod_exp_mont() if the new flag
1197 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1198 will use this BN flag for private exponents unless the flag
1199 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1200 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1202 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1204 *) Change the client implementation for SSLv23_method() and
1205 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1206 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1207 (Previously, the SSL 2.0 backwards compatible Client Hello
1208 message format would be used even with SSL_OP_NO_SSLv2.)
1211 *) Add support for smime-type MIME parameter in S/MIME messages which some
1215 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1216 a threadsafe manner. Modify rsa code to use new function and add calls
1217 to dsa and dh code (which had race conditions before).
1220 *) Include the fixed error library code in the C error file definitions
1221 instead of fixing them up at runtime. This keeps the error code
1222 structures constant.
1225 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1227 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1230 *) Fixes for newer kerberos headers. NB: the casts are needed because
1231 the 'length' field is signed on one version and unsigned on another
1232 with no (?) obvious way to tell the difference, without these VC++
1233 complains. Also the "definition" of FAR (blank) is no longer included
1234 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1235 some needed definitions.
1238 *) Undo Cygwin change.
1241 *) Added support for proxy certificates according to RFC 3820.
1242 Because they may be a security thread to unaware applications,
1243 they must be explicitely allowed in run-time. See
1244 docs/HOWTO/proxy_certificates.txt for further information.
1247 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1249 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1250 server and client random values. Previously
1251 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1252 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1254 This change has negligible security impact because:
1256 1. Server and client random values still have 24 bytes of pseudo random
1259 2. Server and client random values are sent in the clear in the initial
1262 3. The master secret is derived using the premaster secret (48 bytes in
1263 size for static RSA ciphersuites) as well as client server and random
1266 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1269 [Stephen Henson, reported by UK NISCC]
1271 *) Use Windows randomness collection on Cygwin.
1274 *) Fix hang in EGD/PRNGD query when communication socket is closed
1275 prematurely by EGD/PRNGD.
1276 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1278 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1281 *) Back-port of selected performance improvements from development
1282 branch, as well as improved support for PowerPC platforms.
1285 *) Add lots of checks for memory allocation failure, error codes to indicate
1286 failure and freeing up memory if a failure occurs.
1287 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1289 *) Add new -passin argument to dgst.
1292 *) Perform some character comparisons of different types in X509_NAME_cmp:
1293 this is needed for some certificates that reencode DNs into UTF8Strings
1294 (in violation of RFC3280) and can't or wont issue name rollover
1298 *) Make an explicit check during certificate validation to see that
1299 the CA setting in each certificate on the chain is correct. As a
1300 side effect always do the following basic checks on extensions,
1301 not just when there's an associated purpose to the check:
1303 - if there is an unhandled critical extension (unless the user
1304 has chosen to ignore this fault)
1305 - if the path length has been exceeded (if one is set at all)
1306 - that certain extensions fit the associated purpose (if one has
1310 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1312 *) Avoid a race condition when CRLs are checked in a multi threaded
1313 environment. This would happen due to the reordering of the revoked
1314 entries during signature checking and serial number lookup. Now the
1315 encoding is cached and the serial number sort performed under a lock.
1316 Add new STACK function sk_is_sorted().
1319 *) Add Delta CRL to the extension code.
1322 *) Various fixes to s3_pkt.c so alerts are sent properly.
1323 [David Holmes <d.holmes@f5.com>]
1325 *) Reduce the chances of duplicate issuer name and serial numbers (in
1326 violation of RFC3280) using the OpenSSL certificate creation utilities.
1327 This is done by creating a random 64 bit value for the initial serial
1328 number when a serial number file is created or when a self signed
1329 certificate is created using 'openssl req -x509'. The initial serial
1330 number file is created using 'openssl x509 -next_serial' in CA.pl
1331 rather than being initialized to 1.
1334 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1336 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1337 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1338 [Joe Orton, Steve Henson]
1340 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1342 [Joe Orton, Steve Henson]
1344 *) Make it possible to have multiple active certificates with the same
1345 subject in the CA index file. This is done only if the keyword
1346 'unique_subject' is set to 'no' in the main CA section (default
1347 if 'CA_default') of the configuration file. The value is saved
1348 with the database itself in a separate index attribute file,
1349 named like the index file with '.attr' appended to the name.
1352 *) X509 verify fixes. Disable broken certificate workarounds when
1353 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1354 keyUsage extension present. Don't accept CRLs with unhandled critical
1355 extensions: since verify currently doesn't process CRL extensions this
1356 rejects a CRL with *any* critical extensions. Add new verify error codes
1360 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1361 A clarification of RFC2560 will require the use of OCTET STRINGs and
1362 some implementations cannot handle the current raw format. Since OpenSSL
1363 copies and compares OCSP nonces as opaque blobs without any attempt at
1364 parsing them this should not create any compatibility issues.
1367 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1368 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1369 this HMAC (and other) operations are several times slower than OpenSSL
1373 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1374 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1376 *) Use the correct content when signing type "other".
1379 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1381 *) Fix various bugs revealed by running the NISCC test suite:
1383 Stop out of bounds reads in the ASN1 code when presented with
1384 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1386 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1388 If verify callback ignores invalid public key errors don't try to check
1389 certificate signature with the NULL public key.
1393 *) New -ignore_err option in ocsp application to stop the server
1394 exiting on the first error in a request.
1397 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1398 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1402 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1403 extra data after the compression methods not only for TLS 1.0
1404 but also for SSL 3.0 (as required by the specification).
1405 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1407 *) Change X509_certificate_type() to mark the key as exported/exportable
1408 when it's 512 *bits* long, not 512 bytes.
1411 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1412 blocks during encryption.
1415 *) Various fixes to base64 BIO and non blocking I/O. On write
1416 flushes were not handled properly if the BIO retried. On read
1417 data was not being buffered properly and had various logic bugs.
1418 This also affects blocking I/O when the data being decoded is a
1422 *) Various S/MIME bugfixes and compatibility changes:
1423 output correct application/pkcs7 MIME type if
1424 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1425 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1426 of files as .eml work). Correctly handle very long lines in MIME
1430 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1432 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1433 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1434 a protocol version number mismatch like a decryption error
1435 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1438 *) Turn on RSA blinding by default in the default implementation
1439 to avoid a timing attack. Applications that don't want it can call
1440 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1441 They would be ill-advised to do so in most cases.
1442 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1444 *) Change RSA blinding code so that it works when the PRNG is not
1445 seeded (in this case, the secret RSA exponent is abused as
1446 an unpredictable seed -- if it is not unpredictable, there
1447 is no point in blinding anyway). Make RSA blinding thread-safe
1448 by remembering the creator's thread ID in rsa->blinding and
1449 having all other threads use local one-time blinding factors
1450 (this requires more computation than sharing rsa->blinding, but
1451 avoids excessive locking; and if an RSA object is not shared
1452 between threads, blinding will still be very fast).
1455 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1456 ENGINE as defaults for all supported algorithms irrespective of
1457 the 'flags' parameter. 'flags' is now honoured, so applications
1458 should make sure they are passing it correctly.
1461 *) Target "mingw" now allows native Windows code to be generated in
1462 the Cygwin environment as well as with the MinGW compiler.
1465 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1467 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1468 via timing by performing a MAC computation even if incorrrect
1469 block cipher padding has been found. This is a countermeasure
1470 against active attacks where the attacker has to distinguish
1471 between bad padding and a MAC verification error. (CVE-2003-0078)
1473 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1474 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1475 Martin Vuagnoux (EPFL, Ilion)]
1477 *) Make the no-err option work as intended. The intention with no-err
1478 is not to have the whole error stack handling routines removed from
1479 libcrypto, it's only intended to remove all the function name and
1480 reason texts, thereby removing some of the footprint that may not
1481 be interesting if those errors aren't displayed anyway.
1483 NOTE: it's still possible for any application or module to have it's
1484 own set of error texts inserted. The routines are there, just not
1485 used by default when no-err is given.
1488 *) Add support for FreeBSD on IA64.
1489 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1491 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1492 Kerberos function mit_des_cbc_cksum(). Before this change,
1493 the value returned by DES_cbc_cksum() was like the one from
1494 mit_des_cbc_cksum(), except the bytes were swapped.
1495 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1497 *) Allow an application to disable the automatic SSL chain building.
1498 Before this a rather primitive chain build was always performed in
1499 ssl3_output_cert_chain(): an application had no way to send the
1500 correct chain if the automatic operation produced an incorrect result.
1502 Now the chain builder is disabled if either:
1504 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1506 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1508 The reasoning behind this is that an application would not want the
1509 auto chain building to take place if extra chain certificates are
1510 present and it might also want a means of sending no additional
1511 certificates (for example the chain has two certificates and the
1515 *) Add the possibility to build without the ENGINE framework.
1516 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1518 *) Under Win32 gmtime() can return NULL: check return value in
1519 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1522 *) DSA routines: under certain error conditions uninitialized BN objects
1523 could be freed. Solution: make sure initialization is performed early
1524 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1525 Nils Larsch <nla@trustcenter.de> via PR#459)
1528 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1529 checked on reconnect on the client side, therefore session resumption
1530 could still fail with a "ssl session id is different" error. This
1531 behaviour is masked when SSL_OP_ALL is used due to
1532 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1533 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1534 followup to PR #377.
1537 *) IA-32 assembler support enhancements: unified ELF targets, support
1538 for SCO/Caldera platforms, fix for Cygwin shared build.
1541 *) Add support for FreeBSD on sparc64. As a consequence, support for
1542 FreeBSD on non-x86 processors is separate from x86 processors on
1543 the config script, much like the NetBSD support.
1544 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1546 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1548 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1551 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1552 code (06) was taken as the first octet of the session ID and the last
1553 octet was ignored consequently. As a result SSLv2 client side session
1554 caching could not have worked due to the session ID mismatch between
1556 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1560 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1561 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1565 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1566 seems that in spite of existing for more than a year, many application
1567 author have done nothing to provide the necessary callbacks, which
1568 means that this particular engine will not work properly anywhere.
1569 This is a very unfortunate situation which forces us, in the name
1570 of usability, to give the hw_ncipher.c a static lock, which is part
1572 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1573 appear in 0.9.8 or later. We EXPECT application authors to have
1574 dealt properly with this when 0.9.8 is released (unless we actually
1575 make such changes in the libcrypto locking code that changes will
1576 have to be made anyway).
1579 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1580 octets have been read, EOF or an error occurs. Without this change
1581 some truncated ASN1 structures will not produce an error.
1584 *) Disable Heimdal support, since it hasn't been fully implemented.
1585 Still give the possibility to force the use of Heimdal, but with
1586 warnings and a request that patches get sent to openssl-dev.
1589 *) Add the VC-CE target, introduce the WINCE sysname, and add
1590 INSTALL.WCE and appropriate conditionals to make it build.
1591 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1593 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1594 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1595 edit numbers of the version.
1596 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1598 *) Introduce safe string copy and catenation functions
1599 (BUF_strlcpy() and BUF_strlcat()).
1600 [Ben Laurie (CHATS) and Richard Levitte]
1602 *) Avoid using fixed-size buffers for one-line DNs.
1603 [Ben Laurie (CHATS)]
1605 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1606 resizing buffers containing secrets, and use where appropriate.
1607 [Ben Laurie (CHATS)]
1609 *) Avoid using fixed size buffers for configuration file location.
1610 [Ben Laurie (CHATS)]
1612 *) Avoid filename truncation for various CA files.
1613 [Ben Laurie (CHATS)]
1615 *) Use sizeof in preference to magic numbers.
1616 [Ben Laurie (CHATS)]
1618 *) Avoid filename truncation in cert requests.
1619 [Ben Laurie (CHATS)]
1621 *) Add assertions to check for (supposedly impossible) buffer
1623 [Ben Laurie (CHATS)]
1625 *) Don't cache truncated DNS entries in the local cache (this could
1626 potentially lead to a spoofing attack).
1627 [Ben Laurie (CHATS)]
1629 *) Fix various buffers to be large enough for hex/decimal
1630 representations in a platform independent manner.
1631 [Ben Laurie (CHATS)]
1633 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1634 resizing buffers containing secrets, and use where appropriate.
1635 [Ben Laurie (CHATS)]
1637 *) Add BIO_indent() to avoid much slightly worrying code to do
1639 [Ben Laurie (CHATS)]
1641 *) Convert sprintf()/BIO_puts() to BIO_printf().
1642 [Ben Laurie (CHATS)]
1644 *) buffer_gets() could terminate with the buffer only half
1646 [Ben Laurie (CHATS)]
1648 *) Add assertions to prevent user-supplied crypto functions from
1649 overflowing internal buffers by having large block sizes, etc.
1650 [Ben Laurie (CHATS)]
1652 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1654 [Ben Laurie (CHATS)]
1656 *) Eliminate unused copy of key in RC4.
1657 [Ben Laurie (CHATS)]
1659 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1660 [Ben Laurie (CHATS)]
1662 *) Fix off-by-one error in EGD path.
1663 [Ben Laurie (CHATS)]
1665 *) If RANDFILE path is too long, ignore instead of truncating.
1666 [Ben Laurie (CHATS)]
1668 *) Eliminate unused and incorrectly sized X.509 structure
1670 [Ben Laurie (CHATS)]
1672 *) Eliminate unused and dangerous function knumber().
1673 [Ben Laurie (CHATS)]
1675 *) Eliminate unused and dangerous structure, KSSL_ERR.
1676 [Ben Laurie (CHATS)]
1678 *) Protect against overlong session ID context length in an encoded
1679 session object. Since these are local, this does not appear to be
1681 [Ben Laurie (CHATS)]
1683 *) Change from security patch (see 0.9.6e below) that did not affect
1684 the 0.9.6 release series:
1686 Remote buffer overflow in SSL3 protocol - an attacker could
1687 supply an oversized master key in Kerberos-enabled versions.
1689 [Ben Laurie (CHATS)]
1691 *) Change the SSL kerb5 codes to match RFC 2712.
1694 *) Make -nameopt work fully for req and add -reqopt switch.
1695 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1697 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1698 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1700 *) Make sure tests can be performed even if the corresponding algorithms
1701 have been removed entirely. This was also the last step to make
1702 OpenSSL compilable with DJGPP under all reasonable conditions.
1703 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1705 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1706 to allow version independent disabling of normally unselected ciphers,
1707 which may be activated as a side-effect of selecting a single cipher.
1709 (E.g., cipher list string "RSA" enables ciphersuites that are left
1710 out of "ALL" because they do not provide symmetric encryption.
1711 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1712 [Lutz Jaenicke, Bodo Moeller]
1714 *) Add appropriate support for separate platform-dependent build
1715 directories. The recommended way to make a platform-dependent
1716 build directory is the following (tested on Linux), maybe with
1719 # Place yourself outside of the OpenSSL source tree. In
1720 # this example, the environment variable OPENSSL_SOURCE
1721 # is assumed to contain the absolute OpenSSL source directory.
1722 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1723 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1724 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1725 mkdir -p `dirname $F`
1726 ln -s $OPENSSL_SOURCE/$F $F
1729 To be absolutely sure not to disturb the source tree, a "make clean"
1730 is a good thing. If it isn't successfull, don't worry about it,
1731 it probably means the source directory is very clean.
1734 *) Make sure any ENGINE control commands make local copies of string
1735 pointers passed to them whenever necessary. Otherwise it is possible
1736 the caller may have overwritten (or deallocated) the original string
1737 data when a later ENGINE operation tries to use the stored values.
1738 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1740 *) Improve diagnostics in file reading and command-line digests.
1741 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1743 *) Add AES modes CFB and OFB to the object database. Correct an
1744 error in AES-CFB decryption.
1747 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1748 allows existing EVP_CIPHER_CTX structures to be reused after
1749 calling EVP_*Final(). This behaviour is used by encryption
1750 BIOs and some applications. This has the side effect that
1751 applications must explicitly clean up cipher contexts with
1752 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1755 *) Check the values of dna and dnb in bn_mul_recursive before calling
1756 bn_mul_comba (a non zero value means the a or b arrays do not contain
1757 n2 elements) and fallback to bn_mul_normal if either is not zero.
1760 *) Fix escaping of non-ASCII characters when using the -subj option
1761 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1764 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1765 form for "surname", serialNumber has no short form.
1766 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1767 therefore remove "mail" short name for "internet 7".
1768 The OID for unique identifiers in X509 certificates is
1769 x500UniqueIdentifier, not uniqueIdentifier.
1770 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1773 *) Add an "init" command to the ENGINE config module and auto initialize
1774 ENGINEs. Without any "init" command the ENGINE will be initialized
1775 after all ctrl commands have been executed on it. If init=1 the
1776 ENGINE is initailized at that point (ctrls before that point are run
1777 on the uninitialized ENGINE and after on the initialized one). If
1778 init=0 then the ENGINE will not be iniatialized at all.
1781 *) Fix the 'app_verify_callback' interface so that the user-defined
1782 argument is actually passed to the callback: In the
1783 SSL_CTX_set_cert_verify_callback() prototype, the callback
1784 declaration has been changed from
1787 int (*cb)(X509_STORE_CTX *,void *);
1788 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1789 i=s->ctx->app_verify_callback(&ctx)
1790 has been changed into
1791 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1793 To update applications using SSL_CTX_set_cert_verify_callback(),
1794 a dummy argument can be added to their callback functions.
1795 [D. K. Smetters <smetters@parc.xerox.com>]
1797 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1798 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1800 *) Add and OPENSSL_LOAD_CONF define which will cause
1801 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1802 This allows older applications to transparently support certain
1803 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1804 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1805 load the config file and OPENSSL_add_all_algorithms_conf() which will
1806 always load it have also been added.
1809 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1810 Adjust NIDs and EVP layer.
1811 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1813 *) Config modules support in openssl utility.
1815 Most commands now load modules from the config file,
1816 though in a few (such as version) this isn't done
1817 because it couldn't be used for anything.
1819 In the case of ca and req the config file used is
1820 the same as the utility itself: that is the -config
1821 command line option can be used to specify an
1825 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1826 use "openssl_conf" if filename is NULL use default openssl config file.
1829 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1830 config section name. Add a new flag to tolerate a missing config file
1831 and move code to CONF_modules_load_file().
1834 *) Support for crypto accelerator cards from Accelerated Encryption
1835 Processing, www.aep.ie. (Use engine 'aep')
1836 The support was copied from 0.9.6c [engine] and adapted/corrected
1837 to work with the new engine framework.
1838 [AEP Inc. and Richard Levitte]
1840 *) Support for SureWare crypto accelerator cards from Baltimore
1841 Technologies. (Use engine 'sureware')
1842 The support was copied from 0.9.6c [engine] and adapted
1843 to work with the new engine framework.
1846 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1847 make the newer ENGINE framework commands for the CHIL engine work.
1848 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1850 *) Make it possible to produce shared libraries on ReliantUNIX.
1851 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1853 *) Add the configuration target debug-linux-ppro.
1854 Make 'openssl rsa' use the general key loading routines
1855 implemented in apps.c, and make those routines able to
1856 handle the key format FORMAT_NETSCAPE and the variant
1858 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1860 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1861 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1863 *) Add -keyform to rsautl, and document -engine.
1864 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1866 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1867 BIO_R_NO_SUCH_FILE error code rather than the generic
1868 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1871 *) Add new functions
1873 ERR_peek_last_error_line
1874 ERR_peek_last_error_line_data.
1875 These are similar to
1878 ERR_peek_error_line_data,
1879 but report on the latest error recorded rather than the first one
1880 still in the error queue.
1881 [Ben Laurie, Bodo Moeller]
1883 *) default_algorithms option in ENGINE config module. This allows things
1885 default_algorithms = ALL
1886 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1889 *) Prelminary ENGINE config module.
1892 *) New experimental application configuration code.
1895 *) Change the AES code to follow the same name structure as all other
1896 symmetric ciphers, and behave the same way. Move everything to
1897 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1898 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1900 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1901 [Ben Laurie and Theo de Raadt]
1903 *) Add option to output public keys in req command.
1904 [Massimiliano Pala madwolf@openca.org]
1906 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1907 (up to about 10% better than before for P-192 and P-224).
1910 *) New functions/macros
1912 SSL_CTX_set_msg_callback(ctx, cb)
1913 SSL_CTX_set_msg_callback_arg(ctx, arg)
1914 SSL_set_msg_callback(ssl, cb)
1915 SSL_set_msg_callback_arg(ssl, arg)
1917 to request calling a callback function
1919 void cb(int write_p, int version, int content_type,
1920 const void *buf, size_t len, SSL *ssl, void *arg)
1922 whenever a protocol message has been completely received
1923 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1924 protocol version according to which the SSL library interprets
1925 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1926 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1927 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1928 specification (change_cipher_spec(20), alert(21), handshake(22)).
1929 'buf' and 'len' point to the actual message, 'ssl' to the
1930 SSL object, and 'arg' is the application-defined value set by
1931 SSL[_CTX]_set_msg_callback_arg().
1933 'openssl s_client' and 'openssl s_server' have new '-msg' options
1934 to enable a callback that displays all protocol messages.
1937 *) Change the shared library support so shared libraries are built as
1938 soon as the corresponding static library is finished, and thereby get
1939 openssl and the test programs linked against the shared library.
1940 This still only happens when the keyword "shard" has been given to
1941 the configuration scripts.
1943 NOTE: shared library support is still an experimental thing, and
1944 backward binary compatibility is still not guaranteed.
1945 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1947 *) Add support for Subject Information Access extension.
1948 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1950 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1951 additional bytes when new memory had to be allocated, not just
1952 when reusing an existing buffer.
1955 *) New command line and configuration option 'utf8' for the req command.
1956 This allows field values to be specified as UTF8 strings.
1959 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1960 runs for the former and machine-readable output for the latter.
1963 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1964 of the e-mail address in the DN (i.e., it will go into a certificate
1965 extension only). The new configuration file option 'email_in_dn = no'
1966 has the same effect.
1967 [Massimiliano Pala madwolf@openca.org]
1969 *) Change all functions with names starting with des_ to be starting
1970 with DES_ instead. Add wrappers that are compatible with libdes,
1971 but are named _ossl_old_des_*. Finally, add macros that map the
1972 des_* symbols to the corresponding _ossl_old_des_* if libdes
1973 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1974 desired, the des_* symbols will be mapped to DES_*, with one
1977 Since we provide two compatibility mappings, the user needs to
1978 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1979 compatibility is desired. The default (i.e., when that macro
1980 isn't defined) is OpenSSL 0.9.6c compatibility.
1982 There are also macros that enable and disable the support of old
1983 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1984 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1985 are defined, the default will apply: to support the old des routines.
1987 In either case, one must include openssl/des.h to get the correct
1988 definitions. Do not try to just include openssl/des_old.h, that
1991 NOTE: This is a major break of an old API into a new one. Software
1992 authors are encouraged to switch to the DES_ style functions. Some
1993 time in the future, des_old.h and the libdes compatibility functions
1994 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1995 default), and then completely removed.
1998 *) Test for certificates which contain unsupported critical extensions.
1999 If such a certificate is found during a verify operation it is
2000 rejected by default: this behaviour can be overridden by either
2001 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2002 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2003 X509_supported_extension() has also been added which returns 1 if a
2004 particular extension is supported.
2007 *) Modify the behaviour of EVP cipher functions in similar way to digests
2008 to retain compatibility with existing code.
2011 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2012 compatibility with existing code. In particular the 'ctx' parameter does
2013 not have to be to be initialized before the call to EVP_DigestInit() and
2014 it is tidied up after a call to EVP_DigestFinal(). New function
2015 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2016 EVP_MD_CTX_copy() changed to not require the destination to be
2017 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2018 requires the destination to be valid.
