5 Changes between 0.9.8a and 0.9.9 [xx XXX xxxx]
7 *) New option SSL_OP_NO_COMP to disable use of compression selectively
8 in SSL structures. New SSL ctrl to set maximum send fragment size.
9 Save memory by seeting the I/O buffer sizes dynamically instead of
10 using the maximum available value.
13 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
14 in addition to the text details.
17 *) Very, very preliminary EXPERIMENTAL support for printing of general
18 ASN1 structures. This currently produces rather ugly output and doesn't
19 handle several customised structures at all.
22 *) Integrated support for PVK file format and some related formats such
23 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
24 these in the 'rsa' and 'dsa' utilities.
27 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
30 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
31 place for the (very old) "NETSCAPE" format certificates which are now
32 handled using new ASN1 code equivalents.
35 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
36 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
37 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
40 *) Modify CRL distribution points extension code to print out previously
41 unsupported fields. Enhance extension setting code to allow setting of
45 *) Add print and set support for Issuing Distribution Point CRL extension.
48 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
50 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
51 (part of SSL_OP_ALL). This option used to disable the
52 countermeasure against man-in-the-middle protocol-version
53 rollback in the SSL 2.0 server implementation, which is a bad
56 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
57 for Information Security, National Institute of Advanced Industrial
58 Science and Technology [AIST], Japan)]
60 *) Add two function to clear and return the verify parameter flags.
63 *) Keep cipherlists sorted in the source instead of sorting them at
64 runtime, thus removing the need for a lock.
67 *) Avoid some small subgroup attacks in Diffie-Hellman.
68 [Nick Mathewson and Ben Laurie]
70 *) Add functions for well-known primes.
73 *) Extended Windows CE support.
74 [Satoshi Nakamura and Andy Polyakov]
76 *) Initialize SSL_METHOD structures at compile time instead of during
77 runtime, thus removing the need for a lock.
80 *) Make PKCS7_decrypt() work even if no certificate is supplied by
81 attempting to decrypt each encrypted key in turn. Add support to
85 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
87 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
90 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
91 key into the same file any more.
94 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
97 *) Add -utf8 command line and config file option to 'ca'.
98 [Stefan <stf@udoma.org]
100 *) Removed the macro des_crypt(), as it seems to conflict with some
101 libraries. Use DES_crypt().
104 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
105 involves renaming the source and generated shared-libs for
106 both. The engines will accept the corrected or legacy ids
107 ('ncipher' and '4758_cca' respectively) when binding. NB,
108 this only applies when building 'shared'.
109 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
111 *) Add attribute functions to EVP_PKEY structure. Modify
112 PKCS12_create() to recognize a CSP name attribute and
113 use it. Make -CSP option work again in pkcs12 utility.
116 *) Add new functionality to the bn blinding code:
117 - automatic re-creation of the BN_BLINDING parameters after
118 a fixed number of uses (currently 32)
119 - add new function for parameter creation
120 - introduce flags to control the update behaviour of the
121 BN_BLINDING parameters
122 - hide BN_BLINDING structure
123 Add a second BN_BLINDING slot to the RSA structure to improve
124 performance when a single RSA object is shared among several
128 *) Add support for DTLS.
129 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
131 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
132 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
135 *) Remove buggy and incompletet DH cert support from
136 ssl/ssl_rsa.c and ssl/s3_both.c
139 *) Use SHA-1 instead of MD5 as the default digest algorithm for
140 the apps/openssl applications.
143 *) Compile clean with "-Wall -Wmissing-prototypes
144 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
145 DEBUG_SAFESTACK must also be set.
148 *) Change ./Configure so that certain algorithms can be disabled by default.
149 The new counterpiece to "no-xxx" is "enable-xxx".
151 The patented RC5 and MDC2 algorithms will now be disabled unless
152 "enable-rc5" and "enable-mdc2", respectively, are specified.
154 (IDEA remains enabled despite being patented. This is because IDEA
155 is frequently required for interoperability, and there is no license
156 fee for non-commercial use. As before, "no-idea" can be used to
157 avoid this algorithm.)
161 *) Add processing of proxy certificates (see RFC 3820). This work was
162 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
163 EGEE (Enabling Grids for E-science in Europe).
166 *) RC4 performance overhaul on modern architectures/implementations, such
167 as Intel P4, IA-64 and AMD64.
170 *) New utility extract-section.pl. This can be used specify an alternative
171 section number in a pod file instead of having to treat each file as
172 a separate case in Makefile. This can be done by adding two lines to the
175 =for comment openssl_section:XXX
177 The blank line is mandatory.
181 *) New arguments -certform, -keyform and -pass for s_client and s_server
182 to allow alternative format key and certificate files and passphrase
186 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
187 update associated structures and add various utility functions.
189 Add new policy related verify parameters, include policy checking in
190 standard verify code. Enhance 'smime' application with extra parameters
191 to support policy checking and print out.
194 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
195 Nehemiah processors. These extensions support AES encryption in hardware
196 as well as RNG (though RNG support is currently disabled).
197 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
199 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
202 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
203 [Andy Polyakov and a number of other people]
205 *) Improved PowerPC platform support. Most notably BIGNUM assembler
206 implementation contributed by IBM.
207 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
209 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
210 exponent rather than 'unsigned long'. There is a corresponding change to
211 the new 'rsa_keygen' element of the RSA_METHOD structure.
212 [Jelte Jansen, Geoff Thorpe]
214 *) Functionality for creating the initial serial number file is now
215 moved from CA.pl to the 'ca' utility with a new option -create_serial.
217 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
218 number file to 1, which is bound to cause problems. To avoid
219 the problems while respecting compatibility between different 0.9.7
220 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
221 CA.pl for serial number initialization. With the new release 0.9.8,
222 we can fix the problem directly in the 'ca' utility.)
225 *) Reduced header interdepencies by declaring more opaque objects in
226 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
227 give fewer recursive includes, which could break lazy source code - so
228 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
229 developers should define this symbol when building and using openssl to
230 ensure they track the recommended behaviour, interfaces, [etc], but
231 backwards-compatible behaviour prevails when this isn't defined.
234 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
237 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
238 This will generate a random key of the appropriate length based on the
239 cipher context. The EVP_CIPHER can provide its own random key generation
240 routine to support keys of a specific form. This is used in the des and
241 3des routines to generate a key of the correct parity. Update S/MIME
242 code to use new functions and hence generate correct parity DES keys.
243 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
244 valid (weak or incorrect parity).
247 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
248 as looking them up. This is useful when the verified structure may contain
249 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
250 present unless the new PKCS7_NO_CRL flag is asserted.
253 *) Extend ASN1 oid configuration module. It now additionally accepts the
256 shortName = some long name, 1.2.3.4
259 *) Reimplemented the BN_CTX implementation. There is now no more static
260 limitation on the number of variables it can handle nor the depth of the
261 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
262 information can now expand as required, and rather than having a single
263 static array of bignums, BN_CTX now uses a linked-list of such arrays
264 allowing it to expand on demand whilst maintaining the usefulness of
268 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
269 to allow all RSA operations to function using a single BN_CTX.
272 *) Preliminary support for certificate policy evaluation and checking. This
273 is initially intended to pass the tests outlined in "Conformance Testing
274 of Relying Party Client Certificate Path Processing Logic" v1.07.
277 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
278 remained unused and not that useful. A variety of other little bignum
279 tweaks and fixes have also been made continuing on from the audit (see
283 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
284 associated ASN1, EVP and SSL functions and old ASN1 macros.
287 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
288 and this should never fail. So the return value from the use of
289 BN_set_word() (which can fail due to needless expansion) is now deprecated;
290 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
293 *) BN_CTX_get() should return zero-valued bignums, providing the same
294 initialised value as BN_new().
295 [Geoff Thorpe, suggested by Ulf Möller]
297 *) Support for inhibitAnyPolicy certificate extension.
300 *) An audit of the BIGNUM code is underway, for which debugging code is
301 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
302 is considered valid when processing BIGNUMs, and causes execution to
303 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
304 further steps are taken to deliberately pollute unused data in BIGNUM
305 structures to try and expose faulty code further on. For now, openssl will
306 (in its default mode of operation) continue to tolerate the inconsistent
307 forms that it has tolerated in the past, but authors and packagers should
308 consider trying openssl and their own applications when compiled with
309 these debugging symbols defined. It will help highlight potential bugs in
310 their own code, and will improve the test coverage for OpenSSL itself. At
311 some point, these tighter rules will become openssl's default to improve
312 maintainability, though the assert()s and other overheads will remain only
313 in debugging configurations. See bn.h for more details.
314 [Geoff Thorpe, Nils Larsch, Ulf Möller]
316 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
317 that can only be obtained through BN_CTX_new() (which implicitly
318 initialises it). The presence of this function only made it possible
319 to overwrite an existing structure (and cause memory leaks).
322 *) Because of the callback-based approach for implementing LHASH as a
323 template type, lh_insert() adds opaque objects to hash-tables and
324 lh_doall() or lh_doall_arg() are typically used with a destructor callback
325 to clean up those corresponding objects before destroying the hash table
326 (and losing the object pointers). So some over-zealous constifications in
327 LHASH have been relaxed so that lh_insert() does not take (nor store) the
328 objects as "const" and the lh_doall[_arg] callback wrappers are not
329 prototyped to have "const" restrictions on the object pointers they are
330 given (and so aren't required to cast them away any more).
333 *) The tmdiff.h API was so ugly and minimal that our own timing utility
334 (speed) prefers to use its own implementation. The two implementations
335 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
336 its object type properly exposed (MS_TM) instead of casting to/from "char
337 *". This may still change yet if someone realises MS_TM and "ms_time_***"
338 aren't necessarily the greatest nomenclatures - but this is what was used
339 internally to the implementation so I've used that for now.
342 *) Ensure that deprecated functions do not get compiled when
343 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
344 the self-tests were still using deprecated key-generation functions so
345 these have been updated also.
348 *) Reorganise PKCS#7 code to separate the digest location functionality
349 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
350 New function PKCS7_set_digest() to set the digest type for PKCS#7
351 digestedData type. Add additional code to correctly generate the
352 digestedData type and add support for this type in PKCS7 initialization
356 *) New function PKCS7_set0_type_other() this initializes a PKCS7
357 structure of type "other".
360 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
361 sure the loop does correctly stop and breaking ("division by zero")
362 modulus operations are not performed. The (pre-generated) prime
363 table crypto/bn/bn_prime.h was already correct, but it could not be
364 re-generated on some platforms because of the "division by zero"
365 situation in the script.
366 [Ralf S. Engelschall]
368 *) Update support for ECC-based TLS ciphersuites according to
369 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
370 SHA-1 now is only used for "small" curves (where the
371 representation of a field element takes up to 24 bytes); for
372 larger curves, the field element resulting from ECDH is directly
373 used as premaster secret.
374 [Douglas Stebila (Sun Microsystems Laboratories)]
376 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
377 curve secp160r1 to the tests.
378 [Douglas Stebila (Sun Microsystems Laboratories)]
380 *) Add the possibility to load symbols globally with DSO.
381 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
383 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
384 control of the error stack.
387 *) Add support for STORE in ENGINE.
390 *) Add the STORE type. The intention is to provide a common interface
391 to certificate and key stores, be they simple file-based stores, or
392 HSM-type store, or LDAP stores, or...
393 NOTE: The code is currently UNTESTED and isn't really used anywhere.
396 *) Add a generic structure called OPENSSL_ITEM. This can be used to
397 pass a list of arguments to any function as well as provide a way
398 for a function to pass data back to the caller.
401 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
402 works like BUF_strdup() but can be used to duplicate a portion of
403 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
407 *) Add the function sk_find_ex() which works like sk_find(), but will
408 return an index to an element even if an exact match couldn't be
409 found. The index is guaranteed to point at the element where the
410 searched-for key would be inserted to preserve sorting order.
413 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
414 takes an extra flags argument for optional functionality. Currently,
415 the following flags are defined:
417 OBJ_BSEARCH_VALUE_ON_NOMATCH
418 This one gets OBJ_bsearch_ex() to return a pointer to the first
419 element where the comparing function returns a negative or zero
422 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
423 This one gets OBJ_bsearch_ex() to return a pointer to the first
424 element where the comparing function returns zero. This is useful
425 if there are more than one element where the comparing function
429 *) Make it possible to create self-signed certificates with 'openssl ca'
430 in such a way that the self-signed certificate becomes part of the
431 CA database and uses the same mechanisms for serial number generation
432 as all other certificate signing. The new flag '-selfsign' enables
433 this functionality. Adapt CA.sh and CA.pl.in.
436 *) Add functionality to check the public key of a certificate request
437 against a given private. This is useful to check that a certificate
438 request can be signed by that key (self-signing).
441 *) Make it possible to have multiple active certificates with the same
442 subject in the CA index file. This is done only if the keyword
443 'unique_subject' is set to 'no' in the main CA section (default
444 if 'CA_default') of the configuration file. The value is saved
445 with the database itself in a separate index attribute file,
446 named like the index file with '.attr' appended to the name.
449 *) Generate muti valued AVAs using '+' notation in config files for
453 *) Support for nameConstraints certificate extension.
456 *) Support for policyConstraints certificate extension.
459 *) Support for policyMappings certificate extension.
462 *) Make sure the default DSA_METHOD implementation only uses its
463 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
464 and change its own handlers to be NULL so as to remove unnecessary
465 indirection. This lets alternative implementations fallback to the
466 default implementation more easily.
469 *) Support for directoryName in GeneralName related extensions
473 *) Make it possible to link applications using Makefile.shared.
474 Make that possible even when linking against static libraries!
477 *) Support for single pass processing for S/MIME signing. This now
478 means that S/MIME signing can be done from a pipe, in addition
479 cleartext signing (multipart/signed type) is effectively streaming
480 and the signed data does not need to be all held in memory.
482 This is done with a new flag PKCS7_STREAM. When this flag is set
483 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
484 is done after the data is output (and digests calculated) in
488 *) Add full support for -rpath/-R, both in shared libraries and
489 applications, at least on the platforms where it's known how
493 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
494 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
495 will now compute a table of multiples of the generator that
496 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
497 faster (notably in the case of a single point multiplication,
499 [Nils Larsch, Bodo Moeller]
501 *) IPv6 support for certificate extensions. The various extensions
502 which use the IP:a.b.c.d can now take IPv6 addresses using the
503 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
507 *) Added an ENGINE that implements RSA by performing private key
508 exponentiations with the GMP library. The conversions to and from
509 GMP's mpz_t format aren't optimised nor are any montgomery forms
510 cached, and on x86 it appears OpenSSL's own performance has caught up.
511 However there are likely to be other architectures where GMP could
512 provide a boost. This ENGINE is not built in by default, but it can be
513 specified at Configure time and should be accompanied by the necessary
514 linker additions, eg;
515 ./config -DOPENSSL_USE_GMP -lgmp
518 *) "openssl engine" will not display ENGINE/DSO load failure errors when
519 testing availability of engines with "-t" - the old behaviour is
520 produced by increasing the feature's verbosity with "-tt".
523 *) ECDSA routines: under certain error conditions uninitialized BN objects
524 could be freed. Solution: make sure initialization is performed early
525 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
529 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
530 and DH_METHOD (eg. by ENGINE implementations) to override the normal
531 software implementations. For DSA and DH, parameter generation can
532 also be overriden by providing the appropriate method callbacks.
535 *) Change the "progress" mechanism used in key-generation and
536 primality testing to functions that take a new BN_GENCB pointer in
537 place of callback/argument pairs. The new API functions have "_ex"
538 postfixes and the older functions are reimplemented as wrappers for
539 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
540 declarations of the old functions to help (graceful) attempts to
541 migrate to the new functions. Also, the new key-generation API
542 functions operate on a caller-supplied key-structure and return
543 success/failure rather than returning a key or NULL - this is to
544 help make "keygen" another member function of RSA_METHOD etc.
546 Example for using the new callback interface:
548 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
552 BN_GENCB_set(&my_cb, my_callback, my_arg);
554 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
555 /* For the meaning of a, b in calls to my_callback(), see the
556 * documentation of the function that calls the callback.
557 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
558 * my_callback should return 1 if it wants BN_is_prime_ex()
559 * to continue, or 0 to stop.
564 *) Change the ZLIB compression method to be stateful, and make it
565 available to TLS with the number defined in
566 draft-ietf-tls-compression-04.txt.
569 *) Add the ASN.1 structures and functions for CertificatePair, which
570 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
572 CertificatePair ::= SEQUENCE {
573 forward [0] Certificate OPTIONAL,
574 reverse [1] Certificate OPTIONAL,
575 -- at least one of the pair shall be present -- }
577 Also implement the PEM functions to read and write certificate
578 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
580 This needed to be defined, mostly for the sake of the LDAP
581 attribute crossCertificatePair, but may prove useful elsewhere as
585 *) Make it possible to inhibit symlinking of shared libraries in
586 Makefile.shared, for Cygwin's sake.
589 *) Extend the BIGNUM API by creating a function
590 void BN_set_negative(BIGNUM *a, int neg);
591 and a macro that behave like
592 int BN_is_negative(const BIGNUM *a);
594 to avoid the need to access 'a->neg' directly in applications.
597 *) Implement fast modular reduction for pseudo-Mersenne primes
598 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
599 EC_GROUP_new_curve_GFp() will now automatically use this
601 [Nils Larsch <nla@trustcenter.de>]
603 *) Add new lock type (CRYPTO_LOCK_BN).
606 *) Change the ENGINE framework to automatically load engines
607 dynamically from specific directories unless they could be
608 found to already be built in or loaded. Move all the
609 current engines except for the cryptodev one to a new
611 The engines in engines/ are built as shared libraries if
612 the "shared" options was given to ./Configure or ./config.
613 Otherwise, they are inserted in libcrypto.a.
614 /usr/local/ssl/engines is the default directory for dynamic
615 engines, but that can be overriden at configure time through
616 the usual use of --prefix and/or --openssldir, and at run
617 time with the environment variable OPENSSL_ENGINES.
618 [Geoff Thorpe and Richard Levitte]
620 *) Add Makefile.shared, a helper makefile to build shared
621 libraries. Addapt Makefile.org.
624 *) Add version info to Win32 DLLs.
625 [Peter 'Luna' Runestig" <peter@runestig.com>]
627 *) Add new 'medium level' PKCS#12 API. Certificates and keys
628 can be added using this API to created arbitrary PKCS#12
629 files while avoiding the low level API.
631 New options to PKCS12_create(), key or cert can be NULL and
632 will then be omitted from the output file. The encryption
633 algorithm NIDs can be set to -1 for no encryption, the mac
634 iteration count can be set to 0 to omit the mac.
636 Enhance pkcs12 utility by making the -nokeys and -nocerts
637 options work when creating a PKCS#12 file. New option -nomac
638 to omit the mac, NONE can be set for an encryption algorithm.
639 New code is modified to use the enhanced PKCS12_create()
640 instead of the low level API.
643 *) Extend ASN1 encoder to support indefinite length constructed
644 encoding. This can output sequences tags and octet strings in
645 this form. Modify pk7_asn1.c to support indefinite length
646 encoding. This is experimental and needs additional code to
647 be useful, such as an ASN1 bio and some enhanced streaming
650 Extend template encode functionality so that tagging is passed
651 down to the template encoder.
654 *) Let 'openssl req' fail if an argument to '-newkey' is not
655 recognized instead of using RSA as a default.
658 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
659 As these are not official, they are not included in "ALL";
660 the "ECCdraft" ciphersuite group alias can be used to select them.
661 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
663 *) Add ECDH engine support.
664 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
666 *) Add ECDH in new directory crypto/ecdh/.
667 [Douglas Stebila (Sun Microsystems Laboratories)]
669 *) Let BN_rand_range() abort with an error after 100 iterations
670 without success (which indicates a broken PRNG).
673 *) Change BN_mod_sqrt() so that it verifies that the input value
674 is really the square of the return value. (Previously,
675 BN_mod_sqrt would show GIGO behaviour.)
678 *) Add named elliptic curves over binary fields from X9.62, SECG,
679 and WAP/WTLS; add OIDs that were still missing.
681 [Sheueling Chang Shantz and Douglas Stebila
682 (Sun Microsystems Laboratories)]
684 *) Extend the EC library for elliptic curves over binary fields
685 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
688 EC_GF2m_simple_method
692 EC_GROUP_new_curve_GF2m
693 EC_GROUP_set_curve_GF2m
694 EC_GROUP_get_curve_GF2m
695 EC_POINT_set_affine_coordinates_GF2m
696 EC_POINT_get_affine_coordinates_GF2m
697 EC_POINT_set_compressed_coordinates_GF2m
699 Point compression for binary fields is disabled by default for
700 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
703 As binary polynomials are represented as BIGNUMs, various members
704 of the EC_GROUP and EC_POINT data structures can be shared
705 between the implementations for prime fields and binary fields;
706 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
707 are essentially identical to their ..._GFp counterparts.
708 (For simplicity, the '..._GFp' prefix has been dropped from
709 various internal method names.)
711 An internal 'field_div' method (similar to 'field_mul' and
712 'field_sqr') has been added; this is used only for binary fields.
714 [Sheueling Chang Shantz and Douglas Stebila
715 (Sun Microsystems Laboratories)]
717 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
718 through methods ('mul', 'precompute_mult').
720 The generic implementations (now internally called 'ec_wNAF_mul'
721 and 'ec_wNAF_precomputed_mult') remain the default if these
722 methods are undefined.
724 [Sheueling Chang Shantz and Douglas Stebila
725 (Sun Microsystems Laboratories)]
727 *) New function EC_GROUP_get_degree, which is defined through
728 EC_METHOD. For curves over prime fields, this returns the bit
729 length of the modulus.
731 [Sheueling Chang Shantz and Douglas Stebila
732 (Sun Microsystems Laboratories)]
734 *) New functions EC_GROUP_dup, EC_POINT_dup.
735 (These simply call ..._new and ..._copy).
737 [Sheueling Chang Shantz and Douglas Stebila
738 (Sun Microsystems Laboratories)]
740 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
741 Polynomials are represented as BIGNUMs (where the sign bit is not
742 used) in the following functions [macros]:
745 BN_GF2m_sub [= BN_GF2m_add]
746 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
747 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
748 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
750 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
751 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
752 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
753 BN_GF2m_cmp [= BN_ucmp]
755 (Note that only the 'mod' functions are actually for fields GF(2^m).
756 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
758 For some functions, an the irreducible polynomial defining a
759 field can be given as an 'unsigned int[]' with strictly
760 decreasing elements giving the indices of those bits that are set;
761 i.e., p[] represents the polynomial
762 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
764 p[0] > p[1] > ... > p[k] = 0.
765 This applies to the following functions:
770 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
771 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
774 BN_GF2m_mod_solve_quad_arr
778 Conversion can be performed by the following functions:
783 bntest.c has additional tests for binary polynomial arithmetic.
785 Two implementations for BN_GF2m_mod_div() are available.
786 The default algorithm simply uses BN_GF2m_mod_inv() and
787 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
788 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
789 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
791 [Sheueling Chang Shantz and Douglas Stebila
792 (Sun Microsystems Laboratories)]
794 *) Add new error code 'ERR_R_DISABLED' that can be used when some
795 functionality is disabled at compile-time.
796 [Douglas Stebila <douglas.stebila@sun.com>]
798 *) Change default behaviour of 'openssl asn1parse' so that more
799 information is visible when viewing, e.g., a certificate:
801 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
802 mode the content of non-printable OCTET STRINGs is output in a
803 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
804 avoid the appearance of a printable string.
805 [Nils Larsch <nla@trustcenter.de>]
807 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
809 EC_GROUP_set_asn1_flag()
810 EC_GROUP_get_asn1_flag()
811 EC_GROUP_set_point_conversion_form()
812 EC_GROUP_get_point_conversion_form()
813 These control ASN1 encoding details:
814 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
815 has been set to OPENSSL_EC_NAMED_CURVE.
816 - Points are encoded in uncompressed form by default; options for
817 asn1_for are as for point2oct, namely
818 POINT_CONVERSION_COMPRESSED
819 POINT_CONVERSION_UNCOMPRESSED
820 POINT_CONVERSION_HYBRID
822 Also add 'seed' and 'seed_len' members to EC_GROUP with access
826 EC_GROUP_get_seed_len()
827 This is used only for ASN1 purposes (so far).
828 [Nils Larsch <nla@trustcenter.de>]
830 *) Add 'field_type' member to EC_METHOD, which holds the NID
831 of the appropriate field type OID. The new function
832 EC_METHOD_get_field_type() returns this value.
833 [Nils Larsch <nla@trustcenter.de>]
840 providing useful interfaces to EC_POINT_point2oct() and
841 EC_POINT_oct2point().
842 [Nils Larsch <nla@trustcenter.de>]
844 *) Change internals of the EC library so that the functions
845 EC_GROUP_set_generator()
846 EC_GROUP_get_generator()
848 EC_GROUP_get_cofactor()
849 are implemented directly in crypto/ec/ec_lib.c and not dispatched
850 to methods, which would lead to unnecessary code duplication when
851 adding different types of curves.
852 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
854 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
855 arithmetic, and such that modified wNAFs are generated
856 (which avoid length expansion in many cases).
859 *) Add a function EC_GROUP_check_discriminant() (defined via
860 EC_METHOD) that verifies that the curve discriminant is non-zero.
862 Add a function EC_GROUP_check() that makes some sanity tests
863 on a EC_GROUP, its generator and order. This includes
864 EC_GROUP_check_discriminant().
865 [Nils Larsch <nla@trustcenter.de>]
867 *) Add ECDSA in new directory crypto/ecdsa/.
869 Add applications 'openssl ecparam' and 'openssl ecdsa'
870 (these are based on 'openssl dsaparam' and 'openssl dsa').
872 ECDSA support is also included in various other files across the
873 library. Most notably,
874 - 'openssl req' now has a '-newkey ecdsa:file' option;
875 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
876 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
877 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
878 them suitable for ECDSA where domain parameters must be
879 extracted before the specific public key;
880 - ECDSA engine support has been added.
881 [Nils Larsch <nla@trustcenter.de>]
883 *) Include some named elliptic curves, and add OIDs from X9.62,
884 SECG, and WAP/WTLS. Each curve can be obtained from the new
886 EC_GROUP_new_by_curve_name(),
887 and the list of available named curves can be obtained with
888 EC_get_builtin_curves().
889 Also add a 'curve_name' member to EC_GROUP objects, which can be
891 EC_GROUP_set_curve_name()
892 EC_GROUP_get_curve_name()
893 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
895 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
896 was actually never needed) and in BN_mul(). The removal in BN_mul()
897 required a small change in bn_mul_part_recursive() and the addition
898 of the functions bn_cmp_part_words(), bn_sub_part_words() and
899 bn_add_part_words(), which do the same thing as bn_cmp_words(),
900 bn_sub_words() and bn_add_words() except they take arrays with
904 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
906 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
907 (part of SSL_OP_ALL). This option used to disable the
908 countermeasure against man-in-the-middle protocol-version
909 rollback in the SSL 2.0 server implementation, which is a bad
910 idea. (CAN-2005-2969)
912 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
913 for Information Security, National Institute of Advanced Industrial
914 Science and Technology [AIST], Japan)]
916 *) Minimal support for X9.31 signatures and PSS padding modes. This is
917 mainly for FIPS compliance and not fully integrated at this stage.