2020 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2021 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2024 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2025 so that complete 'Handshake' protocol structures are kept in memory
2026 instead of overwriting 'msg_type' and 'length' with 'body' data.
2029 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2030 [Massimo Santin via Richard Levitte]
2032 *) Major restructuring to the underlying ENGINE code. This includes
2033 reduction of linker bloat, separation of pure "ENGINE" manipulation
2034 (initialisation, etc) from functionality dealing with implementations
2035 of specific crypto iterfaces. This change also introduces integrated
2036 support for symmetric ciphers and digest implementations - so ENGINEs
2037 can now accelerate these by providing EVP_CIPHER and EVP_MD
2038 implementations of their own. This is detailed in crypto/engine/README
2039 as it couldn't be adequately described here. However, there are a few
2040 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2041 were changed in the original introduction of ENGINE code have now
2042 reverted back - the hooking from this code to ENGINE is now a good
2043 deal more passive and at run-time, operations deal directly with
2044 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2045 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2046 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2047 they were not being used by the framework as there is no concept of a
2048 BIGNUM_METHOD and they could not be generalised to the new
2049 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2050 ENGINE_cpy() has been removed as it cannot be consistently defined in
2054 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2057 *) Change mkdef.pl to sort symbols that get the same entry number,
2058 and make sure the automatically generated functions ERR_load_*
2059 become part of libeay.num as well.
2062 *) New function SSL_renegotiate_pending(). This returns true once
2063 renegotiation has been requested (either SSL_renegotiate() call
2064 or HelloRequest/ClientHello receveived from the peer) and becomes
2065 false once a handshake has been completed.
2066 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2067 sends a HelloRequest, but does not ensure that a handshake takes
2068 place. SSL_renegotiate_pending() is useful for checking if the
2069 client has followed the request.)
2072 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2073 By default, clients may request session resumption even during
2074 renegotiation (if session ID contexts permit); with this option,
2075 session resumption is possible only in the first handshake.
2077 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2078 more bits available for options that should not be part of
2079 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2082 *) Add some demos for certificate and certificate request creation.
2085 *) Make maximum certificate chain size accepted from the peer application
2086 settable (SSL*_get/set_max_cert_list()), as proposed by
2087 "Douglas E. Engert" <deengert@anl.gov>.
2090 *) Add support for shared libraries for Unixware-7
2091 (Boyd Lynn Gerber <gerberb@zenez.com>).
2094 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2095 be done prior to destruction. Use this to unload error strings from
2096 ENGINEs that load their own error strings. NB: This adds two new API
2097 functions to "get" and "set" this destroy handler in an ENGINE.
2100 *) Alter all existing ENGINE implementations (except "openssl" and
2101 "openbsd") to dynamically instantiate their own error strings. This
2102 makes them more flexible to be built both as statically-linked ENGINEs
2103 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2104 Also, add stub code to each that makes building them as self-contained
2105 shared-libraries easier (see README.ENGINE).
2108 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2109 implementations into applications that are completely implemented in
2110 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2111 commands that can be used to configure what shared-library to load and
2112 to control aspects of the way it is handled. Also, made an update to
2113 the README.ENGINE file that brings its information up-to-date and
2114 provides some information and instructions on the "dynamic" ENGINE
2115 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2118 *) Make it possible to unload ranges of ERR strings with a new
2119 "ERR_unload_strings" function.
2122 *) Add a copy() function to EVP_MD.
2125 *) Make EVP_MD routines take a context pointer instead of just the
2126 md_data void pointer.
2129 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2130 that the digest can only process a single chunk of data
2131 (typically because it is provided by a piece of
2132 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2133 is only going to provide a single chunk of data, and hence the
2134 framework needn't accumulate the data for oneshot drivers.
2137 *) As with "ERR", make it possible to replace the underlying "ex_data"
2138 functions. This change also alters the storage and management of global
2139 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2140 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2141 index counters. The API functions that use this state have been changed
2142 to take a "class_index" rather than pointers to the class's local STACK
2143 and counter, and there is now an API function to dynamically create new
2144 classes. This centralisation allows us to (a) plug a lot of the
2145 thread-safety problems that existed, and (b) makes it possible to clean
2146 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2147 such data would previously have always leaked in application code and
2148 workarounds were in place to make the memory debugging turn a blind eye
2149 to it. Application code that doesn't use this new function will still
2150 leak as before, but their memory debugging output will announce it now
2151 rather than letting it slide.
2153 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2154 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2155 has a return value to indicate success or failure.
2158 *) Make it possible to replace the underlying "ERR" functions such that the
2159 global state (2 LHASH tables and 2 locks) is only used by the "default"
2160 implementation. This change also adds two functions to "get" and "set"
2161 the implementation prior to it being automatically set the first time
2162 any other ERR function takes place. Ie. an application can call "get",
2163 pass the return value to a module it has just loaded, and that module
2164 can call its own "set" function using that value. This means the
2165 module's "ERR" operations will use (and modify) the error state in the
2166 application and not in its own statically linked copy of OpenSSL code.
2169 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2170 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2171 the operation, and provides a more encapsulated way for external code
2172 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2173 to use these functions rather than manually incrementing the counts.
2175 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2178 *) Add EVP test program.
2181 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2184 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2185 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2186 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2187 These allow a CRL to be built without having to access X509_CRL fields
2188 directly. Modify 'ca' application to use new functions.
2191 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2192 bug workarounds. Rollback attack detection is a security feature.
2193 The problem will only arise on OpenSSL servers when TLSv1 is not
2194 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2195 Software authors not wanting to support TLSv1 will have special reasons
2196 for their choice and can explicitly enable this option.
2197 [Bodo Moeller, Lutz Jaenicke]
2199 *) Rationalise EVP so it can be extended: don't include a union of
2200 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2201 (similar to those existing for EVP_CIPHER_CTX).
2206 EVP_MD_CTX_init(&md); /* new function call */
2207 EVP_DigestInit(&md, EVP_sha1());
2208 EVP_DigestUpdate(&md, in, len);
2209 EVP_DigestFinal(&md, out, NULL);
2210 EVP_MD_CTX_cleanup(&md); /* new function call */
2214 *) Make DES key schedule conform to the usual scheme, as well as
2215 correcting its structure. This means that calls to DES functions
2216 now have to pass a pointer to a des_key_schedule instead of a
2217 plain des_key_schedule (which was actually always a pointer
2220 des_key_schedule ks;
2222 des_set_key_checked(..., &ks);
2223 des_ncbc_encrypt(..., &ks, ...);
2225 (Note that a later change renames 'des_...' into 'DES_...'.)
2228 *) Initial reduction of linker bloat: the use of some functions, such as
2229 PEM causes large amounts of unused functions to be linked in due to
2230 poor organisation. For example pem_all.c contains every PEM function
2231 which has a knock on effect of linking in large amounts of (unused)
2232 ASN1 code. Grouping together similar functions and splitting unrelated
2233 functions prevents this.
2236 *) Cleanup of EVP macros.
2239 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2240 correct _ecb suffix.
2243 *) Add initial OCSP responder support to ocsp application. The
2244 revocation information is handled using the text based index
2245 use by the ca application. The responder can either handle
2246 requests generated internally, supplied in files (for example
2247 via a CGI script) or using an internal minimal server.
2250 *) Add configuration choices to get zlib compression for TLS.
2253 *) Changes to Kerberos SSL for RFC 2712 compliance:
2254 1. Implemented real KerberosWrapper, instead of just using
2255 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2256 2. Implemented optional authenticator field of KerberosWrapper.
2258 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2259 and authenticator structs; see crypto/krb5/.
2261 Generalized Kerberos calls to support multiple Kerberos libraries.
2262 [Vern Staats <staatsvr@asc.hpc.mil>,
2263 Jeffrey Altman <jaltman@columbia.edu>
2264 via Richard Levitte]
2266 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2267 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2268 values for each of the key sizes rather than having just
2269 parameters (and 'speed' generating keys each time).
2272 *) Speed up EVP routines.
2275 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2276 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2277 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2278 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2280 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2281 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2282 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2285 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2287 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2290 *) Added the OS2-EMX target.
2291 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2293 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2294 to support NCONF routines in extension code. New function CONF_set_nconf()
2295 to allow functions which take an NCONF to also handle the old LHASH
2296 structure: this means that the old CONF compatible routines can be
2297 retained (in particular wrt extensions) without having to duplicate the
2298 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2301 *) Enhance the general user interface with mechanisms for inner control
2302 and with possibilities to have yes/no kind of prompts.
2305 *) Change all calls to low level digest routines in the library and
2306 applications to use EVP. Add missing calls to HMAC_cleanup() and
2307 don't assume HMAC_CTX can be copied using memcpy().
2308 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2310 *) Add the possibility to control engines through control names but with
2311 arbitrary arguments instead of just a string.
2312 Change the key loaders to take a UI_METHOD instead of a callback
2313 function pointer. NOTE: this breaks binary compatibility with earlier
2314 versions of OpenSSL [engine].
2315 Adapt the nCipher code for these new conditions and add a card insertion
2319 *) Enhance the general user interface with mechanisms to better support
2320 dialog box interfaces, application-defined prompts, the possibility
2321 to use defaults (for example default passwords from somewhere else)
2322 and interrupts/cancellations.
2325 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2326 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2329 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2330 tidy up some unnecessarily weird code in 'sk_new()').
2331 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2333 *) Change the key loading routines for ENGINEs to use the same kind
2334 callback (pem_password_cb) as all other routines that need this
2338 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2339 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2340 than this minimum value is recommended.
2343 *) New random seeder for OpenVMS, using the system process statistics
2344 that are easily reachable.
2347 *) Windows apparently can't transparently handle global
2348 variables defined in DLLs. Initialisations such as:
2350 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2352 wont compile. This is used by the any applications that need to
2353 declare their own ASN1 modules. This was fixed by adding the option
2354 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2355 needed for static libraries under Win32.
2358 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2359 setting of purpose and trust fields. New X509_STORE trust and
2360 purpose functions and tidy up setting in other SSL functions.
2363 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2364 structure. These are inherited by X509_STORE_CTX when it is
2365 initialised. This allows various defaults to be set in the
2366 X509_STORE structure (such as flags for CRL checking and custom
2367 purpose or trust settings) for functions which only use X509_STORE_CTX
2368 internally such as S/MIME.
2370 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2371 trust settings if they are not set in X509_STORE. This allows X509_STORE
2372 purposes and trust (in S/MIME for example) to override any set by default.
2374 Add command line options for CRL checking to smime, s_client and s_server
2378 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2379 are set then the CRL is looked up in the X509_STORE structure and
2380 its validity and signature checked, then if the certificate is found
2381 in the CRL the verify fails with a revoked error.
2383 Various new CRL related callbacks added to X509_STORE_CTX structure.
2385 Command line options added to 'verify' application to support this.
2387 This needs some additional work, such as being able to handle multiple
2388 CRLs with different times, extension based lookup (rather than just
2389 by subject name) and ultimately more complete V2 CRL extension
2393 *) Add a general user interface API (crypto/ui/). This is designed
2394 to replace things like des_read_password and friends (backward
2395 compatibility functions using this new API are provided).
2396 The purpose is to remove prompting functions from the DES code
2397 section as well as provide for prompting through dialog boxes in
2398 a window system and the like.
2401 *) Add "ex_data" support to ENGINE so implementations can add state at a
2402 per-structure level rather than having to store it globally.
2405 *) Make it possible for ENGINE structures to be copied when retrieved by
2406 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2407 This causes the "original" ENGINE structure to act like a template,
2408 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2409 operational state can be localised to each ENGINE structure, despite the
2410 fact they all share the same "methods". New ENGINE structures returned in
2411 this case have no functional references and the return value is the single
2412 structural reference. This matches the single structural reference returned
2413 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2417 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2418 needs to match any other type at all we need to manually clear the
2422 *) Changes to the "openssl engine" utility to include;
2423 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2424 about an ENGINE's available control commands.
2425 - executing control commands from command line arguments using the
2426 '-pre' and '-post' switches. '-post' is only used if '-t' is
2427 specified and the ENGINE is successfully initialised. The syntax for
2428 the individual commands are colon-separated, for example;
2429 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2432 *) New dynamic control command support for ENGINEs. ENGINEs can now
2433 declare their own commands (numbers), names (strings), descriptions,
2434 and input types for run-time discovery by calling applications. A
2435 subset of these commands are implicitly classed as "executable"
2436 depending on their input type, and only these can be invoked through
2437 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2438 can be based on user input, config files, etc). The distinction is
2439 that "executable" commands cannot return anything other than a boolean
2440 result and can only support numeric or string input, whereas some
2441 discoverable commands may only be for direct use through
2442 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2443 pointers, or other custom uses. The "executable" commands are to
2444 support parameterisations of ENGINE behaviour that can be
2445 unambiguously defined by ENGINEs and used consistently across any
2446 OpenSSL-based application. Commands have been added to all the
2447 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2448 control over shared-library paths without source code alterations.
2451 *) Changed all ENGINE implementations to dynamically allocate their
2452 ENGINEs rather than declaring them statically. Apart from this being
2453 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2454 this also allows the implementations to compile without using the
2455 internal engine_int.h header.
2458 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2459 'const' value. Any code that should be able to modify a RAND_METHOD
2460 should already have non-const pointers to it (ie. they should only
2461 modify their own ones).
2464 *) Made a variety of little tweaks to the ENGINE code.
2465 - "atalla" and "ubsec" string definitions were moved from header files
2466 to C code. "nuron" string definitions were placed in variables
2467 rather than hard-coded - allowing parameterisation of these values
2468 later on via ctrl() commands.
2469 - Removed unused "#if 0"'d code.
2470 - Fixed engine list iteration code so it uses ENGINE_free() to release
2471 structural references.
2472 - Constified the RAND_METHOD element of ENGINE structures.
2473 - Constified various get/set functions as appropriate and added
2474 missing functions (including a catch-all ENGINE_cpy that duplicates
2475 all ENGINE values onto a new ENGINE except reference counts/state).
2476 - Removed NULL parameter checks in get/set functions. Setting a method
2477 or function to NULL is a way of cancelling out a previously set
2478 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2479 and doesn't justify the extra error symbols and code.
2480 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2481 flags from engine_int.h to engine.h.
2482 - Changed prototypes for ENGINE handler functions (init(), finish(),
2483 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2486 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2487 to the algorithm using long division. The binary algorithm can be
2488 used only if the modulus is odd. On 32-bit systems, it is faster
2489 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2490 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2491 up to 450 bits. In 64-bit environments, the binary algorithm
2492 appears to be advantageous for much longer moduli; here we use it
2493 for moduli up to 2048 bits.
2496 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2497 could not support the combine flag in choice fields.
2500 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2501 extensions from a certificate request to the certificate.
2504 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2505 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2506 file: this allows the display of the certificate about to be
2507 signed to be customised, to allow certain fields to be included
2508 or excluded and extension details. The old system didn't display
2509 multicharacter strings properly, omitted fields not in the policy
2510 and couldn't display additional details such as extensions.
2513 *) Function EC_POINTs_mul for multiple scalar multiplication
2514 of an arbitrary number of elliptic curve points
2515 \sum scalars[i]*points[i],
2516 optionally including the generator defined for the EC_GROUP:
2517 scalar*generator + \sum scalars[i]*points[i].
2519 EC_POINT_mul is a simple wrapper function for the typical case
2520 that the point list has just one item (besides the optional
2524 *) First EC_METHODs for curves over GF(p):
2526 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2527 operations and provides various method functions that can also
2528 operate with faster implementations of modular arithmetic.
2530 EC_GFp_mont_method() reuses most functions that are part of
2531 EC_GFp_simple_method, but uses Montgomery arithmetic.
2533 [Bodo Moeller; point addition and point doubling
2534 implementation directly derived from source code provided by
2535 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2537 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2538 crypto/ec/ec_lib.c):
2540 Curves are EC_GROUP objects (with an optional group generator)
2541 based on EC_METHODs that are built into the library.
2543 Points are EC_POINT objects based on EC_GROUP objects.
2545 Most of the framework would be able to handle curves over arbitrary
2546 finite fields, but as there are no obvious types for fields other
2547 than GF(p), some functions are limited to that for now.
2550 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2551 that the file contains a complete HTTP response.
2554 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2555 change the def and num file printf format specifier from "%-40sXXX"
2556 to "%-39s XXX". The latter will always guarantee a space after the
2557 field while the former will cause them to run together if the field
2558 is 40 of more characters long.
2561 *) Constify the cipher and digest 'method' functions and structures
2562 and modify related functions to take constant EVP_MD and EVP_CIPHER
2566 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2567 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2570 *) Modify EVP_Digest*() routines so they now return values. Although the
2571 internal software routines can never fail additional hardware versions
2575 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2577 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2578 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2581 ERR_R_NESTED_ASN1_ERROR
2583 ERR_R_MISSING_ASN1_EOS
2584 were 4 .. 9, conflicting with
2585 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2587 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2588 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2590 Add new error code 'ERR_R_INTERNAL_ERROR'.
2593 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2597 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2598 sets the subject name for a new request or supersedes the
2599 subject name in a given request. Formats that can be parsed are
2600 'CN=Some Name, OU=myOU, C=IT'
2602 'CN=Some Name/OU=myOU/C=IT'.
2604 Add options '-batch' and '-verbose' to 'openssl req'.
2605 [Massimiliano Pala <madwolf@hackmasters.net>]
2607 *) Introduce the possibility to access global variables through
2608 functions on platform were that's the best way to handle exporting
2609 global variables in shared libraries. To enable this functionality,
2610 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2611 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2612 is normally done by Configure or something similar).
2614 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2615 in the source file (foo.c) like this:
2617 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2618 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2620 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2621 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2623 OPENSSL_DECLARE_GLOBAL(int,foo);
2624 #define foo OPENSSL_GLOBAL_REF(foo)
2625 OPENSSL_DECLARE_GLOBAL(double,bar);
2626 #define bar OPENSSL_GLOBAL_REF(bar)
2628 The #defines are very important, and therefore so is including the
2629 header file everywhere where the defined globals are used.
2631 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2632 of ASN.1 items, but that structure is a bit different.
2634 The largest change is in util/mkdef.pl which has been enhanced with
2635 better and easier to understand logic to choose which symbols should
2636 go into the Windows .def files as well as a number of fixes and code
2637 cleanup (among others, algorithm keywords are now sorted
2638 lexicographically to avoid constant rewrites).
2641 *) In BN_div() keep a copy of the sign of 'num' before writing the
2642 result to 'rm' because if rm==num the value will be overwritten
2643 and produce the wrong result if 'num' is negative: this caused
2644 problems with BN_mod() and BN_nnmod().
2647 *) Function OCSP_request_verify(). This checks the signature on an
2648 OCSP request and verifies the signer certificate. The signer
2649 certificate is just checked for a generic purpose and OCSP request
2653 *) Add OCSP_check_validity() function to check the validity of OCSP
2654 responses. OCSP responses are prepared in real time and may only
2655 be a few seconds old. Simply checking that the current time lies
2656 between thisUpdate and nextUpdate max reject otherwise valid responses
2657 caused by either OCSP responder or client clock inaccuracy. Instead
2658 we allow thisUpdate and nextUpdate to fall within a certain period of
2659 the current time. The age of the response can also optionally be
2660 checked. Two new options -validity_period and -status_age added to
2664 *) If signature or public key algorithm is unrecognized print out its
2665 OID rather that just UNKNOWN.
2668 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2669 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2670 ID to be generated from the issuer certificate alone which can then be
2671 passed to OCSP_id_issuer_cmp().
2674 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2675 ASN1 modules to export functions returning ASN1_ITEM pointers
2676 instead of the ASN1_ITEM structures themselves. This adds several
2677 new macros which allow the underlying ASN1 function/structure to
2678 be accessed transparently. As a result code should not use ASN1_ITEM
2679 references directly (such as &X509_it) but instead use the relevant
2680 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2681 use of the new ASN1 code on platforms where exporting structures
2682 is problematical (for example in shared libraries) but exporting
2683 functions returning pointers to structures is not.
2686 *) Add support for overriding the generation of SSL/TLS session IDs.
2687 These callbacks can be registered either in an SSL_CTX or per SSL.
2688 The purpose of this is to allow applications to control, if they wish,
2689 the arbitrary values chosen for use as session IDs, particularly as it
2690 can be useful for session caching in multiple-server environments. A
2691 command-line switch for testing this (and any client code that wishes
2692 to use such a feature) has been added to "s_server".
2693 [Geoff Thorpe, Lutz Jaenicke]
2695 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2696 of the form '#if defined(...) || defined(...) || ...' and
2697 '#if !defined(...) && !defined(...) && ...'. This also avoids
2698 the growing number of special cases it was previously handling.
2701 *) Make all configuration macros available for application by making
2702 sure they are available in opensslconf.h, by giving them names starting
2703 with "OPENSSL_" to avoid conflicts with other packages and by making
2704 sure e_os2.h will cover all platform-specific cases together with
2706 Additionally, it is now possible to define configuration/platform-
2707 specific names (called "system identities"). In the C code, these
2708 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2709 macro with the name beginning with "OPENSSL_SYS_", which is determined
2710 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2714 *) New option -set_serial to 'req' and 'x509' this allows the serial
2715 number to use to be specified on the command line. Previously self
2716 signed certificates were hard coded with serial number 0 and the
2717 CA options of 'x509' had to use a serial number in a file which was
2721 *) New options to 'ca' utility to support V2 CRL entry extensions.
2722 Currently CRL reason, invalidity date and hold instruction are
2723 supported. Add new CRL extensions to V3 code and some new objects.
2726 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2727 disable standard block padding (aka PKCS#5 padding) in the EVP
2728 API, which was previously mandatory. This means that the data is
2729 not padded in any way and so the total length much be a multiple
2730 of the block size, otherwise an error occurs.
2733 *) Initial (incomplete) OCSP SSL support.
2736 *) New function OCSP_parse_url(). This splits up a URL into its host,
2737 port and path components: primarily to parse OCSP URLs. New -url
2738 option to ocsp utility.
2741 *) New nonce behavior. The return value of OCSP_check_nonce() now
2742 reflects the various checks performed. Applications can decide
2743 whether to tolerate certain situations such as an absent nonce
2744 in a response when one was present in a request: the ocsp application
2745 just prints out a warning. New function OCSP_add1_basic_nonce()
2746 this is to allow responders to include a nonce in a response even if
2747 the request is nonce-less.
2750 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2751 skipped when using openssl x509 multiple times on a single input file,
2752 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2755 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2756 set string type: to handle setting ASN1_TIME structures. Fix ca
2757 utility to correctly initialize revocation date of CRLs.
2760 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2761 the clients preferred ciphersuites and rather use its own preferences.
2762 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2763 Internet Explorer by ensuring unchanged hash method during stepup.
2764 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2767 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2768 to aes and add a new 'exist' option to print out symbols that don't
2772 *) Additional options to ocsp utility to allow flags to be set and
2773 additional certificates supplied.
2776 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2777 OCSP client a number of certificate to only verify the response
2781 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2782 handle the new API. Currently only ECB, CBC modes supported. Add new
2785 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2786 Encryption Standard (AES) Ciphersuites for Transport Layer
2787 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2788 not enabled by default and were not part of the "ALL" ciphersuite
2789 alias because they were not yet official; they could be
2790 explicitly requested by specifying the "AESdraft" ciphersuite
2791 group alias. In the final release of OpenSSL 0.9.7, the group
2792 alias is called "AES" and is part of "ALL".)
2793 [Ben Laurie, Steve Henson, Bodo Moeller]
2795 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2796 request to response.