920 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
921 the exponentiation using a fixed-length exponent. (Otherwise,
922 the information leaked through timing could expose the secret key
923 after many signatures; cf. Bleichenbacher's attack on DSA with
927 *) Make a new fixed-window mod_exp implementation the default for
928 RSA, DSA, and DH private-key operations so that the sequence of
929 squares and multiplies and the memory access pattern are
930 independent of the particular secret key. This will mitigate
931 cache-timing and potential related attacks.
933 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
934 and this is automatically used by BN_mod_exp_mont() if the new flag
935 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
936 will use this BN flag for private exponents unless the flag
937 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
938 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
940 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
942 *) Change the client implementation for SSLv23_method() and
943 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
944 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
945 (Previously, the SSL 2.0 backwards compatible Client Hello
946 message format would be used even with SSL_OP_NO_SSLv2.)
949 *) Add support for smime-type MIME parameter in S/MIME messages which some
953 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
954 a threadsafe manner. Modify rsa code to use new function and add calls
955 to dsa and dh code (which had race conditions before).
958 *) Include the fixed error library code in the C error file definitions
959 instead of fixing them up at runtime. This keeps the error code
963 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
965 *) Fixes for newer kerberos headers. NB: the casts are needed because
966 the 'length' field is signed on one version and unsigned on another
967 with no (?) obvious way to tell the difference, without these VC++
968 complains. Also the "definition" of FAR (blank) is no longer included
969 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
970 some needed definitions.
973 *) Undo Cygwin change.
976 *) Added support for proxy certificates according to RFC 3820.
977 Because they may be a security thread to unaware applications,
978 they must be explicitely allowed in run-time. See
979 docs/HOWTO/proxy_certificates.txt for further information.
982 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
984 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
985 server and client random values. Previously
986 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
987 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
989 This change has negligible security impact because:
991 1. Server and client random values still have 24 bytes of pseudo random
994 2. Server and client random values are sent in the clear in the initial
997 3. The master secret is derived using the premaster secret (48 bytes in
998 size for static RSA ciphersuites) as well as client server and random
1001 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1004 [Stephen Henson, reported by UK NISCC]
1006 *) Use Windows randomness collection on Cygwin.
1009 *) Fix hang in EGD/PRNGD query when communication socket is closed
1010 prematurely by EGD/PRNGD.
1011 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1013 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1016 *) Back-port of selected performance improvements from development
1017 branch, as well as improved support for PowerPC platforms.
1020 *) Add lots of checks for memory allocation failure, error codes to indicate
1021 failure and freeing up memory if a failure occurs.
1022 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1024 *) Add new -passin argument to dgst.
1027 *) Perform some character comparisons of different types in X509_NAME_cmp:
1028 this is needed for some certificates that reencode DNs into UTF8Strings
1029 (in violation of RFC3280) and can't or wont issue name rollover
1033 *) Make an explicit check during certificate validation to see that
1034 the CA setting in each certificate on the chain is correct. As a
1035 side effect always do the following basic checks on extensions,
1036 not just when there's an associated purpose to the check:
1038 - if there is an unhandled critical extension (unless the user
1039 has chosen to ignore this fault)
1040 - if the path length has been exceeded (if one is set at all)
1041 - that certain extensions fit the associated purpose (if one has
1045 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1047 *) Avoid a race condition when CRLs are checked in a multi threaded
1048 environment. This would happen due to the reordering of the revoked
1049 entries during signature checking and serial number lookup. Now the
1050 encoding is cached and the serial number sort performed under a lock.
1051 Add new STACK function sk_is_sorted().
1054 *) Add Delta CRL to the extension code.
1057 *) Various fixes to s3_pkt.c so alerts are sent properly.
1058 [David Holmes <d.holmes@f5.com>]
1060 *) Reduce the chances of duplicate issuer name and serial numbers (in
1061 violation of RFC3280) using the OpenSSL certificate creation utilities.
1062 This is done by creating a random 64 bit value for the initial serial
1063 number when a serial number file is created or when a self signed
1064 certificate is created using 'openssl req -x509'. The initial serial
1065 number file is created using 'openssl x509 -next_serial' in CA.pl
1066 rather than being initialized to 1.
1069 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1071 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1072 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
1073 [Joe Orton, Steve Henson]
1075 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1077 [Joe Orton, Steve Henson]
1079 *) Make it possible to have multiple active certificates with the same
1080 subject in the CA index file. This is done only if the keyword
1081 'unique_subject' is set to 'no' in the main CA section (default
1082 if 'CA_default') of the configuration file. The value is saved
1083 with the database itself in a separate index attribute file,
1084 named like the index file with '.attr' appended to the name.
1087 *) X509 verify fixes. Disable broken certificate workarounds when
1088 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1089 keyUsage extension present. Don't accept CRLs with unhandled critical
1090 extensions: since verify currently doesn't process CRL extensions this
1091 rejects a CRL with *any* critical extensions. Add new verify error codes
1095 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1096 A clarification of RFC2560 will require the use of OCTET STRINGs and
1097 some implementations cannot handle the current raw format. Since OpenSSL
1098 copies and compares OCSP nonces as opaque blobs without any attempt at
1099 parsing them this should not create any compatibility issues.
1102 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1103 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1104 this HMAC (and other) operations are several times slower than OpenSSL
1108 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1109 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1111 *) Use the correct content when signing type "other".
1114 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1116 *) Fix various bugs revealed by running the NISCC test suite:
1118 Stop out of bounds reads in the ASN1 code when presented with
1119 invalid tags (CAN-2003-0543 and CAN-2003-0544).
1121 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
1123 If verify callback ignores invalid public key errors don't try to check
1124 certificate signature with the NULL public key.
1128 *) New -ignore_err option in ocsp application to stop the server
1129 exiting on the first error in a request.
1132 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1133 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1137 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1138 extra data after the compression methods not only for TLS 1.0
1139 but also for SSL 3.0 (as required by the specification).
1140 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1142 *) Change X509_certificate_type() to mark the key as exported/exportable
1143 when it's 512 *bits* long, not 512 bytes.
1146 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1147 blocks during encryption.
1150 *) Various fixes to base64 BIO and non blocking I/O. On write
1151 flushes were not handled properly if the BIO retried. On read
1152 data was not being buffered properly and had various logic bugs.
1153 This also affects blocking I/O when the data being decoded is a
1157 *) Various S/MIME bugfixes and compatibility changes:
1158 output correct application/pkcs7 MIME type if
1159 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1160 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1161 of files as .eml work). Correctly handle very long lines in MIME
1165 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1167 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1168 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1169 a protocol version number mismatch like a decryption error
1170 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1173 *) Turn on RSA blinding by default in the default implementation
1174 to avoid a timing attack. Applications that don't want it can call
1175 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1176 They would be ill-advised to do so in most cases.
1177 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1179 *) Change RSA blinding code so that it works when the PRNG is not
1180 seeded (in this case, the secret RSA exponent is abused as
1181 an unpredictable seed -- if it is not unpredictable, there
1182 is no point in blinding anyway). Make RSA blinding thread-safe
1183 by remembering the creator's thread ID in rsa->blinding and
1184 having all other threads use local one-time blinding factors
1185 (this requires more computation than sharing rsa->blinding, but
1186 avoids excessive locking; and if an RSA object is not shared
1187 between threads, blinding will still be very fast).
1190 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1191 ENGINE as defaults for all supported algorithms irrespective of
1192 the 'flags' parameter. 'flags' is now honoured, so applications
1193 should make sure they are passing it correctly.
1196 *) Target "mingw" now allows native Windows code to be generated in
1197 the Cygwin environment as well as with the MinGW compiler.
1200 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1202 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1203 via timing by performing a MAC computation even if incorrrect
1204 block cipher padding has been found. This is a countermeasure
1205 against active attacks where the attacker has to distinguish
1206 between bad padding and a MAC verification error. (CAN-2003-0078)
1208 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1209 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1210 Martin Vuagnoux (EPFL, Ilion)]
1212 *) Make the no-err option work as intended. The intention with no-err
1213 is not to have the whole error stack handling routines removed from
1214 libcrypto, it's only intended to remove all the function name and
1215 reason texts, thereby removing some of the footprint that may not
1216 be interesting if those errors aren't displayed anyway.
1218 NOTE: it's still possible for any application or module to have it's
1219 own set of error texts inserted. The routines are there, just not
1220 used by default when no-err is given.
1223 *) Add support for FreeBSD on IA64.
1224 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1226 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1227 Kerberos function mit_des_cbc_cksum(). Before this change,
1228 the value returned by DES_cbc_cksum() was like the one from
1229 mit_des_cbc_cksum(), except the bytes were swapped.
1230 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1232 *) Allow an application to disable the automatic SSL chain building.
1233 Before this a rather primitive chain build was always performed in
1234 ssl3_output_cert_chain(): an application had no way to send the
1235 correct chain if the automatic operation produced an incorrect result.
1237 Now the chain builder is disabled if either:
1239 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1241 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1243 The reasoning behind this is that an application would not want the
1244 auto chain building to take place if extra chain certificates are
1245 present and it might also want a means of sending no additional
1246 certificates (for example the chain has two certificates and the
1250 *) Add the possibility to build without the ENGINE framework.
1251 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1253 *) Under Win32 gmtime() can return NULL: check return value in
1254 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1257 *) DSA routines: under certain error conditions uninitialized BN objects
1258 could be freed. Solution: make sure initialization is performed early
1259 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1260 Nils Larsch <nla@trustcenter.de> via PR#459)
1263 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1264 checked on reconnect on the client side, therefore session resumption
1265 could still fail with a "ssl session id is different" error. This
1266 behaviour is masked when SSL_OP_ALL is used due to
1267 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1268 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1269 followup to PR #377.
1272 *) IA-32 assembler support enhancements: unified ELF targets, support
1273 for SCO/Caldera platforms, fix for Cygwin shared build.
1276 *) Add support for FreeBSD on sparc64. As a consequence, support for
1277 FreeBSD on non-x86 processors is separate from x86 processors on
1278 the config script, much like the NetBSD support.
1279 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1281 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1283 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1286 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1287 code (06) was taken as the first octet of the session ID and the last
1288 octet was ignored consequently. As a result SSLv2 client side session
1289 caching could not have worked due to the session ID mismatch between
1291 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1295 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1296 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1300 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1301 seems that in spite of existing for more than a year, many application
1302 author have done nothing to provide the necessary callbacks, which
1303 means that this particular engine will not work properly anywhere.
1304 This is a very unfortunate situation which forces us, in the name
1305 of usability, to give the hw_ncipher.c a static lock, which is part
1307 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1308 appear in 0.9.8 or later. We EXPECT application authors to have
1309 dealt properly with this when 0.9.8 is released (unless we actually
1310 make such changes in the libcrypto locking code that changes will
1311 have to be made anyway).
1314 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1315 octets have been read, EOF or an error occurs. Without this change
1316 some truncated ASN1 structures will not produce an error.
1319 *) Disable Heimdal support, since it hasn't been fully implemented.
1320 Still give the possibility to force the use of Heimdal, but with
1321 warnings and a request that patches get sent to openssl-dev.
1324 *) Add the VC-CE target, introduce the WINCE sysname, and add
1325 INSTALL.WCE and appropriate conditionals to make it build.
1326 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1328 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1329 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1330 edit numbers of the version.
1331 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1333 *) Introduce safe string copy and catenation functions
1334 (BUF_strlcpy() and BUF_strlcat()).
1335 [Ben Laurie (CHATS) and Richard Levitte]
1337 *) Avoid using fixed-size buffers for one-line DNs.
1338 [Ben Laurie (CHATS)]
1340 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1341 resizing buffers containing secrets, and use where appropriate.
1342 [Ben Laurie (CHATS)]
1344 *) Avoid using fixed size buffers for configuration file location.
1345 [Ben Laurie (CHATS)]
1347 *) Avoid filename truncation for various CA files.
1348 [Ben Laurie (CHATS)]
1350 *) Use sizeof in preference to magic numbers.
1351 [Ben Laurie (CHATS)]
1353 *) Avoid filename truncation in cert requests.
1354 [Ben Laurie (CHATS)]
1356 *) Add assertions to check for (supposedly impossible) buffer
1358 [Ben Laurie (CHATS)]
1360 *) Don't cache truncated DNS entries in the local cache (this could
1361 potentially lead to a spoofing attack).
1362 [Ben Laurie (CHATS)]
1364 *) Fix various buffers to be large enough for hex/decimal
1365 representations in a platform independent manner.
1366 [Ben Laurie (CHATS)]
1368 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1369 resizing buffers containing secrets, and use where appropriate.
1370 [Ben Laurie (CHATS)]
1372 *) Add BIO_indent() to avoid much slightly worrying code to do
1374 [Ben Laurie (CHATS)]
1376 *) Convert sprintf()/BIO_puts() to BIO_printf().
1377 [Ben Laurie (CHATS)]
1379 *) buffer_gets() could terminate with the buffer only half
1381 [Ben Laurie (CHATS)]
1383 *) Add assertions to prevent user-supplied crypto functions from
1384 overflowing internal buffers by having large block sizes, etc.
1385 [Ben Laurie (CHATS)]
1387 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1389 [Ben Laurie (CHATS)]
1391 *) Eliminate unused copy of key in RC4.
1392 [Ben Laurie (CHATS)]
1394 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1395 [Ben Laurie (CHATS)]
1397 *) Fix off-by-one error in EGD path.
1398 [Ben Laurie (CHATS)]
1400 *) If RANDFILE path is too long, ignore instead of truncating.
1401 [Ben Laurie (CHATS)]
1403 *) Eliminate unused and incorrectly sized X.509 structure
1405 [Ben Laurie (CHATS)]
1407 *) Eliminate unused and dangerous function knumber().
1408 [Ben Laurie (CHATS)]
1410 *) Eliminate unused and dangerous structure, KSSL_ERR.
1411 [Ben Laurie (CHATS)]
1413 *) Protect against overlong session ID context length in an encoded
1414 session object. Since these are local, this does not appear to be
1416 [Ben Laurie (CHATS)]
1418 *) Change from security patch (see 0.9.6e below) that did not affect
1419 the 0.9.6 release series:
1421 Remote buffer overflow in SSL3 protocol - an attacker could
1422 supply an oversized master key in Kerberos-enabled versions.
1424 [Ben Laurie (CHATS)]
1426 *) Change the SSL kerb5 codes to match RFC 2712.
1429 *) Make -nameopt work fully for req and add -reqopt switch.
1430 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1432 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1433 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1435 *) Make sure tests can be performed even if the corresponding algorithms
1436 have been removed entirely. This was also the last step to make
1437 OpenSSL compilable with DJGPP under all reasonable conditions.
1438 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1440 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1441 to allow version independent disabling of normally unselected ciphers,
1442 which may be activated as a side-effect of selecting a single cipher.
1444 (E.g., cipher list string "RSA" enables ciphersuites that are left
1445 out of "ALL" because they do not provide symmetric encryption.
1446 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1447 [Lutz Jaenicke, Bodo Moeller]
1449 *) Add appropriate support for separate platform-dependent build
1450 directories. The recommended way to make a platform-dependent
1451 build directory is the following (tested on Linux), maybe with
1454 # Place yourself outside of the OpenSSL source tree. In
1455 # this example, the environment variable OPENSSL_SOURCE
1456 # is assumed to contain the absolute OpenSSL source directory.
1457 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1458 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1459 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1460 mkdir -p `dirname $F`
1461 ln -s $OPENSSL_SOURCE/$F $F
1464 To be absolutely sure not to disturb the source tree, a "make clean"
1465 is a good thing. If it isn't successfull, don't worry about it,
1466 it probably means the source directory is very clean.
1469 *) Make sure any ENGINE control commands make local copies of string
1470 pointers passed to them whenever necessary. Otherwise it is possible
1471 the caller may have overwritten (or deallocated) the original string
1472 data when a later ENGINE operation tries to use the stored values.
1473 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1475 *) Improve diagnostics in file reading and command-line digests.
1476 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1478 *) Add AES modes CFB and OFB to the object database. Correct an
1479 error in AES-CFB decryption.
1482 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1483 allows existing EVP_CIPHER_CTX structures to be reused after
1484 calling EVP_*Final(). This behaviour is used by encryption
1485 BIOs and some applications. This has the side effect that
1486 applications must explicitly clean up cipher contexts with
1487 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1490 *) Check the values of dna and dnb in bn_mul_recursive before calling
1491 bn_mul_comba (a non zero value means the a or b arrays do not contain
1492 n2 elements) and fallback to bn_mul_normal if either is not zero.
1495 *) Fix escaping of non-ASCII characters when using the -subj option
1496 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1499 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1500 form for "surname", serialNumber has no short form.
1501 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1502 therefore remove "mail" short name for "internet 7".
1503 The OID for unique identifiers in X509 certificates is
1504 x500UniqueIdentifier, not uniqueIdentifier.
1505 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1508 *) Add an "init" command to the ENGINE config module and auto initialize
1509 ENGINEs. Without any "init" command the ENGINE will be initialized
1510 after all ctrl commands have been executed on it. If init=1 the
1511 ENGINE is initailized at that point (ctrls before that point are run
1512 on the uninitialized ENGINE and after on the initialized one). If
1513 init=0 then the ENGINE will not be iniatialized at all.
1516 *) Fix the 'app_verify_callback' interface so that the user-defined
1517 argument is actually passed to the callback: In the
1518 SSL_CTX_set_cert_verify_callback() prototype, the callback
1519 declaration has been changed from
1522 int (*cb)(X509_STORE_CTX *,void *);
1523 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1524 i=s->ctx->app_verify_callback(&ctx)
1525 has been changed into
1526 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1528 To update applications using SSL_CTX_set_cert_verify_callback(),
1529 a dummy argument can be added to their callback functions.
1530 [D. K. Smetters <smetters@parc.xerox.com>]
1532 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1533 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1535 *) Add and OPENSSL_LOAD_CONF define which will cause
1536 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1537 This allows older applications to transparently support certain
1538 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1539 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1540 load the config file and OPENSSL_add_all_algorithms_conf() which will
1541 always load it have also been added.
1544 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1545 Adjust NIDs and EVP layer.
1546 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1548 *) Config modules support in openssl utility.
1550 Most commands now load modules from the config file,
1551 though in a few (such as version) this isn't done
1552 because it couldn't be used for anything.
1554 In the case of ca and req the config file used is
1555 the same as the utility itself: that is the -config
1556 command line option can be used to specify an
1560 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1561 use "openssl_conf" if filename is NULL use default openssl config file.
1564 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1565 config section name. Add a new flag to tolerate a missing config file
1566 and move code to CONF_modules_load_file().
1569 *) Support for crypto accelerator cards from Accelerated Encryption
1570 Processing, www.aep.ie. (Use engine 'aep')
1571 The support was copied from 0.9.6c [engine] and adapted/corrected
1572 to work with the new engine framework.
1573 [AEP Inc. and Richard Levitte]
1575 *) Support for SureWare crypto accelerator cards from Baltimore
1576 Technologies. (Use engine 'sureware')
1577 The support was copied from 0.9.6c [engine] and adapted
1578 to work with the new engine framework.
1581 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1582 make the newer ENGINE framework commands for the CHIL engine work.
1583 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1585 *) Make it possible to produce shared libraries on ReliantUNIX.
1586 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1588 *) Add the configuration target debug-linux-ppro.
1589 Make 'openssl rsa' use the general key loading routines
1590 implemented in apps.c, and make those routines able to
1591 handle the key format FORMAT_NETSCAPE and the variant
1593 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1595 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1596 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1598 *) Add -keyform to rsautl, and document -engine.
1599 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1601 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1602 BIO_R_NO_SUCH_FILE error code rather than the generic
1603 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1606 *) Add new functions
1608 ERR_peek_last_error_line
1609 ERR_peek_last_error_line_data.
1610 These are similar to
1613 ERR_peek_error_line_data,
1614 but report on the latest error recorded rather than the first one
1615 still in the error queue.
1616 [Ben Laurie, Bodo Moeller]
1618 *) default_algorithms option in ENGINE config module. This allows things
1620 default_algorithms = ALL
1621 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1624 *) Prelminary ENGINE config module.
1627 *) New experimental application configuration code.
1630 *) Change the AES code to follow the same name structure as all other
1631 symmetric ciphers, and behave the same way. Move everything to
1632 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1633 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1635 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1636 [Ben Laurie and Theo de Raadt]
1638 *) Add option to output public keys in req command.
1639 [Massimiliano Pala madwolf@openca.org]
1641 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1642 (up to about 10% better than before for P-192 and P-224).
1645 *) New functions/macros
1647 SSL_CTX_set_msg_callback(ctx, cb)
1648 SSL_CTX_set_msg_callback_arg(ctx, arg)
1649 SSL_set_msg_callback(ssl, cb)
1650 SSL_set_msg_callback_arg(ssl, arg)
1652 to request calling a callback function
1654 void cb(int write_p, int version, int content_type,
1655 const void *buf, size_t len, SSL *ssl, void *arg)
1657 whenever a protocol message has been completely received
1658 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1659 protocol version according to which the SSL library interprets
1660 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1661 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1662 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1663 specification (change_cipher_spec(20), alert(21), handshake(22)).
1664 'buf' and 'len' point to the actual message, 'ssl' to the
1665 SSL object, and 'arg' is the application-defined value set by
1666 SSL[_CTX]_set_msg_callback_arg().
1668 'openssl s_client' and 'openssl s_server' have new '-msg' options
1669 to enable a callback that displays all protocol messages.
1672 *) Change the shared library support so shared libraries are built as
1673 soon as the corresponding static library is finished, and thereby get
1674 openssl and the test programs linked against the shared library.
1675 This still only happens when the keyword "shard" has been given to
1676 the configuration scripts.
1678 NOTE: shared library support is still an experimental thing, and
1679 backward binary compatibility is still not guaranteed.
1680 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1682 *) Add support for Subject Information Access extension.
1683 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1685 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1686 additional bytes when new memory had to be allocated, not just
1687 when reusing an existing buffer.
1690 *) New command line and configuration option 'utf8' for the req command.
1691 This allows field values to be specified as UTF8 strings.
1694 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1695 runs for the former and machine-readable output for the latter.
1698 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1699 of the e-mail address in the DN (i.e., it will go into a certificate
1700 extension only). The new configuration file option 'email_in_dn = no'
1701 has the same effect.
1702 [Massimiliano Pala madwolf@openca.org]
1704 *) Change all functions with names starting with des_ to be starting
1705 with DES_ instead. Add wrappers that are compatible with libdes,
1706 but are named _ossl_old_des_*. Finally, add macros that map the
1707 des_* symbols to the corresponding _ossl_old_des_* if libdes
1708 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1709 desired, the des_* symbols will be mapped to DES_*, with one
1712 Since we provide two compatibility mappings, the user needs to
1713 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1714 compatibility is desired. The default (i.e., when that macro
1715 isn't defined) is OpenSSL 0.9.6c compatibility.
1717 There are also macros that enable and disable the support of old
1718 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1719 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1720 are defined, the default will apply: to support the old des routines.
1722 In either case, one must include openssl/des.h to get the correct
1723 definitions. Do not try to just include openssl/des_old.h, that
1726 NOTE: This is a major break of an old API into a new one. Software
1727 authors are encouraged to switch to the DES_ style functions. Some
1728 time in the future, des_old.h and the libdes compatibility functions
1729 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1730 default), and then completely removed.
1733 *) Test for certificates which contain unsupported critical extensions.
1734 If such a certificate is found during a verify operation it is
1735 rejected by default: this behaviour can be overridden by either
1736 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1737 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1738 X509_supported_extension() has also been added which returns 1 if a
1739 particular extension is supported.
1742 *) Modify the behaviour of EVP cipher functions in similar way to digests
1743 to retain compatibility with existing code.
1746 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1747 compatibility with existing code. In particular the 'ctx' parameter does
1748 not have to be to be initialized before the call to EVP_DigestInit() and
1749 it is tidied up after a call to EVP_DigestFinal(). New function
1750 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1751 EVP_MD_CTX_copy() changed to not require the destination to be
1752 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1753 requires the destination to be valid.
1755 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1756 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1759 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1760 so that complete 'Handshake' protocol structures are kept in memory
1761 instead of overwriting 'msg_type' and 'length' with 'body' data.
1764 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1765 [Massimo Santin via Richard Levitte]
1767 *) Major restructuring to the underlying ENGINE code. This includes
1768 reduction of linker bloat, separation of pure "ENGINE" manipulation
1769 (initialisation, etc) from functionality dealing with implementations
1770 of specific crypto iterfaces. This change also introduces integrated
1771 support for symmetric ciphers and digest implementations - so ENGINEs
1772 can now accelerate these by providing EVP_CIPHER and EVP_MD
1773 implementations of their own. This is detailed in crypto/engine/README
1774 as it couldn't be adequately described here. However, there are a few
1775 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1776 were changed in the original introduction of ENGINE code have now
1777 reverted back - the hooking from this code to ENGINE is now a good
1778 deal more passive and at run-time, operations deal directly with
1779 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1780 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1781 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1782 they were not being used by the framework as there is no concept of a
1783 BIGNUM_METHOD and they could not be generalised to the new
1784 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1785 ENGINE_cpy() has been removed as it cannot be consistently defined in
1789 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1792 *) Change mkdef.pl to sort symbols that get the same entry number,
1793 and make sure the automatically generated functions ERR_load_*
1794 become part of libeay.num as well.
1797 *) New function SSL_renegotiate_pending(). This returns true once
1798 renegotiation has been requested (either SSL_renegotiate() call
1799 or HelloRequest/ClientHello receveived from the peer) and becomes
1800 false once a handshake has been completed.
1801 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1802 sends a HelloRequest, but does not ensure that a handshake takes
1803 place. SSL_renegotiate_pending() is useful for checking if the
1804 client has followed the request.)
1807 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1808 By default, clients may request session resumption even during
1809 renegotiation (if session ID contexts permit); with this option,
1810 session resumption is possible only in the first handshake.
1812 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1813 more bits available for options that should not be part of
1814 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1817 *) Add some demos for certificate and certificate request creation.
1820 *) Make maximum certificate chain size accepted from the peer application
1821 settable (SSL*_get/set_max_cert_list()), as proposed by
1822 "Douglas E. Engert" <deengert@anl.gov>.
1825 *) Add support for shared libraries for Unixware-7
1826 (Boyd Lynn Gerber <gerberb@zenez.com>).
1829 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1830 be done prior to destruction. Use this to unload error strings from
1831 ENGINEs that load their own error strings. NB: This adds two new API
1832 functions to "get" and "set" this destroy handler in an ENGINE.