2799 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2800 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2801 extract information from a certificate request. OCSP_response_create()
2802 creates a response and optionally adds a basic response structure.
2803 OCSP_basic_add1_status() adds a complete single response to a basic
2804 response and returns the OCSP_SINGLERESP structure just added (to allow
2805 extensions to be included for example). OCSP_basic_add1_cert() adds a
2806 certificate to a basic response and OCSP_basic_sign() signs a basic
2807 response with various flags. New helper functions ASN1_TIME_check()
2808 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2809 (converts ASN1_TIME to GeneralizedTime).
2812 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2813 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2814 structure from a certificate. X509_pubkey_digest() digests the public_key
2815 contents: this is used in various key identifiers.
2818 *) Make sk_sort() tolerate a NULL argument.
2819 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2821 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2822 passed by the function are trusted implicitly. If any of them signed the
2823 response then it is assumed to be valid and is not verified.
2826 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2827 to data. This was previously part of the PKCS7 ASN1 code. This
2828 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2829 [Steve Henson, reported by Kenneth R. Robinette
2830 <support@securenetterm.com>]
2832 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2833 routines: without these tracing memory leaks is very painful.
2834 Fix leaks in PKCS12 and PKCS7 routines.
2837 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2838 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2839 effectively meant GeneralizedTime would never be used. Now it
2840 is initialised to -1 but X509_time_adj() now has to check the value
2841 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2842 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2843 [Steve Henson, reported by Kenneth R. Robinette
2844 <support@securenetterm.com>]
2846 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2847 result in a zero length in the ASN1_INTEGER structure which was
2848 not consistent with the structure when d2i_ASN1_INTEGER() was used
2849 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2850 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2851 where it did not print out a minus for negative ASN1_INTEGER.
2854 *) Add summary printout to ocsp utility. The various functions which
2855 convert status values to strings have been renamed to:
2856 OCSP_response_status_str(), OCSP_cert_status_str() and
2857 OCSP_crl_reason_str() and are no longer static. New options
2858 to verify nonce values and to disable verification. OCSP response
2859 printout format cleaned up.
2862 *) Add additional OCSP certificate checks. These are those specified
2863 in RFC2560. This consists of two separate checks: the CA of the
2864 certificate being checked must either be the OCSP signer certificate
2865 or the issuer of the OCSP signer certificate. In the latter case the
2866 OCSP signer certificate must contain the OCSP signing extended key
2867 usage. This check is performed by attempting to match the OCSP
2868 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2869 in the OCSP_CERTID structures of the response.
2872 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2873 and related routines. This uses the standard OpenSSL certificate
2874 verify routines to perform initial checks (just CA validity) and
2875 to obtain the certificate chain. Then additional checks will be
2876 performed on the chain. Currently the root CA is checked to see
2877 if it is explicitly trusted for OCSP signing. This is used to set
2878 a root CA as a global signing root: that is any certificate that
2879 chains to that CA is an acceptable OCSP signing certificate.
2882 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2883 extensions from a separate configuration file.
2884 As when reading extensions from the main configuration file,
2885 the '-extensions ...' option may be used for specifying the
2887 [Massimiliano Pala <madwolf@comune.modena.it>]
2889 *) New OCSP utility. Allows OCSP requests to be generated or
2890 read. The request can be sent to a responder and the output
2891 parsed, outputed or printed in text form. Not complete yet:
2892 still needs to check the OCSP response validity.
2895 *) New subcommands for 'openssl ca':
2896 'openssl ca -status <serial>' prints the status of the cert with
2897 the given serial number (according to the index file).
2898 'openssl ca -updatedb' updates the expiry status of certificates
2900 [Massimiliano Pala <madwolf@comune.modena.it>]
2902 *) New '-newreq-nodes' command option to CA.pl. This is like
2903 '-newreq', but calls 'openssl req' with the '-nodes' option
2904 so that the resulting key is not encrypted.
2905 [Damien Miller <djm@mindrot.org>]
2907 *) New configuration for the GNU Hurd.
2908 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2910 *) Initial code to implement OCSP basic response verify. This
2911 is currently incomplete. Currently just finds the signer's
2912 certificate and verifies the signature on the response.
2915 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2916 value of OPENSSLDIR. This is available via the new '-d' option
2917 to 'openssl version', and is also included in 'openssl version -a'.
2920 *) Allowing defining memory allocation callbacks that will be given
2921 file name and line number information in additional arguments
2922 (a const char* and an int). The basic functionality remains, as
2923 well as the original possibility to just replace malloc(),
2924 realloc() and free() by functions that do not know about these
2925 additional arguments. To register and find out the current
2926 settings for extended allocation functions, the following
2927 functions are provided:
2929 CRYPTO_set_mem_ex_functions
2930 CRYPTO_set_locked_mem_ex_functions
2931 CRYPTO_get_mem_ex_functions
2932 CRYPTO_get_locked_mem_ex_functions
2934 These work the same way as CRYPTO_set_mem_functions and friends.
2935 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2936 extended allocation function is enabled.
2937 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2938 a conventional allocation function is enabled.
2939 [Richard Levitte, Bodo Moeller]
2941 *) Finish off removing the remaining LHASH function pointer casts.
2942 There should no longer be any prototype-casting required when using
2943 the LHASH abstraction, and any casts that remain are "bugs". See
2944 the callback types and macros at the head of lhash.h for details
2945 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2948 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2949 If /dev/[u]random devices are not available or do not return enough
2950 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2952 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2953 /etc/entropy will be queried once each in this sequence, quering stops
2954 when enough entropy was collected without querying more sockets.
2957 *) Change the Unix RAND_poll() variant to be able to poll several
2958 random devices, as specified by DEVRANDOM, until a sufficient amount
2959 of data has been collected. We spend at most 10 ms on each file
2960 (select timeout) and read in non-blocking mode. DEVRANDOM now
2961 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2962 (previously it was just the string "/dev/urandom"), so on typical
2963 platforms the 10 ms delay will never occur.
2964 Also separate out the Unix variant to its own file, rand_unix.c.
2965 For VMS, there's a currently-empty rand_vms.c.
2968 *) Move OCSP client related routines to ocsp_cl.c. These
2969 provide utility functions which an application needing
2970 to issue a request to an OCSP responder and analyse the
2971 response will typically need: as opposed to those which an
2972 OCSP responder itself would need which will be added later.
2974 OCSP_request_sign() signs an OCSP request with an API similar
2975 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2976 response. OCSP_response_get1_basic() extracts basic response
2977 from response. OCSP_resp_find_status(): finds and extracts status
2978 information from an OCSP_CERTID structure (which will be created
2979 when the request structure is built). These are built from lower
2980 level functions which work on OCSP_SINGLERESP structures but
2981 wont normally be used unless the application wishes to examine
2982 extensions in the OCSP response for example.
2984 Replace nonce routines with a pair of functions.
2985 OCSP_request_add1_nonce() adds a nonce value and optionally
2986 generates a random value. OCSP_check_nonce() checks the
2987 validity of the nonce in an OCSP response.
2990 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2991 This doesn't copy the supplied OCSP_CERTID and avoids the
2992 need to free up the newly created id. Change return type
2993 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2994 This can then be used to add extensions to the request.
2995 Deleted OCSP_request_new(), since most of its functionality
2996 is now in OCSP_REQUEST_new() (and the case insensitive name
2997 clash) apart from the ability to set the request name which
2998 will be added elsewhere.
3001 *) Update OCSP API. Remove obsolete extensions argument from
3002 various functions. Extensions are now handled using the new
3003 OCSP extension code. New simple OCSP HTTP function which
3004 can be used to send requests and parse the response.
3007 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3008 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3009 uses the special reorder version of SET OF to sort the attributes
3010 and reorder them to match the encoded order. This resolves a long
3011 standing problem: a verify on a PKCS7 structure just after signing
3012 it used to fail because the attribute order did not match the
3013 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3014 it uses the received order. This is necessary to tolerate some broken
3015 software that does not order SET OF. This is handled by encoding
3016 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3017 to produce the required SET OF.
3020 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3021 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3022 files to get correct declarations of the ASN.1 item variables.
3025 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3026 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3027 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3028 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3029 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3030 ASN1_ITEM and no wrapper functions.
3033 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3034 replace the old function pointer based I/O routines. Change most of
3035 the *_d2i_bio() and *_d2i_fp() functions to use these.
3038 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3039 lines, recognice more "algorithms" that can be deselected, and make
3040 it complain about algorithm deselection that isn't recognised.
3043 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3044 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3045 to use new functions. Add NO_ASN1_OLD which can be set to remove
3046 some old style ASN1 functions: this can be used to determine if old
3047 code will still work when these eventually go away.
3050 *) New extension functions for OCSP structures, these follow the
3051 same conventions as certificates and CRLs.
3054 *) New function X509V3_add1_i2d(). This automatically encodes and
3055 adds an extension. Its behaviour can be customised with various
3056 flags to append, replace or delete. Various wrappers added for
3057 certifcates and CRLs.
3060 *) Fix to avoid calling the underlying ASN1 print routine when
3061 an extension cannot be parsed. Correct a typo in the
3062 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3065 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3066 entries for variables.
3069 *) Add functionality to apps/openssl.c for detecting locking
3070 problems: As the program is single-threaded, all we have
3071 to do is register a locking callback using an array for
3072 storing which locks are currently held by the program.
3075 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3076 SSL_get_ex_data_X509_STORE_idx(), which is used in
3077 ssl_verify_cert_chain() and thus can be called at any time
3078 during TLS/SSL handshakes so that thread-safety is essential.
3079 Unfortunately, the ex_data design is not at all suited
3080 for multi-threaded use, so it probably should be abolished.
3083 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3084 [Broadcom, tweaked and integrated by Geoff Thorpe]
3086 *) Move common extension printing code to new function
3087 X509V3_print_extensions(). Reorganise OCSP print routines and
3088 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3091 *) New function X509_signature_print() to remove duplication in some
3095 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3096 set (this was treated exactly the same as SET OF previously). This
3097 is used to reorder the STACK representing the structure to match the
3098 encoding. This will be used to get round a problem where a PKCS7
3099 structure which was signed could not be verified because the STACK
3100 order did not reflect the encoded order.
3103 *) Reimplement the OCSP ASN1 module using the new code.
3106 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3107 for its ASN1 operations. The old style function pointers still exist
3108 for now but they will eventually go away.
3111 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3112 completely replaces the old ASN1 functionality with a table driven
3113 encoder and decoder which interprets an ASN1_ITEM structure describing
3114 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3115 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3116 has also been converted to the new form.
3119 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3120 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3121 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3122 for negative moduli.
3125 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3126 of not touching the result's sign bit.
3129 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3133 *) Changed the LHASH code to use prototypes for callbacks, and created
3134 macros to declare and implement thin (optionally static) functions
3135 that provide type-safety and avoid function pointer casting for the
3136 type-specific callbacks.
3139 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3141 [Veers Staats <staatsvr@asc.hpc.mil>,
3142 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3144 *) Reformat the FAQ so the different questions and answers can be divided
3145 in sections depending on the subject.
3148 *) Have the zlib compression code load ZLIB.DLL dynamically under
3152 *) New function BN_mod_sqrt for computing square roots modulo a prime
3153 (using the probabilistic Tonelli-Shanks algorithm unless
3154 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3155 be handled deterministically).
3156 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3158 *) Make BN_mod_inverse faster by explicitly handling small quotients
3159 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3160 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3163 *) New function BN_kronecker.
3166 *) Fix BN_gcd so that it works on negative inputs; the result is
3167 positive unless both parameters are zero.
3168 Previously something reasonably close to an infinite loop was
3169 possible because numbers could be growing instead of shrinking
3170 in the implementation of Euclid's algorithm.
3173 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3174 sign of the number in question.
3176 Fix BN_is_word(a,w) to work correctly for w == 0.
3178 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3179 because its test if the absolute value of 'a' equals 'w'.
3180 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3181 it exists mostly for use in the implementations of BN_is_zero(),
3182 BN_is_one(), and BN_is_word().
3185 *) New function BN_swap.
3188 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3189 the exponentiation functions are more likely to produce reasonable
3190 results on negative inputs.
3193 *) Change BN_mod_mul so that the result is always non-negative.
3194 Previously, it could be negative if one of the factors was negative;
3195 I don't think anyone really wanted that behaviour.
3198 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3199 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3200 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3201 and add new functions:
3210 BN_mod_lshift1_quick
3214 These functions always generate non-negative results.
3216 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3217 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3219 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3220 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3221 be reduced modulo m.
3222 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3225 The following entry accidentily appeared in the CHANGES file
3226 distributed with OpenSSL 0.9.7. The modifications described in
3227 it do *not* apply to OpenSSL 0.9.7.
3229 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3230 was actually never needed) and in BN_mul(). The removal in BN_mul()
3231 required a small change in bn_mul_part_recursive() and the addition
3232 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3233 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3234 bn_sub_words() and bn_add_words() except they take arrays with
3239 *) In 'openssl passwd', verify passwords read from the terminal
3240 unless the '-salt' option is used (which usually means that
3241 verification would just waste user's time since the resulting
3242 hash is going to be compared with some given password hash)
3243 or the new '-noverify' option is used.
3245 This is an incompatible change, but it does not affect
3246 non-interactive use of 'openssl passwd' (passwords on the command
3247 line, '-stdin' option, '-in ...' option) and thus should not
3251 *) Remove all references to RSAref, since there's no more need for it.
3254 *) Make DSO load along a path given through an environment variable
3255 (SHLIB_PATH) with shl_load().
3258 *) Constify the ENGINE code as a result of BIGNUM constification.
3259 Also constify the RSA code and most things related to it. In a
3260 few places, most notable in the depth of the ASN.1 code, ugly
3261 casts back to non-const were required (to be solved at a later
3265 *) Make it so the openssl application has all engines loaded by default.
3268 *) Constify the BIGNUM routines a little more.
3271 *) Add the following functions:
3273 ENGINE_load_cswift()
3275 ENGINE_load_atalla()
3277 ENGINE_load_builtin_engines()
3279 That way, an application can itself choose if external engines that
3280 are built-in in OpenSSL shall ever be used or not. The benefit is
3281 that applications won't have to be linked with libdl or other dso
3282 libraries unless it's really needed.
3284 Changed 'openssl engine' to load all engines on demand.
3285 Changed the engine header files to avoid the duplication of some
3286 declarations (they differed!).
3289 *) 'openssl engine' can now list capabilities.
3292 *) Better error reporting in 'openssl engine'.
3295 *) Never call load_dh_param(NULL) in s_server.
3298 *) Add engine application. It can currently list engines by name and
3299 identity, and test if they are actually available.
3302 *) Improve RPM specification file by forcing symbolic linking and making
3303 sure the installed documentation is also owned by root.root.
3304 [Damien Miller <djm@mindrot.org>]
3306 *) Give the OpenSSL applications more possibilities to make use of
3307 keys (public as well as private) handled by engines.
3310 *) Add OCSP code that comes from CertCo.
3313 *) Add VMS support for the Rijndael code.
3316 *) Added untested support for Nuron crypto accelerator.
3319 *) Add support for external cryptographic devices. This code was
3320 previously distributed separately as the "engine" branch.
3321 [Geoff Thorpe, Richard Levitte]
3323 *) Rework the filename-translation in the DSO code. It is now possible to
3324 have far greater control over how a "name" is turned into a filename
3325 depending on the operating environment and any oddities about the
3326 different shared library filenames on each system.
3329 *) Support threads on FreeBSD-elf in Configure.
3332 *) Fix for SHA1 assembly problem with MASM: it produces
3333 warnings about corrupt line number information when assembling
3334 with debugging information. This is caused by the overlapping
3336 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3339 NCONF_get_number() has no error checking at all. As a replacement,
3340 NCONF_get_number_e() is defined (_e for "error checking") and is
3341 promoted strongly. The old NCONF_get_number is kept around for
3342 binary backward compatibility.
3343 Make it possible for methods to load from something other than a BIO,
3344 by providing a function pointer that is given a name instead of a BIO.
3345 For example, this could be used to load configuration data from an
3349 *) Fix for non blocking accept BIOs. Added new I/O special reason
3350 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3351 with non blocking I/O was not possible because no retry code was
3352 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3356 *) Added the beginnings of Rijndael support.
3359 *) Fix for bug in DirectoryString mask setting. Add support for
3360 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3361 to allow certificate printing to more controllable, additional
3362 'certopt' option to 'x509' to allow new printing options to be
3366 *) Clean old EAY MD5 hack from e_os.h.
3369 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3371 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3372 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3373 [Joe Orton, Steve Henson]
3375 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3377 *) Fix additional bug revealed by the NISCC test suite:
3379 Stop bug triggering large recursion when presented with
3380 certain ASN.1 tags (CVE-2003-0851)
3383 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3385 *) Fix various bugs revealed by running the NISCC test suite:
3387 Stop out of bounds reads in the ASN1 code when presented with
3388 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3390 If verify callback ignores invalid public key errors don't try to check
3391 certificate signature with the NULL public key.
3395 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3396 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3400 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3401 extra data after the compression methods not only for TLS 1.0
3402 but also for SSL 3.0 (as required by the specification).
3403 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3405 *) Change X509_certificate_type() to mark the key as exported/exportable
3406 when it's 512 *bits* long, not 512 bytes.
3409 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3411 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3412 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3413 a protocol version number mismatch like a decryption error
3414 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3417 *) Turn on RSA blinding by default in the default implementation
3418 to avoid a timing attack. Applications that don't want it can call
3419 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3420 They would be ill-advised to do so in most cases.
3421 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3423 *) Change RSA blinding code so that it works when the PRNG is not
3424 seeded (in this case, the secret RSA exponent is abused as
3425 an unpredictable seed -- if it is not unpredictable, there
3426 is no point in blinding anyway). Make RSA blinding thread-safe
3427 by remembering the creator's thread ID in rsa->blinding and
3428 having all other threads use local one-time blinding factors
3429 (this requires more computation than sharing rsa->blinding, but
3430 avoids excessive locking; and if an RSA object is not shared
3431 between threads, blinding will still be very fast).
3434 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3436 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3437 via timing by performing a MAC computation even if incorrrect
3438 block cipher padding has been found. This is a countermeasure
3439 against active attacks where the attacker has to distinguish
3440 between bad padding and a MAC verification error. (CVE-2003-0078)
3442 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3443 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3444 Martin Vuagnoux (EPFL, Ilion)]
3446 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3448 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3449 memory from it's contents. This is done with a counter that will
3450 place alternating values in each byte. This can be used to solve
3451 two issues: 1) the removal of calls to memset() by highly optimizing
3452 compilers, and 2) cleansing with other values than 0, since those can
3453 be read through on certain media, for example a swap space on disk.
3456 *) Bugfix: client side session caching did not work with external caching,
3457 because the session->cipher setting was not restored when reloading
3458 from the external cache. This problem was masked, when
3459 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3460 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3463 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3464 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3465 [Zeev Lieber <zeev-l@yahoo.com>]
3467 *) Undo an undocumented change introduced in 0.9.6e which caused
3468 repeated calls to OpenSSL_add_all_ciphers() and
3469 OpenSSL_add_all_digests() to be ignored, even after calling
3473 *) Change the default configuration reader to deal with last line not
3474 being properly terminated.
3477 *) Change X509_NAME_cmp() so it applies the special rules on handling
3478 DN values that are of type PrintableString, as well as RDNs of type
3479 emailAddress where the value has the type ia5String.
3480 [stefank@valicert.com via Richard Levitte]
3482 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3483 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3484 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3485 the bitwise-OR of the two for use by the majority of applications
3486 wanting this behaviour, and update the docs. The documented
3487 behaviour and actual behaviour were inconsistent and had been
3488 changing anyway, so this is more a bug-fix than a behavioural
3490 [Geoff Thorpe, diagnosed by Nadav Har'El]
3492 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3493 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3496 *) Fix initialization code race conditions in
3497 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3498 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3499 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3500 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3501 ssl2_get_cipher_by_char(),
3502 ssl3_get_cipher_by_char().
3503 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3505 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3506 the cached sessions are flushed, as the remove_cb() might use ex_data
3507 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3508 (see [openssl.org #212]).
3509 [Geoff Thorpe, Lutz Jaenicke]
3511 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3512 length, instead of the encoding length to d2i_ASN1_OBJECT.
3515 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3517 *) [In 0.9.6g-engine release:]
3518 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3519 [Lynn Gazis <lgazis@rainbow.com>]
3521 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3523 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3524 and get fix the header length calculation.
3525 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3526 Alon Kantor <alonk@checkpoint.com> (and others),
3529 *) Use proper error handling instead of 'assertions' in buffer
3530 overflow checks added in 0.9.6e. This prevents DoS (the
3531 assertions could call abort()).
3532 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3534 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3536 *) Add various sanity checks to asn1_get_length() to reject
3537 the ASN1 length bytes if they exceed sizeof(long), will appear
3538 negative or the content length exceeds the length of the
3540 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3542 *) Fix cipher selection routines: ciphers without encryption had no flags
3543 for the cipher strength set and where therefore not handled correctly
3544 by the selection routines (PR #130).
3547 *) Fix EVP_dsa_sha macro.
3551 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3552 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3553 that was added in OpenSSL 0.9.6d.
3555 As the countermeasure turned out to be incompatible with some
3556 broken SSL implementations, the new option is part of SSL_OP_ALL.
3557 SSL_OP_ALL is usually employed when compatibility with weird SSL
3558 implementations is desired (e.g. '-bugs' option to 's_client' and
3559 's_server'), so the new option is automatically set in many
3563 *) Changes in security patch:
3565 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3566 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3567 Air Force Materiel Command, USAF, under agreement number
3570 *) Add various sanity checks to asn1_get_length() to reject
3571 the ASN1 length bytes if they exceed sizeof(long), will appear
3572 negative or the content length exceeds the length of the
3573 supplied buffer. (CVE-2002-0659)
3574 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3576 *) Assertions for various potential buffer overflows, not known to
3578 [Ben Laurie (CHATS)]
3580 *) Various temporary buffers to hold ASCII versions of integers were
3581 too small for 64 bit platforms. (CVE-2002-0655)
3582 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3584 *) Remote buffer overflow in SSL3 protocol - an attacker could
3585 supply an oversized session ID to a client. (CVE-2002-0656)
3586 [Ben Laurie (CHATS)]
3588 *) Remote buffer overflow in SSL2 protocol - an attacker could
3589 supply an oversized client master key. (CVE-2002-0656)
3590 [Ben Laurie (CHATS)]
3592 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3594 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3595 encoded as NULL) with id-dsa-with-sha1.
3596 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3598 *) Check various X509_...() return values in apps/req.c.
3599 [Nils Larsch <nla@trustcenter.de>]
3601 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3602 an end-of-file condition would erronously be flagged, when the CRLF
3603 was just at the end of a processed block. The bug was discovered when
3604 processing data through a buffering memory BIO handing the data to a
3605 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3606 <ptsekov@syntrex.com> and Nedelcho Stanev.
3609 *) Implement a countermeasure against a vulnerability recently found
3610 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3611 before application data chunks to avoid the use of known IVs
3612 with data potentially chosen by the attacker.
3615 *) Fix length checks in ssl3_get_client_hello().
3618 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3619 to prevent ssl3_read_internal() from incorrectly assuming that
3620 ssl3_read_bytes() found application data while handshake
3621 processing was enabled when in fact s->s3->in_read_app_data was
3622 merely automatically cleared during the initial handshake.
3623 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3625 *) Fix object definitions for Private and Enterprise: they were not
3626 recognized in their shortname (=lowercase) representation. Extend
3627 obj_dat.pl to issue an error when using undefined keywords instead
3628 of silently ignoring the problem (Svenning Sorensen
3629 <sss@sss.dnsalias.net>).
3632 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3633 generators, i.e. generators other than 2 and 5. (Previously, the
3634 code did not properly initialise the 'add' and 'rem' values to
3635 BN_generate_prime().)
3637 In the new general case, we do not insist that 'generator' is
3638 actually a primitive root: This requirement is rather pointless;
3639 a generator of the order-q subgroup is just as good, if not
3643 *) Map new X509 verification errors to alerts. Discovered and submitted by
3644 Tom Wu <tom@arcot.com>.
3647 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3648 returning non-zero before the data has been completely received
3649 when using non-blocking I/O.
3650 [Bodo Moeller; problem pointed out by John Hughes]
3652 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3653 [Ben Laurie, Lutz Jaenicke]
3655 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3656 Yoram Zahavi <YoramZ@gilian.com>).
3659 *) Add information about CygWin 1.3 and on, and preserve proper
3660 configuration for the versions before that.
3661 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3663 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3664 check whether we deal with a copy of a session and do not delete from
3665 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3666 <izhar@checkpoint.com>.
3669 *) Do not store session data into the internal session cache, if it
3670 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3671 flag is set). Proposed by Aslam <aslam@funk.com>.
3674 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3678 *) [In 0.9.6d-engine release:]
3679 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3680 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3682 *) Add the configuration target linux-s390x.
3683 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3685 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3686 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3687 variable as an indication that a ClientHello message has been
3688 received. As the flag value will be lost between multiple
3689 invocations of ssl3_accept when using non-blocking I/O, the
3690 function may not be aware that a handshake has actually taken
3691 place, thus preventing a new session from being added to the
3694 To avoid this problem, we now set s->new_session to 2 instead of
3695 using a local variable.
3696 [Lutz Jaenicke, Bodo Moeller]
3698 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3699 if the SSL_R_LENGTH_MISMATCH error is detected.
3700 [Geoff Thorpe, Bodo Moeller]
3702 *) New 'shared_ldflag' column in Configure platform table.
3705 *) Fix EVP_CIPHER_mode macro.
3706 ["Dan S. Camper" <dan@bti.net>]
3708 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3709 type, we must throw them away by setting rr->length to 0.
3710 [D P Chang <dpc@qualys.com>]
3712 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3714 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3715 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3716 worked incorrectly for those cases where range = 10..._2 and
3717 3*range is two bits longer than range.)
3720 *) Only add signing time to PKCS7 structures if it is not already
3724 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3725 OBJ_ld_ce should be OBJ_id_ce.
3726 Also some ip-pda OIDs in crypto/objects/objects.txt were
3727 incorrect (cf. RFC 3039).
3728 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3730 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3731 returns early because it has nothing to do.
3732 [Andy Schneider <andy.schneider@bjss.co.uk>]
3734 *) [In 0.9.6c-engine release:]
3735 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3736 [Andy Schneider <andy.schneider@bjss.co.uk>]
3738 *) [In 0.9.6c-engine release:]
3739 Add support for Cryptographic Appliance's keyserver technology.
3740 (Use engine 'keyclient')
3741 [Cryptographic Appliances and Geoff Thorpe]
3743 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3744 is called via tools/c89.sh because arguments have to be
3745 rearranged (all '-L' options must appear before the first object
3747 [Richard Shapiro <rshapiro@abinitio.com>]
3749 *) [In 0.9.6c-engine release:]
3750 Add support for Broadcom crypto accelerator cards, backported
3752 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3754 *) [In 0.9.6c-engine release:]
3755 Add support for SureWare crypto accelerator cards from
3756 Baltimore Technologies. (Use engine 'sureware')
3757 [Baltimore Technologies and Mark Cox]
3759 *) [In 0.9.6c-engine release:]
3760 Add support for crypto accelerator cards from Accelerated
3761 Encryption Processing, www.aep.ie. (Use engine 'aep')
3762 [AEP Inc. and Mark Cox]
3764 *) Add a configuration entry for gcc on UnixWare.
3765 [Gary Benson <gbenson@redhat.com>]
3767 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3768 messages are stored in a single piece (fixed-length part and
3769 variable-length part combined) and fix various bugs found on the way.
3772 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3773 instead. BIO_gethostbyname() does not know what timeouts are
3774 appropriate, so entries would stay in cache even when they have
3776 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3778 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3779 faced with a pathologically small ClientHello fragment that does
3780 not contain client_version: Instead of aborting with an error,
3781 simply choose the highest available protocol version (i.e.,
3782 TLS 1.0 unless it is disabled). In practice, ClientHello
3783 messages are never sent like this, but this change gives us
3784 strictly correct behaviour at least for TLS.
3787 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3788 never resets s->method to s->ctx->method when called from within
3789 one of the SSL handshake functions.
3790 [Bodo Moeller; problem pointed out by Niko Baric]
3792 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3793 (sent using the client's version number) if client_version is
3794 smaller than the protocol version in use. Also change
3795 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3796 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3797 the client will at least see that alert.
3800 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3804 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3805 client receives HelloRequest while in a handshake.
3806 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3808 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3809 should end in 'break', not 'goto end' which circuments various
3810 cleanups done in state SSL_ST_OK. But session related stuff
3811 must be disabled for SSL_ST_OK in the case that we just sent a
3814 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3815 before just sending a HelloRequest.
3816 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3818 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3819 reveal whether illegal block cipher padding was found or a MAC
3820 verification error occured. (Neither SSLerr() codes nor alerts
3821 are directly visible to potential attackers, but the information
3822 may leak via logfiles.)
3824 Similar changes are not required for the SSL 2.0 implementation
3825 because the number of padding bytes is sent in clear for SSL 2.0,
3826 and the extra bytes are just ignored. However ssl/s2_pkt.c
3827 failed to verify that the purported number of padding bytes is in
3831 *) Add OpenUNIX-8 support including shared libraries
3832 (Boyd Lynn Gerber <gerberb@zenez.com>).
3835 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3836 'wristwatch attack' using huge encoding parameters (cf.
3837 James H. Manger's CRYPTO 2001 paper). Note that the
3838 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3839 encoding parameters and hence was not vulnerable.
3842 *) BN_sqr() bug fix.
3843 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3845 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3846 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3847 followed by modular reduction.
3848 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3850 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3851 equivalent based on BN_pseudo_rand() instead of BN_rand().
3854 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3855 This function was broken, as the check for a new client hello message
3856 to handle SGC did not allow these large messages.
3857 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3860 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3863 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3864 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3867 *) Rework the configuration and shared library support for Tru64 Unix.
3868 The configuration part makes use of modern compiler features and
3869 still retains old compiler behavior for those that run older versions
3870 of the OS. The shared library support part includes a variant that
3871 uses the RPATH feature, and is available through the special
3872 configuration target "alpha-cc-rpath", which will never be selected
3874 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3876 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3877 with the same message size as in ssl3_get_certificate_request().
3878 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3879 messages might inadvertently be reject as too long.
3880 [Petr Lampa <lampa@fee.vutbr.cz>]
3882 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3885 *) Modified SSL library such that the verify_callback that has been set
3886 specificly for an SSL object with SSL_set_verify() is actually being
3887 used. Before the change, a verify_callback set with this function was
3888 ignored and the verify_callback() set in the SSL_CTX at the time of
3889 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3890 to allow the necessary settings.
3893 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3894 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3895 done automatically (in contradiction to the requirements of the C
3896 standard). This made problems when used from OpenSSH.
3899 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3900 dh->length and always used
3902 BN_rand_range(priv_key, dh->p).
3904 BN_rand_range() is not necessary for Diffie-Hellman, and this
3905 specific range makes Diffie-Hellman unnecessarily inefficient if
3906 dh->length (recommended exponent length) is much smaller than the
3907 length of dh->p. We could use BN_rand_range() if the order of
3908 the subgroup was stored in the DH structure, but we only have
3913 BN_rand(priv_key, l, ...)
3915 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3921 RSA_eay_public_encrypt
3922 RSA_eay_private_decrypt
3923 RSA_eay_private_encrypt (signing)
3924 RSA_eay_public_decrypt (signature verification)
3926 (default implementations for RSA_public_encrypt,
3927 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3928 always reject numbers >= n.
3931 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3932 to synchronize access to 'locking_thread'. This is necessary on
3933 systems where access to 'locking_thread' (an 'unsigned long'
3934 variable) is not atomic.
3937 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3938 *before* setting the 'crypto_lock_rand' flag. The previous code had
3939 a race condition if 0 is a valid thread ID.
3940 [Travis Vitek <vitek@roguewave.com>]
3942 *) Add support for shared libraries under Irix.
3943 [Albert Chin-A-Young <china@thewrittenword.com>]
3945 *) Add configuration option to build on Linux on both big-endian and
3947 [Ralf Baechle <ralf@uni-koblenz.de>]
3949 *) Add the possibility to create shared libraries on HP-UX.
3952 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3954 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3955 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3956 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3957 PRNG state recovery was possible based on the output of
3958 one PRNG request appropriately sized to gain knowledge on
3959 'md' followed by enough consecutive 1-byte PRNG requests
3960 to traverse all of 'state'.
3962 1. When updating 'md_local' (the current thread's copy of 'md')
3963 during PRNG output generation, hash all of the previous
3964 'md_local' value, not just the half used for PRNG output.
3966 2. Make the number of bytes from 'state' included into the hash
3967 independent from the number of PRNG bytes requested.
3969 The first measure alone would be sufficient to avoid
3970 Markku-Juhani's attack. (Actually it had never occurred
3971 to me that the half of 'md_local' used for chaining was the
3972 half from which PRNG output bytes were taken -- I had always
3973 assumed that the secret half would be used.) The second
3974 measure makes sure that additional data from 'state' is never
3975 mixed into 'md_local' in small portions; this heuristically
3976 further strengthens the PRNG.
3979 *) Fix crypto/bn/asm/mips3.s.
3982 *) When only the key is given to "enc", the IV is undefined. Print out
3983 an error message in this case.
3986 *) Handle special case when X509_NAME is empty in X509 printing routines.
3989 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3990 positive and less than q.
3993 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3994 used: it isn't thread safe and the add_lock_callback should handle
3996 [Paul Rose <Paul.Rose@bridge.com>]
3998 *) Verify that incoming data obeys the block size in
3999 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4003 [Ulf Möller, Bodo Möller]
4005 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4006 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4007 when fixing the server behaviour for backwards-compatible 'client
4008 hello' messages. (Note that the attack is impractical against
4009 SSL 3.0 and TLS 1.0 anyway because length and version checking
4010 means that the probability of guessing a valid ciphertext is
4011 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4014 Before 0.9.5, the countermeasure (hide the error by generating a
4015 random 'decryption result') did not work properly because
4016 ERR_clear_error() was missing, meaning that SSL_get_error() would
4017 detect the supposedly ignored error.
4019 Both problems are now fixed.
4022 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4023 (previously it was 1024).
4026 *) Fix for compatibility mode trust settings: ignore trust settings
4027 unless some valid trust or reject settings are present.
4030 *) Fix for blowfish EVP: its a variable length cipher.
4033 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4034 parameters in DSA public key structures and return an error in the
4035 DSA routines if parameters are absent.
4038 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4039 in the current directory if neither $RANDFILE nor $HOME was set.
4040 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4041 caused some confusion to Windows users who haven't defined $HOME.
4042 Thus RAND_file_name() is changed again: e_os.h can define a
4043 DEFAULT_HOME, which will be used if $HOME is not set.
4044 For Windows, we use "C:"; on other platforms, we still require
4045 environment variables.
4047 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4048 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4049 having multiple threads call RAND_poll() concurrently.
4052 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4053 combination of a flag and a thread ID variable.
4054 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4055 flag), *other* threads can enter ssleay_add_bytes without obeying
4056 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4057 that they do not hold after the first thread unsets add_do_not_lock).
4060 *) Change bctest again: '-x' expressions are not available in all
4064 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4066 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4067 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4069 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4070 the default extension for executables, if any. Also, make the perl
4071 scripts that use symlink() to test if it really exists and use "cp"
4072 if it doesn't. All this made OpenSSL compilable and installable in
4076 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4077 If SEQUENCE is length is indefinite just set c->slen to the total
4078 amount of data available.
4079 [Steve Henson, reported by shige@FreeBSD.org]
4080 [This change does not apply to 0.9.7.]
4082 *) Change bctest to avoid here-documents inside command substitution
4083 (workaround for FreeBSD /bin/sh bug).
4084 For compatibility with Ultrix, avoid shell functions (introduced
4085 in the bctest version that searches along $PATH).
4088 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4089 with des_encrypt() defined on some operating systems, like Solaris
4093 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4094 On the Importance of Eliminating Errors in Cryptographic
4095 Computations, J. Cryptology 14 (2001) 2, 101-119,
4096 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4099 *) MIPS assembler BIGNUM division bug fix.
4102 *) Disabled incorrect Alpha assembler code.
4105 *) Fix PKCS#7 decode routines so they correctly update the length
4106 after reading an EOC for the EXPLICIT tag.
4108 [This change does not apply to 0.9.7.]
4110 *) Fix bug in PKCS#12 key generation routines. This was triggered
4111 if a 3DES key was generated with a 0 initial byte. Include
4112 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4113 (but broken) behaviour.
4116 *) Enhance bctest to search for a working bc along $PATH and print
4118 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4120 *) Fix memory leaks in err.c: free err_data string if necessary;
4121 don't write to the wrong index in ERR_set_error_data.
4124 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4128 *) Replace rdtsc with _emit statements for VC++ version 5.
4129 [Jeremy Cooper <jeremy@baymoo.org>]
4131 *) Make it possible to reuse SSLv2 sessions.
4134 *) In copy_email() check for >= 0 as a return value for
4135 X509_NAME_get_index_by_NID() since 0 is a valid index.
4136 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4138 *) Avoid coredump with unsupported or invalid public keys by checking if
4139 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4140 PKCS7_verify() fails with non detached data.
4143 *) Don't use getenv in library functions when run as setuid/setgid.
4144 New function OPENSSL_issetugid().
4147 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4148 due to incorrect handling of multi-threading:
4150 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4152 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4154 3. Count how many times MemCheck_off() has been called so that
4155 nested use can be treated correctly. This also avoids
4156 inband-signalling in the previous code (which relied on the
4157 assumption that thread ID 0 is impossible).
4160 *) Add "-rand" option also to s_client and s_server.
4163 *) Fix CPU detection on Irix 6.x.
4164 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4165 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4167 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4170 [This change does not apply to 0.9.7.]
4172 *) Use the cached encoding of an X509_NAME structure rather than
4173 copying it. This is apparently the reason for the libsafe "errors"
4174 but the code is actually correct.
4177 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4178 Bleichenbacher's DSA attack.
4179 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4180 to be set and top=0 forces the highest bit to be set; top=-1 is new
4181 and leaves the highest bit random.
4182 [Ulf Moeller, Bodo Moeller]
4184 *) In the NCONF_...-based implementations for CONF_... queries
4185 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4186 a temporary CONF structure with the data component set to NULL
4187 (which gives segmentation faults in lh_retrieve).
4188 Instead, use NULL for the CONF pointer in CONF_get_string and
4189 CONF_get_number (which may use environment variables) and directly
4190 return NULL from CONF_get_section.
4193 *) Fix potential buffer overrun for EBCDIC.
4196 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4197 keyUsage if basicConstraints absent for a CA.
4200 *) Make SMIME_write_PKCS7() write mail header values with a format that
4201 is more generally accepted (no spaces before the semicolon), since
4202 some programs can't parse those values properly otherwise. Also make
4203 sure BIO's that break lines after each write do not create invalid
4207 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4208 macros previously used would not encode an empty SEQUENCE OF
4209 and break the signature.
4211 [This change does not apply to 0.9.7.]
4213 *) Zero the premaster secret after deriving the master secret in
4217 *) Add some EVP_add_digest_alias registrations (as found in
4218 OpenSSL_add_all_digests()) to SSL_library_init()
4219 aka OpenSSL_add_ssl_algorithms(). This provides improved
4220 compatibility with peers using X.509 certificates
4221 with unconventional AlgorithmIdentifier OIDs.
4224 *) Fix for Irix with NO_ASM.
4225 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4227 *) ./config script fixes.
4228 [Ulf Moeller, Richard Levitte]
4230 *) Fix 'openssl passwd -1'.
4233 *) Change PKCS12_key_gen_asc() so it can cope with non null
4234 terminated strings whose length is passed in the passlen
4235 parameter, for example from PEM callbacks. This was done
4236 by adding an extra length parameter to asc2uni().
4237 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4239 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4240 call failed, free the DSA structure.
4243 *) Fix to uni2asc() to cope with zero length Unicode strings.
4244 These are present in some PKCS#12 files.
4247 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4248 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4249 when writing a 32767 byte record.
4250 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4252 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4253 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4255 (RSA objects have a reference count access to which is protected
4256 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4257 so they are meant to be shared between threads.)
4258 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4259 "Reddie, Steven" <Steven.Reddie@ca.com>]
4261 *) Fix a deadlock in CRYPTO_mem_leaks().
4264 *) Use better test patterns in bntest.
4267 *) rand_win.c fix for Borland C.
4270 *) BN_rshift bugfix for n == 0.
4273 *) Add a 'bctest' script that checks for some known 'bc' bugs
4274 so that 'make test' does not abort just because 'bc' is broken.
4277 *) Store verify_result within SSL_SESSION also for client side to
4278 avoid potential security hole. (Re-used sessions on the client side
4279 always resulted in verify_result==X509_V_OK, not using the original
4280 result of the server certificate verification.)
4283 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4284 SSL3_RT_APPLICATION_DATA, return 0.
4285 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4289 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4290 releases, have been re-implemented by renaming the previous
4291 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4292 and ssl3_read_internal, respectively, and adding 'peek' parameters
4293 to them. The new ssl[23]_{read,peek} functions are calls to
4294 ssl[23]_read_internal with the 'peek' flag set appropriately.
4295 A 'peek' parameter has also been added to ssl3_read_bytes, which
4296 does the actual work for ssl3_read_internal.
4299 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4300 the method-specific "init()" handler. Also clean up ex_data after
4301 calling the method-specific "finish()" handler. Previously, this was
4302 happening the other way round.
4305 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4306 The previous value, 12, was not always sufficient for BN_mod_exp().
4309 *) Make sure that shared libraries get the internal name engine with
4310 the full version number and not just 0. This should mark the
4311 shared libraries as not backward compatible. Of course, this should
4312 be changed again when we can guarantee backward binary compatibility.
4315 *) Fix typo in get_cert_by_subject() in by_dir.c
4316 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4318 *) Rework the system to generate shared libraries:
4320 - Make note of the expected extension for the shared libraries and
4321 if there is a need for symbolic links from for example libcrypto.so.0
4322 to libcrypto.so.0.9.7. There is extended info in Configure for
4325 - Make as few rebuilds of the shared libraries as possible.
4327 - Still avoid linking the OpenSSL programs with the shared libraries.
4329 - When installing, install the shared libraries separately from the
4333 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4335 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4336 and not in SSL_clear because the latter is also used by the
4337 accept/connect functions; previously, the settings made by
4338 SSL_set_read_ahead would be lost during the handshake.
4339 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4341 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4342 Previously, it would create entries for disableed algorithms no
4346 *) Added several new manual pages for SSL_* function.
4349 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4351 *) In ssl23_get_client_hello, generate an error message when faced
4352 with an initial SSL 3.0/TLS record that is too small to contain the
4353 first two bytes of the ClientHello message, i.e. client_version.
4354 (Note that this is a pathologic case that probably has never happened
4355 in real life.) The previous approach was to use the version number
4356 from the record header as a substitute; but our protocol choice
4357 should not depend on that one because it is not authenticated
4358 by the Finished messages.
4361 *) More robust randomness gathering functions for Windows.
4362 [Jeffrey Altman <jaltman@columbia.edu>]
4364 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4365 not set then we don't setup the error code for issuer check errors
4366 to avoid possibly overwriting other errors which the callback does
4367 handle. If an application does set the flag then we assume it knows
4368 what it is doing and can handle the new informational codes
4372 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4373 a general "ANY" type, as such it should be able to decode anything
4374 including tagged types. However it didn't check the class so it would
4375 wrongly interpret tagged types in the same way as their universal
4376 counterpart and unknown types were just rejected. Changed so that the
4377 tagged and unknown types are handled in the same way as a SEQUENCE:
4378 that is the encoding is stored intact. There is also a new type
4379 "V_ASN1_OTHER" which is used when the class is not universal, in this
4380 case we have no idea what the actual type is so we just lump them all
4384 *) On VMS, stdout may very well lead to a file that is written to
4385 in a record-oriented fashion. That means that every write() will
4386 write a separate record, which will be read separately by the
4387 programs trying to read from it. This can be very confusing.
4389 The solution is to put a BIO filter in the way that will buffer
4390 text until a linefeed is reached, and then write everything a
4391 line at a time, so every record written will be an actual line,
4392 not chunks of lines and not (usually doesn't happen, but I've
4393 seen it once) several lines in one record. BIO_f_linebuffer() is
4396 Currently, it's a VMS-only method, because that's where it has
4397 been tested well enough.
4400 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4401 it can return incorrect results.
4402 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4403 but it was in 0.9.6-beta[12].)
4406 *) Disable the check for content being present when verifying detached
4407 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4408 include zero length content when signing messages.
4411 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4412 BIO_ctrl (for BIO pairs).
4415 *) Add DSO method for VMS.
4418 *) Bug fix: Montgomery multiplication could produce results with the
4422 *) Add RPM specification openssl.spec and modify it to build three
4423 packages. The default package contains applications, application
4424 documentation and run-time libraries. The devel package contains
4425 include files, static libraries and function documentation. The
4426 doc package contains the contents of the doc directory. The original
4427 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4430 *) Add a large number of documentation files for many SSL routines.
4431 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4433 *) Add a configuration entry for Sony News 4.
4434 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4436 *) Don't set the two most significant bits to one when generating a
4437 random number < q in the DSA library.
4440 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4441 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4442 the underlying transport is blocking) if a handshake took place.
4443 (The default behaviour is needed by applications such as s_client
4444 and s_server that use select() to determine when to use SSL_read;
4445 but for applications that know in advance when to expect data, it
4446 just makes things more complicated.)
4449 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4453 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4454 work better on such systems.
4455 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4457 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4458 Update PKCS12_parse() so it copies the friendlyName and the
4459 keyid to the certificates aux info.
4462 *) Fix bug in PKCS7_verify() which caused an infinite loop
4463 if there was more than one signature.
4464 [Sven Uszpelkat <su@celocom.de>]
4466 *) Major change in util/mkdef.pl to include extra information
4467 about each symbol, as well as presentig variables as well
4468 as functions. This change means that there's n more need
4469 to rebuild the .num files when some algorithms are excluded.
4472 *) Allow the verify time to be set by an application,
4473 rather than always using the current time.
4476 *) Phase 2 verify code reorganisation. The certificate
4477 verify code now looks up an issuer certificate by a
4478 number of criteria: subject name, authority key id
4479 and key usage. It also verifies self signed certificates
4480 by the same criteria. The main comparison function is
4481 X509_check_issued() which performs these checks.
4483 Lot of changes were necessary in order to support this
4484 without completely rewriting the lookup code.
4486 Authority and subject key identifier are now cached.
4488 The LHASH 'certs' is X509_STORE has now been replaced
4489 by a STACK_OF(X509_OBJECT). This is mainly because an
4490 LHASH can't store or retrieve multiple objects with
4491 the same hash value.