1835 *) Alter all existing ENGINE implementations (except "openssl" and
1836 "openbsd") to dynamically instantiate their own error strings. This
1837 makes them more flexible to be built both as statically-linked ENGINEs
1838 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1839 Also, add stub code to each that makes building them as self-contained
1840 shared-libraries easier (see README.ENGINE).
1843 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1844 implementations into applications that are completely implemented in
1845 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1846 commands that can be used to configure what shared-library to load and
1847 to control aspects of the way it is handled. Also, made an update to
1848 the README.ENGINE file that brings its information up-to-date and
1849 provides some information and instructions on the "dynamic" ENGINE
1850 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1853 *) Make it possible to unload ranges of ERR strings with a new
1854 "ERR_unload_strings" function.
1857 *) Add a copy() function to EVP_MD.
1860 *) Make EVP_MD routines take a context pointer instead of just the
1861 md_data void pointer.
1864 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1865 that the digest can only process a single chunk of data
1866 (typically because it is provided by a piece of
1867 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1868 is only going to provide a single chunk of data, and hence the
1869 framework needn't accumulate the data for oneshot drivers.
1872 *) As with "ERR", make it possible to replace the underlying "ex_data"
1873 functions. This change also alters the storage and management of global
1874 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1875 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1876 index counters. The API functions that use this state have been changed
1877 to take a "class_index" rather than pointers to the class's local STACK
1878 and counter, and there is now an API function to dynamically create new
1879 classes. This centralisation allows us to (a) plug a lot of the
1880 thread-safety problems that existed, and (b) makes it possible to clean
1881 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1882 such data would previously have always leaked in application code and
1883 workarounds were in place to make the memory debugging turn a blind eye
1884 to it. Application code that doesn't use this new function will still
1885 leak as before, but their memory debugging output will announce it now
1886 rather than letting it slide.
1888 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1889 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1890 has a return value to indicate success or failure.
1893 *) Make it possible to replace the underlying "ERR" functions such that the
1894 global state (2 LHASH tables and 2 locks) is only used by the "default"
1895 implementation. This change also adds two functions to "get" and "set"
1896 the implementation prior to it being automatically set the first time
1897 any other ERR function takes place. Ie. an application can call "get",
1898 pass the return value to a module it has just loaded, and that module
1899 can call its own "set" function using that value. This means the
1900 module's "ERR" operations will use (and modify) the error state in the
1901 application and not in its own statically linked copy of OpenSSL code.
1904 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1905 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1906 the operation, and provides a more encapsulated way for external code
1907 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1908 to use these functions rather than manually incrementing the counts.
1910 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1913 *) Add EVP test program.
1916 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1919 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1920 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1921 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1922 These allow a CRL to be built without having to access X509_CRL fields
1923 directly. Modify 'ca' application to use new functions.
1926 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1927 bug workarounds. Rollback attack detection is a security feature.
1928 The problem will only arise on OpenSSL servers when TLSv1 is not
1929 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1930 Software authors not wanting to support TLSv1 will have special reasons
1931 for their choice and can explicitly enable this option.
1932 [Bodo Moeller, Lutz Jaenicke]
1934 *) Rationalise EVP so it can be extended: don't include a union of
1935 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1936 (similar to those existing for EVP_CIPHER_CTX).
1941 EVP_MD_CTX_init(&md); /* new function call */
1942 EVP_DigestInit(&md, EVP_sha1());
1943 EVP_DigestUpdate(&md, in, len);
1944 EVP_DigestFinal(&md, out, NULL);
1945 EVP_MD_CTX_cleanup(&md); /* new function call */
1949 *) Make DES key schedule conform to the usual scheme, as well as
1950 correcting its structure. This means that calls to DES functions
1951 now have to pass a pointer to a des_key_schedule instead of a
1952 plain des_key_schedule (which was actually always a pointer
1955 des_key_schedule ks;
1957 des_set_key_checked(..., &ks);
1958 des_ncbc_encrypt(..., &ks, ...);
1960 (Note that a later change renames 'des_...' into 'DES_...'.)
1963 *) Initial reduction of linker bloat: the use of some functions, such as
1964 PEM causes large amounts of unused functions to be linked in due to
1965 poor organisation. For example pem_all.c contains every PEM function
1966 which has a knock on effect of linking in large amounts of (unused)
1967 ASN1 code. Grouping together similar functions and splitting unrelated
1968 functions prevents this.
1971 *) Cleanup of EVP macros.
1974 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1975 correct _ecb suffix.
1978 *) Add initial OCSP responder support to ocsp application. The
1979 revocation information is handled using the text based index
1980 use by the ca application. The responder can either handle
1981 requests generated internally, supplied in files (for example
1982 via a CGI script) or using an internal minimal server.
1985 *) Add configuration choices to get zlib compression for TLS.
1988 *) Changes to Kerberos SSL for RFC 2712 compliance:
1989 1. Implemented real KerberosWrapper, instead of just using
1990 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1991 2. Implemented optional authenticator field of KerberosWrapper.
1993 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1994 and authenticator structs; see crypto/krb5/.
1996 Generalized Kerberos calls to support multiple Kerberos libraries.
1997 [Vern Staats <staatsvr@asc.hpc.mil>,
1998 Jeffrey Altman <jaltman@columbia.edu>
1999 via Richard Levitte]
2001 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2002 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2003 values for each of the key sizes rather than having just
2004 parameters (and 'speed' generating keys each time).
2007 *) Speed up EVP routines.
2010 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2011 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2012 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2013 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2015 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2016 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2017 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2020 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2022 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2025 *) Added the OS2-EMX target.
2026 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2028 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2029 to support NCONF routines in extension code. New function CONF_set_nconf()
2030 to allow functions which take an NCONF to also handle the old LHASH
2031 structure: this means that the old CONF compatible routines can be
2032 retained (in particular wrt extensions) without having to duplicate the
2033 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2036 *) Enhance the general user interface with mechanisms for inner control
2037 and with possibilities to have yes/no kind of prompts.
2040 *) Change all calls to low level digest routines in the library and
2041 applications to use EVP. Add missing calls to HMAC_cleanup() and
2042 don't assume HMAC_CTX can be copied using memcpy().
2043 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2045 *) Add the possibility to control engines through control names but with
2046 arbitrary arguments instead of just a string.
2047 Change the key loaders to take a UI_METHOD instead of a callback
2048 function pointer. NOTE: this breaks binary compatibility with earlier
2049 versions of OpenSSL [engine].
2050 Adapt the nCipher code for these new conditions and add a card insertion
2054 *) Enhance the general user interface with mechanisms to better support
2055 dialog box interfaces, application-defined prompts, the possibility
2056 to use defaults (for example default passwords from somewhere else)
2057 and interrupts/cancellations.
2060 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2061 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2064 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2065 tidy up some unnecessarily weird code in 'sk_new()').
2066 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2068 *) Change the key loading routines for ENGINEs to use the same kind
2069 callback (pem_password_cb) as all other routines that need this
2073 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2074 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2075 than this minimum value is recommended.
2078 *) New random seeder for OpenVMS, using the system process statistics
2079 that are easily reachable.
2082 *) Windows apparently can't transparently handle global
2083 variables defined in DLLs. Initialisations such as:
2085 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2087 wont compile. This is used by the any applications that need to
2088 declare their own ASN1 modules. This was fixed by adding the option
2089 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2090 needed for static libraries under Win32.
2093 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2094 setting of purpose and trust fields. New X509_STORE trust and
2095 purpose functions and tidy up setting in other SSL functions.
2098 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2099 structure. These are inherited by X509_STORE_CTX when it is
2100 initialised. This allows various defaults to be set in the
2101 X509_STORE structure (such as flags for CRL checking and custom
2102 purpose or trust settings) for functions which only use X509_STORE_CTX
2103 internally such as S/MIME.
2105 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2106 trust settings if they are not set in X509_STORE. This allows X509_STORE
2107 purposes and trust (in S/MIME for example) to override any set by default.
2109 Add command line options for CRL checking to smime, s_client and s_server
2113 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2114 are set then the CRL is looked up in the X509_STORE structure and
2115 its validity and signature checked, then if the certificate is found
2116 in the CRL the verify fails with a revoked error.
2118 Various new CRL related callbacks added to X509_STORE_CTX structure.
2120 Command line options added to 'verify' application to support this.
2122 This needs some additional work, such as being able to handle multiple
2123 CRLs with different times, extension based lookup (rather than just
2124 by subject name) and ultimately more complete V2 CRL extension
2128 *) Add a general user interface API (crypto/ui/). This is designed
2129 to replace things like des_read_password and friends (backward
2130 compatibility functions using this new API are provided).
2131 The purpose is to remove prompting functions from the DES code
2132 section as well as provide for prompting through dialog boxes in
2133 a window system and the like.
2136 *) Add "ex_data" support to ENGINE so implementations can add state at a
2137 per-structure level rather than having to store it globally.
2140 *) Make it possible for ENGINE structures to be copied when retrieved by
2141 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2142 This causes the "original" ENGINE structure to act like a template,
2143 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2144 operational state can be localised to each ENGINE structure, despite the
2145 fact they all share the same "methods". New ENGINE structures returned in
2146 this case have no functional references and the return value is the single
2147 structural reference. This matches the single structural reference returned
2148 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2152 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2153 needs to match any other type at all we need to manually clear the
2157 *) Changes to the "openssl engine" utility to include;
2158 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2159 about an ENGINE's available control commands.
2160 - executing control commands from command line arguments using the
2161 '-pre' and '-post' switches. '-post' is only used if '-t' is
2162 specified and the ENGINE is successfully initialised. The syntax for
2163 the individual commands are colon-separated, for example;
2164 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2167 *) New dynamic control command support for ENGINEs. ENGINEs can now
2168 declare their own commands (numbers), names (strings), descriptions,
2169 and input types for run-time discovery by calling applications. A
2170 subset of these commands are implicitly classed as "executable"
2171 depending on their input type, and only these can be invoked through
2172 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2173 can be based on user input, config files, etc). The distinction is
2174 that "executable" commands cannot return anything other than a boolean
2175 result and can only support numeric or string input, whereas some
2176 discoverable commands may only be for direct use through
2177 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2178 pointers, or other custom uses. The "executable" commands are to
2179 support parameterisations of ENGINE behaviour that can be
2180 unambiguously defined by ENGINEs and used consistently across any
2181 OpenSSL-based application. Commands have been added to all the
2182 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2183 control over shared-library paths without source code alterations.
2186 *) Changed all ENGINE implementations to dynamically allocate their
2187 ENGINEs rather than declaring them statically. Apart from this being
2188 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2189 this also allows the implementations to compile without using the
2190 internal engine_int.h header.
2193 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2194 'const' value. Any code that should be able to modify a RAND_METHOD
2195 should already have non-const pointers to it (ie. they should only
2196 modify their own ones).
2199 *) Made a variety of little tweaks to the ENGINE code.
2200 - "atalla" and "ubsec" string definitions were moved from header files
2201 to C code. "nuron" string definitions were placed in variables
2202 rather than hard-coded - allowing parameterisation of these values
2203 later on via ctrl() commands.
2204 - Removed unused "#if 0"'d code.
2205 - Fixed engine list iteration code so it uses ENGINE_free() to release
2206 structural references.
2207 - Constified the RAND_METHOD element of ENGINE structures.
2208 - Constified various get/set functions as appropriate and added
2209 missing functions (including a catch-all ENGINE_cpy that duplicates
2210 all ENGINE values onto a new ENGINE except reference counts/state).
2211 - Removed NULL parameter checks in get/set functions. Setting a method
2212 or function to NULL is a way of cancelling out a previously set
2213 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2214 and doesn't justify the extra error symbols and code.
2215 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2216 flags from engine_int.h to engine.h.
2217 - Changed prototypes for ENGINE handler functions (init(), finish(),
2218 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2221 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2222 to the algorithm using long division. The binary algorithm can be
2223 used only if the modulus is odd. On 32-bit systems, it is faster
2224 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2225 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2226 up to 450 bits. In 64-bit environments, the binary algorithm
2227 appears to be advantageous for much longer moduli; here we use it
2228 for moduli up to 2048 bits.
2231 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2232 could not support the combine flag in choice fields.
2235 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2236 extensions from a certificate request to the certificate.
2239 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2240 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2241 file: this allows the display of the certificate about to be
2242 signed to be customised, to allow certain fields to be included
2243 or excluded and extension details. The old system didn't display
2244 multicharacter strings properly, omitted fields not in the policy
2245 and couldn't display additional details such as extensions.
2248 *) Function EC_POINTs_mul for multiple scalar multiplication
2249 of an arbitrary number of elliptic curve points
2250 \sum scalars[i]*points[i],
2251 optionally including the generator defined for the EC_GROUP:
2252 scalar*generator + \sum scalars[i]*points[i].
2254 EC_POINT_mul is a simple wrapper function for the typical case
2255 that the point list has just one item (besides the optional
2259 *) First EC_METHODs for curves over GF(p):
2261 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2262 operations and provides various method functions that can also
2263 operate with faster implementations of modular arithmetic.
2265 EC_GFp_mont_method() reuses most functions that are part of
2266 EC_GFp_simple_method, but uses Montgomery arithmetic.
2268 [Bodo Moeller; point addition and point doubling
2269 implementation directly derived from source code provided by
2270 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2272 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2273 crypto/ec/ec_lib.c):
2275 Curves are EC_GROUP objects (with an optional group generator)
2276 based on EC_METHODs that are built into the library.
2278 Points are EC_POINT objects based on EC_GROUP objects.
2280 Most of the framework would be able to handle curves over arbitrary
2281 finite fields, but as there are no obvious types for fields other
2282 than GF(p), some functions are limited to that for now.
2285 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2286 that the file contains a complete HTTP response.
2289 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2290 change the def and num file printf format specifier from "%-40sXXX"
2291 to "%-39s XXX". The latter will always guarantee a space after the
2292 field while the former will cause them to run together if the field
2293 is 40 of more characters long.
2296 *) Constify the cipher and digest 'method' functions and structures
2297 and modify related functions to take constant EVP_MD and EVP_CIPHER
2301 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2302 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2305 *) Modify EVP_Digest*() routines so they now return values. Although the
2306 internal software routines can never fail additional hardware versions
2310 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2312 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2313 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2316 ERR_R_NESTED_ASN1_ERROR
2318 ERR_R_MISSING_ASN1_EOS
2319 were 4 .. 9, conflicting with
2320 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2322 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2323 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2325 Add new error code 'ERR_R_INTERNAL_ERROR'.
2328 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2332 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2333 sets the subject name for a new request or supersedes the
2334 subject name in a given request. Formats that can be parsed are
2335 'CN=Some Name, OU=myOU, C=IT'
2337 'CN=Some Name/OU=myOU/C=IT'.
2339 Add options '-batch' and '-verbose' to 'openssl req'.
2340 [Massimiliano Pala <madwolf@hackmasters.net>]
2342 *) Introduce the possibility to access global variables through
2343 functions on platform were that's the best way to handle exporting
2344 global variables in shared libraries. To enable this functionality,
2345 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2346 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2347 is normally done by Configure or something similar).
2349 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2350 in the source file (foo.c) like this:
2352 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2353 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2355 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2356 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2358 OPENSSL_DECLARE_GLOBAL(int,foo);
2359 #define foo OPENSSL_GLOBAL_REF(foo)
2360 OPENSSL_DECLARE_GLOBAL(double,bar);
2361 #define bar OPENSSL_GLOBAL_REF(bar)
2363 The #defines are very important, and therefore so is including the
2364 header file everywhere where the defined globals are used.
2366 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2367 of ASN.1 items, but that structure is a bit different.
2369 The largest change is in util/mkdef.pl which has been enhanced with
2370 better and easier to understand logic to choose which symbols should
2371 go into the Windows .def files as well as a number of fixes and code
2372 cleanup (among others, algorithm keywords are now sorted
2373 lexicographically to avoid constant rewrites).
2376 *) In BN_div() keep a copy of the sign of 'num' before writing the
2377 result to 'rm' because if rm==num the value will be overwritten
2378 and produce the wrong result if 'num' is negative: this caused
2379 problems with BN_mod() and BN_nnmod().
2382 *) Function OCSP_request_verify(). This checks the signature on an
2383 OCSP request and verifies the signer certificate. The signer
2384 certificate is just checked for a generic purpose and OCSP request
2388 *) Add OCSP_check_validity() function to check the validity of OCSP
2389 responses. OCSP responses are prepared in real time and may only
2390 be a few seconds old. Simply checking that the current time lies
2391 between thisUpdate and nextUpdate max reject otherwise valid responses
2392 caused by either OCSP responder or client clock inaccuracy. Instead
2393 we allow thisUpdate and nextUpdate to fall within a certain period of
2394 the current time. The age of the response can also optionally be
2395 checked. Two new options -validity_period and -status_age added to
2399 *) If signature or public key algorithm is unrecognized print out its
2400 OID rather that just UNKNOWN.
2403 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2404 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2405 ID to be generated from the issuer certificate alone which can then be
2406 passed to OCSP_id_issuer_cmp().
2409 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2410 ASN1 modules to export functions returning ASN1_ITEM pointers
2411 instead of the ASN1_ITEM structures themselves. This adds several
2412 new macros which allow the underlying ASN1 function/structure to
2413 be accessed transparently. As a result code should not use ASN1_ITEM
2414 references directly (such as &X509_it) but instead use the relevant
2415 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2416 use of the new ASN1 code on platforms where exporting structures
2417 is problematical (for example in shared libraries) but exporting
2418 functions returning pointers to structures is not.
2421 *) Add support for overriding the generation of SSL/TLS session IDs.
2422 These callbacks can be registered either in an SSL_CTX or per SSL.
2423 The purpose of this is to allow applications to control, if they wish,
2424 the arbitrary values chosen for use as session IDs, particularly as it
2425 can be useful for session caching in multiple-server environments. A
2426 command-line switch for testing this (and any client code that wishes
2427 to use such a feature) has been added to "s_server".
2428 [Geoff Thorpe, Lutz Jaenicke]
2430 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2431 of the form '#if defined(...) || defined(...) || ...' and
2432 '#if !defined(...) && !defined(...) && ...'. This also avoids
2433 the growing number of special cases it was previously handling.
2436 *) Make all configuration macros available for application by making
2437 sure they are available in opensslconf.h, by giving them names starting
2438 with "OPENSSL_" to avoid conflicts with other packages and by making
2439 sure e_os2.h will cover all platform-specific cases together with
2441 Additionally, it is now possible to define configuration/platform-
2442 specific names (called "system identities"). In the C code, these
2443 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2444 macro with the name beginning with "OPENSSL_SYS_", which is determined
2445 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2449 *) New option -set_serial to 'req' and 'x509' this allows the serial
2450 number to use to be specified on the command line. Previously self
2451 signed certificates were hard coded with serial number 0 and the
2452 CA options of 'x509' had to use a serial number in a file which was
2456 *) New options to 'ca' utility to support V2 CRL entry extensions.
2457 Currently CRL reason, invalidity date and hold instruction are
2458 supported. Add new CRL extensions to V3 code and some new objects.
2461 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2462 disable standard block padding (aka PKCS#5 padding) in the EVP
2463 API, which was previously mandatory. This means that the data is
2464 not padded in any way and so the total length much be a multiple
2465 of the block size, otherwise an error occurs.
2468 *) Initial (incomplete) OCSP SSL support.
2471 *) New function OCSP_parse_url(). This splits up a URL into its host,
2472 port and path components: primarily to parse OCSP URLs. New -url
2473 option to ocsp utility.
2476 *) New nonce behavior. The return value of OCSP_check_nonce() now
2477 reflects the various checks performed. Applications can decide
2478 whether to tolerate certain situations such as an absent nonce
2479 in a response when one was present in a request: the ocsp application
2480 just prints out a warning. New function OCSP_add1_basic_nonce()
2481 this is to allow responders to include a nonce in a response even if
2482 the request is nonce-less.
2485 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2486 skipped when using openssl x509 multiple times on a single input file,
2487 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2490 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2491 set string type: to handle setting ASN1_TIME structures. Fix ca
2492 utility to correctly initialize revocation date of CRLs.
2495 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2496 the clients preferred ciphersuites and rather use its own preferences.
2497 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2498 Internet Explorer by ensuring unchanged hash method during stepup.
2499 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2502 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2503 to aes and add a new 'exist' option to print out symbols that don't
2507 *) Additional options to ocsp utility to allow flags to be set and
2508 additional certificates supplied.
2511 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2512 OCSP client a number of certificate to only verify the response
2516 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2517 handle the new API. Currently only ECB, CBC modes supported. Add new
2520 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2521 Encryption Standard (AES) Ciphersuites for Transport Layer
2522 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2523 not enabled by default and were not part of the "ALL" ciphersuite
2524 alias because they were not yet official; they could be
2525 explicitly requested by specifying the "AESdraft" ciphersuite
2526 group alias. In the final release of OpenSSL 0.9.7, the group
2527 alias is called "AES" and is part of "ALL".)
2528 [Ben Laurie, Steve Henson, Bodo Moeller]
2530 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2531 request to response.
2534 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2535 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2536 extract information from a certificate request. OCSP_response_create()
2537 creates a response and optionally adds a basic response structure.
2538 OCSP_basic_add1_status() adds a complete single response to a basic
2539 response and returns the OCSP_SINGLERESP structure just added (to allow
2540 extensions to be included for example). OCSP_basic_add1_cert() adds a
2541 certificate to a basic response and OCSP_basic_sign() signs a basic
2542 response with various flags. New helper functions ASN1_TIME_check()
2543 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2544 (converts ASN1_TIME to GeneralizedTime).
2547 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2548 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2549 structure from a certificate. X509_pubkey_digest() digests the public_key
2550 contents: this is used in various key identifiers.
2553 *) Make sk_sort() tolerate a NULL argument.
2554 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2556 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2557 passed by the function are trusted implicitly. If any of them signed the
2558 response then it is assumed to be valid and is not verified.
2561 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2562 to data. This was previously part of the PKCS7 ASN1 code. This
2563 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2564 [Steve Henson, reported by Kenneth R. Robinette
2565 <support@securenetterm.com>]
2567 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2568 routines: without these tracing memory leaks is very painful.
2569 Fix leaks in PKCS12 and PKCS7 routines.
2572 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2573 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2574 effectively meant GeneralizedTime would never be used. Now it
2575 is initialised to -1 but X509_time_adj() now has to check the value
2576 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2577 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2578 [Steve Henson, reported by Kenneth R. Robinette
2579 <support@securenetterm.com>]
2581 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2582 result in a zero length in the ASN1_INTEGER structure which was
2583 not consistent with the structure when d2i_ASN1_INTEGER() was used
2584 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2585 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2586 where it did not print out a minus for negative ASN1_INTEGER.
2589 *) Add summary printout to ocsp utility. The various functions which
2590 convert status values to strings have been renamed to:
2591 OCSP_response_status_str(), OCSP_cert_status_str() and
2592 OCSP_crl_reason_str() and are no longer static. New options
2593 to verify nonce values and to disable verification. OCSP response
2594 printout format cleaned up.
2597 *) Add additional OCSP certificate checks. These are those specified
2598 in RFC2560. This consists of two separate checks: the CA of the
2599 certificate being checked must either be the OCSP signer certificate
2600 or the issuer of the OCSP signer certificate. In the latter case the
2601 OCSP signer certificate must contain the OCSP signing extended key
2602 usage. This check is performed by attempting to match the OCSP
2603 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2604 in the OCSP_CERTID structures of the response.
2607 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2608 and related routines. This uses the standard OpenSSL certificate
2609 verify routines to perform initial checks (just CA validity) and
2610 to obtain the certificate chain. Then additional checks will be
2611 performed on the chain. Currently the root CA is checked to see
2612 if it is explicitly trusted for OCSP signing. This is used to set
2613 a root CA as a global signing root: that is any certificate that
2614 chains to that CA is an acceptable OCSP signing certificate.
2617 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2618 extensions from a separate configuration file.
2619 As when reading extensions from the main configuration file,
2620 the '-extensions ...' option may be used for specifying the
2622 [Massimiliano Pala <madwolf@comune.modena.it>]
2624 *) New OCSP utility. Allows OCSP requests to be generated or
2625 read. The request can be sent to a responder and the output
2626 parsed, outputed or printed in text form. Not complete yet:
2627 still needs to check the OCSP response validity.
2630 *) New subcommands for 'openssl ca':
2631 'openssl ca -status <serial>' prints the status of the cert with
2632 the given serial number (according to the index file).
2633 'openssl ca -updatedb' updates the expiry status of certificates
2635 [Massimiliano Pala <madwolf@comune.modena.it>]
2637 *) New '-newreq-nodes' command option to CA.pl. This is like
2638 '-newreq', but calls 'openssl req' with the '-nodes' option
2639 so that the resulting key is not encrypted.
2640 [Damien Miller <djm@mindrot.org>]
2642 *) New configuration for the GNU Hurd.
2643 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2645 *) Initial code to implement OCSP basic response verify. This
2646 is currently incomplete. Currently just finds the signer's
2647 certificate and verifies the signature on the response.
2650 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2651 value of OPENSSLDIR. This is available via the new '-d' option
2652 to 'openssl version', and is also included in 'openssl version -a'.
2655 *) Allowing defining memory allocation callbacks that will be given
2656 file name and line number information in additional arguments
2657 (a const char* and an int). The basic functionality remains, as
2658 well as the original possibility to just replace malloc(),
2659 realloc() and free() by functions that do not know about these
2660 additional arguments. To register and find out the current
2661 settings for extended allocation functions, the following
2662 functions are provided:
2664 CRYPTO_set_mem_ex_functions
2665 CRYPTO_set_locked_mem_ex_functions
2666 CRYPTO_get_mem_ex_functions
2667 CRYPTO_get_locked_mem_ex_functions
2669 These work the same way as CRYPTO_set_mem_functions and friends.
2670 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2671 extended allocation function is enabled.
2672 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2673 a conventional allocation function is enabled.
2674 [Richard Levitte, Bodo Moeller]
2676 *) Finish off removing the remaining LHASH function pointer casts.
2677 There should no longer be any prototype-casting required when using
2678 the LHASH abstraction, and any casts that remain are "bugs". See
2679 the callback types and macros at the head of lhash.h for details
2680 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2683 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2684 If /dev/[u]random devices are not available or do not return enough
2685 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2687 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2688 /etc/entropy will be queried once each in this sequence, quering stops
2689 when enough entropy was collected without querying more sockets.
2692 *) Change the Unix RAND_poll() variant to be able to poll several
2693 random devices, as specified by DEVRANDOM, until a sufficient amount
2694 of data has been collected. We spend at most 10 ms on each file
2695 (select timeout) and read in non-blocking mode. DEVRANDOM now
2696 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2697 (previously it was just the string "/dev/urandom"), so on typical
2698 platforms the 10 ms delay will never occur.