4493 As a result various functions (which were all internal
4494 use only) have changed to handle the new X509_STORE
4495 structure. This will break anything that messed round
4496 with X509_STORE internally.
4498 The functions X509_STORE_add_cert() now checks for an
4499 exact match, rather than just subject name.
4501 The X509_STORE API doesn't directly support the retrieval
4502 of multiple certificates matching a given criteria, however
4503 this can be worked round by performing a lookup first
4504 (which will fill the cache with candidate certificates)
4505 and then examining the cache for matches. This is probably
4506 the best we can do without throwing out X509_LOOKUP
4507 entirely (maybe later...).
4509 The X509_VERIFY_CTX structure has been enhanced considerably.
4511 All certificate lookup operations now go via a get_issuer()
4512 callback. Although this currently uses an X509_STORE it
4513 can be replaced by custom lookups. This is a simple way
4514 to bypass the X509_STORE hackery necessary to make this
4515 work and makes it possible to use more efficient techniques
4516 in future. A very simple version which uses a simple
4517 STACK for its trusted certificate store is also provided
4518 using X509_STORE_CTX_trusted_stack().
4520 The verify_cb() and verify() callbacks now have equivalents
4521 in the X509_STORE_CTX structure.
4523 X509_STORE_CTX also has a 'flags' field which can be used
4524 to customise the verify behaviour.
4527 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4528 excludes S/MIME capabilities.
4531 *) When a certificate request is read in keep a copy of the
4532 original encoding of the signed data and use it when outputing
4533 again. Signatures then use the original encoding rather than
4534 a decoded, encoded version which may cause problems if the
4535 request is improperly encoded.
4538 *) For consistency with other BIO_puts implementations, call
4539 buffer_write(b, ...) directly in buffer_puts instead of calling
4542 In BIO_puts, increment b->num_write as in BIO_write.
4543 [Peter.Sylvester@EdelWeb.fr]
4545 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4546 BN_zero, we may not return a BIGNUM with an array consisting of
4550 *) Avoid calling abort() from within the library when problems are
4551 detected, except if preprocessor symbols have been defined
4552 (such as REF_CHECK, BN_DEBUG etc.).
4555 *) New openssl application 'rsautl'. This utility can be
4556 used for low level RSA operations. DER public key
4557 BIO/fp routines also added.
4560 *) New Configure entry and patches for compiling on QNX 4.
4561 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4563 *) A demo state-machine implementation was sponsored by
4564 Nuron (http://www.nuron.com/) and is now available in
4565 demos/state_machine.
4568 *) New options added to the 'dgst' utility for signature
4569 generation and verification.
4572 *) Unrecognized PKCS#7 content types are now handled via a
4573 catch all ASN1_TYPE structure. This allows unsupported
4574 types to be stored as a "blob" and an application can
4575 encode and decode it manually.
4578 *) Fix various signed/unsigned issues to make a_strex.c
4580 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4582 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4583 length if passed a buffer. ASN1_INTEGER_to_BN failed
4584 if passed a NULL BN and its argument was negative.
4585 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4587 *) Modification to PKCS#7 encoding routines to output definite
4588 length encoding. Since currently the whole structures are in
4589 memory there's not real point in using indefinite length
4590 constructed encoding. However if OpenSSL is compiled with
4591 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4594 *) Added BIO_vprintf() and BIO_vsnprintf().
4597 *) Added more prefixes to parse for in the the strings written
4598 through a logging bio, to cover all the levels that are available
4599 through syslog. The prefixes are now:
4601 PANIC, EMERG, EMR => LOG_EMERG
4602 ALERT, ALR => LOG_ALERT
4603 CRIT, CRI => LOG_CRIT
4604 ERROR, ERR => LOG_ERR
4605 WARNING, WARN, WAR => LOG_WARNING
4606 NOTICE, NOTE, NOT => LOG_NOTICE
4607 INFO, INF => LOG_INFO
4608 DEBUG, DBG => LOG_DEBUG
4610 and as before, if none of those prefixes are present at the
4611 beginning of the string, LOG_ERR is chosen.
4613 On Win32, the LOG_* levels are mapped according to this:
4615 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4616 LOG_WARNING => EVENTLOG_WARNING_TYPE
4617 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4621 *) Made it possible to reconfigure with just the configuration
4622 argument "reconf" or "reconfigure". The command line arguments
4623 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4624 and are retrieved from there when reconfiguring.
4628 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4630 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4633 *) The obj_dat.pl script was messing up the sorting of object
4634 names. The reason was that it compared the quoted version
4635 of strings as a result "OCSP" > "OCSP Signing" because
4636 " > SPACE. Changed script to store unquoted versions of
4637 names and add quotes on output. It was also omitting some
4638 names from the lookup table if they were given a default
4639 value (that is if SN is missing it is given the same
4640 value as LN and vice versa), these are now added on the
4641 grounds that if an object has a name we should be able to
4642 look it up. Finally added warning output when duplicate
4643 short or long names are found.
4646 *) Changes needed for Tandem NSK.
4647 [Scott Uroff <scott@xypro.com>]
4649 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4650 RSA_padding_check_SSLv23(), special padding was never detected
4651 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4652 version rollback attacks was not effective.
4654 In s23_clnt.c, don't use special rollback-attack detection padding
4655 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4656 client; similarly, in s23_srvr.c, don't do the rollback check if
4657 SSL 2.0 is the only protocol enabled in the server.
4660 *) Make it possible to get hexdumps of unprintable data with 'openssl
4661 asn1parse'. By implication, the functions ASN1_parse_dump() and
4662 BIO_dump_indent() are added.
4665 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4666 these print out strings and name structures based on various
4667 flags including RFC2253 support and proper handling of
4668 multibyte characters. Added options to the 'x509' utility
4669 to allow the various flags to be set.
4672 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4673 Also change the functions X509_cmp_current_time() and
4674 X509_gmtime_adj() work with an ASN1_TIME structure,
4675 this will enable certificates using GeneralizedTime in validity
4676 dates to be checked.
4679 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4680 negative public key encodings) on by default,
4681 NO_NEG_PUBKEY_BUG can be set to disable it.
4684 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4685 content octets. An i2c_ASN1_OBJECT is unnecessary because
4686 the encoding can be trivially obtained from the structure.
4689 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4690 not read locks (CRYPTO_r_[un]lock).
4693 *) A first attempt at creating official support for shared
4694 libraries through configuration. I've kept it so the
4695 default is static libraries only, and the OpenSSL programs
4696 are always statically linked for now, but there are
4697 preparations for dynamic linking in place.
4698 This has been tested on Linux and Tru64.
4701 *) Randomness polling function for Win9x, as described in:
4702 Peter Gutmann, Software Generation of Practically Strong
4706 *) Fix so PRNG is seeded in req if using an already existing
4710 *) New options to smime application. -inform and -outform
4711 allow alternative formats for the S/MIME message including
4712 PEM and DER. The -content option allows the content to be
4713 specified separately. This should allow things like Netscape
4714 form signing output easier to verify.
4717 *) Fix the ASN1 encoding of tags using the 'long form'.
4720 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4721 STRING types. These convert content octets to and from the
4722 underlying type. The actual tag and length octets are
4723 already assumed to have been read in and checked. These
4724 are needed because all other string types have virtually
4725 identical handling apart from the tag. By having versions
4726 of the ASN1 functions that just operate on content octets
4727 IMPLICIT tagging can be handled properly. It also allows
4728 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4729 and ASN1_INTEGER are identical apart from the tag.
4732 *) Change the handling of OID objects as follows:
4734 - New object identifiers are inserted in objects.txt, following
4735 the syntax given in objects.README.
4736 - objects.pl is used to process obj_mac.num and create a new
4738 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4741 This is currently kind of a hack, and the perl code in objects.pl
4742 isn't very elegant, but it works as I intended. The simplest way
4743 to check that it worked correctly is to look in obj_dat.h and
4744 check the array nid_objs and make sure the objects haven't moved
4745 around (this is important!). Additions are OK, as well as
4746 consistent name changes.
4749 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4752 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4753 The given file adds to whatever has already been seeded into the
4754 random pool through the RANDFILE configuration file option or
4755 environment variable, or the default random state file.
4758 *) mkstack.pl now sorts each macro group into lexical order.
4759 Previously the output order depended on the order the files
4760 appeared in the directory, resulting in needless rewriting
4764 *) Patches to make OpenSSL compile under Win32 again. Mostly
4765 work arounds for the VC++ problem that it treats func() as
4766 func(void). Also stripped out the parts of mkdef.pl that
4767 added extra typesafe functions: these no longer exist.
4770 *) Reorganisation of the stack code. The macros are now all
4771 collected in safestack.h . Each macro is defined in terms of
4772 a "stack macro" of the form SKM_<name>(type, a, b). The
4773 DEBUG_SAFESTACK is now handled in terms of function casts,
4774 this has the advantage of retaining type safety without the
4775 use of additional functions. If DEBUG_SAFESTACK is not defined
4776 then the non typesafe macros are used instead. Also modified the
4777 mkstack.pl script to handle the new form. Needs testing to see
4778 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4779 the default if no major problems. Similar behaviour for ASN1_SET_OF
4780 and PKCS12_STACK_OF.
4783 *) When some versions of IIS use the 'NET' form of private key the
4784 key derivation algorithm is different. Normally MD5(password) is
4785 used as a 128 bit RC4 key. In the modified case
4786 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4787 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4788 as the old Netscape_RSA functions except they have an additional
4789 'sgckey' parameter which uses the modified algorithm. Also added
4790 an -sgckey command line option to the rsa utility. Thanks to
4791 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4792 algorithm to openssl-dev.
4795 *) The evp_local.h macros were using 'c.##kname' which resulted in
4796 invalid expansion on some systems (SCO 5.0.5 for example).
4797 Corrected to 'c.kname'.
4798 [Phillip Porch <root@theporch.com>]
4800 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4801 a STACK of email addresses from a certificate or request, these look
4802 in the subject name and the subject alternative name extensions and
4803 omit any duplicate addresses.
4806 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4807 This makes DSA verification about 2 % faster.
4810 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4811 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4812 plus overhead for 1024 bit moduli).
4813 This makes exponentiations about 0.5 % faster for 1024 bit
4814 exponents (as measured by "openssl speed rsa2048").
4817 *) Rename memory handling macros to avoid conflicts with other
4819 Malloc => OPENSSL_malloc
4820 Malloc_locked => OPENSSL_malloc_locked
4821 Realloc => OPENSSL_realloc
4822 Free => OPENSSL_free
4825 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4826 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4829 *) CygWin32 support.
4830 [John Jarvie <jjarvie@newsguy.com>]
4832 *) The type-safe stack code has been rejigged. It is now only compiled
4833 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4834 by default all type-specific stack functions are "#define"d back to
4835 standard stack functions. This results in more streamlined output
4836 but retains the type-safety checking possibilities of the original
4840 *) The STACK code has been cleaned up, and certain type declarations
4841 that didn't make a lot of sense have been brought in line. This has
4842 also involved a cleanup of sorts in safestack.h to more correctly
4843 map type-safe stack functions onto their plain stack counterparts.
4844 This work has also resulted in a variety of "const"ifications of
4845 lots of the code, especially "_cmp" operations which should normally
4846 be prototyped with "const" parameters anyway.
4849 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4850 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4851 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4852 where all of 'md' is used each time the PRNG is used, but 'state'
4853 is used only indexed by a cyclic counter. As entropy may not be
4854 well distributed from the beginning, 'md' is important as a
4855 chaining variable. However, the output function chains only half
4856 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4857 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4858 in all of 'state' being rewritten, with the new values depending
4859 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4862 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4863 the handshake is continued after ssl_verify_cert_chain();
4864 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4865 can lead to 'unexplainable' connection aborts later.
4866 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4868 *) Major EVP API cipher revision.
4869 Add hooks for extra EVP features. This allows various cipher
4870 parameters to be set in the EVP interface. Support added for variable
4871 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4872 setting of RC2 and RC5 parameters.
4874 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4877 Remove lots of duplicated code from the EVP library. For example *every*
4878 cipher init() function handles the 'iv' in the same way according to the
4879 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4880 for CFB and OFB modes they zero ctx->num.
4882 New functionality allows removal of S/MIME code RC2 hack.
4884 Most of the routines have the same form and so can be declared in terms
4887 By shifting this to the top level EVP_CipherInit() it can be removed from
4888 all individual ciphers. If the cipher wants to handle IVs or keys
4889 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4892 Change lots of functions like EVP_EncryptUpdate() to now return a
4893 value: although software versions of the algorithms cannot fail
4894 any installed hardware versions can.
4897 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4898 this option is set, tolerate broken clients that send the negotiated
4899 protocol version number instead of the requested protocol version
4903 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4904 i.e. non-zero for export ciphersuites, zero otherwise.
4905 Previous versions had this flag inverted, inconsistent with
4906 rsa_tmp_cb (..._TMP_RSA_CB).
4907 [Bodo Moeller; problem reported by Amit Chopra]
4909 *) Add missing DSA library text string. Work around for some IIS
4910 key files with invalid SEQUENCE encoding.
4913 *) Add a document (doc/standards.txt) that list all kinds of standards
4914 and so on that are implemented in OpenSSL.
4917 *) Enhance c_rehash script. Old version would mishandle certificates
4918 with the same subject name hash and wouldn't handle CRLs at all.
4919 Added -fingerprint option to crl utility, to support new c_rehash
4923 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4926 *) Fix for SSL server purpose checking. Server checking was
4927 rejecting certificates which had extended key usage present
4928 but no ssl client purpose.
4929 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4931 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4932 is a little unclear about how a blank password is handled.
4933 Since the password in encoded as a BMPString with terminating
4934 double NULL a zero length password would end up as just the
4935 double NULL. However no password at all is different and is
4936 handled differently in the PKCS#12 key generation code. NS
4937 treats a blank password as zero length. MSIE treats it as no
4938 password on export: but it will try both on import. We now do
4939 the same: PKCS12_parse() tries zero length and no password if
4940 the password is set to "" or NULL (NULL is now a valid password:
4941 it wasn't before) as does the pkcs12 application.
4944 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4945 perror when PEM_read_bio_X509_REQ fails, the error message must
4946 be obtained from the error queue.
4949 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4950 it in ERR_remove_state if appropriate, and change ERR_get_state
4951 accordingly to avoid race conditions (this is necessary because
4952 thread_hash is no longer constant once set).
4955 *) Bugfix for linux-elf makefile.one.
4958 *) RSA_get_default_method() will now cause a default
4959 RSA_METHOD to be chosen if one doesn't exist already.
4960 Previously this was only set during a call to RSA_new()
4961 or RSA_new_method(NULL) meaning it was possible for
4962 RSA_get_default_method() to return NULL.
4965 *) Added native name translation to the existing DSO code
4966 that will convert (if the flag to do so is set) filenames
4967 that are sufficiently small and have no path information
4968 into a canonical native form. Eg. "blah" converted to
4969 "libblah.so" or "blah.dll" etc.
4972 *) New function ERR_error_string_n(e, buf, len) which is like
4973 ERR_error_string(e, buf), but writes at most 'len' bytes
4974 including the 0 terminator. For ERR_error_string_n, 'buf'
4976 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4978 *) CONF library reworked to become more general. A new CONF
4979 configuration file reader "class" is implemented as well as a
4980 new functions (NCONF_*, for "New CONF") to handle it. The now
4981 old CONF_* functions are still there, but are reimplemented to
4982 work in terms of the new functions. Also, a set of functions
4983 to handle the internal storage of the configuration data is
4984 provided to make it easier to write new configuration file
4985 reader "classes" (I can definitely see something reading a
4986 configuration file in XML format, for example), called _CONF_*,
4987 or "the configuration storage API"...
4989 The new configuration file reading functions are:
4991 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4992 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4994 NCONF_default, NCONF_WIN32
4996 NCONF_dump_fp, NCONF_dump_bio
4998 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4999 NCONF_new creates a new CONF object. This works in the same way
5000 as other interfaces in OpenSSL, like the BIO interface.
5001 NCONF_dump_* dump the internal storage of the configuration file,
5002 which is useful for debugging. All other functions take the same
5003 arguments as the old CONF_* functions wth the exception of the
5004 first that must be a `CONF *' instead of a `LHASH *'.
5006 To make it easer to use the new classes with the old CONF_* functions,
5007 the function CONF_set_default_method is provided.
5010 *) Add '-tls1' option to 'openssl ciphers', which was already
5011 mentioned in the documentation but had not been implemented.
5012 (This option is not yet really useful because even the additional
5013 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5016 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5017 OpenSSL-based applications) load shared libraries and bind to
5018 them in a portable way.
5019 [Geoff Thorpe, with contributions from Richard Levitte]
5021 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5023 *) Make sure _lrotl and _lrotr are only used with MSVC.
5025 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5026 (the default implementation of RAND_status).
5028 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5029 to '-clrext' (= clear extensions), as intended and documented.
5030 [Bodo Moeller; inconsistency pointed out by Michael Attili
5033 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5034 was larger than the MD block size.
5035 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5037 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5038 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5039 using the passed key: if the passed key was a private key the result
5040 of X509_print(), for example, would be to print out all the private key
5044 *) des_quad_cksum() byte order bug fix.
5045 [Ulf Möller, using the problem description in krb4-0.9.7, where
5046 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5048 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5050 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5052 *) For easily testing in shell scripts whether some command
5053 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5054 returns with exit code 0 iff no command of the given name is available.
5055 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5056 the output goes to stdout and nothing is printed to stderr.
5057 Additional arguments are always ignored.
5059 Since for each cipher there is a command of the same name,
5060 the 'no-cipher' compilation switches can be tested this way.
5062 ('openssl no-XXX' is not able to detect pseudo-commands such
5063 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5066 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5069 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5070 is set; it will be thrown away anyway because each handshake creates
5072 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5073 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5074 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5075 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5078 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5079 'Q' and 'R' lose their special meanings (quit/renegotiate).
5080 This is part of what -quiet does; unlike -quiet, -ign_eof
5081 does not suppress any output.
5084 *) Add compatibility options to the purpose and trust code. The
5085 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5086 accepts a certificate or CA, this was the previous behaviour,
5087 with all the associated security issues.
5089 X509_TRUST_COMPAT is the old trust behaviour: only and
5090 automatically trust self signed roots in certificate store. A
5091 new trust setting X509_TRUST_DEFAULT is used to specify that
5092 a purpose has no associated trust setting and it should instead
5093 use the value in the default purpose.
5096 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5097 and fix a memory leak.
5100 *) In util/mkerr.pl (which implements 'make errors'), preserve
5101 reason strings from the previous version of the .c file, as
5102 the default to have only downcase letters (and digits) in
5103 automatically generated reasons codes is not always appropriate.
5106 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5107 using strerror. Previously, ERR_reason_error_string() returned
5108 library names as reason strings for SYSerr; but SYSerr is a special
5109 case where small numbers are errno values, not library numbers.
5112 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5113 converts DSA parameters into DH parameters. (When creating parameters,
5114 DSA_generate_parameters is used.)
5117 *) Include 'length' (recommended exponent length) in C code generated
5118 by 'openssl dhparam -C'.
5121 *) The second argument to set_label in perlasm was already being used
5122 so couldn't be used as a "file scope" flag. Moved to third argument
5126 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5127 instead of RAND_bytes for encryption IVs and salts.
5130 *) Include RAND_status() into RAND_METHOD instead of implementing
5131 it only for md_rand.c Otherwise replacing the PRNG by calling
5132 RAND_set_rand_method would be impossible.
5135 *) Don't let DSA_generate_key() enter an infinite loop if the random
5136 number generation fails.
5139 *) New 'rand' application for creating pseudo-random output.
5142 *) Added configuration support for Linux/IA64
5143 [Rolf Haberrecker <rolf@suse.de>]
5145 *) Assembler module support for Mingw32.
5148 *) Shared library support for HPUX (in shlib/).
5149 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5151 *) Shared library support for Solaris gcc.
5152 [Lutz Behnke <behnke@trustcenter.de>]
5154 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5156 *) PKCS7_encrypt() was adding text MIME headers twice because they
5157 were added manually and by SMIME_crlf_copy().
5160 *) In bntest.c don't call BN_rand with zero bits argument.
5161 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5163 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5164 case was implemented. This caused BN_div_recp() to fail occasionally.
5167 *) Add an optional second argument to the set_label() in the perl
5168 assembly language builder. If this argument exists and is set
5169 to 1 it signals that the assembler should use a symbol whose
5170 scope is the entire file, not just the current function. This
5171 is needed with MASM which uses the format label:: for this scope.
5172 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5174 *) Change the ASN1 types so they are typedefs by default. Before
5175 almost all types were #define'd to ASN1_STRING which was causing
5176 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5180 *) Change names of new functions to the new get1/get0 naming
5181 convention: After 'get1', the caller owns a reference count
5182 and has to call ..._free; 'get0' returns a pointer to some
5183 data structure without incrementing reference counters.
5184 (Some of the existing 'get' functions increment a reference
5185 counter, some don't.)
5186 Similarly, 'set1' and 'add1' functions increase reference
5187 counters or duplicate objects.
5190 *) Allow for the possibility of temp RSA key generation failure:
5191 the code used to assume it always worked and crashed on failure.
5194 *) Fix potential buffer overrun problem in BIO_printf().
5195 [Ulf Möller, using public domain code by Patrick Powell; problem
5196 pointed out by David Sacerdote <das33@cornell.edu>]
5198 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5199 RAND_egd() and RAND_status(). In the command line application,
5200 the EGD socket can be specified like a seed file using RANDFILE
5204 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5205 Some CAs (e.g. Verisign) distribute certificates in this form.
5208 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5209 list to exclude them. This means that no special compilation option
5210 is needed to use anonymous DH: it just needs to be included in the
5214 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5215 EVP_MD_type. The old functionality is available in a new macro called
5216 EVP_MD_md(). Change code that uses it and update docs.
5219 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5220 where the 'void *' argument is replaced by a function pointer argument.
5221 Previously 'void *' was abused to point to functions, which works on
5222 many platforms, but is not correct. As these functions are usually
5223 called by macros defined in OpenSSL header files, most source code
5224 should work without changes.
5227 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5228 sections with information on -D... compiler switches used for
5229 compiling the library so that applications can see them. To enable
5230 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5231 must be defined. E.g.,
5232 #define OPENSSL_ALGORITHM_DEFINES
5233 #include <openssl/opensslconf.h>
5234 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5235 [Richard Levitte, Ulf and Bodo Möller]
5237 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5241 *) Change the 'other' type in certificate aux info to a STACK_OF
5242 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5243 the required ASN1 format: arbitrary types determined by an OID.
5246 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5247 argument to 'req'. This is not because the function is newer or
5248 better than others it just uses the work 'NEW' in the certificate
5249 request header lines. Some software needs this.
5252 *) Reorganise password command line arguments: now passwords can be
5253 obtained from various sources. Delete the PEM_cb function and make
5254 it the default behaviour: i.e. if the callback is NULL and the
5255 usrdata argument is not NULL interpret it as a null terminated pass
5256 phrase. If usrdata and the callback are NULL then the pass phrase
5257 is prompted for as usual.
5260 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5261 the support is automatically enabled. The resulting binaries will
5262 autodetect the card and use it if present.
5263 [Ben Laurie and Compaq Inc.]
5265 *) Work around for Netscape hang bug. This sends certificate request
5266 and server done in one record. Since this is perfectly legal in the
5267 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5268 the bugs/SSLv3 entry for more info.
5271 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5274 *) Add -rand argument to smime and pkcs12 applications and read/write
5278 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5281 *) Add command line password options to the remaining applications.