2699 Also separate out the Unix variant to its own file, rand_unix.c.
2700 For VMS, there's a currently-empty rand_vms.c.
2703 *) Move OCSP client related routines to ocsp_cl.c. These
2704 provide utility functions which an application needing
2705 to issue a request to an OCSP responder and analyse the
2706 response will typically need: as opposed to those which an
2707 OCSP responder itself would need which will be added later.
2709 OCSP_request_sign() signs an OCSP request with an API similar
2710 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2711 response. OCSP_response_get1_basic() extracts basic response
2712 from response. OCSP_resp_find_status(): finds and extracts status
2713 information from an OCSP_CERTID structure (which will be created
2714 when the request structure is built). These are built from lower
2715 level functions which work on OCSP_SINGLERESP structures but
2716 wont normally be used unless the application wishes to examine
2717 extensions in the OCSP response for example.
2719 Replace nonce routines with a pair of functions.
2720 OCSP_request_add1_nonce() adds a nonce value and optionally
2721 generates a random value. OCSP_check_nonce() checks the
2722 validity of the nonce in an OCSP response.
2725 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2726 This doesn't copy the supplied OCSP_CERTID and avoids the
2727 need to free up the newly created id. Change return type
2728 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2729 This can then be used to add extensions to the request.
2730 Deleted OCSP_request_new(), since most of its functionality
2731 is now in OCSP_REQUEST_new() (and the case insensitive name
2732 clash) apart from the ability to set the request name which
2733 will be added elsewhere.
2736 *) Update OCSP API. Remove obsolete extensions argument from
2737 various functions. Extensions are now handled using the new
2738 OCSP extension code. New simple OCSP HTTP function which
2739 can be used to send requests and parse the response.
2742 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2743 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2744 uses the special reorder version of SET OF to sort the attributes
2745 and reorder them to match the encoded order. This resolves a long
2746 standing problem: a verify on a PKCS7 structure just after signing
2747 it used to fail because the attribute order did not match the
2748 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2749 it uses the received order. This is necessary to tolerate some broken
2750 software that does not order SET OF. This is handled by encoding
2751 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2752 to produce the required SET OF.
2755 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2756 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2757 files to get correct declarations of the ASN.1 item variables.
2760 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2761 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2762 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2763 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2764 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2765 ASN1_ITEM and no wrapper functions.
2768 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2769 replace the old function pointer based I/O routines. Change most of
2770 the *_d2i_bio() and *_d2i_fp() functions to use these.
2773 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2774 lines, recognice more "algorithms" that can be deselected, and make
2775 it complain about algorithm deselection that isn't recognised.
2778 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2779 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2780 to use new functions. Add NO_ASN1_OLD which can be set to remove
2781 some old style ASN1 functions: this can be used to determine if old
2782 code will still work when these eventually go away.
2785 *) New extension functions for OCSP structures, these follow the
2786 same conventions as certificates and CRLs.
2789 *) New function X509V3_add1_i2d(). This automatically encodes and
2790 adds an extension. Its behaviour can be customised with various
2791 flags to append, replace or delete. Various wrappers added for
2792 certifcates and CRLs.
2795 *) Fix to avoid calling the underlying ASN1 print routine when
2796 an extension cannot be parsed. Correct a typo in the
2797 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2800 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2801 entries for variables.
2804 *) Add functionality to apps/openssl.c for detecting locking
2805 problems: As the program is single-threaded, all we have
2806 to do is register a locking callback using an array for
2807 storing which locks are currently held by the program.
2810 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2811 SSL_get_ex_data_X509_STORE_idx(), which is used in
2812 ssl_verify_cert_chain() and thus can be called at any time
2813 during TLS/SSL handshakes so that thread-safety is essential.
2814 Unfortunately, the ex_data design is not at all suited
2815 for multi-threaded use, so it probably should be abolished.
2818 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2819 [Broadcom, tweaked and integrated by Geoff Thorpe]
2821 *) Move common extension printing code to new function
2822 X509V3_print_extensions(). Reorganise OCSP print routines and
2823 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2826 *) New function X509_signature_print() to remove duplication in some
2830 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2831 set (this was treated exactly the same as SET OF previously). This
2832 is used to reorder the STACK representing the structure to match the
2833 encoding. This will be used to get round a problem where a PKCS7
2834 structure which was signed could not be verified because the STACK
2835 order did not reflect the encoded order.
2838 *) Reimplement the OCSP ASN1 module using the new code.
2841 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2842 for its ASN1 operations. The old style function pointers still exist
2843 for now but they will eventually go away.
2846 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2847 completely replaces the old ASN1 functionality with a table driven
2848 encoder and decoder which interprets an ASN1_ITEM structure describing
2849 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2850 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2851 has also been converted to the new form.
2854 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2855 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2856 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2857 for negative moduli.
2860 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2861 of not touching the result's sign bit.
2864 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2868 *) Changed the LHASH code to use prototypes for callbacks, and created
2869 macros to declare and implement thin (optionally static) functions
2870 that provide type-safety and avoid function pointer casting for the
2871 type-specific callbacks.
2874 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2876 [Veers Staats <staatsvr@asc.hpc.mil>,
2877 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2879 *) Reformat the FAQ so the different questions and answers can be divided
2880 in sections depending on the subject.
2883 *) Have the zlib compression code load ZLIB.DLL dynamically under
2887 *) New function BN_mod_sqrt for computing square roots modulo a prime
2888 (using the probabilistic Tonelli-Shanks algorithm unless
2889 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2890 be handled deterministically).
2891 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2893 *) Make BN_mod_inverse faster by explicitly handling small quotients
2894 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2895 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2898 *) New function BN_kronecker.
2901 *) Fix BN_gcd so that it works on negative inputs; the result is
2902 positive unless both parameters are zero.
2903 Previously something reasonably close to an infinite loop was
2904 possible because numbers could be growing instead of shrinking
2905 in the implementation of Euclid's algorithm.
2908 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2909 sign of the number in question.
2911 Fix BN_is_word(a,w) to work correctly for w == 0.
2913 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2914 because its test if the absolute value of 'a' equals 'w'.
2915 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2916 it exists mostly for use in the implementations of BN_is_zero(),
2917 BN_is_one(), and BN_is_word().
2920 *) New function BN_swap.
2923 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2924 the exponentiation functions are more likely to produce reasonable
2925 results on negative inputs.
2928 *) Change BN_mod_mul so that the result is always non-negative.
2929 Previously, it could be negative if one of the factors was negative;
2930 I don't think anyone really wanted that behaviour.
2933 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2934 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2935 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2936 and add new functions:
2945 BN_mod_lshift1_quick
2949 These functions always generate non-negative results.
2951 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2952 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2954 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2955 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2956 be reduced modulo m.
2957 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2960 The following entry accidentily appeared in the CHANGES file
2961 distributed with OpenSSL 0.9.7. The modifications described in
2962 it do *not* apply to OpenSSL 0.9.7.
2964 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2965 was actually never needed) and in BN_mul(). The removal in BN_mul()
2966 required a small change in bn_mul_part_recursive() and the addition
2967 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2968 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2969 bn_sub_words() and bn_add_words() except they take arrays with
2974 *) In 'openssl passwd', verify passwords read from the terminal
2975 unless the '-salt' option is used (which usually means that
2976 verification would just waste user's time since the resulting
2977 hash is going to be compared with some given password hash)
2978 or the new '-noverify' option is used.
2980 This is an incompatible change, but it does not affect
2981 non-interactive use of 'openssl passwd' (passwords on the command
2982 line, '-stdin' option, '-in ...' option) and thus should not
2986 *) Remove all references to RSAref, since there's no more need for it.
2989 *) Make DSO load along a path given through an environment variable
2990 (SHLIB_PATH) with shl_load().
2993 *) Constify the ENGINE code as a result of BIGNUM constification.
2994 Also constify the RSA code and most things related to it. In a
2995 few places, most notable in the depth of the ASN.1 code, ugly
2996 casts back to non-const were required (to be solved at a later
3000 *) Make it so the openssl application has all engines loaded by default.
3003 *) Constify the BIGNUM routines a little more.
3006 *) Add the following functions:
3008 ENGINE_load_cswift()
3010 ENGINE_load_atalla()
3012 ENGINE_load_builtin_engines()
3014 That way, an application can itself choose if external engines that
3015 are built-in in OpenSSL shall ever be used or not. The benefit is
3016 that applications won't have to be linked with libdl or other dso
3017 libraries unless it's really needed.
3019 Changed 'openssl engine' to load all engines on demand.
3020 Changed the engine header files to avoid the duplication of some
3021 declarations (they differed!).
3024 *) 'openssl engine' can now list capabilities.
3027 *) Better error reporting in 'openssl engine'.
3030 *) Never call load_dh_param(NULL) in s_server.
3033 *) Add engine application. It can currently list engines by name and
3034 identity, and test if they are actually available.
3037 *) Improve RPM specification file by forcing symbolic linking and making
3038 sure the installed documentation is also owned by root.root.
3039 [Damien Miller <djm@mindrot.org>]
3041 *) Give the OpenSSL applications more possibilities to make use of
3042 keys (public as well as private) handled by engines.
3045 *) Add OCSP code that comes from CertCo.
3048 *) Add VMS support for the Rijndael code.
3051 *) Added untested support for Nuron crypto accelerator.
3054 *) Add support for external cryptographic devices. This code was
3055 previously distributed separately as the "engine" branch.
3056 [Geoff Thorpe, Richard Levitte]
3058 *) Rework the filename-translation in the DSO code. It is now possible to
3059 have far greater control over how a "name" is turned into a filename
3060 depending on the operating environment and any oddities about the
3061 different shared library filenames on each system.
3064 *) Support threads on FreeBSD-elf in Configure.
3067 *) Fix for SHA1 assembly problem with MASM: it produces
3068 warnings about corrupt line number information when assembling
3069 with debugging information. This is caused by the overlapping
3071 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3074 NCONF_get_number() has no error checking at all. As a replacement,
3075 NCONF_get_number_e() is defined (_e for "error checking") and is
3076 promoted strongly. The old NCONF_get_number is kept around for
3077 binary backward compatibility.
3078 Make it possible for methods to load from something other than a BIO,
3079 by providing a function pointer that is given a name instead of a BIO.
3080 For example, this could be used to load configuration data from an
3084 *) Fix for non blocking accept BIOs. Added new I/O special reason
3085 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3086 with non blocking I/O was not possible because no retry code was
3087 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3091 *) Added the beginnings of Rijndael support.
3094 *) Fix for bug in DirectoryString mask setting. Add support for
3095 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3096 to allow certificate printing to more controllable, additional
3097 'certopt' option to 'x509' to allow new printing options to be
3101 *) Clean old EAY MD5 hack from e_os.h.
3104 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3106 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3107 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
3108 [Joe Orton, Steve Henson]
3110 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3112 *) Fix additional bug revealed by the NISCC test suite:
3114 Stop bug triggering large recursion when presented with
3115 certain ASN.1 tags (CAN-2003-0851)
3118 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3120 *) Fix various bugs revealed by running the NISCC test suite:
3122 Stop out of bounds reads in the ASN1 code when presented with
3123 invalid tags (CAN-2003-0543 and CAN-2003-0544).
3125 If verify callback ignores invalid public key errors don't try to check
3126 certificate signature with the NULL public key.
3130 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3131 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3135 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3136 extra data after the compression methods not only for TLS 1.0
3137 but also for SSL 3.0 (as required by the specification).
3138 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3140 *) Change X509_certificate_type() to mark the key as exported/exportable
3141 when it's 512 *bits* long, not 512 bytes.
3144 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3146 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3147 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3148 a protocol version number mismatch like a decryption error
3149 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3152 *) Turn on RSA blinding by default in the default implementation
3153 to avoid a timing attack. Applications that don't want it can call
3154 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3155 They would be ill-advised to do so in most cases.
3156 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3158 *) Change RSA blinding code so that it works when the PRNG is not
3159 seeded (in this case, the secret RSA exponent is abused as
3160 an unpredictable seed -- if it is not unpredictable, there
3161 is no point in blinding anyway). Make RSA blinding thread-safe
3162 by remembering the creator's thread ID in rsa->blinding and
3163 having all other threads use local one-time blinding factors
3164 (this requires more computation than sharing rsa->blinding, but
3165 avoids excessive locking; and if an RSA object is not shared
3166 between threads, blinding will still be very fast).
3169 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3171 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3172 via timing by performing a MAC computation even if incorrrect
3173 block cipher padding has been found. This is a countermeasure
3174 against active attacks where the attacker has to distinguish
3175 between bad padding and a MAC verification error. (CAN-2003-0078)
3177 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3178 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3179 Martin Vuagnoux (EPFL, Ilion)]
3181 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3183 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3184 memory from it's contents. This is done with a counter that will
3185 place alternating values in each byte. This can be used to solve
3186 two issues: 1) the removal of calls to memset() by highly optimizing
3187 compilers, and 2) cleansing with other values than 0, since those can
3188 be read through on certain media, for example a swap space on disk.
3191 *) Bugfix: client side session caching did not work with external caching,
3192 because the session->cipher setting was not restored when reloading
3193 from the external cache. This problem was masked, when
3194 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3195 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3198 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3199 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3200 [Zeev Lieber <zeev-l@yahoo.com>]
3202 *) Undo an undocumented change introduced in 0.9.6e which caused
3203 repeated calls to OpenSSL_add_all_ciphers() and
3204 OpenSSL_add_all_digests() to be ignored, even after calling
3208 *) Change the default configuration reader to deal with last line not
3209 being properly terminated.
3212 *) Change X509_NAME_cmp() so it applies the special rules on handling
3213 DN values that are of type PrintableString, as well as RDNs of type
3214 emailAddress where the value has the type ia5String.
3215 [stefank@valicert.com via Richard Levitte]
3217 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3218 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3219 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3220 the bitwise-OR of the two for use by the majority of applications
3221 wanting this behaviour, and update the docs. The documented
3222 behaviour and actual behaviour were inconsistent and had been
3223 changing anyway, so this is more a bug-fix than a behavioural
3225 [Geoff Thorpe, diagnosed by Nadav Har'El]
3227 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3228 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3231 *) Fix initialization code race conditions in
3232 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3233 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3234 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3235 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3236 ssl2_get_cipher_by_char(),
3237 ssl3_get_cipher_by_char().
3238 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3240 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3241 the cached sessions are flushed, as the remove_cb() might use ex_data
3242 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3243 (see [openssl.org #212]).
3244 [Geoff Thorpe, Lutz Jaenicke]
3246 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3247 length, instead of the encoding length to d2i_ASN1_OBJECT.
3250 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3252 *) [In 0.9.6g-engine release:]
3253 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3254 [Lynn Gazis <lgazis@rainbow.com>]
3256 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3258 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3259 and get fix the header length calculation.
3260 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3261 Alon Kantor <alonk@checkpoint.com> (and others),
3264 *) Use proper error handling instead of 'assertions' in buffer
3265 overflow checks added in 0.9.6e. This prevents DoS (the
3266 assertions could call abort()).
3267 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3269 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3271 *) Add various sanity checks to asn1_get_length() to reject
3272 the ASN1 length bytes if they exceed sizeof(long), will appear
3273 negative or the content length exceeds the length of the
3275 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3277 *) Fix cipher selection routines: ciphers without encryption had no flags
3278 for the cipher strength set and where therefore not handled correctly
3279 by the selection routines (PR #130).
3282 *) Fix EVP_dsa_sha macro.
3286 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3287 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3288 that was added in OpenSSL 0.9.6d.
3290 As the countermeasure turned out to be incompatible with some
3291 broken SSL implementations, the new option is part of SSL_OP_ALL.
3292 SSL_OP_ALL is usually employed when compatibility with weird SSL
3293 implementations is desired (e.g. '-bugs' option to 's_client' and
3294 's_server'), so the new option is automatically set in many
3298 *) Changes in security patch:
3300 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3301 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3302 Air Force Materiel Command, USAF, under agreement number
3305 *) Add various sanity checks to asn1_get_length() to reject
3306 the ASN1 length bytes if they exceed sizeof(long), will appear
3307 negative or the content length exceeds the length of the
3308 supplied buffer. (CAN-2002-0659)
3309 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3311 *) Assertions for various potential buffer overflows, not known to
3313 [Ben Laurie (CHATS)]
3315 *) Various temporary buffers to hold ASCII versions of integers were
3316 too small for 64 bit platforms. (CAN-2002-0655)
3317 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3319 *) Remote buffer overflow in SSL3 protocol - an attacker could
3320 supply an oversized session ID to a client. (CAN-2002-0656)
3321 [Ben Laurie (CHATS)]
3323 *) Remote buffer overflow in SSL2 protocol - an attacker could
3324 supply an oversized client master key. (CAN-2002-0656)
3325 [Ben Laurie (CHATS)]
3327 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3329 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3330 encoded as NULL) with id-dsa-with-sha1.
3331 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3333 *) Check various X509_...() return values in apps/req.c.
3334 [Nils Larsch <nla@trustcenter.de>]
3336 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3337 an end-of-file condition would erronously be flagged, when the CRLF
3338 was just at the end of a processed block. The bug was discovered when
3339 processing data through a buffering memory BIO handing the data to a
3340 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3341 <ptsekov@syntrex.com> and Nedelcho Stanev.
3344 *) Implement a countermeasure against a vulnerability recently found
3345 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3346 before application data chunks to avoid the use of known IVs
3347 with data potentially chosen by the attacker.
3350 *) Fix length checks in ssl3_get_client_hello().
3353 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3354 to prevent ssl3_read_internal() from incorrectly assuming that
3355 ssl3_read_bytes() found application data while handshake
3356 processing was enabled when in fact s->s3->in_read_app_data was
3357 merely automatically cleared during the initial handshake.
3358 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3360 *) Fix object definitions for Private and Enterprise: they were not
3361 recognized in their shortname (=lowercase) representation. Extend
3362 obj_dat.pl to issue an error when using undefined keywords instead
3363 of silently ignoring the problem (Svenning Sorensen
3364 <sss@sss.dnsalias.net>).
3367 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3368 generators, i.e. generators other than 2 and 5. (Previously, the
3369 code did not properly initialise the 'add' and 'rem' values to
3370 BN_generate_prime().)
3372 In the new general case, we do not insist that 'generator' is
3373 actually a primitive root: This requirement is rather pointless;
3374 a generator of the order-q subgroup is just as good, if not
3378 *) Map new X509 verification errors to alerts. Discovered and submitted by
3379 Tom Wu <tom@arcot.com>.
3382 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3383 returning non-zero before the data has been completely received
3384 when using non-blocking I/O.
3385 [Bodo Moeller; problem pointed out by John Hughes]
3387 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3388 [Ben Laurie, Lutz Jaenicke]
3390 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3391 Yoram Zahavi <YoramZ@gilian.com>).
3394 *) Add information about CygWin 1.3 and on, and preserve proper
3395 configuration for the versions before that.
3396 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3398 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3399 check whether we deal with a copy of a session and do not delete from
3400 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3401 <izhar@checkpoint.com>.
3404 *) Do not store session data into the internal session cache, if it
3405 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3406 flag is set). Proposed by Aslam <aslam@funk.com>.
3409 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3413 *) [In 0.9.6d-engine release:]
3414 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3415 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3417 *) Add the configuration target linux-s390x.
3418 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3420 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3421 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3422 variable as an indication that a ClientHello message has been
3423 received. As the flag value will be lost between multiple
3424 invocations of ssl3_accept when using non-blocking I/O, the
3425 function may not be aware that a handshake has actually taken
3426 place, thus preventing a new session from being added to the
3429 To avoid this problem, we now set s->new_session to 2 instead of
3430 using a local variable.
3431 [Lutz Jaenicke, Bodo Moeller]
3433 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3434 if the SSL_R_LENGTH_MISMATCH error is detected.
3435 [Geoff Thorpe, Bodo Moeller]
3437 *) New 'shared_ldflag' column in Configure platform table.
3440 *) Fix EVP_CIPHER_mode macro.
3441 ["Dan S. Camper" <dan@bti.net>]
3443 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3444 type, we must throw them away by setting rr->length to 0.
3445 [D P Chang <dpc@qualys.com>]
3447 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3449 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3450 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3451 worked incorrectly for those cases where range = 10..._2 and
3452 3*range is two bits longer than range.)
3455 *) Only add signing time to PKCS7 structures if it is not already
3459 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3460 OBJ_ld_ce should be OBJ_id_ce.
3461 Also some ip-pda OIDs in crypto/objects/objects.txt were
3462 incorrect (cf. RFC 3039).
3463 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3465 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3466 returns early because it has nothing to do.
3467 [Andy Schneider <andy.schneider@bjss.co.uk>]
3469 *) [In 0.9.6c-engine release:]
3470 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3471 [Andy Schneider <andy.schneider@bjss.co.uk>]
3473 *) [In 0.9.6c-engine release:]
3474 Add support for Cryptographic Appliance's keyserver technology.
3475 (Use engine 'keyclient')
3476 [Cryptographic Appliances and Geoff Thorpe]
3478 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3479 is called via tools/c89.sh because arguments have to be
3480 rearranged (all '-L' options must appear before the first object
3482 [Richard Shapiro <rshapiro@abinitio.com>]
3484 *) [In 0.9.6c-engine release:]
3485 Add support for Broadcom crypto accelerator cards, backported
3487 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3489 *) [In 0.9.6c-engine release:]
3490 Add support for SureWare crypto accelerator cards from
3491 Baltimore Technologies. (Use engine 'sureware')
3492 [Baltimore Technologies and Mark Cox]
3494 *) [In 0.9.6c-engine release:]
3495 Add support for crypto accelerator cards from Accelerated
3496 Encryption Processing, www.aep.ie. (Use engine 'aep')
3497 [AEP Inc. and Mark Cox]
3499 *) Add a configuration entry for gcc on UnixWare.
3500 [Gary Benson <gbenson@redhat.com>]
3502 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3503 messages are stored in a single piece (fixed-length part and
3504 variable-length part combined) and fix various bugs found on the way.
3507 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3508 instead. BIO_gethostbyname() does not know what timeouts are
3509 appropriate, so entries would stay in cache even when they have
3511 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3513 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3514 faced with a pathologically small ClientHello fragment that does
3515 not contain client_version: Instead of aborting with an error,
3516 simply choose the highest available protocol version (i.e.,
3517 TLS 1.0 unless it is disabled). In practice, ClientHello
3518 messages are never sent like this, but this change gives us
3519 strictly correct behaviour at least for TLS.
3522 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3523 never resets s->method to s->ctx->method when called from within
3524 one of the SSL handshake functions.
3525 [Bodo Moeller; problem pointed out by Niko Baric]
3527 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3528 (sent using the client's version number) if client_version is
3529 smaller than the protocol version in use. Also change
3530 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3531 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3532 the client will at least see that alert.
3535 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3539 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3540 client receives HelloRequest while in a handshake.
3541 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3543 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3544 should end in 'break', not 'goto end' which circuments various
3545 cleanups done in state SSL_ST_OK. But session related stuff
3546 must be disabled for SSL_ST_OK in the case that we just sent a
3549 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3550 before just sending a HelloRequest.
3551 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3553 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3554 reveal whether illegal block cipher padding was found or a MAC
3555 verification error occured. (Neither SSLerr() codes nor alerts
3556 are directly visible to potential attackers, but the information
3557 may leak via logfiles.)
3559 Similar changes are not required for the SSL 2.0 implementation
3560 because the number of padding bytes is sent in clear for SSL 2.0,
3561 and the extra bytes are just ignored. However ssl/s2_pkt.c
3562 failed to verify that the purported number of padding bytes is in
3566 *) Add OpenUNIX-8 support including shared libraries
3567 (Boyd Lynn Gerber <gerberb@zenez.com>).
3570 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3571 'wristwatch attack' using huge encoding parameters (cf.
3572 James H. Manger's CRYPTO 2001 paper). Note that the
3573 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3574 encoding parameters and hence was not vulnerable.
3577 *) BN_sqr() bug fix.
3578 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3580 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3581 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3582 followed by modular reduction.
3583 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3585 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3586 equivalent based on BN_pseudo_rand() instead of BN_rand().
3589 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3590 This function was broken, as the check for a new client hello message
3591 to handle SGC did not allow these large messages.
3592 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3595 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3598 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3599 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3602 *) Rework the configuration and shared library support for Tru64 Unix.
3603 The configuration part makes use of modern compiler features and
3604 still retains old compiler behavior for those that run older versions
3605 of the OS. The shared library support part includes a variant that
3606 uses the RPATH feature, and is available through the special
3607 configuration target "alpha-cc-rpath", which will never be selected
3609 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3611 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3612 with the same message size as in ssl3_get_certificate_request().
3613 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3614 messages might inadvertently be reject as too long.
3615 [Petr Lampa <lampa@fee.vutbr.cz>]
3617 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3620 *) Modified SSL library such that the verify_callback that has been set
3621 specificly for an SSL object with SSL_set_verify() is actually being
3622 used. Before the change, a verify_callback set with this function was
3623 ignored and the verify_callback() set in the SSL_CTX at the time of
3624 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3625 to allow the necessary settings.
3628 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3629 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3630 done automatically (in contradiction to the requirements of the C
3631 standard). This made problems when used from OpenSSH.
3634 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3635 dh->length and always used
3637 BN_rand_range(priv_key, dh->p).
3639 BN_rand_range() is not necessary for Diffie-Hellman, and this
3640 specific range makes Diffie-Hellman unnecessarily inefficient if
3641 dh->length (recommended exponent length) is much smaller than the
3642 length of dh->p. We could use BN_rand_range() if the order of
3643 the subgroup was stored in the DH structure, but we only have
3648 BN_rand(priv_key, l, ...)
3650 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3656 RSA_eay_public_encrypt
3657 RSA_eay_private_decrypt
3658 RSA_eay_private_encrypt (signing)
3659 RSA_eay_public_decrypt (signature verification)
3661 (default implementations for RSA_public_encrypt,
3662 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3663 always reject numbers >= n.
3666 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3667 to synchronize access to 'locking_thread'. This is necessary on
3668 systems where access to 'locking_thread' (an 'unsigned long'
3669 variable) is not atomic.
3672 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3673 *before* setting the 'crypto_lock_rand' flag. The previous code had
3674 a race condition if 0 is a valid thread ID.
3675 [Travis Vitek <vitek@roguewave.com>]
3677 *) Add support for shared libraries under Irix.
3678 [Albert Chin-A-Young <china@thewrittenword.com>]
3680 *) Add configuration option to build on Linux on both big-endian and
3682 [Ralf Baechle <ralf@uni-koblenz.de>]
3684 *) Add the possibility to create shared libraries on HP-UX.
3687 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3689 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3690 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3691 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3692 PRNG state recovery was possible based on the output of
3693 one PRNG request appropriately sized to gain knowledge on
3694 'md' followed by enough consecutive 1-byte PRNG requests
3695 to traverse all of 'state'.
3697 1. When updating 'md_local' (the current thread's copy of 'md')
3698 during PRNG output generation, hash all of the previous
3699 'md_local' value, not just the half used for PRNG output.
3701 2. Make the number of bytes from 'state' included into the hash
3702 independent from the number of PRNG bytes requested.
3704 The first measure alone would be sufficient to avoid
3705 Markku-Juhani's attack. (Actually it had never occurred
3706 to me that the half of 'md_local' used for chaining was the
3707 half from which PRNG output bytes were taken -- I had always
3708 assumed that the secret half would be used.) The second
3709 measure makes sure that additional data from 'state' is never
3710 mixed into 'md_local' in small portions; this heuristically
3711 further strengthens the PRNG.
3714 *) Fix crypto/bn/asm/mips3.s.
3717 *) When only the key is given to "enc", the IV is undefined. Print out
3718 an error message in this case.
3721 *) Handle special case when X509_NAME is empty in X509 printing routines.
3724 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3725 positive and less than q.
3728 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3729 used: it isn't thread safe and the add_lock_callback should handle
3731 [Paul Rose <Paul.Rose@bridge.com>]
3733 *) Verify that incoming data obeys the block size in
3734 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3738 [Ulf Möller, Bodo Möller]
3740 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3741 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3742 when fixing the server behaviour for backwards-compatible 'client
3743 hello' messages. (Note that the attack is impractical against
3744 SSL 3.0 and TLS 1.0 anyway because length and version checking
3745 means that the probability of guessing a valid ciphertext is
3746 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3749 Before 0.9.5, the countermeasure (hide the error by generating a
3750 random 'decryption result') did not work properly because
3751 ERR_clear_error() was missing, meaning that SSL_get_error() would
3752 detect the supposedly ignored error.
3754 Both problems are now fixed.
3757 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3758 (previously it was 1024).
3761 *) Fix for compatibility mode trust settings: ignore trust settings
3762 unless some valid trust or reject settings are present.
3765 *) Fix for blowfish EVP: its a variable length cipher.
3768 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3769 parameters in DSA public key structures and return an error in the
3770 DSA routines if parameters are absent.
3773 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3774 in the current directory if neither $RANDFILE nor $HOME was set.
3775 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3776 caused some confusion to Windows users who haven't defined $HOME.
3777 Thus RAND_file_name() is changed again: e_os.h can define a
3778 DEFAULT_HOME, which will be used if $HOME is not set.
3779 For Windows, we use "C:"; on other platforms, we still require
3780 environment variables.
3782 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3783 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3784 having multiple threads call RAND_poll() concurrently.
3787 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3788 combination of a flag and a thread ID variable.
3789 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3790 flag), *other* threads can enter ssleay_add_bytes without obeying
3791 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3792 that they do not hold after the first thread unsets add_do_not_lock).
3795 *) Change bctest again: '-x' expressions are not available in all
3799 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3801 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3802 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3804 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3805 the default extension for executables, if any. Also, make the perl
3806 scripts that use symlink() to test if it really exists and use "cp"
3807 if it doesn't. All this made OpenSSL compilable and installable in
3811 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3812 If SEQUENCE is length is indefinite just set c->slen to the total
3813 amount of data available.
3814 [Steve Henson, reported by shige@FreeBSD.org]
3815 [This change does not apply to 0.9.7.]
3817 *) Change bctest to avoid here-documents inside command substitution
3818 (workaround for FreeBSD /bin/sh bug).
3819 For compatibility with Ultrix, avoid shell functions (introduced
3820 in the bctest version that searches along $PATH).
3823 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3824 with des_encrypt() defined on some operating systems, like Solaris
3828 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3829 On the Importance of Eliminating Errors in Cryptographic
3830 Computations, J. Cryptology 14 (2001) 2, 101-119,
3831 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3834 *) MIPS assembler BIGNUM division bug fix.
3837 *) Disabled incorrect Alpha assembler code.
3840 *) Fix PKCS#7 decode routines so they correctly update the length
3841 after reading an EOC for the EXPLICIT tag.
3843 [This change does not apply to 0.9.7.]
3845 *) Fix bug in PKCS#12 key generation routines. This was triggered
3846 if a 3DES key was generated with a 0 initial byte. Include
3847 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3848 (but broken) behaviour.
3851 *) Enhance bctest to search for a working bc along $PATH and print
3853 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3855 *) Fix memory leaks in err.c: free err_data string if necessary;
3856 don't write to the wrong index in ERR_set_error_data.
3859 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3863 *) Replace rdtsc with _emit statements for VC++ version 5.
3864 [Jeremy Cooper <jeremy@baymoo.org>]
3866 *) Make it possible to reuse SSLv2 sessions.
3869 *) In copy_email() check for >= 0 as a return value for
3870 X509_NAME_get_index_by_NID() since 0 is a valid index.
3871 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3873 *) Avoid coredump with unsupported or invalid public keys by checking if
3874 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3875 PKCS7_verify() fails with non detached data.
3878 *) Don't use getenv in library functions when run as setuid/setgid.
3879 New function OPENSSL_issetugid().
3882 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3883 due to incorrect handling of multi-threading:
3885 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3887 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3889 3. Count how many times MemCheck_off() has been called so that
3890 nested use can be treated correctly. This also avoids
3891 inband-signalling in the previous code (which relied on the
3892 assumption that thread ID 0 is impossible).
3895 *) Add "-rand" option also to s_client and s_server.
3898 *) Fix CPU detection on Irix 6.x.
3899 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3900 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3902 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3905 [This change does not apply to 0.9.7.]
3907 *) Use the cached encoding of an X509_NAME structure rather than
3908 copying it. This is apparently the reason for the libsafe "errors"
3909 but the code is actually correct.
3912 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3913 Bleichenbacher's DSA attack.
3914 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3915 to be set and top=0 forces the highest bit to be set; top=-1 is new
3916 and leaves the highest bit random.
3917 [Ulf Moeller, Bodo Moeller]
3919 *) In the NCONF_...-based implementations for CONF_... queries
3920 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3921 a temporary CONF structure with the data component set to NULL
3922 (which gives segmentation faults in lh_retrieve).
3923 Instead, use NULL for the CONF pointer in CONF_get_string and
3924 CONF_get_number (which may use environment variables) and directly
3925 return NULL from CONF_get_section.
3928 *) Fix potential buffer overrun for EBCDIC.
3931 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3932 keyUsage if basicConstraints absent for a CA.
3935 *) Make SMIME_write_PKCS7() write mail header values with a format that
3936 is more generally accepted (no spaces before the semicolon), since
3937 some programs can't parse those values properly otherwise. Also make
3938 sure BIO's that break lines after each write do not create invalid
3942 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3943 macros previously used would not encode an empty SEQUENCE OF
3944 and break the signature.
3946 [This change does not apply to 0.9.7.]
3948 *) Zero the premaster secret after deriving the master secret in
3952 *) Add some EVP_add_digest_alias registrations (as found in
3953 OpenSSL_add_all_digests()) to SSL_library_init()
3954 aka OpenSSL_add_ssl_algorithms(). This provides improved
3955 compatibility with peers using X.509 certificates
3956 with unconventional AlgorithmIdentifier OIDs.
3959 *) Fix for Irix with NO_ASM.
3960 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3962 *) ./config script fixes.
3963 [Ulf Moeller, Richard Levitte]
3965 *) Fix 'openssl passwd -1'.
3968 *) Change PKCS12_key_gen_asc() so it can cope with non null
3969 terminated strings whose length is passed in the passlen
3970 parameter, for example from PEM callbacks. This was done
3971 by adding an extra length parameter to asc2uni().
3972 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3974 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3975 call failed, free the DSA structure.
3978 *) Fix to uni2asc() to cope with zero length Unicode strings.
3979 These are present in some PKCS#12 files.
3982 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3983 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3984 when writing a 32767 byte record.
3985 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3987 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3988 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3990 (RSA objects have a reference count access to which is protected
3991 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3992 so they are meant to be shared between threads.)
3993 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3994 "Reddie, Steven" <Steven.Reddie@ca.com>]
3996 *) Fix a deadlock in CRYPTO_mem_leaks().
3999 *) Use better test patterns in bntest.
4002 *) rand_win.c fix for Borland C.
4005 *) BN_rshift bugfix for n == 0.
4008 *) Add a 'bctest' script that checks for some known 'bc' bugs
4009 so that 'make test' does not abort just because 'bc' is broken.
4012 *) Store verify_result within SSL_SESSION also for client side to
4013 avoid potential security hole. (Re-used sessions on the client side
4014 always resulted in verify_result==X509_V_OK, not using the original
4015 result of the server certificate verification.)
4018 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4019 SSL3_RT_APPLICATION_DATA, return 0.
4020 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4024 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4025 releases, have been re-implemented by renaming the previous
4026 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4027 and ssl3_read_internal, respectively, and adding 'peek' parameters
4028 to them. The new ssl[23]_{read,peek} functions are calls to
4029 ssl[23]_read_internal with the 'peek' flag set appropriately.
4030 A 'peek' parameter has also been added to ssl3_read_bytes, which
4031 does the actual work for ssl3_read_internal.
4034 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4035 the method-specific "init()" handler. Also clean up ex_data after
4036 calling the method-specific "finish()" handler. Previously, this was
4037 happening the other way round.
4040 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4041 The previous value, 12, was not always sufficient for BN_mod_exp().
4044 *) Make sure that shared libraries get the internal name engine with
4045 the full version number and not just 0. This should mark the
4046 shared libraries as not backward compatible. Of course, this should
4047 be changed again when we can guarantee backward binary compatibility.
4050 *) Fix typo in get_cert_by_subject() in by_dir.c
4051 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4053 *) Rework the system to generate shared libraries:
4055 - Make note of the expected extension for the shared libraries and
4056 if there is a need for symbolic links from for example libcrypto.so.0
4057 to libcrypto.so.0.9.7. There is extended info in Configure for
4060 - Make as few rebuilds of the shared libraries as possible.
4062 - Still avoid linking the OpenSSL programs with the shared libraries.
4064 - When installing, install the shared libraries separately from the
4068 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4070 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4071 and not in SSL_clear because the latter is also used by the
4072 accept/connect functions; previously, the settings made by
4073 SSL_set_read_ahead would be lost during the handshake.
4074 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4076 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4077 Previously, it would create entries for disableed algorithms no
4081 *) Added several new manual pages for SSL_* function.
4084 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4086 *) In ssl23_get_client_hello, generate an error message when faced
4087 with an initial SSL 3.0/TLS record that is too small to contain the
4088 first two bytes of the ClientHello message, i.e. client_version.
4089 (Note that this is a pathologic case that probably has never happened
4090 in real life.) The previous approach was to use the version number
4091 from the record header as a substitute; but our protocol choice
4092 should not depend on that one because it is not authenticated
4093 by the Finished messages.
4096 *) More robust randomness gathering functions for Windows.
4097 [Jeffrey Altman <jaltman@columbia.edu>]
4099 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4100 not set then we don't setup the error code for issuer check errors
4101 to avoid possibly overwriting other errors which the callback does
4102 handle. If an application does set the flag then we assume it knows
4103 what it is doing and can handle the new informational codes
4107 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4108 a general "ANY" type, as such it should be able to decode anything
4109 including tagged types. However it didn't check the class so it would
4110 wrongly interpret tagged types in the same way as their universal
4111 counterpart and unknown types were just rejected. Changed so that the
4112 tagged and unknown types are handled in the same way as a SEQUENCE:
4113 that is the encoding is stored intact. There is also a new type
4114 "V_ASN1_OTHER" which is used when the class is not universal, in this
4115 case we have no idea what the actual type is so we just lump them all
4119 *) On VMS, stdout may very well lead to a file that is written to
4120 in a record-oriented fashion. That means that every write() will
4121 write a separate record, which will be read separately by the
4122 programs trying to read from it. This can be very confusing.
4124 The solution is to put a BIO filter in the way that will buffer
4125 text until a linefeed is reached, and then write everything a
4126 line at a time, so every record written will be an actual line,
4127 not chunks of lines and not (usually doesn't happen, but I've
4128 seen it once) several lines in one record. BIO_f_linebuffer() is
4131 Currently, it's a VMS-only method, because that's where it has
4132 been tested well enough.
4135 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4136 it can return incorrect results.
4137 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4138 but it was in 0.9.6-beta[12].)
4141 *) Disable the check for content being present when verifying detached
4142 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4143 include zero length content when signing messages.
4146 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4147 BIO_ctrl (for BIO pairs).
4150 *) Add DSO method for VMS.
4153 *) Bug fix: Montgomery multiplication could produce results with the
4157 *) Add RPM specification openssl.spec and modify it to build three
4158 packages. The default package contains applications, application
4159 documentation and run-time libraries. The devel package contains
4160 include files, static libraries and function documentation. The
4161 doc package contains the contents of the doc directory. The original
4162 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4165 *) Add a large number of documentation files for many SSL routines.
4166 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4168 *) Add a configuration entry for Sony News 4.
4169 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4171 *) Don't set the two most significant bits to one when generating a
4172 random number < q in the DSA library.
4175 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4176 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4177 the underlying transport is blocking) if a handshake took place.
4178 (The default behaviour is needed by applications such as s_client
4179 and s_server that use select() to determine when to use SSL_read;
4180 but for applications that know in advance when to expect data, it
4181 just makes things more complicated.)
4184 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4188 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4189 work better on such systems.
4190 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4192 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4193 Update PKCS12_parse() so it copies the friendlyName and the
4194 keyid to the certificates aux info.
4197 *) Fix bug in PKCS7_verify() which caused an infinite loop
4198 if there was more than one signature.
4199 [Sven Uszpelkat <su@celocom.de>]
4201 *) Major change in util/mkdef.pl to include extra information
4202 about each symbol, as well as presentig variables as well
4203 as functions. This change means that there's n more need
4204 to rebuild the .num files when some algorithms are excluded.
4207 *) Allow the verify time to be set by an application,
4208 rather than always using the current time.
4211 *) Phase 2 verify code reorganisation. The certificate
4212 verify code now looks up an issuer certificate by a
4213 number of criteria: subject name, authority key id
4214 and key usage. It also verifies self signed certificates
4215 by the same criteria. The main comparison function is
4216 X509_check_issued() which performs these checks.
4218 Lot of changes were necessary in order to support this
4219 without completely rewriting the lookup code.
4221 Authority and subject key identifier are now cached.
4223 The LHASH 'certs' is X509_STORE has now been replaced
4224 by a STACK_OF(X509_OBJECT). This is mainly because an
4225 LHASH can't store or retrieve multiple objects with
4226 the same hash value.
4228 As a result various functions (which were all internal
4229 use only) have changed to handle the new X509_STORE
4230 structure. This will break anything that messed round
4231 with X509_STORE internally.
4233 The functions X509_STORE_add_cert() now checks for an
4234 exact match, rather than just subject name.
4236 The X509_STORE API doesn't directly support the retrieval
4237 of multiple certificates matching a given criteria, however
4238 this can be worked round by performing a lookup first
4239 (which will fill the cache with candidate certificates)
4240 and then examining the cache for matches. This is probably
4241 the best we can do without throwing out X509_LOOKUP
4242 entirely (maybe later...).
4244 The X509_VERIFY_CTX structure has been enhanced considerably.
4246 All certificate lookup operations now go via a get_issuer()
4247 callback. Although this currently uses an X509_STORE it
4248 can be replaced by custom lookups. This is a simple way
4249 to bypass the X509_STORE hackery necessary to make this
4250 work and makes it possible to use more efficient techniques
4251 in future. A very simple version which uses a simple
4252 STACK for its trusted certificate store is also provided
4253 using X509_STORE_CTX_trusted_stack().
4255 The verify_cb() and verify() callbacks now have equivalents
4256 in the X509_STORE_CTX structure.
4258 X509_STORE_CTX also has a 'flags' field which can be used
4259 to customise the verify behaviour.
4262 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4263 excludes S/MIME capabilities.
4266 *) When a certificate request is read in keep a copy of the
4267 original encoding of the signed data and use it when outputing
4268 again. Signatures then use the original encoding rather than
4269 a decoded, encoded version which may cause problems if the
4270 request is improperly encoded.
4273 *) For consistency with other BIO_puts implementations, call
4274 buffer_write(b, ...) directly in buffer_puts instead of calling
4277 In BIO_puts, increment b->num_write as in BIO_write.
4278 [Peter.Sylvester@EdelWeb.fr]
4280 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4281 BN_zero, we may not return a BIGNUM with an array consisting of
4285 *) Avoid calling abort() from within the library when problems are
4286 detected, except if preprocessor symbols have been defined
4287 (such as REF_CHECK, BN_DEBUG etc.).
4290 *) New openssl application 'rsautl'. This utility can be
4291 used for low level RSA operations. DER public key
4292 BIO/fp routines also added.
4295 *) New Configure entry and patches for compiling on QNX 4.
4296 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4298 *) A demo state-machine implementation was sponsored by
4299 Nuron (http://www.nuron.com/) and is now available in
4300 demos/state_machine.
4303 *) New options added to the 'dgst' utility for signature
4304 generation and verification.
4307 *) Unrecognized PKCS#7 content types are now handled via a
4308 catch all ASN1_TYPE structure. This allows unsupported
4309 types to be stored as a "blob" and an application can
4310 encode and decode it manually.
4313 *) Fix various signed/unsigned issues to make a_strex.c
4315 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4317 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4318 length if passed a buffer. ASN1_INTEGER_to_BN failed
4319 if passed a NULL BN and its argument was negative.
4320 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4322 *) Modification to PKCS#7 encoding routines to output definite
4323 length encoding. Since currently the whole structures are in
4324 memory there's not real point in using indefinite length
4325 constructed encoding. However if OpenSSL is compiled with
4326 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4329 *) Added BIO_vprintf() and BIO_vsnprintf().
4332 *) Added more prefixes to parse for in the the strings written
4333 through a logging bio, to cover all the levels that are available
4334 through syslog. The prefixes are now:
4336 PANIC, EMERG, EMR => LOG_EMERG
4337 ALERT, ALR => LOG_ALERT
4338 CRIT, CRI => LOG_CRIT
4339 ERROR, ERR => LOG_ERR
4340 WARNING, WARN, WAR => LOG_WARNING
4341 NOTICE, NOTE, NOT => LOG_NOTICE
4342 INFO, INF => LOG_INFO
4343 DEBUG, DBG => LOG_DEBUG
4345 and as before, if none of those prefixes are present at the
4346 beginning of the string, LOG_ERR is chosen.
4348 On Win32, the LOG_* levels are mapped according to this:
4350 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4351 LOG_WARNING => EVENTLOG_WARNING_TYPE
4352 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4356 *) Made it possible to reconfigure with just the configuration
4357 argument "reconf" or "reconfigure". The command line arguments
4358 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4359 and are retrieved from there when reconfiguring.
4363 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4365 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4368 *) The obj_dat.pl script was messing up the sorting of object
4369 names. The reason was that it compared the quoted version
4370 of strings as a result "OCSP" > "OCSP Signing" because
4371 " > SPACE. Changed script to store unquoted versions of
4372 names and add quotes on output. It was also omitting some
4373 names from the lookup table if they were given a default
4374 value (that is if SN is missing it is given the same
4375 value as LN and vice versa), these are now added on the
4376 grounds that if an object has a name we should be able to
4377 look it up. Finally added warning output when duplicate
4378 short or long names are found.
4381 *) Changes needed for Tandem NSK.
4382 [Scott Uroff <scott@xypro.com>]
4384 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4385 RSA_padding_check_SSLv23(), special padding was never detected
4386 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4387 version rollback attacks was not effective.
4389 In s23_clnt.c, don't use special rollback-attack detection padding
4390 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4391 client; similarly, in s23_srvr.c, don't do the rollback check if
4392 SSL 2.0 is the only protocol enabled in the server.
4395 *) Make it possible to get hexdumps of unprintable data with 'openssl
4396 asn1parse'. By implication, the functions ASN1_parse_dump() and
4397 BIO_dump_indent() are added.
4400 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4401 these print out strings and name structures based on various
4402 flags including RFC2253 support and proper handling of
4403 multibyte characters. Added options to the 'x509' utility
4404 to allow the various flags to be set.
4407 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4408 Also change the functions X509_cmp_current_time() and
4409 X509_gmtime_adj() work with an ASN1_TIME structure,
4410 this will enable certificates using GeneralizedTime in validity
4411 dates to be checked.
4414 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4415 negative public key encodings) on by default,
4416 NO_NEG_PUBKEY_BUG can be set to disable it.
4419 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4420 content octets. An i2c_ASN1_OBJECT is unnecessary because
4421 the encoding can be trivially obtained from the structure.
4424 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4425 not read locks (CRYPTO_r_[un]lock).
4428 *) A first attempt at creating official support for shared
4429 libraries through configuration. I've kept it so the
4430 default is static libraries only, and the OpenSSL programs
4431 are always statically linked for now, but there are
4432 preparations for dynamic linking in place.
4433 This has been tested on Linux and Tru64.
4436 *) Randomness polling function for Win9x, as described in:
4437 Peter Gutmann, Software Generation of Practically Strong
4441 *) Fix so PRNG is seeded in req if using an already existing
4445 *) New options to smime application. -inform and -outform
4446 allow alternative formats for the S/MIME message including
4447 PEM and DER. The -content option allows the content to be
4448 specified separately. This should allow things like Netscape
4449 form signing output easier to verify.
4452 *) Fix the ASN1 encoding of tags using the 'long form'.
4455 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4456 STRING types. These convert content octets to and from the
4457 underlying type. The actual tag and length octets are
4458 already assumed to have been read in and checked. These
4459 are needed because all other string types have virtually
4460 identical handling apart from the tag. By having versions
4461 of the ASN1 functions that just operate on content octets
4462 IMPLICIT tagging can be handled properly. It also allows
4463 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4464 and ASN1_INTEGER are identical apart from the tag.
4467 *) Change the handling of OID objects as follows:
4469 - New object identifiers are inserted in objects.txt, following
4470 the syntax given in objects.README.
4471 - objects.pl is used to process obj_mac.num and create a new
4473 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4476 This is currently kind of a hack, and the perl code in objects.pl
4477 isn't very elegant, but it works as I intended. The simplest way
4478 to check that it worked correctly is to look in obj_dat.h and
4479 check the array nid_objs and make sure the objects haven't moved
4480 around (this is important!). Additions are OK, as well as
4481 consistent name changes.
4484 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4487 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4488 The given file adds to whatever has already been seeded into the
4489 random pool through the RANDFILE configuration file option or
4490 environment variable, or the default random state file.
4493 *) mkstack.pl now sorts each macro group into lexical order.
4494 Previously the output order depended on the order the files
4495 appeared in the directory, resulting in needless rewriting
4499 *) Patches to make OpenSSL compile under Win32 again. Mostly
4500 work arounds for the VC++ problem that it treats func() as
4501 func(void). Also stripped out the parts of mkdef.pl that
4502 added extra typesafe functions: these no longer exist.
4505 *) Reorganisation of the stack code. The macros are now all
4506 collected in safestack.h . Each macro is defined in terms of
4507 a "stack macro" of the form SKM_<name>(type, a, b). The
4508 DEBUG_SAFESTACK is now handled in terms of function casts,
4509 this has the advantage of retaining type safety without the
4510 use of additional functions. If DEBUG_SAFESTACK is not defined
4511 then the non typesafe macros are used instead. Also modified the
4512 mkstack.pl script to handle the new form. Needs testing to see
4513 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4514 the default if no major problems. Similar behaviour for ASN1_SET_OF
4515 and PKCS12_STACK_OF.
4518 *) When some versions of IIS use the 'NET' form of private key the
4519 key derivation algorithm is different. Normally MD5(password) is
4520 used as a 128 bit RC4 key. In the modified case
4521 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4522 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4523 as the old Netscape_RSA functions except they have an additional
4524 'sgckey' parameter which uses the modified algorithm. Also added
4525 an -sgckey command line option to the rsa utility. Thanks to
4526 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4527 algorithm to openssl-dev.
4530 *) The evp_local.h macros were using 'c.##kname' which resulted in
4531 invalid expansion on some systems (SCO 5.0.5 for example).
4532 Corrected to 'c.kname'.
4533 [Phillip Porch <root@theporch.com>]
4535 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4536 a STACK of email addresses from a certificate or request, these look
4537 in the subject name and the subject alternative name extensions and
4538 omit any duplicate addresses.
4541 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4542 This makes DSA verification about 2 % faster.
4545 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4546 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4547 plus overhead for 1024 bit moduli).
4548 This makes exponentiations about 0.5 % faster for 1024 bit
4549 exponents (as measured by "openssl speed rsa2048").
4552 *) Rename memory handling macros to avoid conflicts with other
4554 Malloc => OPENSSL_malloc
4555 Malloc_locked => OPENSSL_malloc_locked
4556 Realloc => OPENSSL_realloc
4557 Free => OPENSSL_free
4560 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4561 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4564 *) CygWin32 support.
4565 [John Jarvie <jjarvie@newsguy.com>]
4567 *) The type-safe stack code has been rejigged. It is now only compiled
4568 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4569 by default all type-specific stack functions are "#define"d back to
4570 standard stack functions. This results in more streamlined output
4571 but retains the type-safety checking possibilities of the original
4575 *) The STACK code has been cleaned up, and certain type declarations
4576 that didn't make a lot of sense have been brought in line. This has
4577 also involved a cleanup of sorts in safestack.h to more correctly
4578 map type-safe stack functions onto their plain stack counterparts.
4579 This work has also resulted in a variety of "const"ifications of
4580 lots of the code, especially "_cmp" operations which should normally
4581 be prototyped with "const" parameters anyway.
4584 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4585 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4586 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4587 where all of 'md' is used each time the PRNG is used, but 'state'
4588 is used only indexed by a cyclic counter. As entropy may not be
4589 well distributed from the beginning, 'md' is important as a
4590 chaining variable. However, the output function chains only half
4591 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4592 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4593 in all of 'state' being rewritten, with the new values depending
4594 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4597 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4598 the handshake is continued after ssl_verify_cert_chain();
4599 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4600 can lead to 'unexplainable' connection aborts later.
4601 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4603 *) Major EVP API cipher revision.
4604 Add hooks for extra EVP features. This allows various cipher
4605 parameters to be set in the EVP interface. Support added for variable
4606 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4607 setting of RC2 and RC5 parameters.
4609 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4612 Remove lots of duplicated code from the EVP library. For example *every*
4613 cipher init() function handles the 'iv' in the same way according to the
4614 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4615 for CFB and OFB modes they zero ctx->num.
4617 New functionality allows removal of S/MIME code RC2 hack.
4619 Most of the routines have the same form and so can be declared in terms
4622 By shifting this to the top level EVP_CipherInit() it can be removed from
4623 all individual ciphers. If the cipher wants to handle IVs or keys
4624 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4627 Change lots of functions like EVP_EncryptUpdate() to now return a
4628 value: although software versions of the algorithms cannot fail
4629 any installed hardware versions can.
4632 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4633 this option is set, tolerate broken clients that send the negotiated
4634 protocol version number instead of the requested protocol version
4638 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4639 i.e. non-zero for export ciphersuites, zero otherwise.
4640 Previous versions had this flag inverted, inconsistent with
4641 rsa_tmp_cb (..._TMP_RSA_CB).
4642 [Bodo Moeller; problem reported by Amit Chopra]
4644 *) Add missing DSA library text string. Work around for some IIS
4645 key files with invalid SEQUENCE encoding.
4648 *) Add a document (doc/standards.txt) that list all kinds of standards
4649 and so on that are implemented in OpenSSL.