5284 *) Bug fix for BN_div_recp() for numerators with an even number of
5288 *) More tests in bntest.c, and changed test_bn output.
5291 *) ./config recognizes MacOS X now.
5294 *) Bug fix for BN_div() when the first words of num and divsor are
5295 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5298 *) Add support for various broken PKCS#8 formats, and command line
5299 options to produce them.
5302 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5303 get temporary BIGNUMs from a BN_CTX.
5306 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5310 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5311 include a #define from the old name to the new. The original intent
5312 was that statically linked binaries could for example just call
5313 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5314 link with digests. This never worked becayse SSLeay_add_all_digests()
5315 and SSLeay_add_all_ciphers() were in the same source file so calling
5316 one would link with the other. They are now in separate source files.
5319 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5322 *) Use a less unusual form of the Miller-Rabin primality test (it used
5323 a binary algorithm for exponentiation integrated into the Miller-Rabin
5324 loop, our standard modexp algorithms are faster).
5327 *) Support for the EBCDIC character set completed.
5328 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5330 *) Source code cleanups: use const where appropriate, eliminate casts,
5331 use void * instead of char * in lhash.
5334 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5335 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5336 this the server could overwrite ephemeral keys that the client
5340 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5341 using 50 iterations of the Rabin-Miller test.
5343 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5344 iterations of the Rabin-Miller test as required by the appendix
5345 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5346 As BN_is_prime_fasttest includes trial division, DSA parameter
5347 generation becomes much faster.
5349 This implies a change for the callback functions in DSA_is_prime
5350 and DSA_generate_parameters: The callback function is called once
5351 for each positive witness in the Rabin-Miller test, not just
5352 occasionally in the inner loop; and the parameters to the
5353 callback function now provide an iteration count for the outer
5354 loop rather than for the current invocation of the inner loop.
5355 DSA_generate_parameters additionally can call the callback
5356 function with an 'iteration count' of -1, meaning that a
5357 candidate has passed the trial division test (when q is generated
5358 from an application-provided seed, trial division is skipped).
5361 *) New function BN_is_prime_fasttest that optionally does trial
5362 division before starting the Rabin-Miller test and has
5363 an additional BN_CTX * argument (whereas BN_is_prime always
5364 has to allocate at least one BN_CTX).
5365 'callback(1, -1, cb_arg)' is called when a number has passed the
5366 trial division stage.
5369 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5373 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5376 *) New function BN_pseudo_rand().
5379 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5380 bignum version of BN_from_montgomery() with the working code from
5381 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5385 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5386 made it impossible to use the same SSL_SESSION data structure in
5387 SSL2 clients in multiple threads.
5390 *) The return value of RAND_load_file() no longer counts bytes obtained
5391 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5392 to seed the PRNG (previously an explicit byte count was required).
5393 [Ulf Möller, Bodo Möller]
5395 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5396 used (char *) instead of (void *) and had casts all over the place.
5399 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5402 *) Retain source code compatibility for BN_prime_checks macro:
5403 BN_is_prime(..., BN_prime_checks, ...) now uses
5404 BN_prime_checks_for_size to determine the appropriate number of
5405 Rabin-Miller iterations.
5408 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5409 DH_CHECK_P_NOT_SAFE_PRIME.
5410 (Check if this is true? OpenPGP calls them "strong".)
5413 *) Merge the functionality of "dh" and "gendh" programs into a new program
5414 "dhparam". The old programs are retained for now but will handle DH keys
5415 (instead of parameters) in future.
5418 *) Make the ciphers, s_server and s_client programs check the return values
5419 when a new cipher list is set.
5422 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5423 ciphers. Before when the 56bit ciphers were enabled the sorting was
5426 The syntax for the cipher sorting has been extended to support sorting by
5427 cipher-strength (using the strength_bits hard coded in the tables).
5428 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5430 Fix a bug in the cipher-command parser: when supplying a cipher command
5431 string with an "undefined" symbol (neither command nor alphanumeric
5432 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5433 an error is flagged.
5435 Due to the strength-sorting extension, the code of the
5436 ssl_create_cipher_list() function was completely rearranged. I hope that
5437 the readability was also increased :-)
5438 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5440 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5441 for the first serial number and places 2 in the serial number file. This
5442 avoids problems when the root CA is created with serial number zero and
5443 the first user certificate has the same issuer name and serial number
5447 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5448 the new code. Add documentation for this stuff.
5451 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5452 X509_*() to X509at_*() on the grounds that they don't handle X509
5453 structures and behave in an analagous way to the X509v3 functions:
5454 they shouldn't be called directly but wrapper functions should be used
5457 So we also now have some wrapper functions that call the X509at functions
5458 when passed certificate requests. (TO DO: similar things can be done with
5459 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5460 things. Some of these need some d2i or i2d and print functionality
5461 because they handle more complex structures.)
5464 *) Add missing #ifndefs that caused missing symbols when building libssl
5465 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5466 NO_RSA in ssl/s2*.c.
5467 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5469 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5470 has a return value which indicates the quality of the random data
5471 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5472 error queue. New function RAND_pseudo_bytes() generates output that is
5473 guaranteed to be unique but not unpredictable. RAND_add is like
5474 RAND_seed, but takes an extra argument for an entropy estimate
5475 (RAND_seed always assumes full entropy).
5478 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5479 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5480 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5481 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5482 false-positive rate of at most 2^-80 for random input.
5485 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5488 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5489 in the 0.9.5 release), this returns the chain
5490 from an X509_CTX structure with a dup of the stack and all
5491 the X509 reference counts upped: so the stack will exist
5492 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5495 Also make SSL_SESSION_print() print out the verify return
5499 *) Add manpage for the pkcs12 command. Also change the default
5500 behaviour so MAC iteration counts are used unless the new
5501 -nomaciter option is used. This improves file security and
5502 only older versions of MSIE (4.0 for example) need it.
5505 *) Honor the no-xxx Configure options when creating .DEF files.
5508 *) Add PKCS#10 attributes to field table: challengePassword,
5509 unstructuredName and unstructuredAddress. These are taken from
5510 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5511 international characters are used.
5513 More changes to X509_ATTRIBUTE code: allow the setting of types
5514 based on strings. Remove the 'loc' parameter when adding
5515 attributes because these will be a SET OF encoding which is sorted
5519 *) Initial changes to the 'req' utility to allow request generation
5520 automation. This will allow an application to just generate a template
5521 file containing all the field values and have req construct the
5524 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5525 used all over the place including certificate requests and PKCS#7
5526 structures. They are currently handled manually where necessary with
5527 some primitive wrappers for PKCS#7. The new functions behave in a
5528 manner analogous to the X509 extension functions: they allow
5529 attributes to be looked up by NID and added.
5531 Later something similar to the X509V3 code would be desirable to
5532 automatically handle the encoding, decoding and printing of the
5533 more complex types. The string types like challengePassword can
5534 be handled by the string table functions.
5536 Also modified the multi byte string table handling. Now there is
5537 a 'global mask' which masks out certain types. The table itself
5538 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5539 is useful when for example there is only one permissible type
5540 (as in countryName) and using the mask might result in no valid
5544 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5545 SSL_get_peer_finished to allow applications to obtain the latest
5546 Finished messages sent to the peer or expected from the peer,
5547 respectively. (SSL_get_peer_finished is usually the Finished message
5548 actually received from the peer, otherwise the protocol will be aborted.)
5550 As the Finished message are message digests of the complete handshake
5551 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5552 be used for external authentication procedures when the authentication
5553 provided by SSL/TLS is not desired or is not enough.
5556 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5557 the host supports BWX extension and if Compaq C is present on the
5558 $PATH. Just exploiting of the BWX extension results in 20-30%
5559 performance kick for some algorithms, e.g. DES and RC4 to mention
5560 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5564 *) Add support for MS "fast SGC". This is arguably a violation of the
5565 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5566 weak crypto and after checking the certificate is SGC a second one
5567 with strong crypto. MS SGC stops the first handshake after receiving
5568 the server certificate message and sends a second client hello. Since
5569 a server will typically do all the time consuming operations before
5570 expecting any further messages from the client (server key exchange
5571 is the most expensive) there is little difference between the two.
5573 To get OpenSSL to support MS SGC we have to permit a second client
5574 hello message after we have sent server done. In addition we have to
5575 reset the MAC if we do get this second client hello.
5578 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5579 if a DER encoded private key is RSA or DSA traditional format. Changed
5580 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5581 format DER encoded private key. Newer code should use PKCS#8 format which
5582 has the key type encoded in the ASN1 structure. Added DER private key
5583 support to pkcs8 application.
5586 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5587 ciphersuites has been selected (as required by the SSL 3/TLS 1
5588 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5589 is set, we interpret this as a request to violate the specification
5590 (the worst that can happen is a handshake failure, and 'correct'
5591 behaviour would result in a handshake failure anyway).
5594 *) In SSL_CTX_add_session, take into account that there might be multiple
5595 SSL_SESSION structures with the same session ID (e.g. when two threads
5596 concurrently obtain them from an external cache).
5597 The internal cache can handle only one SSL_SESSION with a given ID,
5598 so if there's a conflict, we now throw out the old one to achieve
5602 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5603 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5604 some routines that use cipher OIDs: some ciphers do not have OIDs
5605 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5609 *) Simplify the trust setting structure and code. Now we just have
5610 two sequences of OIDs for trusted and rejected settings. These will
5611 typically have values the same as the extended key usage extension
5612 and any application specific purposes.
5614 The trust checking code now has a default behaviour: it will just
5615 check for an object with the same NID as the passed id. Functions can
5616 be provided to override either the default behaviour or the behaviour
5617 for a given id. SSL client, server and email already have functions
5618 in place for compatibility: they check the NID and also return "trusted"
5619 if the certificate is self signed.
5622 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5623 traditional format into an EVP_PKEY structure.
5626 *) Add a password callback function PEM_cb() which either prompts for
5627 a password if usr_data is NULL or otherwise assumes it is a null
5628 terminated password. Allow passwords to be passed on command line
5629 environment or config files in a few more utilities.
5632 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5633 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5634 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5635 Update documentation.
5638 *) Support for ASN1 "NULL" type. This could be handled before by using
5639 ASN1_TYPE but there wasn't any function that would try to read a NULL
5640 and produce an error if it couldn't. For compatibility we also have
5641 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5642 don't allocate anything because they don't need to.
5645 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5647 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5649 *) Rebuild of the memory allocation routines used by OpenSSL code and
5650 possibly others as well. The purpose is to make an interface that
5651 provide hooks so anyone can build a separate set of allocation and
5652 deallocation routines to be used by OpenSSL, for example memory
5653 pool implementations, or something else, which was previously hard
5654 since Malloc(), Realloc() and Free() were defined as macros having
5655 the values malloc, realloc and free, respectively (except for Win32
5656 compilations). The same is provided for memory debugging code.
5657 OpenSSL already comes with functionality to find memory leaks, but
5658 this gives people a chance to debug other memory problems.
5660 With these changes, a new set of functions and macros have appeared:
5662 CRYPTO_set_mem_debug_functions() [F]
5663 CRYPTO_get_mem_debug_functions() [F]
5664 CRYPTO_dbg_set_options() [F]
5665 CRYPTO_dbg_get_options() [F]
5666 CRYPTO_malloc_debug_init() [M]
5668 The memory debug functions are NULL by default, unless the library
5669 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5670 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5671 gives the standard debugging functions that come with OpenSSL) or
5672 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5673 provided by the library user) must be used. When the standard
5674 debugging functions are used, CRYPTO_dbg_set_options can be used to
5675 request additional information:
5676 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5677 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5679 Also, things like CRYPTO_set_mem_functions will always give the
5680 expected result (the new set of functions is used for allocation
5681 and deallocation) at all times, regardless of platform and compiler
5684 To finish it up, some functions that were never use in any other
5685 way than through macros have a new API and new semantic:
5688 CRYPTO_dbg_realloc()
5691 All macros of value have retained their old syntax.
5692 [Richard Levitte and Bodo Moeller]
5694 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5695 ordering of SMIMECapabilities wasn't in "strength order" and there
5696 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5700 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5701 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5702 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5704 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5705 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5706 functionality to handle multipart/signed properly) and a utility
5707 called 'smime' to call all this stuff. This is based on code I
5708 originally wrote for Celo who have kindly allowed it to be
5709 included in OpenSSL.
5712 *) Add variants des_set_key_checked and des_set_key_unchecked of
5713 des_set_key (aka des_key_sched). Global variable des_check_key
5714 decides which of these is called by des_set_key; this way
5715 des_check_key behaves as it always did, but applications and
5716 the library itself, which was buggy for des_check_key == 1,
5717 have a cleaner way to pick the version they need.
5720 *) New function PKCS12_newpass() which changes the password of a
5724 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5725 dynamic mix. In both cases the ids can be used as an index into the
5726 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5727 functions so they accept a list of the field values and the
5728 application doesn't need to directly manipulate the X509_TRUST
5732 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5736 *) Modify the way the V3 extension code looks up extensions. This now
5737 works in a similar way to the object code: we have some "standard"
5738 extensions in a static table which is searched with OBJ_bsearch()
5739 and the application can add dynamic ones if needed. The file
5740 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5741 updated whenever a new extension is added to the core code and kept
5742 in ext_nid order. There is a simple program 'tabtest.c' which checks
5743 this. New extensions are not added too often so this file can readily
5744 be maintained manually.
5746 There are two big advantages in doing things this way. The extensions
5747 can be looked up immediately and no longer need to be "added" using
5748 X509V3_add_standard_extensions(): this function now does nothing.
5749 [Side note: I get *lots* of email saying the extension code doesn't
5750 work because people forget to call this function]
5751 Also no dynamic allocation is done unless new extensions are added:
5752 so if we don't add custom extensions there is no need to call
5753 X509V3_EXT_cleanup().
5756 *) Modify enc utility's salting as follows: make salting the default. Add a
5757 magic header, so unsalted files fail gracefully instead of just decrypting
5758 to garbage. This is because not salting is a big security hole, so people
5759 should be discouraged from doing it.
5762 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5763 digest to be passed on the command line but it only used this
5764 parameter when signing a certificate. Modified so all relevant
5765 operations are affected by the digest parameter including the
5766 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5767 DSA key was used because it didn't fix the digest.
5770 *) Initial certificate chain verify code. Currently tests the untrusted
5771 certificates for consistency with the verify purpose (which is set
5772 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5774 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5775 this is because it will reject chains with invalid extensions whereas
5776 every previous version of OpenSSL and SSLeay made no checks at all.
5778 Trust code: checks the root CA for the relevant trust settings. Trust
5779 settings have an initial value consistent with the verify purpose: e.g.
5780 if the verify purpose is for SSL client use it expects the CA to be
5781 trusted for SSL client use. However the default value can be changed to
5782 permit custom trust settings: one example of this would be to only trust
5783 certificates from a specific "secure" set of CAs.
5785 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5786 which should be used for version portability: especially since the
5787 verify structure is likely to change more often now.
5789 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5790 to set them. If not set then assume SSL clients will verify SSL servers
5793 Two new options to the verify program: -untrusted allows a set of
5794 untrusted certificates to be passed in and -purpose which sets the
5795 intended purpose of the certificate. If a purpose is set then the
5796 new chain verify code is used to check extension consistency.
5799 *) Support for the authority information access extension.
5802 *) Modify RSA and DSA PEM read routines to transparently handle
5803 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5804 public keys in a format compatible with certificate
5805 SubjectPublicKeyInfo structures. Unfortunately there were already
5806 functions called *_PublicKey_* which used various odd formats so
5807 these are retained for compatibility: however the DSA variants were
5808 never in a public release so they have been deleted. Changed dsa/rsa
5809 utilities to handle the new format: note no releases ever handled public
5810 keys so we should be OK.
5812 The primary motivation for this change is to avoid the same fiasco
5813 that dogs private keys: there are several incompatible private key
5814 formats some of which are standard and some OpenSSL specific and
5815 require various evil hacks to allow partial transparent handling and
5816 even then it doesn't work with DER formats. Given the option anything
5817 other than PKCS#8 should be dumped: but the other formats have to
5818 stay in the name of compatibility.
5820 With public keys and the benefit of hindsight one standard format
5821 is used which works with EVP_PKEY, RSA or DSA structures: though
5822 it clearly returns an error if you try to read the wrong kind of key.
5824 Added a -pubkey option to the 'x509' utility to output the public key.
5825 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5826 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5827 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5828 that do the same as the EVP_PKEY_assign_*() except they up the
5829 reference count of the added key (they don't "swallow" the
5833 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5834 CRLs would fail if the file contained no certificates or no CRLs:
5835 added a new function to read in both types and return the number
5836 read: this means that if none are read it will be an error. The
5837 DER versions of the certificate and CRL reader would always fail
5838 because it isn't possible to mix certificates and CRLs in DER format
5839 without choking one or the other routine. Changed this to just read
5840 a certificate: this is the best we can do. Also modified the code
5841 in apps/verify.c to take notice of return codes: it was previously
5842 attempting to read in certificates from NULL pointers and ignoring
5843 any errors: this is one reason why the cert and CRL reader seemed
5844 to work. It doesn't check return codes from the default certificate
5845 routines: these may well fail if the certificates aren't installed.
5848 *) Code to support otherName option in GeneralName.
5851 *) First update to verify code. Change the verify utility
5852 so it warns if it is passed a self signed certificate:
5853 for consistency with the normal behaviour. X509_verify
5854 has been modified to it will now verify a self signed
5855 certificate if *exactly* the same certificate appears
5856 in the store: it was previously impossible to trust a
5857 single self signed certificate. This means that:
5858 openssl verify ss.pem
5859 now gives a warning about a self signed certificate but
5860 openssl verify -CAfile ss.pem ss.pem
5864 *) For servers, store verify_result in SSL_SESSION data structure
5865 (and add it to external session representation).
5866 This is needed when client certificate verifications fails,
5867 but an application-provided verification callback (set by
5868 SSL_CTX_set_cert_verify_callback) allows accepting the session
5869 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5870 but returns 1): When the session is reused, we have to set
5871 ssl->verify_result to the appropriate error code to avoid
5873 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5875 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5876 case in PKCS7_dataInit() where the signed PKCS7 structure
5877 didn't contain any existing data because it was being created.
5878 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5880 *) Add a salt to the key derivation routines in enc.c. This
5881 forms the first 8 bytes of the encrypted file. Also add a
5882 -S option to allow a salt to be input on the command line.
5885 *) New function X509_cmp(). Oddly enough there wasn't a function
5886 to compare two certificates. We do this by working out the SHA1
5887 hash and comparing that. X509_cmp() will be needed by the trust
5891 *) SSL_get1_session() is like SSL_get_session(), but increments
5892 the reference count in the SSL_SESSION returned.
5893 [Geoff Thorpe <geoff@eu.c2.net>]
5895 *) Fix for 'req': it was adding a null to request attributes.
5896 Also change the X509_LOOKUP and X509_INFO code to handle
5897 certificate auxiliary information.
5900 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5904 *) Add the possibility to add extra information to the memory leak
5905 detecting output, to form tracebacks, showing from where each
5906 allocation was originated: CRYPTO_push_info("constant string") adds
5907 the string plus current file name and line number to a per-thread
5908 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5909 is like calling CYRPTO_pop_info() until the stack is empty.
5910 Also updated memory leak detection code to be multi-thread-safe.
5913 *) Add options -text and -noout to pkcs7 utility and delete the
5914 encryption options which never did anything. Update docs.
5917 *) Add options to some of the utilities to allow the pass phrase
5918 to be included on either the command line (not recommended on
5919 OSes like Unix) or read from the environment. Update the
5920 manpages and fix a few bugs.
5923 *) Add a few manpages for some of the openssl commands.
5926 *) Fix the -revoke option in ca. It was freeing up memory twice,
5927 leaking and not finding already revoked certificates.
5930 *) Extensive changes to support certificate auxiliary information.
5931 This involves the use of X509_CERT_AUX structure and X509_AUX
5932 functions. An X509_AUX function such as PEM_read_X509_AUX()
5933 can still read in a certificate file in the usual way but it
5934 will also read in any additional "auxiliary information". By
5935 doing things this way a fair degree of compatibility can be
5936 retained: existing certificates can have this information added
5937 using the new 'x509' options.
5939 Current auxiliary information includes an "alias" and some trust
5940 settings. The trust settings will ultimately be used in enhanced
5941 certificate chain verification routines: currently a certificate
5942 can only be trusted if it is self signed and then it is trusted
5946 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5947 The problem was that one of the replacement routines had not been working
5948 since SSLeay releases. For now the offending routine has been replaced
5949 with non-optimised assembler. Even so, this now gives around 95%
5950 performance improvement for 1024 bit RSA signs.
5953 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5954 handling. Most clients have the effective key size in bits equal to
5955 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5956 A few however don't do this and instead use the size of the decrypted key
5957 to determine the RC2 key length and the AlgorithmIdentifier to determine
5958 the effective key length. In this case the effective key length can still
5959 be 40 bits but the key length can be 168 bits for example. This is fixed
5960 by manually forcing an RC2 key into the EVP_PKEY structure because the
5961 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5962 the key length and effective key length are equal.
5965 *) Add a bunch of functions that should simplify the creation of
5966 X509_NAME structures. Now you should be able to do:
5967 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5968 and have it automatically work out the correct field type and fill in
5969 the structures. The more adventurous can try:
5970 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5971 and it will (hopefully) work out the correct multibyte encoding.
5974 *) Change the 'req' utility to use the new field handling and multibyte
5975 copy routines. Before the DN field creation was handled in an ad hoc
5976 way in req, ca, and x509 which was rather broken and didn't support
5977 BMPStrings or UTF8Strings. Since some software doesn't implement
5978 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5979 using the dirstring_type option. See the new comment in the default
5980 openssl.cnf for more info.
5983 *) Make crypto/rand/md_rand.c more robust:
5984 - Assure unique random numbers after fork().
5985 - Make sure that concurrent threads access the global counter and
5986 md serializably so that we never lose entropy in them
5987 or use exactly the same state in multiple threads.
5988 Access to the large state is not always serializable because
5989 the additional locking could be a performance killer, and
5990 md should be large enough anyway.
5993 *) New file apps/app_rand.c with commonly needed functionality
5994 for handling the random seed file.
5996 Use the random seed file in some applications that previously did not:
5998 dsaparam -genkey (which also ignored its '-rand' option),
6001 x509 (when signing).
6002 Except on systems with /dev/urandom, it is crucial to have a random
6003 seed file at least for key creation, DSA signing, and for DH exchanges;
6004 for RSA signatures we could do without one.
6006 gendh and gendsa (unlike genrsa) used to read only the first byte
6007 of each file listed in the '-rand' option. The function as previously
6008 found in genrsa is now in app_rand.c and is used by all programs
6009 that support '-rand'.
6012 *) In RAND_write_file, use mode 0600 for creating files;
6013 don't just chmod when it may be too late.
6016 *) Report an error from X509_STORE_load_locations
6017 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6020 *) New function ASN1_mbstring_copy() this copies a string in either
6021 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6022 into an ASN1_STRING type. A mask of permissible types is passed
6023 and it chooses the "minimal" type to use or an error if not type
6027 *) Add function equivalents to the various macros in asn1.h. The old
6028 macros are retained with an M_ prefix. Code inside the library can
6029 use the M_ macros. External code (including the openssl utility)
6030 should *NOT* in order to be "shared library friendly".
6033 *) Add various functions that can check a certificate's extensions
6034 to see if it usable for various purposes such as SSL client,
6035 server or S/MIME and CAs of these types. This is currently
6036 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6037 verification. Also added a -purpose flag to x509 utility to
6038 print out all the purposes.