4652 *) Enhance c_rehash script. Old version would mishandle certificates
4653 with the same subject name hash and wouldn't handle CRLs at all.
4654 Added -fingerprint option to crl utility, to support new c_rehash
4658 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4661 *) Fix for SSL server purpose checking. Server checking was
4662 rejecting certificates which had extended key usage present
4663 but no ssl client purpose.
4664 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4666 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4667 is a little unclear about how a blank password is handled.
4668 Since the password in encoded as a BMPString with terminating
4669 double NULL a zero length password would end up as just the
4670 double NULL. However no password at all is different and is
4671 handled differently in the PKCS#12 key generation code. NS
4672 treats a blank password as zero length. MSIE treats it as no
4673 password on export: but it will try both on import. We now do
4674 the same: PKCS12_parse() tries zero length and no password if
4675 the password is set to "" or NULL (NULL is now a valid password:
4676 it wasn't before) as does the pkcs12 application.
4679 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4680 perror when PEM_read_bio_X509_REQ fails, the error message must
4681 be obtained from the error queue.
4684 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4685 it in ERR_remove_state if appropriate, and change ERR_get_state
4686 accordingly to avoid race conditions (this is necessary because
4687 thread_hash is no longer constant once set).
4690 *) Bugfix for linux-elf makefile.one.
4693 *) RSA_get_default_method() will now cause a default
4694 RSA_METHOD to be chosen if one doesn't exist already.
4695 Previously this was only set during a call to RSA_new()
4696 or RSA_new_method(NULL) meaning it was possible for
4697 RSA_get_default_method() to return NULL.
4700 *) Added native name translation to the existing DSO code
4701 that will convert (if the flag to do so is set) filenames
4702 that are sufficiently small and have no path information
4703 into a canonical native form. Eg. "blah" converted to
4704 "libblah.so" or "blah.dll" etc.
4707 *) New function ERR_error_string_n(e, buf, len) which is like
4708 ERR_error_string(e, buf), but writes at most 'len' bytes
4709 including the 0 terminator. For ERR_error_string_n, 'buf'
4711 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4713 *) CONF library reworked to become more general. A new CONF
4714 configuration file reader "class" is implemented as well as a
4715 new functions (NCONF_*, for "New CONF") to handle it. The now
4716 old CONF_* functions are still there, but are reimplemented to
4717 work in terms of the new functions. Also, a set of functions
4718 to handle the internal storage of the configuration data is
4719 provided to make it easier to write new configuration file
4720 reader "classes" (I can definitely see something reading a
4721 configuration file in XML format, for example), called _CONF_*,
4722 or "the configuration storage API"...
4724 The new configuration file reading functions are:
4726 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4727 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4729 NCONF_default, NCONF_WIN32
4731 NCONF_dump_fp, NCONF_dump_bio
4733 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4734 NCONF_new creates a new CONF object. This works in the same way
4735 as other interfaces in OpenSSL, like the BIO interface.
4736 NCONF_dump_* dump the internal storage of the configuration file,
4737 which is useful for debugging. All other functions take the same
4738 arguments as the old CONF_* functions wth the exception of the
4739 first that must be a `CONF *' instead of a `LHASH *'.
4741 To make it easer to use the new classes with the old CONF_* functions,
4742 the function CONF_set_default_method is provided.
4745 *) Add '-tls1' option to 'openssl ciphers', which was already
4746 mentioned in the documentation but had not been implemented.
4747 (This option is not yet really useful because even the additional
4748 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4751 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4752 OpenSSL-based applications) load shared libraries and bind to
4753 them in a portable way.
4754 [Geoff Thorpe, with contributions from Richard Levitte]
4756 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4758 *) Make sure _lrotl and _lrotr are only used with MSVC.
4760 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4761 (the default implementation of RAND_status).
4763 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4764 to '-clrext' (= clear extensions), as intended and documented.
4765 [Bodo Moeller; inconsistency pointed out by Michael Attili
4768 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4769 was larger than the MD block size.
4770 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4772 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4773 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4774 using the passed key: if the passed key was a private key the result
4775 of X509_print(), for example, would be to print out all the private key
4779 *) des_quad_cksum() byte order bug fix.
4780 [Ulf Möller, using the problem description in krb4-0.9.7, where
4781 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4783 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4785 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4787 *) For easily testing in shell scripts whether some command
4788 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4789 returns with exit code 0 iff no command of the given name is available.
4790 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4791 the output goes to stdout and nothing is printed to stderr.
4792 Additional arguments are always ignored.
4794 Since for each cipher there is a command of the same name,
4795 the 'no-cipher' compilation switches can be tested this way.
4797 ('openssl no-XXX' is not able to detect pseudo-commands such
4798 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4801 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4804 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4805 is set; it will be thrown away anyway because each handshake creates
4807 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4808 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4809 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4810 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4813 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4814 'Q' and 'R' lose their special meanings (quit/renegotiate).
4815 This is part of what -quiet does; unlike -quiet, -ign_eof
4816 does not suppress any output.
4819 *) Add compatibility options to the purpose and trust code. The
4820 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4821 accepts a certificate or CA, this was the previous behaviour,
4822 with all the associated security issues.
4824 X509_TRUST_COMPAT is the old trust behaviour: only and
4825 automatically trust self signed roots in certificate store. A
4826 new trust setting X509_TRUST_DEFAULT is used to specify that
4827 a purpose has no associated trust setting and it should instead
4828 use the value in the default purpose.
4831 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4832 and fix a memory leak.
4835 *) In util/mkerr.pl (which implements 'make errors'), preserve
4836 reason strings from the previous version of the .c file, as
4837 the default to have only downcase letters (and digits) in
4838 automatically generated reasons codes is not always appropriate.
4841 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4842 using strerror. Previously, ERR_reason_error_string() returned
4843 library names as reason strings for SYSerr; but SYSerr is a special
4844 case where small numbers are errno values, not library numbers.
4847 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4848 converts DSA parameters into DH parameters. (When creating parameters,
4849 DSA_generate_parameters is used.)
4852 *) Include 'length' (recommended exponent length) in C code generated
4853 by 'openssl dhparam -C'.
4856 *) The second argument to set_label in perlasm was already being used
4857 so couldn't be used as a "file scope" flag. Moved to third argument
4861 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4862 instead of RAND_bytes for encryption IVs and salts.
4865 *) Include RAND_status() into RAND_METHOD instead of implementing
4866 it only for md_rand.c Otherwise replacing the PRNG by calling
4867 RAND_set_rand_method would be impossible.
4870 *) Don't let DSA_generate_key() enter an infinite loop if the random
4871 number generation fails.
4874 *) New 'rand' application for creating pseudo-random output.
4877 *) Added configuration support for Linux/IA64
4878 [Rolf Haberrecker <rolf@suse.de>]
4880 *) Assembler module support for Mingw32.
4883 *) Shared library support for HPUX (in shlib/).
4884 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4886 *) Shared library support for Solaris gcc.
4887 [Lutz Behnke <behnke@trustcenter.de>]
4889 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4891 *) PKCS7_encrypt() was adding text MIME headers twice because they
4892 were added manually and by SMIME_crlf_copy().
4895 *) In bntest.c don't call BN_rand with zero bits argument.
4896 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4898 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4899 case was implemented. This caused BN_div_recp() to fail occasionally.
4902 *) Add an optional second argument to the set_label() in the perl
4903 assembly language builder. If this argument exists and is set
4904 to 1 it signals that the assembler should use a symbol whose
4905 scope is the entire file, not just the current function. This
4906 is needed with MASM which uses the format label:: for this scope.
4907 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4909 *) Change the ASN1 types so they are typedefs by default. Before
4910 almost all types were #define'd to ASN1_STRING which was causing
4911 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4915 *) Change names of new functions to the new get1/get0 naming
4916 convention: After 'get1', the caller owns a reference count
4917 and has to call ..._free; 'get0' returns a pointer to some
4918 data structure without incrementing reference counters.
4919 (Some of the existing 'get' functions increment a reference
4920 counter, some don't.)
4921 Similarly, 'set1' and 'add1' functions increase reference
4922 counters or duplicate objects.
4925 *) Allow for the possibility of temp RSA key generation failure:
4926 the code used to assume it always worked and crashed on failure.
4929 *) Fix potential buffer overrun problem in BIO_printf().
4930 [Ulf Möller, using public domain code by Patrick Powell; problem
4931 pointed out by David Sacerdote <das33@cornell.edu>]
4933 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4934 RAND_egd() and RAND_status(). In the command line application,
4935 the EGD socket can be specified like a seed file using RANDFILE
4939 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4940 Some CAs (e.g. Verisign) distribute certificates in this form.
4943 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4944 list to exclude them. This means that no special compilation option
4945 is needed to use anonymous DH: it just needs to be included in the
4949 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4950 EVP_MD_type. The old functionality is available in a new macro called
4951 EVP_MD_md(). Change code that uses it and update docs.
4954 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4955 where the 'void *' argument is replaced by a function pointer argument.
4956 Previously 'void *' was abused to point to functions, which works on
4957 many platforms, but is not correct. As these functions are usually
4958 called by macros defined in OpenSSL header files, most source code
4959 should work without changes.
4962 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4963 sections with information on -D... compiler switches used for
4964 compiling the library so that applications can see them. To enable
4965 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4966 must be defined. E.g.,
4967 #define OPENSSL_ALGORITHM_DEFINES
4968 #include <openssl/opensslconf.h>
4969 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4970 [Richard Levitte, Ulf and Bodo Möller]
4972 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4976 *) Change the 'other' type in certificate aux info to a STACK_OF
4977 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4978 the required ASN1 format: arbitrary types determined by an OID.
4981 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4982 argument to 'req'. This is not because the function is newer or
4983 better than others it just uses the work 'NEW' in the certificate
4984 request header lines. Some software needs this.
4987 *) Reorganise password command line arguments: now passwords can be
4988 obtained from various sources. Delete the PEM_cb function and make
4989 it the default behaviour: i.e. if the callback is NULL and the
4990 usrdata argument is not NULL interpret it as a null terminated pass
4991 phrase. If usrdata and the callback are NULL then the pass phrase
4992 is prompted for as usual.
4995 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4996 the support is automatically enabled. The resulting binaries will
4997 autodetect the card and use it if present.
4998 [Ben Laurie and Compaq Inc.]
5000 *) Work around for Netscape hang bug. This sends certificate request
5001 and server done in one record. Since this is perfectly legal in the
5002 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5003 the bugs/SSLv3 entry for more info.
5006 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5009 *) Add -rand argument to smime and pkcs12 applications and read/write
5013 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5016 *) Add command line password options to the remaining applications.
5019 *) Bug fix for BN_div_recp() for numerators with an even number of
5023 *) More tests in bntest.c, and changed test_bn output.
5026 *) ./config recognizes MacOS X now.
5029 *) Bug fix for BN_div() when the first words of num and divsor are
5030 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5033 *) Add support for various broken PKCS#8 formats, and command line
5034 options to produce them.
5037 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5038 get temporary BIGNUMs from a BN_CTX.
5041 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5045 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5046 include a #define from the old name to the new. The original intent
5047 was that statically linked binaries could for example just call
5048 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5049 link with digests. This never worked becayse SSLeay_add_all_digests()
5050 and SSLeay_add_all_ciphers() were in the same source file so calling
5051 one would link with the other. They are now in separate source files.
5054 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5057 *) Use a less unusual form of the Miller-Rabin primality test (it used
5058 a binary algorithm for exponentiation integrated into the Miller-Rabin
5059 loop, our standard modexp algorithms are faster).
5062 *) Support for the EBCDIC character set completed.
5063 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5065 *) Source code cleanups: use const where appropriate, eliminate casts,
5066 use void * instead of char * in lhash.
5069 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5070 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5071 this the server could overwrite ephemeral keys that the client
5075 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5076 using 50 iterations of the Rabin-Miller test.
5078 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5079 iterations of the Rabin-Miller test as required by the appendix
5080 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5081 As BN_is_prime_fasttest includes trial division, DSA parameter
5082 generation becomes much faster.
5084 This implies a change for the callback functions in DSA_is_prime
5085 and DSA_generate_parameters: The callback function is called once
5086 for each positive witness in the Rabin-Miller test, not just
5087 occasionally in the inner loop; and the parameters to the
5088 callback function now provide an iteration count for the outer
5089 loop rather than for the current invocation of the inner loop.
5090 DSA_generate_parameters additionally can call the callback
5091 function with an 'iteration count' of -1, meaning that a
5092 candidate has passed the trial division test (when q is generated
5093 from an application-provided seed, trial division is skipped).
5096 *) New function BN_is_prime_fasttest that optionally does trial
5097 division before starting the Rabin-Miller test and has
5098 an additional BN_CTX * argument (whereas BN_is_prime always
5099 has to allocate at least one BN_CTX).
5100 'callback(1, -1, cb_arg)' is called when a number has passed the
5101 trial division stage.
5104 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5108 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5111 *) New function BN_pseudo_rand().
5114 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5115 bignum version of BN_from_montgomery() with the working code from
5116 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5120 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5121 made it impossible to use the same SSL_SESSION data structure in
5122 SSL2 clients in multiple threads.
5125 *) The return value of RAND_load_file() no longer counts bytes obtained
5126 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5127 to seed the PRNG (previously an explicit byte count was required).
5128 [Ulf Möller, Bodo Möller]
5130 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5131 used (char *) instead of (void *) and had casts all over the place.
5134 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5137 *) Retain source code compatibility for BN_prime_checks macro:
5138 BN_is_prime(..., BN_prime_checks, ...) now uses
5139 BN_prime_checks_for_size to determine the appropriate number of
5140 Rabin-Miller iterations.
5143 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5144 DH_CHECK_P_NOT_SAFE_PRIME.
5145 (Check if this is true? OpenPGP calls them "strong".)
5148 *) Merge the functionality of "dh" and "gendh" programs into a new program
5149 "dhparam". The old programs are retained for now but will handle DH keys
5150 (instead of parameters) in future.
5153 *) Make the ciphers, s_server and s_client programs check the return values
5154 when a new cipher list is set.
5157 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5158 ciphers. Before when the 56bit ciphers were enabled the sorting was
5161 The syntax for the cipher sorting has been extended to support sorting by
5162 cipher-strength (using the strength_bits hard coded in the tables).
5163 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5165 Fix a bug in the cipher-command parser: when supplying a cipher command
5166 string with an "undefined" symbol (neither command nor alphanumeric
5167 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5168 an error is flagged.
5170 Due to the strength-sorting extension, the code of the
5171 ssl_create_cipher_list() function was completely rearranged. I hope that
5172 the readability was also increased :-)
5173 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5175 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5176 for the first serial number and places 2 in the serial number file. This
5177 avoids problems when the root CA is created with serial number zero and
5178 the first user certificate has the same issuer name and serial number
5182 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5183 the new code. Add documentation for this stuff.
5186 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5187 X509_*() to X509at_*() on the grounds that they don't handle X509
5188 structures and behave in an analagous way to the X509v3 functions:
5189 they shouldn't be called directly but wrapper functions should be used
5192 So we also now have some wrapper functions that call the X509at functions
5193 when passed certificate requests. (TO DO: similar things can be done with
5194 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5195 things. Some of these need some d2i or i2d and print functionality
5196 because they handle more complex structures.)
5199 *) Add missing #ifndefs that caused missing symbols when building libssl
5200 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5201 NO_RSA in ssl/s2*.c.
5202 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5204 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5205 has a return value which indicates the quality of the random data
5206 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5207 error queue. New function RAND_pseudo_bytes() generates output that is
5208 guaranteed to be unique but not unpredictable. RAND_add is like
5209 RAND_seed, but takes an extra argument for an entropy estimate
5210 (RAND_seed always assumes full entropy).
5213 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5214 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5215 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5216 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5217 false-positive rate of at most 2^-80 for random input.
5220 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5223 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5224 in the 0.9.5 release), this returns the chain
5225 from an X509_CTX structure with a dup of the stack and all
5226 the X509 reference counts upped: so the stack will exist
5227 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5230 Also make SSL_SESSION_print() print out the verify return
5234 *) Add manpage for the pkcs12 command. Also change the default
5235 behaviour so MAC iteration counts are used unless the new
5236 -nomaciter option is used. This improves file security and
5237 only older versions of MSIE (4.0 for example) need it.
5240 *) Honor the no-xxx Configure options when creating .DEF files.
5243 *) Add PKCS#10 attributes to field table: challengePassword,
5244 unstructuredName and unstructuredAddress. These are taken from
5245 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5246 international characters are used.
5248 More changes to X509_ATTRIBUTE code: allow the setting of types
5249 based on strings. Remove the 'loc' parameter when adding
5250 attributes because these will be a SET OF encoding which is sorted
5254 *) Initial changes to the 'req' utility to allow request generation
5255 automation. This will allow an application to just generate a template
5256 file containing all the field values and have req construct the
5259 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5260 used all over the place including certificate requests and PKCS#7
5261 structures. They are currently handled manually where necessary with
5262 some primitive wrappers for PKCS#7. The new functions behave in a
5263 manner analogous to the X509 extension functions: they allow
5264 attributes to be looked up by NID and added.
5266 Later something similar to the X509V3 code would be desirable to
5267 automatically handle the encoding, decoding and printing of the
5268 more complex types. The string types like challengePassword can
5269 be handled by the string table functions.
5271 Also modified the multi byte string table handling. Now there is
5272 a 'global mask' which masks out certain types. The table itself
5273 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5274 is useful when for example there is only one permissible type
5275 (as in countryName) and using the mask might result in no valid
5279 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5280 SSL_get_peer_finished to allow applications to obtain the latest
5281 Finished messages sent to the peer or expected from the peer,
5282 respectively. (SSL_get_peer_finished is usually the Finished message
5283 actually received from the peer, otherwise the protocol will be aborted.)
5285 As the Finished message are message digests of the complete handshake
5286 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5287 be used for external authentication procedures when the authentication
5288 provided by SSL/TLS is not desired or is not enough.
5291 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5292 the host supports BWX extension and if Compaq C is present on the
5293 $PATH. Just exploiting of the BWX extension results in 20-30%
5294 performance kick for some algorithms, e.g. DES and RC4 to mention
5295 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5299 *) Add support for MS "fast SGC". This is arguably a violation of the
5300 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5301 weak crypto and after checking the certificate is SGC a second one
5302 with strong crypto. MS SGC stops the first handshake after receiving
5303 the server certificate message and sends a second client hello. Since
5304 a server will typically do all the time consuming operations before
5305 expecting any further messages from the client (server key exchange
5306 is the most expensive) there is little difference between the two.
5308 To get OpenSSL to support MS SGC we have to permit a second client
5309 hello message after we have sent server done. In addition we have to
5310 reset the MAC if we do get this second client hello.
5313 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5314 if a DER encoded private key is RSA or DSA traditional format. Changed
5315 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5316 format DER encoded private key. Newer code should use PKCS#8 format which
5317 has the key type encoded in the ASN1 structure. Added DER private key
5318 support to pkcs8 application.
5321 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5322 ciphersuites has been selected (as required by the SSL 3/TLS 1
5323 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5324 is set, we interpret this as a request to violate the specification
5325 (the worst that can happen is a handshake failure, and 'correct'
5326 behaviour would result in a handshake failure anyway).
5329 *) In SSL_CTX_add_session, take into account that there might be multiple
5330 SSL_SESSION structures with the same session ID (e.g. when two threads
5331 concurrently obtain them from an external cache).
5332 The internal cache can handle only one SSL_SESSION with a given ID,
5333 so if there's a conflict, we now throw out the old one to achieve
5337 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5338 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5339 some routines that use cipher OIDs: some ciphers do not have OIDs
5340 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5344 *) Simplify the trust setting structure and code. Now we just have
5345 two sequences of OIDs for trusted and rejected settings. These will
5346 typically have values the same as the extended key usage extension
5347 and any application specific purposes.
5349 The trust checking code now has a default behaviour: it will just
5350 check for an object with the same NID as the passed id. Functions can
5351 be provided to override either the default behaviour or the behaviour
5352 for a given id. SSL client, server and email already have functions
5353 in place for compatibility: they check the NID and also return "trusted"
5354 if the certificate is self signed.
5357 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5358 traditional format into an EVP_PKEY structure.
5361 *) Add a password callback function PEM_cb() which either prompts for
5362 a password if usr_data is NULL or otherwise assumes it is a null
5363 terminated password. Allow passwords to be passed on command line
5364 environment or config files in a few more utilities.
5367 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5368 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5369 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5370 Update documentation.
5373 *) Support for ASN1 "NULL" type. This could be handled before by using
5374 ASN1_TYPE but there wasn't any function that would try to read a NULL
5375 and produce an error if it couldn't. For compatibility we also have
5376 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5377 don't allocate anything because they don't need to.
5380 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5382 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5384 *) Rebuild of the memory allocation routines used by OpenSSL code and
5385 possibly others as well. The purpose is to make an interface that
5386 provide hooks so anyone can build a separate set of allocation and
5387 deallocation routines to be used by OpenSSL, for example memory
5388 pool implementations, or something else, which was previously hard
5389 since Malloc(), Realloc() and Free() were defined as macros having
5390 the values malloc, realloc and free, respectively (except for Win32
5391 compilations). The same is provided for memory debugging code.
5392 OpenSSL already comes with functionality to find memory leaks, but
5393 this gives people a chance to debug other memory problems.
5395 With these changes, a new set of functions and macros have appeared:
5397 CRYPTO_set_mem_debug_functions() [F]
5398 CRYPTO_get_mem_debug_functions() [F]
5399 CRYPTO_dbg_set_options() [F]
5400 CRYPTO_dbg_get_options() [F]
5401 CRYPTO_malloc_debug_init() [M]
5403 The memory debug functions are NULL by default, unless the library
5404 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5405 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5406 gives the standard debugging functions that come with OpenSSL) or
5407 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5408 provided by the library user) must be used. When the standard
5409 debugging functions are used, CRYPTO_dbg_set_options can be used to
5410 request additional information:
5411 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5412 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5414 Also, things like CRYPTO_set_mem_functions will always give the
5415 expected result (the new set of functions is used for allocation
5416 and deallocation) at all times, regardless of platform and compiler
5419 To finish it up, some functions that were never use in any other
5420 way than through macros have a new API and new semantic:
5423 CRYPTO_dbg_realloc()
5426 All macros of value have retained their old syntax.
5427 [Richard Levitte and Bodo Moeller]
5429 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5430 ordering of SMIMECapabilities wasn't in "strength order" and there
5431 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5435 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5436 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5437 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5439 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5440 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5441 functionality to handle multipart/signed properly) and a utility
5442 called 'smime' to call all this stuff. This is based on code I
5443 originally wrote for Celo who have kindly allowed it to be
5444 included in OpenSSL.
5447 *) Add variants des_set_key_checked and des_set_key_unchecked of
5448 des_set_key (aka des_key_sched). Global variable des_check_key
5449 decides which of these is called by des_set_key; this way
5450 des_check_key behaves as it always did, but applications and
5451 the library itself, which was buggy for des_check_key == 1,
5452 have a cleaner way to pick the version they need.
5455 *) New function PKCS12_newpass() which changes the password of a
5459 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5460 dynamic mix. In both cases the ids can be used as an index into the
5461 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5462 functions so they accept a list of the field values and the
5463 application doesn't need to directly manipulate the X509_TRUST
5467 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5471 *) Modify the way the V3 extension code looks up extensions. This now
5472 works in a similar way to the object code: we have some "standard"
5473 extensions in a static table which is searched with OBJ_bsearch()
5474 and the application can add dynamic ones if needed. The file
5475 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5476 updated whenever a new extension is added to the core code and kept
5477 in ext_nid order. There is a simple program 'tabtest.c' which checks
5478 this. New extensions are not added too often so this file can readily
5479 be maintained manually.
5481 There are two big advantages in doing things this way. The extensions
5482 can be looked up immediately and no longer need to be "added" using
5483 X509V3_add_standard_extensions(): this function now does nothing.
5484 [Side note: I get *lots* of email saying the extension code doesn't
5485 work because people forget to call this function]
5486 Also no dynamic allocation is done unless new extensions are added:
5487 so if we don't add custom extensions there is no need to call
5488 X509V3_EXT_cleanup().
5491 *) Modify enc utility's salting as follows: make salting the default. Add a
5492 magic header, so unsalted files fail gracefully instead of just decrypting
5493 to garbage. This is because not salting is a big security hole, so people
5494 should be discouraged from doing it.
5497 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5498 digest to be passed on the command line but it only used this
5499 parameter when signing a certificate. Modified so all relevant
5500 operations are affected by the digest parameter including the
5501 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5502 DSA key was used because it didn't fix the digest.
5505 *) Initial certificate chain verify code. Currently tests the untrusted
5506 certificates for consistency with the verify purpose (which is set
5507 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5509 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5510 this is because it will reject chains with invalid extensions whereas
5511 every previous version of OpenSSL and SSLeay made no checks at all.
5513 Trust code: checks the root CA for the relevant trust settings. Trust
5514 settings have an initial value consistent with the verify purpose: e.g.
5515 if the verify purpose is for SSL client use it expects the CA to be
5516 trusted for SSL client use. However the default value can be changed to
5517 permit custom trust settings: one example of this would be to only trust
5518 certificates from a specific "secure" set of CAs.
5520 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5521 which should be used for version portability: especially since the
5522 verify structure is likely to change more often now.
5524 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5525 to set them. If not set then assume SSL clients will verify SSL servers
5528 Two new options to the verify program: -untrusted allows a set of
5529 untrusted certificates to be passed in and -purpose which sets the
5530 intended purpose of the certificate. If a purpose is set then the
5531 new chain verify code is used to check extension consistency.
5534 *) Support for the authority information access extension.
5537 *) Modify RSA and DSA PEM read routines to transparently handle
5538 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5539 public keys in a format compatible with certificate
5540 SubjectPublicKeyInfo structures. Unfortunately there were already
5541 functions called *_PublicKey_* which used various odd formats so
5542 these are retained for compatibility: however the DSA variants were
5543 never in a public release so they have been deleted. Changed dsa/rsa
5544 utilities to handle the new format: note no releases ever handled public
5545 keys so we should be OK.
5547 The primary motivation for this change is to avoid the same fiasco
5548 that dogs private keys: there are several incompatible private key
5549 formats some of which are standard and some OpenSSL specific and
5550 require various evil hacks to allow partial transparent handling and
5551 even then it doesn't work with DER formats. Given the option anything
5552 other than PKCS#8 should be dumped: but the other formats have to
5553 stay in the name of compatibility.
5555 With public keys and the benefit of hindsight one standard format
5556 is used which works with EVP_PKEY, RSA or DSA structures: though
5557 it clearly returns an error if you try to read the wrong kind of key.
5559 Added a -pubkey option to the 'x509' utility to output the public key.
5560 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5561 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5562 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5563 that do the same as the EVP_PKEY_assign_*() except they up the
5564 reference count of the added key (they don't "swallow" the
5568 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5569 CRLs would fail if the file contained no certificates or no CRLs:
5570 added a new function to read in both types and return the number
5571 read: this means that if none are read it will be an error. The
5572 DER versions of the certificate and CRL reader would always fail
5573 because it isn't possible to mix certificates and CRLs in DER format
5574 without choking one or the other routine. Changed this to just read
5575 a certificate: this is the best we can do. Also modified the code
5576 in apps/verify.c to take notice of return codes: it was previously
5577 attempting to read in certificates from NULL pointers and ignoring
5578 any errors: this is one reason why the cert and CRL reader seemed
5579 to work. It doesn't check return codes from the default certificate
5580 routines: these may well fail if the certificates aren't installed.
5583 *) Code to support otherName option in GeneralName.
5586 *) First update to verify code. Change the verify utility
5587 so it warns if it is passed a self signed certificate:
5588 for consistency with the normal behaviour. X509_verify
5589 has been modified to it will now verify a self signed
5590 certificate if *exactly* the same certificate appears
5591 in the store: it was previously impossible to trust a
5592 single self signed certificate. This means that:
5593 openssl verify ss.pem
5594 now gives a warning about a self signed certificate but
5595 openssl verify -CAfile ss.pem ss.pem
5599 *) For servers, store verify_result in SSL_SESSION data structure
5600 (and add it to external session representation).
5601 This is needed when client certificate verifications fails,
5602 but an application-provided verification callback (set by
5603 SSL_CTX_set_cert_verify_callback) allows accepting the session
5604 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5605 but returns 1): When the session is reused, we have to set
5606 ssl->verify_result to the appropriate error code to avoid
5608 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5610 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5611 case in PKCS7_dataInit() where the signed PKCS7 structure
5612 didn't contain any existing data because it was being created.
5613 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5615 *) Add a salt to the key derivation routines in enc.c. This
5616 forms the first 8 bytes of the encrypted file. Also add a
5617 -S option to allow a salt to be input on the command line.
5620 *) New function X509_cmp(). Oddly enough there wasn't a function
5621 to compare two certificates. We do this by working out the SHA1
5622 hash and comparing that. X509_cmp() will be needed by the trust
5626 *) SSL_get1_session() is like SSL_get_session(), but increments
5627 the reference count in the SSL_SESSION returned.
5628 [Geoff Thorpe <geoff@eu.c2.net>]
5630 *) Fix for 'req': it was adding a null to request attributes.
5631 Also change the X509_LOOKUP and X509_INFO code to handle
5632 certificate auxiliary information.
5635 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5639 *) Add the possibility to add extra information to the memory leak
5640 detecting output, to form tracebacks, showing from where each
5641 allocation was originated: CRYPTO_push_info("constant string") adds
5642 the string plus current file name and line number to a per-thread
5643 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5644 is like calling CYRPTO_pop_info() until the stack is empty.
5645 Also updated memory leak detection code to be multi-thread-safe.
5648 *) Add options -text and -noout to pkcs7 utility and delete the
5649 encryption options which never did anything. Update docs.
5652 *) Add options to some of the utilities to allow the pass phrase
5653 to be included on either the command line (not recommended on
5654 OSes like Unix) or read from the environment. Update the
5655 manpages and fix a few bugs.
5658 *) Add a few manpages for some of the openssl commands.
5661 *) Fix the -revoke option in ca. It was freeing up memory twice,
5662 leaking and not finding already revoked certificates.
5665 *) Extensive changes to support certificate auxiliary information.
5666 This involves the use of X509_CERT_AUX structure and X509_AUX
5667 functions. An X509_AUX function such as PEM_read_X509_AUX()
5668 can still read in a certificate file in the usual way but it
5669 will also read in any additional "auxiliary information". By
5670 doing things this way a fair degree of compatibility can be
5671 retained: existing certificates can have this information added
5672 using the new 'x509' options.
5674 Current auxiliary information includes an "alias" and some trust
5675 settings. The trust settings will ultimately be used in enhanced
5676 certificate chain verification routines: currently a certificate
5677 can only be trusted if it is self signed and then it is trusted
5681 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5682 The problem was that one of the replacement routines had not been working
5683 since SSLeay releases. For now the offending routine has been replaced
5684 with non-optimised assembler. Even so, this now gives around 95%
5685 performance improvement for 1024 bit RSA signs.
5688 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5689 handling. Most clients have the effective key size in bits equal to
5690 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5691 A few however don't do this and instead use the size of the decrypted key
5692 to determine the RC2 key length and the AlgorithmIdentifier to determine
5693 the effective key length. In this case the effective key length can still
5694 be 40 bits but the key length can be 168 bits for example. This is fixed
5695 by manually forcing an RC2 key into the EVP_PKEY structure because the
5696 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5697 the key length and effective key length are equal.
5700 *) Add a bunch of functions that should simplify the creation of
5701 X509_NAME structures. Now you should be able to do:
5702 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5703 and have it automatically work out the correct field type and fill in
5704 the structures. The more adventurous can try:
5705 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5706 and it will (hopefully) work out the correct multibyte encoding.
5709 *) Change the 'req' utility to use the new field handling and multibyte
5710 copy routines. Before the DN field creation was handled in an ad hoc
5711 way in req, ca, and x509 which was rather broken and didn't support
5712 BMPStrings or UTF8Strings. Since some software doesn't implement
5713 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5714 using the dirstring_type option. See the new comment in the default
5715 openssl.cnf for more info.
5718 *) Make crypto/rand/md_rand.c more robust:
5719 - Assure unique random numbers after fork().
5720 - Make sure that concurrent threads access the global counter and
5721 md serializably so that we never lose entropy in them
5722 or use exactly the same state in multiple threads.
5723 Access to the large state is not always serializable because
5724 the additional locking could be a performance killer, and
5725 md should be large enough anyway.
5728 *) New file apps/app_rand.c with commonly needed functionality
5729 for handling the random seed file.
5731 Use the random seed file in some applications that previously did not:
5733 dsaparam -genkey (which also ignored its '-rand' option),
5736 x509 (when signing).
5737 Except on systems with /dev/urandom, it is crucial to have a random
5738 seed file at least for key creation, DSA signing, and for DH exchanges;
5739 for RSA signatures we could do without one.
5741 gendh and gendsa (unlike genrsa) used to read only the first byte
5742 of each file listed in the '-rand' option. The function as previously
5743 found in genrsa is now in app_rand.c and is used by all programs
5744 that support '-rand'.
5747 *) In RAND_write_file, use mode 0600 for creating files;
5748 don't just chmod when it may be too late.
5751 *) Report an error from X509_STORE_load_locations
5752 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5755 *) New function ASN1_mbstring_copy() this copies a string in either
5756 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5757 into an ASN1_STRING type. A mask of permissible types is passed
5758 and it chooses the "minimal" type to use or an error if not type
5762 *) Add function equivalents to the various macros in asn1.h. The old
5763 macros are retained with an M_ prefix. Code inside the library can
5764 use the M_ macros. External code (including the openssl utility)
5765 should *NOT* in order to be "shared library friendly".
5768 *) Add various functions that can check a certificate's extensions
5769 to see if it usable for various purposes such as SSL client,
5770 server or S/MIME and CAs of these types. This is currently
5771 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5772 verification. Also added a -purpose flag to x509 utility to
5773 print out all the purposes.
5776 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5780 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5781 for, obtain and decode and extension and obtain its critical flag.
5782 This allows all the necessary extension code to be handled in a
5783 single function call.
5786 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5787 platforms. See crypto/rc4/rc4_enc.c for further details.
5790 *) New -noout option to asn1parse. This causes no output to be produced
5791 its main use is when combined with -strparse and -out to extract data
5792 from a file (which may not be in ASN.1 format).
5795 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5796 when producing the local key id.
5797 [Richard Levitte <levitte@stacken.kth.se>]
5799 *) New option -dhparam in s_server. This allows a DH parameter file to be
5800 stated explicitly. If it is not stated then it tries the first server
5801 certificate file. The previous behaviour hard coded the filename
5805 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5806 a public key to be input or output. For example:
5807 openssl rsa -in key.pem -pubout -out pubkey.pem
5808 Also added necessary DSA public key functions to handle this.
5811 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5812 in the message. This was handled by allowing
5813 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5814 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5816 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5817 to the end of the strings whereas this didn't. This would cause problems
5818 if strings read with d2i_ASN1_bytes() were later modified.
5819 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5821 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5822 data and it contains EOF it will end up returning an error. This is
5823 caused by input 46 bytes long. The cause is due to the way base64
5824 BIOs find the start of base64 encoded data. They do this by trying a
5825 trial decode on each line until they find one that works. When they
5826 do a flag is set and it starts again knowing it can pass all the
5827 data directly through the decoder. Unfortunately it doesn't reset
5828 the context it uses. This means that if EOF is reached an attempt
5829 is made to pass two EOFs through the context and this causes the
5830 resulting error. This can also cause other problems as well. As is
5831 usual with these problems it takes *ages* to find and the fix is
5832 trivial: move one line.
5833 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5835 *) Ugly workaround to get s_client and s_server working under Windows. The
5836 old code wouldn't work because it needed to select() on sockets and the
5837 tty (for keypresses and to see if data could be written). Win32 only
5838 supports select() on sockets so we select() with a 1s timeout on the
5839 sockets and then see if any characters are waiting to be read, if none
5840 are present then we retry, we also assume we can always write data to
5841 the tty. This isn't nice because the code then blocks until we've
5842 received a complete line of data and it is effectively polling the
5843 keyboard at 1s intervals: however it's quite a bit better than not
5844 working at all :-) A dedicated Windows application might handle this
5845 with an event loop for example.
5848 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5849 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5850 will be called when RSA_sign() and RSA_verify() are used. This is useful
5851 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5852 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5853 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5854 This necessitated the support of an extra signature type NID_md5_sha1
5855 for SSL signatures and modifications to the SSL library to use it instead
5856 of calling RSA_public_decrypt() and RSA_private_encrypt().
5859 *) Add new -verify -CAfile and -CApath options to the crl program, these
5860 will lookup a CRL issuers certificate and verify the signature in a
5861 similar way to the verify program. Tidy up the crl program so it
5862 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5863 less strict. It will now permit CRL extensions even if it is not
5864 a V2 CRL: this will allow it to tolerate some broken CRLs.
5867 *) Initialize all non-automatic variables each time one of the openssl
5868 sub-programs is started (this is necessary as they may be started
5869 multiple times from the "OpenSSL>" prompt).
5870 [Lennart Bang, Bodo Moeller]
5872 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5873 removing all other RSA functionality (this is what NO_RSA does). This
5874 is so (for example) those in the US can disable those operations covered
5875 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5879 *) Non-copying interface to BIO pairs.
5880 (still largely untested)
5883 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5884 ASCII string. This was handled independently in various places before.
5887 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5888 UTF8 strings a character at a time.
5891 *) Use client_version from client hello to select the protocol
5892 (s23_srvr.c) and for RSA client key exchange verification
5893 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5896 *) Add various utility functions to handle SPKACs, these were previously
5897 handled by poking round in the structure internals. Added new function
5898 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5899 print, verify and generate SPKACs. Based on an original idea from
5900 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5903 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5906 *) Allow the config file extension section to be overwritten on the
5907 command line. Based on an original idea from Massimiliano Pala
5908 <madwolf@comune.modena.it>. The new option is called -extensions
5909 and can be applied to ca, req and x509. Also -reqexts to override
5910 the request extensions in req and -crlexts to override the crl extensions
5914 *) Add new feature to the SPKAC handling in ca. Now you can include
5915 the same field multiple times by preceding it by "XXXX." for example:
5918 this is the same syntax as used in the req config file.
5921 *) Allow certificate extensions to be added to certificate requests. These
5922 are specified in a 'req_extensions' option of the req section of the
5923 config file. They can be printed out with the -text option to req but
5924 are otherwise ignored at present.
5927 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5928 data read consists of only the final block it would not decrypted because
5929 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5930 A misplaced 'break' also meant the decrypted final block might not be
5931 copied until the next read.
5934 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5935 a few extra parameters to the DH structure: these will be useful if
5936 for example we want the value of 'q' or implement X9.42 DH.
5939 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5940 provides hooks that allow the default DSA functions or functions on a
5941 "per key" basis to be replaced. This allows hardware acceleration and
5942 hardware key storage to be handled without major modification to the
5943 library. Also added low level modexp hooks and CRYPTO_EX structure and
5944 associated functions.
5947 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5948 as "read only": it can't be written to and the buffer it points to will
5949 not be freed. Reading from a read only BIO is much more efficient than
5950 a normal memory BIO. This was added because there are several times when
5951 an area of memory needs to be read from a BIO. The previous method was
5952 to create a memory BIO and write the data to it, this results in two
5953 copies of the data and an O(n^2) reading algorithm. There is a new
5954 function BIO_new_mem_buf() which creates a read only memory BIO from
5955 an area of memory. Also modified the PKCS#7 routines to use read only
5959 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5960 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5961 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5962 but a retry condition occured while trying to read the rest.
5965 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5966 NID_pkcs7_encrypted by default: this was wrong since this should almost
5967 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5968 the encrypted data type: this is a more sensible place to put it and it
5969 allows the PKCS#12 code to be tidied up that duplicated this
5973 *) Changed obj_dat.pl script so it takes its input and output files on
5974 the command line. This should avoid shell escape redirection problems
5978 *) Initial support for certificate extension requests, these are included
5979 in things like Xenroll certificate requests. Included functions to allow
5980 extensions to be obtained and added.
5983 *) -crlf option to s_client and s_server for sending newlines as
5984 CRLF (as required by many protocols).
5987 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5989 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5990 [Ralf S. Engelschall]
5992 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5993 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5995 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5999 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6000 DH parameters/keys (q is lost during that conversion, but the resulting
6001 DH parameters contain its length).
6003 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6004 much faster than DH_generate_parameters (which creates parameters
6005 where p = 2*q + 1), and also the smaller q makes DH computations
6006 much more efficient (160-bit exponentiation instead of 1024-bit
6007 exponentiation); so this provides a convenient way to support DHE
6008 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6009 utter importance to use
6010 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6012 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6013 when such DH parameters are used, because otherwise small subgroup
6014 attacks may become possible!
6017 *) Avoid memory leak in i2d_DHparams.
6020 *) Allow the -k option to be used more than once in the enc program:
6021 this allows the same encrypted message to be read by multiple recipients.
6024 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6025 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6026 it will always use the numerical form of the OID, even if it has a short
6030 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6031 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6032 otherwise bn_mod_exp was called. In the case of hardware keys for example
6033 no private key components need be present and it might store extra data
6034 in the RSA structure, which cannot be accessed from bn_mod_exp.
6035 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6036 private key operations.
6039 *) Added support for SPARC Linux.
6042 *) pem_password_cb function type incompatibly changed from
6043 typedef int pem_password_cb(char *buf, int size, int rwflag);
6045 ....(char *buf, int size, int rwflag, void *userdata);
6046 so that applications can pass data to their callbacks:
6047 The PEM[_ASN1]_{read,write}... functions and macros now take an
6048 additional void * argument, which is just handed through whenever
6049 the password callback is called.
6050 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6052 New function SSL_CTX_set_default_passwd_cb_userdata.
6054 Compatibility note: As many C implementations push function arguments
6055 onto the stack in reverse order, the new library version is likely to
6056 interoperate with programs that have been compiled with the old
6057 pem_password_cb definition (PEM_whatever takes some data that
6058 happens to be on the stack as its last argument, and the callback
6059 just ignores this garbage); but there is no guarantee whatsoever that
6062 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6063 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6064 problems not only on Windows, but also on some Unix platforms.
6065 To avoid problematic command lines, these definitions are now in an
6066 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6067 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6070 *) MIPS III/IV assembler module is reimplemented.
6073 *) More DES library cleanups: remove references to srand/rand and
6074 delete an unused file.
6077 *) Add support for the the free Netwide assembler (NASM) under Win32,
6078 since not many people have MASM (ml) and it can be hard to obtain.
6079 This is currently experimental but it seems to work OK and pass all
6080 the tests. Check out INSTALL.W32 for info.
6083 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6084 without temporary keys kept an extra copy of the server key,
6085 and connections with temporary keys did not free everything in case
6089 *) New function RSA_check_key and new openssl rsa option -check
6090 for verifying the consistency of RSA keys.
6091 [Ulf Moeller, Bodo Moeller]
6093 *) Various changes to make Win32 compile work:
6094 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6095 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6096 comparison" warnings.
6097 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6100 *) Add a debugging option to PKCS#5 v2 key generation function: when
6101 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6102 derived keys are printed to stderr.
6105 *) Copy the flags in ASN1_STRING_dup().
6106 [Roman E. Pavlov <pre@mo.msk.ru>]
6108 *) The x509 application mishandled signing requests containing DSA
6109 keys when the signing key was also DSA and the parameters didn't match.
6111 It was supposed to omit the parameters when they matched the signing key:
6112 the verifying software was then supposed to automatically use the CA's
6113 parameters if they were absent from the end user certificate.
6115 Omitting parameters is no longer recommended. The test was also
6116 the wrong way round! This was probably due to unusual behaviour in
6117 EVP_cmp_parameters() which returns 1 if the parameters match.
6118 This meant that parameters were omitted when they *didn't* match and
6119 the certificate was useless. Certificates signed with 'ca' didn't have
6121 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6123 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6124 The interface is as follows:
6125 Applications can use
6126 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6127 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6128 "off" is now the default.
6129 The library internally uses
6130 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6131 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6132 to disable memory-checking temporarily.
6134 Some inconsistent states that previously were possible (and were
6135 even the default) are now avoided.
6137 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6138 with each memory chunk allocated; this is occasionally more helpful
6139 than just having a counter.
6141 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6143 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6147 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6148 which largely parallels "options", but is for changing API behaviour,
6149 whereas "options" are about protocol behaviour.
6150 Initial "mode" flags are:
6152 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6153 a single record has been written.
6154 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6155 retries use the same buffer location.
6156 (But all of the contents must be
6160 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6163 *) Fix problems with no-hmac etc.
6164 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6166 *) New functions RSA_get_default_method(), RSA_set_method() and
6167 RSA_get_method(). These allows replacement of RSA_METHODs without having
6168 to mess around with the internals of an RSA structure.
6171 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6172 Also really enable memory leak checks in openssl.c and in some
6174 [Chad C. Mulligan, Bodo Moeller]
6176 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6177 up the length of negative integers. This has now been simplified to just
6178 store the length when it is first determined and use it later, rather
6179 than trying to keep track of where data is copied and updating it to
6181 [Steve Henson, reported by Brien Wheeler
6182 <bwheeler@authentica-security.com>]
6184 *) Add a new function PKCS7_signatureVerify. This allows the verification
6185 of a PKCS#7 signature but with the signing certificate passed to the
6186 function itself. This contrasts with PKCS7_dataVerify which assumes the
6187 certificate is present in the PKCS#7 structure. This isn't always the
6188 case: certificates can be omitted from a PKCS#7 structure and be
6189 distributed by "out of band" means (such as a certificate database).
6192 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6193 function prototypes in pem.h, also change util/mkdef.pl to add the
6194 necessary function names.
6197 *) mk1mf.pl (used by Windows builds) did not properly read the
6198 options set by Configure in the top level Makefile, and Configure
6199 was not even able to write more than one option correctly.
6200 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6203 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6204 file to be loaded from a BIO or FILE pointer. The BIO version will
6205 for example allow memory BIOs to contain config info.
6208 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6209 Whoever hopes to achieve shared-library compatibility across versions
6210 must use this, not the compile-time macro.
6211 (Exercise 0.9.4: Which is the minimum library version required by
6213 Note: All this applies only to multi-threaded programs, others don't
6217 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6218 through a BIO pair triggered the default case, i.e.
6219 SSLerr(...,SSL_R_UNKNOWN_STATE).
6222 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6223 can use the SSL library even if none of the specific BIOs is
6227 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6228 for the encoded length.
6229 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6231 *) Add initial documentation of the X509V3 functions.
6234 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6235 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6236 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6237 secure PKCS#8 private key format with a high iteration count.
6240 *) Fix determination of Perl interpreter: A perl or perl5
6241 _directory_ in $PATH was also accepted as the interpreter.
6242 [Ralf S. Engelschall]
6244 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6245 wrong with it but it was very old and did things like calling
6246 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6250 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6251 to use the new extension code.
6254 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6255 with macros. This should make it easier to change their form, add extra
6256 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6260 *) Add to configuration table a new entry that can specify an alternative
6261 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6262 according to Mark Crispin <MRC@Panda.COM>.
6266 *) DES CBC did not update the IV. Weird.
6269 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6270 Changing the behaviour of the former might break existing programs --
6271 where IV updating is needed, des_ncbc_encrypt can be used.
6274 *) When bntest is run from "make test" it drives bc to check its
6275 calculations, as well as internally checking them. If an internal check
6276 fails, it needs to cause bc to give a non-zero result or make test carries
6277 on without noticing the failure. Fixed.
6280 *) DES library cleanups.
6283 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6284 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6285 ciphers. NOTE: although the key derivation function has been verified
6286 against some published test vectors it has not been extensively tested
6287 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6291 *) Instead of "mkdir -p", which is not fully portable, use new
6292 Perl script "util/mkdir-p.pl".
6295 *) Rewrite the way password based encryption (PBE) is handled. It used to
6296 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6297 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6298 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6299 the 'parameter' field of the AlgorithmIdentifier is passed to the
6300 underlying key generation function so it must do its own ASN1 parsing.
6301 This has also changed the EVP_PBE_CipherInit() function which now has a
6302 'parameter' argument instead of literal salt and iteration count values
6303 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6306 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6307 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6308 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6309 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6310 value was just used as a "magic string" and not used directly its
6311 value doesn't matter.
6314 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6318 *) "linux-sparc64" configuration (ultrapenguin).
6319 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6320 "linux-sparc" configuration.
6321 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6323 *) config now generates no-xxx options for missing ciphers.
6326 *) Support the EBCDIC character set (work in progress).
6327 File ebcdic.c not yet included because it has a different license.
6328 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6330 *) Support BS2000/OSD-POSIX.
6331 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6333 *) Make callbacks for key generation use void * instead of char *.
6336 *) Make S/MIME samples compile (not yet tested).
6339 *) Additional typesafe stacks.
6342 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6346 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6348 *) New configuration variant "sco5-gcc".
6350 *) Updated some demos.
6351 [Sean O Riordain, Wade Scholine]
6353 *) Add missing BIO_free at exit of pkcs12 application.
6356 *) Fix memory leak in conf.c.
6359 *) Updates for Win32 to assembler version of MD5.
6362 *) Set #! path to perl in apps/der_chop to where we found it
6363 instead of using a fixed path.
6366 *) SHA library changes for irix64-mips4-cc.
6369 *) Improvements for VMS support.
6373 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6375 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6376 This also avoids the problems with SC4.2 and unpatched SC5.
6377 [Andy Polyakov <appro@fy.chalmers.se>]
6379 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6380 These are required because of the typesafe stack would otherwise break
6381 existing code. If old code used a structure member which used to be STACK
6382 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6383 sk_num or sk_value it would produce an error because the num, data members
6384 are not present in STACK_OF. Now it just produces a warning. sk_set
6385 replaces the old method of assigning a value to sk_value
6386 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6387 that does this will no longer work (and should use sk_set instead) but
6388 this could be regarded as a "questionable" behaviour anyway.
6391 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6392 correctly handle encrypted S/MIME data.
6395 *) Change type of various DES function arguments from des_cblock
6396 (which means, in function argument declarations, pointer to char)
6397 to des_cblock * (meaning pointer to array with 8 char elements),
6398 which allows the compiler to do more typechecking; it was like
6399 that back in SSLeay, but with lots of ugly casts.
6401 Introduce new type const_des_cblock.
6404 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6405 problems: find RecipientInfo structure that matches recipient certificate
6406 and initialise the ASN1 structures properly based on passed cipher.
6409 *) Belatedly make the BN tests actually check the results.
6412 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6413 to and from BNs: it was completely broken. New compilation option
6414 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6415 key elements as negative integers.
6418 *) Reorganize and speed up MD5.
6419 [Andy Polyakov <appro@fy.chalmers.se>]
6422 [Richard Levitte <richard@levitte.org>]
6424 *) New option -out to asn1parse to allow the parsed structure to be
6425 output to a file. This is most useful when combined with the -strparse
6426 option to examine the output of things like OCTET STRINGS.
6429 *) Make SSL library a little more fool-proof by not requiring any longer
6430 that SSL_set_{accept,connect}_state be called before
6431 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6432 in many applications because usually everything *appeared* to work as
6433 intended anyway -- now it really works as intended).
6436 *) Move openssl.cnf out of lib/.
6439 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6440 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6441 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6442 [Ralf S. Engelschall]
6444 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6445 handle PKCS#7 enveloped data properly.
6446 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6448 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6449 copying pointers. The cert_st handling is changed by this in
6450 various ways (and thus what used to be known as ctx->default_cert
6451 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6452 any longer when s->cert does not give us what we need).
6453 ssl_cert_instantiate becomes obsolete by this change.
6454 As soon as we've got the new code right (possibly it already is?),
6455 we have solved a couple of bugs of the earlier code where s->cert
6456 was used as if it could not have been shared with other SSL structures.
6458 Note that using the SSL API in certain dirty ways now will result
6459 in different behaviour than observed with earlier library versions:
6460 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6461 does not influence s as it used to.
6463 In order to clean up things more thoroughly, inside SSL_SESSION
6464 we don't use CERT any longer, but a new structure SESS_CERT
6465 that holds per-session data (if available); currently, this is
6466 the peer's certificate chain and, for clients, the server's certificate
6467 and temporary key. CERT holds only those values that can have
6468 meaningful defaults in an SSL_CTX.
6471 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6472 from the internal representation. Various PKCS#7 fixes: remove some
6473 evil casts and set the enc_dig_alg field properly based on the signing
6477 *) Allow PKCS#12 password to be set from the command line or the
6478 environment. Let 'ca' get its config file name from the environment
6479 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6483 *) Allow certificate policies extension to use an IA5STRING for the
6484 organization field. This is contrary to the PKIX definition but
6485 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6489 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6490 without disallowing inline assembler and the like for non-pedantic builds.
6493 *) Support Borland C++ builder.
6494 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6499 *) SHA-1 cleanups and performance enhancements.
6500 [Andy Polyakov <appro@fy.chalmers.se>]
6502 *) Sparc v8plus assembler for the bignum library.
6503 [Andy Polyakov <appro@fy.chalmers.se>]
6505 *) Accept any -xxx and +xxx compiler options in Configure.
6508 *) Update HPUX configuration.
6511 *) Add missing sk_<type>_unshift() function to safestack.h
6512 [Ralf S. Engelschall]
6514 *) New function SSL_CTX_use_certificate_chain_file that sets the
6515 "extra_cert"s in addition to the certificate. (This makes sense
6516 only for "PEM" format files, as chains as a whole are not
6520 *) Support verify_depth from the SSL API.
6521 x509_vfy.c had what can be considered an off-by-one-error:
6522 Its depth (which was not part of the external interface)
6523 was actually counting the number of certificates in a chain;
6524 now it really counts the depth.