6041 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6045 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6046 for, obtain and decode and extension and obtain its critical flag.
6047 This allows all the necessary extension code to be handled in a
6048 single function call.
6051 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6052 platforms. See crypto/rc4/rc4_enc.c for further details.
6055 *) New -noout option to asn1parse. This causes no output to be produced
6056 its main use is when combined with -strparse and -out to extract data
6057 from a file (which may not be in ASN.1 format).
6060 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6061 when producing the local key id.
6062 [Richard Levitte <levitte@stacken.kth.se>]
6064 *) New option -dhparam in s_server. This allows a DH parameter file to be
6065 stated explicitly. If it is not stated then it tries the first server
6066 certificate file. The previous behaviour hard coded the filename
6070 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6071 a public key to be input or output. For example:
6072 openssl rsa -in key.pem -pubout -out pubkey.pem
6073 Also added necessary DSA public key functions to handle this.
6076 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6077 in the message. This was handled by allowing
6078 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6079 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6081 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6082 to the end of the strings whereas this didn't. This would cause problems
6083 if strings read with d2i_ASN1_bytes() were later modified.
6084 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6086 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6087 data and it contains EOF it will end up returning an error. This is
6088 caused by input 46 bytes long. The cause is due to the way base64
6089 BIOs find the start of base64 encoded data. They do this by trying a
6090 trial decode on each line until they find one that works. When they
6091 do a flag is set and it starts again knowing it can pass all the
6092 data directly through the decoder. Unfortunately it doesn't reset
6093 the context it uses. This means that if EOF is reached an attempt
6094 is made to pass two EOFs through the context and this causes the
6095 resulting error. This can also cause other problems as well. As is
6096 usual with these problems it takes *ages* to find and the fix is
6097 trivial: move one line.
6098 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6100 *) Ugly workaround to get s_client and s_server working under Windows. The
6101 old code wouldn't work because it needed to select() on sockets and the
6102 tty (for keypresses and to see if data could be written). Win32 only
6103 supports select() on sockets so we select() with a 1s timeout on the
6104 sockets and then see if any characters are waiting to be read, if none
6105 are present then we retry, we also assume we can always write data to
6106 the tty. This isn't nice because the code then blocks until we've
6107 received a complete line of data and it is effectively polling the
6108 keyboard at 1s intervals: however it's quite a bit better than not
6109 working at all :-) A dedicated Windows application might handle this
6110 with an event loop for example.
6113 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6114 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6115 will be called when RSA_sign() and RSA_verify() are used. This is useful
6116 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6117 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6118 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6119 This necessitated the support of an extra signature type NID_md5_sha1
6120 for SSL signatures and modifications to the SSL library to use it instead
6121 of calling RSA_public_decrypt() and RSA_private_encrypt().
6124 *) Add new -verify -CAfile and -CApath options to the crl program, these
6125 will lookup a CRL issuers certificate and verify the signature in a
6126 similar way to the verify program. Tidy up the crl program so it
6127 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6128 less strict. It will now permit CRL extensions even if it is not
6129 a V2 CRL: this will allow it to tolerate some broken CRLs.
6132 *) Initialize all non-automatic variables each time one of the openssl
6133 sub-programs is started (this is necessary as they may be started
6134 multiple times from the "OpenSSL>" prompt).
6135 [Lennart Bang, Bodo Moeller]
6137 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6138 removing all other RSA functionality (this is what NO_RSA does). This
6139 is so (for example) those in the US can disable those operations covered
6140 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6144 *) Non-copying interface to BIO pairs.
6145 (still largely untested)
6148 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6149 ASCII string. This was handled independently in various places before.
6152 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6153 UTF8 strings a character at a time.
6156 *) Use client_version from client hello to select the protocol
6157 (s23_srvr.c) and for RSA client key exchange verification
6158 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6161 *) Add various utility functions to handle SPKACs, these were previously
6162 handled by poking round in the structure internals. Added new function
6163 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6164 print, verify and generate SPKACs. Based on an original idea from
6165 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6168 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6171 *) Allow the config file extension section to be overwritten on the
6172 command line. Based on an original idea from Massimiliano Pala
6173 <madwolf@comune.modena.it>. The new option is called -extensions
6174 and can be applied to ca, req and x509. Also -reqexts to override
6175 the request extensions in req and -crlexts to override the crl extensions
6179 *) Add new feature to the SPKAC handling in ca. Now you can include
6180 the same field multiple times by preceding it by "XXXX." for example:
6183 this is the same syntax as used in the req config file.
6186 *) Allow certificate extensions to be added to certificate requests. These
6187 are specified in a 'req_extensions' option of the req section of the
6188 config file. They can be printed out with the -text option to req but
6189 are otherwise ignored at present.
6192 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6193 data read consists of only the final block it would not decrypted because
6194 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6195 A misplaced 'break' also meant the decrypted final block might not be
6196 copied until the next read.
6199 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6200 a few extra parameters to the DH structure: these will be useful if
6201 for example we want the value of 'q' or implement X9.42 DH.
6204 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6205 provides hooks that allow the default DSA functions or functions on a
6206 "per key" basis to be replaced. This allows hardware acceleration and
6207 hardware key storage to be handled without major modification to the
6208 library. Also added low level modexp hooks and CRYPTO_EX structure and
6209 associated functions.
6212 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6213 as "read only": it can't be written to and the buffer it points to will
6214 not be freed. Reading from a read only BIO is much more efficient than
6215 a normal memory BIO. This was added because there are several times when
6216 an area of memory needs to be read from a BIO. The previous method was
6217 to create a memory BIO and write the data to it, this results in two
6218 copies of the data and an O(n^2) reading algorithm. There is a new
6219 function BIO_new_mem_buf() which creates a read only memory BIO from
6220 an area of memory. Also modified the PKCS#7 routines to use read only
6224 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6225 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6226 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6227 but a retry condition occured while trying to read the rest.
6230 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6231 NID_pkcs7_encrypted by default: this was wrong since this should almost
6232 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6233 the encrypted data type: this is a more sensible place to put it and it
6234 allows the PKCS#12 code to be tidied up that duplicated this
6238 *) Changed obj_dat.pl script so it takes its input and output files on
6239 the command line. This should avoid shell escape redirection problems
6243 *) Initial support for certificate extension requests, these are included
6244 in things like Xenroll certificate requests. Included functions to allow
6245 extensions to be obtained and added.
6248 *) -crlf option to s_client and s_server for sending newlines as
6249 CRLF (as required by many protocols).
6252 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6254 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6255 [Ralf S. Engelschall]
6257 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6258 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6260 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6264 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6265 DH parameters/keys (q is lost during that conversion, but the resulting
6266 DH parameters contain its length).
6268 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6269 much faster than DH_generate_parameters (which creates parameters
6270 where p = 2*q + 1), and also the smaller q makes DH computations
6271 much more efficient (160-bit exponentiation instead of 1024-bit
6272 exponentiation); so this provides a convenient way to support DHE
6273 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6274 utter importance to use
6275 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6277 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6278 when such DH parameters are used, because otherwise small subgroup
6279 attacks may become possible!
6282 *) Avoid memory leak in i2d_DHparams.
6285 *) Allow the -k option to be used more than once in the enc program:
6286 this allows the same encrypted message to be read by multiple recipients.
6289 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6290 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6291 it will always use the numerical form of the OID, even if it has a short
6295 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6296 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6297 otherwise bn_mod_exp was called. In the case of hardware keys for example
6298 no private key components need be present and it might store extra data
6299 in the RSA structure, which cannot be accessed from bn_mod_exp.
6300 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6301 private key operations.
6304 *) Added support for SPARC Linux.
6307 *) pem_password_cb function type incompatibly changed from
6308 typedef int pem_password_cb(char *buf, int size, int rwflag);
6310 ....(char *buf, int size, int rwflag, void *userdata);
6311 so that applications can pass data to their callbacks:
6312 The PEM[_ASN1]_{read,write}... functions and macros now take an
6313 additional void * argument, which is just handed through whenever
6314 the password callback is called.
6315 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6317 New function SSL_CTX_set_default_passwd_cb_userdata.
6319 Compatibility note: As many C implementations push function arguments
6320 onto the stack in reverse order, the new library version is likely to
6321 interoperate with programs that have been compiled with the old
6322 pem_password_cb definition (PEM_whatever takes some data that
6323 happens to be on the stack as its last argument, and the callback
6324 just ignores this garbage); but there is no guarantee whatsoever that
6327 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6328 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6329 problems not only on Windows, but also on some Unix platforms.
6330 To avoid problematic command lines, these definitions are now in an
6331 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6332 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6335 *) MIPS III/IV assembler module is reimplemented.
6338 *) More DES library cleanups: remove references to srand/rand and
6339 delete an unused file.
6342 *) Add support for the the free Netwide assembler (NASM) under Win32,
6343 since not many people have MASM (ml) and it can be hard to obtain.
6344 This is currently experimental but it seems to work OK and pass all
6345 the tests. Check out INSTALL.W32 for info.
6348 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6349 without temporary keys kept an extra copy of the server key,
6350 and connections with temporary keys did not free everything in case
6354 *) New function RSA_check_key and new openssl rsa option -check
6355 for verifying the consistency of RSA keys.
6356 [Ulf Moeller, Bodo Moeller]
6358 *) Various changes to make Win32 compile work:
6359 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6360 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6361 comparison" warnings.
6362 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6365 *) Add a debugging option to PKCS#5 v2 key generation function: when
6366 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6367 derived keys are printed to stderr.
6370 *) Copy the flags in ASN1_STRING_dup().
6371 [Roman E. Pavlov <pre@mo.msk.ru>]
6373 *) The x509 application mishandled signing requests containing DSA
6374 keys when the signing key was also DSA and the parameters didn't match.
6376 It was supposed to omit the parameters when they matched the signing key:
6377 the verifying software was then supposed to automatically use the CA's
6378 parameters if they were absent from the end user certificate.
6380 Omitting parameters is no longer recommended. The test was also
6381 the wrong way round! This was probably due to unusual behaviour in
6382 EVP_cmp_parameters() which returns 1 if the parameters match.
6383 This meant that parameters were omitted when they *didn't* match and
6384 the certificate was useless. Certificates signed with 'ca' didn't have
6386 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6388 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6389 The interface is as follows:
6390 Applications can use
6391 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6392 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6393 "off" is now the default.
6394 The library internally uses
6395 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6396 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6397 to disable memory-checking temporarily.
6399 Some inconsistent states that previously were possible (and were
6400 even the default) are now avoided.
6402 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6403 with each memory chunk allocated; this is occasionally more helpful
6404 than just having a counter.
6406 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6408 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6412 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6413 which largely parallels "options", but is for changing API behaviour,
6414 whereas "options" are about protocol behaviour.
6415 Initial "mode" flags are:
6417 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6418 a single record has been written.
6419 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6420 retries use the same buffer location.
6421 (But all of the contents must be
6425 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6428 *) Fix problems with no-hmac etc.
6429 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6431 *) New functions RSA_get_default_method(), RSA_set_method() and
6432 RSA_get_method(). These allows replacement of RSA_METHODs without having
6433 to mess around with the internals of an RSA structure.
6436 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6437 Also really enable memory leak checks in openssl.c and in some
6439 [Chad C. Mulligan, Bodo Moeller]
6441 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6442 up the length of negative integers. This has now been simplified to just
6443 store the length when it is first determined and use it later, rather
6444 than trying to keep track of where data is copied and updating it to
6446 [Steve Henson, reported by Brien Wheeler
6447 <bwheeler@authentica-security.com>]
6449 *) Add a new function PKCS7_signatureVerify. This allows the verification
6450 of a PKCS#7 signature but with the signing certificate passed to the
6451 function itself. This contrasts with PKCS7_dataVerify which assumes the
6452 certificate is present in the PKCS#7 structure. This isn't always the
6453 case: certificates can be omitted from a PKCS#7 structure and be
6454 distributed by "out of band" means (such as a certificate database).
6457 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6458 function prototypes in pem.h, also change util/mkdef.pl to add the
6459 necessary function names.
6462 *) mk1mf.pl (used by Windows builds) did not properly read the
6463 options set by Configure in the top level Makefile, and Configure
6464 was not even able to write more than one option correctly.
6465 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6468 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6469 file to be loaded from a BIO or FILE pointer. The BIO version will
6470 for example allow memory BIOs to contain config info.
6473 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6474 Whoever hopes to achieve shared-library compatibility across versions
6475 must use this, not the compile-time macro.
6476 (Exercise 0.9.4: Which is the minimum library version required by
6478 Note: All this applies only to multi-threaded programs, others don't
6482 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6483 through a BIO pair triggered the default case, i.e.
6484 SSLerr(...,SSL_R_UNKNOWN_STATE).
6487 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6488 can use the SSL library even if none of the specific BIOs is
6492 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6493 for the encoded length.
6494 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6496 *) Add initial documentation of the X509V3 functions.
6499 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6500 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6501 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6502 secure PKCS#8 private key format with a high iteration count.
6505 *) Fix determination of Perl interpreter: A perl or perl5
6506 _directory_ in $PATH was also accepted as the interpreter.
6507 [Ralf S. Engelschall]
6509 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6510 wrong with it but it was very old and did things like calling
6511 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6515 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6516 to use the new extension code.
6519 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6520 with macros. This should make it easier to change their form, add extra
6521 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6525 *) Add to configuration table a new entry that can specify an alternative
6526 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6527 according to Mark Crispin <MRC@Panda.COM>.
6531 *) DES CBC did not update the IV. Weird.
6534 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6535 Changing the behaviour of the former might break existing programs --
6536 where IV updating is needed, des_ncbc_encrypt can be used.
6539 *) When bntest is run from "make test" it drives bc to check its
6540 calculations, as well as internally checking them. If an internal check
6541 fails, it needs to cause bc to give a non-zero result or make test carries
6542 on without noticing the failure. Fixed.
6545 *) DES library cleanups.
6548 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6549 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6550 ciphers. NOTE: although the key derivation function has been verified
6551 against some published test vectors it has not been extensively tested
6552 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6556 *) Instead of "mkdir -p", which is not fully portable, use new
6557 Perl script "util/mkdir-p.pl".
6560 *) Rewrite the way password based encryption (PBE) is handled. It used to
6561 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6562 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6563 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6564 the 'parameter' field of the AlgorithmIdentifier is passed to the
6565 underlying key generation function so it must do its own ASN1 parsing.
6566 This has also changed the EVP_PBE_CipherInit() function which now has a
6567 'parameter' argument instead of literal salt and iteration count values
6568 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6571 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6572 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6573 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6574 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6575 value was just used as a "magic string" and not used directly its
6576 value doesn't matter.
6579 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6583 *) "linux-sparc64" configuration (ultrapenguin).
6584 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6585 "linux-sparc" configuration.
6586 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6588 *) config now generates no-xxx options for missing ciphers.
6591 *) Support the EBCDIC character set (work in progress).
6592 File ebcdic.c not yet included because it has a different license.
6593 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6595 *) Support BS2000/OSD-POSIX.
6596 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6598 *) Make callbacks for key generation use void * instead of char *.
6601 *) Make S/MIME samples compile (not yet tested).
6604 *) Additional typesafe stacks.
6607 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6611 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6613 *) New configuration variant "sco5-gcc".
6615 *) Updated some demos.
6616 [Sean O Riordain, Wade Scholine]
6618 *) Add missing BIO_free at exit of pkcs12 application.
6621 *) Fix memory leak in conf.c.
6624 *) Updates for Win32 to assembler version of MD5.
6627 *) Set #! path to perl in apps/der_chop to where we found it
6628 instead of using a fixed path.
6631 *) SHA library changes for irix64-mips4-cc.
6634 *) Improvements for VMS support.
6638 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6640 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6641 This also avoids the problems with SC4.2 and unpatched SC5.
6642 [Andy Polyakov <appro@fy.chalmers.se>]
6644 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6645 These are required because of the typesafe stack would otherwise break
6646 existing code. If old code used a structure member which used to be STACK
6647 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6648 sk_num or sk_value it would produce an error because the num, data members
6649 are not present in STACK_OF. Now it just produces a warning. sk_set
6650 replaces the old method of assigning a value to sk_value
6651 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6652 that does this will no longer work (and should use sk_set instead) but
6653 this could be regarded as a "questionable" behaviour anyway.
6656 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6657 correctly handle encrypted S/MIME data.
6660 *) Change type of various DES function arguments from des_cblock
6661 (which means, in function argument declarations, pointer to char)
6662 to des_cblock * (meaning pointer to array with 8 char elements),
6663 which allows the compiler to do more typechecking; it was like
6664 that back in SSLeay, but with lots of ugly casts.
6666 Introduce new type const_des_cblock.
6669 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6670 problems: find RecipientInfo structure that matches recipient certificate
6671 and initialise the ASN1 structures properly based on passed cipher.
6674 *) Belatedly make the BN tests actually check the results.
6677 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6678 to and from BNs: it was completely broken. New compilation option
6679 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6680 key elements as negative integers.
6683 *) Reorganize and speed up MD5.
6684 [Andy Polyakov <appro@fy.chalmers.se>]
6687 [Richard Levitte <richard@levitte.org>]
6689 *) New option -out to asn1parse to allow the parsed structure to be
6690 output to a file. This is most useful when combined with the -strparse
6691 option to examine the output of things like OCTET STRINGS.
6694 *) Make SSL library a little more fool-proof by not requiring any longer
6695 that SSL_set_{accept,connect}_state be called before
6696 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6697 in many applications because usually everything *appeared* to work as
6698 intended anyway -- now it really works as intended).
6701 *) Move openssl.cnf out of lib/.
6704 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6705 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6706 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6707 [Ralf S. Engelschall]
6709 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6710 handle PKCS#7 enveloped data properly.
6711 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6713 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6714 copying pointers. The cert_st handling is changed by this in
6715 various ways (and thus what used to be known as ctx->default_cert
6716 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6717 any longer when s->cert does not give us what we need).
6718 ssl_cert_instantiate becomes obsolete by this change.
6719 As soon as we've got the new code right (possibly it already is?),
6720 we have solved a couple of bugs of the earlier code where s->cert
6721 was used as if it could not have been shared with other SSL structures.
6723 Note that using the SSL API in certain dirty ways now will result
6724 in different behaviour than observed with earlier library versions:
6725 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6726 does not influence s as it used to.
6728 In order to clean up things more thoroughly, inside SSL_SESSION
6729 we don't use CERT any longer, but a new structure SESS_CERT
6730 that holds per-session data (if available); currently, this is
6731 the peer's certificate chain and, for clients, the server's certificate
6732 and temporary key. CERT holds only those values that can have
6733 meaningful defaults in an SSL_CTX.
6736 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6737 from the internal representation. Various PKCS#7 fixes: remove some
6738 evil casts and set the enc_dig_alg field properly based on the signing
6742 *) Allow PKCS#12 password to be set from the command line or the
6743 environment. Let 'ca' get its config file name from the environment
6744 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6748 *) Allow certificate policies extension to use an IA5STRING for the
6749 organization field. This is contrary to the PKIX definition but
6750 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6754 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6755 without disallowing inline assembler and the like for non-pedantic builds.
6758 *) Support Borland C++ builder.
6759 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6764 *) SHA-1 cleanups and performance enhancements.
6765 [Andy Polyakov <appro@fy.chalmers.se>]
6767 *) Sparc v8plus assembler for the bignum library.
6768 [Andy Polyakov <appro@fy.chalmers.se>]
6770 *) Accept any -xxx and +xxx compiler options in Configure.
6773 *) Update HPUX configuration.
6776 *) Add missing sk_<type>_unshift() function to safestack.h
6777 [Ralf S. Engelschall]
6779 *) New function SSL_CTX_use_certificate_chain_file that sets the
6780 "extra_cert"s in addition to the certificate. (This makes sense
6781 only for "PEM" format files, as chains as a whole are not
6785 *) Support verify_depth from the SSL API.
6786 x509_vfy.c had what can be considered an off-by-one-error:
6787 Its depth (which was not part of the external interface)
6788 was actually counting the number of certificates in a chain;
6789 now it really counts the depth.
6792 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6793 instead of X509err, which often resulted in confusing error
6794 messages since the error codes are not globally unique
6795 (e.g. an alleged error in ssl3_accept when a certificate
6796 didn't match the private key).
6798 *) New function SSL_CTX_set_session_id_context that allows to set a default
6799 value (so that you don't need SSL_set_session_id_context for each
6800 connection using the SSL_CTX).
6803 *) OAEP decoding bug fix.
6806 *) Support INSTALL_PREFIX for package builders, as proposed by
6810 *) New Configure options "threads" and "no-threads". For systems
6811 where the proper compiler options are known (currently Solaris
6812 and Linux), "threads" is the default.
6815 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6818 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6819 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6820 such as /usr/local/bin.
6823 *) "make linux-shared" to build shared libraries.
6824 [Niels Poppe <niels@netbox.org>]
6826 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6829 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6830 extension adding in x509 utility.
6833 *) Remove NOPROTO sections and error code comments.
6836 *) Partial rewrite of the DEF file generator to now parse the ANSI
6840 *) New Configure options --prefix=DIR and --openssldir=DIR.
6843 *) Complete rewrite of the error code script(s). It is all now handled
6844 by one script at the top level which handles error code gathering,
6845 header rewriting and C source file generation. It should be much better
6846 than the old method: it now uses a modified version of Ulf's parser to
6847 read the ANSI prototypes in all header files (thus the old K&R definitions
6848 aren't needed for error creation any more) and do a better job of
6849 translating function codes into names. The old 'ASN1 error code imbedded
6850 in a comment' is no longer necessary and it doesn't use .err files which
6851 have now been deleted. Also the error code call doesn't have to appear all
6852 on one line (which resulted in some large lines...).
6855 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6858 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6859 0 (which usually indicates a closed connection), but continue reading.
6862 *) Fix some race conditions.
6865 *) Add support for CRL distribution points extension. Add Certificate
6866 Policies and CRL distribution points documentation.
6869 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6872 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6873 8 of keying material. Merlin has also confirmed interop with this fix
6874 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6875 [Merlin Hughes <merlin@baltimore.ie>]
6877 *) Fix lots of warnings.
6878 [Richard Levitte <levitte@stacken.kth.se>]
6880 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6881 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6882 [Richard Levitte <levitte@stacken.kth.se>]
6884 *) Fix problems with sizeof(long) == 8.
6885 [Andy Polyakov <appro@fy.chalmers.se>]
6887 *) Change functions to ANSI C.
6890 *) Fix typos in error codes.
6891 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6893 *) Remove defunct assembler files from Configure.
6896 *) SPARC v8 assembler BIGNUM implementation.
6897 [Andy Polyakov <appro@fy.chalmers.se>]
6899 *) Support for Certificate Policies extension: both print and set.
6900 Various additions to support the r2i method this uses.
6903 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6904 return a const string when you are expecting an allocated buffer.
6907 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6908 types DirectoryString and DisplayText.
6911 *) Add code to allow r2i extensions to access the configuration database,
6912 add an LHASH database driver and add several ctx helper functions.
6915 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6916 fail when they extended the size of a BIGNUM.
6919 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6920 support typesafe stack.
6923 *) Fix typo in SSL_[gs]et_options().
6924 [Nils Frostberg <nils@medcom.se>]
6926 *) Delete various functions and files that belonged to the (now obsolete)
6927 old X509V3 handling code.
6930 *) New Configure option "rsaref".
6933 *) Don't auto-generate pem.h.
6936 *) Introduce type-safe ASN.1 SETs.
6939 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6940 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6942 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6943 that links with OpenSSL (well at least cause lots of warnings), but fear
6944 not: the conversion is trivial, and it eliminates loads of evil casts. A
6945 few STACKed things have been converted already. Feel free to convert more.
6946 In the fullness of time, I'll do away with the STACK type altogether.
6949 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6950 specified in <certfile> by updating the entry in the index.txt file.
6951 This way one no longer has to edit the index.txt file manually for
6952 revoking a certificate. The -revoke option does the gory details now.
6953 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6955 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6956 `-text' option at all and this way the `-noout -text' combination was
6957 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6958 [Ralf S. Engelschall]
6960 *) Make sure a corresponding plain text error message exists for the
6961 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6962 verify callback function determined that a certificate was revoked.
6963 [Ralf S. Engelschall]
6965 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6966 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6967 all available cipers including rc5, which was forgotten until now.
6968 In order to let the testing shell script know which algorithms
6969 are available, a new (up to now undocumented) command
6970 "openssl list-cipher-commands" is used.
6973 *) Bugfix: s_client occasionally would sleep in select() when
6974 it should have checked SSL_pending() first.
6977 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6978 the raw DSA values prior to ASN.1 encoding.
6981 *) Tweaks to Configure
6982 [Niels Poppe <niels@netbox.org>]
6984 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6988 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6991 *) New config option to avoid instructions that are illegal on the 80386.
6992 The default code is faster, but requires at least a 486.
6995 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6996 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6997 same as SSL2_VERSION anyway.
7000 *) New "-showcerts" option for s_client.
7003 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7004 application. Various cleanups and fixes.
7007 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7008 modify error routines to work internally. Add error codes and PBE init
7009 to library startup routines.
7012 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7013 packing functions to asn1 and evp. Changed function names and error
7014 codes along the way.
7017 *) PKCS12 integration: and so it begins... First of several patches to
7018 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7019 objects to objects.h
7022 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7023 and display support for Thawte strong extranet extension.
7026 *) Add LinuxPPC support.
7027 [Jeff Dubrule <igor@pobox.org>]
7029 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7030 bn_div_words in alpha.s.
7031 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7033 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7034 OAEP isn't supported when OpenSSL is built with RSAref.
7035 [Ulf Moeller <ulf@fitug.de>]
7037 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7038 so they no longer are missing under -DNOPROTO.
7039 [Soren S. Jorvang <soren@t.dk>]
7042 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7044 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7045 doesn't work when the session is reused. Coming soon!
7048 *) Fix a security hole, that allows sessions to be reused in the wrong
7049 context thus bypassing client cert protection! All software that uses
7050 client certs and session caches in multiple contexts NEEDS PATCHING to
7051 allow session reuse! A fuller solution is in the works.
7052 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7054 *) Some more source tree cleanups (removed obsolete files
7055 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7056 permission on "config" script to be executable) and a fix for the INSTALL
7058 [Ulf Moeller <ulf@fitug.de>]
7060 *) Remove some legacy and erroneous uses of malloc, free instead of
7062 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7064 *) Make rsa_oaep_test return non-zero on error.
7065 [Ulf Moeller <ulf@fitug.de>]
7067 *) Add support for native Solaris shared libraries. Configure
7068 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7069 if someone would make that last step automatic.
7070 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7072 *) ctx_size was not built with the right compiler during "make links". Fixed.
7075 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7076 except NULL ciphers". This means the default cipher list will no longer
7077 enable NULL ciphers. They need to be specifically enabled e.g. with
7078 the string "DEFAULT:eNULL".
7081 *) Fix to RSA private encryption routines: if p < q then it would
7082 occasionally produce an invalid result. This will only happen with
7083 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7086 *) Be less restrictive and allow also `perl util/perlpath.pl
7087 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7088 because this way one can also use an interpreter named `perl5' (which is
7089 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7090 installed as `perl').
7091 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7093 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7094 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7096 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7097 advapi32.lib to Win32 build and change the pem test comparision
7098 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7099 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7100 and crypto/des/ede_cbcm_enc.c.
7103 *) DES quad checksum was broken on big-endian architectures. Fixed.
7106 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7107 Win32 test batch file so it (might) work again. The Win32 test batch file
7108 is horrible: I feel ill....
7111 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7112 in e_os.h. Audit of header files to check ANSI and non ANSI
7113 sections: 10 functions were absent from non ANSI section and not exported
7114 from Windows DLLs. Fixed up libeay.num for new functions.
7117 *) Make `openssl version' output lines consistent.
7118 [Ralf S. Engelschall]
7120 *) Fix Win32 symbol export lists for BIO functions: Added
7121 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7122 to ms/libeay{16,32}.def.
7123 [Ralf S. Engelschall]
7125 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7126 fine under Unix and passes some trivial tests I've now added. But the
7127 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7128 added to make sure no one expects that this stuff really works in the
7129 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7130 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7132 [Ralf S. Engelschall]
7134 *) Fix the generation of two part addresses in perl.
7135 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7137 *) Add config entry for Linux on MIPS.
7138 [John Tobey <jtobey@channel1.com>]
7140 *) Make links whenever Configure is run, unless we are on Windoze.
7143 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7144 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7148 *) Add a useful kludge to allow package maintainers to specify compiler and
7149 other platforms details on the command line without having to patch the
7150 Configure script everytime: One now can use ``perl Configure
7151 <id>:<details>'', i.e. platform ids are allowed to have details appended
7152 to them (seperated by colons). This is treated as there would be a static
7153 pre-configured entry in Configure's %table under key <id> with value
7154 <details> and ``perl Configure <id>'' is called. So, when you want to
7155 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7156 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7157 now, which overrides the FreeBSD-elf entry on-the-fly.
7158 [Ralf S. Engelschall]
7160 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7163 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7164 on the `perl Configure ...' command line. This way one can compile
7165 OpenSSL libraries with Position Independent Code (PIC) which is needed
7166 for linking it into DSOs.
7167 [Ralf S. Engelschall]
7169 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7173 *) Cleaned up the LICENSE document: The official contact for any license
7174 questions now is the OpenSSL core team under openssl-core@openssl.org.
7175 And add a paragraph about the dual-license situation to make sure people
7176 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7177 to the OpenSSL toolkit.
7178 [Ralf S. Engelschall]
7180 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7181 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7182 Additonally cleaned up the `make links' target: Remove unnecessary
7183 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7184 to speed processing and no longer clutter the display with confusing
7185 stuff. Instead only the actually done links are displayed.
7186 [Ralf S. Engelschall]
7188 *) Permit null encryption ciphersuites, used for authentication only. It used
7189 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7190 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7194 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7195 signed attributes when verifying signatures (this would break them),
7196 the detached data encoding was wrong and public keys obtained using
7197 X509_get_pubkey() weren't freed.
7200 *) Add text documentation for the BUFFER functions. Also added a work around
7201 to a Win95 console bug. This was triggered by the password read stuff: the
7202 last character typed gets carried over to the next fread(). If you were
7203 generating a new cert request using 'req' for example then the last
7204 character of the passphrase would be CR which would then enter the first
7208 *) Added the new `Includes OpenSSL Cryptography Software' button as
7209 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7210 button and can be used by applications based on OpenSSL to show the
7211 relationship to the OpenSSL project.
7212 [Ralf S. Engelschall]
7214 *) Remove confusing variables in function signatures in files
7215 ssl/ssl_lib.c and ssl/ssl.h.
7216 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7218 *) Don't install bss_file.c under PREFIX/include/
7219 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7221 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7222 functions that return function pointers and has support for NT specific
7223 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7224 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7225 unsigned to signed types: this was killing the Win32 compile.
7228 *) Add new certificate file to stack functions,
7229 SSL_add_dir_cert_subjects_to_stack() and
7230 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7231 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7232 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7233 This means that Apache-SSL and similar packages don't have to mess around
7234 to add as many CAs as they want to the preferred list.
7237 *) Experiment with doxygen documentation. Currently only partially applied to
7239 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7240 openssl.doxy as the configuration file.
7243 *) Get rid of remaining C++-style comments which strict C compilers hate.
7244 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7246 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7247 compiled in by default: it has problems with large keys.
7250 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7251 DH private keys and/or callback functions which directly correspond to
7252 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7253 is needed for applications which have to configure certificates on a
7254 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7256 For the RSA certificate situation is makes no difference, but
7257 for the DSA certificate situation this fixes the "no shared cipher"
7258 problem where the OpenSSL cipher selection procedure failed because the
7259 temporary keys were not overtaken from the context and the API provided
7260 no way to reconfigure them.
7261 The new functions now let applications reconfigure the stuff and they
7262 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7263 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7264 non-public-API function ssl_cert_instantiate() is used as a helper
7265 function and also to reduce code redundancy inside ssl_rsa.c.
7266 [Ralf S. Engelschall]
7268 *) Move s_server -dcert and -dkey options out of the undocumented feature
7269 area because they are useful for the DSA situation and should be
7270 recognized by the users.
7271 [Ralf S. Engelschall]
7273 *) Fix the cipher decision scheme for export ciphers: the export bits are
7274 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7275 SSL_EXP_MASK. So, the original variable has to be used instead of the
7276 already masked variable.
7277 [Richard Levitte <levitte@stacken.kth.se>]
7279 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7280 [Richard Levitte <levitte@stacken.kth.se>]
7282 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7283 from `int' to `unsigned int' because it's a length and initialized by
7284 EVP_DigestFinal() which expects an `unsigned int *'.
7285 [Richard Levitte <levitte@stacken.kth.se>]
7287 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7288 script. Instead use the usual Shell->Perl transition trick.
7289 [Ralf S. Engelschall]
7291 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7292 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7293 -noout -modulus' as it's already the case for `openssl rsa -noout
7294 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7295 currently the public key is printed (a decision which was already done by
7296 `openssl dsa -modulus' in the past) which serves a similar purpose.
7297 Additionally the NO_RSA no longer completely removes the whole -modulus
7298 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7300 [Ralf S. Engelschall]
7302 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7303 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7304 [Arne Ansper <arne@ats.cyber.ee>]
7306 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7307 to be added. Now both 'req' and 'ca' can use new objects defined in the
7311 *) Add cool BIO that does syslog (or event log on NT).
7312 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7314 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7315 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7316 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7317 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7320 *) Add preliminary config info for new extension code.
7323 *) Make RSA_NO_PADDING really use no padding.
7324 [Ulf Moeller <ulf@fitug.de>]
7326 *) Generate errors when private/public key check is done.
7329 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7330 for some CRL extensions and new objects added.
7333 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7334 key usage extension and fuller support for authority key id.
7337 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7338 padding method for RSA, which is recommended for new applications in PKCS
7339 #1 v2.0 (RFC 2437, October 1998).
7340 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7341 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7342 against Bleichbacher's attack on RSA.
7343 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7346 *) Updates to the new SSL compression code
7347 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7349 *) Fix so that the version number in the master secret, when passed
7350 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7351 (because the server will not accept higher), that the version number
7352 is 0x03,0x01, not 0x03,0x00
7353 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7355 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7356 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7357 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7360 *) Support for RAW extensions where an arbitrary extension can be
7361 created by including its DER encoding. See apps/openssl.cnf for
7365 *) Make sure latest Perl versions don't interpret some generated C array
7366 code as Perl array code in the crypto/err/err_genc.pl script.
7367 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7369 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7370 not many people have the assembler. Various Win32 compilation fixes and
7371 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7375 *) Modify configure script 'Configure' to automatically create crypto/date.h
7376 file under Win32 and also build pem.h from pem.org. New script
7377 util/mkfiles.pl to create the MINFO file on environments that can't do a
7378 'make files': perl util/mkfiles.pl >MINFO should work.
7381 *) Major rework of DES function declarations, in the pursuit of correctness
7382 and purity. As a result, many evil casts evaporated, and some weirdness,
7383 too. You may find this causes warnings in your code. Zapping your evil
7384 casts will probably fix them. Mostly.
7387 *) Fix for a typo in asn1.h. Bug fix to object creation script
7388 obj_dat.pl. It considered a zero in an object definition to mean
7389 "end of object": none of the objects in objects.h have any zeros
7390 so it wasn't spotted.
7391 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7393 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7394 Masking (CBCM). In the absence of test vectors, the best I have been able
7395 to do is check that the decrypt undoes the encrypt, so far. Send me test
7396 vectors if you have them.
7399 *) Correct calculation of key length for export ciphers (too much space was
7400 allocated for null ciphers). This has not been tested!
7403 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7404 message is now correct (it understands "crypto" and "ssl" on its
7405 command line). There is also now an "update" option. This will update
7406 the util/ssleay.num and util/libeay.num files with any new functions.
7408 perl util/mkdef.pl crypto ssl update
7409 it will update them.
7412 *) Overhauled the Perl interface (perl/*):
7413 - ported BN stuff to OpenSSL's different BN library
7414 - made the perl/ source tree CVS-aware
7415 - renamed the package from SSLeay to OpenSSL (the files still contain
7416 their history because I've copied them in the repository)
7417 - removed obsolete files (the test scripts will be replaced
7418 by better Test::Harness variants in the future)
7419 [Ralf S. Engelschall]
7421 *) First cut for a very conservative source tree cleanup:
7422 1. merge various obsolete readme texts into doc/ssleay.txt
7423 where we collect the old documents and readme texts.
7424 2. remove the first part of files where I'm already sure that we no
7425 longer need them because of three reasons: either they are just temporary
7426 files which were left by Eric or they are preserved original files where
7427 I've verified that the diff is also available in the CVS via "cvs diff
7428 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7429 the crypto/md/ stuff).
7430 [Ralf S. Engelschall]
7432 *) More extension code. Incomplete support for subject and issuer alt
7433 name, issuer and authority key id. Change the i2v function parameters
7434 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7435 what that's for :-) Fix to ASN1 macro which messed up
7436 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7439 *) Preliminary support for ENUMERATED type. This is largely copied from the
7443 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7444 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7446 *) Make sure `make rehash' target really finds the `openssl' program.
7447 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7449 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7450 like to hear about it if this slows down other processors.
7453 *) Add CygWin32 platform information to Configure script.
7454 [Alan Batie <batie@aahz.jf.intel.com>]
7456 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7457 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7459 *) New program nseq to manipulate netscape certificate sequences
7462 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7466 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7467 but the BN code had some problems that would cause failures when
7468 doing certificate verification and some other functions.
7469 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7471 *) Add ASN1 and PEM code to support netscape certificate sequences.
7474 *) Add ASN1 and PEM code to support netscape certificate sequences.
7477 *) Add several PKIX and private extended key usage OIDs.
7480 *) Modify the 'ca' program to handle the new extension code. Modify
7481 openssl.cnf for new extension format, add comments.
7484 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7485 and add a sample to openssl.cnf so req -x509 now adds appropriate
7489 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7490 error code, add initial support to X509_print() and x509 application.
7493 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7494 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7495 stuff is currently isolated and isn't even compiled yet.
7498 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7499 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7500 Removed the versions check from X509 routines when loading extensions:
7501 this allows certain broken certificates that don't set the version
7502 properly to be processed.
7505 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7506 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7507 can still be regenerated with "make depend".
7510 *) Spelling mistake in C version of CAST-128.
7511 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7513 *) Changes to the error generation code. The perl script err-code.pl
7514 now reads in the old error codes and retains the old numbers, only
7515 adding new ones if necessary. It also only changes the .err files if new
7516 codes are added. The makefiles have been modified to only insert errors
7517 when needed (to avoid needlessly modifying header files). This is done
7518 by only inserting errors if the .err file is newer than the auto generated
7519 C file. To rebuild all the error codes from scratch (the old behaviour)
7520 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7521 or delete all the .err files.
7524 *) CAST-128 was incorrectly implemented for short keys. The C version has
7525 been fixed, but is untested. The assembler versions are also fixed, but
7526 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7527 to regenerate it if needed.
7528 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7529 Hagino <itojun@kame.net>]
7531 *) File was opened incorrectly in randfile.c.
7532 [Ulf Möller <ulf@fitug.de>]
7534 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7535 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7536 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7537 al: it's just almost always a UTCTime. Note this patch adds new error
7538 codes so do a "make errors" if there are problems.
7541 *) Correct Linux 1 recognition in config.
7542 [Ulf Möller <ulf@fitug.de>]
7544 *) Remove pointless MD5 hash when using DSA keys in ca.
7545 [Anonymous <nobody@replay.com>]
7547 *) Generate an error if given an empty string as a cert directory. Also
7548 generate an error if handed NULL (previously returned 0 to indicate an
7549 error, but didn't set one).
7550 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7552 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7555 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7556 parameters. This was causing a warning which killed off the Win32 compile.
7559 *) Remove C++ style comments from crypto/bn/bn_local.h.
7560 [Neil Costigan <neil.costigan@celocom.com>]
7562 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7563 based on a text string, looking up short and long names and finally
7564 "dot" format. The "dot" format stuff didn't work. Added new function
7565 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7566 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7567 OID is not part of the table.
7570 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7571 X509_LOOKUP_by_alias().
7574 *) Sort openssl functions by name.
7577 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7578 encryption from sample DSA keys (in case anyone is interested the password
7582 *) Make _all_ *_free functions accept a NULL pointer.
7583 [Frans Heymans <fheymans@isaserver.be>]
7585 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7587 [Anonymous <nobody@replay.com>]
7589 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7590 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7592 *) Don't blow it for numeric -newkey arguments to apps/req.
7593 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7595 *) Temp key "for export" tests were wrong in s3_srvr.c.
7596 [Anonymous <nobody@replay.com>]
7598 *) Add prototype for temp key callback functions
7599 SSL_CTX_set_tmp_{rsa,dh}_callback().
7602 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7603 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7606 *) X509_name_add_entry() freed the wrong thing after an error.
7607 [Arne Ansper <arne@ats.cyber.ee>]
7609 *) rsa_eay.c would attempt to free a NULL context.
7610 [Arne Ansper <arne@ats.cyber.ee>]
7612 *) BIO_s_socket() had a broken should_retry() on Windoze.
7613 [Arne Ansper <arne@ats.cyber.ee>]
7615 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7616 [Arne Ansper <arne@ats.cyber.ee>]
7618 *) Make sure the already existing X509_STORE->depth variable is initialized
7619 in X509_STORE_new(), but document the fact that this variable is still
7620 unused in the certificate verification process.
7621 [Ralf S. Engelschall]
7623 *) Fix the various library and apps files to free up pkeys obtained from
7624 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7627 *) Fix reference counting in X509_PUBKEY_get(). This makes
7628 demos/maurice/example2.c work, amongst others, probably.
7629 [Steve Henson and Ben Laurie]
7631 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7632 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7633 are no longer created. This way we have a single and consistent command
7634 line interface `openssl <command>', similar to `cvs <command>'.
7635 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7637 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7638 BIT STRING wrapper always have zero unused bits.
7641 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7644 *) Make the top-level INSTALL documentation easier to understand.
7647 *) Makefiles updated to exit if an error occurs in a sub-directory
7648 make (including if user presses ^C) [Paul Sutton]
7650 *) Make Montgomery context stuff explicit in RSA data structure.
7653 *) Fix build order of pem and err to allow for generated pem.h.
7656 *) Fix renumbering bug in X509_NAME_delete_entry().
7659 *) Enhanced the err-ins.pl script so it makes the error library number
7660 global and can add a library name. This is needed for external ASN1 and
7661 other error libraries.
7664 *) Fixed sk_insert which never worked properly.
7667 *) Fix ASN1 macros so they can handle indefinite length construted
7668 EXPLICIT tags. Some non standard certificates use these: they can now
7672 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7673 into a single doc/ssleay.txt bundle. This way the information is still
7674 preserved but no longer messes up this directory. Now it's new room for
7675 the new set of documenation files.
7676 [Ralf S. Engelschall]
7678 *) SETs were incorrectly DER encoded. This was a major pain, because they
7679 shared code with SEQUENCEs, which aren't coded the same. This means that
7680 almost everything to do with SETs or SEQUENCEs has either changed name or
7681 number of arguments.
7682 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7684 *) Fix test data to work with the above.
7687 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7688 was already fixed by Eric for 0.9.1 it seems.
7689 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7691 *) Autodetect FreeBSD3.
7694 *) Fix various bugs in Configure. This affects the following platforms:
7698 unixware-2.0-pentium
7702 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7703 before they are needed.
7706 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7710 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7712 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7713 changed SSLeay to OpenSSL in version strings.
7714 [Ralf S. Engelschall]
7716 *) Some fixups to the top-level documents.
7719 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7720 because the symlink to include/ was missing.
7721 [Ralf S. Engelschall]
7723 *) Incorporated the popular no-RSA/DSA-only patches
7724 which allow to compile a RSA-free SSLeay.
7725 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7727 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7728 when "ssleay" is still not found.
7729 [Ralf S. Engelschall]
7731 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7732 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7734 *) Updated the README file.
7735 [Ralf S. Engelschall]
7737 *) Added various .cvsignore files in the CVS repository subdirs
7738 to make a "cvs update" really silent.
7739 [Ralf S. Engelschall]
7741 *) Recompiled the error-definition header files and added
7742 missing symbols to the Win32 linker tables.
7743 [Ralf S. Engelschall]
7745 *) Cleaned up the top-level documents;
7746 o new files: CHANGES and LICENSE
7747 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7748 o merged COPYRIGHT into LICENSE
7749 o removed obsolete TODO file
7750 o renamed MICROSOFT to INSTALL.W32
7751 [Ralf S. Engelschall]
7753 *) Removed dummy files from the 0.9.1b source tree:
7754 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7755 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7756 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7757 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7758 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7759 [Ralf S. Engelschall]
7761 *) Added various platform portability fixes.
7764 *) The Genesis of the OpenSSL rpject:
7765 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7766 Young and Tim J. Hudson created while they were working for C2Net until
7768 [The OpenSSL Project]
7771 Changes between 0.9.0b and 0.9.1b [not released]
7773 *) Updated a few CA certificates under certs/
7776 *) Changed some BIGNUM api stuff.
7779 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7780 DGUX x86, Linux Alpha, etc.
7783 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7784 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7788 *) Add -strparse option to asn1pars program which parses nested
7790 [Dr Stephen Henson <shenson@bigfoot.com>]
7792 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7795 *) DSA fix for "ca" program.
7798 *) Added "-genkey" option to "dsaparam" program.
7801 *) Added RIPE MD160 (rmd160) message digest.
7804 *) Added -a (all) option to "ssleay version" command.
7807 *) Added PLATFORM define which is the id given to Configure.
7810 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7813 *) Extended the ASN.1 parser routines.
7816 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7819 *) Added a BN_CTX to the BN library.
7822 *) Fixed the weak key values in DES library
7825 *) Changed API in EVP library for cipher aliases.
7828 *) Added support for RC2/64bit cipher.
7831 *) Converted the lhash library to the crypto/mem.c functions.
7834 *) Added more recognized ASN.1 object ids.
7837 *) Added more RSA padding checks for SSL/TLS.
7840 *) Added BIO proxy/filter functionality.
7843 *) Added extra_certs to SSL_CTX which can be used
7844 send extra CA certificates to the client in the CA cert chain sending
7845 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7848 *) Now Fortezza is denied in the authentication phase because
7849 this is key exchange mechanism is not supported by SSLeay at all.
7852 *) Additional PKCS1 checks.
7855 *) Support the string "TLSv1" for all TLS v1 ciphers.
7858 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7859 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7862 *) Fixed a few memory leaks.
7865 *) Fixed various code and comment typos.
7868 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7869 bytes sent in the client random.
7870 [Edward Bishop <ebishop@spyglass.com>]