6527 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6528 instead of X509err, which often resulted in confusing error
6529 messages since the error codes are not globally unique
6530 (e.g. an alleged error in ssl3_accept when a certificate
6531 didn't match the private key).
6533 *) New function SSL_CTX_set_session_id_context that allows to set a default
6534 value (so that you don't need SSL_set_session_id_context for each
6535 connection using the SSL_CTX).
6538 *) OAEP decoding bug fix.
6541 *) Support INSTALL_PREFIX for package builders, as proposed by
6545 *) New Configure options "threads" and "no-threads". For systems
6546 where the proper compiler options are known (currently Solaris
6547 and Linux), "threads" is the default.
6550 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6553 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6554 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6555 such as /usr/local/bin.
6558 *) "make linux-shared" to build shared libraries.
6559 [Niels Poppe <niels@netbox.org>]
6561 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6564 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6565 extension adding in x509 utility.
6568 *) Remove NOPROTO sections and error code comments.
6571 *) Partial rewrite of the DEF file generator to now parse the ANSI
6575 *) New Configure options --prefix=DIR and --openssldir=DIR.
6578 *) Complete rewrite of the error code script(s). It is all now handled
6579 by one script at the top level which handles error code gathering,
6580 header rewriting and C source file generation. It should be much better
6581 than the old method: it now uses a modified version of Ulf's parser to
6582 read the ANSI prototypes in all header files (thus the old K&R definitions
6583 aren't needed for error creation any more) and do a better job of
6584 translating function codes into names. The old 'ASN1 error code imbedded
6585 in a comment' is no longer necessary and it doesn't use .err files which
6586 have now been deleted. Also the error code call doesn't have to appear all
6587 on one line (which resulted in some large lines...).
6590 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6593 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6594 0 (which usually indicates a closed connection), but continue reading.
6597 *) Fix some race conditions.
6600 *) Add support for CRL distribution points extension. Add Certificate
6601 Policies and CRL distribution points documentation.
6604 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6607 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6608 8 of keying material. Merlin has also confirmed interop with this fix
6609 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6610 [Merlin Hughes <merlin@baltimore.ie>]
6612 *) Fix lots of warnings.
6613 [Richard Levitte <levitte@stacken.kth.se>]
6615 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6616 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6617 [Richard Levitte <levitte@stacken.kth.se>]
6619 *) Fix problems with sizeof(long) == 8.
6620 [Andy Polyakov <appro@fy.chalmers.se>]
6622 *) Change functions to ANSI C.
6625 *) Fix typos in error codes.
6626 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6628 *) Remove defunct assembler files from Configure.
6631 *) SPARC v8 assembler BIGNUM implementation.
6632 [Andy Polyakov <appro@fy.chalmers.se>]
6634 *) Support for Certificate Policies extension: both print and set.
6635 Various additions to support the r2i method this uses.
6638 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6639 return a const string when you are expecting an allocated buffer.
6642 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6643 types DirectoryString and DisplayText.
6646 *) Add code to allow r2i extensions to access the configuration database,
6647 add an LHASH database driver and add several ctx helper functions.
6650 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6651 fail when they extended the size of a BIGNUM.
6654 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6655 support typesafe stack.
6658 *) Fix typo in SSL_[gs]et_options().
6659 [Nils Frostberg <nils@medcom.se>]
6661 *) Delete various functions and files that belonged to the (now obsolete)
6662 old X509V3 handling code.
6665 *) New Configure option "rsaref".
6668 *) Don't auto-generate pem.h.
6671 *) Introduce type-safe ASN.1 SETs.
6674 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6675 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6677 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6678 that links with OpenSSL (well at least cause lots of warnings), but fear
6679 not: the conversion is trivial, and it eliminates loads of evil casts. A
6680 few STACKed things have been converted already. Feel free to convert more.
6681 In the fullness of time, I'll do away with the STACK type altogether.
6684 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6685 specified in <certfile> by updating the entry in the index.txt file.
6686 This way one no longer has to edit the index.txt file manually for
6687 revoking a certificate. The -revoke option does the gory details now.
6688 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6690 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6691 `-text' option at all and this way the `-noout -text' combination was
6692 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6693 [Ralf S. Engelschall]
6695 *) Make sure a corresponding plain text error message exists for the
6696 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6697 verify callback function determined that a certificate was revoked.
6698 [Ralf S. Engelschall]
6700 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6701 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6702 all available cipers including rc5, which was forgotten until now.
6703 In order to let the testing shell script know which algorithms
6704 are available, a new (up to now undocumented) command
6705 "openssl list-cipher-commands" is used.
6708 *) Bugfix: s_client occasionally would sleep in select() when
6709 it should have checked SSL_pending() first.
6712 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6713 the raw DSA values prior to ASN.1 encoding.
6716 *) Tweaks to Configure
6717 [Niels Poppe <niels@netbox.org>]
6719 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6723 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6726 *) New config option to avoid instructions that are illegal on the 80386.
6727 The default code is faster, but requires at least a 486.
6730 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6731 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6732 same as SSL2_VERSION anyway.
6735 *) New "-showcerts" option for s_client.
6738 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6739 application. Various cleanups and fixes.
6742 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6743 modify error routines to work internally. Add error codes and PBE init
6744 to library startup routines.
6747 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6748 packing functions to asn1 and evp. Changed function names and error
6749 codes along the way.
6752 *) PKCS12 integration: and so it begins... First of several patches to
6753 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6754 objects to objects.h
6757 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6758 and display support for Thawte strong extranet extension.
6761 *) Add LinuxPPC support.
6762 [Jeff Dubrule <igor@pobox.org>]
6764 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6765 bn_div_words in alpha.s.
6766 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6768 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6769 OAEP isn't supported when OpenSSL is built with RSAref.
6770 [Ulf Moeller <ulf@fitug.de>]
6772 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6773 so they no longer are missing under -DNOPROTO.
6774 [Soren S. Jorvang <soren@t.dk>]
6777 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6779 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6780 doesn't work when the session is reused. Coming soon!
6783 *) Fix a security hole, that allows sessions to be reused in the wrong
6784 context thus bypassing client cert protection! All software that uses
6785 client certs and session caches in multiple contexts NEEDS PATCHING to
6786 allow session reuse! A fuller solution is in the works.
6787 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6789 *) Some more source tree cleanups (removed obsolete files
6790 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6791 permission on "config" script to be executable) and a fix for the INSTALL
6793 [Ulf Moeller <ulf@fitug.de>]
6795 *) Remove some legacy and erroneous uses of malloc, free instead of
6797 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6799 *) Make rsa_oaep_test return non-zero on error.
6800 [Ulf Moeller <ulf@fitug.de>]
6802 *) Add support for native Solaris shared libraries. Configure
6803 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6804 if someone would make that last step automatic.
6805 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6807 *) ctx_size was not built with the right compiler during "make links". Fixed.
6810 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6811 except NULL ciphers". This means the default cipher list will no longer
6812 enable NULL ciphers. They need to be specifically enabled e.g. with
6813 the string "DEFAULT:eNULL".
6816 *) Fix to RSA private encryption routines: if p < q then it would
6817 occasionally produce an invalid result. This will only happen with
6818 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6821 *) Be less restrictive and allow also `perl util/perlpath.pl
6822 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6823 because this way one can also use an interpreter named `perl5' (which is
6824 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6825 installed as `perl').
6826 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6828 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6829 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6831 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6832 advapi32.lib to Win32 build and change the pem test comparision
6833 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6834 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6835 and crypto/des/ede_cbcm_enc.c.
6838 *) DES quad checksum was broken on big-endian architectures. Fixed.
6841 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6842 Win32 test batch file so it (might) work again. The Win32 test batch file
6843 is horrible: I feel ill....
6846 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6847 in e_os.h. Audit of header files to check ANSI and non ANSI
6848 sections: 10 functions were absent from non ANSI section and not exported
6849 from Windows DLLs. Fixed up libeay.num for new functions.
6852 *) Make `openssl version' output lines consistent.
6853 [Ralf S. Engelschall]
6855 *) Fix Win32 symbol export lists for BIO functions: Added
6856 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6857 to ms/libeay{16,32}.def.
6858 [Ralf S. Engelschall]
6860 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6861 fine under Unix and passes some trivial tests I've now added. But the
6862 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6863 added to make sure no one expects that this stuff really works in the
6864 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6865 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6867 [Ralf S. Engelschall]
6869 *) Fix the generation of two part addresses in perl.
6870 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6872 *) Add config entry for Linux on MIPS.
6873 [John Tobey <jtobey@channel1.com>]
6875 *) Make links whenever Configure is run, unless we are on Windoze.
6878 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6879 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6883 *) Add a useful kludge to allow package maintainers to specify compiler and
6884 other platforms details on the command line without having to patch the
6885 Configure script everytime: One now can use ``perl Configure
6886 <id>:<details>'', i.e. platform ids are allowed to have details appended
6887 to them (seperated by colons). This is treated as there would be a static
6888 pre-configured entry in Configure's %table under key <id> with value
6889 <details> and ``perl Configure <id>'' is called. So, when you want to
6890 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6891 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6892 now, which overrides the FreeBSD-elf entry on-the-fly.
6893 [Ralf S. Engelschall]
6895 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6898 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6899 on the `perl Configure ...' command line. This way one can compile
6900 OpenSSL libraries with Position Independent Code (PIC) which is needed
6901 for linking it into DSOs.
6902 [Ralf S. Engelschall]
6904 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6908 *) Cleaned up the LICENSE document: The official contact for any license
6909 questions now is the OpenSSL core team under openssl-core@openssl.org.
6910 And add a paragraph about the dual-license situation to make sure people
6911 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6912 to the OpenSSL toolkit.
6913 [Ralf S. Engelschall]
6915 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6916 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6917 Additonally cleaned up the `make links' target: Remove unnecessary
6918 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6919 to speed processing and no longer clutter the display with confusing
6920 stuff. Instead only the actually done links are displayed.
6921 [Ralf S. Engelschall]
6923 *) Permit null encryption ciphersuites, used for authentication only. It used
6924 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6925 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6929 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6930 signed attributes when verifying signatures (this would break them),
6931 the detached data encoding was wrong and public keys obtained using
6932 X509_get_pubkey() weren't freed.
6935 *) Add text documentation for the BUFFER functions. Also added a work around
6936 to a Win95 console bug. This was triggered by the password read stuff: the
6937 last character typed gets carried over to the next fread(). If you were
6938 generating a new cert request using 'req' for example then the last
6939 character of the passphrase would be CR which would then enter the first
6943 *) Added the new `Includes OpenSSL Cryptography Software' button as
6944 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6945 button and can be used by applications based on OpenSSL to show the
6946 relationship to the OpenSSL project.
6947 [Ralf S. Engelschall]
6949 *) Remove confusing variables in function signatures in files
6950 ssl/ssl_lib.c and ssl/ssl.h.
6951 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6953 *) Don't install bss_file.c under PREFIX/include/
6954 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6956 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6957 functions that return function pointers and has support for NT specific
6958 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6959 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6960 unsigned to signed types: this was killing the Win32 compile.
6963 *) Add new certificate file to stack functions,
6964 SSL_add_dir_cert_subjects_to_stack() and
6965 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6966 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6967 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6968 This means that Apache-SSL and similar packages don't have to mess around
6969 to add as many CAs as they want to the preferred list.
6972 *) Experiment with doxygen documentation. Currently only partially applied to
6974 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6975 openssl.doxy as the configuration file.
6978 *) Get rid of remaining C++-style comments which strict C compilers hate.
6979 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6981 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6982 compiled in by default: it has problems with large keys.
6985 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6986 DH private keys and/or callback functions which directly correspond to
6987 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6988 is needed for applications which have to configure certificates on a
6989 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6991 For the RSA certificate situation is makes no difference, but
6992 for the DSA certificate situation this fixes the "no shared cipher"
6993 problem where the OpenSSL cipher selection procedure failed because the
6994 temporary keys were not overtaken from the context and the API provided
6995 no way to reconfigure them.
6996 The new functions now let applications reconfigure the stuff and they
6997 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6998 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6999 non-public-API function ssl_cert_instantiate() is used as a helper
7000 function and also to reduce code redundancy inside ssl_rsa.c.
7001 [Ralf S. Engelschall]
7003 *) Move s_server -dcert and -dkey options out of the undocumented feature
7004 area because they are useful for the DSA situation and should be
7005 recognized by the users.
7006 [Ralf S. Engelschall]
7008 *) Fix the cipher decision scheme for export ciphers: the export bits are
7009 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7010 SSL_EXP_MASK. So, the original variable has to be used instead of the
7011 already masked variable.
7012 [Richard Levitte <levitte@stacken.kth.se>]
7014 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7015 [Richard Levitte <levitte@stacken.kth.se>]
7017 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7018 from `int' to `unsigned int' because it's a length and initialized by
7019 EVP_DigestFinal() which expects an `unsigned int *'.
7020 [Richard Levitte <levitte@stacken.kth.se>]
7022 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7023 script. Instead use the usual Shell->Perl transition trick.
7024 [Ralf S. Engelschall]
7026 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7027 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7028 -noout -modulus' as it's already the case for `openssl rsa -noout
7029 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7030 currently the public key is printed (a decision which was already done by
7031 `openssl dsa -modulus' in the past) which serves a similar purpose.
7032 Additionally the NO_RSA no longer completely removes the whole -modulus
7033 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7035 [Ralf S. Engelschall]
7037 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7038 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7039 [Arne Ansper <arne@ats.cyber.ee>]
7041 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7042 to be added. Now both 'req' and 'ca' can use new objects defined in the
7046 *) Add cool BIO that does syslog (or event log on NT).
7047 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7049 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7050 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7051 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7052 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7055 *) Add preliminary config info for new extension code.
7058 *) Make RSA_NO_PADDING really use no padding.
7059 [Ulf Moeller <ulf@fitug.de>]
7061 *) Generate errors when private/public key check is done.
7064 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7065 for some CRL extensions and new objects added.
7068 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7069 key usage extension and fuller support for authority key id.
7072 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7073 padding method for RSA, which is recommended for new applications in PKCS
7074 #1 v2.0 (RFC 2437, October 1998).
7075 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7076 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7077 against Bleichbacher's attack on RSA.
7078 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7081 *) Updates to the new SSL compression code
7082 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7084 *) Fix so that the version number in the master secret, when passed
7085 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7086 (because the server will not accept higher), that the version number
7087 is 0x03,0x01, not 0x03,0x00
7088 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7090 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7091 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7092 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7095 *) Support for RAW extensions where an arbitrary extension can be
7096 created by including its DER encoding. See apps/openssl.cnf for
7100 *) Make sure latest Perl versions don't interpret some generated C array
7101 code as Perl array code in the crypto/err/err_genc.pl script.
7102 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7104 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7105 not many people have the assembler. Various Win32 compilation fixes and
7106 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7110 *) Modify configure script 'Configure' to automatically create crypto/date.h
7111 file under Win32 and also build pem.h from pem.org. New script
7112 util/mkfiles.pl to create the MINFO file on environments that can't do a
7113 'make files': perl util/mkfiles.pl >MINFO should work.
7116 *) Major rework of DES function declarations, in the pursuit of correctness
7117 and purity. As a result, many evil casts evaporated, and some weirdness,
7118 too. You may find this causes warnings in your code. Zapping your evil
7119 casts will probably fix them. Mostly.
7122 *) Fix for a typo in asn1.h. Bug fix to object creation script
7123 obj_dat.pl. It considered a zero in an object definition to mean
7124 "end of object": none of the objects in objects.h have any zeros
7125 so it wasn't spotted.
7126 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7128 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7129 Masking (CBCM). In the absence of test vectors, the best I have been able
7130 to do is check that the decrypt undoes the encrypt, so far. Send me test
7131 vectors if you have them.
7134 *) Correct calculation of key length for export ciphers (too much space was
7135 allocated for null ciphers). This has not been tested!
7138 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7139 message is now correct (it understands "crypto" and "ssl" on its
7140 command line). There is also now an "update" option. This will update
7141 the util/ssleay.num and util/libeay.num files with any new functions.
7143 perl util/mkdef.pl crypto ssl update
7144 it will update them.
7147 *) Overhauled the Perl interface (perl/*):
7148 - ported BN stuff to OpenSSL's different BN library
7149 - made the perl/ source tree CVS-aware
7150 - renamed the package from SSLeay to OpenSSL (the files still contain
7151 their history because I've copied them in the repository)
7152 - removed obsolete files (the test scripts will be replaced
7153 by better Test::Harness variants in the future)
7154 [Ralf S. Engelschall]
7156 *) First cut for a very conservative source tree cleanup:
7157 1. merge various obsolete readme texts into doc/ssleay.txt
7158 where we collect the old documents and readme texts.
7159 2. remove the first part of files where I'm already sure that we no
7160 longer need them because of three reasons: either they are just temporary
7161 files which were left by Eric or they are preserved original files where
7162 I've verified that the diff is also available in the CVS via "cvs diff
7163 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7164 the crypto/md/ stuff).
7165 [Ralf S. Engelschall]
7167 *) More extension code. Incomplete support for subject and issuer alt
7168 name, issuer and authority key id. Change the i2v function parameters
7169 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7170 what that's for :-) Fix to ASN1 macro which messed up
7171 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7174 *) Preliminary support for ENUMERATED type. This is largely copied from the
7178 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7179 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7181 *) Make sure `make rehash' target really finds the `openssl' program.
7182 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7184 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7185 like to hear about it if this slows down other processors.
7188 *) Add CygWin32 platform information to Configure script.
7189 [Alan Batie <batie@aahz.jf.intel.com>]
7191 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7192 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7194 *) New program nseq to manipulate netscape certificate sequences
7197 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7201 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7202 but the BN code had some problems that would cause failures when
7203 doing certificate verification and some other functions.
7204 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7206 *) Add ASN1 and PEM code to support netscape certificate sequences.
7209 *) Add ASN1 and PEM code to support netscape certificate sequences.
7212 *) Add several PKIX and private extended key usage OIDs.
7215 *) Modify the 'ca' program to handle the new extension code. Modify
7216 openssl.cnf for new extension format, add comments.
7219 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7220 and add a sample to openssl.cnf so req -x509 now adds appropriate
7224 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7225 error code, add initial support to X509_print() and x509 application.
7228 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7229 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7230 stuff is currently isolated and isn't even compiled yet.
7233 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7234 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7235 Removed the versions check from X509 routines when loading extensions:
7236 this allows certain broken certificates that don't set the version
7237 properly to be processed.
7240 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7241 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7242 can still be regenerated with "make depend".
7245 *) Spelling mistake in C version of CAST-128.
7246 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7248 *) Changes to the error generation code. The perl script err-code.pl
7249 now reads in the old error codes and retains the old numbers, only
7250 adding new ones if necessary. It also only changes the .err files if new
7251 codes are added. The makefiles have been modified to only insert errors
7252 when needed (to avoid needlessly modifying header files). This is done
7253 by only inserting errors if the .err file is newer than the auto generated
7254 C file. To rebuild all the error codes from scratch (the old behaviour)
7255 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7256 or delete all the .err files.
7259 *) CAST-128 was incorrectly implemented for short keys. The C version has
7260 been fixed, but is untested. The assembler versions are also fixed, but
7261 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7262 to regenerate it if needed.
7263 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7264 Hagino <itojun@kame.net>]
7266 *) File was opened incorrectly in randfile.c.
7267 [Ulf Möller <ulf@fitug.de>]
7269 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7270 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7271 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7272 al: it's just almost always a UTCTime. Note this patch adds new error
7273 codes so do a "make errors" if there are problems.
7276 *) Correct Linux 1 recognition in config.
7277 [Ulf Möller <ulf@fitug.de>]
7279 *) Remove pointless MD5 hash when using DSA keys in ca.
7280 [Anonymous <nobody@replay.com>]
7282 *) Generate an error if given an empty string as a cert directory. Also
7283 generate an error if handed NULL (previously returned 0 to indicate an
7284 error, but didn't set one).
7285 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7287 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7290 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7291 parameters. This was causing a warning which killed off the Win32 compile.
7294 *) Remove C++ style comments from crypto/bn/bn_local.h.
7295 [Neil Costigan <neil.costigan@celocom.com>]
7297 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7298 based on a text string, looking up short and long names and finally
7299 "dot" format. The "dot" format stuff didn't work. Added new function
7300 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7301 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7302 OID is not part of the table.
7305 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7306 X509_LOOKUP_by_alias().
7309 *) Sort openssl functions by name.
7312 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7313 encryption from sample DSA keys (in case anyone is interested the password
7317 *) Make _all_ *_free functions accept a NULL pointer.
7318 [Frans Heymans <fheymans@isaserver.be>]
7320 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7322 [Anonymous <nobody@replay.com>]
7324 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7325 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7327 *) Don't blow it for numeric -newkey arguments to apps/req.
7328 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7330 *) Temp key "for export" tests were wrong in s3_srvr.c.
7331 [Anonymous <nobody@replay.com>]
7333 *) Add prototype for temp key callback functions
7334 SSL_CTX_set_tmp_{rsa,dh}_callback().
7337 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7338 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7341 *) X509_name_add_entry() freed the wrong thing after an error.
7342 [Arne Ansper <arne@ats.cyber.ee>]
7344 *) rsa_eay.c would attempt to free a NULL context.
7345 [Arne Ansper <arne@ats.cyber.ee>]
7347 *) BIO_s_socket() had a broken should_retry() on Windoze.
7348 [Arne Ansper <arne@ats.cyber.ee>]
7350 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7351 [Arne Ansper <arne@ats.cyber.ee>]
7353 *) Make sure the already existing X509_STORE->depth variable is initialized
7354 in X509_STORE_new(), but document the fact that this variable is still
7355 unused in the certificate verification process.
7356 [Ralf S. Engelschall]
7358 *) Fix the various library and apps files to free up pkeys obtained from
7359 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7362 *) Fix reference counting in X509_PUBKEY_get(). This makes
7363 demos/maurice/example2.c work, amongst others, probably.
7364 [Steve Henson and Ben Laurie]
7366 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7367 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7368 are no longer created. This way we have a single and consistent command
7369 line interface `openssl <command>', similar to `cvs <command>'.
7370 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7372 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7373 BIT STRING wrapper always have zero unused bits.
7376 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7379 *) Make the top-level INSTALL documentation easier to understand.
7382 *) Makefiles updated to exit if an error occurs in a sub-directory
7383 make (including if user presses ^C) [Paul Sutton]
7385 *) Make Montgomery context stuff explicit in RSA data structure.
7388 *) Fix build order of pem and err to allow for generated pem.h.
7391 *) Fix renumbering bug in X509_NAME_delete_entry().
7394 *) Enhanced the err-ins.pl script so it makes the error library number
7395 global and can add a library name. This is needed for external ASN1 and
7396 other error libraries.
7399 *) Fixed sk_insert which never worked properly.
7402 *) Fix ASN1 macros so they can handle indefinite length construted
7403 EXPLICIT tags. Some non standard certificates use these: they can now
7407 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7408 into a single doc/ssleay.txt bundle. This way the information is still
7409 preserved but no longer messes up this directory. Now it's new room for
7410 the new set of documenation files.
7411 [Ralf S. Engelschall]
7413 *) SETs were incorrectly DER encoded. This was a major pain, because they
7414 shared code with SEQUENCEs, which aren't coded the same. This means that
7415 almost everything to do with SETs or SEQUENCEs has either changed name or
7416 number of arguments.
7417 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7419 *) Fix test data to work with the above.
7422 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7423 was already fixed by Eric for 0.9.1 it seems.
7424 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7426 *) Autodetect FreeBSD3.
7429 *) Fix various bugs in Configure. This affects the following platforms:
7433 unixware-2.0-pentium
7437 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7438 before they are needed.
7441 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7445 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7447 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7448 changed SSLeay to OpenSSL in version strings.
7449 [Ralf S. Engelschall]
7451 *) Some fixups to the top-level documents.
7454 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7455 because the symlink to include/ was missing.
7456 [Ralf S. Engelschall]
7458 *) Incorporated the popular no-RSA/DSA-only patches
7459 which allow to compile a RSA-free SSLeay.
7460 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7462 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7463 when "ssleay" is still not found.
7464 [Ralf S. Engelschall]
7466 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7467 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7469 *) Updated the README file.
7470 [Ralf S. Engelschall]
7472 *) Added various .cvsignore files in the CVS repository subdirs
7473 to make a "cvs update" really silent.
7474 [Ralf S. Engelschall]
7476 *) Recompiled the error-definition header files and added
7477 missing symbols to the Win32 linker tables.
7478 [Ralf S. Engelschall]
7480 *) Cleaned up the top-level documents;
7481 o new files: CHANGES and LICENSE
7482 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7483 o merged COPYRIGHT into LICENSE
7484 o removed obsolete TODO file
7485 o renamed MICROSOFT to INSTALL.W32
7486 [Ralf S. Engelschall]
7488 *) Removed dummy files from the 0.9.1b source tree:
7489 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7490 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7491 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7492 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7493 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7494 [Ralf S. Engelschall]
7496 *) Added various platform portability fixes.
7499 *) The Genesis of the OpenSSL rpject:
7500 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7501 Young and Tim J. Hudson created while they were working for C2Net until
7503 [The OpenSSL Project]
7506 Changes between 0.9.0b and 0.9.1b [not released]
7508 *) Updated a few CA certificates under certs/
7511 *) Changed some BIGNUM api stuff.
7514 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7515 DGUX x86, Linux Alpha, etc.
7518 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7519 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7523 *) Add -strparse option to asn1pars program which parses nested
7525 [Dr Stephen Henson <shenson@bigfoot.com>]
7527 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7530 *) DSA fix for "ca" program.
7533 *) Added "-genkey" option to "dsaparam" program.
7536 *) Added RIPE MD160 (rmd160) message digest.
7539 *) Added -a (all) option to "ssleay version" command.
7542 *) Added PLATFORM define which is the id given to Configure.
7545 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7548 *) Extended the ASN.1 parser routines.
7551 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7554 *) Added a BN_CTX to the BN library.
7557 *) Fixed the weak key values in DES library
7560 *) Changed API in EVP library for cipher aliases.
7563 *) Added support for RC2/64bit cipher.
7566 *) Converted the lhash library to the crypto/mem.c functions.
7569 *) Added more recognized ASN.1 object ids.
7572 *) Added more RSA padding checks for SSL/TLS.
7575 *) Added BIO proxy/filter functionality.
7578 *) Added extra_certs to SSL_CTX which can be used
7579 send extra CA certificates to the client in the CA cert chain sending
7580 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7583 *) Now Fortezza is denied in the authentication phase because
7584 this is key exchange mechanism is not supported by SSLeay at all.
7587 *) Additional PKCS1 checks.
7590 *) Support the string "TLSv1" for all TLS v1 ciphers.
7593 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7594 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7597 *) Fixed a few memory leaks.
7600 *) Fixed various code and comment typos.
7603 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7604 bytes sent in the client random.
7605 [Edward Bishop <ebishop@spyglass.com>]